A method for controlling power units.

专利摘要:
The invention relates to a control method for optimizing the operation of a power plant with power plant blocks during a selected operating period, which is subdivided to have regular intervals within which each of the power plant blocks has either an on state or an off state. The control method includes: determining a preferred case for each of the competing modes of operation for each of the intervals; based on the preferred cases, selecting proposed part load operating sequences for the selected operating period; Determining a shutdown operation for each of the power plant blocks having the off state for one or more intervals during the selected operating period and calculating therefrom an economic shutdown result; Determining a partial load operation for each of the power plant blocks having the on state for one or more intervals during the selected operating period and calculating therefrom an economic partial load result; Calculating an economic sequence result for each of the proposed partial load operating sequences; and comparing the economic sequence results.
公开号:CH710433A2
申请号:CH01710/15
申请日:2015-11-23
公开日:2016-06-15
发明作者:Anne Wichmann Lisa；Kumar Pandey Achalesh；Michael Raczynski Christopher
申请人:Gen Electric；
IPC主号:

专利说明:

CROSS-REFERENCE TO RELATED APPLICATIONS
This application claims priority over US Provisional Patent Application No. 61 / 922,555 entitled "TURBINE ENGINE AND PLANT FLEXIBILITY AND ECONOMIC OPTIMIZATION SYSTEMS AND ASSOCIATED PROCESSES" filed Dec. 31, 2013, the provisional application being incorporated by reference in its Entity is included herein; This application claims the benefit of the filing date of the provisional application according to 35 U.S.C.119 (e).
BACKGROUND OF THE INVENTION
The invention of the present application generally relates to power generation, and more particularly to methods and systems related to optimizing and / or enhancing the economics and performance of power plant power plants.
In power systems generate a number of participants or power plants electricity, which is then distributed over common transmission lines to residential and commercial customers. As will be understood, this is still mostly done by thermal power plants such as e.g. Gas turbines, steam turbines and combined cycle power plants, which generate a significant proportion of the power that such systems require. Each of the power plants within such systems has one or more power generation units, and each of these units typically includes a control system that controls operation, and in the case of more than one block power plants, the power of the power plant as a whole. As an example, one of the responsibilities of the plant operator is to create a supply curve that represents the cost of electricity production. A supply curve typically includes an incremental variable cost curve, an average variable cost curve, or other appropriate variable power generation cost that is typically expressed in dollars per megawatt hour compared to megawatts. It will be understood that an average variable cost curve may represent cumulative costs divided by a cumulative current output for a given point and an incremental variable cost curve may represent a change in cost divided by a change in current output. For example, an incremental variable cost curve may be obtained by taking a first derivative of an input-output curve of the power plant that represents the cost per hour compared to the generated current. In a combined cycle power plant where waste heat from a fuel burning generator is used to generate steam for driving a supplementary steam turbine, an incremental variable cost curve can also be obtained by known techniques, but their derivative can be more complex.
In most power systems, a competitive process, commonly referred to as economic load sharing, is used to divide the system load among the power plants over a future period of time. As part of this process, power plants periodically generate supply curves and send the supply curves to a power system authority or dispatcher. Such supply curves represent bids from the power plants to generate a share of the electricity required by the power system over a future market period. The load distribution authority receives the supply curves from the power plants within their system and evaluates them to determine the level at which each power plant will be integrated to most efficiently meet the predicted load requirements of the system. To do this, the load-distribution authority analyzes the supply curves and, with the aim of finding the lowest production costs for the system, creates a commitment timetable that describes the extent to which each of the power plants will be involved over the relevant period.
After the commitment schedule has been transmitted to the power plants, each power plant can determine the most efficient and cost effective way by which it complies with its load obligation. It will be understood that the plant blocks of the power plant have control systems that monitor and control the operation. When the power plant blocks include heat generators, such control systems are responsible for the combustion systems and other aspects of operation. (Illustratively, both a gas turbine and a combined cycle power plant will be described herein, but it will be understood that certain embodiments of the present invention may be applied to or used in conjunction with other types of power generation units.) The control system may implement scheduling algorithms. which adjust the fuel flow, swirl chokes and other control inputs to ensure efficient operation of the engine. However, the actual performance and efficiency of a power plant are affected by external factors, such as variable environmental conditions that can not be fully foreseen. As will be appreciated, the complexity of such systems and the variability of operating conditions make it difficult to predict and control performance, often resulting in inefficient operation.
Machine wear that occurs over time is another difficult factor to quantify, which can have a significant impact on power plant block performance. It will be understood that wear rate, replacement of worn out components, scheduling of maintenance routines, and other factors affect the short-term performance of the plant, and thus in generating the cost curves during the load distribution process, as well as evaluating the long-term value for money Plant must be considered. As an example, gas turbine life typically has limits that are expressed in terms of both operating hours and the number of plant starts. If a gas turbine or a component thereof reaches its limit for plant starts before the hour limit, it must be repaired or replaced, even if there is still an hourly life left. The hourly lifetime of a gas turbine can be extended by reducing the firing temperature, but this reduces the efficiency of the gas turbine, which in turn increases operating costs. Conversely, increasing the firing temperature increases efficiency, but shortens gas turbine life and increases maintenance and / or replacement costs. As will be appreciated, the life cycle costs of a heat engine are dependent upon many complex factors, while also representing a significant consideration for the economic efficiency of the power plant.
In view of the complexity of modern power plants, especially those with multiple power plant blocks, and the market in which they compete, power plant operators are constantly struggling to maximize economic profitability. For example, grid compatibility and load distribution planning for a power plant is controlled by the control of thermal power plants in an excessively static manner, i. using static control profiles, e.g. Curves of specific heat consumption derived only from periodic performance tests, adversely affected. Between these periodic updates, the performance of the turbine engine may change (for example, due to wear), which in turn may affect startup and load performance. In addition, changes in the external factors during the day, without taking them into account in the turbine control profiles, can lead to inefficient operation. To compensate for this type of variability, power plant operators often become overly conservative in planning future operations, resulting in under-utilized power plant units. Other times, plant operators are forced to operate units inefficiently to meet high commitments.
Without identification of short-term inefficiencies and / or long-term degradation, as recognized, conventional power plant control systems either need to be retuned frequently, which is an expensive process, or they must be operated conservatively so as to take preventive account of component wear. The alternative would be to risk a breach of operating limits, resulting in excessive attenuation or failure. Similarly, conventional power plant control systems lack the ability to account for changing conditions in the most cost effective manner. As will be understood, this results in power plant utilization, which is often far from optimum. Thus, there is a need for improved methods and systems for monitoring, modeling, and controlling power plant operation, particularly those that provide a more complete understanding of the myriad operating modes available to the operators of complex, modern power plants and the economic tradeoffs associated with each of them.
BRIEF SUMMARY OF THE INVENTION
The present application thus describes a control method for optimizing the operation of a power plant with power plant blocks during a selected operating period. The selected operating period may be subdivided to include regular intervals within which each of the power plant blocks has either an on state or an off state. The unique combinations that the power plant blocks have on-state and off-state define competing modes of operation within the intervals. The control method may include the steps of: determining a preferred case for each of the contending modes of operation for each of the intervals; based on the data relating to the preferred cases, selecting proposed part-load operating sequences for the selected operating period, each of the proposed part-load operating sequences describing a unique progression of the off-state and on-state for the power plant blocks over the intervals of the selected operating period ; for each of the proposed partial load operating sequences, determining a shutdown operation for each of the power plant blocks having the off state for one or more intervals during the selected operating period, and calculating an economic shutdown result; for each of the proposed partial load operating sequences, determining a partial load operation for each of the power plant blocks having the on-state for one or more intervals during the selected operating period and calculating therefrom an economic partial load result; in view of the economic shutdown and partial load result, calculating an economic sequence result for each of the proposed partial load operating sequences; and comparing the economic sequence results and based on outputting a preferred part-load operating sequence.
These and other features of the present application will become more apparent upon a review of the following detailed description of the preferred embodiments, taken in conjunction with the drawings and the appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011]<Tb> FIG. 1 <SEP> is a schematic diagram of a power system according to aspects of the present invention;<Tb> FIG. 2 <SEP> illustrates a schematic representation of an exemplary thermal power plant as may be used within power plants in accordance with embodiments of the present invention;<Tb> FIG. 3 <SEP> is a schematic representation of an exemplary multi-gas turbine power plant in accordance with embodiments of the present invention;<Tb> FIG. 4 <SEP> illustrates an example system configuration of a plant controller and optimizer in accordance with aspects of the present invention;<Tb> FIG. 5 <SEP> illustrates a schematic of a power plant having a plant controller and optimizer with a system configuration in accordance with certain aspects of the present invention;<Tb> FIG. FIG. 6 shows a computer system with an exemplary user interface according to certain aspects of the present invention; FIG.<Tb> FIG. 7 <SEP> is an exemplary incremental specific heat consumption curve and an effect that may have an error on the economic load distribution process;<Tb> FIG. 8 <SEP> is a schematic illustration of an exemplary plant controller having a power system in accordance with aspects of the present invention;<Tb> FIG. 9 <SEP> illustrates a flowchart of a power plant control method in accordance with aspects of the present invention;<Tb> FIG. 10 illustrates a data flow diagram describing an asset optimization system architecture for a combined cycle power plant in accordance with aspects of the present invention;<Tb> FIG. Figure 11 provides a simplified block diagram of a computer system as may be used with a real-time optimization system in accordance with aspects of the present invention;<Tb> FIG. FIG. 12 is a flowchart of an exemplary method for solving parameterized simultaneous equations and constraints in accordance with the present invention; FIG.<Tb> FIG. Fig. 13 shows a simplified configuration of a computer system according to the control methodology of embodiments of the present invention;<Tb> FIG. 14 <SEP> illustrates an alternative configuration of a computer system in accordance with the control methodology of embodiments of the present invention;<Tb> FIG. 15 <SEP> is a flowchart of an example control methodology according to exemplary aspects of the present invention;<Tb> FIG. 16 <SEP> is a flowchart of an alternative control methodology according to exemplary aspects of the present invention;<Tb> FIG. 17 <SEP> is a flowchart of an alternative control methodology according to example aspects of the present invention;<Tb> FIG. Fig. 18 illustrates a flowchart in which an alternative embodiment of the present invention is provided concerning the optimization of the partial load operation;<Tb> FIG. 19 <SEP> illustrates a flowchart in which an alternative embodiment of the present invention is provided relating to optimization between part load operation and shut down operation;<Tb> FIG. 20 <SEP> is a diagram illustrating the available operating modes of a gas turbine during a selected operating period at defined intervals in accordance with aspects of an exemplary embodiment of the present invention;<Tb> FIG. 21 <SEP> is a diagram illustrating the available operating modes of a gas turbine during a selected operating period at defined intervals in accordance with aspects of an alternative embodiment of the present invention;<Tb> FIG. 22 <SEP> illustrates a flowchart according to a power plant inventory optimization process according to an alternative embodiment of the present invention;<Tb> FIG. 23 <SEP> illustrates a schematic representation of a power plant inventory optimization system in accordance with aspects of the present invention;<Tb> FIG. 24 <SEP> illustrates a schematic representation of a power plant inventory optimization system in accordance with alternative aspects of the present invention;<Tb> FIG. 25 <SEP> illustrates a schematic representation of a power plant inventory optimization system in accordance with alternative aspects of the present invention;<Tb> FIG. 26 <SEP> illustrates a schematic of a power plant block optimization system that includes a block controller;<Tb> FIG. Figure 27 illustrates a schematic representation of an alternative power plant block optimization system including a block controller;<Tb> FIG. 28 <SEP> is a flowchart illustrating one embodiment of a process for optimizing the shutdown of a combined cycle power plant; and<Tb> FIG. FIG. 29 illustrates an exemplary control system in which a model-less adaptive controller according to aspects of the present invention is used.
DETAILED DESCRIPTION OF THE INVENTION
Example embodiments of the invention will now be described in more detail with reference to the accompanying drawings, in which some, but not all embodiments are shown. In fact, the invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided for this disclosure to meet applicable regulatory requirements. Same numbers can refer to the same elements throughout the document.
In accordance with aspects of the present invention, systems and methods are disclosed which may be used to optimize the performance of power systems, power plants, and / or thermal power generation units. In exemplary embodiments, this optimization involves economic optimization in which the operator of a power plant decides between alternative modes of operation to increase profitability. The embodiments may be employed within a particular power system to secure a competitive advantage through advantageous economic commitment conditions during the load-sharing process. A consultancy function may allow operators to choose between operating modes based on accurate economic comparisons and forecasts. As another feature, the process of anticipating fuel purchase for future generation periods may be improved so that fuel inventory is minimized while the risk of shortage does not increase. Other configurations of the present invention, as described below, provide computer-implemented methods and apparatus for modeling power systems and power plants having multiple thermal power plants. The technical effects of some configurations of the present invention include the creation and resolution of power system models that predict performance under varying physical, operational, and / or economic conditions. Exemplary embodiments of the present invention combine a power plant model that predicts performance under varying environmental and operating conditions with an economic model that includes economic constraints, goals, and market conditions to optimize profitability. Thereby, the optimization system of the present invention can predict optimized setpoints that maximize profitability for certain combinations of environmental, operating, contract, regulatory, legal and / or economic and market conditions.
Fig. 1 illustrates a schematic representation of a power system 10 incorporating aspects of the present invention, as well as an exemplary environment in which embodiments may be operated. The power system 10 may include power generators or power plants 12, such as the illustrated wind and thermal power plants. It will be understood that thermal power plants may comprise power plant blocks, such as e.g. Gas turbines, coal-fired steam turbines and / or CCGT plants. In addition, the power system 10 may include other types of power plants (not shown), such as those shown in FIG. Solar power installations, hydroelectric, geothermal, nuclear and / or any other suitable power sources currently known or hereafter discovered. Overland lines 14 may connect various power plants 12 to customers or consumers 16 of the power system 10. It should be understood that power lines 14 may represent a network or distribution network for the power system and may include multiple sections and / or substations as desired or required. The power generated by the power plants 12 can be supplied via power lines 14 to the consumers 16, which may include, for example, municipalities, residential or commercial customers. The power system 10 may also include memory devices 18 connected to the transmission lines 14 for storing energy during periods of excess generation.
The power system 10 also includes control systems or controllers 22, 23, 25 that manage or control the operation of a plurality of the components contained therein. For example, a plant controller 22 may control the operation of each of the power plants 12. Load controllers 23 may control the operation of the various loads 16 that are part of the power system 10. For example, a load controller 23 may manage the nature or timing of a customer's power purchase. A load sharing authority 24 may manage certain aspects of the operation of the power system 10, and may include a power system controller 25 that controls the economic load sharing process by which load commitments are distributed among participating power plants. The controllers 22, 23, 25, which are represented by rectangular boxes, can be connected via communication lines or connections 21 to a communication network 20, via which data is exchanged. The connections 21 may be wired or wireless. It will be understood that the communication network 20 may be connected to or part of a larger communication system or network, such as e.g. the Internet or a private computer network. In addition, the controllers 22, 23, 25 may receive information, data, and commands from data libraries and resources, which may be referred to herein generically as "data resources 26," and / or send information, data, and commands thereto Communication network 20, or alternatively they may locally store or house one or more of such data repositories. The data resources 26 may include several types of data, including, but not limited to, market data, operational data, and environmental data. The market data includes information on market conditions, such as market conditions. Energy selling price, fuel costs, labor costs, regulations, etc. The operating data includes information related to the operating conditions of the power plant or its power plant blocks, such as power plant units. Temperature or pressure measurements within the power plant, air flow rate, fuel flow rate, etc. The environmental data includes information related to environmental conditions at the facility, e.g. Ambient air temperature, humidity and / or pressure. The market, operational and environmental data may each include historical records, current condition data and / or data relating to forecasts. For example, the data resources 26 may include current and predicted meteorological / climate information, current and predicted market conditions, utility and performance history records of the operation of the power plant, and / or measured parameters regarding the operation of other power plants with similar components and / or configurations, and other data that may be required and / or desired include. For example, during operation, the power distribution system controller 25 of the load distribution authority 24 may receive or issue commands to the other controllers 22, 23 within the power system 10. Each of the plant and load controllers then controls the system component for which it is responsible, forwards information about them to the power system controller 25 and receives commands from it.
FIG. 2 is a schematic illustration of an exemplary thermal power plant, a gas turbine system 30, that may be used within a power plant according to the present invention. As illustrated, the gas turbine system 30 includes a compressor 32, a combustor 34 and a turbine 36 that may be drivingly coupled to the compressor 32, and a component controller 31. The component controller 31 may be connected to the plant controller 22, which in turn may be connected to a user input device may be connected to receive communication from an operator 39. Alternatively, it will be understood that the component controller 31 and the plant controller 22 may be combined into a single controller. An inlet duct 40 directs ambient air to the compressor 32. As discussed in FIG. 3, water injected through the inlet duct 40 and / or other humidifying agent may be directed to the compressor. Intake passage 40 may include filters, strainers and mufflers that contribute to a pressure loss of ambient air flowing through intake passage 40 into swirl chokes 41 of compressor 32. An exhaust duct 42 routes combustion gases from an outlet of the turbine 36, for example, through emission control and silencing devices. The soundproofing materials and emission control devices may create a back pressure on the turbine 36. The turbine 36 may drive a generator 44 that generates electrical power that may then be distributed over the transmission lines 14 by the power system 10.
The operation of the gas turbine system 30 may be monitored by a plurality of sensors 46 that detect various operating conditions or parameters throughout the system, including, for example, conditions within the compressor 32, the combustor 34, the turbine 36, the generator 44, and the surrounding Environment 33. For example, temperature sensors 46 may monitor ambient temperature, compressor exit temperature, turbine exhaust temperature, and other temperatures within the flow path of the gas turbine system 30. Similarly, pressure sensors 46 may monitor ambient pressure, static and dynamic pressure levels at the compressor inlet, compressor outlet, turbine outlet, and at other suitable locations within the gas turbine system. Moisture sensors 46, such as e.g. Moisture and dry thermometers, can measure the ambient humidity in the inlet channel of the compressor. The sensors 46 may also be flow sensors, velocity sensors, flame sensor sensors, valve position sensors, vane angle sensors, and other sensors commonly used to measure various operating parameters and conditions relative to the operation of the gas turbine system 30. As used herein, the term "parameter" refers to measurable physical operating characteristics that are used to define operating conditions within a system, such as a system. The gas turbine system 30 or other generation system described herein may be used. Operating parameters may include temperature, pressure, humidity, and gas flow characteristics at positions defined along the path of the working fluid, as well as environmental conditions, fuel properties, and other measures, as well as without limitation. It will be understood that the control system 31 also includes a plurality of actuators 47 by which it mechanically controls the operation of the gas turbine system 30. The actuators 47 may include variable setpoint electro-mechanical devices or adjustments that allow the manipulation of certain process inputs (i.e., manipulated variables) to control process outputs (i.e., controlled variables) in accordance with a desired result or mode of operation. For example, commands generated by the component controller 31 may cause one or more actuators 47 within the turbine system 30 to adjust valves between the fuel supply and the combustor 34 that regulate the flow level, fuel splits, and / or type of combusted fuel. As another example, commands generated by the control system 31 may cause one or more actuators to adjust a swirl throttle setting that changes their orientation angle.
The component controller 31 may be a computer system having a processor executing program code for controlling the operation of the gas turbine system 30 using sensor measurements and instructions from the user or plant operator (hereinafter "operator 39"). As discussed in more detail below, software executed by the controller 31 may include scheduling algorithms for regulating each of the subsystems described herein. The component controller 31 may regulate the gas turbine system 30 in part based on algorithms stored in its digital memory. The algorithms may, for example, allow the component controller 31 to maintain NOx and CO emissions in the turbine exhaust within certain predefined limits or, in another example, maintain the combustion chamber firing temperature within predefined limits. It will be understood that the algorithms may include inputs for parameter variables, such as those shown in FIG. Compressor pressure ratio, ambient humidity, inlet pressure loss, turbine outlet back pressure, as well as any other suitable parameters may include. The schedules and algorithms performed by the component controller 31 accommodate variations in environmental conditions that affect emissions, combustor dynamics, firing temperature limits at full and part load operating conditions, and so forth. As discussed in greater detail below, the component controller 31 may employ gas turbine timing algorithms, such as those described with reference to FIG. those that set desired turbine exit temperatures and combustor fuel splits with the goal of meeting performance targets while maintaining gas turbine system operability limits. For example, the component controller 31 may determine the combustor temperature rise and NOx output during a partial load operation to raise the operating margin to the combustion dynamics limit and thereby improve the operability, reliability and availability of the block.
Referring to Figure 3, a schematic representation of an exemplary power plant 12 having multiple power plant blocks or plant components 49 in accordance with aspects of the present invention is provided. The illustrated power plant 12 of FIG. 3 is a common configuration and is used to discuss several of the exemplary embodiments of the present invention set forth below. However, as will be appreciated, the methods and systems described herein may be more generally applicable and scalable to power plants having more power plant blocks than those shown in FIG. 3, while also applicable to power plants having only a single generation component, such as eg which illustrated in Fig. 2. It will be understood that the power plant 12 of FIG. 3 is a combined cycle power plant having multiple plant components 49, including a gas turbine system 30 and a steam turbine system 50. The power generation may be increased by other plant components 49, e.g. an inlet conditioning system 51 and / or a heat recovery steam generator (HRSG) with a duct firing system (hereinafter "HRSG duct firing system 52"). It will be understood that each of the gas turbine system 30, the steam turbine system 50 including the HRSG ducting system 52, and the intake conditioning system 51 includes a control system or component controller 31 that communicates electronically with the sensors 46 and actuators 47, each Plant component are assigned. As used herein, unless otherwise stated, the intake conditioning system 51 may refer to components used to condition air prior to entering the compressor, including an inlet cooling system or system, an evaporator, nebulizer, water injection system, and / or. or, in some alternative cases, count a heating element.
In operation, the intake conditioning system 51 cools the air entering the gas turbine system 30 to increase the power generation capacity of the unit. The HRSG duct firing system 52 burns fuel to provide additional heat to increase the supply of steam that is expanded by a turbine 53. In this manner, the HRSG duct firing system 52 increases the energy provided by the hot exhaust gases 55 from the gas turbine system, thereby increasing the power generation capacity of the steam turbine system.
In an exemplary operation, the power plant 12 of FIG. 3 directs a fuel stream for combustion to the combustor 34 of the gas turbine system 30. The turbine 36 is driven by combustion gases and drives the compressor 32 and the generator 44 which supplies electrical power to the transmission lines 14 of the power supply system 10 supplies. The component controller 31 of the gas turbine system 30 may adjust commands for the gas turbine system regarding fuel flow and receive sensor data from the gas turbine system, such as a gas turbine system. Air intake temperature, humidity, power output, shaft speed and exhaust gas temperatures. The component controller 31 may also receive other operational data from pressure and temperature sensors, flow control devices, and other devices that monitor the operation of the gas turbine system. The component controller 31 may send data regarding the operation of the gas turbine system and receive instructions from the plant controller 22 regarding setpoints for actuators that control process inputs.
During certain modes of operation, the air entering the gas turbine system 30 may be cooled or otherwise conditioned by the intake conditioning system 51 to increase the generating capability of the gas turbine system. The inlet conditioning system 51 may include a refrigeration system 65 for cooling water, and a component controller 31 that controls its operation. In this case, the component controller 31 may receive information regarding the temperature of the cooling water as well as instructions regarding the desired injection level that may come from the plant controller 22. The component controller 31 of the inlet conditioning system 51 may also issue commands that cause the refrigeration system 65 to produce cooling water having a particular temperature and flow. The component controller 31 of the intake conditioning system 51 may send data regarding the operation of the intake conditioning system 51.
The steam turbine system 50 may include the turbine 53 and the HRSG channel firing system 52, as well as a component controller 31, as illustrated, dedicated to the control of its operation. Hot exhaust gases 55 from the exhaust ducts of the gas turbine system 30 may be directed into the steam turbine system 50 to produce the steam that is expanded by the turbine 53. As will be appreciated, HRSG duct firing systems are used regularly to provide additional energy for the production of steam to increase the generating capacity of a steam turbine system. It will be understood that the rotation induced within the turbine 53 by the steam drives a generator 44 to generate electrical energy which can then be sold within the power system 10 via power lines 14. The component controller 31 of the steam turbine system 50 can adjust the flow of fuel burned by the duct burner 52, thereby increasing the production of steam above the amount that can be generated with exhaust gases 55 alone. The component controller 31 of the steam turbine system 50 may send data regarding the operation of this plant component 49 and receive instructions on how it should operate.
The plant controller 22 of FIG. 3, as illustrated, may be connected to each of the component controllers 31 and communicate via these ports with sensors 46 and actuators 47 of the multiple plant components 49. As part of the control of the power plant 12, the plant controller 22 can simulate its operation. More specifically, the plant controller 22 may include or communicate with digital models (or simply "models") that simulate the operation of each plant component 49. The model may include algorithms that correlate process input variables with process output variables. The algorithms may include instruction sets, logic, mathematical formulas, function relationship descriptions, schedules, data collections, and / or the like. In this case, the plant controller 22 includes a gas turbine model 60 that models the operation of the gas turbine system 30; an intake conditioning system model 61 modeling the operation of the intake conditioning system 51; and a steam turbine model 62 modeling the operation of the steam turbine system 50 and the HRSG duct firing system 52. Generally, it will be understood that the systems and their relevant models, as well as the individual steps of the methods provided herein, may be variously divided and / or combined without materially departing from the scope of the present invention, and that the manner in which each of which is described, by way of example unless otherwise specified or claimed. Using these models, the plant controller 22 can control the operation, e.g. simulate the thermodynamic performance or parameters that describe the operation of the power plant 12.
The plant controller 22 may then use results from the simulations to determine optimized operating modes. Such optimized operating modes can be described by parameter sets, which include several operating parameters and / or setpoints for actuators and / or other operating conditions. As used herein, the optimized mode of operation is one that is preferred according to defined criteria or performance counters, at least to at least one alternative mode of operation that may be selected by an operator to evaluate plant operation. More specifically, optimized operating modes, as used herein, are those that are evaluated against one or more other possible operating modes, which were also simulated by the plant model, as preferred. The optimized operating modes are determined by evaluating how the model predicts how the power plant will work under each of them. As discussed below, an optimizer 64, e.g. a digital software optimization program, run the digital power plant model according to different sets of parameters and then identify preferred or optimized operating modes by evaluating the results. The variations in the setpoints may be generated by perturbations applied around the setpoints selected for analysis. These can be partly based on the past operation. It will be understood that the optimized mode of operation may be determined by the optimizer 64 based on one or more defined cost functions. Such cost functions may, for example, take into account the cost of generating electricity, profitability, efficiency or some other criteria as defined by the operator 39.
To determine the cost and profitability, the plant controller 22 may include or be in communication with an economic model 63 which tracks the price of electricity and certain other variable costs, e.g. the cost of the fuel used in the gas turbine system, intake conditioning system and HRSG duct combustion system. The economics model 63 may provide the data used by the plant controller 22 to judge which of the proposed setpoints (i.e., those chosen setpoints for which the optimized model setpoint operation is modeled) represents minimum production cost or maximum profitability. According to further embodiments, as discussed in greater detail with FIG. 4, the optimizer 64 of the plant controller 22 may include a filter, such as a filter. include, or operate in conjunction with, a Kalman filter to assist with the adjustment, adjustment and calibration of the digital models so that the models accurately simulate the operation of the power plant 12. As discussed below, the model may be a dynamic model that includes a learning mode in which it compares between actual operation (ie, values for measured operating parameters that reflect the actual operation of the power plant 12) and the predicted operation (ie, values for the same operating parameters that the model predicted) are adjusted or tuned. As part of the control system, the filter can also be used to adjust or calibrate the models in real time or near real time, such every few minutes or hours or as specified.
The optimized setpoints generated by the plant controller 22 represent a recommended operating mode and may include, for example, fuel and air settings for the gas turbine system, temperature and water mass flow for the inlet conditioning system, or the level of duct firing within the steam turbine system 50 count. According to certain embodiments, these proposed operating setpoints may be provided via an interface device, such as an interface device. a computer display screen, printer or speaker are provided to the operator 39. Once the optimized setpoints are known, the operator may then input the setpoints to the plant controller 22 and / or the component controller 31, which then generates control information to achieve the recommended operating mode. In such embodiments, where the optimized setpoints do not include specified control information to achieve the operating mode, the component controllers may provide the necessary control information therefor and, as discussed in greater detail below, control the plant component in a closed manner according to the recommended operating mode up to continue next optimization cycle. Depending on the operator preference, the plant controller 22 may also implement directly or automatically optimized setpoints without operator intervention.
In an exemplary operation, the power plant 12 of FIG. 3 directs a fuel stream for combustion to the combustor 34 of the gas turbine system 30. The turbine 36 is driven by combustion gases to drive the compressor 32 and the generator 44 which supply electrical power to the transmission lines 14 of the power supply system 10 supplies. The component controller 31 may adjust commands for the gas turbine system 30 regarding fuel flow and receive sensor data from the gas turbine system 30, such as, for example, FIG. Air inlet temperature and humidity, power output, shaft speed and exhaust gas temperatures. The component controller 31 may also collect other operating data from pressure and temperature sensors, flow control devices, and other devices that monitor the gas turbine system 30. The component controller 31 of the gas turbine system 30 may send data regarding the operation of the system and receive instructions from the plant controller 22 regarding the setpoints for actuators that control process inputs.
During certain modes of operation, the air entering the gas turbine system 30 may be cooled by cold water provided to the intake air passage 42 from the intake conditioning system 51. It will be understood that the cooling of the air entering a gas turbine may occur to increase the capacity of the gas turbine engine to generate electricity. The intake conditioning system 51 includes a refrigeration system or chiller 65 for cooling water and a component controller 31. In this case, the component controller 31 receives information regarding the temperature of the cooling water and commands regarding the desired cooling of the intake air. These commands may come from the plant controller 22. The component controller 31 of the intake conditioning system 51 may also issue commands to cause the refrigeration system 65 to produce cooling water having a particular temperature and flow. The component controller 31 of the inlet conditioning system 51 may send data regarding the operation of the inlet conditioning system 51 and receive instructions from the controller 22.
The steam turbine system 50 includes a HRSG having a channel firing device 52, a steam turbine 53, and a component controller 31 that may be dedicated to its operation. Hot exhaust gases 55 from an exhaust duct 42 of the gas turbine system 30 are directed into the steam turbine system 50 to produce the steam that drives it. The HRSG channel firing system 52 may be used to provide additional heat energy to generate steam to increase the generating capability of the steam turbine system 50. The steam turbine 53 drives the generator 44 to generate electrical energy that is supplied to the power system 10 via the power lines 14. The component controller 31 of the steam turbine system 50 may adjust the flow of the fuel burned by the duct burner 52. The heat generated by the channel firing device increases the generation of steam beyond the amount generated by the exhaust gases 55 from the turbine 36 alone. The component controller 31 of the steam turbine system 50 may send data regarding the operation of the system to the plant controller 22 and receive instructions therefrom.
The plant controller 22 may communicate with the operator 39 and data resources 26, for example, to obtain data on market conditions, e.g. Prices and demand for delivered electricity. According to certain embodiments, the plant controller 22 issues recommendations regarding desired operating setpoints for the gas turbine system 30, the intake conditioning system 51, and the steam turbine system 50 to the operator 39. The plant controller 22 may receive and store data for operating the components and subsystems of the power plant 12. The plant controller 22 may be a computer system having a processor and memory for storing data, the digital models 60, 61, 62, 63, the optimizer 64, and other computer programs. The computer system may be implemented in a single physical or virtual computing device or distributed to local or remote computing devices. The digital models 60, 61, 62, 63 may be implemented as a set of algorithms, e.g. Transfer functions that relate the operating parameters of each of the systems. The models may include a physics-based aero-thermodynamic computer model, a regression fit model, or another suitable computer-implemented model. According to preferred embodiments, the models 60, 61, 62, 63 may be regularly tuned, adjusted or calibrated automatically and in real time or near real time, or tuned according to ongoing comparisons between the predicted operation and the measured parameters of the actual operation. Models 60, 61, 62, 63 may include filters that receive data inputs regarding actual physical and thermodynamic operating conditions of the combined cycle power plant. These data inputs may be provided to the filter in real time or periodically every 5 minutes, 15 minutes, hours, days, etc. during operation of the power plant 12. The data inputs may be compared to data predicted by digital models 60, 61, 62, 63, and based on the comparisons, the models may be continually refined.
FIG. 4 illustrates a schematic system configuration of a plant controller 22 including a filter 70, an artificial neural network configuration 71 ("neural network 71"), and an optimizer 64 in accordance with aspects of the present invention. The filter 70, which may be, for example, a Kalman filter, may obtain the actual data 72 of measured operating parameters from the sensors 46 of the power plant 12 with predicted data 73 of the same operating parameters by the models 60, 61, 62, 63 and the neural network 71, which simulates the operation of the power plant 12, compare. Differences between the actual data and predicted data may then be used by the filter 70 to tune the model of the power plant simulated by the neural network 71 and the digital models.
It should be understood that while certain aspects of the present invention are described herein with reference to models in the form of models based on a neural network, it is contemplated that the present invention may be practiced using other types can be implemented by models including, but not limited to, physics-based models, data-driven models, empirically-developed models, heuristic-based models, support vector machine models, models developed by linear regression, models developed using knowledge of "principles", etc. In addition, for properly detecting the relationship between the manipulated variables / variables and the controlled variables according to certain preferred embodiments, the power plant model may have one or more of the following characteristics: 1) nonlinearity (a nonlinear model is able to plot a curve ansta a straight-line relationship between manipulated variables / disturbance variables and controlled variables); 2) multiple input / multiple output (the model may be able to capture the relationships between multiple inputs - the manipulated variables and disturbances - and multiple outputs of control variables); 3) dynamic (changes in inputs may not affect outputs immediately, but may result in a time delay followed by a dynamic response to the changes, for example, it may take several minutes for input changes to progress fully through the system Since optimization systems operate at a predetermined frequency, the model must account for and consider the effects of these changes over time; 4) adaptive (the model can be updated at the beginning of each optimization to reflect the current operating conditions); and 5) derived from empirical data (since each power plant is unique, the model may be derived from empirical data obtained from the power generation unit). In view of the above requirements, the neural network based approach is a preferred technology for implementing the necessary plant models. Neural networks can be developed based on empirical data using advanced regression algorithms. As will be understood, neural networks are capable of detecting the nonlinearity that often occurs in the operation of the power plant components. Neural networks can also be used to represent systems with multiple inputs and outputs. In addition, neural networks can be updated using either feedback or adaptive online learning. Dynamic models can also be implemented in a neural network-based structure. A variety of different types of model architectures have been used to implement dynamic neural networks. Many of the model architectures of neural networks require a large amount of data to successfully train the dynamic neural network. Given a robust power plant model, it is possible to calculate the effects of changes in the control variables on the controlled variables. Further, since the plant model is dynamic, it is possible to calculate the effects of changes in the manipulated variables over a future time horizon.
The filter 70 may generate power multipliers applied to inputs or outputs of the digital models and the neural network, or the weights applied to the logical units and algorithms used by the digital models and the neural network be, modify. These actions through the filter reduce the differences between the actual condition data and the predicted data. The filter continues working to further reduce the differences or deals with fluctuations that may occur. For example, the filter 70 may generate power multiples for the predicted data related to compressor gasket pressure and temperature in the gas turbine, the efficiency of the gas and steam turbine, the fuel flow to the gas turbine system, the inlet conditioning system and HRSG channel firing system, and / or other suitable parameters. It will be understood that these categories of operational data reflect operational parameters that are subject to performance degradation over time. By providing performance multipliers for these types of data, the filter 70 may be particularly useful in accommodating the models and the neural network for accommodating degradation in power plant performance.
As illustrated in FIG. 4, according to certain embodiments of the present invention, each of the digital models 60, 61, 62, 63 of the plurality of plant components 49 of the power plant of FIG. 3 includes algorithms represented by the plurality of curves associated with Modeling the appropriate systems can be used. The models interact and communicate within the neural network 71, and it will be understood that thereby the neural network 71 forms a model of the entire combined cycle power plant 12. In this way, the neural network simulates the thermodynamic and economic operation of the plant. As indicated by the solid arrow lines in Fig. 4, the neural network 71 collects data output by the models 60, 61, 62, 63 and provides data to be used as inputs by the digital models.
The plant controller 22 of FIG. 4 also includes an optimizer 64, such as an inverter. a computer program that interacts with the neural network 71 to search for optimal setpoints for the gas turbine system, intake conditioning system, steam turbine system, and HRSG channel firing system to achieve a defined performance goal. For example, the performance goal may be maximizing the profitability of the power plant. The optimizer 64 may cause the neural network 71 to execute the digital models 60, 61, 62, 63 with different operating setpoints. The optimizer 64 may include perturbation algorithms that help to vary the operating setpoints of the models. The perturbation algorithms cause the simulation of the combined cycle power plant provided by the digital models and the neural network to operate with setpoints that are different from the current plant setpoints for the plant. By simulating the plant's operation with different setpoints, the optimizer 64 searches for operating setpoints that would cause the plant to operate more efficiently or improve performance through some other criteria that can be defined by the operator 39.
According to exemplary embodiments, the business model 63 provides data that is used by the optimizer 64 to determine which setpoints are most profitable. For example, the business model 63 may receive and store fuel cost data, e.g. how a graphic 630 is formatted, which estimates the fuel cost over time, e.g. during the seasons of a year, correlates. Another graph 631 may correlate the price received for electric power at different times of a day, a week, or a month. The economics model 63 may provide data regarding the price received for electricity and the cost of the fuel (gas turbine fuel, duct combustion fuel, and intake conditioning system fuel) used to produce it. The data from the business model 63 may be used by the optimizer 64 to evaluate each of the operating states of the power plant according to the performance goals defined by the operator. The optimizer 64 may identify which of the operating conditions of the power plant 12 is optimal given the performance goals defined by the operator 39 (which, as used herein, is at least preferred to an alternative operating condition). As described, the digital models may be used to simulate the operation of the plant components 49 of the power plant 12, such as e.g. modeling the thermodynamic operation of the gas turbine system, the intake conditioning system or the steam turbine system. The models may include algorithms, e.g. mathematical equations and look-up tables that may be stored locally and periodically updated or remotely acquired via data resources 26 that simulate the response of plant components 49 to specific input conditions. Such look-up tables may include measured operating parameters that describe the operation of the same type of components used in remote power plant installations.
The thermal model 60 of the gas turbine system 30 includes, for example, an algorithm 600 that correlates the effect of the temperature of the intake air on the power output. It will be understood that this algorithm may show that the current output decreases from a maximum value 601 as the intake air temperature increases above a threshold temperature 602. The model 60 may also include an algorithm 603 that correlates the specific heat consumption of the gas turbine at different current output levels of the engine. As discussed, specific heat consumption is the efficiency of a gas turbine engine or other power generation unit, and conversely, it is related to efficiency. Lower specific heat consumption indicates higher thermodynamic power efficiency. The digital model 61 may simulate the thermodynamic operation of the intake conditioning system 51. In this case, the digital model 61 includes, for example, an algorithm 610 that correlates cooling capacity based on energy applied to operating the refrigeration system 65 of the intake conditioning system 51, such that the calculated cooling capacity indicates the amount of cooling that is due to the cooling Air is applied, which enters the gas turbine. There may be a maximum cooling capacity value 611 that can be achieved by the refrigeration system 65. In another case, a related algorithm 612 may correlate the energy used to operate the refrigeration system 65 with the temperature of the cooled air entering the compressor 32 of the gas turbine system 30. The model 61 may, for example, show that the energy required to operate the intake conditioning system increases dramatically as the temperature of the air entering the gas turbine drops below the dew point 613 of the ambient air. In the case of the steam turbine system 50, the digital model 62 may include an algorithm 620 that correlates the power output of the steam turbine system with the energy supplied by the HRSG channel firing system 52, such as, for example, FIG. the amount of fuel consumed by the duct firing. The model 62 may indicate, for example, that there is an upper threshold level 621 for the increase in steam turbine system output that may be achieved by the HRSG channel firing system, which may be included in the algorithm 620.
According to certain embodiments of the present invention, as illustrated in FIG. 4, the neural network 71 may interact and provide communication through each of the digital models of the multiple plant components 49 of the power plant 12 of FIG. 3. The interaction may include collecting output data from the models and generating input data used by the models to generate additional output data. The neural network 71 may be a digital network of connected logical elements. The logical elements may each embody an algorithm that accepts data inputs to generate one or more data outputs. A simple logical element can sum the values of the inputs to produce output data. Other logical elements may multiply values of the inputs or apply other mathematical relationships to the input data. The data inputs to each of the logic elements of the neural network 71 may be assigned a weight, such as a weight. a multiplier between one and zero. The weights may be adjusted during a learning mode that adjusts the neural network to better model the power plant's performance. The weights may also be adjusted based on commands provided by the filter. Adapting the weights of the data inputs to the logical units in the neural network is one example of the manner in which the neural network can be dynamically modified during operation of the combined cycle power plant. Other examples include modifying the weights of data inputs in algorithms (which is an example of a logical unit) in each of the thermodynamic digital models for the steam turbine system, the intake conditioning system, and the gas turbine. The plant controller 22 may also be modified in other ways, e.g. by adjustments in the logical units and algorithms, based on the data provided by the optimizer and / or filter.
The plant controller 22 may generate an output of recommended or optimized target values 74 for the combined cycle power plant 12, which, as illustrated, are first routed to an operator 39 for approval before being communicated and implemented by the power plant actuators 47. As illustrated, the optimized setpoints 74 may include or be approved by an operator 39 input through a computer system, such as a computer. described below with reference to FIG. For example, to the optimized set values 74, a temperature and mass flow for the cooling water may be generated by the inlet conditioning system and used to cool the air entering the gas turbine system; a fuel flow to the gas turbine system; and counting a channel firing rate. It will be understood that the optimized setpoints 74 may then also be used by the neural network 71 and the models 60, 61, 62, 63, such that the ongoing system simulation may predict operating data that can later be compared to actual operating data, so that Plant model can be continuously refined.
FIG. 5 illustrates a simplified system configuration of a plant controller 22 having an optimizer 64 and the power plant model 75. In this exemplary embodiment, the plant controller 22 is shown as a system with the optimizer 64 and the power plant model 75 (which, for example, the above Referring to Fig. 4, neural network 71 and models 60, 61, 62, 63 are included). The power plant model 75 may simulate the overall operation of a power plant 12. In accordance with the illustrated embodiment, the power plant 12 includes multiple power plant blocks or plant components 49. The plant component 49 may include, for example, thermal power plants, or other plant subsystems, as previously described, which may include all corresponding component controllers 31. The plant controller 22 may communicate with the component controllers 31 and may control the operation of the power plant 12 via and through the component controllers 31 via connections to the sensors 46 and actuators 47.
It will be understood that power plants have numerous variables that affect their operation. All of these variables can generally be categorized as either input variables or output variables. Input variables represent process inputs, and include variables that can be manipulated by plant operators, e.g. the air and fuel flow. Input variables also include those variables that can not be manipulated, such as Environmental conditions. Output variables are variables, e.g. the power output that is controlled by manipulating those input variables that can be manipulated. A power plant model is configured to include the algorithmic relationship between input variables, including those that can be manipulated, or "manipulated variables", and those that can not be manipulated, or "disturbances," and output or controlled variables, which are referred to as "controlled variables" represents. More specifically, manipulated variables are those that can be varied by the plant controller 22 to affect the controlled variables. Actuating variables include such things as valve setpoints that control the flow of fuel and air. Disturbances refer to variables that influence controlled variables but can not be manipulated or controlled. Exciter variables include environmental conditions, fuel characteristics, etc. The optimizer 64 determines an optimal set of setpoint values for the given manipulated variables: (1) power plant performance goals (e.g., meeting load requirements while maximizing profitability); and (2) limitations associated with the operation of the power plant (e.g., emissions and equipment limitations).
According to the present invention, an "optimization cycle" may begin at a predetermined frequency (e.g., every 5 to 60 seconds or 1 to 30 minutes). At the beginning of an optimization cycle, the plant controller 22 may receive available data for manipulated variables, controlled variables and disturbances from the component controllers 31 and / or directly from the sensors 46 of each of the plant components 49. The plant controller 22 may then use the power plant model 75 to determine optimal set point values for the manipulated variables based on the present data. As a result, the plant controller 22 may execute the plant model 75 at various operating setpoints to determine which set of operating setpoints is most preferred in view of the power plant performance goals, which may be referred to as "simulation runs". For example, a performance goal may be to maximize profitability. By simulating the plant's operation with different setpoints, the optimizer 64 searches for the set of setpoints predicted by the plant model 75, which causes the plant to operate in an optimal (or at least preferred) manner. As indicated, this optimal set of setpoints may be referred to as "optimized setpoints" or as an "optimized mode of operation". Typically, optimizer 64, when arriving at the optimized setpoints, has compared numerous sets of setpoints, and the optimized setpoints are ranked superior to each of the other sets given the performance objections defined by the operator. The operator 39 of the power plant 12 may have the option to approve the optimized setpoints, or the optimized setpoints may be automatically approved. The system controller 22 can send the optimized setpoint values to the component controller 31 or, alternatively, directly to the actuators 47 of the system components 49, so that settings according to the optimized setpoint values can be adjusted. The plant controller 22 may operate in a closed-loop manner to adjust the setpoint values of the manipulated variables at a predetermined frequency (e.g., every 10-30 seconds or more) based on the measured current operating conditions.
The optimizer 64 may be used to minimize a "cost function" that is subject to a number of limitations. The cost function is essentially a mathematical representation of an asset performance goal, and the limitations are limits within which the power plant must operate. Such limits may be legal, regulatory, environmental, equipment or physical limitations. For example, the cost function of minimizing NOx emissions includes a term that decreases as the NOx level decreases. A common method for minimizing such a cost function is known, for example, as "gradient descent optimization". Gradient descent is an optimization algorithm that approximates a local minimum of a function by taking steps proportional to the negative of the gradient (or approximate gradient) of the function at the current point. It should be understood that a variety of different optimization techniques may be used, depending on the shape of the model and the cost and constraints. For example, it is contemplated that the present invention may be implemented by using, singly or in combination, a variety of different types of optimization approaches. These optimization approaches include, but are not limited to, linear programming, quadratic programming, mixed integer nonlinear programming, stochastic programming, global nonlinear programming, genetic algorithms, and particle / swarm techniques. In addition, the asset model 75 can be dynamic, so that effects of changes over a future time horizon are taken into account. Therefore, the cost function includes terms over a future horizon. Since the model is used to predict over a time horizon, this approach is referred to as model predictive control, which is described in S. Piche, B. Sayyar-Rodsari, D. Johnson and M. Gerules, "Nonlinear model predictive control using neural networks", IEEE Control Systems Magazine, Vol. 20, No. 2, pp. 53-62, 2000 and which is incorporated herein by reference in its entirety.
Restrictions may apply both to process inputs (which control variables count) and process outputs (to which control variables count) of the power plant over the future time horizon. Usually, restrictions that are consistent with limits associated with the plant controller apply to the manipulated variables. Limitations of the outputs can be determined by the problem being solved. According to embodiments of the present invention and as a step in the optimization cycle, the optimizer 64 may calculate the complete curve over which the manipulated variable moves over the future time horizon, for example one hour. Thus, for an optimization system running every 30 seconds, 120 values can be calculated over a one-hour future time horizon for each manipulated variable. Because the plant model or performance targets or constraints may change prior to the next optimization cycle, the plant controller 22 / optimizer 64 may output only the first value in the time horizon for each manipulated variable to the component controllers 31 as optimized setpoints for each respective manipulated variable. At the next optimization cycle, the plant model 75 can be updated based on the current conditions. The cost function and limitations can also be updated if they have changed. The optimizer 64 may then be used to recalculate the set of values for the manipulated variables over the time horizon, and the first value in the time horizon for each manipulated variable is output to the component controller 31 as setpoint values for each respective manipulated variable. The optimizer 64 may repeat this process for each optimization cycle, thereby constantly maintaining the optimum performance, even if the power plant 12 is affected by unanticipated changes in such points as load, environmental conditions, fuel properties, and so on.
Referring to FIG. 6, an illustrative environment and user input device for a plant controller and control program according to an exemplary embodiment is shown. Although other configurations are possible, the embodiment includes a computer system 80 having a display 81, a processor 82, a user input device 83, and a memory 84. Aspects of the computer system 80 may reside in the power plant 12 while other aspects are removed and communicated via a communication network 20 can be connected. As discussed, the computer system 80 may be associated with each block or other plant component 49 of the power plant 12. The power plant components 49 may include the gas turbine system 30, the steam turbine system 50, the intake conditioning system 51, the HRSG duct firing system 52, and / or any subsystems or subcomponents related thereto, or any combination thereof. Computer system 80 may also be connected to one or more sensors 46 and actuators 47, as may be necessary or desirable. As indicated, the sensors 46 may be configured to detect operating conditions and parameters of the components and pass signals to the computer system 80 regarding these conditions. The computer system 80 may be configured to receive these signals and use them in a manner described herein that may include transmitting signals to one or more of the actuators 47. However, unless otherwise required, the present invention may include embodiments that are not configured to directly control the power plant 12 and / or detect operating conditions. In configurations of the present invention that control the power plant 12 and / or detect operating conditions, such input or control may be accomplished by receiving and / or transmitting signals from one or more separate software or hardware systems. n, which interact more directly with physical components of the power plant and its sensors and actuators. The computer system 80 may include a power plant control program ("control program") by which the computer system 80 may manage data in a plant controller by performing the processes described herein.
In general, processor 82 executes program code that defines the control program that is at least partially stored in memory 84. During execution of the program code, processor 82 may process data, which may result in reading and / or writing transformed data from memory 84. The display 81 and input device 83 may enable a human user to interact with the computer system 80 and / or one or more communication devices to enable a system user to communicate with the computer system 80 using any type of communication link. In embodiments, a communication network, such as e.g. Network hardware / software, allow computer system 80 to communicate with other devices inside and outside a node in which it is installed. To this end, the control program of the present invention may manage a set of interfaces that allow human and / or system users to interact with the control program. Further, as discussed below, the control program may store data, such as data. Manage control data (e.g., store, retrieve, generate, manipulate, organize, present, etc.) through any solution.
The computer system 80 may include one or more general purpose computing products capable of executing program code, such as the like. the control programs defined herein installed thereon. As used herein, it is understood that "program code" means any collection of instructions, in any language, code or notation, that cause a computing device having information processing capability to perform a particular action, either directly or after any combination of the following : (a) conversion to another language, code or notation; (b) reproduction in a different material form; and / or (c) decompression. In addition, the computer code may include object code, source code, and / or executable code, and may form part of a computer program product when residing on at least one computer-readable medium. It is understood that the term "computer readable medium" includes one or more of any type of tangible expression medium currently known or to be developed in the future, from which a copy of the program code can be recognized, reproduced, or otherwise communicated by a computing device. When the computer executes the computer program code, it becomes an apparatus for practicing the invention, and on a general purpose microprocessor, specific logical circuits are created by configuring the microprocessor with computer code segments. A technical effect of the executable instructions is the implementation of a power plant control method and / or system and / or computer program product that utilizes models to increase or increase or optimize the operational characteristics of power plants to meet anticipated environmental and / or market conditions, performance parameters, and / or the like associated lifecycle costs to use the economic yield of a power plant more efficiently. In addition to using up-to-date information, historical and / or predicted information can be used, and a feedback loop can be established to more dynamically operate the plant during fluctuating conditions. The computer code of the control program may be written in computer instructions executable by the plant controller 22. For this, the control program executed by the computer system 80 may be implemented as any combination of system software and / or application software. Furthermore, the control program may be implemented using a module set. In this case, one module may allow the computer system 80 to perform a set of tasks used by the control program, and may be implemented separately and / or implemented remotely from other portions of the control program. As used herein, the term "component" means any configuration of hardware, with or without software, that implements the functionality described in connection therewith through any solution, while the term "module" means program code that enables the computer system that is in communication implement any action described by any solution. When stored in the memory 84 of the computer system 80 that includes the processor 82, a module is an integral portion of a component that implements the actions. Notwithstanding, it is understood that two or more components, modules, and / or systems may share some / all of their respective hardware and / or software. Further, it is understood that some of the functionality discussed herein may not be implemented or that additional functionality may be included as part of the computer system 80. If the computer system 80 includes multiple computing devices, only one portion of the control program may be stored on each computing device (e.g., one or more modules). Regardless, if the computer system 80 includes multiple computing devices, the computing devices may communicate over any type of communication link. Further, while performing a process as described herein, computer system 80 may communicate with one or more other computer systems by any type of communication link.
As discussed herein, the control program allows computer system 80 to implement a power plant control product and / or method. The computer system 80 may receive power plant control data by any means. For example, the computer system 80 may generate and / or use power plant control data, retrieve power plant control data from one or more data stores, databases or sources, power plant control data from another system or device within or outside a power plant, plant controller, component controller / s and / or the like received. In a further embodiment, the invention provides a method of providing a copy of program code, such as e.g. for the power plant control program, which may implement part or all of a process described herein. It is understood that aspects of the invention may be implemented as part of a business process that performs a process described herein on a subscription, advertising and / or fee basis. A service provider might offer to implement a power plant control program and / or method as described herein. In this case, the service provider may use a computer system, such as a computer system. the computer system 80, manage (e.g., create, maintain, support, etc.) that performs a process described herein for one or more customers.
Computer models of power plants can be constructed and then used to control and optimize power plant operation. Such plant models may be dynamic and iteratively updated through continuous comparison between actual (i.e., measured) operating parameters versus the same parameters as predicted by the plant model. In creating and maintaining such models, instructions may be written or otherwise provided that instruct the processor 82 of the computer system 80 to create a library of power system power plant blocks and components ("library of components") in response to user input. In some configurations, the user input and the created library include properties of the component with the library, as well as rules for creating scripts in accordance with operational and property values. These property values may be compiled from data stored locally in memory 84 and / or extracted from a central database maintained at a remote location. The library of components may contain non-physical components, such as e.g. economic or legal components. Examples of economic components are fuel purchases and sales, and examples of legal components are emission limits and credits. These non-physical components can be modeled using mathematical rules, just as components that represent physical equipment can be modeled using mathematical rules. The instructions may be configured to compile a configuration of power system components from the library, as may be configured by an operator. A library of power system components may be provided so that a user may select components therefrom to replicate the actual power plant or create a hypothetical one. It will be understood that each component may have multiple characteristics that may be used by the user to input specific values that match the operating conditions of an actual or hypothetical power plant being modeled. Scripts can be created for the assembled power system components and their configuration. The created scripts may include mathematical relationships within and / or between the power system components, including economic and / or legal components, if used in the power system component configuration. Computer system 80 may then resolve mathematical relationships and display results of the solution on display 81. In configurations in which signals may be transmitted from the computer 80, the signals may be used to control a power system in accordance with the results of the solution. Otherwise, the results may be displayed or printed and used for the adjustment of physical equipment parameters and / or determination and / or use of certain non-physical parameters, e.g. Fuel purchase and / or sales, are used, so that a preferred or optimized mode of operation is achieved. The library of plant components may include a central database that represents a constant accumulation of data related to how each plant component operates under different parameters and conditions. The central database can be used to provide, for example, "line data" when sensor data is unreliably determined.
Referring to Figs. 7-9, a more detailed discussion of the economic load balancing process is provided, including ways in which the control systems discussed above may be used to perform such load distribution operations from the point of view of both a central power system authority and individual power plants, as appropriate involved in such systems. It will be understood that, from the point of view of a central authority dispatcher, the goal of the economic load balancing process is to dynamically respond to changing variables, including changing load requirements or environmental conditions, while still minimizing production costs within the system. For the power plants involved, it will be understood that, in general, the goal is to use the available capacity while minimizing generation costs so as to maximize the economic yield. Given the complexities of power systems, the process of economic load sharing usually involves the frequent adaptation of the load in the participating power plants by the dispatcher. If successful, the process results in the available power plants being operated with loads in which their incremental generation costs are about the same - resulting in a minimization of the production cost - while also considering system constraints, such as those of the prior art. maximum and minimum allowable loads, system stability, etc. It will be understood that accurate incremental cost data is necessary for optimal economic load balancing. Such incremental cost data includes primary components, which include fuel costs and incremental fuel consumption. The data of the incremental fuel consumption is usually given as a graph of the incremental specific heat consumption versus the power output. Specifically, the incremental heat rate (IHR) of a thermal power plant is defined as the increase in the specific heat consumption curve, the unit's specific heat consumption being the ratio of the heat input plotted against the electrical output at each load. Errors in these data result in the load distribution on units, which does not minimize the total cost of production.
A series of points may introduce errors into the curve of incremental specific heat consumption. These can be grouped into two categories. A first category includes points that generate errors that exist at the time the data is submitted to the dispatcher. For example, when the data is collected through testing, errors due to instrument inaccuracy are entered in all calculations performed on them. As discussed in more detail below, certain aspects of the present invention include ways to confirm sensor accuracy during data collection and timely identification of instances where collected data may be unreliable due to sensor malfunction. A second category of errors includes points that cause data to lose accuracy over time. For example, if the performance of a block changes due to equipment wear or repair or changes in environmental conditions, the incremental specific heat consumption data used for the distribution will be in error until such data is updated. One aspect of the present invention is to identify those parameters of thermal power plants that may have a significant impact on the calculations of incremental specific heat consumption. The knowledge of such parameters and their relative significance can then be used to determine how often load distribution data should be updated to reflect true system performance.
Errors in the data of the incremental specific heat consumption lead to situations in which the load is incorrectly distributed to the power plants, which usually results in increased generation costs for the power system. For example, referring to the graph of FIG. 7, a situation is provided in which the true incremental specific heat consumption differs from the incremental specific heat consumption used in the load distribution process. When distributing load among the units, the load-distribution authority uses the incremental specific heat consumption data that is erroneous by "E" as indicated. (It should be noted that FIG. 7 assumes that incremental specific heat consumption of the power system is not affected by the load assigned to the given unit, which may be substantially correct when the power system is compared to the size of the given block is a large one.) As shown, the load distribution on the block is at Li, which is the load at which the incremental specific heat consumption of the unit and system is the same based on the information available. If the correct information of the incremental specific heat consumption was used, the load at L2 would be assigned to the unit where the true incremental specific heat consumption of the plant is equal to the incremental specific heat consumption of the power supply system. As will be understood, the failure results in the underuse of the power plant. In cases where the alternative is true, i. in which the positioning of the plot of incorrect incremental specific heat consumption is reversed relative to the plot of true incremental specific heat consumption, the error results in the unit being excessively stressed, possibly requiring it to operate inefficiently, due to the load commitment assigned to it to fulfill. From the point of view of the central load distribution authority of the power system, it will be understood that reducing errors in the data used in the load sharing process will lower overall system fuel costs, increase system efficiency, and / or reduce the risk that load requirements will not be met. For the operators of power plants within the system, reducing such errors should support the full utilization of the plant and improve the economic yield.
FIGS. 8 and 9 respectively illustrate a schematic representation of a plant controller 22 and a flowchart 169 of a control method according to aspects of the present invention. In these examples, methods are provided that illustrate economic optimization within a power system that utilizes economic load sharing to distribute the load among potential providers. The basic process of economic load balancing is a process that can be used in different ways, as well as between any two levels defined within the layered hierarchy common to many power systems. For example, in one case, the economic load balancing process may be used as part of a concurrent process by which a central government agency or an industrial cooperation association distributes the burden among several competing companies. Alternatively, the same basics of economic load sharing can be used to divide the load among co-owned power plants to minimize generation costs for the owner of the plants. It can also be used at the plant level, as a way for an operator or plant controller to split the load requirements among the different local power plant blocks that are available to him. It will be understood that unless otherwise stated, the systems and methods of the present invention are generally applicable to any of these possible manifestations of the economic load sharing process.
In general, the load sharing process seeks to minimize generation costs within a power system by establishing a load-sharing schedule in which the incremental generation cost for each involved power plant or unit is about the same. As will be understood, multiple terms are often used to describe the economic load balancing process and are defined as follows. A "forecast horizon" is a predefined period of time over which an optimization is to be carried out. For example, a typical forecast horizon can take from a few hours to a few days. An "interval" within the prediction horizon is a predefined time resolution of the optimization, i. the aforementioned "optimization cycle", which describes how often an optimization is to be carried out during the forecasting horizon. For example, a typical time interval for an optimization cycle may take from several seconds to several minutes. Finally, a "prediction time" is the number of time intervals for which optimization is to be performed, and it can be obtained by dividing the prediction horizon by the time interval. Thus, for a 12-hour forecast horizon and a 5-minute time interval, a prediction time is 144 time intervals.
Aspects of the present invention provide methods for controlling and / or controlling power plants, as well as methods and systems for optimizing performance, cost-effectiveness, and efficiency. For example, according to the present invention, minimum variable operating costs for a thermal power plant or power plant can be achieved including variable performance characteristics and cost parameters (ie, fuel costs, environmental conditions, market conditions, etc.) with life cycle costs (ie, variable operation and its effect on maintenance schedules, part replacement, etc .). By varying one or more parameters of a thermal power plant, taking into account such factors, a more economical benefit can be drawn from the unit over its lifetime. For example, at power plants that include a gas turbine, the firing temperature may be varied to provide a desired load level that is more economical based on operating profile, environmental conditions, market conditions, predictions, power plant performance, and / or other factors. Because of this, disposal of parts that still have a residual hourly life left in run-limited units can be reduced. Further, a power plant control system incorporating a feedback loop, which is updated substantially with real-time data from sensors that are periodically tested and confirmed to be working correctly, allows further plant optimization. That is, according to certain embodiments of the present invention, the introduction of a real time feedback loop between the power plant control system and the load distribution authority, target load, and unit commitment may be based on highly accurate supply curves created based on real time engine performance parameters.
FIG. 8 illustrates a schematic design of an exemplary plant controller 22 in accordance with aspects of the present invention. It will be understood that the plant controller 22 may be particularly well suited to the implementation of the method 169 of FIG. 9. Thus, FIGS. 8 and 9 will be discussed together, although it will be understood that each may have aspects that are applicable to more general usage. The power system 10 illustrated in FIG. 8 includes a "power plant 12a" for which the plant controller 22 is dedicated, as well as "other power plants 12b," which may be power plants within the power system competing with the power plant 12a. As illustrated, the power system 10 also includes a load sharing authority 24 that, through a dedicated system controller 25, manages the load sharing process between all participating power plants 12a, 12b within the system.
The power plant 12a may include numerous sensors 46 and actuators 47, by which the plant controller 22 monitors operating conditions and controls the operation of the plant. The plant controller 22 may communicate with numerous data resources 26 which may be remote therefrom and are accessible via a communication network and / or located locally and accessible via a local area network. As illustrated, the schematic of the plant controller 22 includes several subsystems defined by the multiple boxes. These subsystems or "boxes" were separated mainly by function to simplify the description. However, it will be understood that separate boxes may or may not represent individual chips or processors or other discrete hardware elements, and may or may not constitute separate portions of computer program code executed within the plant controller unless otherwise specified. Similarly, while method 169 is divided into two main sections or blocks, for the sake of simplicity and ease of description. It will be understood that all of the separate boxes shown in FIG. 8 may be combined into one or more sections in the plant controller 22, as well as all of the separate boxes or steps shown in FIG. 9.
For example, the method 169 of FIG. 9 may begin with a control section 170 that receives and collects existing information and data for use (at step 171), which may include market data, operational data, and / or environmental data. Within the plant controller 22, a corresponding control module 110 may be arranged to request / receive this type of data from the data resources 26 or any other suitable source. The control module 110 may also be configured to receive a destination load 128 from the load-balancing authority 24 (although at an initial run such a target load may not be available and a predefined initial target load may be used). Environment data may be received from remote or local databases and / or forecast services, and may be included as a component of the data resources 26. Environmental data may also be collected via environmental sensors operating around the power plant 12a and received via a communication link with the load distribution authority 24. According to aspects of the present invention, environmental data includes historical, present, and / or predicted data describing environmental conditions for the power plant 12a, which may include, for example, air temperature, relative humidity, pressure, and so forth. Market data may be received from remote or local databases and / or forecast services, and may be included as a component of the data resources 26. Market data may also be received via a communication link with the load distribution authority 24. In accordance with aspects of the present invention, market data includes historical, present, and / or predicted data describing market conditions for the power plant 12a, which include, for example, energy sales prices, fuel costs, labor costs, and so forth. Operating data may also be received from databases and / or predictive services, and may be included as a component of the data resources 26. Operating data may include data collected by a plurality of sensors 46 operating within the power plant 12 and its plant components 49 which measure physical parameters related to plant operation. Operating data may include historical, present, and / or predicted data, as well as a variety of process inputs and outputs.
As can be seen in Fig. 9, an initial set point for the power plant 12 may be determined, such as e.g. with a control model III in the system controller 22 of FIG. 8. For example, the control model III may be used to utilize thermodynamic and / or physical details of the power plant 12 and additional information, such as power. Environment data or market data or process data, for determining a value of an operating parameter for the power plant 12 to be configured (at step 172 of FIG. 9). For example, in one case, the value of an operating parameter may be a value that would be required to achieve a power output sufficient to meet a target load. The particular value may be used as an initial setpoint for the corresponding operating parameter of the power plant 12 (also step 172 of FIG. 9). It will be understood that examples of such operating parameters may include: fuel flow, firing temperature, location for the swirl throttles (if vanes are present), vapor pressure, steam temperature, and vapor flow. A performance indicator may then be determined (at step 173 of FIG. 9) using a performance model 112 of the asset controller 22. The performance indicator may have an operational characteristic, such as Efficiency of the power plant 12. The performance model 112 may be configured to use thermodynamic and / or physical details of the power plant 12 and the setpoint determined by the control model 111 to determine a value of an operating characteristic of the power plant 12. The performance model 112 may be used to consider additional information, such as e.g. Environmental conditions, market conditions, process conditions and / or other relevant information to be configured.
Furthermore, according to certain aspects of the present invention, an estimate of the life cycle cost (LCC) of the power plant 12 may be made (at step 174 of FIG. 9), e.g. with an LCC model 113 included in the plant controller 22 of FIG. The LCC model 113, which may be a computer program or the like, may be configured to use physical and / or cost information about the power plant 12 as well as setpoints from the control model 111 to determine the estimated life cycle cost of the power plant 12 , Life cycle costs may include, for example, total cost, maintenance costs and / or operating costs of the power plant 12 over its lifetime. The LCC model 113 may also be configured for increased accuracy to account for the results of the performance model 112. The LLC model 113 may therefore use the determined setpoint values of the control model 111 and the operating characteristic from the performance model 112 as well as other information as desired to estimate the operating life of the power plant 12 and how much it costs the power plant 12 during its operational life to operate and / or to maintain. As noted above, the operating life of a power plant may be expressed in terms of operating hours and / or number of starts, and a given power plant has an expected operating life that may be provided by a manufacturer of the power plant. Thus, predefined values of the expected operating life may be used at least as a starting point for the LCC model 113 and / or an expansion module 114.
Using information from other embodiments of the invention, e.g. Results from determining an initial setpoint, a performance indicator, and an estimated life cycle may resolve an optimization problem for the power plant 12 (at step 175) as described below. Such an optimization problem may involve multiple equations and variables depending on a desired depth of analysis, and may include an objective function, which in certain embodiments may be an LCC-based objective function. The solution may include providing an improved or increased operating parameter of the power plant 12, such as by minimizing an LCC based target function (also step 175). In embodiments, the solution to the optimization problem may be provided by an expansion module 114 of the plant controller 22 of FIG. 8.
As is known from optimization theory, an objective function represents a property or parameter to be optimized, and may take into account many variables and / or parameters, depending on how the optimization problem is defined. In an optimization problem, an objective function can be maximized or minimized, depending on the particular problem and / or parameter represented by the objective function. For example, as stated above, according to embodiments, an objective function expressing LCC would be minimized to generate at least one operating parameter that may be used to operate the power plant 12 to keep the LCC as low as feasible. An optimization problem for the power plant 12, or at least an objective function, may include such factors as power plant characteristics, location parameters, customer specifications, results from the control model III, performance model 112 and / or LCC model 113, environmental conditions, market conditions and / or process conditions, and any additional information, which may be suitable and / or desired. Such factors may be collected in terms of an objective function such that, for example, an LCC-based objective function includes maintenance costs and operating costs over time, the time being a prediction horizon based on an estimated component service life. It will be understood that complex objective functions and / or optimization problems may be used in implementations of the present invention, as each may include many or all of the various functions and / or factors described herein.
For example, maintenance costs may be estimated by modeling parts of power plant 12 for estimating wear based on various parameters, such as: already discussed. It will be understood that for this purpose each part of the power plant 12 can be modeled. In a practical application, however, the parts may be modeled in connection with fewer, larger sections or fewer, selected portions of the power plant 12 and / or constant or connected values may possibly be used for some parts instead of the modeling. No matter what level of detail is employed, the minimization of such LCC-based objective function is part of an optimization problem that may arise for a given power plant due to many factors, such as: may vary, and may include at least one improved or increased operating parameter of the power plant 12, such as, for example, in accordance with the minimization of the LCC. In addition, those skilled in the art will recognize that at least one constraint can be imposed on the optimization problem, such as e.g. a predefined operating time and / or downtime, a predefined maximum and / or minimum temperature at various locations in the power plant 12, a predefined torque, a predefined power output, and / or other limitations, as desired and / or appropriate. Unless otherwise stated, it is within the purview of those skilled in the art to determine what limitations should be applied in what manner for a given optimization problem. Further, those skilled in the art will recognize situations in which additional optimization theory techniques may be employed, such as e.g. the addition of a slip variable to allow a feasible solution to the optimization problem.
Known techniques, such as e.g. by the expansion module 114 (FIG. 8) may be used to solve an optimization problem for the operation of the power plant 12. For example, as appropriate and / or desired, integer programming, linear, mixed integer linear, mixed integer nonlinear, and / or another technique may be used. Furthermore, as can be seen in the example objective function, the optimization problem can be solved over a prediction horizon, providing an array of values for at least one operating parameter of the power plant 12. While improvement or enhancement can be made over a relatively short prediction horizon, such as e.g. 24 hours, or even on the order of minutes, extension module 114 (Figure 8) may use a longer prediction horizon, such as, depending on a desired depth of analysis. to an estimated operating life of the power plant 12. In embodiments, initial setpoints such as, for example, determined by the control model 111 (FIG. 8) may be adjusted in response to and / or as part of the solution to the optimization problem to obtain an improved or increased or optimized setpoint. In addition, upon determining an initial setpoint, determining a value of a performance indicator, determining estimated LCC costs, and improving or increasing (in step 172-175 of FIG. 9) iteration may be used to refine the results and / or control setpoints of the power plant 12 better increase or increase.
As will be described, an offer curve section 180 may produce a supply curve or set of offer curves, an example of which has been previously shown with reference to FIG. In the plant controller 22, control information 115 may be received by the control module 110 and / or data resources 26 through a supply curve module 120 (in step 181 of FIG. 9). According to certain embodiments, the control information 115 includes control setpoints, power, environmental conditions, and / or market conditions. This information may also be known as "history" information. In addition, an environmental condition prediction 121 and / or market condition prediction 122 may be received (at step 182). According to certain embodiments, a database 123 may be included and may store up-to-date information, history information, and / or historical information locally, including all environmental conditions, market conditions, power plant performance information, offer curves, control setpoints, and / or any other information that may be appropriate can. The database 123 may be used to provide information for simulating the operation of the power plant 12 (in step 183), such as in the example of FIG. with an offline model 124 of the power plant 12.
The offline model 124 may include a model similar to the control model 111, but may also include additional modeling information. For example, the offline model 124 may include portions or the entirety of the control model 111, performance model 112, LCC model 113, and / or additional modeling information. By performing the off-line model 124 with setpoints and / or information from improving or increasing the LCC, the output of the offline model 124 may be used to provide estimated values for the cost of power production for each time interval in a prediction horizon and for different ones Power output values of power plant 12 may be determined (at step 184) to generate one or more supply curves 125 that may be sent to or otherwise provided to load sharing authority 24 (at step 185). The offline model 124 may contain any suitable information, such as Use historical, current and / or predicted information in determining estimated operating costs and / or conditions of the power plant 12. In addition, the offline model 124 may be tuned in embodiments (at step 186), such as e.g. For example, tuning may include periodically adjusting parameters for the offline model 124 based on information received from and / or provided by other portions of the plant controller 22 about the actual operation of the power plant 12 to better reflect and simulate the operation of the power plant 12 better. Thus, for a given set of operating parameters, if the plant controller 12 monitors an actual process condition that is different from what the offline model 124 had predicted, the plant controller 12 may change the offline model 124 accordingly.
In addition to the offer curves 125 from the power plant 12a, as illustrated, the load distribution authority 24 may receive supply curves 125 from other power plants 12b under its control. The load distribution authority 24 may evaluate the supply curves 125 and may create a load-sharing schedule to distribute the load in the power system 10. The load balancing authority 24 may also, as appropriate and / or desirable, take into account predicted environmental conditions, load prediction, and / or other information that it may receive from various local or remote data resources 26 to which it has access. As illustrated, the load-sharing schedule established by the load-sharing authority 24 includes a control signal for the power plant 12 that includes a target load 128 to which the plant controller 22 may respond as described above.
It will be understood that the inclusion of life cycle cost considerations, as described herein, may serve to increase the scope and accuracy of the plant models used in the optimization process, thereby enabling improvements to the process. The supply curves 125, as described above, may represent variable costs (measured in dollars per megawatt hour versus power plant output in megawatts). The supply curves 125 may include a supply curve of the incremental variable cost and a supply curve of the average variable cost. As can be seen, embodiments of the present invention can provide accurate variable cost assessments over their generated offer curves 125. Using embodiments of the present invention, the incremental variable cost supply curves have been shown to predict very accurately the actual incremental variable cost curves, while the average variable cost supply curves have been shown to accurately predict the actual average variable cost curves. The accuracy of the offer curves established by embodiments of the present invention indicates that the various models used in the plant controller 22 of FIG. 8 provide a suitably representative model for the purposes set forth.
[0070] Referring to Figures 10 through 12, further aspects of the present invention will be described with reference to and including certain systems and methods provided above. 10 is a data flow diagram demonstrating architecture for a plant optimization system 200 that may be used in a gas and steam turbine power plant combined cycle power plant. In the embodiment provided, a system 200 includes monitoring and control instruments 202, 204, such as those shown in FIG. the above-discussed sensors and actuators associated with both the gas turbine (202) and steam turbine systems (204). Each of the monitoring and control instruments 202, 204 may transmit signals indicative of measured operating parameters to a plant controller 208. The plant controller 208 receives the signals, processes the signals in accordance with predetermined algorithms, and transmits control signals to the monitoring and control instruments 202, 204 to effect changes in plant operations.
The plant controller 208 interfaces with a data acquisition module 210. The data acquisition module 210 may be communicatively coupled to a database / historian 212 that maintains archive data for future reference and analysis. A heat balance module 214 may as requested receive data from the data acquisition module 210 and the database / historian 212 to process algorithms that tune a power and energy balance model of the power plant to be as close as possible to measured data. Discrepancies between the model and the measured data may indicate errors in the data. As will be understood, a power module 216 may use plant equipment models to predict the expected performance of main plant components and equipment. The difference between expected and current performance may represent deterioration in the condition of plant equipment, parts, and components, such as, but not limited to, contamination, encrustation, corrosion, and fractures. In accordance with aspects of the present invention, the power module 216 may track degradation with time so that performance issues that have the most significant impact on system performance are identified.
As illustrated, an optimizer module 218 may be included. The optimizer module 218 may include a methodology for optimizing economic load distribution of the plant. For example, according to embodiments, the load distribution to the power plants may be based on the specific heat consumption or the incremental specific heat consumption based on the assumption that the specific heat consumption is equivalent to monetary resources. In an alternative scenario in which the power plant includes an additional manufacturing process (not shown) for which steam is used directly (ie where the generated steam can be diverted from power generation in the steam turbine to another production use), it will be understood that the optimizer module 218 may solve an optimization problem whereby a component having a higher specific heat consumption may be assigned. For example, in certain situations, a demand for steam may develop faster than a demand for electricity, or electrical power may be limited by electrical system requirements. In such cases, assigning a gas turbine engine with lower efficiency may allow more heat to be gained without raising electrical power beyond a limit. In such scenarios, the allocation of the component with a higher specific heat consumption is the economically optimized alternative.
The optimizer module 218 may be selectable between an online (automatic) and an offline (manual) mode. In online mode, optimizer 218 automatically calculates current economic asset parameters, such as Cost of electricity generated, incremental cost at each generation level, cost of process steam and plant operating revenue at a predetermined periodicity, for example in real time or once every five minutes. An off-line mode can be used to simulate continuous power, analyze what-if scenarios, analyze budget and upgrade options, and current power generation capability, specific target heat usage, and how to validate current asset operation for warranty conditions to predict the impact of operating restrictions and maintenance operations and fuel consumption. Optimizer 218 calculates yield optimized power for the power plant based on real-time economic cost data, output prices, load levels, and equipment wear rather than efficiency-based performance by combining plant heat balances with a plant financial model. The optimizer 218 may be tuned to account for the degradation of each component individually, and may generate an advisory output 220 and / or may generate a closed feedback loop control output 222. The Advisory Issue 220 recommends operators to set controllable parameters of the power plant to optimize each plant component and to maximize profitability. In the exemplary embodiment, the advisory output 220 is a computer display screen that is communicatively coupled to a computer-executing optimizer module 218. In an alternative embodiment, the advisory output is a remote workstation screen, with the workstation accessing the optimizer module 218 via a network. The closed feedback loop control output 222 may receive data from the optimizer module 218 and calculate optimized setpoints and / or presets for the modules of the system 200 to implement real time feedback control.
FIG. 11 is a simplified block diagram of a real-time optimization system for a thermal power plant 230 which, in accordance with aspects of the present invention, includes a server system 231 and a plurality of client subsystems, also referred to as client systems 234, communicatively coupled to the client Server system 231 are coupled. As used herein, real-time refers to events that occur in a substantially short period of time after a change in the inputs affects the result, for example computer calculations. The period represents the amount of time between each iteration of a regularly repeated task. Such repeated tasks may be referred to herein as periodic tasks or cycles. The period is a design parameter of the real-time system that may be selected based on the meaning of the result and / or the ability of the system implementing the processing of the inputs to generate the result. In addition, events that take place in real time take place without a significant deliberate delay. In the exemplary embodiment, calculations may be updated in real time with a periodicity of one minute or less. The client systems 234 may be computers that include a web browser such that the server system 231 is accessible to the client systems 234 via the Internet or other network. The client systems 234 may be connected to the Internet via many interfaces. The client systems 234 could be any device that is capable of connecting to the Internet. A database server 236 is connected to a database 239 which contains information regarding multiple issues, as described in more detail below. In one embodiment, a centralized database 239, which includes aspects of the data resources 26 discussed above, is stored on the server system 231, and potential users on one of the client systems 234 can access them by logging in to the server system 231 via the client systems 234. In an alternative embodiment, database 239 is stored remotely from server system 231 and may be remote.
[0075] In accordance with aspects of the present invention, certain of the control methods discussed above may be developed for use in conjunction with the system diagrams of FIGS. 10 and 11. For example, one method includes simulating power plant performance using a plant power module of a software code segment that receives data from the power plant monitoring instrument. The data may be received over a network from a plant controller or a database / historian software program running on a server. Any additional equipment components, such as An intake conditioning system or HRSG duct combustion system may be simulated in a similar manner as power plant performance simulation. Determining the performance of each plant component in the same manner allows treatment of the entire power plant as a single plant to determine optimized setpoint values for the power plant rather than separately determining such setpoints for each component separately. Measurable quantities for each plant component can be parameterized to express performance or power plant efficiency on a component-by-component basis. Parameterizing plant equipment and plant performance involves calculating the efficiency for components such as, but not limited to, a gas turbine compressor, gas turbine, HRSG, blower, cooling tower, condenser, feedwater heater, evaporator Similarly, it will be understood that specific heat consumption and power calculations can be parameterized and the resulting simultaneous equations solved in real time so that the calculated results are available without any intentional delay from the time each parameter was acquired , Solving parameterized simultaneous equations and constraints may also include determining a current heat balance for the power plant and determining an expected power using constraints in the operation of the power plant such as, but not limited to, reserve power requirements, electrical system requirements, maintenance activities, fresh water demand, and Component failures include. Solving parameterized equations and constraints may also include determining parameters to adjust and modify the current heat balance such that a future heat balance equals the determined expected performance. In an alternative embodiment, solving parameterized simultaneous equations and constraints includes determining inlet conditions to the power plant, predicting power plant performance based on the determined inlet conditions and a predetermined power plant model, determining current power plant performance, comparing the predicted ones Performance at the specified power and adjusting system parameters until the particular power equals the predicted power. In exemplary embodiments, the method also includes correlating controllable plant parameters, plant equipment, and plant performance using parameterized equations, defining the goal of the optimization using an objective function that includes minimizing the specific heat consumption of the power plant and / or maximizing the yield of the power plant and defining the physically possible range of operation of each individual piece of equipment and / or the total limits using constraints, the total limits counting maximum power production, maximum fuel consumption, and so on.
FIG. 12 is a flowchart of an example method 250 for solving parameterized simultaneous equations and constraints in accordance with the present invention. The method 250 includes determining (at 252) a current heat balance for the power plant, determining (at 254) an expected power using current limitations of operation, and determining (at 256) parameters to adjust and modify the current heat balance, such that a future heat balance is equal to the particular expected performance. The method 250 also includes determining power plant inlet conditions 258, predicting power plant performance 260 based on the determined intake conditions and a predetermined power plant model, determining current power plant power 262, comparing the predicted power with the determined power 264 Power and adjusting 266 plant parameters until the particular power equals the predicted power. It will be understood that the described method and systems discussed with respect to FIGS. 10 and 11 provide a cost effective and reliable means of optimizing CCPPs.
Referring to FIGS. 13-16, several flowcharts and system configurations illustrating a control methodology according to certain aspects of the present invention are considered. In general, according to an example embodiment, a control system for a thermal power plant, such as e.g. the gas turbine system, or a power plant, may include first and second examples of a model that models the operation of the turbines, such as a turbine; by exploiting physics-based models or mathematical modeling (e.g., transfer functions, etc.). The first model (which may also be referred to as the "primary model") may provide present operating parameters of the gas turbine system describing the turbine operating mode and the corresponding operating conditions. As used herein, "parameters" refer to points that may be used to define the operating conditions of the turbine, such as, but not limited to, temperatures, pressures, gas flows at defined locations in the turbine, compressor, combustor and turbine efficiencies, etc. Performance parameters may also be referred to as "model correction factors" which refer to factors used to adapt the first or second model to reflect the operation of the turbine. Inputs to the first model may be captured or measured and provided by an operator. In addition to current performance parameters, the method of the present invention may include receiving or otherwise obtaining information about external factors or disturbances such as, for example, Environmental conditions, which may affect the current or future operation of the gas turbine system.
The second model (also referred to as a "secondary model" or a "predictive model") is constructed to provide one or more operational parameters, such as a. To identify or predict the control variables of the gas turbine system, taking into account the existing operating parameters, such as. Actuating variables, and the one or more disturbances. Exemplary operating parameters of the turbine include, but are not limited to, actual turbine operating conditions, such as those shown in FIG. Exhaust gas temperature, turbine power, compressor pressure ratios, specific heat consumption, emissions, fuel consumption, expected yields, and the like. Therefore, this second or predictive model can be used to indicate or predict turbine behavior at certain operating setpoints, performance goals, or operating conditions that differ from the present operating conditions. As used herein, the term "model" generally refers to the act of modeling, simulating, predicting or indicating based on the output of the model. It is understood that while the term "second model" is used herein, in some cases there may not be any difference between the formulation of the first and second models, so that the "second model" means performing the first model with adjusted parameters or a second model represents additional or different input.
Accordingly, by modeling the turbine performance using the second or predictive model taking into account external factors and / or different operating conditions, the turbine controller may be adjusted to operate more efficiently under these different operating conditions or in view of the unanticipated external factors. This system therefore allows for automated turbine control based on modeled behaviors and operational characteristics. In addition, the modeling system described allows for the creation of operator-specified scenarios, inputs, operating points, operating goals, and / or operating conditions for predicting turbine behavior and operating characteristics under these operator-specified conditions. The prediction of such hypothetical scenarios allows operators to make more informed control and operational decisions, e.g. Scheduling, Load, Partload, etc. As used herein, the term "operating points" generally refers to operating points, conditions, and / or objectives, and is not intended to be limiting. Thus, an operating point may refer to a target or a setpoint, such as e.g. Base load, partial load point, top fire and the like.
An example use of the turbine modeling system described involves adjusting the turbine operation to meet grid satisfaction requirements while still operating at the most efficient level. For example, regional network authorities typically dictate requirements that power generation facilities must be able to support a network in the event of frequency noise. Supporting the network in the event of disturbances involves increasing or decreasing the turbine load under certain conditions, depending on the network condition. For example, in the event of a failure, a power plant is expected to increase its power generation capacity (e.g., up to 2%) to compensate for other delivery shortfalls. Thus, turbine operation usually restricts the base load point to allow the turbine to operate at a limited power level (also referred to as the "reserved margin") so that the increased load can be provided if needed without the additional maintenance factor in the engine Connection with overfeed arises. As an example, the reserved margin may be 98% of what would normally be the base load, which would allow increasing the load to meet network requirements (e.g., increase by 2%) without exceeding the 100% base load. However, unanticipated external factors, e.g. Temperature, humidity or pressure, have a negative impact on turbine efficiency. If temperatures rise sharply in a day, a turbine may no longer have that 2% reserve it needs because the heat has caused the turbine to run less efficiently and the turbine can not reach that 100% load as originally planned , To compensate for this, given the possible loss of engine efficiency (e.g., 96%, etc.), conventional specific heat consumption curves cause the turbine to operate more efficiently throughout the day. However, the turbine modeling system described herein allows modeling of turbine behavior in real time according to the current external factors (e.g., temperature, humidity, pressure, etc.) and thus controlling turbine operation for the most efficient operation in the current environmental conditions. Similarly, future turbine behavior may be predicted, such as e.g. predicting turbine behavior in response to heat fluctuation during a day, thereby allowing turbine operations planning to achieve the most efficient and economical feasible operation. As another example, power plants typically make decisions about whether to shut down gas turbines at night or simply to lower power levels (e.g., partial load). The turbine operating characteristics, such as Emissions, exhaust gas temperature and the like have an influence on this decision. Utilizing the turbine modeling system described herein, decisions can be made on a smarter basis, either in advance or in real time, or near-real time. External factors and expected turbine operating parameters may be provided to the second model to determine what the turbine operating characteristics would be. Thus, the modeled properties can be used to determine whether a turbine should be shut down or operated at part load taking into account these characteristics (e.g., efficiency, emissions, costs, etc.).
As yet another example, a turbine modeling system may be utilized to evaluate the benefits of performing turbine maintenance at a given time. The turbine modeling system of the present invention may be used to model the operating characteristics of the turbine at its current capabilities based on current performance parameters. Then, a user-specified scenario may be generated that models the operating characteristics of the turbines if maintenance is performed (e.g., improving the performance parameter values to indicate an expected performance boost). For example, as the turbines degrade over time, the performance parameters reflect engine wear. In some cases, maintenance may be performed to improve these performance parameters and thus the operating characteristics of the turbine. By modeling or predicting the improved operating characteristics, a cost-benefit analysis can be performed to compare the benefits gained from performing maintenance against the costs incurred.
FIG. 13 illustrates an exemplary system 300 that may be used to model turbine performance. According to this embodiment, a power plant 302 is provided, which has a gas turbine with a compressor and a combustion chamber. An inlet channel to the compressor supplies ambient air and possibly injected water to the compressor. The configuration of the intake passage contributes to a pressure loss of the ambient air flowing into the compressor. An exhaust duct for the power plant 302 directs combustion gases from the outlet of the power plant 302, for example, through emission control and sound attenuation devices. The amount of inlet pressure loss and backpressure may vary over time due to the addition of components to the inlet and exhaust air ducts as well as blockage of the inlet and exhaust air ducts.
The operation of the power plant 302 may be monitored by one or more sensors sensing one or more detectable conditions or operating or performance parameters of the power plant 302. In addition, external factors, such as the surrounding environment, be measured by one or more sensors. In many cases, two or three redundant sensors can measure the same parameter. For example, groups of redundant temperature sensors may monitor the ambient temperature surrounding the power plant 302, the compressor exit temperature, the turbine exhaust temperature, and other temperatures throughout the power plant 302. Similarly, groups of redundant pressure sensors can monitor ambient pressure and static and dynamic pressure levels at the compressor inlet and outlet, at the turbine outlet, and at other locations throughout the engine. Groups of redundant moisture sensors can measure the ambient humidity in the inlet channel of the compressor. Groups of redundant sensors may also include flow sensors, velocity sensors, flame sensor sensors, valve position sensors, vane angle sensors, or the like, which detect various parameters relevant to the operation of the power plant 302. A fuel control system may regulate the fuel flowing from a fuel supply to the combustor. The fuel controller may also select the type of fuel for the combustion chamber.
As noted, "operating parameters" refer to points that may be used to define the operating conditions of the turbine system, such as those described in US Pat. Temperatures, pressures, compressor pressure ratios, gas flows at defined positions in the turbine, load setpoint, firing temperature, and one or more conditions that correspond to the extent of the turbine or compressor wear and / or the level of turbine or compressor efficiency. Some parameters are measured directly. Other parameters are estimated by the turbine models or are indirectly known. Still other parameters may represent hypothetical or future conditions and may be defined by the plant operator. The measured and estimated parameters may be used to represent a given turbine operating condition. As used herein, "performance indicators" are operating parameters that are derived from the values of certain measured operating parameters and are a performance criterion for the operation of the power plant over a defined period of time. For example, counters include specific heat consumption, output level, etc.
As illustrated in FIG. 13, the system 300 includes one or more controllers 303a, 303b, each of which may be a computer system having one or more processors that execute programs to control the operation of a power plant or block 302 , Although FIG. 13 illustrates two controllers, it will be understood that a single controller 303 may also be provided. According to a preferred embodiment, multiple controllers may be included to provide redundant and / or distributed processing. For example, the control actions may depend on sensor inputs or instructions from plant operators. The programs executed by the controller 303 may include scheduling algorithms, such as e.g. those for regulating fuel flow to the combustor, managing network requirements, part load, etc. The commands generated by the controller 303 may cause actuators on the turbine, for example, to adjust valves between the fuel supply and combustion chambers to control fuel flow, partitions, and to regulate the type of fuel. The actuators can adjust swirl chokes on the compressor or activate other control setpoints on the turbine. It will be understood that the controller 303 may be used to create the first and / or second model as described herein, in addition to simplifying control of the power plant. The controller 303 may receive user-specified and / or present modeled outputs (or any other system outputs). As previously described, the controller 303 may include a memory that stores programmed logic (e.g., software) and may store data, such as memory. detected operating parameters, modeled operating parameters, operating limits and targets, operating profiles and the like. A processor may use the operating system to execute the programmed logic and thereby also use data stored thereon. Users receive interfaces with the controller 303 via at least one user interface device. The controller 303 may be in communication with the power plant online while it is operating via an I / O interface and may also be offline in communication with the power plant while it is not operating. It will be understood that one or more of the controllers 303 may perform the implementation of the model-based control system described herein which may include, but is not limited to: acquiring, modeling, and / or receiving operating parameters and performance parameters; Creating a first power plant model that reflects the current turbine operation; Collecting, modeling and / or receiving information about external factors; Receiving operator input, e.g. Performance goals and other variables; Creating a second power plant model that reflects operation in light of the additional data provided; Controlling the current or future turbine operation and / or presenting modeled operational characteristics. It should also be understood that other external devices or multiple other power plants or power plant blocks may communicate with the controller 303 via I / O interfaces. The controller 303 may be located remotely with respect to the power plant that it controls. Further, the controller 303 and the programmed logic implemented thereby may include software, hardware, firmware, or any combination thereof.
The first controller 303a (which, as indicated, may be the same or a different controller than the second controller 303b) may be suitable for modeling the power plant 302 by a first or primary model 305, including modeling the current performance parameters of the model Turbine. The second controller 303b may be adapted via a second or predictive model 306 to model turbine operating characteristics under different conditions. The first model 305 and the second model 306 may each be an arrangement of one or more mathematical representations of the turbine behavior. Each of these representations may be based on input values to produce an estimated value of a modeled operating parameter. In some cases, the mathematical representations may generate a replacement operating parameter value that may be used in cases where no measured parameter value is available. The first model 305 may then be used to provide a baseline and / or input to the second model 306 for determining turbine operating characteristics based on the current performance parameters of the power plant 302 and any other factors, such as power plant 302. external factors, operator supplied commands or conditions, and / or adjusted operating conditions. As described above, it is understood that "the second model 306" may simply be an example of the same model as the first model 305, which takes into account additional or different inputs, such as e.g. external factors, different operating points to model different performance parameters or turbine behavior given the different inputs. The system 301 may further include an interface 307.
With further reference to Figure 13, a brief description of the interrelationship between the system components is provided. As described, the first or primary model 305 models current performance parameters 308 of the power plant 302. These current performance parameters 308 may include, but are not limited to, conditions that correspond to the level of turbine wear, conditions that correspond to the level of turbine efficiency (eg specific heat consumption or fuel to power output ratio), swirl throttle angle, amount of fuel flow, turbine rotational speed, compressor inlet pressure and temperature, compressor outlet pressure and temperature, turbine exhaust temperature, generator power output, compressor airflow, combustor fuel / air ratio, firing temperature (turbine inlet) , Combustion chamber flame temperature, fuel system pressure ratios and acoustic properties. Some of these performance parameters 308 may be measured or captured directly from turbine operation, and some may be modeled based on other measured or sensed parameters. The performance parameters may be provided by the first model 305 and / or may generally be provided by the controller, e.g. if they are detected and / or measured by the controller. When creating the first model 305, the performance parameters 308 (which are to be referenced by the model for any turbine behavior) are provided for constructing the second or predictive model 306. Depending on its purpose, other variables 309 may also be provided to the second model 306. For example, the other variables may include external factors, e.g. Environmental conditions that are generally uncontrollable and that are easy to accept. In addition, to the other variables 309, a controller-specified scenario or operating point (eg, a turbine operating point generated by or otherwise provided via the controller 303, such as turbine control based on the first model 305, etc.) may be measured inputs, which may be some or all of the same measured inputs may count, possibly as described as modeled by the first model 305. As described below with reference to Figure 14, an operator-specified scenario 313 (e.g., one or more operator-provided instructions that specify different turbine operating points or conditions) may also be provided to the second model 306 via operator input. For example, as an example use, the other variables 309 may include a controller-specified scenario provided as one or more inputs to the second model 306 when attempting real-time or near-real-time actual turbine behavior based on additional inputs , such as external factors or measured inputs. By utilizing a controller-specified scenario of the first model in addition to one or more of these additional inputs, the expected real-time behavior of the power plant 302 may be modeled by the second model 306 taking into account these additional inputs, which in turn may be used to power plant 302 or to adapt the first model 305 by control profile inputs 310.
Referring to FIG. 14, a user-specified operating mode or scenario 313 is provided as one or more inputs via the interface 307 to the second or predictive model 306, which then models or predicts future turbine behavior under a variety of conditions. For example, an operator may provide commands to the interface 307 to create a scenario in which the power plant 302 operates at a different operating point (e.g., different loads, configuration, efficiency, etc.). As an illustrative example, a set of operating conditions may be provided via the operator-specified scenario 313, which may represent conditions expected for the following day (or another future time frame), such as the following: Environmental conditions or demand requirements. These conditions may then be used by the second model 306 to generate expected or predicted turbine operating characteristics 314 for the power plant 302 during this time frame. In executing the second model 306 under the operator-specified scenario, the predicted operating characteristics 314 represent turbine behaviors such as, but not limited to, base load output capability, peak output capability, minimum component load points, emission levels, specific heat consumption, and the like. These modeled or predicted operational characteristics 313 may be useful in the planning of and commitment to power generation levels, such as, for example, for day-ahead market planning and offers.
Fig. 15 illustrates an example method 320 by which an embodiment of the invention may operate. Provided is a flow chart of the basic operation of a system for modeling a turbine, as may be performed by one or more controllers, such as a turbine. by referring to FIGS. 13 and 14. The method 320 may begin at step 325 where the controller, through a first or primary model, may model one or more current performance parameters of a turbine according to the current operation. To create this first model, the controller may receive one or more operating parameters as inputs to the model indicating the current operation of the turbine. As described above, these operating parameters may be detected or measured and / or they may be modeled, e.g. can take place if the parameters can not be detected. The current operating parameters may include any parameters indicative of the current turbine operation, as described above. It is understood that the methods and systems disclosed herein do not directly depend on whether the operating parameters are measured or modeled. For example, the controller may include a created model of the gas turbine. The model may be an arrangement of one or more mathematical representations of the operating parameters. Each of these representations may be based on input values for generating an estimated value of a modeled operating parameter. The mathematical representations may generate a replacement operating parameter value that may be used in cases where no measured parameter value is available.
In step 330, the controller may receive or otherwise determine one or more external factors that may affect the current and / or future operation. As described above, these external factors are usually (but need not be) uncontrollable, and therefore the inclusion of their influence in the second model is beneficial in providing the desired turbine control profile and / or performance. External factors may include, but are not limited to, ambient temperature, humidity or air pressure, and fuel consumption and / or supply pressure, which may affect turbine performance. These external factors may be measured or detected, estimated or otherwise manually provided by an operator (such as when the operator requests predicted behavior based on hypothetical scenarios or future conditions), and / or may be provided by third-party information sources (eg, weather service, etc.). ,
At step 335, the controller may receive adjusted operating points and / or other variables for predicting turbine behavior at a condition different from the current turbine condition. Adjusted operating points may include, but are not limited to, identifying the desired output level, such as the desired output level. when modeling the turbine with a reserved margin (e.g., 98% of the base load) or when modeling the turbine with, for example, a peak load or at part load. Operating limits may also include operating limits such as, but not limited to, hot gas trailability (or firing temperature), exhaust friability, NOx emissions, CO emissions, combustor lean extinction limit, combustion dynamics, compressor surge, compressor frosting, aeromechanical compressor limits, compressor clearances, and compressor outlet temperature. Thus, by providing these adjusted operating points or other variables, the operator may provide hypothetical scenarios for which the turbine model predicts the operating characteristics under those scenarios, which may be useful for controlling the future operation of the turbine and / or for planning future power generation and obligations can.
Step 335 is followed by step 340, in which a second or predictive model of the turbine based on the first model created in step 325 and, optionally, the external factors and / or adjusted operating points or other variables that are described in Step 335 is provided is created. This second or predictive model may thus accurately specify or predict operating parameters and therefrom performance indicators for the turbine during a future operating period.
In step 345, the modeled power may be used to adjust the current or future turbine operation and / or to display the modeled power to an operator. Accordingly, if the current turbine operation is being adjusted, the turbine controller may receive the modeled performance parameters as inputs to change a current control model (e.g., the first model) or a current control profile, such as a control model. by modifying various setpoints and / or references used for the current turbine control. It is anticipated that this real-time or near real-time control of the turbine would be performed if the inputs to the second model made in step 340 are representative of the current turbine conditions or current external factors. For example, a real-time or near real-time adjustment may be made in step 345 if the second model represents performance characteristics that take into account the current temperature, pressure, or humidity and / or considers turbine operating parameters or performance parameters that / detail the turbine wear and / or efficiency. Fig. 16 describes an example embodiment that may optionally receive operator-specific inputs and generate predicted behavior under a different operating condition. The output of the model created in step 340 may also be displayed to an operator via an interface or otherwise presented. For example, in an embodiment where the operator provides hypothetical operating scenarios in step 335, the predicted turbine operating characteristics may be displayed for analysis and possible inclusion in future control or scheduling activities. Accordingly, the method 320 may terminate after step 345 after the turbine's current performance parameters have been modeled by a first model and then the same turbine has been modeled considering additional external factors, adjusted operating points, or other additional turbine operation prediction data based on that additional data ,
FIG. 16 illustrates an example method 400 by which an alternative embodiment may operate. Provided is an example flow diagram of the operation of a system for modeling a turbine, such as may be performed by one or more controllers, such as a turbine. with reference to Figs. 13 and 14. The method 400 illustrates the use of the system 301 in which an operator may optionally provide additional variables to use the modeling capabilities to predict turbine behavior under hypothetical scenarios. The method 400 may begin at decision step 405, where it is determined whether the turbine is to be modeled according to current turbine operating parameters and performance parameters, or whether parameters provided by the operator should be taken into account when creating the model. For example, if the system is used to predict hypothetical operating scenarios, then current performance parameters may not be needed as inputs to the model (assuming that the model already reflects the basic turbine operation and behavior). Accordingly, if it is determined in decision step 405 that no current parameters are to be utilized, operation proceeds to step 410, in which the operator provides different performance parameters that allow the turbine to be modeled at a different operating point and under a different operating condition (eg in a worse condition, at a different level of performance, etc.). Otherwise the current performance parameters and / or operating parameters are used, such as with reference to step 325 of FIG. 15, and operation proceeds to step 415. In step 415, the controller may model by a first or primary model one or more performance parameters of a turbine according to either the operator provided input from step 410 or the current turbine operation. For example, if the model is created based at least in part on parameters provided by the operator in step 410, the model created in step 415 is representative of predicted turbine behavior under those performance parameters.
Step 415 is followed by decision step 420, in which it is determined whether subsequent modeling (e.g., the "second model" or the "predictive model") should be based on current external factors, e.g. current temperature, pressure or humidity, or on various external factors provided by the operator. For example, in one scenario, the controller may model the turbine performance based on the additional one of the one or more current external factors, which would allow further prediction of turbine behavior in light of current conditions. However, in another scenario, the controller may be used to further model the turbine according to the operator provided conditions, allowing for predicting turbine operating characteristics under various hypothetical scenarios. Accordingly, if it is determined in step 320 that the data of the external factors provided by the operator are to be considered in the modeling, the operation proceeds to step 425. Otherwise, the operation proceeds to step 430 utilizing the current external factors. In step 430, the controller receives external factors to be considered in creating the second or predictive model, whether they represent the current state or hypothetical factors. Step 430 is followed by steps 435-445, which optionally include consideration of different operating points, constructing the predictive model based on the received data, and displaying the predicted behavior in the same or a similar manner as in relation to step 325-345 of Fig. 15. The method 400 may end after step 445 after the turbine operating behavior has optionally been modeled based on operator-specified scenarios.
Accordingly, embodiments described herein permit the use of turbine models to indicate turbine behavior and corresponding operating parameters of an actual turbine, in addition to predicting turbine behavior taking into account the current performance parameters and one or more identified external factors. These embodiments therefore provide a technical effect of indicating or predicting turbine behavior at operating points or under operating conditions that are different than current turbine operation. An additional technical effect is provided which allows the automated turbine controller to have turbine performance and operating characteristics based at least in part on modeled behaviors and operating characteristics, which may optionally include the creation of operator-specified scenarios, inputs, operating points, and / or operating conditions to predict user-specified conditions. Another implemented technical effect includes the ability to predict various hypothetical scenarios that allow the operator to make more informed control and operational decisions, such as those described in the following. As will be understood, reference is made herein to step-by-step diagrams of systems, methods, apparatus, and computer program products according to example embodiments of the invention.
Referring to Figure 17, a flow diagram 500 in accordance with an alternative embodiment of the present invention is illustrated. As will be appreciated, the flowchart 500 includes aspects that may be used as a control method or as part of a control system to facilitate the optimization of a power plant 501. Power plant 501 may be similar to any of those discussed with respect to FIGS. 2 and 3, however, unless otherwise limited in the appended claims, it should be understood that the present invention is also applicable to other types of power plants can come. In a preferred embodiment, the power plant 501 may include a plurality of thermal power plants that generate electricity generated within a power system market, such as a power plant. which is discussed with reference to FIG. 1. The power plant 501 may include many possible types of operating modes, including, for example, the different ways in which plant thermal power plants are integrated or operated, the output level of the plant, the ways in which the plant responds to changing environmental conditions, during load requirements It will be understood that the operating modes may be described and defined by operating parameters that consider physical characteristics of certain aspects of the operation of the power plant 501. As further illustrated in FIG. 17, the present invention may include a power plant model 502. The power plant model 502 may include a computerized representation of the power plant that correlates process inputs and outputs as part of a simulation that is intended to mimic plant operation. As shown, the present invention further includes a voting module 503; a plant controller 505; a tuned power plant model 507; a plant operator module 509; and an optimizer 510, which are discussed in detail below.
The power plant 501 may include the sensors 511 which measure operating parameters. These sensors 511, as well as the operating parameters that they measure, may include any of those already discussed herein. As part of the present method, the sensors 511 may make measurements of the operating parameters during an initial, current, or first period of operation (hereafter "first period of operation"), and these measurements may be used to tune a mathematical model of the power plant, which then, as below discussed as part of an optimization process for controlling power plant 501 in an improved or optimized manner of operation during a subsequent or second operating period (hereinafter "second operating period"). The measured operating parameters themselves can also be used to evaluate plant performance, or they can be used in calculations to derive performance indicators relating to specific aspects of the plant's operation and performance. As will be understood, performance indicators of this type may include specific heat consumption, efficiency, generation capability and others. Accordingly, as an initial step, operating parameters measured by the sensors 511 during the first period of operation may be used as one or more performance indicators (or to calculate values for them). As used herein, such values for performance indicators (i.e., those based on measured values of operating parameters) are referred to herein as "measured values." The measurements of the operating parameters and / or the measured values for the performance indicators may, as shown, be communicated to both the plant controller 505 and the voting module 503 (512). The reconciliation module 503, as discussed in more detail below, may be configured to calculate feedback from a reconciliation process for use in tuning the power plant model 502 to configure the tuned power plant model 507.
The power plant model 502 may, as discussed, be a computerized model configured to simulate the operation of the power plant 501. According to the present method, the power plant model 502 may be configured to simulate power plant operation corresponding to the first operating period of the power plant 501. To accomplish this, the power plant model 502 may be provided with information and data concerning the operating parameters of the first operating period. While this information may include all of the operating parameters measured during the first operating period, it will be appreciated that the input data to the power plant model 502 may be limited to a subset of the measured operating parameters. In this manner, the power plant model 502 may then be used to calculate values for selected operating parameters that were excluded from the input data set. More specifically, the power plant model may be provided with input data for the simulation that includes many of the values measured for the operating parameters, but from which certain measured values for selected operating parameters have been omitted. As an output, the simulation may be configured to predict a simulated value for the selected operating parameter. The present method may then use the simulated values to predict values for the performance indicators. In this case, these values for the performance indicators are referred to herein as the "predicted values". In this way, the measured values for the performance indicators, which were determined directly from measured power plant operating parameters, may have corresponding predicted values. As illustrated, the predicted values for the performance counters may be communicated to the voting module 503 (514).
The voting module 503 may be configured to compare the corresponding measured and predicted values for the performance indicators to determine a difference therebetween. As will be understood, the difference thus calculated reflects an error plane between the actual power (or measurements thereof) and the power simulated by the power plant model. Power plant model 502 may be tuned based on this difference or feedback 515. In this way, the tuned power plant model 507 is configured. The tuned power plant model 507, which may also be referred to as an off-line or predictive model, may then be used to determine optimized operating modes for a subsequent operating period by simulating proposed or possible operating modes. The simulations may include estimates or predictions about future unknown operating conditions, such as Environmental conditions include. As will be appreciated, optimization may be based on one or more performance goals 516 in which a cost function is defined. As illustrated, the performance targets 516 may be communicated to the optimizer 510 via the plant operator module 509.
The process of tuning the plant model may be configured as a repetitive process involving several steps. As will be appreciated, power plant model 502 may, according to certain embodiments, include algorithms in which logic inputs and / or parameterized equations correlate process inputs (i.e., fuel supply, air supply, etc.) with process outputs (generated electricity, plant efficiency, etc.). The step of tuning the power plant model 502 may include adjusting one of the algorithms in the power plant model 502 and then simulating the operation of the power plant 501 for the first operating period using the adjusted power plant model 502 to determine the effect of the adaptation. More specifically, the predicted value for the performance indicator may be recalculated to determine the effect of adapting the power plant model to the calculated difference. If the difference turns out to be lower using the adjusted power plant model 502, the power plant model 502 may be updated or "tuned" to further include the fit. It will also be understood that the power plant model 502 may be constructed with multiple logic statements that include power multipliers used to reflect changes in the way the power plant operates under certain conditions. In such cases, tuning the power plant model 502 based on the calculated difference may include the steps of: a) making adjustments to one or more of the power multipliers; b) simulating the power plant operation for the first period of operation with the power plant model 502 with the adjusted power multiplier; and c) recalculating the predicted value for the performance indicator using the power plant model 502 as adjusted by the power multiplier to determine whether the recalculation results in a decreased difference. These steps may be repeated until an adjustment made to one of the performance multipliers results in a reduction in the difference, which would indicate that the model is more accurately simulating actual performance. It will be understood, for example, that the power multiplier may refer to the expected performance degradation based on the accumulated operating hours of the equipment. In another example, in which the performance indicator includes generating capability, the step of tuning the power plant model 502 may include recommending adjustments to factors based on a difference between a measured generation capability and a predicted generation capability. Such adjustments may include changes that ultimately result in the predicted generation capability, which is substantially equal to the measured generation capability. Accordingly, the step of tuning the power plant model 502 may include modifying one or more correlations within the power plant model 502 until the predicted or simulated value for a performance indicator is substantially equal to (or within a margin of) the measured value for the performance indicator.
After the tuning, the method may then use the tuned model 507 to simulate a proposed operation of the power plant. According to certain embodiments, a next step of the present method includes determining which simulated operation is preferable in view of the defined performance goals 516. In this way, optimized operating modes of the power plant can be determined. According to a preferred embodiment, the process of determining an optimized mode of operation may include several steps. First, several proposed modes of operation can be selected or selected among the many possible ones. For each of the proposed modes of operation, corresponding suggested parameter sets 517 may be generated for the second operating period. As used herein, a parameter set defines values for a plurality of operating parameters such that, taken together, the parameter set defines and describes aspects of a particular operating mode. As such, the suggested parameter sets may be configured to describe or reference many of the potential operating modes of the power plant 501, and may be configured as input data sets for the tuned power plant model 507 for simulating the operation. Once the operating parameters have been generated and organized into the suggested parameter sets, the tuned power plant model 507 may simulate the operation of the power plant 501 according to each of them. The optimizer 510 may then evaluate the results of the simulated operation 519 for each of the proposed parameter sets 517. The evaluation may be in accordance with the performance objectives defined by the plant operator and the cost functions defined herein. The optimization process may include any of the methods described herein.
Cost functions defined by the performance goals may be used to evaluate economic performance of the simulated operation of the power plant 501 over the second period of operation. On the basis of the evaluations, one of the proposed sets of parameters may be considered to produce a simulated operation that is preferable to that produced by the other sets of parameters proposed. According to the present invention, the mode of operation corresponding to or described by the proposed parameter set generating the most preferred simulated operation is referred to as the optimized mode of operation. After determining, as discussed in more detail below, the optimized mode of operation may be passed on to the plant operator for consideration or communicated to the plant controller for automatic implementation.
In accordance with a preferred embodiment, methods of the present invention may be used to evaluate specific operating modes for determining and recommending preferred alternatives. As will be understood, the power plant blocks of the power plant 501 are controlled by variable set point actuators controllably connected to a control system, such as those shown in FIG. the plant controller 505. The operating parameters of the power plant 501 can be divided into three categories: manipulated variables, disturbances and controlled variables. The manipulated variables look at controllable process inputs, which can be manipulated via actuators to control the controlled variables, whereas the disturbances consider non-controllable process inputs, which affect the controlled variables. The controlled variables are the process outputs, which are controlled relative to defined target levels. According to preferred embodiments, the control method may include receiving predicted values for the disturbances for the second operating period (i.e., the operating period for which an optimized operating mode is calculated). The disturbances may include environmental conditions, e.g. Ambient temperature, pressure and humidity. In such cases, the suggested parameter sets generated for the second period of operation may include values for the disturbances related to the predicted values for the disturbances. More specifically, the generated values for each environmental condition parameter may include a range of values for each of the environmental condition parameters. The area may include, for example, a low case, a middle case, and a high case. It will be appreciated that the availability of multiple cases may allow a plant operator to plan for the worst / worst scenarios. The predicted values may include probabilistic ratings corresponding to different cases that may further assist the operator of the plant to plan and / or protect against losses for different operational eventualities.
The step of generating the proposed parameter sets may include generating target levels for the controlled variables. The target levels may be generated to correspond to competing or alternative modes of operation of the power plant 501, and may include operator input. Such operator inputs may be queried by the plant operator module 509. According to a preferred embodiment, such target levels may include a desired output level for the power plant 501, which may be based on likely output levels given past usage patterns for the plant. As used herein, an "output level" reflects a load level or level of electricity generated by the power plant 501 for commercial distribution during the second period of operation. The step of generating the suggested parameter sets may involve generating multiple cases, with the output level remaining the same or constant. Such a constant output level may reflect a base load for the plant or a block set. Several target levels may be generated, each corresponding to a different commitment level for each of the power plant blocks, and these may be pulled towards more likely operating modes in the light of prioritized use. The method may then determine the most efficient mode of operation, taking into account the known limitations. Moreover, the proposed parameter sets can be generated such that the disturbances maintain a constant level for the multiple cases generated for each target level. The constant level for the disturbances may be based on predicted values that have been received. In such cases, in one aspect of the present invention, the step of generating the proposed parameter sets includes generating multiple instances wherein the manipulated variables are varied over ranges to determine an optimized mode of operation for achieving a base load level, taking into account the predicted or expected environmental conditions , According to exemplary embodiments, the cost function is defined as a plant efficiency or a specific heat consumption, or may include a more direct economic indicator, such as: Operating costs, income or profit. In this way, the most efficient method of controlling the power plant 501 can be determined in situations where a base load is known and disturbances with a relatively high degree of accuracy can be predicted. The optimized mode of operation determined in such cases by the present invention may be configured to include a specific control solution (ie, specific setpoints and / or ranges therefor for the actuators that control the power plant's manipulated variables) determined by the control system Plant controller 505 could be used to achieve a more optimal function. Calculated in this manner, the control solution represents the optimized mode of operation for meeting a defined or agreed target load, taking into account the predicted values for the various disturbances. This type of functionality may be considered as optimization consulting or testing for a period during the day or serve between the markets, which analyzes the ongoing operations in the background for the purpose of recognizing more efficient operating modes that still meet predefined load levels. For example, as the market period covered by the previous load-distribution bid progresses, ambient conditions are known or at least the confidence level increases in their accurate prediction of what was estimated during the bidding process. In view of this, the present method can be used to optimize control solutions to meet the distributed load requirement given the safer knowledge of environmental conditions. This particular functionality is illustrated in FIG. 17 as the second parameter sets 517 and the simulated operation 519 with respect to the second parameter sets 517. In this way, the optimization process of the present invention may also include a "fine-tuning" aspect by which simulation runs on the tuned power plant model 507 propose more efficient control solutions that can then be communicated to and implemented by the plant controller.
Another aspect of the present invention involves its use to optimize fuel purchase for the power plant 501. It will be understood that power plants typically make regular fuel purchases from fuel markets that operate in a particular manner. Specifically, such fuel markets typically operate on a prospective basis, with the power plants 501 predicting the amount of fuel needed for a future operating period and then making purchases based on the prediction. In such systems, power plants 501 seek to maximize their profits by keeping low fuel stocks. However, the power plants 501 regularly buy additional fuel quantities to avoid the expensive situation of having an inadequate supply of purchased fuel to generate the amount of electricity that the plant has committed to provide during the load sharing process. This type of situation can occur when, for example, changing environmental conditions result in less efficient power generation than predicted or the true generating capability of the power plants is overestimated. It will be understood that several aspects of the present application already discussed may be used to determine an optimized mode of operation and to calculate, using it, a high accuracy fuel required prediction. That is, the present optimization processes may provide a more accurate prediction of plant efficiency and load capabilities that may be used to estimate the amount of fuel needed for a future period of operation. This allows plant operators to maintain a narrower fuel purchase margin, which in turn benefits the plant's economic performance.
The present invention includes, according to an alternative embodiment, a method for optimizing plant performance in which a prediction horizon is defined and used in the optimization process. As will be understood, a forecast horizon is a future operating period, which is subdivided into regularly repeating intervals for the purpose of determining an optimized operating mode for an initial time interval of the forecast horizon. Specifically, power plant operation is optimized by optimizing performance over the entire forecast horizon, which is then used to determine an optimized operating mode for the initial time interval. As will be understood, the process is then repeated to determine how the power plant should operate during the next time interval, which, as will be understood, becomes the initial time interval relative to the next repetition of the optimization cycle. For this subsequent optimization, the prediction horizon may remain the same, but is redefined relative to what is just defined as the initial time interval. This means that the forecast horizon is effectively pushed forward by an additional time interval each time it is repeated. As already mentioned, a "suggested parameter set" refers to a data set that contains values for several operating parameters and thereby defines or describes one of the possible operating modes for the power plant 501. According to a preferred embodiment, the process of determining the optimized mode of operation in cases including a prediction horizon may comprise one of the several of the following steps. First, several suggested horizon parameter sets are generated for the forecast horizon. As used herein, a "suggested horizon parameter set" includes a suggested parameter set for each of the time intervals of the prediction horizon. For example, a 24-hour forecast horizon may be defined to include 241 hour time intervals, which means that the proposed horizon parameter set includes suggested parameter sets for each of the 24 time intervals. As a next step, the proposed horizon parameter sets are used to simulate the operation over the forecast horizon. Then, for each of the simulation runs, the cost function for evaluating economic performance is used to determine which of the proposed horizon parameter sets is most appropriate or, as used herein, an "optimized horizon simulation run". According to exemplary embodiments, the operating mode described within the optimized horizon simulation run for the initial time interval of the prediction horizon may then be referred to as the optimized operating mode for the operating period corresponding to the initial time interval. The optimization process can then be repeated for subsequent time intervals. The present invention may include receiving predicted values for the disturbances for each of the time intervals defined within the prediction horizon. The proposed horizon parameter sets may then be generated such that the proposed parameter set corresponding to each of the time intervals includes values for the disturbances related to the predicted values received for the disturbances.
As will be understood, the proposed horizon parameter sets may be generated to cover a range of values for the disturbances. As before, this range may include multiple cases for each of the disturbances, and may include high and low values corresponding to cases above and below the predicted values. It will be understood that, in accordance with each of the described embodiments, the steps of simulating operating modes and determining optimized operating modes therefrom may be repeated and configured into a repetitive process. As used herein, each repetition is referred to as an "optimization cycle." It will be understood that each repetition may include defining a subsequent or next operating period for optimization. This subsequent period may take place shortly after the operating period optimized by the previous cycle, or may include an operating period corresponding to a future period, for example, when the present method is for the purpose of preparing load distribution bids or consulting on the economic Impact of alternative maintenance schedules.
The steps of tuning the power plant model 502 may be repeated to update the tuned power plant model 507. In this way, a tuned power plant model 507 that reflects a recent tune can be used with optimization cycles to produce more effective results. According to alternative embodiments, the optimization cycle and the tuning cycle of the power plant model 502 may be relatively separate, such that each cycle occurs according to its own schedule. In other embodiments, the power plant model 502 may be updated or tuned after a predefined number of repetitions of the optimization cycle. The updated tuned power plant model 507 is then used in subsequent optimization cycles until the predefined number of repetitions have occurred to initiate another tuning cycle. In certain embodiments, the tuning cycle occurs after each optimization cycle. According to alternative embodiments, the number of optimization cycles that initiate tuning of the power plant model 502 is related to the number of time intervals of the prediction horizon.
The present invention, as indicated, may optimize the operation of the power plants 501 in accordance with performance objectives that may be defined by the plant operator. According to preferred embodiments, the present method is used to economically optimize power plant operation. In such cases, the performance goals include and define a cost function that provides the criteria for economic optimization. According to exemplary embodiments, the simulated operation includes, for each of the proposed parameter sets, as an output, predicted values for selected performance indicators. The cost function may include an algorithm that correlates the predicted values for the performance indicators with operating costs or another indication of economic performance. Other performance indicators that may be used in this manner include, for example, specific heat consumption of the power plant and / or fuel consumption. According to alternative embodiments, the simulation outputs include predicted values for the hot gas path temperatures for one or more of the thermal power plants of the power plant 501, which may be used to calculate the cost of consumed component life. These costs reflect predicted wear costs associated with the hot gas path components resulting from the simulated operation. The cost function may further include an algorithm that correlates predicted values for the performance indicators with an operating revenue. In such cases, the operating yield may then be compared to the operating costs to reflect a net yield or gain for the power plant 501. The present method may further include the step of receiving a predicted price of electricity sold within the market for the optimized period, and the selected counters may count an output level of the electricity, which is then used to calculate an expected operating yield for the next one Operating period can be used. In this way, the present method can be used to maximize economic yield by comparing operating costs to revenue.
Further, as will be appreciated, performance objectives may be defined to include selected functionality limitations. According to certain alternative embodiments, the present method includes the step of disqualifying any of the proposed parameter sets that produce a simulated operation that violates one of the defined health limitations. Operational limitations may include, for example, emission thresholds, maximum operating temperatures, maximum mechanical stress levels, etc., as well as legal or environmental regulations, contract terms, safety regulations, and / or machine or component health thresholds and limitations.
The present method includes, as already mentioned, generating suggested parameter sets 517 describing alternative or possible operating modes of the power plant 501. As illustrated, the suggested parameter sets 517 may be generated in the plant operator module 509 and may include input from a plant manager or human operators. Generally speaking, the possible modes of operation may be considered as competing modes for which simulation is performed to determine the mode of operation that best meets the performance goals and anticipated conditions. According to exemplary embodiments, these alternative modes of operation may be selected or defined in a variety of ways. According to a preferred embodiment, the alternative operating modes include different output levels for the power plant 501. Output level, as used herein, refers to the level of electricity generated by the power plant 501 for commercial distribution within the market during a defined market period. The suggested parameter sets may be configured to define multiple cases at each of the different output levels. Multiple output levels may be covered by the suggested parameter sets, and the ones selected may be configured to match a range of possible outputs for the power plant 501. It will be understood that the range of possible output levels may not be linear. Specifically, due to the multiple power plant blocks of the power plant and the associated scalability limitations, the proposed parameter sets may be grouped or concentrated at levels that are more achievable or preferred given the particular configuration of the power plant 501.
As indicated, each of the competing modes of operation may include multiple cases. For example, where the competing modes of operation are defined at different levels, the multiple instances may be chosen to reflect a different manner by which the output level is achieved. If the power plant has multiple power plant blocks, the multiple cases at each output level can be differentiated by how each of the thermal power plants is operated and / or involved. According to one embodiment, the multiple generated cases are differentiated by varying the percentage of the output level provided by each of the power plant blocks. For example, the power plant 501 may include a combined cycle power plant 501 in which thermal power plants have gas and steam turbines. In addition, the gas and steam turbines may be replaced by an inlet conditioning system, such as an air conditioning system. a refrigeration system, and a HRSG channel firing system can be improved. As will be appreciated, for example, the intake conditioning system may be configured to cool the intake air of the gas turbine to crank its generating capability, and the HRSG channel combustion system may be configured as a secondary heat source to the boiler to boost the steam turbine's generating capability. According to this example, the thermal power plants include the gas turbine or, alternatively, the gas turbine cranked by the intake conditioning system; and the steam turbine or, alternatively, the steam turbine cranked by the HRSG duct firing system. The multiple cases covered by the proposed sets of parameters may then include examples in which these particular thermal power plants are incorporated in different ways while still meeting the different output levels chosen as competing modes of operation. The simulated operation may then be analyzed to determine which of them reflects an optimized mode of operation according to a defined criterion.
According to an alternative embodiment, the proposed parameter sets may be pulled towards different operating modes to calculate economic benefits of maintenance operations. To accomplish this, one of the competing modes of operation may be defined as one assuming that the maintenance operation was completed before the period of operation chosen for the optimization. This mode of operation may be defined to reflect a surge in performance that is expected to accompany completion of this maintenance operation. An alternative mode of operation may be defined as one in which the maintenance mode is not performed, meaning that the simulation of the multiple cases for that mode of operation would not have the expected performance overhead. The results from the simulations can then be analyzed so that the economic effects are better understood, and the multiple cases can be used to show how different scenarios (such as fluctuations in fuel prices or unexpected environmental conditions) affect the outcome. As will be appreciated, the competitive modes of operation may include a part-load mode and a shut-down mode using the same principles.
The present invention also includes different ways in which the optimization process by power plant operators can be used to automate processes and improve efficiency and performance. According to one embodiment, as illustrated in FIG. 17, the method includes the step of communicating a calculated optimized operation mode 521 to the plant operator module 509 for approval by a human operator before the power plant 501 is controlled according to the optimized operation mode. In a consulting mode, the present method may be configured to present alternative operating modes and the economic consequences associated with each of these to alert the plant operator to such alternatives. Alternatively, the control system of the present invention may be capable of automatically implementing optimized solutions. In such cases, the optimized operating mode may be electronically communicated to the plant controller 505 to cause control of the power plant 501 in a manner consistent therewith. In power systems that include an economic load sharing system for distributing electricity generation to a group of power plants 501, the optimization method of the present invention may be used to provide more accurate and competitive bids for submission to the central authority or central dispatcher. As one of ordinary skill in the art will understand, the optimization features already described may be used to create offerings that reflect true generation capability, efficiency, and specific heat consumption, while also providing useful economic compromise information to plant operators who operate the power plant future market periods by choosing between different operating modes. The increased accuracy of this type and the additional analysis help to ensure that the power plant remains competitive in the bidding process while also minimizing the risk of highly unprofitable load balancing resulting from unforeseen eventualities.
Figs. 18 to 21 illustrate exemplary embodiments of the present invention relating to the partial load and / or shutdown operation of a power plant. The first embodiment, as illustrated in flowchart 600 of FIG. 18 - which may be referred to as a "partial load advisor" - teaches methods and systems for simulating and optimizing a partial load level for the power plant during a defined or selected operating period ("selected operating period"). ). In preferred embodiments, the present method is used in power plants with multiple gas turbines, which may include gas turbine power plants with multiple gas turbines and one or more steam turbines. The tuned power plant model may be used to determine an optimized minimum load to operate the power plant at a part-load level during the selected operating period. As previously indicated, an "optimized" mode of operation may be defined as one deemed preferable or evaluated over one or more other possible modes of operation. An operating mode according to these embodiments may include assigning particular power generation units to meet a load obligation or other performance goals, as well as the physical configurations of the power plant units within a power plant. Such functionality means that by using an optimized or improved mode of operation, the present invention can accommodate a variety of equipment combinations that take into account the different part load configurations of each block, as well as configurations that disable one or more of the units while others on a full or partial basis Continue working at partial load level. The method may also have other limitations such as e.g. Functional constraints, performance objectives, cost functions, operator inputs, and environmental conditions are considered in its calculation of an improved part-load operating mode for the power plant that increases performance and / or efficiency. The present method, as described herein and / or in the appended claims, may include present and predicted environmental conditions for optimizing the part-load mode of operation as well as a change in unit configuration and / or control to dynamically adjust the operation of one or more of the power plant units if the actual conditions differ from those predicted. According to a preferred embodiment, such performance is defined, at least in part, as that which minimizes the level of fuel use or consumption over the proposed partial load operating period.
The part-load advisor of the present invention may take into account several factors, criteria and / or operating parameters in achieving an optimized or improved part-load solution and / or recommended part-load action. Preferred embodiments include, but are not limited to, gas turbine engine operating limits (i.e., temperature, aerodynamics, fuel distribution, lean extinction limit, mechanical, and emission limits); Gas turbine and steam turbine control systems; Minimum steam turbine throttle temperature; the maintenance of the vacuum seal on the capacitor as well as other factors, e.g. the configuration or installation of the systems or their control. One of the outputs of the optimization may include a recommended operating mode and configuration of the power plant or multiple plants, where the multiple plants may include different types of power plants, including wind, solar, piston engine, atom and / or other species. It will be understood that the recommended operating mode may be automatically initiated or electronically communicated to a plant operator for approval. Such control may be implemented via external or internal control systems configured to control the operation of the power plant blocks. In addition, in situations where the power plant includes multiple gas turbine engines, the output of the present method may include identifying which of the gas turbines should continue operating during the part load period and which should be shut down, which is a process which will be discussed in more detail with reference to FIG. For each of the gas turbines that the consultant recommends for continued operation during the part load period, the present method may further calculate a load level. Another output may include calculating the total load for the power plant during the part load period, as well as the hourly target load profile based on the predicted environmental conditions, which may be adjusted as indicated as conditions change. The present invention may also calculate the predicted fuel consumption and emissions of the power plant during the part load operating period. The output of the disclosed method may include operational planning / configuration in view of the control setpoints available to the power plant units and the plant to more efficiently achieve the target generation levels.
As discussed above, traders and / or asset managers (hereafter "plant operators", if no distinction is made between them) that are not tied to pre-existing contract terms typically offer their power plants in a prospective market, such as, for example. a day ahead market. As an additional consideration, plant operators are responsible for ensuring that an adequate supply of fuel is maintained so that the power plant is able to meet target or contracted production levels. However, fuel markets in many cases work prospectively so that advantageous pricing is available to power plants willing or able to commit in advance to future fuel purchases. More specifically, the further the fuel is purchased in advance, the more advantageous the pricing. Given this market momentum, in order for a power plant to achieve an optimized or high level of economic return, the plant operator must offer the plant competitive power to other power plant units in order to exploit its production capacity, while also accurately estimating the fuel needed for future production periods : 1) the fuel can be bought in advance to secure the low price allocation; and 2) no large fuel buffer is needed, so a small fuel inventory can be maintained. If successful, the plant operator will secure better pricing by committing early on to future fuel purchases while at the same time not making excessive purchases, which would require unnecessary and expensive fuel storage, or under-buying, thus risking a fuel supply deficit ,
The methods of the present invention can optimize or increase the efficiency and profitability of the power generation activities by specifying an IHR profile for a block or the particular configuration of a plant, especially as these relate to the preparation of a load sharing offer to accommodate a generation market share to back up. The present method may include specifying an optimal generation allocation across multiple power plant blocks within a power plant or across multiple plants. The present method can take into account the operating and control configurations available to these power plant blocks, permuting possible arrangements and thereby achieving an offer that, when selected, allows the generation of power over the bidding period at reduced or minimized cost. As a result, the present method can take into account all applicable physical, regulatory and / or contractual constraints. As part of this overall process, the present method may be used to optimize or improve part-load and shut-down operation for a multiple power plant power plant. This process may include considering anticipated exogenous conditions, such as weather or environmental conditions, gas quality, reliability of the power plant blocks, and ancillary obligations, such as, for example, Steam production. The present method can be used to create IHR profiles for multiple multi-configuration power plant blocks, as well as control settings for the selected part-load configuration, and then control for the anticipated exogenous conditions in preparing the plant load-distribution offer.
A frequent decision for operators concerns partial load operation or shutdown of the power plant during light load periods, such as low load periods. overnight when the need or load requirements are minimal. As will be understood, the result of this decision depends significantly on the operator's understanding of the economic consequences for each of these possible operating modes. In certain cases, the partial load operation of the power plant may be readily apparent, while the optimum minimum load at which the power plant should be maintained during the part load period remains uncertain. That is, while the plant operator has made the decision to operate the plant at partial load for a certain period of time, the operator is unsure about the partial load operating points at which the plant's multiple power plant units are run in the most cost effective manner.
The part-load advisor of FIG. 18 may be used as part of a process for recommending an optimal minimum load with which the power plant is to be operated. This advisor function may also recommend the best course of action for the power plant given a specific scenario of environmental conditions, economic inputs and operating parameters and constraints. From these inputs, the process may calculate the best operating levels and may then recommend the necessary operating parameters to control the power plant, as will be discussed in greater detail with respect to FIG. As will be appreciated, this functionality may result in several additional benefits, including extended part life, more efficient part-load operation, improved economical performance, and improved accuracy in fuel purchasing.
As illustrated in flowchart 600, certain information and relevant criteria may be collected during the initial steps. In step 602, data, variables, and other factors associated with power plant systems and power plant blocks may be determined. This may include any of the above factors or information. According to a preferred embodiment, an environmental profile may be received, which may include a prediction of environmental conditions during the selected period of operation. Relevant emission data may also be collected as part of this step, which may include emission limits as well as current emissions for the power plant. Another factor includes data related to the potential sale of electricity and / or steam during the selected operating period. Other variables that may be determined as part of this step include the number of gas turbines in the plant, the combustion and control systems for each of the gas turbines, and any other plant-specific limitations that may be relevant to the calculations discussed below.
In step 604, the period of the proposed partial load operation (or the "selected operation period") may be defined with care. As will be understood, this may be defined by a user or plant operator and include a selected operating period during which an analysis of available partial load operating modes is desired. The definition of the selected operating period may include an anticipated length and a user specified start time (i.e., the time at which the selected operating period starts) and / or a stop time (i.e., the time at which the selected operating period ends). This step may further include defining an interval within the selected operating period. The interval may be configured to divide the selected operating period into a plurality of sequential and periodically spaced periods. In terms of the example provided herein, the interval is defined as one hour and the selected operating period is defined as including more of the one-hour intervals.
In step 606, the number of gas turbines involved in the optimization process for the selected period of operation may be selected. This can include all gas turbines in the power plant or in a section thereof. The method may further include the consideration of other power plant units in the power plant, e.g. Steam turbine systems, and take into account their operating conditions during the selected operating period, as will be described in more detail below. The determination of the gas turbines involved in the partial load operation may include requesting or receiving inputs from the plant operator.
In step 608, the present method may configure a permutation matrix in view of the number of gas turbines that were determined as part of the proposed part-load operation during the selected operating period. As will be understood, the permutation matrix is a matrix that includes the various possibilities of how the multiple gas turbine engines can be integrated or operated during the selected period of operation. For example, as illustrated in the example permutation matrix 609 of FIG. 18, in the case of two gas turbines, the permutation matrix includes four different combinations covering each of the possible configurations. Specifically, if the power plant comprises a first and a second gas turbine, the permutation matrix includes the following series or cases: a) both the first and second gas turbines are "on", i. they are operated in a partial load mode; 2) both the first and the second gas turbine are "off", i. they are operated in a shutdown mode; 3) the first gas turbine is "on" and the second gas turbine is "off"; and 4) the first gas turbine is "off" and the second gas turbine is "on". As will be appreciated, in the case of a single gas turbine, only two permutations are possible, while for three gas turbines, seven different series or cases would be possible, each representing a different configuration, such as the three gas turbine engines during a given timeframe "Off" operating states can be involved. Referring to Fig. 17 and the optimization process discussed in the accompanying text, each case or row of a permutation matrix may be considered to represent a different or competing mode of operation.
As part of the steps represented by steps 610, 613, 614, 616, and 618, the present method may configure proposed parameter sets for the proposed part-load operation. As indicated, the selected operating period may be divided into the several one-hour time intervals. The process of configuring the suggested parameter sets may begin in step 610 where it is determined whether each of the intervals has been addressed. If the answer to this question is yes, the process as illustrated may proceed to an issue step (i.e., step 611), where the output of the partial load analysis is provided to an operator 612. If not all of the intervals have been covered, the process may proceed to step 613, selecting one of the intervals that has not yet been covered. Then, in step 614, the environmental conditions for the selected interval may be adjusted based on received predictions. Continuing with step 616, the process may select a row from the permutation matrix and adjust the on / off state of the gas turbines according to the particular row at step 618.
From there, the present method can continue along two different paths. Specifically, the method may proceed to an optimization step represented by step 620, while at step 621 it may also proceed to a decision step wherein the process determines whether all permutations or rows of the permutation matrix have been covered for the selected interval. If the answer to this is "no", the process may return to step 616, where a different permutation series is selected for the interval. If the answer to this is yes, then the process, as illustrated, may proceed to step 610 to determine if all intervals have been covered. As will be understood, after all rows of the permutation matrix have been addressed for each interval, the process may proceed to the output step of step 611.
In step 620, the present method may optimize performance using the tuned power plant model as discussed previously in FIG. 17. Consistent with this approach, multiple cases can be generated for each of the competing modes of operation, i. each of the rows of the permutation matrix for each of the intervals of the selected period of operation. According to a preferred embodiment, the present method provides suggested sets of parameters in which a plurality of operating parameters are varied to determine the effect on a selected operating parameter or performance indicator. For example, according to this embodiment, the proposed parameter sets may include manipulating settings for "IGV" ("inlet guide vanes") and / or exhaust gas temperature of the turbine ("Texh") to determine which combination would be appropriate in the light of the on / off State of the particular series and environmental condition prediction for the particular interval results in a minimized total fuel consumption rate for the power plant. As will be appreciated, operation that minimizes fuel consumption while meeting the other constraints associated with part-load operation is one way in which part-load power is economically optimized or at least economically enhanced relative to one or more alternative operating modes can.
As shown, cost functions, performance goals and / or functionality limitations may be used by the present invention during this optimization process, according to certain embodiments. These may be provided via a plant operator, represented by step 622. These limits may include limits on IGV settings, Texh limits, combustion limits, etc., as well as those associated with the other thermal systems that may be part of the power plant. For example, in power plants with CCGT systems, the operation or maintenance of the steam turbine during part-load operation may present certain limitations, such as maintaining a minimum steam temperature or a condenser vacuum seal. Another capability limitation may include the necessary logic that certain ancillary systems may be degraded in certain modes of operation and / or certain subsystems may be mutually exclusive, such as e.g. Evaporative cooler and refrigeration system.
After the present process has passed through the repetitions through the intervals and the different rows of the permutation matrix, the results of the optimization can be communicated to the plant operator in step 611. These results may include an optimized case for each of the rows of the permutation matrix for each of the time intervals. As an example, the output describes optimized operation defined by a cost function of the fuel consumption for the power plant for each of the permutations for each of the intervals. Specifically, the output may meet the minimum fuel required (as optimized using the tuned power plant model according to previously described methods) for each of the possible plant configurations (as represented by the rows of the permutation matrix) for each interval, while also meeting health limitations, performance goals, and anticipated environmental conditions will involve. According to another embodiment, the output includes an optimization that minimizes a generation output level (i.e., megawatts) for the possible plant configurations for each of the intervals in the same manner. As will be understood, certain of the possible plant configurations (as represented by permutations of the permutation matrix) may not be able to meet performance limitations, regardless of the fueling for the generation output level. Such results may be discarded and are no longer considered or reported as part of the outcome of step 611.
FIGS. 19 and 20 graphically illustrate ways in which a gas turbine of a power plant may be operated over a selected period of operation including defined intervals ("I" in the figures) in view of typical transient operating limitations. As will be understood, transient operation involves switching a block between different modes of operation, including those involving transition into or out of a shutdown mode of operation. As shown, multiple operating paths or sequences 639 may be achieved, depending on: 1) an initial state 640 of the gas turbine; and 2) the decisions made as to whether operating modes are being changed in the intervals in which changes are possible in view of the limitations of transient operation. As will be appreciated, the plurality of different sequences 639 represent the multiple ways in which the block can operate over the intervals shown.
As will be appreciated, the output of the method of FIG. 18 may be used in conjunction with the diagrams of FIGS. 19 and 20 to configure proposed partial load operating sequences for power plant power plant units. That Figs. 19 and 20 illustrate examples of how a block of a power plant can be integrated and how its modes of operation are modified when the time intervals expire, which may include examples if the mode of operation of the block remains unchanged; Shut down operating mode is modified in a partial load operating mode, and examples, when the operating mode of the unit is modified from a shutdown operating mode in a partial load operating mode. As illustrated, the transient operating constraint used in this example that modifying an operating mode requires the unit to remain in the modified operating mode for a minimum of at least two of the intervals. The many sequences (or paths) through which the block arrives at the last interval represents the possible part load operating sequences available to the unit in view of the transient operating limitations.
As will be appreciated, the analytical results of Fig. 18 - i. the optimized partial load operation for each of the matrix permutations - may be used to select a plurality of preferred cases from the possible part load operating sequences, which may be referred to as the proposed partial load operating sequences. Specifically, in view of the results of the method described with reference to FIG. 18, the proposed part-load operating sequences may be selected from cases of part-load operation that satisfy plant performance goals and constraints, while also providing power according to a selected cost function (such as MW output or Fuel consumption) is optimized. The considerations illustrated in FIGS. 19 and 20 represent one way to determine whether partial load operating sequences are feasible in view of transient operating limitations. That is, the proposed part-load operating sequences achieved by the combined analysis of Figs. 18-20 are operating sequences that coincide with time constraints associated with the transition of a unit from one operating mode to another.
Referring to Fig. 21, a method for further modeling and analyzing part-load operation of a power plant is provided. As will be understood, this method can be used to analyze the partial load costs versus shutdown costs for specific cases involving a single block over a defined time interval. However, it may also be used to analyze plant-level costs, with a recommendation on how to control the operation of multiple power plant blocks over a selected operating period at multiple intervals. In this manner, the output of Figs. 18 and 20 may be assembled to configure possible modes of operation or sequences over the span of multiple intervals, which may then be analyzed according to the method of Fig. 21, as will be shown to provide a more complete understanding of part-load operation over a wider operating period.
As already discussed, plant operators must regularly decide between part-load and shut-down modes of operation during the low load hours. While certain conditions make the decision uncomplicated, it is often difficult, especially given the increased complexity of modern power plants and the multiple thermal power plants that are typically included within each individual. As will be understood, the decision on partial load versus shutdown of a power plant will depend significantly on a full understanding of the economic benefits associated with each mode of operation. The present invention, according to the alternative embodiment illustrated in Figure 21, may be used by the plant operator to obtain an improved understanding of the trade-offs associated with each of these different operating modes to facilitate decision-making. According to certain embodiments, the method of FIG. 21 may be used in conjunction with the part-load advisor of FIG. 18 to enable a combined advisor function that: 1) Given the best course of action between part-load and shut-down modes of operation for the power plant power plant units recommended conditions and economic factors; and 2) when part load operation is the best course of action for some of these units, the optimum part-load minimum load level is recommended. In this way, the plant operator can more easily identify situations in which the units of the power plants should be switched off at partial load, or vice versa, based on which of these modes is best in view of a specific scenario of environmental conditions, economic inputs and operating parameters economic approach for the power plant. Secondary benefits, such as Extending the component part life is also possible. It should also be understood that the methods and systems described with respect to FIGS. 18 and 21 may also be used separately.
In general, the method of flowchart 700 - which may also be part of or is referred to as a "part-load advisor" - applies user inputs and data from analytical operations to perform calculations related to costs evaluate with the switching of a power plant in the partial load operation over those for its shutdown. As will be understood, flow chart 700 of FIG. 21 provides this advisory feature by, in accordance with certain preferred embodiments, supporting the tuned power plant model discussed in detail above. As part of this functionality, the present invention can advise on the various results, both economical and otherwise, between part load operation and shutdown of a power plant during light load demand periods. The present invention may provide relevant data that indicates whether a partial load operation of the power plant is preferable to its shutdown for a specified market period. According to certain embodiments, operation with the lower cost may then be recommended as the appropriate action to the plant operator, although, as also illustrated herein, additional aspects or other considerations may be communicated to the plant operator that may influence the decision. The present method may present potential costs as well as the likelihood of such costs being incurred, and these considerations may affect the ultimate decision as to which mode of operation is preferred. Such considerations may include, for example, a complete analysis of both short term operating costs and long term operating costs associated with plant maintenance, operating efficiencies, emissions levels, equipment upgrades, and so on.
As will be appreciated, the part-load advisor may be implemented using many of the systems and methods described above, particularly those discussed with respect to FIGS. 16-20. The part load advisor of Figure 21 may collect and use one or more of the following types of data: user specified start and stop time for the proposed part load operating period (i.e., the period for which the part load operating mode is analyzed or considered); Fuel costs; Environmental conditions; Break-breaker; alternative power consumption; Sale / price of electricity or steam during the relevant period; Operating and maintenance costs over the period; User input; calculated part load load; predicted emissions for operation; current emission levels generated by the power plant and limits for defined regulatory periods; Specifications regarding the operation of the rotary device; Control and equipment relating to the flushing process; fixed costs for modes of power plant operation; Costs relating to the starting operation; Anlagenanfahrzuverlässigkeit; Imbalance costs or penalties for delayed start-up; Emissions in terms of start-up; used fuel rate for auxiliary boilers when a steam turbine is present; and historical data regarding how the power plant's gas turbines had previously operated in part-load and shut-down modes of operation. In certain embodiments, as discussed below, the outputs of the present invention may include: a recommended operating mode (i.e., partial load and shutdown mode of operation) for the power plant over the relevant time period; Costs associated with each mode of operation; a recommended plant operating load and load profile over time; a recommended time to initiate unit startup; as well as emissions that have been discharged in the current year to date and allow emission levels remaining for the remainder of the year. According to certain embodiments, the present invention may calculate or predict the fuel consumption and emissions of the power plant over the relevant time period, which information may then be used to calculate the cost of partial load versus shutdown for one or more particular gas turbine engines. The present method may use the cost of each gas turbine in shutdown and part load modes to determine the combination that has the lowest cost of ownership. Such optimization may be based on different criteria that may be defined by the plant operator. For example, the criteria may be based on yield, net yield, emissions, efficiency, fuel consumption, etc. Moreover, according to alternative embodiments, the present method may recommend specific actions, such as e.g. whether a flush bonus should be accepted or not; the gas turbine units that should be shut down and / or those that should be switched to part load (which may be based, for example, on past start-up reliability and potential imbalance costs that may be incurred due to a delayed start). The present invention may also be used to improve fuel consumption predictions to more accurately track prospective fuel purchases or, alternatively, to facilitate fuel buying for market periods that are more in the future, which has a positive impact on fuel pricing and / or maintaining a smaller fuel stock or margin.
FIG. 19 illustrates an exemplary embodiment of a part-load advisor according to an exemplary embodiment of the present invention, which is in the form of a flowchart 700. The part load advisor may be used for relative cost recommendations over a future period of operation of a power plant shutdown or portion thereof during operation of another of the power plant units in a partial load mode. According to this exemplary embodiment, the potential costs associated with the shutdown and the partial load mode of operation may be analyzed and communicated to the operator for the appropriate action.
As initial steps, certain data or operating parameters may be collected that may affect or may be used to determine operating costs during the selected part-load operating period. These are grouped as illustrated, as illustrated: partial load data 701; Shutdown data 702; and shared data 703. The shared data 703 includes those cost elements that relate to both the shutdown and partial load modes of operation. The common data 703 includes, for example, the selected period of operation for which the partial load mode of operation analysis is performed. It will be understood that more than one selected operating period may be defined and analyzed separately for contiguous modes of partial load operation so that a wider optimization over an extended time frame is achieved. As will be understood, defining the selected operating period may include defining the length of the period and its start or end point. Other common data 703 may include, as shown, the following: the fuel price; the different emission limits for the power plant; and data related to environmental conditions. With regard to emission limits, the data collected may include amounts that may be generated during a defined regulatory period, such as, for example, a year, as well as the amounts already accumulated for the power plant and the extent to which the current regulatory period has already expired. Furthermore, emission data may include penalties or other costs associated with exceeding any of the limits. In this way, the present method may be informed of the current status of the power plant relative to annual or periodic regulatory limits, as well as the likelihood of potential infringement and penalties associated with such non-compliance. This information may be relevant to the decision as to whether power plant blocks are shut down or run at partial load, as each operating mode has a different impact on system emissions. With respect to the environmental condition data, such data may be obtained and used in accordance with the processes already described herein.
As will be understood, the partial load mode of operation includes data that is solely relevant to determining operating costs associated therewith. Such partial load data 701, as illustrated, includes the revenue that can be earned from the power generated while the power plant is operating at part-load levels. More specifically, there is the potential that because the partial load mode of operation is one in which power generation continues, albeit at a lower level, this power will generate revenue for the power plant. As this is done, the revenue may be used to offset some of the other operating costs associated with the partial load mode of operation. Accordingly, the present method involves receiving a price or other economic indication in connection with the sale or commercial use of the power the equipment is producing while operating in the partial load mode. This may be based on historical data, and the revenue earned may depend on the partial load level at which the power plant is operating.
The partial load data 701 may further include operation and maintenance associated with operating the equipment at the part-load level during the selected operating period. This may also be based on historical data, and such costs may depend on the partial load level for the power plant and how the power plant is configured. In some cases, these costs may be reflected as hourly costs, which depend on the load level and historical records of a similar operation. The partial load data 701 may further include data related to plant emissions during operation in part-load mode.
The shutdown data 702 also includes several elements unique to the shutdown mode of operation, and this type of data may be collected at this stage of the current process. According to certain embodiments, some of these are data related to the operation of the rotary device during the shutdown period. In addition, data are defined with regard to the various phases of the shutdown operation. These may include, for example, data related to: the shutdown operation itself, which includes historical data on the amount of time necessary to bring the power plant blocks from a regular load level to a state in which the turntable is engaged; the length of time the power plant remains switched off according to the selected period of operation; the amount of time that the block usually remains on the rotating device; and data regarding the process by which the power plant units are restarted or reintegrated after shutdown and the time required for this; Starting fuel requirements and starting emission data. In determining the start-up time, such information as the types of start-up possible for the block and specifications relating thereto may be determined. As one skilled in the art will understand, start-up processes may depend on the time for which the power plant remains turned on. Another consideration that affects the startup time is whether the power plant has certain features that may affect or shorten the startup time and / or whether the power plant operator chooses to use those features. For example, a flushing process, if necessary, can extend the startup time. However, a purge bonus may be available if the power plant has been shut down in a particular manner. Fixed costs associated with shutdown operation, including those associated with startup, may be determined during this step, as well as costs specific to one of the relevant power plant units. Emission data in connection with the startup and / or shutdown of the power plant can also be determined. These may be based on historical records of the operation or other data. Finally, data relating to the starting reliability for each of the thermal power plants can also be determined. As will be understood, power plants can incur costs, sanctions and / or penalties if the process of reconnecting has delays resulting in the power plant being unable to meet its load obligations. These costs may be determined and, as discussed in more detail below, may be considered in light of the historical data related to start-up reliability. In this way, such costs may be discounted to reflect the likelihood of their occurrence and / or to include an expense by which the risk of such costs is taken into account or hedged.
From the initial data acquisition steps 701 to 703, the exemplary embodiment illustrated in FIG. 19 may proceed via a partial load analyzer 710 and a shutdown analyzer 719, each configured to calculate operating costs for the operating mode to which it corresponds. As illustrated, each of these analyzers 710, 719 may continue to provide cost, emissions, and / or other data at step 730, where data regarding possible part-load and unit shutdown scenarios is compiled and compared, so that at step 731 a Output can be made to a power plant operator. As will be discussed, this issue may include costs and other considerations for one or more of the possible scenarios, and may eventually suggest a particular action and reasons for it.
With regard to the partial load analyzer 710, the method may first determine the load level for the proposed partial load operation during the selected operating period. As further discussed below, many of the costs associated with part-load operation may significantly depend on the load level on which the power plant is operating, as well as how the plant is configured to generate that load, which may include, for example, how the various thermal power plants are involved are (ie which are driven in partial load and which are switched off). The partial load load level for the proposed partial load operation may be determined in various ways according to alternative embodiments of the present invention. First, the plant operator can select the partial load load level. Second, the load level can be selected by analyzing historical records for past partial load levels on which the plant has been operating efficiently. From these records, a proposed load level may be analyzed and selected based on operator-provided criteria, such as efficiency, emissions, meeting one or more location-specific targets, availability of alternative commercial uses for the flow generated during part-load, environmental conditions, and other factors.
As a third method of selecting the part-load level for the proposed part-load operation, a computer-implemented optimization program, such as the one shown in FIG. described with reference to Fig. 18 for calculating an optimized partial load level. In Fig. 19, this process is represented by steps 711 and 712. An optimized partial load level may be calculated by proposing part-load operating modes in step 711 and then analyzing in step 712 if the operating limits for the power plant are met. As will be understood, a detailed description of how this is achieved is provided above with reference to FIG. By using a process such as this, for optimizing the partial load level, it will be understood that the partial load modes of operation selected for comparison to the switch-off alternatives for the selected period of operation represent the optimized case and, in view of this, the comparison between the partial load and the shutdown alternative is meaningful. As indicated in relation to FIG. 18, the minimum load partial level may be calculated via an optimization process that calculates the partial load level according to operator selected criteria and / or cost functions. One of the functions may be the level of fuel consumption during the proposed part-load operating period. That is, the optimized partial load level may be determined by optimizing fuel consumption towards a minimum level while also meeting all other operating limits or site specific performance goals.
From there, the present method of FIG. 19 may determine the costs associated with the proposed partial load operating mode for the selected operating period according to the part load operating mode characteristics determined via steps 711 and 712. As illustrated, step 713 may calculate the fuel consumption and therefrom the fuel costs for the proposed part-load operation. In accordance with the illustrative embodiment just discussed, which describes optimization based on minimizing fuel consumption, fuel cost may be deduced simply by taking the fuel level calculated as part of the optimization step and then multiplying it by the anticipated or known price of fuel becomes. In a next step (step 715), the output derived from the current generated during the selected operating period may be calculated in view of the proposed partial load level and the availability of a commercial demand during the selected operating period. Then, in step 716, the operating and maintenance costs may be determined. The operating and maintenance costs associated with the proposed partial load operation may be calculated by any conventional method and may depend on the partial load level. Operating and maintenance costs may be reflected as hourly costs derived from partial load operation historical records, and may include component usage costs that reflect a portion of the expected life of various component systems operating during the proposed part load operation. In a next step, indicated by step 717, net costs for the proposed partial load operating mode for the selected operating period may be calculated by adding the cost (fuel, operation, and maintenance) and subtracting the yield.
The present method may also include step 718 which determines the plant emissions over the selected operating period taking into account the proposed partial load operating mode, which may be referred to as the "emissions impact". The net cost and emissions impact may then be provided to a compose and compare step, shown as step 730, so that the cost and emissions impact of different part load scenarios may be analyzed so that ultimately, in an issue step 731, a recommendation may be made, as below will be discussed further.
Referring to the shutdown analyzer 719, it may be used to calculate aspects related to the operation of one or more of the power plant power plant units in a shutdown mode of operation during the selected operating period. As part of this aspect of the invention, operations, including the procedures by which the power plant is shut down and then restarted at the end of the selected period, may be analyzed for costs and emissions. In accordance with a preferred embodiment, the shutdown analyzer 719, as part of the initial steps 720 and 721, may determine a suggested shutdown mode of operation, which may represent an optimized shutdown mode of operation. The proposed shutdown mode of operation includes processes whereby one or more of the power plant blocks are shut down and then restarted to reconnect the units at the end of the selected operating period. As will be understood, the duration of the period during which a block is not operating determines the type of possible startup processes that are available to it. For example, the availability of a hot or cold startup depends on whether the shutdown period is short or long. By determining the proposed shutdown mode of operation, the present method may calculate the time necessary for the startup process to bring the block back to an operating load level. In step 721, the method of the present invention may check to ensure that the proposed shutdown operation complies with all operating limits of the power plant. If one of the operating limits is not met, the method may return to step 720 to calculate an alternative startup procedure. This can be repeated until an optimized start-up operation has been calculated, which complies with the operating limits of the power plant. As will be appreciated, the tuned power plant model according to the methods and systems discussed above may be used to simulate alternative shutdown modes of operation to determine optimized cases taking into account the relevant operating period and project environment conditions.
Given the proposed shutdown mode of operation of steps 720 and 721, the process may be continued by determining the cost associated therewith. Initial steps include analyzing the type of startup process that includes the shutdown mode of operation. In step 722, the process may determine the specific operating parameters of the startup, including a determination of whether flushing is required or not, or requested by a plant operator. Given the particular startup, fuel cost may be determined in step 723. According to an exemplary embodiment, the shutdown analyzer 719 then calculates the costs associated with the delays that occasionally occur during the startup process. Specifically, as indicated in step 724, the process may calculate the likelihood of such a delay. This calculation may include inputs such as the type of start-up as well as historical records of past start-up of the relevant power plant units in the power plant as well as data regarding the start-up of such power plant units in other power plants. As part of this, the process may calculate costs related to the proposed shutdown mode of operation, which includes the likelihood of having a startup delay and penalties, such as: Contractual penalties that would arise reflect. Such costs may include any costs associated with a hedging policy by which the power company passes on part of the risk of incurring such penalties to a service provider or other insurer.
In step 726, the current method may determine costs associated with operation of the rotary device during the shutdown process. The method may calculate a speed profile for the turning device taking into account the turn-off period and, using that, the cost of the self-consumption current needed to operate the turning device. As will be appreciated, this is the current required to continue rotating the rotor blades of the gas turbine as it cools, which is done to avoid warping or distortion that would otherwise occur if the blades were placed in a stationary position would cool down. In step 727, as illustrated, the operating and maintenance costs for the shutdown operation may be determined. The operating and maintenance costs associated with the proposed shutdown can be calculated by any conventional method. The operating and maintenance costs may include component usage costs that reflect a portion of the expected life of various component systems used during the proposed shutdown operation. In a next step, indicated by step 728, net cost for the proposed shutdown mode of operation for the selected operating period may be calculated by adding the calculated cost of fuel, rotating device, and operation and maintenance. The present method may also include step 729 in which the plant emissions are determined over the selected period of operation in view of the proposed shutdown mode of operation, which, as previously indicated, may be referred to as the "emissions impact" of the operating mode. The net cost and emissions impact may then be provided to compilation and comparison step 730.
In step 730, the current method may assemble and compare various plant part-load modes of operation for the selected period of operation. In one embodiment, the current method may analyze competing partial load modes of operation identified as part of the methods and processes described with respect to FIGS. 18-20. In step 730, the aggregated cost data and emissions impact for each of the competing partial load modes of operation may be compared and provided as an output as part of step 731. In this way, according to the result of the comparison of the competing operating modes, a recommendation can be made as to how the power plant should operate during the selected part-load operating period, including which of the turbines should be shut down and which of the turbines should be in partial load should be switched, as well as the part-load level, on which they should be operated.
The emission data may also be provided as part of the output of step 731, especially in cases where the analyzed competing modes of operation have similar economic results. As will be understood, notification may also be provided as to what impact each alternative has on asset issues and, given the impact, the likelihood of non-compliance during the current regulatory period, and related economic performance. Specifically, the accumulated emissions of one or more plant pollutants during the regulatory period may be compared to the total limits allowed during that time frame. According to certain preferred embodiments, the step of communicating the result of the comparison may derive an emission rate of the power plant from averaging a cumulative emission level for the power plant over a portion of a current regulatory emission period relative to an emission rate derived by averaging a cumulative emission limit above the current one include the regulatory issue period. This can be done to determine what the power plant is like when compared to the average emission rates that are allowed without a violation occurring. The method may also determine the emissions still permitted to the power plant during the current regulatory period and whether or not there is a sufficient margin for one of the proposed modes of operation, or rather, if the emission impact increases improperly, the likelihood of a future breach the regulations exist.
As an output, the present method can provide a recommended action advising of advantages / disadvantages, both economical and otherwise, between the proposed part-load and shut-down modes of operation. The recommendation may include cost reporting and a detailed breakdown between the categories in which these costs were incurred and the assumptions made in their calculation. In addition, the recommended action may include a summary of any other considerations that could influence the decision by which the most favorable mode of operation is selected. This may include information regarding applicable emission limits and regulatory periods, as well as where the current cumulative emissions of the power plant are related. This may include notifying the power plant operator of any operating mode that unreasonably increases the risk that will violate emission thresholds and the cost of such violations.
The present invention may further include a unified system architecture or integrated computational control system that efficiently enables and improves the performance of many of the functional aspects described above. Power plants - even jointly owned - often operate across different markets, government jurisdictions and time zones, have many types of stakeholders and decision-makers involved in their management, and exist under varying types of service and other contractual arrangements. Within such diverse frameworks, a single owner can control and operate a number of power plants, each with multiple power plant blocks and types across overlapping markets. Owners may also have different criteria for the evaluation of effective power plant operations, including, for example, unique cost models, response time, availability, flexibility, cyber security, functionality, and differences inherent in the ways in which separate markets operate. However, as will be appreciated, most current electricity trading markets are based on various off-line generated files shared by multiple parties and decision makers, including those conveyed between traders, asset managers and regulators. In view of such complexities, the capabilities of power plants and / or power plants within a market segment may not be fully understood, particularly across the multi-layered hierarchy, for example, from individual power plant units to power plants or from individual power plants to multiple such facilities. As such, each of the successive levels of the electricity trading market typically secures the power reported by the underlying level. This results in inefficiencies and lost revenue for the owners, as repeated hedging increases systemic underutilization. Another aspect of the present invention, as discussed below, is working to reduce the gaps that are the source of these problems. In one embodiment, a system or platform is developed that can perform analyzes, collect and evaluate historical data, and perform what-if or alternative scenario analysis on a unified system architecture. The unified architecture may have different functions, different components, e.g. Power plant modeling, operational decision support tools, the prediction of power plant operation and performance and the optimization according to performance goals more efficiently. According to certain aspects, the unified architecture may achieve this by integrating components locally in the power plant with those remote from it, such as those housed in a centrally located or cloud-based infrastructure. As will be appreciated, aspects of such integration may enable improved and more accurate power plant models while not compromising consistency, effectiveness, or timeliness of the results. This can include exploiting the previously discussed co-ordinated power plant models in local and external computing systems. Given their use in an externally located infrastructure, the system architecture can be scaled to accommodate additional sites and units.
Referring to Figs. 22-25, scalable architecture and control systems are illustrated which may be used to support the many requirements associated with control, management, and optimization of a power plant inventory in which multiple power plant blocks are distributed over multiple locations. A local / remote hybrid architecture, as provided herein, may be employed based on particular criteria or parameters that are situation- or case-specific. For example, an owner or operator of a number of power plants may want certain aspects of the system functionality to be located locally, while others may represent a centrally located environment, such as e.g. a cloud-based infrastructure to pool data from all powerhouse blocks and serve as a common database that can be used to clear the data on cross-referenced values of common equipment, configurations, and conditions while also supporting analytical functions. The method of selecting the appropriate architecture for each of the various types of owner / operator may focus on the significant concerns that drive the operation of the power plants as well as the specific characteristics of the electricity market in which the plants operate. According to certain embodiments, as provided below, performance calculations may be performed locally to assist the closed loop of a particular power plant, to enhance cyber security, or to provide the response speed necessary to perform near real time processing. On the other hand, the present system may be configured such that the data stream between local and remote systems includes local data and model tuning parameters that are communicated to the centrally located infrastructure to create a tuned power plant model that is then used for analyzes such as e.g. the analysis of alternative scenarios is used. Remote or centrally located infrastructure can be used to tailor interactions with a common asset model according to the unique needs of the different types of users who need access. In addition, a scaling strategy can be determined based on response time and service agreements that depend on the unique aspects of a particular market. If faster reaction times are required for the availability of final results, the analytical processes can be scaled in terms of both software and hardware resources. The system architecture also supports redundancy. If a system running analysis goes out of service, processing can continue on a redundant node that contains the same power plant models and historical data. The unified architecture can bring together applications and processes to enhance performance and increase the level of functionality for both technical and commercial benefits. As will be understood, such benefits include: convenient integration of new power plant models; Separation of procedures and models; enabling different operators to share the same data in real time, while also presenting the data in a unique way according to the needs of each operator; comfortable upgrades; and compliance with the NERC CIP restrictions for sending supervisory control.
Figure 22 illustrates a high-level logic flowchart or method for inventory-level optimization according to certain aspects of the present invention. As shown, the plant stock may include multiple power plant blocks or assets 802 that may represent separate power plant units across multiple power plants or the power plants themselves. Asset assets 802 may be owned by a single owner or a legal entity and may compete with other such assets across one or more markets for contract rights to generate shares of the load required by a customer network. The assets 802 may include multiple power plant blocks having the same type of configurations. In step 803, performance data collected by the sensors in the various assets of the facilities may be electronically communicated to a central database. Then, in step 804, the measured data may be adjusted or filtered so that, as described below, a more accurate or truer indication of the power level for each asset is determined.
As described in detail above, one way this balance can be done is to compare the measured data with corresponding data predicted by power plant models, which, as discussed, may be configured to simulate the operation of one of the assets , Such models, which may also be referred to as off-line or predictive models, may include physics-based models, and the adjustment process may be used to periodically tune the models to maintain and / or improve the accuracy with which the models, through which the models Simulation, represent the actual operation. That is, as discussed in detail above, in step 805, the method may use the most recent collected data to tune the power plant models. This process may involve tuning the models for each of the assets, i. each of the power plant units and / or power plants, as well as more general models that cover the operation of multiple power plants or aspects of existing operations. The matching process may also include comparing the collected data between similar assets 802 to resolve discrepancies and / or identify anomalies, particularly data collected from the same type of asset with similar configurations. During this process, gross errors can be eliminated given the collective and redundant nature of the compiled data. For example, sensors having higher accuracy capabilities or those known to be recently tested and whose operation has been proven to be correct may be preferred. In this way, the collected data can be cross-checked, verified and reconciled to produce a single unified record that can be used to calculate more accurate actual inventory performance. This data set can then be used to reconcile offline asset models, which can then be used to simulate and determine optimized plant asset control solutions during a future market period, which can be used, for example, to assess the competitiveness of the power plant during load distribution Bidding procedure.
In step 806, as illustrated, the true capabilities of the power plant are determined from the adjusted performance data and the matched models of step 805. Then, assets 802 of the asset pool may be collectively optimized in step 807 in view of a selected optimization criterion. As will be understood, this may involve the same processes discussed above in detail. In step 808, an optimized supply curve or asset schedule may be created. This may describe the manner in which the assets are scheduled or operated, as well as the level at which each is involved, for example, to meet a proposed or hypothetical load level for the power plant stock. The optimization criteria may be selected by the operator or owner of the assets. For example, the optimization criteria may include efficiency, yield, profitability, or some other measure.
[0159] As illustrated, subsequent steps may include communicating the optimized asset schedule as part of an offer for load generation contracts for future market periods. To do so, in step 809, the optimized asset schedule may be communicated to energy traders who then bid according to the optimized asset schedule. As will be appreciated, the offers may be used in step 810 to participate in a power system wide load distribution process by which the load is shared among multiple power plants and power plant blocks located within the system, many of which are owned by competing owners can. The bids or quotes for the load-balancing process may be configured according to a defined criterion, such as: variable generation cost or efficiency as determined by the particular dispatcher of the power system. In step 811, the power system optimization results may be used to create an asset schedule that reflects how the various assets should be integrated into the power system to meet the predicted demand. The asset schedule of step 811, which reflects the result of the system-wide optimization or load distribution process, may then be communicated back to the owners of the assets 802 such that in step 812, operating setpoints (or, in particular, operating modes), for example the load, in which each of the assets is operated, may be counted, communicated to a controller controlling the operation of the assets 802. In step 813, the controller may calculate a control solution and then communicate and / or directly control the assets 802 to meet the load requirements that it has committed to during the load-sharing process. Power plant owners can adjust the way one or more power plants work as conditions change to optimize profitability.
Fig. 23 illustrates the data flow between local and remote systems according to an alternative embodiment. As stated, certain functions may be located locally while other functions are located remotely in a centrally located environment. The method of selecting the appropriate architecture according to the present invention involves determining the considerations that are significant drivers of operating the assets within the asset pool. Accordingly, considerations such as e.g. Cybersecurity concerns, possibly that certain systems remain located locally. Time-consuming performance calculations are also stored locally, so that the necessary up-to-dateness is maintained. As illustrated in FIG. 23, the local asset control system 816 may receive sensor measurements and communicate the data to a reconciliation module 817 where, as previously discussed, with particular reference to FIG. 17, a reconciliation process may be completed using performance calculations that compare actual or measured values to those provided by the asset or asset model. As illustrated, via the data router 818, the model matching parameters and adjusted data may be sent to a centrally located infrastructure, such as the Internet. the remote central database 819. From there, the model tuning parameters are used to tune the offline power plant model 820, which can then be used as described above to optimize future plant inventory operation, provide alternative scenarios or what-if analyzes, and between possible or competing ones To advise on operating modes of the asset portfolio.
The results of the analyzes performed using the offline power plant model 820 can be communicated to the existing plant operators via a web portal 821, as illustrated. Web portal 821 may provide custom access 822 to users for inventory management. Such users may include plant operators, energy dealers, owners, plant operators, engineers, and other stakeholders. According to the user interaction via the web portal access, decisions can be made regarding the recommendations offered by the analyzes performed using the offline power plant model 820.
Figures 24 and 25 illustrate a schematic system configuration of a unified architecture according to certain alternative aspects of the present invention. As illustrated in FIG. 25, a remote central inventory and analysis component 825 may receive the power and measured operating parameters of multiple assets 802 to perform inventory level optimization. Inventory-level optimization may be based on additional input data, which may include, for example: the fuel quantities currently stored and available at each power plant, the site-specific price of fuel for each power plant, the site-specific price of electricity in each power plant the current weather forecast and disparities between remote assets and / or outage and maintenance schedules. For example, a planned component overhaul for a gas turbine may mean that short-term operation at higher temperatures is more economical. The process can then calculate a supply curve that includes optimized variable generation costs for the power plant inventory. In addition, as illustrated, the present invention may allow for more automated quotation preparation so that, at least in some circumstances, the quotation may be transmitted directly to the systemwide load balancing authority 826, thereby bypassing the energy traders 809. As illustrated in FIG. 25, the power system optimization results (via the systemwide load balancing authority) may be used to create an asset schedule that reflects how the various assets in the power system should be integrated with the predicted demand fulfill. This asset schedule may reflect a system-wide optimization and, as illustrated, may be communicated back to the asset pool 802 owners so that asset setpoints and asset operation modes can be communicated to the controller controlling each asset in the system.
Accordingly, methods and systems of FIGS. 22-25 may be developed whereby a number of power plants operating within a competitive power system are optimized for increased performance and improved bidding for future market periods. Up-to-date data regarding operating conditions and parameters can be received in real time from each of the power plants within the plant inventory. The power plant and / or asset inventory models can then be tuned according to current data, further improving model accuracy and predictive range. As will be appreciated, this can be achieved by comparing measured performance indicators and corresponding values predicted by power plant models. As a next step, the plant level co-ordinated power plant models and / or models may be used to calculate the true production capabilities for each of the power plants within the inventory based on competing operating modes simulated with the tuned models. Optimization then takes place using the true plant performance capabilities and optimization criteria defined by the asset or asset owner. When determining an optimized operating mode, an asset schedule can be created that calculates optimal operating points for each of the power plants within the asset inventory. As will be understood, the operating points may then be transferred to different power plants to control each of them in accordance therewith or, alternatively, the operating points may also serve as the basis on which tenders are prepared for submission to the central load distribution authority.
Referring also to the centralized control and optimization of multiple power units, FIGS. 26 and 27 illustrate a power system 850 in which a block controller 855 is used to control multiple power plant blocks 860. The power plant blocks 860, as indicated, may define an asset pool 861 of the assets. As will be understood, these embodiments provide a further exemplary application of the optimization and control methods described in detail above, although they include an extension of the optimization perspective to an asset inventory level. Thus, the present invention can also provide opportunities to reduce certain inefficiencies that still impact modern power generation systems, particularly those that have a large number of remote and diverse thermal power plants. Each of these assets may represent any of the thermal power plants discussed herein, such as gas and steam turbines, and associated subcomponents, such as HRSGs, intake conditioners, duct burners, etc. The assets may operate according to multiple generation configurations, depending on how the subcomponents are involved. The power generation from the plurality of power plant blocks 860 may be centrally controlled by a block controller 855. With respect to the system in FIG. 27, discussed in more detail below, the block controller 855 may control the system according to optimization processes that take into account the asset and power plant status as well as generation schedules, maintenance schedules, and other factors relevant to one of the assets or power plant units 860, including location-dependent variables. In addition, insights from operational data collected from similarly configured assets and power plant blocks, which are not part of the asset pool, can be used to refine control strategies.
Conventionally, conventional asset controllers (indicated as "DCS" in Figure 26) are local to the generation assets and operate in substantial isolation. As a result, such controllers do not take into account the current state of the other assets that make up power plant block 860 and / or asset 861. As will be understood, this lack of perspective leads to suboptimal power generation for asset stock 861, taking into account this perspective. Still referring to the methods and systems already described, particularly those with respect to FIGS. 3, 4, and 17 to 25, the present exemplary embodiment teaches an asset level control system that provides several system-wide benefits, including improved common power usage strategies, cost effectiveness and improved efficiency across grouped assets or powerhouse blocks.
As indicated, the control system may interact with the asset controllers as illustrated by the block controller 855. The block controller 855 may also communicate with the network 862, as well as with a central load balancing or other competent authority associated with its management. In this way, for example, supply and demand information can be exchanged between asset stock 861 and a central authority. According to an exemplary embodiment, supply information, such as e.g. Load balancing offers based on the optimization of asset stock 861 by the block controller. The present invention may further include optimization processes that occur between offer periods, which may be periodically used to optimize the way the asset stock 861 is configured to meet an already established load level. Specifically, such optimization between offers may be used to address dynamic and unanticipated operating variables. Appropriate control actions for the assets of the power plant blocks 860 may be communicated by the block controller 855 to the control systems within each of the power plant blocks 860 or more directly to the assets. According to preferred embodiments, the implementation of control solutions of the block controller 855 may include allowing it to override the asset controllers if certain predefined conditions are met. Factors that influence such transition may include variable generation costs for each of the power plant blocks / assets, the remaining part life of hot gas path components, changing demand levels, changing environmental conditions, and others.
The block controller 855, as illustrated, may be communicatively connected to the multiple power plant blocks 860 of asset 861, as well as directly to the assets, and thereby may receive many data inputs on which the control solutions described herein are based. The optimizations may take into account one or more of the following inputs: degradation of state and performance; Electricity production schedules; Power frequency; Maintenance and inspection schedules; Fuel availability; Fuel costs; Fuel consumption patterns and forecasts; Problems in the past and equipment breakdowns; true capabilities; Life models; Startup and connection characteristics; Measurement of operating parameter data, past and current; Weather information; Cost data, etc. As discussed in more detail with respect to other embodiments, the inputs may include detailed current and past data regarding measured operating parameters for each of the assets asset 861 assets. All such inputs, both past and present, may for example be stored in a central database according to conventional methods and thereby made available upon request by the block controller 855, possibly as required by one of the method steps described herein.
A cost function may be developed according to the preferences of a plant asset operator. In accordance with a preferred embodiment, a weighted average sum of a plant health robustness index may be used to determine preferred or optimized common power usage configurations. The asset robustness index may include, for example, optimization according to several factors that apply to a given demand or asset level output levels. These factors may include: thermal and mechanical loads; Deterioration or loss, including deterioration rate; Production costs and / or fuel consumption. In this way, the present embodiment can be used to address multiple ongoing plant inventory control issues, particularly optimizing performance across multiple power plant blocks with multiple and multiple generation assets.
The data inputs may include the types discussed herein, including those relating to computer modeling, maintenance, optimization and model-free adaptive learning. For example, in accordance with the present embodiment, computer models, transfer functions, or algorithms may be developed and maintained so that the operation (or aspects of operation) of the assets and / or collectively, power plant units, or asset inventory may be simulated under a variety of scenarios. The results of the simulations may include values for particular performance indicators that represent predictions regarding operating and performance of asset, power plant block, or asset performance over the selected operating period. The performance counters may be selected based on a known or developed correlation with one or more cost results, and may thus be used to compare the economic aspects of each simulation. A "cost" result, as used herein, may include any economic consequence, both positive and negative, associated with the operation of asset stock 861 over the selected operating period. Cost results may thus include any revenue earned from the generation of electricity over the period, as well as any operating and maintenance costs incurred by the asset inventory. These operating and maintenance costs may include a resulting deterioration of assets of the asset inventory in light of the scenarios and the simulated operation resulting therefrom. As will be understood, data extracted from the simulation results may be used to calculate which of the alternative operating modes for the asset inventory is more desirable or less costly.
The models for the assets, blocks, or assets may include algorithms or transfer functions developed by physics-based models, adaptive or learned "model-free" process input / output correlations, or combinations thereof. Baseline degradation or loss models can be developed that correlate process inputs / outputs with degradation or loss data for each asset type. The deterioration or loss data and related cost outcome may thus be calculable based on the predicted values for the operating parameters of the proposed alternative or competing operating modes for the inventory, which, according to certain embodiments, by the manner in which the assets and Powerhouse blocks are involved, the way in which the generation is divided over the asset inventory assets, and other factors described herein may be differentiated. As noted, the learning of similarly configured assets can be used to inform or further refine the models used as part of this process. For example, a degradation model may be developed that calculates accumulated equipment wear and losses in light of the values for selected performance indicators. Such degradation can then be used to calculate the economic consequences or the cost outcome for each of the competing modes of operation. These economic consequences may include asset degradation, component wear, consumed part life (ie, the portion of the life of a component consumed during a period of use), and other measures of value, such as emissions, regulatory costs Cost, fuel consumption and other variable costs that depend on the output level are included. As will be appreciated, since the degradation and consumption of part life for a particular asset may be non-linear, as well as dynamic and / or site-specific variables, significant cost savings over time can be achieved by distributing the output level of the asset pool to the asset pool Minimizing deterioration of the overall asset floor, especially if this minimization is spread across the assets, minimizing the impact on the productivity and efficiency of the overall asset portfolio.
Thus, considering conditions that are predicted for a future market period, which anticipated demand and environmental condition predictions may count, multiple concurrent asset operation modes may be selected for analysis and / or simulation to optimize, or at least preferred to determine asset inventory operating mode. Each of the competing plant inventory modes of operation may describe a unique plant configuration generation 861. The competing plant inventory modes of operation may be developed to include parameter sets and / or control settings that define the unique generation configurations by which a particular asset inventory output level is achieved. As mentioned earlier, there are a number of ways to select asset level output. First, it can be selected to reflect an already known asset level output level, for example, an output level established via a recently completed load sharing process, so that the optimization process can be used to determine an optimized asset inventory configuration that meets that particular output level , The asset level output level may also be selected according to an expected load level in the light of historical generation records, expected customer demand, and / or other predicted conditions. Alternatively, the asset balance output level can be varied over a selected range. In this way, variable production costs for asset stock 861 may be calculated and then used, for example, as part of a bidding process to feed into the preparation of a competitive bid as information. Thus, the way the plant inventory output level is defined by the plant operator can be used to support, in one case, activities related to preparing a competitive offer, while in another case the output level can be selected that it supports an advisory function that optimizes asset performance when actual conditions may differ from those anticipated.
[0172] In accordance with an example operation as indicated by the more detailed system of FIG. 27, parameter sets describing each of the competing plant inventory modes of operation may be developed, and for each of the competing plant inventory modes of operation, different scenarios or cases may be developed within which manipulable variables are varied over a selected range to determine the effect of the variation on the overall operation of the asset inventory. The different cases for the contending asset inventory operating modes may be configured to cover alternative ways in which the asset inventory output level is shared across the powerhouse blocks 860 and / or assets. As another example, the different cases may be selected based on alternative configurations available to particular ones of the assets, including the different ways in which each of the assets is involved. For example, some cases may involve the incorporation of certain subcomponents of the assets, such as e.g. Duct burners or inlet conditioners to increase power generation capabilities while recommending that other assets operate at the shutdown or part load levels. Other scenarios may explore situations in which these asset configurations vary somewhat or are reversed overall.
As illustrated in FIG. 27, the block controller 855 may communicate with a data and analysis component 865, which may include a plurality of modules through which relevant data is collected, normalized, stored, and made available to the block controller 855 upon request. A data recording module may receive real-time and historical data inputs from a monitoring system associated with generating assets. A performance monitoring module may also be included, and one or more offline models may be maintained with respect thereto. Each of these modules may operate substantially in accordance with other embodiments discussed herein. A learning module may also be included to collect operational data from similarly configured assets or power plant blocks that are not operating within asset stock 861. As will be understood, these data may support a learning function that will provide a deeper and more thorough understanding of the assets of the assets. Such data may also be used to normalize measured data collected from asset stock 861 so that performance degradation of the generation asset can be accurately calculated, which allows for the consideration of the effects of other variables, such as, e.g. Fuel properties, environmental conditions, etc., which may also affect performance capacity and efficiency.
As described with respect to FIGS. 24 and 25, asset level optimization may be based on location-dependent variables. These variables may reflect conditions that are unique and apply to particular particular assets or power plant blocks, and may include, for example, the following: the actual amounts of fuel stored and available in each asset; the site-specific price of fuel for each asset; the site-specific market price for electricity generated in each asset; current weather forecasts and the differences between remote assets within the asset portfolio; and outage and maintenance schedules for each asset. For example, a planned component overhaul for a gas turbine asset may mean that short-term operation at higher temperatures is more economically advantageous. As illustrated, the data and analysis component 865 may include a module to account for these differences.
The block controller 855 may further include, as indicated, modules directed to power generation models (which may include asset models, block models, asset inventory models, and degradation or loss models), an optimizer, and a cost function. The asset, power plant block, and / or asset inventory model may be created, reconciled, and / or reconciled and maintained in accordance with the methods already described herein. These models may be used to simulate or otherwise predict the operation of the asset stock, or a selected portion thereof, over the selected operating period so that the optimizer module is able to determine a preferred scenario according to a defined cost function. More specifically, the results from the simulations may be used to calculate a cost outcome for each of them, which may include summation of revenue, operating costs, degradation, consumed part life, and other costs referred to herein across the power plant and / or asset stocks. As will be understood, the revenue may be determined by an extrapolated output level multiplied by a market unit price. The calculation of the costs as indicated may include degradation models or algorithms that correlate an economic result with the way the assets operate within the simulations. The performance data from the simulation results may be used to determine plant-wide operating costs, degradation, and other losses, as previously described.As will be understood, certain cost considerations, such as e.g. fixed aspects of operating costs, not appreciably between the competing asset inventory operating modes, and thus can be excluded from such calculations. In addition, the simulations described herein may be configured to include all or part of the assets of assets, and may focus on limited aspects of asset operations that have been shown to be particularly relevant to the prediction of cost outcomes, as provided herein are.
[0176] According to certain embodiments, the cost function module may include a plant inventory robustness index to efficiently distinguish between alternative operating modes. The asset robustness index may represent an averaged accumulation of losses accumulated within the power plant units. The robustness index may include a factor indicative of cost in terms of consumed part life, which is a summation of the consumed part life over the assets, such as the part lifetimes. Hot gas path parts and compressor blades in gas turbines, can act away. For example, a generation set scheduled for shutdown during the selected operating period according to one of the competing plant inventory modes of operation results in an economic loss corresponding to the consumed part life for each shutdown / startup operation. Instead, a production asset scheduled to operate at full load during the same period of operation may incur a loss equal to those operating hours. As will be understood, such losses may be further calibrated to specifically reflect the thermal and mechanical loads expected in view of the load level and operating parameters predicted to meet a given load level, for example, from such factors as predicted Environmental conditions, fuel properties, etc. may depend. Additional economic losses may be included in the accumulation of asset inventory losses to derive a cost outcome for each of the competing asset inventory operating modes. This may include a summation of fuel consumption for asset inventory assets, and, for example, the economic impact of predicted emission levels given the simulation results.
[0177] After completing summation of asset-wide revenue and / or losses for each of the simulated scenarios, the present method may include the step of calculating one or more preferred or optimized cases. The present method may then include one or more outputs relating to the preferred or optimized cases. For example, the preferred or optimized cases may be communicated electronically to a plant inventory operator, such as an operator. via the user interface 866. In such cases, the outputs of the present method may include: a power plant block / asset condition advisory; a performance sharing recommendation; a failure planner; an optimal set point control solution for the power plant blocks; DCS bridging; and / or an expected generation schedule. The output may also have automated control behavior, which may include automatically skipping one of the asset controllers. According to another alternative, an output may include creating a load distribution offer according to one or more of the preferred or optimized cases. As will be appreciated, the outputs of the method, as indicated at the user interface 866, may provide various possibilities for asset savings. First, for example, preferred power division configurations may minimize, reduce, or advantageously split plant inventory degradation, which may have a significant impact on generation capability and efficiency over future operating periods. Second, an advisory function using the described components may be configured to optimize or at least improve maintenance intervals, thereby reducing degradation losses, both recoverable and unrecoverable. The monitoring and prediction of the deterioration rate and the effective scheduling / execution of maintenance operations, e.g. Compressor washes or filter cleaning ensure that the gas turbine works as efficiently as possible.
Referring to Figure 28, another related aspect of the present invention is discussed which describes the more specific example of controlling the operation of multiple gas turbine engines operating as a power plant block. As will be understood, the gas turbine engines may be located in a particular power plant or across multiple remote power plants. As discussed earlier, controlling a gas turbine engine block to optimize or improve power sharing is a challenge. Actual control systems do not effectively synchronize across a block of multiple engines, and instead essentially engage each of the motors individually based on a simple allocation of the output level for which the power plant block is collectively responsible. As will be understood, this often leads to imbalances and inefficient rates of deterioration. Accordingly, there is a need for more optimal control strategies, and in particular, a system controller that provides efficient power sharing strategies across multiple gas turbines that promote a more cost effective loss or degradation rate when the units are collectively considered to be a power plant block. For example, if a gas turbine block has multiple engines of the same power, the present invention may provide recommendations based on the current state of wear of the engines, which units should operate at higher output levels and which should operate at reduced levels. The present invention may achieve this in accordance with aspects already discussed herein, particularly those discussed with respect to FIGS. 24-27. As one of ordinary skill in the art will understand, the benefits of such functionality include: increased life and performance of the gas turbines; improved lifetime prediction, which can enable more competitive and / or risk-sharing service contracts; higher operational flexibility for the power plant block as a whole; and robust multi-objective optimization that efficiently accounts for operational compromises, including, for example, the consumption of hot gas path component life, current degradation levels and degradation rates, and the current power generation performance, such as power consumption. Demand, efficiency, fuel consumption, etc.
One way in which this can be achieved is according to a system 900, which will now be described with reference to FIG. 28. As indicated, multiple gas turbines 901 may be operated as part of a power plant block or "block" 902. As discussed as part of the systems above, the operating parameters 903 for each of the assets 901 may be collected and electronically communicated to a block controller 904. According to a preferred embodiment, the operating parameters may include a rotor speed, a compressor pumping limit and a blade tip spacing. As will be appreciated, the compressor pumping limit may be calculated relative to the measured rotor speed, and the blade tip spacing may be measured according to any conventional method, including, for example, microwave sensors. As another input, the block controller may receive the recordings 905 from a database component, such as a user. any of those already discussed which may record current and past operational parameter measurements, including rotor speed, surge limit, blade tip distance, control settings, environmental condition data, etc., to adaptively correlate process inputs and outputs.
[0180] According to preferred embodiments, the block controller 904 may be configured to operate as a model-free adaptive controller. The model-less adaptive controller may have a neural network-based structure having inputs (for example, via the records 905) of each of the gas turbines that meet demand, specific heat consumption, and so on. As will be appreciated, modelless adaptive control is a particularly effective control method for unknown time discrete nonlinear systems with time varying parameters and time varying structures. The design and analysis of model-free adaptive control places a focus on process inputs and outputs to "learn" predictive correlations or algorithms that explain the relationships between them. Correlations between measured inputs and outputs of the system are controlled. By operating the block controller 904 in this manner, control commands or recommendations are derived, and these may be communicated as an output 906 for implementation to a master controller 907. According to a preferred embodiment, the output 906 from the block controller 904 includes a preferred or optimized power sharing command or recommendation. According to further embodiments, the output 906 may include commands or recommendations regarding a modulated coolant flow for hot gas path components of the gas turbines 901 and / or modulated IGV settings for the compressor units of the gas turbines 901.
The master controller 907 may be communicatively coupled to the gas turbines 901 of the power plant block 902 to implement control solutions in the face of the output 906. As illustrated, the master controller 907 may also communicate such information to the block controller 904. Thus configured, the control system of FIG. 28 may be operated to control the multiple gas turbines of the power plant block 902 to provide a combined load or output level, such as a contract output level, as may be determined by a load sharing offer process the gas turbines are jointly responsible - is generated in an improved or optimized manner according to a defined cost function. This control solution may include recommending a percentage of the combined output level that each of the gas turbines should contribute. In addition, the master controller 907 may include a physics-based model for controlling the gas turbines according to the optimized mode of operation, as previously discussed.
For example, according to an exemplary embodiment, distance and surge limit data may be tracked for each of the gas turbines. If the margin or surge margin data for one of the gas turbines is determined to be above a predefined threshold, that particular turbine may be operated at a reduced load. If operating this gas turbine at a reduced level is not possible, other recommendations may be made, such as: modulating IGV settings or the coolant flow to hot gas path components. On the other hand, if one of the gas turbines has been selected for operation at a reduced level, the optimized generation configuration may include recommending that one or more of the other gas turbines operate at a higher / peak load to compensate for any deficit. The method may select the higher / peak load turbines based on surge margin and distance data, with the desired effect of balancing current degradation levels and degradation rates between the gas turbine engines of the power plant block to collectively extend operating life while achieving higher fuel efficiency Block output level and efficiency are maintained. As mentioned, because performance degradation rates and the consumption of part life may be non-linear and depend on parameters that are variable over geographically dispersed units, savings may be made by using the block level perspective described herein to split the load in a manner that requires Cost result for the block optimized to be achieved. The power generation can thus be distributed such that the cost over the block 902 is accommodated by taking into account real-time data (in particular surge margin and distance data) that is considered highly disposable and efficient in evaluating the power degradation levels, degradation rates, remaining part life, and true power capacity for the gas turbines Blocks were determined to be optimized.
Referring to FIG. 29, another exemplary embodiment of the present invention includes systems and methods that provide for more efficient and / or more optimized shutdown of combined cycle power plants. As will be appreciated, during shutdown of a combined cycle gas turbine, a controller typically gradually reduces the fuel flow to the gas turbine to reduce the rotor speed to a minimum speed. This minimum speed may be referred to as the "rotary speed" because it represents the speed at which the rotor engages a rotating device and is thereby rotated to prevent thermal bending of the rotor during the turn-off period. Depending on the nature of the gas turbine engine, the fuel flow may be stopped at about twenty percent of the normal full speed with the rotating device engaged at about one percent of full speed. Reducing the fuel flow in this gradual manner, however, does not provide a direct relationship to reducing rotor speed. Rather, large and rigid variations in the speed of the rotor over the shutdown period are common. The variations in rotor speed may then cause significant differences in fuel-to-air ratio due to the fact that air ingress is a function of rotor speed, whereas fuel flow is not. Such variations can then lead to significant and abrupt variations in firing temperatures, transient temperature gradients, emissions, coolant flow, and others. The variations in turn-off behavior may have an effect on the turbine spacing and thus on the overall turbine performance and component life.
Therefore, there is a desire for a CCGT shutdown controller that improves system shutdown by overcoming one or more of these problems. Preferably, such a controller would control the deceleration rate of the turbine rotor and related components over time to minimize non-uniform shutdown variations and thereby minimize the negative impact on the engine systems and components. According to certain embodiments, a more effective controller functions to optimize the rotor loads and rotor speed slew rate and torque. The shutdown controller may also correct variability in subsystems such that, for example, the coolant flow and the wheelspace temperature remain at preferred levels. According to preferred embodiments, the present method may control the deceleration rate of the rotor and related components over time to minimize the shutdown variations so as to reduce cost, equipment losses, and other adverse effects. According to the systems and methods already described, the control methodology may operate to optimize factors that affect shutdown costs according to user-defined criteria or cost functions. One way in which this can be achieved is according to a process 920, which will now be described with reference to FIG. As will be apparent to one of ordinary skill in the art, aspects of the process 920 are based on articles discussed herein, particularly with reference to the discussion of FIGS. 3 and 4, which, for the sake of brevity, are summarized but not completely repeated.
[0185] According to one embodiment, the shutdown operations and / or the GuD shutdown controller of the present invention are configured as a conventional looping controller. The controller of the present invention may include model-free adaptive control and model-based control aspects as described in the appended claims. The GuD shutdown controller may include a target shutdown time controller and an actual shutdown time controller, and may control most, if not all, aspects of system shutdown. The controller may receive inputs, e.g. Exhaust gas distribution, wheelspace temperature, distance, surge limit, steam and gas turbine rotor loads, gas turbine rotor deceleration rate, demand, fuel flow, current power production, grid frequency, secondary firing, drum level and so on. Based on these inputs, the shutdown controller may calculate a time range for shutdown (e.g., shutdown speed), rotor deceleration rate, corrected coolant flow, and corrected swirl throttle profile, and / or a generator reverse torque desired during shutdown, as described in greater detail below. According to certain embodiments, each of these outputs may be used to compensate for potentially harmful shutdown variations detected by one of the power plant sensors. The CCGT controller may provide a trajectory of the profile of the RPM / slew rate over time, as well as a profile of the deceleration rate compared to the current power production that is better suited for shutdown operation, and both may account for CCGT systems, e.g. HRSG, steam turbines, boilers and the like. The shutdown controller may control the components described herein until the rotator speed is reached, thereby providing for more optimal steam turbine and HRSG operability conditions while also reducing component loads.
Fig. 29 is a flowchart showing one embodiment of a process 920 used to shut down a combined cycle power plant, such as a gas turbine power plant. the power plant 12 described with reference to FIG. 3, is suitable. The process 920 may be implemented as computer code executable by the GuD shutdown controller and may be initiated upon receipt (step 921) of a shutdown command. The shutdown command may be received, for example, based on a maintenance event, a fuel change event, and so on. The process 920 may then retrieve a current state of the plant components (step 922) that may be captured, compiled, stored, and retrieved by any of the sensors, systems, and / or methods already described herein. The current state of plant components may include, for example, turbine rotor speed, component temperature, exhaust gas temperature, pressures, flow rates, distances (i.e., distances between rotating and stationary components), vibration measurements, and the like. The state of the plant may also include current electricity production and cost data, such as costs of not producing electricity, costs of electricity to market rates, green credits (e.g., carbon credits), and the like.
In a next step, costing or loss data may be retrieved (step 923), for example by querying a variety of systems, including accounting systems, futures trading systems, power market systems, or a combination thereof. Historical data can also be retrieved (step 924). The historical data may include system performance data, maintenance data, historical asset data (e.g., logs from other components in facilities located at different geographic locations), inspection reports, and / or historical cost data.
The process may then derive algorithms regarding plant shutdown degradation or loss related to the shutdown operation (step 925). Such discharges may be determined using multiple input types, including, for example, historical operating data related to gas turbine systems, steam turbine systems, HRSG units, as well as any other subcomponents that may be present. According to certain embodiments, a variety of models or algorithms may be developed by which GTO shutdown losses are derived. As discussed more fully in the discussion with respect to FIG. 4, such algorithms may be used to calculate a summation of power plant shutdown losses based on values for selected operating parameters or performance indicators, such as those shown in FIG. Temperatures, pressures, flows, distances, loads, vibration, turn-off time and the like to provide. As will be understood, in accordance with the other embodiments described herein alternative, proposed or competitive shutdown modes for the combined cycle power plant may be simulated in a combined cycle power plant model. That is, a combined cycle power plant model can be developed, tuned and maintained and then used to simulate alternative or competitive shutdown modes of operation to derive predicted values for certain predefined performance counters. The predicted values for the performance parameter can then be used to calculate shutdown costs according to the derived loss algorithms.
For example, algorithms may be developed that correlate shutdown losses and a predicted thermal stress profile that may be determined from the predicted values for particular performance parameters in view of the operating parameters related to one of the competitive shutdown modes of operation. According to certain embodiments, such losses may reflect an overall economic consequence of the competitive shutdown mode of operation, and may, for example, account for degradation of hot gas path components and / or a percentage of consumed part life in the light of the shutdown mode, as well as any resulting performance degradation of the equipment, and may vary from the initiation of shutdown may be calculated at an achieved shutdown time, which may be, for example, the time when the rotary speed is reached. Similarly, loss algorithms can be developed to determine losses related to: mechanical compressor and turbine loads; Distances between stationary and rotating parts; Abschaltemissionen; Abschaltbelastungen; thermal steam turbine rotor / stator loads; Boiler drum pressure gradient etc.
The process 920 may then derive an improved or optimized shutdown mode of operation (step 926) which includes a profile of the RPM / slew rate with time and / or a profile of the deceleration rate versus current power production required to shut down the CCGT. Power plant is particularly well suited, may include. For example, the optimized shutdown mode of operation may be determined as the one that best performs and accounts for the operation of the steam turbine, HRSG unit, boiler, and / or other components of the gas and steam combined cycle plant. In one embodiment, the controller receives the aforementioned inputs and derives therefrom expected conditions having different RPM / rise time parameters to derive a RPM / slope curve plotted along a time axis that minimizes loads and / or optimizes shutdown costs. Also, the profiles of deceleration rate versus current power production may include a fuel flow that more desirably (compared to other proposed shutdown modes) improves actual power production based on the deceleration of the shaft. According to further embodiments, cost functions may be defined by which other, more preferred or optimized shutdown modes of operation are derived and selected. As will be understood, scenarios may be derived which, for example, minimize gas turbine loads and / or losses, minimize steam turbine loads and / or losses, minimize loads and / or losses for the HRSG, or a combination thereof. Depending on how the cost function is defined, further optimized shutdown modes may be based on criteria such as e.g. Costs of electricity production during the shutdown period, plant emissions, fuel consumption and / or combinations thereof, without limitation, are determined.
The CCGT shutdown controller may further include a control system for shutting down the power plant according to the optimized shutdown mode. According to a preferred embodiment, this control system may include a physics-based modeler or model-based controller, which then derives a control solution in the light of the optimized shutdown mode. The model-based controller can derive control inputs and settings for the control of actuators and control devices so that the combined cycle power plant is operated during the shutdown period in accordance with the preferred or optimized shutdown mode of operation. For example, the shutdown controller may actuate fuel valves to effect a desired fuel flow while also controlling the gas turbine exhaust gas swirl throttles to control the exhaust flow into the HRSG while also controlling the steam valves of the steam turbine to control the steam turbine shutdown. By combining the control of a plurality of components at approximately the same time, the shutdown for the plant can be improved and can match desired scenarios.
While the invention has been described in conjunction with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not intended to be limited to the disclosed embodiment, but to cover various modifications and equivalent arrangements which are included within the spirit and scope of the appended claims.

权利要求:
Claims (20)
[1]
A control method for optimizing the operation of a power plant having power plant blocks during a selected operating period, the selected operating period being subdivided to include periodic intervals within which each of the power plant blocks has either an on state or an off state, wherein unique combinations of which of the power plant blocks have the on-state and which the off-state define competing modes of operation within the intervals, the control method comprising the steps of:Determining a preferred case for each of the competing modes of operation for each of the intervals;based on the data relating to the preferred cases, selecting proposed part-load operating sequences for the selected operating period, each of the proposed part-load operating sequences describing a unique progression of the off-state and on-state for the power plant blocks over the intervals of the selected operating period ;for each of the proposed partial load operating sequences, determining a shutdown operation for each of the power plant blocks having the off state for one or more intervals during the selected operating period, and calculating an economic shutdown result;for each of the proposed partial load operating sequences, determining a partial load operation for each of the power plant blocks having the on state for one or more intervals during the selected operating period, and calculating therefrom an economic partial load result;in view of the economic shutdown and partial load result, calculating an economic sequence result for each of the proposed partial load operating sequences; andComparing the economic sequence results, and based thereon, outputting a preferred part-load operating sequence.
[2]
2. The control method of claim 1, wherein the power plant blocks comprise gas turbines;wherein the selected operating period includes a future low load operating period for the power plant and the operation comprises a power plant part load; andwherein at least one of the proposed part-load operating sequences includes the off-state for at least one of the gas turbines during at least one of the intervals.
[3]
3. The control method of claim 2, wherein the economic sequence result is a summation of the respective economic cutoff results and partial partial cost results for the shutdown and part load operations described by the progression of the off-state and on-state for the gas turbines, specific to one of the proposed Part load operating sequences includes.
[4]
The control method of claim 3, wherein the step of determining the preferred cases includes the steps of:Selecting the competing modes of operation by configuring various possible onesCombinations as to which of the gas turbines have the on-state and which the off-state during the intervals;Defining a plurality of cases for each of the competing modes of operation, the plurality of cases including varying a value of an operating parameter over a range;Receiving performance goals that include a cost function for evaluating the operation of the power plant during the intervals of the selected operating period;Receiving an environmental condition prediction for each of the intervals of the selected operating period;for each of the multiple cases of competing modes of operation, simulating the operation of the plant over each of the intervals with a power plant model according to the value of the operating parameter and the environmental condition prediction;Evaluate a simulation result from each of the simulations according to the cost function to select therefrom a preferred case from the multiple cases for each of the competing modes of operation.
[5]
The method of claim 4, wherein the step of defining the competing modes of operation includes:Permuting the gas turbines to configure an on / off permutation matrix for each of the intervals, wherein permutations of the permutation matrix describe a unique combination in terms of which of the gas turbines have the on state and which the off state during the interval; andfor each interval, defining each of the permutations of the permutation matrix as one of the competing modes of operation.
[6]
6. The control method of claim 5, wherein the permutation matrix for each of the intervals is configured such that the gas turbines each have only one of the on state and the off state, excluding each other for the entire duration of the interval; andwherein the on / off permutation matrix comprises each possible unique combination as to which of the gas turbines has the on-state and which the off-state during one of the intervals of the selected operating period.
[7]
7. The control method of claim 5, wherein the step of defining a plurality of cases for each of the competing modes of operation includes varying a value for a first operating parameter over a first range and a value for a second operating parameter over a second range.
[8]
8. The control method of claim 7, wherein the step of simulating with the power plant model of each of the plurality of competing modes of operation includes generating suggested parameter sets for each particular case of the plurality of cases as input data for the power plant model;where, for each particular case, the proposed parameter set includes:the value within the first range for the first operating parameter and the value within the second range for the second operating parameter for the particular case;the on-state and off-state for the gas turbines for the competitive mode of operation, which corresponds to the particular case; andEnvironmental condition prediction data for the interval corresponding to the particular case.
[9]
9. The control method of claim 8, wherein the simulating step with the power plant model of each of the plurality of cases includes performing a simulation run with the power plant model in accordance with the suggested parameter sets, the simulation run simulating the operation of the power plant during the interval according to the input data of the power plant model configured parameter sets is configured; andthe performance goals further comprising functional limitations; andwherein the step of evaluating the simulation results from the simulation runs comprises determining which, if any, of the simulation results violates any of the functional limitations, and disqualifying for consideration as one of the preferred cases of any of the multiple cases that yielded the simulation results, which have violated the limitations of functionality.
[10]
10. The control method of claim 8, wherein the first operating parameter includes a swirl throttle adjustment and the second operating parameter comprises a turbine exhaust temperature.
[11]
11. The control method of claim 10, wherein the cost function comprises a total fuel consumption by the gas turbines such that determining the preferred case includes determining which of the plurality of cases yielded the simulation results that minimize the overall fuel consumption during the interval.
[12]
12. The control method of claim 10, wherein the cost function comprises the generation output level for the gas turbines such that determining the preferred case comprises determining which of the plurality of cases yielded the simulation results that minimize the generation output level during the interval.
[13]
13. The control method of claim 8, wherein the step of selecting the proposed partial load operating sequences comprises determining possible partial load operating sequences according to transient operating limitations for each of the gas turbines; and theSelecting the proposed part-load operating sequences from the possible part-load operating sequences based on the comparison of an economic aspect of the preferred cases of a first of the intervals with the economic aspect of the preferred cases of a second of the intervals.
[14]
14. The control method of claim 12, wherein the step of determining shutdown operation for the gas turbines for each of the proposed partial load operating sequences includes:Determining cases of continuous shutdown operation for each of the gas turbines, the continuous shutdown operation comprising one of the gas turbines with the off state over two or more consecutive ones of the intervals; andDetermining a shutdown period for each case of the shutdown operation for each of the gas turbines, wherein in the cases of continuous shutdown operation, the shutdown period comprises summing the intervals over which the continuous shutdown operation occurs; andwherein the step of determining the partial load operation for the gas turbines for each of the proposed partial load operating sequences includes determining a generation output level for each of the gas turbines for each of the intervals.
[15]
15. The control method of claim 14, wherein calculating the economic cutoff result comprises calculating the cutoff operation and maintenance costs in view of the cutoff operation set for the gas turbines in each of the proposed part load operating sequences;wherein calculating the partial operating economical result comprises calculating the partial load operating and maintenance costs in view of the partial load operation set for the gas turbines in each of the proposed partial load operating sequences.
[16]
16. The control method of claim 15, wherein calculating the shutdown operation and maintenance costs comprises:based on a length of the shutdown periods, determining a shutdown / start up type for each of the gas turbines for each of the shutdown periods;wherein the shutdown / start up type includes a turning device procedure, the turning device procedure comprising rotating turbine rotor wheels of the gas turbines during cooling of the turbine rotor wheels; andwherein the shutdown / start up type comprises a start up procedure for restarting the gas turbines.
[17]
17. The control method of claim 16, wherein calculating the shutdown operation and maintenance costs comprises:Calculating fuel costs in view of the rotary device procedure and the startup procedure for each of the shutdown periods;Calculating component life cost for each of the shutdown periods, wherein the component life cost includes a component lifetime portion of a gas turbine component consumed in the light of the shutdown operation;Calculating the cost of a delayed start-up that reflects a penalty for a delayed start and a probability of occurrence of the penalty in view of a historical start-up reliability of each of the gas turbines;Calculating costs related to an emissions impact for each of the shutdown periods.
[18]
18. The control method of claim 17, wherein calculating the partial load operation and maintenance costs comprises:Calculating fuel costs in view of the generation output level for each case of partial load operation;Calculating a yield for the generation output level for each case of the partial load operation;Calculating component life cost for each case of part load operation, wherein the component life cost comprises a component lifetime portion of a gas turbine component consumed in view of the part load operation; andCalculate costs related to an emissions impact for each case of partial load operation.
[19]
19. The control method of claim 18, wherein the emissions impact includes an indication of potential emissions costs including at least one predicted emission level for each of the cases of part load operation and shut down operation within the selected operation period.
[20]
The control method of claim 18, wherein the emissions impact includes an indication of potential emission costs, including:cumulative power plant emission levels generated for the power plant during a current regulatory period, and regulatory limits for the current regulatory period; anda penalty for violation of regulatory boundaries.

类似技术:

---

公开号 | 公开日 | 专利标题

---

CH710433A2|2016-06-15|A method for controlling power units.

---

CH710432A2|2016-05-31|A method of operation of power-heat generating units.

---

US10287988B2|2019-05-14|Methods and systems for enhancing operation of power plant generating units and systems

---

US10534328B2|2020-01-14|Methods and systems for enhancing control of power plant generating units

---

US9957843B2|2018-05-01|Methods and systems for enhancing control of power plant generating units

---

US9926852B2|2018-03-27|Methods and systems for enhancing control of power plant generating units

---

JP6755673B2|2020-09-16|Power plants Methods and systems for increased control of power generation units

---

US9960598B2|2018-05-01|Methods and systems for enhancing control of power plant generating units

---

US9945264B2|2018-04-17|Methods and systems for enhancing control of power plant generating units

---

JP6679281B2|2020-04-15|Method and system for enhancing control of a power plant power generation unit

---

DE102011051671A1|2012-01-12|Optimization system using an iterative expert engine

---

DE102011051673A1|2012-01-12|Power Management System

---

US10620598B2|2020-04-14|Methods and systems for controlling generating units and power plants for improved performance

---

DE102015120446A1|2016-06-02|Methods and systems for improving the control of power generator units

---

DE102018106606A1|2018-10-04|Real-time command and real-time operations for a gas turbine delivery optimizer

---

Jackson2008|Energy Budgets at Risk |: a risk management approach to energy purchase and efficiency choices

---

Kaufman et al.2009|Industrial energy optimization: managing energy consumption for higher profitability

---

Julin2019|Demand response in commercial buildings

---

Dean et al.2021|Enhancing Performance Contracts with Monitoring-Based Commissioning |

---

Rantaniemi2015|Productizing of advanced service for power plants

同族专利:

---

公开号 | 公开日

---

US20160146118A1|2016-05-26|

---

US9404426B2|2016-08-02|

---

JP6774180B2|2020-10-21|

---

GB2534981A|2016-08-10|

---

GB201520797D0|2016-01-06|

---

JP2016105687A|2016-06-09|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

DE102020210840A1|2020-08-27|2022-03-03|Siemens Aktiengesellschaft|Method for estimating future power consumption within a consumer network using a power prediction system, computer program product, computer-readable storage medium and power prediction system|US5278772A|1992-05-06|1994-01-11|Honeywell Inc.|Real-time economic load allocation|

---

US6278899B1|1996-05-06|2001-08-21|Pavilion Technologies, Inc.|Method for on-line optimization of a plant|

---

US6167547A|1996-06-20|2000-12-26|Ce Nuclear Power Llc|Automatic self-test system utilizing multi-sensor, multi-channel redundant monitoring and control circuits|

---

JP3764269B2|1998-03-20|2006-04-05|株式会社東芝|Disk storage|

---

JPH11273254A|1998-03-20|1999-10-08|Toshiba Corp|Disk storage device|

---

US6164057A|1999-03-16|2000-12-26|General Electric Co.|Gas turbine generator having reserve capacity controller|

---

EP1217473A1|2000-12-21|2002-06-26|Abb Research Ltd.|Optimizing plant control values of a power plant|

---

JP2004048961A|2002-07-15|2004-02-12|Tokyo Electric Power Co Inc:The|Apparatus, method, and program for computing cost for generator operation|

---

US7454270B2|2003-05-13|2008-11-18|Siemens Power Transmission & Distribution, Inc.|Dynamic economic dispatch for the management of a power distribution system|

---

JP4115958B2|2004-03-26|2008-07-09|株式会社東芝|Plant operation schedule optimization method and optimization system|

---

US7288921B2|2004-06-25|2007-10-30|Emerson Process Management Power & Water Solutions, Inc.|Method and apparatus for providing economic analysis of power generation and distribution|

---

US7356383B2|2005-02-10|2008-04-08|General Electric Company|Methods and apparatus for optimizing combined cycle/combined process facilities|

---

US7974826B2|2005-09-23|2011-07-05|General Electric Company|Energy system modeling apparatus and methods|

---

US8807055B2|2005-11-05|2014-08-19|Clearchem Development, Llc|Control of combustion system emissions|

---

US8140279B2|2007-09-24|2012-03-20|Budderfly Ventures, Llc|Computer based energy management|

---

US8396608B2|2007-09-24|2013-03-12|Budderfly Ventures Llc|Computer based energy management|

---

US7870299B1|2008-02-06|2011-01-11|Westinghouse Electric Co Llc|Advanced logic system|

---

US8504407B2|2008-02-12|2013-08-06|Justin Sacks|Economic impact analysis and supplier interface system|

---

US7805207B2|2008-03-28|2010-09-28|Mitsubishi Electric Research Laboratories, Inc.|Method and apparatus for adaptive parallel proportional-integral-derivative controller|

---

US8126629B2|2008-04-25|2012-02-28|General Electric Company|Method and system for operating gas turbine engine systems|

---

US8352148B2|2008-05-21|2013-01-08|General Electric Company|System for controlling input profiles of combined cycle power generation system|

---

WO2009156299A2|2008-06-26|2009-12-30|Alstom Technology Ltd|A method of estimating the maximum power generation capacity and for controlling a specified power reserve of a single cycle or combined cycle gas turbine power plant, and a power generating system for use with said method|

---

US8080887B2|2008-09-19|2011-12-20|Brad Radl|Closed loop control of hydrogen cooling of an electric power generator|

---

US8510013B2|2009-05-04|2013-08-13|General Electric Company|Gas turbine shutdown|

---

US8255066B2|2009-05-18|2012-08-28|Imb Controls Inc.|Method and apparatus for tuning a PID controller|

---

CN102893227A|2009-12-31|2013-01-23|Abb研究有限公司|Method and control system for scheduling load of a power plant|

---

US8135568B2|2010-06-25|2012-03-13|General Electric Company|Turbomachine airfoil life management system and method|

---

US8538593B2|2010-07-02|2013-09-17|Alstom Grid Inc.|Method for integrating individual load forecasts into a composite load forecast to present a comprehensive synchronized and harmonized load forecast|

---

US8639480B2|2010-09-20|2014-01-28|General Electric Company|Methods and systems for modeling turbine operation|

---

US8816531B2|2011-01-27|2014-08-26|General Electric Company|Systems, methods, and apparatus for integrated volt/VAR control in power distribution networks|

---

US9348325B2|2012-01-30|2016-05-24|Johnson Controls Technology Company|Systems and methods for detecting a control loop interaction|

---

EP2642089B1|2012-03-19|2016-08-24|General Electric Technology GmbH|Method for operating a power plant|

---

US9002530B2|2012-04-23|2015-04-07|General Electric Company|Power plant operation enhancement|

---

US20140150438A1|2012-11-30|2014-06-05|General Electric Company|System and method for operating a gas turbine in a turndown mode|

---

US9477214B2|2013-03-07|2016-10-25|General Electric Company|Plant control systems and methods|

---

US8797199B1|2013-05-16|2014-08-05|Amazon Technologies, Inc.|Continuous adaptive digital to analog control|

---

US20150184549A1|2013-12-31|2015-07-02|General Electric Company|Methods and systems for enhancing control of power plant generating units|

---

US9957843B2|2013-12-31|2018-05-01|General Electric Company|Methods and systems for enhancing control of power plant generating units|JP6094675B2|2013-07-11|2017-03-15|トヨタ自動車株式会社|Hybrid vehicle|

---

JP6366259B2|2013-11-18|2018-08-01|三菱日立パワーシステムズ株式会社|Control device and control method for two-shaft gas turbine|

---

JP6375585B2|2014-03-31|2018-08-22|三菱日立パワーシステムズ株式会社|Combined cycle plant, control method thereof, and control device thereof|

---

US9985437B2|2014-09-26|2018-05-29|Enrichment Technology Company Ltd., Zweigniederlassung Deutschland|Combined electrical power plant|

---

US10626748B2|2014-12-08|2020-04-21|General Electric Company|System and method for predicting and managing life consumption of gas turbine parts|

---

US9927820B2|2015-04-09|2018-03-27|Honeywell International Inc.|Heating, ventilation, and air conditioning system boiler controller|

---

US10126810B2|2015-09-29|2018-11-13|Emerson Process Management Power & Water Solutions, Inc.|Method for controlling power generation unit to desired output as specified by load demand signal by using modified control signal|

---

JP6786233B2|2016-03-22|2020-11-18|三菱パワー株式会社|Gas turbine characterization device and gas turbine characterization method|

---

US10033316B2|2016-09-30|2018-07-24|General Electric Company|System and method for model based turbine shaft power predictor|

---

US10774753B2|2016-10-21|2020-09-15|General Electric Company|Indirect monitoring of aircraft combustor dynamics|

---

US10566881B2|2017-01-27|2020-02-18|Franklin Electric Co., Inc.|Motor drive system including removable bypass circuit and/or cooling features|

---

US10452041B2|2017-03-31|2019-10-22|General Electric Company|Gas turbine dispatch optimizer real-time command and operations|

---

KR101893689B1|2017-04-26|2018-08-30|두산중공업 주식회사|Gas Turbine System and Controlling Method thereof|

---

US10620598B2|2017-06-12|2020-04-14|General Electric Company|Methods and systems for controlling generating units and power plants for improved performance|

---

KR102103324B1|2017-07-21|2020-04-22|가부시끼가이샤 도시바|Plant control apparatus, plant control method and power generation plant|

---

US10789657B2|2017-09-18|2020-09-29|Innio Jenbacher Gmbh & Co Og|System and method for compressor scheduling|

---

US11041446B2|2018-02-23|2021-06-22|Rolls-Royce Corporation|Gas turbine engine fuel additive control system|

---

US10642262B2|2018-02-27|2020-05-05|Woodward, Inc.|Anomaly detection and anomaly-based control|

---

CN109118024B|2018-09-25|2021-07-27|国网福建省电力有限公司|Power system multi-resource scheduling method considering multi-stage state transfer of thermal power generating unit|

---

CN109687470A|2018-12-03|2019-04-26|三峡大学|Based on electricity price-load coordination and response co-generation unit wind electricity digestion dispatching method|

---

US11069926B1|2019-02-14|2021-07-20|Vcritonc Alpha, Inc.|Controlling ongoing battery system usage via parametric linear approximation|

---

US20220065184A1|2020-08-31|2022-03-03|Garrett Transportation I Inc.|Control system with diagnostics monitoring for engine control|

---

CN112684778A|2020-12-24|2021-04-20|武汉第二船舶设计研究所|Steam generator water supply system diagnosis method based on multi-source information reinforcement learning|

法律状态:
2017-03-15| NV| New agent|Representative=s name: GENERAL ELECTRIC TECHNOLOGY GMBH GLOBAL PATENT, CH |

---

2019-03-15| AZW| Rejection (application)|

优先权:

---

申请号 | 申请日 | 专利标题

---

US14/555,130|US9404426B2|2013-12-31|2014-11-26|Methods and systems for enhancing control of power plant generating units|

---