• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    A system includes machines (70) and a protection monitoring system (72) that includes a process configured to: analyze a tendency of one or more data measurements of the machines to one or more patterns that is a potential future error within the machines to indicate the tendency; and providing an error prediction based on the analysis of the trend.
    公开号:CH710181A2
    申请号:CH01377/15
    申请日:2015-09-22
    公开日:2016-03-31
    发明作者:William Randolph Shinkle；Hardev Singh；David Michael Boudreaux
    申请人:Gen Electric；
    IPC主号:
    专利说明:

    State of the art
    The subject matter disclosed herein relates to industrial gas turbine control systems and, more particularly, to system level error prediction in the industrial control system.
    Certain systems, such as industrial control systems, may provide control capabilities that enable the execution of control instructions in various types of devices, such as sensors, pumps, valves, and the like. In addition, certain industrial control systems may include one or more graphical user interfaces that may be used to present to the operator details about the various devices present in the control system network. For example, an operator graphical user interface may present warnings that may include alarm or diagnostic information about a device in the control system network.
    Short description
    Certain embodiments that are within the scope of the originally claimed invention are summarized below. These embodiments are not intended to limit the scope of the claimed invention, but these embodiments are intended to provide only a brief illustration of possible forms of the invention. In fact, the invention may include many forms that may be the same or different than the embodiments below.
    In one embodiment, a system includes: machines; and a protection monitoring system comprising a process configured to: analyze a tendency of one or more data measurements of the machines to one or more patterns indicative of a potential future error within the machines; and providing an error prediction based on the analysis of the trend.
    [0005] In a second embodiment, a tangible, non-transitory, machine-readable medium includes instructions for: obtaining data measurement trends relating to one or more characteristics of a piece of equipment; Analyzing the data measurement trends to identify one or more patterns indicative of a potential future error within the machines; and providing an error prediction based on the analysis of the trend.
    In a third embodiment, a method includes: predicting, via a computer processor, a potential machine error by: obtaining data measurement trends relating to one or more indicia of a piece of machinery; Analyzing the data measurement trends to identify one or more patterns indicative of a potential future error within the machines; and providing an error prediction based on the analysis of the trend.
    Brief description of the drawings
    These and other features, aspects and advantages of the present invention will become more apparent upon reading the following detailed description with reference to the accompanying drawings, in which like reference characters represent like parts throughout the drawings; show it:<Tb> FIG. 1 <SEP> is a schematic diagram of an industrial control system, including an error prediction / protection system, according to one embodiment;<Tb> FIG. 2 <SEP> is a schematic diagram of an industrial control system, including a system-level error prediction / protection system, according to one embodiment;<Tb> FIG. 3 <SEP> is a schematic drawing of an industrial control system having a turbine system including an error prediction / protection system, according to one embodiment;<Tb> FIG. 4 is a flowchart illustrating a process for executing control within the industrial control system based on a predicted fault according to an embodiment;<Tb> FIG. 5 is a flowchart illustrating a system level prediction process for an error according to one embodiment;<Tb> FIG. FIG. 6 is a diagrammatic view illustrating data that may be used to predict an error, according to an embodiment; FIG.<Tb> FIG. 7 <SEP> is a schematic view of a cloud-based fault prediction / protection system according to one embodiment; and<Tb> FIG. 8 is a flowchart illustrating a process for controlling the industrial control system based on prediction confidence and / or urgency according to an embodiment.
    Detailed description
    Referring to FIG. 1, an embodiment of an industrial process control system 10 is illustrated. The control system 10 may include a computer 12 adapted to perform various field device configuration and monitoring applications and provide a user interface through which an engineer or technician may monitor the components of the control system 10. The computer 12 may be any type of computing device that is suitable for executing software applications, such as a laptop, a workstation, a tablet computer, or a portable handheld device (eg, PDA or cellular phone). In fact, the computer 12 may include any of a variety of hardware and / or operating system platforms. In one embodiment, the computer 12 may host industrial control software, such as human machine interface (HMI) software 14, manufacturing execution system (MES) 16, distributed control system (DCS) 18, and / or supervisor control and data Acquisition (SCADA) system 20. For example, computer 12 may host the ControlST ™ software distributed by General Electric Co., Schenectady, New York.
    Further, the computer 12 is communicably connected to a plant data highway 22, which is suitable for enabling communication between the computer 12 and other computers 12 shown in the system. In fact, the industrial control system may include a plurality of computers 12 interconnected by the plant data highway 22. The computer 12 may also be communicably connected to a unit data highway 24 suitable for communicatively coupling the computer 12 to industrial controllers 16. The system 10 may include other computers coupled to the plant data highway 22 and / or the unit data highway 16. For example, embodiments of the system 10 may include a computer 28 that performs virtual control, a computer 30 that has an Ethernet Global Data (EGD) configuration server, an Object Linking and Process Control (OPC) Data Access (DA) server, an alarm server or a combination thereof, a computer 32 hosting a General Electric Device System Standard Message (GSM) server, a computer 34 hosting an OPC Alarm and Events (AE) server, and a computer 36 hosting an alarm viewer , Other computers coupled to the plant data highway 22 and / or the unitary data highway 24 may include computers hosting Cimplicity ™, ControlST ™ and ToolboxST ™ distributed by General Electric Co., Schenectady, New York.
    The system 10 may include any number and suitable configuration of industrial controllers 26. For example, in some embodiments, the system 10 may include one industrial controller 26, two industrial controllers 26, three or more industrial controllers for redundancy. The industrial controls 26 may facilitate control logic that may be useful in automating a variety of plant equipment, such as a turbine system 38. In fact, the industrial controls 26 can communicate with a variety of devices, including, but not limited to, Temperature sensors, flow meters, temperature sensors, vibration sensors, clearance sensors (for example, measuring distances between a rotating component and a fixed component) and pressure sensors. The industrial controller 26 may also communicate with electric actuators, switches (eg, Hall switches, magnetic switches, relay switches, limit switches, etc.).
    In the illustrated embodiment, the turbine system 38 is communicably coupled to the automation controller 26 using connection devices 46 and 48 suitable for connection between an I / O network 50 and a HI network 52. For example, the connection devices 46 and 48 may include the FG-100 connection device sold by Softing AG, Haar, Germany. In some embodiments, a connection device, such as the connection device 48, may be coupled to the I / O NET via a switch 54. In such an embodiment, other components that are also coupled to the I / O network 50, such as one of the industrial controllers 26, may also be coupled to the switch 54. Accordingly, data transmitted and received through the I / O network 50, such as a 100 megabit (MB) high-speed Ethernet (HSE) network, may again be passed through the HI network 52, such as 31.25 kilobits per second Network, transmitted and received. That is, the connection devices 46 and 48 may act as bridges between the I / O network 50 and the HI network 52.
    Many devices may be connected to the industrial controller 26 and the computer 12. For example, the devices, such as the turbine system 38, may include industrial devices such as Foundation Fieldbus devices that include support for the bidirectional Foundation HI communication protocol. In such an embodiment, in addition, a Foundation Fieldbus power supply 53, such as a Phoenix Contact Fieldbus Power Supply, sold by Phoenix Contact, Middletown, PA, may also be coupled to the HI network 52 and may be connected to a power source such as DC or AC. be coupled. The power supply 53 may be suitable for powering the turbine 38 and for facilitating communication between the turbine 38 and other devices of the system 10. Conveniently, the HI network 52 may carry both power and communication signals (eg, alerts) over the same wiring with minimal communication interference. The turbine 38 may also include support for other communication protocols, such as those included in the HART® Communications Foundation (HCF) protocol and the Profibus User Organization e.V. (PNO) protocol.
    Each of the connection devices 46 and 48 may include one or more segment ports 56 and 58 useful in segmenting the HI network 52. For example, the connection device 46 may use the segment port 56 for communicable coupling to the turbine 38, while the connection device 48 may use the segment port 58 for communicable coupling to the other devices of the system 10. Distributing the input / output between the turbine 38 using, for example, the segment ports 56 and 58 may provide physical separation that is useful in maintaining fault tolerance, redundancy, and improving communication time. In some embodiments, additional devices may be coupled to the I / O network 50. For example, in one embodiment, an I / O pack 60 may be coupled to the I / O network 50. The I / O pack 60 may provide attachment of additional sensors and actuators to the system 10.
    The system 10 may include one or more error prediction / protection systems 62 that act to predict future errors within the industrial control system 10 and / or to provide control of the industrial control system 10 to prevent future errors. For example, the error prediction / protection system 62, as discussed in more detail below, may predict and / or prevent stalls of the turbine 38 using data obtained from the sensors of the turbine 38. To do so, fault prediction / protection system 62 may be communicably coupled to turbine system 38 (eg, via direct data collection from sensors of turbine system 38 or via coupling to other components (eg, one or more controllers 26 in communication with turbine system 38) In the event of a future failure, the fault prediction / protection systems 62 may provide for notification (eg, trigger an alarm or alert) and / or take other preventative measures, such as controlling one or more operations of the turbine system 38.
    As indicated above, the fault prediction / protection system 62 is communicatively coupled to the system (eg, turbine system 38) that it is tasked with monitoring. FIG. 2 illustrates a high level view of the control system 10 that is automated to predict and / or protect against errors. As shown in FIG. 2, the fault prediction / protection system may be distributed in one or more areas of the control system 10.
    The illustrated industrial control system 10 includes machines 70, a protection monitoring system 72, and a control system 26. In other embodiments, the industrial automation system 10 may include additional devices, such as monitors. As illustrated, the error prediction / protection system 62 may include machine-readable instructions stored on a tangible, non-transitory, machine-readable medium that may be integrated with or incorporated into one of the components of the control system 10 (eg, the machines 70 and / or the controller 26) can.
    In addition, certain embodiments of the industrial control system 10 may incorporate the error prediction / protection system 62 as incorporated into the machines 70 and / or the controller 26. In some embodiments, the error prediction / protection system 62 may be separate and distinct from other portions of the industrial automation system.
    The error prediction / protection system 62 limits the controller 26 to operate the machines 70 to achieve desired goals within various limitations of the machines 70. In other words, the controller 26 and / or the error prediction / protection system 62 may be used to protect the machine Machines 70 may be used against damage due to operating parameters that vary beyond safety values. For example, in certain embodiments, controller 26 may include a MARK® VI or MARK® VIe controller from General Electric®, Schenectady, New York. In some embodiments, the protection monitoring system 72 may include a protection monitoring system that is similar to a 3500 Series Machinery Protection System with Bently Nevada <TM> Asset Condition Monitoring, also distributed by General Electric®, Schenectady, New York. The machines 70 may include mechanically functioning parts of the industrial automation system 10 or subsystem (eg, the turbine system 38). For example, the engines may include engines, compressors, combustors, conveyors, generators, inlet guide vanes (IGVs), pumps, turboexpanders, etc. In embodiments where the industrial control system 10 includes the gas turbine system 38, the engines 70 may include a turbine and / or other mechanically functioning parts (eg, compressors).
    Additionally, the fault prediction / protection system 62 may include various system diagnostic and monitoring devices (eg, sensors, transducers, connections therebetween, etc.). The protection monitoring system 72 monitors functionality and connectivity of the fault prediction / protection system 62. In other words, the protection monitoring system 72 verifies that the machines 70 are being properly protected by the protection system 62. In some embodiments, the fault prediction / protection system 62 may include the protection monitoring system 72 as substantially incorporated into a system that protects the machines 70 and monitors the status of the protection. In certain embodiments, the protection monitoring system 72 may be implemented using software stored on a computing device (eg, an electronic device having a processor). For example, in some embodiments, the protection monitoring system 72 may be stored as instructions stored on a machine-readable medium, such as a memory, hard disk, drive, optical drive, or other type of memory. In certain embodiments, these instructions may be stored and / or executed by the controller 26, an operator machine, or a remote server. In other embodiments, the protection monitoring system 72 may be implemented using hardware circuitry. For example, in some embodiments, the protection monitoring system 72 may be located in a housing that encases the controller 26. As discussed below, if the protection monitoring system 72 determines that the error prediction / protection system 62 has made a mistake, the protection monitoring system 72 may address the error and / or warn a user / operator to indicate that the machines 70 are not protected ,
    FIG. 3 illustrates one embodiment of a control system 10 with a more detailed view of the turbine system 38. The turbine system 38 may be used to provide power, such as electrical and / or mechanical power. Certain of the turbine systems 38, such as the LMS100 turbine system 38 sold by General Electric Co. of Schenectady, New York, may include an intercooler 80. It will be appreciated that the turbine system 38 may be any turbine system that is configured to convert fuel to a rotational force. Accordingly, various arrangements of the turbine components may be used, and FIG. 3 describes a representative example. The intercooler 80 may increase the efficiency of the turbine system 38 by, for example, cooling a hot fluid (eg, compressed air) supplied from a low pressure (LP) compressor 82 and passing the cooled fluid (eg, compressed air) into a high pressure (LP) compressor. HP) compressor 84. For example, a fluid, such as air, may enter an inlet 86 and be compressed by LP compressor 82. The air compressed by LP compressor 82 may gain thermal energy (i.e., heat) during compression. For example, the compressed air may be at a temperature of approximately between 50 ° and 300 ° Celsius. The higher temperature air may then be directed into the intercooler 80. The intercooler 80 may include two chambers, such as an inner tube bundle chamber 88 and an outer shell chamber 90. The hot air may enter the outer shell chamber 90 and exchange heat with the inner tube bundle chamber 88, thereby reducing the temperature of the hot air. In certain embodiments, the inner tube bundle chamber 88 may flow a cooling fluid, such as water, to transfer heat from the hot air to produce cooler air. The cooler air may then be directed into the HP compressor 84. By cooling the air supplied to the HP compressor 84, higher energy efficiencies can be achieved. For example, the cooler air may reduce the compression work in the HP compressor 84 as mass flow of air into the turbine system 38 is increased, thereby increasing the overall efficiency.
    A seal 92 is disposed between the inner tube chamber 88 and the outer shell chamber 90 to block fluid flow (eg, leakage) between both chambers 20 and 22. The chamber 88 and / or 22 may thermally expand and contract, in some cases having a movement between about 0.5 cm and 10 cm. In certain embodiments, the seal 92 may also expand to accommodate expansion of the chamber 88 while also maintaining a suitable barrier against fluid flow between the chambers 88 and 90. In fact, the seal 92 may expand and contract to properly block fluid flow (eg, leakage); also at sealed locations with bends or bends, thereby improving the overall efficiency of the turbine system 38.
    As shown, the turbine system 38 may include a combustor 94 that receives and burns a fuel / air mixture for generating hot pressurized exhaust gases. The turbine system 38 directs the exhaust gases through a high pressure (HP) turbine 96 and a low pressure (LP) turbine 98 toward an exhaust gas outlet 100. The HP turbine 96 may be part of an HP rotor. Likewise, LP turbine 98 may be part of an LP rotor. As the exhaust gases flow through the HP turbine 96 and the LP turbine 98, the gases force turbine blades to rotate about a drive shaft 102 along an axis of the turbine system 38. As shown, the drive shaft 102 is connected to various components of the turbine system 38, including the turbine HP compressor 84 and LP compressor 82. It should be understood that other turbine systems may include intermediate pressure compressors, intermediate pressure turbines, and systems having a different arrangement of components, including shaft assemblies and couplings to the generator 104.
    The drive shaft 102 may include one or more shafts that may be concentrically aligned, for example. The drive shaft 102 may include a shaft that connects the HP turbine 96 to the HP compressor 84 to form an HP rotor. The HP compressor 84 may include blades coupled to the drive shaft 102. Therefore, rotation of the turbine blades in the HP turbine 96 causes the shaft connecting the HP turbine 96 to the HP compressor 84 to rotate blades within the HP compressor 84. This compresses air in the HP compressor 84. Likewise, the IP drive shaft 102 includes a shaft that connects the IP turbine 97 to the LP compressor 82 to form an LP rotor. The LP compressor 82 includes blades coupled to the IP shaft 102. Therefore, rotation of turbine blades in the IP turbine 97 causes the shaft connecting the IP turbine 97 to the LP compressor 82 to rotate blades within the LP compressor 82. The pressurized air is supplied to the combustor 94 and mixed with fuel to facilitate combustion at a higher efficiency. Thus, the turbine system 38 may include a dual concentric waveguide arrangement with the LP turbine 98 being drivably connected to the generator 104 by the drive shaft 103, while the HP turbine 96 is driven by a second shaft in the drive shaft 102 within the first shaft is and thus concentric, is drivingly connected to the HP compressor 84. The shaft 102 may also be connected to a power generator 104 or any other load, such as a mechanical load. The generator 104 may be connected to a power distribution network 106 that is suitable for distributing the electricity generated by the generator 104.
    As indicated above, the control system 10 may include one or more sensors 106 used in monitoring and / or controlling the control system 10. For example, in the current embodiment, the turbine system 38 includes a plurality of sensors 106 configured to provide operating data relating to one or more components of the turbine system 38 (eg, LP compressors 82, HP compressors 84, combustors 94, HP turbines 96, LP turbine 98, etc.). Data from the sensors may be communicated to the controller 26, HMI 14, or the error prediction / protection system 62 for monitoring and controlling the turbine system 38.
    As described in greater detail below, the error prediction / protection system 62 may use these data from the sensors 106 to detect the likelihood of future failure (eg turbine stall, etc.). By predicting errors prior to their occurrence, measures can be taken to reduce the likelihood of the error actually occurring. Accordingly, costly failures can be reduced.
    FIG. 4 illustrates an embodiment of a process 110 for protecting a turbine system 38 from a fault (eg, a power dip, shutdown, or trip). First, operating parameters are obtained from the engines by the fault prediction / protection system 62 (block 112) ). As noted above, this data may be obtained from sensors 106 of the turbine system 38 that are tasked with monitoring parameters of one or more components of the turbine system 38.
    The machine data 114 is analyzed to detect data patterns associated with machine failure (block 114). For example, in some cases, individual process parameter thresholds may indicate a potential machine error within the turbine system 38. In some cases, two or more process parameters for determining a potential machine error may be analyzed together. Embodiments of certain patterns that may correlate to a potential machine error will be described in more detail below.
    At decision block 116, if no data is detected that is correlative to a potential machine error, the process 110 returns to block 112 to obtain subsequent machine data. However, if data is detected that is correlative to a potential machine error, additional reporting and / or control may occur. In some embodiments, upon detecting such correlative data, an indication of a potential defect / error may be provided to the operator of the turbine block 38 (block 118). For example, an alarm or warning may be provided to the computer 30 (of FIGURE 1) that hosts the alarm server where an alarm or alert may finally be provided to the operator (eg, via the HMI 14).
    Additionally or alternatively, upon detecting data that is correlated with a potential machine error, the system 10 may control the machines to forestall the error (block 120). For example, the controller 26 (of FIG. 1) may operate control components (eg, electric actuators, Hall switches, solenoid switches, relay switches, limit switches, or other components) to affect one or more operational changes within the turbine system 38. As described below with respect to FIG. 8, the control system 10 may cause the machines to shut down prior to an error, thereby protecting the machines and / or the environment from error-related damage.
    FIG. 5 illustrates a process 130 for predicting a failure (eg, a power dip, shutdown, or a stall) of the turbine system 38 in the industrial control system 10 using current and historical data. As discussed above, the controllers 26 monitor and collect engine data of components of the system 10 (block 132). The data is stored in a historical database (for example, a database or file on non-transitory machine-readable media) (block 134). For example, the data may be stored on the computer 12 that hosts the HMI software 14, a computer 12 that is dedicated to storing and providing historical error data, or any other computer 12 within the system 10. The collected data can be reviewed and analyzed (block 136). This process of reviewing and analyzing the collected data may continue as the system 10 is operating, or may occur at set intervals (for example, every second, 1, 10, or 30 minutes). Based on the analyzed data and predefined relationships or correlations, the controller 26 or other processor-based processor may derive a prediction value based on the current data (eg, a running process prediction state for each process parameter) (block 138). The relationships or correlations may include mathematical equations, look-up tables, software models, or a combination thereof. For example, in some embodiments, the predefined ratios (eg, mathematical equations) may be weighed based on the severity of a deviation from normal operating parameters.
    The controller 26 or other computer with processor may then retrieve any stored historical data and derive a final prediction state based on the current process prediction state and historical data 100 (block 140). For example, the historical data 140 may include a historical error that may be applied directly to the current process inventory to obtain final status. For example, a number of times and / or a severity in which a parameter has deviated from normal operating parameters may be useful in predicting a future error. The more often and / or more seriously the deviation occurs, the more unstable the system is. Therefore, the final score may increase as the number of times and / or magnitude of severity of one or more parameters increases within the deviation of the system 10 from normal ranges.
    In some embodiments, the final predictive score may be based on several independent calculations. In other words, high prediction results resulting from anomalies (eg, sensor offsets or other sensor problems) can be filtered out using several independent means of creating verifiable stalls. If the several independent calculations verify each other, a prediction level may increase, while the level may be relatively lower in cases where the independent calculations do not verify each other.
    In some embodiments, the historical data events and parameters of the system 10 may correlate to previous performance drops, shutdowns, and / or shutdowns within the turbine system 38. It should also be noted that from time to time it is desirable to reset the historical data so that at least part of the historical data does not affect the final forecast level. In some embodiments, the operator may be able to reset at least a portion of the historical data so that the historical data will not be used in future predictions. In certain embodiments, the HMI software 14 allows the operator to archive or remove historical data relating to a specific parameter or component of the system 10. For example, the operator may select an option in the HMI software 14 to archive the historical data relating to the engine failing. After issuing a request to archive the historical data, a processor within the computer 12 may cause the historical data of the failed motor to be moved to archived storage (eg, a dedicated archive file or storage system). In some embodiments, the operator may simply request that the data be removed, thereby causing the processor to erase the historical data associated with the malfunctioning engine. Because historical data is very useful in predicting system-level errors, it may be beneficial to prevent unauthorized resets of historical data. Therefore, such functionality, whether in HMI 14 or elsewhere, may include password algorithms to ensure that historical data is only reset by authorized persons.
    After discussing certain features of the error prediction / protection system 62, the discussion now turns to certain data patterns that may indicate a potential failure of a turbine system 38. FIG. 6 is a diagrammatic view illustrating tendencies of data that may be used to predict an error, according to one embodiment. Diagram view 150 includes six diagrams: a high pressure compressor efficiency diagram 152, an exhaust temperature diagram 154 (eg, at low pressure turbine 98), power turbine inlet pressure diagram 156, high pressure compressor discharge pressure diagram 157, power output diagram 158, and turbine coefficient diagram 159 that provides a measurement = 15.0 * (T48-1100) / PS48, where T48 is the power turbine inlet temperature and PS48 is the power turbine inlet pressure.
    The high pressure compressor efficiency diagram 152 illustrates a plot of compressor efficiency over time. The x-axis 160 represents a compressor efficiency percentage, and the y-axis represents time (eg, in seconds or minutes). The inlet temperature diagram 154 illustrates a graph of temperature measurements (eg in degrees Fahrenheit) at a power turbine inlet 164 over time 162. The high pressure compressor exhaust pressure diagram illustrates pressure measurements 169 over time 162. The power output diagram 158 illustrates power output (eg, in megawatts) 172 over time 162 Turbine Coefficient Graph represents 15.0 * (T48-1100) / PS48173 over time 162.
    The pressure, temperature, efficiency and / or power output data may be useful alone or in relation to each other in predicting a failure. For example, analyzing data measurement over time may be useful in predicting errors. In one embodiment, a marked decrease in high pressure compressor efficiency 160, pressure 168, and / or power output 172 that exceeds an established threshold (eg, either preset or dynamically changing) may indicate the likelihood of an error. Further, a marked increase in temperature 164 that exceeds an established threshold may also indicate a likelihood of failure. In some embodiments (for example, a General Electric ™ LM6000 gas turbine), the threshold may be set to 1% change or a data reading indicating a rate of change of 1%. In embodiments in which a frame machine is monitored, the threshold may be set to 0.5% change or a data reading indicating a rate of change of 0.5%. In other words, the threshold may vary depending on the machines being monitored. Each piece of equipment / type of machine being monitored may include its own thresholds (eg, 5% change, 1% change, 15% change, etc.), which may be the same as the thresholds of other pieces / types of machines or not. For example, machines operating at higher temperatures / pressures may experience higher magnitude occurrences when a failure is likely to occur. The thresholds can be modified accordingly.
    A combination of time data measurements can increase the reliability of an error prediction. For example, in certain embodiments, a ratio of temperature 164 to pressure 168 may be calculated periodically and / or frequently. As indicated above, just prior to an error, temperature values can swing up and down pressure values. Accordingly, these data points can diverge when the rashes occur. Therefore, a quotient of the ratio of temperature 164L to_pressure 168 may indicate an error. For example, if a quotient changes beyond a certain threshold or percentage of change, the system may provide a prediction that an error will occur. Likewise, the quotient of a ratio of temperature 164 to power output 172 may be used similarly to predicting the error.
    Various data measurements may also be used to validate a prediction based on one or more indicators. For example, as the compressor efficiency 160 lowers, indicating that an error may occur, the system may validate this fault indication by searching for an increase in temperature 164, a decrease in pressure 168, and / or a decrease in power output 172. In one embodiment, a ratio of exhaust gas temperatures to a power turbine inlet pressure (eg, exhaust gas temperature - 1100 / power turbine inlet pressure) may be used to validate a prediction.
    It is important to note that while certain relationships between data measurements have been discussed, the discussion is not intended to limit the prediction to these specific data measurement ratios. In fact, many data point patterns, data measurement ratios, etc. can be used to predict an error. For example, ratios between one or more of compressor efficiency 160, temperature 164, pressure 168, and / or power output 172 (eg, with a threshold of change) may be used to predict an error.
    In some embodiments, data patterns, relationships, etc. that are useful for predicting potential machine errors may be recognized by alternative control systems. FIG. 7 illustrates a schematic view of a cloud-based error prediction / protection system 200 according to one embodiment. The cloud-based error prediction / protection system 200 may be used to communicate error prediction correlations between independent control systems 10.
    In the embodiment of FIG. 7, three independent control systems 10A, 10B, and 10C transmit error data 202 (eg, process parameters prior to a machine error) to a cloud computing unit 204. The cloud-based computing unit 204 may include one or more processors that recognize the error data 202 receive and analyze patterns among the received data. These data patterns may be useful for predicting subsequent potential errors within one of the control systems 10A, 10B and IOC. For example, if the control system 10A produces an error, an extract of process data prior to the incident (eg, 1 minute, 1 hour, 1 day, and so on prior to the incident) may be captured and communicated to the cloud 204. The cloud 204 may exploit the data 202 to determine potential patterns and / or anomalies of the data prior to the failure. Based on this exhaustion, the cloud may broadcast a potential error analysis 206 to the collection of independent control systems 10A, 10B, and 10C, or send the potential failure analysis 206 to a subset of the control systems 10A, 10B, and 10C (e.g., participating control systems, etc.).
    In some embodiments, one or more of the control systems 10 may perform the data analysis locally. For example, after experiencing a fault, the control system 10A may perform a local analysis of process data recorded prior to the fault. The analysis results 208 may be provided to the cloud 204, where the potential error analysis may be broadcast and / or sent as discussed above.
    By incorporating this cloud-based approach, additional data points may be used in the error prediction / protection system 62. Further, by determining whether the data analysis of a control system 10 is similar to the data analysis of a second control system 10, prediction confidence may be measured. If the data analysis between control systems 10 supports similar results and / or the results are repeatable on one or more of the control systems, predictive confidence may increase.
    By using the data measurements to predict errors, corrective actions can be taken before the error. This can increase system performance and reduce costs. For example, a foreign object may enter machinery and cause partial loss of wings. Over time, larger wing pieces may be lost, causing damage to the system. Using the prediction techniques described herein, errors can be predicted before the larger wing pieces cause damage. This can lead to reduced repair costs and increased system usage.
    In some embodiments, control actions may be selected based on error predictions based on an urgency of the prediction and / or a confidence level of the prediction. FIG. 8 illustrates a process 250 for controlling the engines after predicting an error (block 252) prior to the occurrence of the failure. As noted above, one or more staged approaches may be determined based on an urgency and / or trust level determination (eg, decision block 254). For example, the control system 10 may determine how likely (eg, trust) and / or how impending (eg, urgency) may be a predicted error. For example, urgency could be detected based on a number of data swings and / or the magnitude of one or more data swips. If there are data patterns that indicate that an error might occur but the confidence of the prediction is low, or the data indicates that the error is likely to occur at a much later time, the urgency level may be determined to be "low" , In circumstances where there is some confidence in the prediction (for example, a data pattern has been correlated with a past error) and / or the data indicates that the error is likely to occur in the near future but leaves enough time for a machine idle period, the Urgency to be set to "medium". In situations where the trust level is high (for example, many similar data patterns have been correlated to an error) and / or the data indicates that the error is likely to occur in the near future without enough time for a machine idle period, the urgency can be set to high become. In alternative embodiments, an urgency level determination could not be used.
    On the basis of the trust and / or the urgency, a certain control process can be carried out. For example, if the urgency and / or trust level is "low," it may be desirable to provide a message (eg, an alarm) (block 256) without changing the operation of the machines. In some embodiments, an alarm may be triggered on the HMI 14 (eg, FIG. 1). A remote alarm monitoring system may provide e-mail, SMS or other messages to control system operators and / or machine manufacturers. In some embodiments, if the trust level is "low," a manufacturer may be notified while a customer is not notified of the predicted failure. This can ensure that the prediction is conveyed while reducing false positives presented to the operators.
    In circumstances where the urgency and / or trust level is "middle", the control system 10 may control the machines to enter an idle phase (block 258). The idle phase may prepare the machines for the subsequent shutdown (block 260), for example, by reducing the operating speeds of the machines. By gradually reducing the speed of the machines, the integrity of certain components (eg, wings) of the machines can be maintained. Once the end of the idle phase is reached, the engines may be shut down (block 260).
    In circumstances where the thingness and / or trust level is "high", the control system 10 may control the shutdown engines (block 260) without entering the idle phase (block 258). This can help with quick shutdown of the machines to help avoid a mistake that is approaching quickly.
    In embodiments where no urgency level determination is used, any combination of the controls may be performed as desired by an operator and / or a manufacturer of the control system 10. For example, in such embodiments, the control system 10 may be configured to alert (block 256) and / or shut down the engines (block 260) to the prediction of a failure.
    In determining operations for potential errors within the control system 10, the control system 10 may use the prediction trust to determine a particular protection operation. For example, for a lower confidence level prediction analysis, the control system 10 may only alert the operator of the control system 10, while the control system 10 may change the operation of the turbine system 38 using a high confidence level prediction analysis when an error is predicted.
    Technical effects of the invention include a control system capable of preventing turbine failure by predicting errors prior to their occurrence. Based on this prediction, the operators of the control system may be warned of the potential problem and / or the control system may automatically move a change of operation within the control system. Preventive reporting and prevention of turbine failures can reduce repair and associated costs.
    This written description uses examples to disclose the invention, including the best mode, and to enable one skilled in the art to practice the invention, including making and using any apparatus or systems and performing any incorporated method. The patentable scope of the invention is defined by the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal language of the claims.
    权利要求:
    Claims (20)
    [1]
    A system comprising:Machinery; anda protection monitoring system comprising a processor configured to:Analyzing a tendency of one or more data measurements of the machines to one or more patterns indicative of a potential future error within the machines; andProvide an error prediction based on the analysis of the trend.
    [2]
    2. The system of claim 1, wherein the engines comprise a gas turbine.
    [3]
    3. The system of claim 1, wherein the data measurements relate to high pressure compressor efficiency, power turbine inlet temperature, power output, or any combination thereof.
    [4]
    4. The system of claim 3, wherein the processor is configured to calculate a ratio of trends from at least two data measurements.
    [5]
    The system of claim 4, wherein a quotient of the ratio is used in the analysis.
    [6]
    6. The system of claim 1, comprising a controller configured to invoke one or more control actions based on the error prediction.
    [7]
    7. The system of claim 1, wherein the processor is configured to determine a confidence level of the error prediction based on a frequency of prior occurrence of the one or more patterns associated with a subsequent error.
    [8]
    8. The system of claim 1, wherein the processor is configured to determine an urgency level with respect to the error prediction.
    [9]
    9. The system of claim 1, wherein the system comprises:several independent industrial control systems; anda cloud computing environment configured to:Receiving data measurements and error data from at least one of the independent industrial control systems;Analyzing the data measurements of the at least one independent industrial control system to associate at least one data pattern of the data measurements with the future potential error; andSupplying the at least one data pattern to the other industrial control systems so that the other independent control systems know that the at least one data pattern is associated with the future potential error.
    [10]
    10. A tangible non-transitory machine-readable medium comprising instructions for:Obtaining the data measurement trends relating to one or more identifiers of a piece of machinery;Analyzing the data measurement trends to identify one or more patterns indicative of a potential future error within the machines; and providing an error prediction based on the analysis of the trend.
    [11]
    The tangible, non-transitory, machine-readable medium of claim 10, comprising instructions for: receiving historical error data; and identifying the one or more patterns indicative of a potential future error based at least in part on the historical error data.
    [12]
    12. The tangible, non-transitory, machine-readable medium of claim 10, wherein the machines comprise a gas turbine.
    [13]
    The tangible, non-transitory, machine-readable medium of claim 12, wherein the characteristics of the piece of machinery include: high-pressure compressor efficiency, low-pressure inlet temperature, combustor inlet pressure, or output of the gas turbine.
    [14]
    14. The tangible, non-transitory, machine-readable medium of claim 10, wherein the instructions for analyzing the data measurement trends to identify one or more patterns include instructions for: identifying data excursions in the data measurement trends over a certain threshold of change.
    [15]
    15. The tangible, non-transitory, machine-readable medium of claim 10, wherein the determined change threshold is greater than 0.5%.
    [16]
    16. Method, comprising:Predicting a computer processor of a potential machine error by:Obtaining the data measurement trends relating to one or more identifiers of a piece of machinery;Analyzing the data measurement trends to identify one or more patterns indicative of a potential future error within the machines; and providing an error prediction based on the analysis of the trend.
    [17]
    17. The method of claim 16, comprising:Determining, via the computer processor, a confidence level of the error prediction; andSelecting a particular control action from a set of control actions based on the confidence level of the error prediction.
    [18]
    18. The method of claim 16, comprising:Determining, via the computer processor, an urgency level of the error prediction; andSelecting a particular control action from a set of control actions based on the urgency level of the error prediction.
    [19]
    19. The method of claim 16, wherein the one or more characteristics include: a high pressure compressor efficiency, a combustor inlet pressure, a low pressure inlet temperature, a power output, or any combination thereof.
    [20]
    20. The method of claim 16, comprising:Calculating a ratio of two or more of the one or more characteristics and using theRelationships for identifying the one or more patterns indicating the potential future error.
    类似技术:
    公开号 | 公开日 | 专利标题
    CH710181A2|2016-03-31|Systems and methods for error analysis.
    JP5295482B2|2013-09-18|Machine monitoring method and system
    KR102073912B1|2020-02-06|Method and system for diagnostic rules for heavy duty gas turbines
    CN103576640B|2019-04-30|System and method for monitoring pump cavitation
    US9360864B2|2016-06-07|Turbine fault prediction
    US6260004B1|2001-07-10|Method and apparatus for diagnosing a pump system
    US7840332B2|2010-11-23|Systems and methods for steam turbine remote monitoring, diagnosis and benchmarking
    EP2193413B1|2013-12-11|System for preserving and displaying process control data associated with an abnormal situation
    DK2175256T3|2016-03-29|Systems and methods for sensor-level machine monitoring
    US20100257838A1|2010-10-14|Model based health monitoring of aeroderivatives, robust to sensor failure and profiling
    EP1706718A2|2006-10-04|Method and apparatus to diagnose mechanical problems in machinery
    EP2975480A1|2016-01-20|Automated data overlay in industrial monitoring systems
    EP2026159A2|2009-02-18|A method and system for automatically evaluating the performance of a power plant machine
    JP2009115090A|2009-05-28|Automatic detection and notification of degradation of turbine internal component
    Kovalchuk et al.2019|Ensuring the reliable operation of the pumping units by efficient state diagnosis
    JP2006105053A|2006-04-20|Gas turbine monitoring device and gas turbine monitoring system
    Stanchi et al.2019|From Remote Monitoring to Predictive Diagnostic of Combined Cycle Power Plants
    Vilcu et al.2017|“HolderCPS”-A new type data recorder system for proactive maintenance to rotary blade machines
    Castellani et al.2014|Advanced data mining techniques for power performance verification of an on-shore wind farm
    Martin et al.1997|Integrating vibration, motor current, and wear particle analysis with machine operating state for on-line machinery prognostics/diagnostics systems |
    Vincent et al.2017|Health Management Solution Thanks to Wireless Vibration Sensors Network
    WO2012076427A1|2012-06-14|Method and device for the model-based monitoring of a turbomachine
    Snyder et al.2010|Predictive Analytics and Diagnostics Drive Effectiveness in Condition Based Monitoring
    Sheng2015|Improving Component Reliability Through Performance and Condition Monitoring Data Analysis; NREL |
    You et al.2004|Unified framework for facility condition monitoring, detection, and diagnostics
    同族专利:
    公开号 | 公开日
    JP6619186B2|2019-12-11|
    US20160091397A1|2016-03-31|
    US10037026B2|2018-07-31|
    JP2016070272A|2016-05-09|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    DE102017104884A1|2017-03-08|2018-09-27|Mts Consulting & Engineering Gmbh|System and method for determining error images from sensor data in product validation and manufacturing processes|JPH05187973A|1991-04-09|1993-07-27|Texas Instr Inc <Ti>|System and apparatus for estimation, monitoring, diagnosis and maintenance using portable computer system and radio communication|
    WO1996034207A1|1995-04-24|1996-10-31|United Technologies Corporation|Compressor stall diagnostics and avoidance|
    US5748500A|1995-11-14|1998-05-05|Electric Power Research Institute, Inc.|System to assess the starting performance of a turbine|
    US6301531B1|1999-08-23|2001-10-09|General Electric Company|Vehicle maintenance management system and method|
    US6917839B2|2000-06-09|2005-07-12|Intellectual Assets Llc|Surveillance system and method having an operating mode partitioned fault classification model|
    US6532433B2|2001-04-17|2003-03-11|General Electric Company|Method and apparatus for continuous prediction, monitoring and control of compressor health via detection of precursors to rotating stall and surge|
    US7089154B2|2001-08-09|2006-08-08|Rovsing Dynamics A/S|Automatic machinery fault diagnostic method and apparatus|
    US6843060B2|2001-11-13|2005-01-18|Goodrich Pump & Engine Control Systems, Inc.|Fault management system for gas turbine engines|
    AU2002353097A1|2001-12-07|2003-07-09|Battelle Memorial Institute|Methods and systems for analyzing the degradation and failure of mechanical systems|
    US6839660B2|2002-04-22|2005-01-04|Csi Technology, Inc.|On-line rotating equipment monitoring device|
    JP2004085316A|2002-08-26|2004-03-18|Mitsubishi Electric Corp|Failure determination device|
    GB0301707D0|2003-01-24|2003-02-26|Rolls Royce Plc|Fault diagnosis|
    US6871487B2|2003-02-14|2005-03-29|Kulite Semiconductor Products, Inc.|System for detecting and compensating for aerodynamic instabilities in turbo-jet engines|
    US7035763B2|2003-09-03|2006-04-25|Siemens Westinghouse Power Corporation|Systems and methods for selecting training data and generating fault models for use in use sensor-based monitoring|
    US7096159B2|2003-09-05|2006-08-22|Siemens Corporate Research Corp.|System and method for detecting and excluding outlier sensors in sensor-based monitoring|
    JP2006057595A|2004-08-23|2006-03-02|Hitachi Ltd|Gas turbine performance diagnosing system and its method|
    US7159401B1|2004-12-23|2007-01-09|Kulite Semiconductor Products, Inc.|System for detecting and compensating for aerodynamic instabilities in turbo-jet engines|
    US7222048B2|2005-04-21|2007-05-22|General Electric Company|Methods and systems for diagnosing machinery|
    EP1768007A1|2005-09-22|2007-03-28|Abb Research Ltd.|Monitoring a system having degrading components|
    US7949497B2|2007-04-12|2011-05-24|Siemens Corporation|Machine condition monitoring using discontinuity detection|
    US8155820B2|2008-03-10|2012-04-10|Eklund Neil H|Method, apparatus and computer program product for predicting and avoiding a fault|
    US7788209B2|2008-05-05|2010-08-31|United Technologies Corporation|Hybrid fault reasoning and guided troubleshooting system that uses case-based reasoning and model-based reasoning|
    TWI424832B|2008-12-15|2014-02-01|Proteus Digital Health Inc|Body-associated receiver and method|
    US8352152B2|2009-02-10|2013-01-08|Honeywell International Inc.|System, method, apparatus and computer program product for providing improved engine control|
    US20110313726A1|2009-03-05|2011-12-22|Honeywell International Inc.|Condition-based maintenance system for wind turbines|
    US8322145B2|2009-04-23|2012-12-04|General Electric Company|Systems and methods for providing surge protection to a turbine component|
    US8996334B2|2011-03-02|2015-03-31|General Electric Company|Method and system for analysis of turbomachinery|
    US9256224B2|2011-07-19|2016-02-09|GE Intelligent Platforms, Inc|Method of sequential kernel regression modeling for forecasting and prognostics|
    US20130046507A1|2011-08-19|2013-02-21|General Electric Company|Method and system for analysis of turbomachinery|
    US9110452B2|2011-09-19|2015-08-18|Fisher-Rosemount Systems, Inc.|Inferential process modeling, quality prediction and fault detection using multi-stage data segregation|
    US9360864B2|2012-04-11|2016-06-07|General Electric Company|Turbine fault prediction|
    SE536922C2|2013-02-19|2014-10-28|Basim Al-Najjar|A method and apparatus for predicting the condition of a machine or component of the machine|
    US9091616B2|2013-06-06|2015-07-28|Honeywell International Inc.|Engine operations support systems and methods for reducing fuel flow|EP3479239A4|2016-06-21|2020-02-19|Winant, Thomas, Arthur|System and method for determining the risk of failure of a structure|
    US10845792B2|2017-03-24|2020-11-24|Kabushiki Kaisha Toshiba|Failure diagnosis apparatus, monitoring apparatus, failure diagnosis method and recording medium|
    DE102017108496A1|2017-04-21|2018-10-25|Windmöller & Hölscher Kg|Method and devices and system for winding and unwinding a roll|
    US11073540B2|2018-04-24|2021-07-27|Hewlett Packard Enterprise Development Lp|Load disparity detection|
    US10819584B2|2018-05-03|2020-10-27|Servicenow, Inc.|System and method for performing actions based on future predicted metric values generated from time-series data|
    EP3582051A1|2018-06-12|2019-12-18|Siemens Aktiengesellschaft|Comprehensive fault analysis for control devices and industrial technical installations|
    CN109634233A|2018-12-06|2019-04-16|南京邮电大学|Industrial big data intellectual analysis decision-making technique, readable storage medium storing program for executing and terminal|
    CN112118260B|2020-09-16|2021-09-14|北京天融信网络安全技术有限公司|OPCDA message processing method, device, electronic equipment and storage medium|
    法律状态:
    2017-03-15| NV| New agent|Representative=s name: GENERAL ELECTRIC TECHNOLOGY GMBH GLOBAL PATENT, CH |
    2018-11-30| AZW| Rejection (application)|
    优先权:
    申请号 | 申请日 | 专利标题
    US14/497,156|US10037026B2|2014-09-25|2014-09-25|Systems and methods for fault analysis|
    [返回顶部]