• 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:
    SYSTEM FOR ESTIMATING AN ENGINE EVENT SITE AND METHOD FOR TRAINING A CONTROLLER These are systems and methods for estimating an engine event location. In one embodiment, a controller is configured to receive a signal from at least one knock sensor coupled to a reciprocating engine, transform the signal, using a multivariate transformation algorithm, into a spectral power density, transform the spectral density of power in a plurality of resource vectors with the use of predictive frequency bands, predict the location of the motor event with the use of at least the plurality of resource vectors and a predictive model and adjust the operation of the reciprocating engine with based on the engine event location.
    公开号:BR102016005985B1
    申请号:R102016005985-2
    申请日:2016-03-18
    公开日:2021-01-26
    发明作者:Iyad Batal；Jeffrey Jacob Bizub；Brett Alexander Matthews
    申请人:General Electric Company；
    IPC主号:
    专利说明:

    [0001] [001] The present invention relates to systems and methods for estimating an engine event location on a combustion engine. BACKGROUND OF THE INVENTION
    [0002] [002] Combustion engines typically burn a carbonaceous fuel, such as natural gas, gasoline, diesel and the like and use the corresponding expansion of high temperature and pressure gases to apply a force to certain engine components (for example, disposed piston in a cylinder) to move the components over a distance. Each cylinder can include one or more valves that open and close together with the combustion of the carbonaceous fuel. For example, an intake valve can direct an oxidizer, such as air, to the cylinder. A fuel mixes with the oxidizer and burns (for example, spark ignition) to generate combustion fluids (for example, hot gases), which then exit the cylinder through an exhaust valve.
    [0003] [003] The location (for example, timing or crank angle) of some engine events (for example, peak trigger pressure or intake and / or exhaust valve opening and closing) can affect fuel economy , power and other operating parameters. Unfortunately, using cylinder sensors to determine the location of such events can be expensive and uneconomical. DESCRIPTION OF THE INVENTION
    [0004] [004] A summary of certain achievements revealed in this document is presented below. It should be understood that these aspects are presented merely to provide the reader with a brief summary of these achievements and that these aspects are not intended to limit the scope of this invention. In fact, this invention can cover a variety of aspects that may not be presented below.
    [0005] [005] In one embodiment, a system for estimating a motor event location includes a controller configured to receive a signal from at least one knock sensor coupled to a reciprocating motor, transform the signal, using a transformation algorithm multivariate, in a spectral power density, transform the spectral power density into a plurality of resource vectors with the use of predictive frequency bands, predict the location of the motor event with the use of at least the plurality of resource vectors and a predictive model and adjust reciprocating engine operation based on the engine event location.
    [0006] [006] In another embodiment, a method for training a controller to estimate the peak trigger pressure location on a reciprocating engine includes receiving a first signal from at least one knock sensor, where the signal comprises at least data corresponding to a peak trigger pressure event. The method also includes receiving a second signal from a pressure sensor corresponding to a true peak trigger pressure location. Additionally, the method includes transforming the first signal into a power spectral density and comparing the power spectral density to the second signal to form predictive frequency bands. Finally, the method includes converting the spectral power density into a plurality of resource vectors and executing an algorithm to generate a predictive model that uses the plurality of resource vectors and the second sign, in which the predictive model is configured to estimate the peak trigger pressure location on the reciprocating engine during common engine operation.
    [0007] [007] In another embodiment, a system includes a reciprocating motor controller configured to receive a signal from at least one knock sensor coupled to the reciprocating motor and to transform the signal into a spectral power density using a transformation algorithm multivariate. The controller also transforms the power spectral density into a plurality of resource vectors using predictive frequency bands and predicts a peak trigger pressure location using at least the plurality of resource vectors and a model predictive. Finally, the controller is configured to issue a control action to at least the reciprocating engine based on the location of the peak trigger pressure. BRIEF DESCRIPTION OF THE DRAWINGS
    [0008] - a Figura 1 ilustra um diagrama em blocos de uma porção de um sistema de geração de potência acionado por motor que tem um motor de combustão reciprocante interno, de acordo com os aspectos da presente invenção; - a Figura 2 ilustra uma vista lateral em corte transversal de um conjunto de pistão-cilindro que tem um pistão disposto dentro de um cilindro do motor reciprocante da Figura 1, de acordo com os aspectos da presente invenção; - a Figura 3 ilustra um fluxograma de um método para treinar um sistema para estimar um local de um evento de motor através do desenvolvimento de um modelo preditivo e de bandas de frequência preditivas (“PFBs”), de acordo com os aspectos da presente invenção; - a Figura 4 ilustra um fluxograma de um método para usar ou testar o modelo preditivo e as PFBs da Figura 3 para determinar um local do evento de motor durante a operação comum de motor, de acordo com os aspectos da presente invenção; - a Figura 5 ilustra uma plotagem de sinal de pressão indicativa de um local verdadeiro do evento de motor, um sinal de sensor de detonação e um espectrograma com base no sinal de sensor de detonação, de acordo com os aspectos da presente invenção; - a Figura 6 é um esquema que ilustra como o sistema pode determinar um índice D de uma frequência específica, de acordo com os aspectos da presente invenção; - a Figura 7 é um diagrama que ilustra como o sistema constrói as PFBs com base em um valor de índice D, de acordo com os aspectos da presente invenção; - a Figura 8 ilustra três gráficos que podem ser usados para prever um tempo do evento de motor desejado, de acordo com os aspectos da presente invenção; - a Figura 9 é um histograma que inclui valores de erro para uma pluralidade de testes executados com o uso do modelo preditivo e das PFBs da Figura 3, de acordo com os aspectos da presente invenção; - a Figura 10 ilustra um primeiro gráfico representativo de uma pressão verdadeira dentro de um cilindro de motor ao longo do tempo e um segundo gráfico que exibe uma probabilidade de que um evento de pressão de disparo de pico ocorra ao longo do tempo para estimar um local da pressão de disparo de pico no motor, de acordo com os aspectos da presente invenção; e - a Figura 11 ilustra uma tabela que inclui valores que comparam um tempo estimado (por exemplo, grau de ângulo de manivela) do evento de motor com o uso do modelo preditivo e das PFBs da Figura 3 e de um tempo verdadeiro (por exemplo, grau de ângulo de manivela) do evento de motor, de acordo com os aspectos da presente invenção. [008] These and other features, aspects and advantages of the present invention will be better understood when the following detailed description is read with reference to the accompanying drawings in which similar characters represent similar parts throughout the drawings, in which: Figure 1 illustrates a block diagram of a portion of an engine-driven power generation system that has an internal reciprocating combustion engine, in accordance with aspects of the present invention; Figure 2 shows a cross-sectional side view of a piston-cylinder assembly that has a piston disposed within a cylinder of the reciprocating engine of Figure 1, according to the aspects of the present invention; - Figure 3 illustrates a flowchart of a method for training a system to estimate a motor event location by developing a predictive model and predictive frequency bands ("PFBs"), in accordance with the aspects of the present invention. ; Figure 4 illustrates a flow chart of a method for using or testing the predictive model and PFBs of Figure 3 to determine a location of the motor event during common motor operation, in accordance with aspects of the present invention; Figure 5 illustrates a pressure signal plot indicative of a true engine event location, a knock sensor signal and a spectrogram based on the knock sensor signal, according to the aspects of the present invention; Figure 6 is a schematic illustrating how the system can determine a D index of a specific frequency, in accordance with aspects of the present invention; Figure 7 is a diagram illustrating how the system builds PFBs based on an index value D, according to the aspects of the present invention; Figure 8 illustrates three graphs that can be used to predict a desired engine event time, in accordance with aspects of the present invention; - Figure 9 is a histogram that includes error values for a plurality of tests performed using the predictive model and the PFBs of Figure 3, according to the aspects of the present invention; - Figure 10 illustrates a first graph representing a true pressure inside an engine cylinder over time and a second graph showing a probability that a peak trigger pressure event will occur over time to estimate a location the peak trigger pressure in the engine, according to the aspects of the present invention; and - Figure 11 illustrates a table that includes values that compare an estimated time (for example, degree of crank angle) of the motor event with the use of the predictive model and the PFBs in Figure 3 and a true time (for example, degree of crank angle) of the motor event, in accordance with aspects of the present invention.
    [0009] [009] One or more specific embodiments of the present invention will be described below. In an effort to provide a concise description of these achievements, all features of an actual deployment may not be described in the specification. It should be understood that in the development of such an actual deployment, as in any engineering project or project, several specific implementation decisions must be made to achieve the specific objectives of the developers, such as compliance with the restrictions related to the system and business, which may vary from one deployment to the next. In addition, it must be understood that such a development effort can be complex and time-consuming, but it would, in any case, be a routine application of design, manufacture and production for those skilled in the art who enjoy the benefit of this invention.
    [0010] [010] When presenting the elements of various embodiments of the present invention, the articles "one", "one", "the" and "said" are intended to mean that there are one or more of the elements. The terms "that understands", "that includes" and "that has" are intended to be inclusive and mean that there may be elements in addition to the elements listed.
    [0011] [011] The systems and methods disclosed herein refer to estimating a location (for example, timing) of an engine event (for example, peak trigger pressure or closing an intake / exhaust valve) on an engine. internal and reciprocating combustion using one or more sensors, such as a knock sensor. A knock sensor can include an acoustic or audible sensor, a vibration sensor or any combination thereof. For example, the knock sensor can be a piezoelectric accelerometer, a microelectromechanical system sensor (MEMS), a Hall effect sensor, a magnetostrictive sensor and / or any other sensor designed to perceive vibration, acceleration, acoustics, sound and / or movement. The knock sensor can monitor acoustics and / or vibrations associated with combustion in the engine to detect a knock condition (for example, combustion at an unexpected time other than during a normal combustion time window) or other engine events that can create acoustic and / or vibration signals. In other embodiments, the sensor may not be a knock sensor, but any sensor that can sense vibration, pressure, acceleration, deflection or movement.
    [0012] [012] In certain cases, it may be desirable to estimate the timing of various engine events (for example, peak trigger pressure or closing an intake / exhaust valve) that are indicative of engine execution. Locating such events can allow a user or a controller to adjust various parameters based on the operating condition information to optimize engine execution. However, sensors (for example, pressure sensors) positioned inside an engine cylinder and configured to locate such events can be significantly more expensive than detonation sensors and can be more susceptible to damage. Therefore, it may be advantageous to train (for example, through machine learning) a controller to convert or transform a knock sensor signal into a form that can allow an accurate prediction of the location (for example, timing) of an event. motor. Such a system can estimate the location (eg, timing) of the engine event with accuracy comparable to that of a cylinder sensor (eg, pressure sensor), while having the benefit of being less expensive and more robust.
    [0013] [013] Due to the percussive nature of combustion engines, detonation sensors may be able to detect particularities even when mounted on the outside of an engine cylinder. However, detonation sensors can also be arranged in several locations on or near one or more cylinders. Knock sensors detect, for example, cylinder vibrations, and a controller can convert a vibrating cylinder profile, provided by a knock sensor, into parameters useful for estimating the location of an engine event. It is now recognized that detonation sensors detect vibrations in or near the cylinder, and can communicate a signal indicative of the vibrating profile to a controller, which can convert the signal and do various computations to produce the estimated location. The present invention relates to systems and methods for determining a location (e.g., timing) of an engine event (e.g., peak trigger pressure or closing an intake / exhaust valve) by training a controller or another computing device to locate a desired engine event in a knock sensor signal.
    [0014] [014] Referring to the drawings, Figure 1 illustrates a block diagram of an embodiment of a portion of an engine driven power system that has a reciprocating internal combustion engine, which can experience an engine event that can be located using the system and methods disclosed here. As described in detail below, system 8 includes an engine 10 (for example, a reciprocating internal combustion engine) that has one or more combustion chambers 12 (for example, 1, 2, 3, 4, 5, 6, 7 , 8, 10, 12, 14, 16, 18, 20 or more combustion chambers 12). An oxidant supply 14 (for example, an air supply) is configured to provide a pressurized oxidant 16, such as air, oxygen, oxygen-enriched air, oxygen-reduced air or any combination thereof, for each combustion chamber 12 The combustion chamber 12 is also configured to receive a fuel 18 (for example, a liquid and / or gaseous fuel) from a fuel supply 19 and a fuel-air mixture ignites and burns within each combustion chamber 12. The hot and pressurized combustion gases cause a piston 20 adjacent to each combustion chamber 12 to move linearly inside a cylinder 26, which converts the pressure exerted by the gases into rotational movement, thereby causing an axis 22 rotate. In addition, shaft 22 can be coupled to a load 24, which is fed through rotation on shaft 22. For example, load 24 can be any suitable device that can generate power through the rotational output of system 10, such as a electric generator. In addition, although the following discussion refers to air as oxidizer 16, any suitable oxidizer can be used with the disclosed embodiments. Similarly, fuel 18 can be any suitable gaseous fuel, such as natural gas, associated gas and oil, propane, biogas, sewage gas, landfill gas, coal mine gas, for example. Fuel 18 may also include a variety of liquid fuels, such as gasoline or diesel fuel.
    [0015] [015] System 8 disclosed in this document can be adapted for use in stationary applications (for example, in industrial power generation engines) or in mobile applications (for example, in cars or aircraft). Engine 10 can be a two-stroke engine, a three-stroke engine, a four-stroke engine, a five-stroke engine, or a six-stroke engine. The engine 10 can also include any number of combustion chambers 12, pistons 20 and associated cylinders 26 (for example, 1 to 24). For example, in certain embodiments, system 8 may include a large-scale industrial reciprocating engine that has 4, 6, 8, 10, 16, 24 or more pistons 20 that reciprocate in cylinders 26. In some cases, cylinders 26 and / or the pistons 20 can have a diameter of approximately between 13.5 to 34 centimeters (cm). In some embodiments, cylinders 26 and / or pistons 20 may have a diameter of approximately 10 to 40 cm, 15 to 25 cm or about 15 cm. System 10 can generate power in a range from 10 kW to 10 MW. In some embodiments, engine 10 can operate at less than approximately 1,800 revolutions per minute (RPM). In some embodiments, engine 10 can operate at less than approximately 2,000 RPM, 1,900 RPM, 1,700 RPM, 1,600 RPM, 1,500 RPM, 1,400 RPM, 1,300 RPM, 1,200 RPM, 1,000 RPM, 900 RPM or 750 RPM. In some embodiments, engine 10 can operate between approximately 750 to 2,000 RPM, 900 to 1,800 RPM or 1,000 to 1,600 RPM. In some embodiments, engine 10 can operate at approximately 1,800 RPM, 1,500 RPM, 1,200 RPM, 1,000 RPM or 900 RPM. Engines 10 can include Jenbacher Engines from General Electric Company (for example, Jenbacher Type 2, Type 3, Type 4, Type 6 or J920 FleXtra) or Waukesha Engines (for example, Waukesha VGF, VHP, APG, 275GL), for example
    [0016] [016] The motor-driven power generation system 8 may include one or more knock sensors 23 suitable for detecting an engine “knock”. The knock sensor 23 can detect vibrations, acoustics or sound caused by combustion in the engine 10, as well as vibrations, acoustics or sound due to detonation, pre-ignition and / or crash. The knock sensor 23 can also detect vibrations, acoustics or sound caused by closing the exhaust or intake valve. Therefore, the knock sensor 23 can include an audible or acoustic sensor, a vibration sensor or a combination thereof. For example, knock sensor 23 may include a piezoelectric vibration sensor. The knock sensor 23 is shown communicatively coupled to a system 25 (for example, a control system, a monitoring system, a controller or an engine control unit “ECU”). During operations, signals from the knock sensor 23 are communicated to the system 25 to determine whether knock conditions (for example, strike) exist. System 25 can adjust engine 10 operating parameters to optimize engine execution. For example, system 25 can adjust an engine timing map for engine 10, an oxidant / fuel ratio for engine 10, an exhaust gas recirculation fuse for engine 10, an intake or exhaust valve position or other engine operating parameter 10.
    [0017] [017] Figure 2 is a cross-sectional side view of an embodiment of a piston-cylinder assembly that has a piston 20 disposed within a cylinder 26 (for example, an engine cylinder) of the reciprocating engine 10. The cylinder 26 has an internal annular wall 28 that defines a cylindrical cavity 30 (e.g., hole). The piston 20 can be defined by a geometry axis or an axial direction 34, a geometry axis or radial direction 36 and a geometry axis or circumferential direction 38. The piston 20 includes a top portion 40 (for example, a top area) . The top portion 40 generally blocks fuel 18 and air 16 or a fuel-air mixture 32 from escaping from the combustion chamber 12 during the reciprocating movement of the piston 20.
    [0018] [018] As shown, piston 20 is attached to a crankshaft 54 via a connecting rod 56 and a pin 58. Crankshaft 54 translates the reciprocating linear motion of piston 24 in a rotational motion. As piston 20 moves, crankshaft 54 rotates to feed load 24 (shown in Figure 1), as discussed above. As shown, the combustion chamber 12 is positioned adjacent to the top area 40 of piston 24. A fuel injector 60 supplies fuel 18 to the combustion chamber 12 and an intake valve 62 controls the delivery of oxidant (for example, air 16) to the combustion chamber 12. An exhaust valve 64 controls the discharge of an exhaust from the engine 10. However, it must be understood that any elements and / or techniques suitable for supplying fuel 18 and air 16 to the combustion chamber 12 and / or to discharge the exhaust can be used and, in some embodiments, no fuel injection is used. In operation, combustion of fuel 18 with oxidizer 16 in the combustion chamber 12 can cause piston 20 to move reciprocally (for example, back and forth) in the axial direction 34 within cavity 30 of cylinder 26 .
    [0019] [019] During operations, when piston 20 is at the highest point of cylinder 26, it is in a position called upper dead center (TDC). When piston 20 is at its lowest point on cylinder 26, it is in a position called lower dead center (BDC). As piston 20 moves from the TDC to the BDC or from the BDC to the TDC, the crankshaft 54 rotates half a revolution. Each movement of piston 20 from the TDC to the BDC or from the BDC to the TDC is called a stroke, and engine achievements 10 can include two-stroke engines, three-stroke engines, four-stroke engines, five-stroke engines, six times or more.
    [0020] [020] During engine 10 operations, a sequence that includes an intake process, a compression process, a feed process and an exhaust process typically occurs. The intake process allows a fuel mixture, such as fuel 18 and oxidizer 16 (for example, air), to be pulled into cylinder 26, in this way, intake valve 62 is opened and exhaust valve 64 is closed . The compression process compresses the fuel mixture in a smaller space, so that both the intake valve 62 and the exhaust valve 64 are closed. The feeding process ignores the compressed fuel-air mixture, which may include a spark ignition through a spark plug system and / or a compression ignition through compression heat. The pressure resulting from the combustion then impels piston 20 to the BDC. The exhaust process typically returns piston 20 to the TDC while keeping exhaust valve 64 open. The exhaust process thus expels the spent fuel-air mixture through the exhaust valve 64. It should be noted that more than one intake valve 62 and exhaust valve 64 can be used per cylinder 26.
    [0021] [021] The engine shown 10 can include a crankshaft sensor 66, a knock sensor 23 and system 25, which includes a processor 72 and a memory unit 74. Crankshaft sensor 66 detects the position and / or the crankshaft rotational speed 54. Consequently, a crank angle or crank timing information can be derived. That is, when monitoring combustion engines, timing is often expressed in terms of crankshaft angle. For example, a complete cycle of a four-stroke engine 10 can be measured as a cycle of 720 °. The knock sensor 23 can be a piezoelectric accelerometer, a microelectromechanical system sensor (MEMS), a Hall effect sensor, a magnetostrictive sensor and / or any other sensor designed to detect vibration, acceleration, acoustics, sound and / or movement. In other embodiments, sensor 23 may not be a knock sensor, but any sensor that can detect vibration, pressure, acceleration, deflection or movement.
    [0022] [022] Due to the percussive nature of the engine 10, the knock sensor 23 may be able to detect particularities even when mounted outside the cylinder 26. However, the knock sensor 23 may be arranged in various locations on the cylinder or close to the even 26. Additionally, in some embodiments, a single knock sensor 23 may be shared, for example, with one or more adjacent cylinders 26. In other embodiments, each cylinder may include one or more knock sensor 23. The crankshaft sensor 66 and the knock sensor 23 are shown in electrical communication with system 25 (for example, a control system, a monitoring system, a controller or an “ECU” engine control unit). System 25 may include a non-transitory code or instructions stored in a machine-readable medium (for example, memory unit 74) and used by a processor (for example, processor 72) to implement the techniques disclosed herein. The memory can store computer instructions that can be executed by the processor 72. In addition, the memory can store reference tables and / or other relevant data. System 25 monitors and controls the operation of engine 10, for example, by adjusting the ignition timing, opening / closing timing of valves 62 and 64, adjusting the delivery of fuel and oxidant (for example, air) and so on. .
    [0023] [023] In certain embodiments, other sensors can also be included in system 8 and coupled to system 25. For example, sensors can include atmospheric and motor sensors, such as pressure sensors, temperature sensors, speed sensors and so on. onwards. For example, sensors can include knock sensors, crankshaft sensors, oxygen or lambda sensors, engine air intake temperature sensors, engine air intake pressure sensors, jacket water temperature sensors, sensors engine exhaust temperature, engine exhaust pressure sensors and exhaust gas composition sensors. Other sensors may also include compressor inlet and outlet sensors for temperature and pressure.
    [0024] [024] During the process of feeding an engine operation, a force (for example, a pressure force) is exerted on piston 20 through the expansion of flue gases. The maximum force exerted on piston 20 is described as the peak trigger pressure (“PFP”). It may be desirable for PFP to occur a few degrees of crank angle after piston 20 has reached the TDC so that the maximum amount of force can be exerted on piston 20. Therefore, having the ability to estimate the location (for example, angle timing or crank) of the PFP using the knock sensor 23 is desirable due to the fact that the PFP site can be compared to the TDC site to assess whether engine 10 is operating at optimal efficiency. In addition, if the PFP timing is not at an optimal level, several engine parameters (for example, ignition timing, fuel / air ratio, intake or exhaust valve closing timing, etc.) can be adjusted to optimize engine execution. For example, system 25 can set an engine timing map for engine 10, an oxidant / fuel ratio, an exhaust gas recirculation flow, an inlet 62 or exhaust valve position 64, or another operating parameter of the engine 10.
    [0025] [025] Additionally, it may be desirable to also estimate a location (for example, timing) of other motor events. For example, estimating the location of the exhaust valve closure 64 may allow a user or system 25 to determine whether the exhaust valve 64 is functioning properly or whether it is stuck in an open position or in a closed position. Keeping the exhaust valve 64 open for a certain period of time can improve the efficiency of the engine. Accordingly, while the present invention focuses primarily on estimating the PFP location during engine operation, it should be noted that the systems and methods disclosed can be used to estimate a location for other engine events (for example, closing the exhaust 64).
    [0026] [026] The present invention relates to predicting the timing of an engine event (for example, PFP or closing of the exhaust valve 64 or the intake valve 62) with the use of a knock sensor signal 23. In certain realizations, system 25 is trained (for example, through machine learning) to associate the resources of a knock sensor signal with an occurrence of a desired engine event.
    [0027] [027] Figure 3 illustrates a flow chart 100 of a method for training system 25 (for example, a control system, a monitoring system, a controller or an “ECU” engine control unit) through the development of a predictive model and predictive frequency bands (“PFBs”) to estimate a desired engine event time (for example, PFP). In block 102, a true location (for example, timing) of the desired motor event (s) is received or entered into system 25. The actual locations can be determined by a pressure signal plot, such as as shown in Figure 5, or another signal that measures a relevant engine operating parameter (intake valve 62 or exhaust valve position 64, oxidant flow rate or exhaust gas flow rate) indicative of the desired engine over time. Additionally, system 25 receives a knock sensor signal at block 104. The knock sensor signal is also indicative of the motor event in which it may include a response to the motor event displayed by the knock sensor 23. However, the knock sensor signal may not be used to directly estimate the time of the motor event. In this way, the knock sensor signal can be transformed through a short time Fourier Transform 106 (“STFT”) into a power spectral density 108 (“PSD”) over time. The STFT computes the PSDs 108 of individual windows (for example, subsets) of the knock sensor signal and plots the computed PSDs 108 over time. Each PSD 108 can include the energy content of the knock sensor subsign as a frequency function (for example, a frequency energy plot). The formation of a PSD over time is described in more detail in this document with reference to Figure 5.
    [0028] [028] After PSD 108 has been acquired over time, system 25 can undermine predictive frequency bands (“PFBs”) 112 in block 110. PFBs 112 are frequency ranges of the knock sensor signal that are indicative of the occurrence of the desired motor event. To undermine PFBs 112, the detonation signal is broken into a variety of subsigns. A given subsignal can include the engine event or the subsign can belong to a time before or after the engine event. The number of subsignals that includes the motor event can be represented as "N". In certain embodiments, a discrete frequency value of the knock sensor signal can occur more than once for the entire course of the entire knock sensor signal, so that the discrete frequency value is present in more than one subsignal of the signal of whole knock sensor. For example, a discrete frequency value can be present in a first subsign that includes the motor event and in a second subsign that corresponds to a time before or after the motor event occurs. However, although the discrete frequency value can occur multiple times throughout the entire knock sensor signal, different energy values can be associated with each occurrence of the discrete frequency value. Therefore, PSD 108 over time may include multiple occurrences of the same distinct frequency; however, each occurrence may not have the same energy content. In certain embodiments, each occurrence of the distinct frequency can be arranged in order to increase energy. In addition, each occurrence can be classified as either a positive or a negative. In certain embodiments, a positive occurrence corresponds to a subsign in which the motor event actually occurs. On the other hand, a negative occurrence corresponds to a subsign in which the motor event has not occurred. System 25 can know whether an occurrence of the distinct frequency is positive or negative due to the fact that system 25 has received the true location of the motor event in block 102.
    [0029] [029] Since the occurrences of each distinct frequency are arranged in order to increase energy, a discriminative index (“index D”) can be calculated for each distinct frequency of the knock sensor signal. In certain embodiments, the D index is computed by selecting “N” occurrences from the distinct frequency that has the highest energy. Of those selected, the number of positive occurrences can be divided by N to receive the index D. The calculation of index D is described in more detail in this document with reference to Figure 8.
    [0030] [030] Since the index D is computed for a frequency of the knock sensor signal, system 25 can combine different hard frequencies and compute a second index D for the frequency range. If the D index of the frequency range is greater than the D index of the individual discrete frequency, system 25 can combine the two distinct frequencies into one frequency range. In addition, additional discrete frequency values can be combined with the frequency range, until the D index cannot be further improved. At that point, the system can use the distinct frequency or frequency range like PFB 112. PFB 112 can be indicative of the frequency ranges of a knock sensor signal that correspond to the occurrence of the desired motor event.
    [0031] [031] In block 114, system 25 can convert each sub-sign into a resource vector using PFBs 112. Each resource in the resource vector can correspond to a specific PFB (for example, through an energy). Consequently, the resource vector can have a length, "i", in which the value of the i-th resource corresponds to the energy in the nth PFB. In this way, the system can compare the resource vectors to the actual motor event location and is subjected to model 116 learning, so that system 25 can associate certain resources with the motor event and / or the location of the motor event. motor. For example, system 25 can use a logistic regression classifier, a support vector machine, or other machine learning algorithm configured to generate a predictive model 118 using resource vectors and the actual location of the motor event. In this way, system 25 can store predictive model 118 and use predictive model 118 to determine when a subsignal includes the engine event and to estimate a location of the engine event.
    [0032] [032] Figure 4 illustrates a flow chart 130 of a method for using or testing predictive model 118 to determine a motor event location. Similar to block diagram 100, system 25 (for example, a control system, a monitoring system, a controller or an engine control unit “ECU”) can receive a knock sensor signal at block 104 and transform the signal in a PSD 108 over time with the use of an STFT 106. Additionally, system 25 can transform PSD 108 over time into resource vectors 132 using the PFBs 112 determined in flowchart 100. The system 25 can use the resulting resource vectors 132 and predictive model 118 to determine a location of the motor event 134. For example, resource vectors 132 can include information that system 25 can associate with the location of the motor event. If the motor event was likely to occur during a given subsignal, predictive model 118 can compute a probability that the motor event occurred at each time in the sub window. System 25 can use the time with the highest probability 134 to control engine operating parameters and improve engine execution. For example, system 25 can adjust an engine timing map (for example, ignition timing) of engine 10, an oxidant / fuel ratio, an exhaust gas recirculation flow, an inlet valve position 62, or exhaust 64 or other engine 10 operating parameter.
    [0033] [033] In certain embodiments, system 25 will be subjected to the process in flowchart 130 (for example, test mode of the predictive model) immediately after the process in flowchart 100 (for example, formation of the predictive model). Depending on the difference between the predicted location (for example, timing) of the motor event (for example, from flow chart 130) and the actual location of the motor event, system 25 can repeat the process in flow chart 100 until the difference between the estimated location and the actual location is at a desirable level (for example, less than 1 ° from the crankshaft). In other words, system 25 can continue to perform the process on flowchart 100 to refine the predictive model and PFBs until the timing of the motor event can be estimated to a desired degree of accuracy.
    [0034] [034] Additionally, the predictive model 118 generated by the process of flowchart 100 can be specific to a particular type of engine. For example, the predictive model 118 used to estimate the location of the motor event on a Jenbacher Type 2 Engine may not accurately estimate the location of the motor event on a Jenbacher Type 3 Engine. Thus, the process of flowchart 100 can be performed for each type of engine on which the engine event location will be estimated. As non-limiting examples, the flowchart 100 process can be performed on General Electric Company's Jenbacher Engines (eg Jenbacher Type 2, Type 3, Type 4, Type 6 or J920 FleXtra), Waukesha Engines (eg, Waukesha VGF, VHP, APG, 275GL) or any other reciprocating internal combustion engines.
    [0035] [035] Figure 5 illustrates a realization of a pressure signal plot 150 indicative of the true time of the motor event, a knock sensor signal 152 and a spectrogram (for example, PSD over time) 154 based on knock sensor signal 152. Pressure signal plot 150 can originate from a cylinder pressure sensor and provide a true motor event time to be detected by system 25. As shown in the illustrated embodiment, the Y 156 geometric axis of the pressure signal plot 150 represents a pressure in the cylinder 26 of the motor 10. Additionally, the geometric axis X 158 represents time (for example, crank angle). In this way, the pressure signal plot 150 illustrates the pressure in the cylinder 26 over a given time interval. The pressure increases until it reaches a maximum point 160, which represents the actual time of the motor event (for example, PFP). It should be noted that in other embodiments, the timing of an engine event (for example, closing an exhaust valve 62), in addition to PFP, can be estimated. Therefore, the actual time of such an engine event can be determined using a plot, in addition to the pressure plot 150, which corresponds to a relevant measurement (for example, exhaust valve angle, exhaust gas flow rate exhaust valve, etc.) of the engine event along.
    [0036] [036] As illustrated, the knock sensor signal 152 has a geometric axis Y 162 that represents a voltage, resistance or other representative quantity of the response displayed by the knock sensor 23 for a change in vibration, sound, acoustics, etc. on cylinder 26. The knock sensor signal also has an X 164 geometry axis that represents time (for example, crank angle), which is substantially aligned with the X 158 geometry axis of the pressure signal plot 150. As shown, knock sensor 23 exhibits a greater response before the PFP's true time (for example, knock sensor signal 152 exhibits the largest change in magnitude at a time delay of approximately 0.03, while PFP occurs after 0.04) . Therefore, it may not be necessary to estimate the timing of the desired motor event by simply computing the time at which knock sensor signal 152 exhibits the highest rate of change. Consequently, other computations and / or manipulations can be applied to the knock sensor signal 152 to estimate the time of the desired motor event.
    [0037] [037] Spectrogram 154 illustrates a computation that can be performed on the knock sensor signal 152. For example, spectrogram 154 may represent a spectral power density of the knock sensor signal 152 over time. The spectral power density can refer to the energy content of the knock sensor signal 152 as a frequency function. In other words, the power spectral density is a function of frequency, not time. Therefore, spectrogram 154 can illustrate spectral power densities of subsins (e.g., windows) of the knock sensor signal 152 as a function of timing (e.g., crank angle). In other embodiments, spectrogram 154 can categorize different frequencies of the knock sensor signal 152 according to a frequency intensity (for example, different hues in spectrogram 154 refer to different intensities of a given frequency). To transform the knock sensor signal 152 to spectrogram 154, a multivariate transformation algorithm can be applied to the knock sensor signal 152. In certain embodiments, spectrogram 154 is produced using the short time Fourier Transform ( “STFT”) 106. In other embodiments, spectrogram 154 can be generated using another type of Fourier Transform, a discrete cosine transform, a Laplace Transform, a Mellin Transform, a Hartley Transform, a Transform Chirplet, a Hankel Transform or any combination thereof. Spectrogram 154 can be used to create the predictive model 118 that estimates a location of the motor event (e.g., PFP) as described above.
    [0038] [038] It should be noted that, in certain embodiments, system 25 may not physically generate spectrogram 154. System 25 may encapsulate, or hide, the functionality provided by the spectrogram in the processing steps performed by processor 72 and / or stored in the memory unit 74, so that the spectrogram is never displayed or even obtainable by a user. For example, system 25 can directly convert the signal from the knock sensor 23 into the resource vectors, or it can incorporate the functionality provided by the spectrogram into one or more transform functions or comparable mathematical constructions in order to streamline certain steps among those discussed in the this document. Additionally, spectrogram 154 should not be limited to the realization illustrated in Figure 5. In other embodiments, spectrogram 154 (or the equivalence of the data construction of the same) can be any data, data table, algorithm, graph, schema or similar which are intended to represent the frequencies and the intensity of such frequencies, in the knock sensor signal 152 over time. For example, spectrogram 154 can categorize frequency intensities in another way besides color (for example, shapes, letters, numbers, tinting, etc.).
    [0039] [039] Figure 6 is a schematic illustrating how system 25 (for example, a control system, a monitoring system, a controller or an engine control unit “ECU”) can determine an index D of a frequency distinct or a frequency range. As illustrated, the occurrences of the specific distinct frequency are arranged in such a way as to increase energy as well as to be characterized as positive (for example, in a sub-station where the motor event occurred) or negative (for example, in a sub-station where the event engine failure has not occurred). Scheme 180 includes multiple occurrences that correspond to the same distinct frequency, but have different energy values due to the fact that each occurrence may belong to a different subsign of the detonation sensor signal 152. As illustrated, scheme 180 includes a geometric axis 182 that represents the energy associated with each occurrence of the distinct frequency. The energy increases from the left to the right so that the events with the lowest energy are on the left and the events with the highest energy are on the right. As described above, the D index can be computed by selecting “N” occurrences with the highest energy values, where “N” represents the number of subsignals that include the motor event. In the illustrated embodiment, “N” is equal to five due to the fact that there are five total occurrences that are classified as positive (that is, they are within a subsign in which the motor event occurred). However, of the five selected occurrences 184, only four are positive. Thus, the index D of the illustrated realization is 4 divided by 5 or 0.8. It should be noted that although “N” is equal to five in scheme 180, “N” can be any positive integer, such as 1, 2, 3, 4, 5, 6, 7 8, 9, 10, 12, 14 , 16, 18, 20, 30, 40, 50 or more. As previously discussed, the D index can be calculated for a range of different frequencies, in addition to the different frequencies themselves, to determine PFBs 112.
    [0040] [040] Figure 7 is a diagram that illustrates how the system (for example, a control system, a monitoring system, a controller or an engine control unit “ECU”) builds PFBs 112 based on a value of index D. In certain embodiments, diagram 200 includes three layers (for example, levels); however, other achievements can have less than three levels (for example, 1 or 2), while other achievements can have more than three levels (for example, 4, 5, 6 or more). In diagram 200, the first layer 202 includes all distinct frequencies in the spectrum (for example, all frequencies in the knock sensor signal 152). The second layer 204 is a combination of two distinct frequencies from the first layer 202. For example, a distinct frequency of 100 Hertz ("Hz") and one of 200 Hz are joined together in a frequency range of 100 to 200 Hz. As discussed previously, the distinct frequencies of a layer can be joined when the D index of the frequency range is greater than the D index of the individual distinct frequency. Consequently, in certain embodiments, the D index of the 100 to 200 Hz frequency range is greater than the individual D indices of the different frequencies of 100 Hz and 200 Hz.
    [0041] [041] Similarly, a separate frequency of 400 Hz and one of 500 Hz can be joined in a frequency range of 400 to 500 Hz, as shown in diagram 200. Again, this may be due to the fact that the index D the frequency range 400 to 500 Hz is greater than the individual D indices of the distinct frequencies of 400 Hz and 500 Hz. If no combination of the distinct frequencies occurs, then the D index of the individual distinct frequency may have been greater than the D index of the combined frequency range. For example, a distinct frequency of 600 Hz has not been combined with any other distinct frequency or frequency range. Therefore, the 600 Hz D index may have been greater than the D index of the 500 to 600 Hz frequency range or the 400 to 600 Hz frequency range.
    [0042] [042] Diagram 200 also has a third layer 206. The third layer represents a frequency range that is larger (for example, wider) than the frequency range of the second layer (for example, the third layer has a range of frequency of 200 Hz while the second layer has a frequency range of 100 Hz). As shown in diagram 200, a distinct frequency of 300 Hz has been combined with the frequency range of the second layer from 100 to 200 Hz to create a frequency range of the third layer of 100 to 300 Hz. frequency from 100 to 300 Hz can be greater than the D index of the 100 to 200 Hz frequency range as well as the D index for each of the individual discrete frequencies (for example, the D index for 100 Hz, 200 Hz and 300 Hz ).
    [0043] [043] Since the D index can no longer be increased by combining another distinct individual frequency, a PFB 112 has been determined. For example, if the D index for a 100 to 400 Hz frequency range is less than the D index for the 100 to 300 Hz frequency range, then the 400 Hz discrete frequency is not combined in PFB 112 and 100 to 300 Hz is the frequency range for PFB 112.
    [0044] [044] Figure 8 further illustrates how knock sensor signal 152 can be transformed to estimate a desired engine event location (eg, PFP) during common engine operation. In other words, Figure 8 illustrates and elaborates the process in a flow chart 130. Figure 8 illustrates a subsignal 220 of the knock sensor signal 152, a power spectral density plot 222 (“PSD plot”) and a scheme of a resource vector 224.
    [0045] [045] In certain embodiments, system 25 is trained to identify the locations of the desired engine event (for example, PFP) using PFBs 112 and the predictive model 118. Therefore, when the engine operates under common conditions (for example, example, it does not operate to collect a signal from the desired motor event) the system 25 receives a signal from the knock sensor 23, but does not receive the pressure signal 150 or other signal indicative of the motor event (for example, PFP). Consequently, system 25 does not know the actual timing of the desired motor event. In certain embodiments, system 25 extracts a subsignal 220 from the knock sensor signal 150. Additionally, system 25 can produce the PSD 222 plot by applying the STFT 106 to the subsign 220. In other embodiments, the PSD 222 plot can be generated using another type of Fourier Transform, a discrete cosine transform, a Laplace Transform, a Mellin Transform, a Hartley Transform, a Chirplet Transform, a Hankel Transform or any other transform configured for generate a plot of PSDs over time.
    [0046] [046] The PSD 222 plot can be separated into the PFBs (for example, lines 226) determined in flowchart 100. As described above, the PSD 222 plot includes the energies of subsign 220 as a frequency function.
    [0047] [047] Resource vector 224 can be created from PSD plot 222 and PFB lines 226. In certain embodiments, the number of resources 228 can correspond to the number of PFBs. For example, the PSD plot 222 is separated into five portions by lines 226 (for example, five PFBs). Thus, five resources 228 are included in resource vector 224. As shown in the illustrated embodiment, the five resources v 228 do not have equal resource values. However, in other embodiments, resources 228 can have the same resource values. As described above, the i-th resource corresponds to the sub-energy in the i-th PFB. It should be understood that although the illustrated resource vector 224 includes five resources 228, more or less resources can be formed and included in resource vector 224. For example, resource vector 224 can have 1, 2, 3, 4, 6, 7, 8, 9, 10, 12, 15, 20, 25, 30, 40, 50 or more resources.
    [0048] [048] In certain embodiments, predictive model 118 can be applied to the resource vector to compute a probability that the desired engine event has occurred at each location. The location that is most likely can be used by system 25 to adjust various operational parameters (for example, an engine timing map 10, an oxidant / fuel ratio, an exhaust gas recirculation flow, a valve position intake 62 or exhaust 64 or other operating parameter of engine 10) of engine 10 to optimize engine execution.
    [0049] [049] Figures 9 to 11 illustrate diagrams and tables that compare the time estimated by the predictive model 118 to the real time of the motor event determined by the pressure signal plot 102 or another signal that measures a relevant motor operating parameter (angle of intake valve 62 or exhaust valve 64) indicative of the desired engine event over time. Figures 9 and 10 were generated by testing the techniques shown here on a VHP L5794GSI engine (base). Figure 11 was generated by testing the techniques disclosed here on a VHP L5794GSI engine (base), a VHP L5794GSI engine (detonation), a Jenbacher Type 4 engine and a CFR-RON engine.
    [0050] [050] For example, Figure 9 is a histogram 240 that plots the error for a plurality of tests performed using the predictive model. Figure 9 is intended to be representative of what can be achieved by the revealed achievements and, therefore, it is not intended to limit the achievements disclosed here only to such results. Specifically, histogram 240 plots a number of times that predictive model 118 estimated the timing of the desired motor event within a certain value of the real time of the motor event. The geometric axis X 242 of the histogram 240 represents the difference between the estimated time (for example, degree of crank angle) using the predictive model 118 and the true time (for example, degree of crank angle) of the motor event . The geometric axis Y 244 represents the number of tests that reached the difference value. A line 246 represents the most accurate estimates of the location of the motor event by the predictive model 118 due to the fact that line 246 corresponds to a difference of zero between the estimated location and the true location. As seen from histogram 240, the predictive model almost always estimates the time of the engine event within 5 ° of the crankshaft and, therefore, is relatively accurate.
    [0051] [051] Figure 10 illustrates a first graph 260 that represents the true pressure inside the cylinder 26 over time and, thus, the true locations 262 of the engine event. In addition, a second graph 264 represents a probability, or estimate, that the desired engine event will occur at a time 266 using the predictive model 118. Again, Figure 10 is intended to be representative of what can be achieved by the achievements disclosed and, therefore, it is not intended to limit the achievements disclosed here to just such results. However, as can be seen in Figure 10, predictive model 118 is relatively accurate in estimating the timing of the motor event due to the fact that peaks 266 of the second graph 264 (for example, the estimated probability of the trigger pressure event peak) finds very closely the true locations 262 illustrated in the first graph 260.
    [0052] [052] Figure 11 illustrates a table 270 that includes valves that compare an estimated location (for example, timing) of the motor event with the use of predictive model 118 and the actual location of the motor event. Figure 11 is also intended to be representative of what can be achieved by the revealed achievements and, therefore, it is not intended to limit the achievements disclosed here to just such results. Figure 11 includes data for a number of cycles 272 performed, an average absolute error 274 of the tests performed, an average detection error 276, a standard deviation of detection error 278 and a failure rate 280. Figure 11 shows that the Predictive model 118 can potentially predict the location (eg, timing) the engine event within 2 ° of the crankshaft (eg, the mean detection error column 276). In certain embodiments, the predictive model 118 can estimate a location (for example, timing) of the motor event within 0 and 30 degrees of crank angle; 0 and 25 degrees of crank angle; 0.05 and 15 degrees crank angle; or anyone in between. In addition, Figure 11 illustrates the failure rate 280 of the predictive model 118, in which the failure rate 280 can be defined as the percentage of total motor cycles in which the estimated location (for example, timing) of the motor event was more than 10 ° from the handle away from the actual engine event location. In other embodiments, failure rate 280 can be defined as the percentage of total engine cycles where the estimated location (e.g., timing) of the engine event was more than 15 ° from crankshaft 54 away from the actual engine location. In even more realizations, the failure rate 280 can be defined as the percentage of total motor cycles where the estimated location (for example, timing) of the motor event was more than 1, 2, 3, 4, 5, 6 , 7, 8, 9, 11, 12, 13, 14, 20, 25, 30, 35, 40, 45, 50 or more crank angles away from the actual engine location. As illustrated by the data in Figure 11, predictive model 118 can be applied within the engine 10 during common operation and accurately estimate the location of the desired engine event (for example, from the failure rate column 280).
    [0053] [053] The effects of the inventive technique include receiving a signal from a knock sensor that refers to an engine event. The signal can be used to estimate a motor event location using a predictive model and PFBs. Engine parameters can be adjusted based on the estimated location to improve fuel efficiency, improve power output, etc.
    [0054] [054] This written description uses examples to reveal the invention, which include the best way and also to allow any person skilled in the art to practice the invention, which includes making and using any devices or systems and executing any incorporated methods. 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 are not different from the literal language of the claims or if they include equivalent structural elements with non-substantial differences from the literal language of the claims.
    权利要求:
    Claims (14)
    [0001]
    SYSTEM (8) FOR ESTIMATING AN ENGINE EVENT PLACE (134), characterized by comprising: a controller (25) configured to: receiving (104) a signal (152) from at least one knock sensor (23) coupled to a reciprocating engine (10), transform (106) the signal (152), using a multivariate transformation algorithm, into a power spectral density (108), transform the spectral power density (108) into a plurality of resource vectors (132) using predictive frequency bands (112), predict the engine event location (134) using at least the plurality of resource vectors (132) and a predictive model (118) and adjust the operation of the reciprocating engine (10) based on the engine event location (134).
    [0002]
    SYSTEM (8) according to claim 1, characterized in that the motor event comprises a peak firing pressure of a reciprocating engine cylinder (10).
    [0003]
    SYSTEM (8), according to claim 2, characterized in that the controller (25) is configured to adjust an engine timing map of the reciprocating engine (10), an oxidant / fuel ratio of the reciprocating engine (10), a flow reciprocating engine exhaust recirculation gas (10), a reciprocating engine inlet or exhaust valve position (10) or other reciprocating engine operating parameter (10) in response to the peak trigger pressure location.
    [0004]
    SYSTEM (8) according to any one of claims 1 to 3, characterized in that the engine event comprises closing a reciprocating engine intake and / or exhaust valve (10).
    [0005]
    SYSTEM (8) according to any one of claims 1 to 4, characterized in that the multivariate transformation algorithm comprises a short time Fourier Transform.
    [0006]
    SYSTEM (8) according to any one of claims 1 to 5, characterized in that the multivariate transformation algorithm comprises a discrete cosine transform.
    [0007]
    SYSTEM (8) according to any one of claims 1 to 6, characterized in that at least one knock sensor (23) comprises a piezoelectric knock sensor configured to sense vibrations and / or acoustics in a reciprocating engine cylinder (10) .
    [0008]
    SYSTEM (8) according to any one of claims 1 to 7, characterized in that the predictive model (118) is trained to predict the engine event location (134) for a specific reciprocating engine (10).
    [0009]
    METHOD (100) TO TRAIN A CONTROLLER (25), to estimate the location of peak trigger pressure in a reciprocating engine (10), characterized by understanding the steps of: receiving (104) a first signal (152) from at least one knock sensor (23), wherein the first signal (152) comprises data corresponding to a peak trigger pressure event; receiving (102) a second signal (160) from a pressure sensor corresponding to the true peak trigger pressure location; transforming (106) the first signal (152) into a spectral power density (108) comprising frequency energies over time; comparing (110) the spectral power density to the second signal (160) to form predictive frequency bands (112); converting (114) the power spectral density into a plurality of resource vectors (132); and execute (116) an algorithm to generate a predictive model (118) using the plurality of resource vectors (132) and the second signal (160), in which the predictive model (118) is configured to estimate the pressure site peak trip on the reciprocating engine (10) during common engine operation.
    [0010]
    METHOD (100), according to claim 9, characterized by the predictive frequency bands (112) being formed by calculating a discriminative index.
    [0011]
    METHOD (100), according to claim 10, characterized by the calculation of the discriminative index comprising the steps of: extracting a first subset of the frequency energies over time, where the first subset corresponds to the frequency energies for a first frequency range; classify the first subset in order to increase energy; characterizing each frequency energy of the first subset as positive and negative, where the positives correspond to the frequency energies that occurred within a subsign of the first signal (152) that includes the peak trigger pressure; selecting a segment of the frequency energies of the first subset that comprises the highest energy, where the segment comprises a number of frequency energies equal to the number of positives in the first subset; determine the number of frequency energies in the segment that are true positives to calculate the discriminative index; and repeat the calculation of the discriminative index for a second subset of the frequency energies over time until the discriminatory index fails to increase, where the second set corresponds to the frequency energies for a second frequency range and where the second frequency range frequency is higher than the first frequency range.
    [0012]
    METHOD (100) according to any one of claims 9 to 11, characterized in that the pressure sensor is a cylinder pressure sensor.
    [0013]
    METHOD (100), according to any one of claims 9 to 12, characterized in that the algorithm comprises a logistic regression classifier, a support vector machine or another machine learning algorithm configured to generate a predictive model (118) that uses resource vectors (132) and a pressure signal.
    [0014]
    METHOD (100) according to any one of claims 9 to 13, characterized in that the method (100) is repeated until the predictive model (118) estimates the peak trigger pressure within 10 degrees of the actual peak trigger pressure .
    类似技术:
    公开号 | 公开日 | 专利标题
    BR102016005985B1|2021-01-26|system for estimating a motor event location and method for training a controller
    BR102016007569A2|2016-10-11|system and methods for training a controller and for estimating a motor event time
    US9915217B2|2018-03-13|Methods and systems to derive health of mating cylinder using knock sensors
    JP2016145571A|2016-08-12|System and method for detecting operating events of engine
    KR20160082484A|2016-07-08|System and method for locating engine noise
    US9556810B2|2017-01-31|System and method for regulating exhaust gas recirculation in an engine
    BR102015032783A2|2016-07-05|method of analyzing a computer readable noise, system and medium
    BR102016002606A2|2016-10-11|diagnostic method of a non-transient computer-readable beat, system and media sensor
    BR102016007574A2|2016-10-25|computer readable method, system and medium
    BR102015032241A2|2016-07-12|systems and method
    US9869257B2|2018-01-16|System and method for predicting peak pressure values using knock sensor
    US10961942B2|2021-03-30|System and method for determining the timing of an engine event
    BR102016001908B1|2021-12-07|METHOD OF MONITORING AN OPERATION EVENT OF A COMBUSTION ENGINE, COMPUTER-READable, NON-TRANSITORY, AND SYSTEM
    同族专利:
    公开号 | 公开日
    KR20160114535A|2016-10-05|
    CN106014658B|2021-04-02|
    CN106014658A|2016-10-12|
    JP2016183671A|2016-10-20|
    US9593631B2|2017-03-14|
    CA2923277A1|2016-09-24|
    US20160281617A1|2016-09-29|
    EP3095992A1|2016-11-23|
    BR102016005985A2|2016-09-27|
    KR102261809B1|2021-06-07|
    AU2016201507A1|2016-10-13|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US4446722A|1981-07-07|1984-05-08|Robert Bosch Gmbh|Vibration sensor, particularly knock sensor for internal combustion engines|
    US4491010A|1983-06-20|1985-01-01|General Motors Corporation|Dynamic combustion characteristic sensor for internal combustion engine|
    US5052214A|1989-01-03|1991-10-01|Luxtron Corporation|Knock detector using optical fiber thermometer|
    GB8908531D0|1989-04-14|1989-06-01|Lucas Ind Plc|Engine knock sensing system|
    US5539638A|1993-08-05|1996-07-23|Pavilion Technologies, Inc.|Virtual emissions monitor for automobile|
    JPH07280637A|1994-04-01|1995-10-27|Ngk Insulators Ltd|Fire sensor|
    US5675505A|1995-07-10|1997-10-07|Chrysler Corporation|Sine on random data analysis method for simulating engine vibration|
    US5763769A|1995-10-16|1998-06-09|Kluzner; Michael|Fiber optic misfire, knock and LPP detector for internal combustion engines|
    JPH10185677A|1996-12-19|1998-07-14|Unisia Jecs Corp|Knock sensor|
    JP3092552B2|1997-09-16|2000-09-25|トヨタ自動車株式会社|Compression ignition type internal combustion engine|
    US6275765B1|1999-10-28|2001-08-14|Brunswick Corporation|System for providing a prognosis of future engine faults|
    US6285947B1|1999-10-28|2001-09-04|Brunswick Corporation|Prognostication of future failure of an engine indicator parameter|
    US6651432B1|2002-08-08|2003-11-25|The United States Of America As Represented By The Administrator Of The Environmental Protection Agency|Controlled temperature combustion engine|
    JP4086602B2|2002-09-17|2008-05-14|株式会社日立製作所|Control device and control method for multi-cylinder engine|
    JP4451180B2|2004-03-26|2010-04-14|富士重工業株式会社|Engine combustion detection interval setting device|
    US7222607B2|2005-08-05|2007-05-29|Gm Global Technology Operations, Inc.|DSP-based engine knock detection including knock sensor and circuit diagnostics|
    US7317994B2|2005-08-10|2008-01-08|General Electric Company|Method and apparatus for signal signature analysis for event detection in rotating machinery|
    US7383799B1|2005-10-13|2008-06-10|Brunswick Corporation|Method for monitoring the operating condition of an engine valve system|
    FR2900745B1|2006-05-05|2008-10-10|Eurocopter France|METHOD AND DEVICE FOR DIAGNOSING A MECHANISM|
    JP4773888B2|2006-06-02|2011-09-14|本田技研工業株式会社|Ignition timing control device for internal combustion engine|
    JP2008045476A|2006-08-14|2008-02-28|Toyota Motor Corp|Method for and device of adapting internal combustion engine|
    JP2008275336A|2007-04-25|2008-11-13|Toyota Central R&D Labs Inc|Evaluation device and program|
    FR2924219B1|2007-11-22|2009-12-25|Siemens Vdo Automotive|METHOD FOR DETERMINING A REPRESENTATIVE PRESSURE VALUE IN A COMBUSTION CHAMBER OF AN INTERNAL COMBUSTION ENGINE.|
    JP2009138691A|2007-12-10|2009-06-25|Denso Corp|Valve control device and abnormality detection device for internal combustion engine|
    JP4550917B2|2008-05-12|2010-09-22|三菱電機株式会社|Control device for internal combustion engine|
    US8055432B2|2008-08-07|2011-11-08|The United States Of America As Represented By The Administrator Of The U.S. Environmental Protection Agency|Method and system of transient control for homogeneous charge compression ignition engines|
    EP2202504B1|2008-12-29|2013-08-21|STMicroelectronics Srl|Method and device for detecting peak values of pressure in a cylinder of an internal combustion engine|
    CA2673216C|2009-07-31|2011-05-03|Westport Power Inc.|Method and apparatus for reconstructing in-cylinder pressure and correcting for signal decay|
    CN102365446B|2010-02-08|2015-04-01|松下电器产业株式会社|Rpm increase/decrease determination device and method|
    JP5707090B2|2010-10-20|2015-04-22|カヤバ工業株式会社|Railway vehicle vibration analyzer|
    FR2976074B1|2011-06-01|2013-05-17|IFP Energies Nouvelles|METHOD OF ESTIMATING THE INTENSITY OF THE CLICKS OF AN INTERNAL COMBUSTION ENGINE BY INVERTING A WAVE EQUATION|
    JP5837341B2|2011-06-24|2015-12-24|株式会社ブリヂストン|Road surface condition determination method and apparatus|
    JP5924910B2|2011-11-25|2016-05-25|ダイハツ工業株式会社|Control device for internal combustion engine|
    KR101316281B1|2011-12-13|2013-10-08|아주대학교산학협력단|Method for controlling of combustion in diesel engine|
    EP2864919A1|2012-06-21|2015-04-29|Philip Morris Products S.a.s.|Systems and methods for generating biomarker signatures with integrated dual ensemble and generalized simulated annealing techniques|
    JP5546595B2|2012-08-07|2014-07-09|三菱電機株式会社|Knock control device for internal combustion engine|
    JP5937921B2|2012-08-09|2016-06-22|株式会社ブリヂストン|Road surface condition determination method and apparatus|
    JP5942862B2|2013-01-16|2016-06-29|マツダ株式会社|Catalyst early warm-up controller for spark ignition engine|
    CN103217280B|2013-03-18|2015-10-28|西安交通大学|The multivariable support vector machine prediction method of aero-engine rotor residual life|EP3074742B1|2013-11-25|2019-11-20|Sem AB|Measurement of knock intensity|
    JP6312618B2|2015-03-13|2018-04-18|日立オートモティブシステムズ株式会社|Internal combustion engine control device and abnormal combustion detection method|
    US9897021B2|2015-08-06|2018-02-20|General Electric Company|System and method for determining location and value of peak firing pressure|
    US9869257B2|2015-08-19|2018-01-16|General Electric Company|System and method for predicting peak pressure values using knock sensor|
    WO2018044294A1|2016-08-31|2018-03-08|General Electric Company|System and method for determining the timing of an engine event|
    US10393050B2|2017-10-30|2019-08-27|Ai Alpine Us Bidco Inc.|Estimation of cylinder conditions using a knock sensor|
    DE102018126501B3|2018-10-24|2019-12-19|Iav Gmbh Ingenieurgesellschaft Auto Und Verkehr|Method for predicting the maintenance of components of an internal combustion engine using a structure-borne noise sensor|
    CN109404152B|2018-10-26|2021-08-20|长安大学|Fuel supply closed-loop control method for vehicle electric control methanol-diesel dual-fuel engine|
    KR102315405B1|2019-11-21|2021-10-21|한국철도기술연구원|Apparatus and method for inspecting railway line|
    CN111075682A|2019-12-26|2020-04-28|龚明瀚|Compressor structure driven by crank|
    法律状态:
    2016-09-27| B03A| Publication of a patent application or of a certificate of addition of invention [chapter 3.1 patent gazette]|
    2020-05-12| B06U| Preliminary requirement: requests with searches performed by other patent offices: procedure suspended [chapter 6.21 patent gazette]|
    2020-12-08| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|
    2021-01-26| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 18/03/2016, OBSERVADAS AS CONDICOES LEGAIS. |
    2021-07-20| B25A| Requested transfer of rights approved|Owner name: AI ALPINE US BIDCO INC. (US) |
    2021-12-07| B25L| Entry of change of name and/or headquarter and transfer of application, patent and certificate of addition of invention: publication cancelled|Owner name: GENERAL ELECTRIC COMPANY (US) Free format text: ANULADA A PUBLICACAO CODIGO 25.1 NA RPI NO 2637 DE 20/07/2021 POR TER SIDO INDEVIDA. |
    2021-12-21| B25A| Requested transfer of rights approved|Owner name: AI ALPINE US BIDCO INC. (US) |
    优先权:
    申请号 | 申请日 | 专利标题
    US14/667,275|2015-03-24|
    US14/667,275|US9593631B2|2015-03-24|2015-03-24|System and method for locating an engine event|
    [返回顶部]