A combustion and electrical online identification method and system for an internal combustion power generation system

Abstract

This invention provides a method and system for online identification of combustion and electrical characteristics in an internal combustion power generation system, relating to the field of data processing technology. The method includes: denoising motion characteristic signals, combustion pressure signals, and electrical response signals to obtain corresponding denoised signals; performing time-frequency analysis on the denoised motion characteristic signals to extract piston acceleration characteristic signals; constructing a combustion pressure reconstruction model and an electrical response identification model based on the relevant denoised signals and piston acceleration characteristic signals; acquiring and denoising target motion characteristic signals in real time, and extracting target piston acceleration characteristic signals through time-frequency analysis; inputting the denoised target motion characteristic signals into the combustion pressure reconstruction model, and inputting the denoised target motion characteristic signals and target piston acceleration characteristic signals into the electrical response identification model to obtain combustion pressure reconstruction results and electrical response identification results.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for online identification of combustion and electrical properties in an internal combustion power generation system. Background Technology

[0002] Internal combustion power generation systems are crucial power supply devices. Online identification of their combustion state and electrical response is a core element in ensuring stable system operation, providing critical data support for system monitoring and control. This has broad application scenarios and research value in the field of power engineering. Reconstructing the combustion pressure and accurately identifying the electrical response of internal combustion power generation systems are essential foundations for achieving efficient system operation.

[0003] Currently, there are existing methods for signal acquisition and analysis of internal combustion power generation systems. These methods collect signals such as system motion characteristics, combustion pressure, and electrical response through sensors, and combine them with signal processing techniques and modeling methods to achieve preliminary monitoring of the system's operating status. This provides technical reference for optimizing the operation of internal combustion power generation systems, promotes the standardized development of internal combustion power generation technology, and is of great significance for improving the reliability of system operation.

[0004] However, existing technologies have not yet developed an integrated online method that takes into account both combustion pressure reconstruction and electrical response identification. The signal processing is not targeted enough, and the model construction does not fully incorporate the phase characteristics and timing features of the system operation, making it difficult to achieve accurate and synchronous online identification of combustion state and electrical response. Summary of the Invention

[0005] To address the technical challenges of the lack of an integrated online method that combines combustion pressure reconstruction and electrical response identification in existing technologies, insufficient targeting of signal processing, and inadequate model construction that does not fully incorporate the phase characteristics and timing features of system operation, thus hindering the accurate and synchronous online identification of combustion state and electrical response.

[0006] The technical solution provided by this invention is as follows: A first aspect of this invention provides a method for online identification of combustion and electrical systems in an internal combustion power generation system, comprising: S1: Acquire motion characteristic signals, combustion pressure signals, and electrical response signals of the internal combustion power generation system; S2: Denoise the motion characteristic signal, combustion pressure signal and electrical response signal to obtain the denoised motion characteristic signal, denoised combustion pressure signal and denoised electrical response signal respectively; S3: Perform time-frequency analysis on the noise-reduced motion characteristic signal to extract the piston acceleration characteristic signal of the internal combustion power generation system; S4: Construct a combustion pressure reconstruction model based on the noise-reduced motion characteristic signal and the noise-reduced combustion pressure signal; S5: Construct an electrical response identification model based on the noise-reduced motion characteristic signal, piston acceleration characteristic signal, and noise-reduced electrical response signal; S6: Real-time acquisition of target motion characteristic signals of the internal combustion power generation system, and noise reduction processing of the target motion characteristic signals to obtain noise-reduced target motion characteristic signals; S7: Perform time-frequency analysis on the noise reduction target motion characteristic signal to extract the target piston acceleration characteristic signal; S8: Input the noise reduction target motion characteristic signal into the combustion pressure reconstruction model, and input the noise reduction target motion characteristic signal and the target piston acceleration characteristic signal into the electrical response identification model to obtain the combustion pressure reconstruction result and electrical response identification result of the internal combustion power generation system.

[0007] A second aspect of the present invention provides an online combustion and electrical identification system for an internal combustion power generation system, comprising: a processor and a memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the online combustion and electrical identification method for internal combustion power generation systems as described in the first aspect.

[0008] The beneficial effects of the technical solution provided by this invention include: In this embodiment of the invention, addressing the difficulty of achieving accurate and synchronous online identification of combustion state and electrical response in existing technologies, noise reduction processing is first applied to various signals to enhance the targeting of signal processing. Then, time-frequency analysis is performed on the noise-reduced motion characteristic signals to extract piston acceleration characteristic signals, providing a reliable basis for model construction. Next, a combustion pressure reconstruction model and an electrical response identification model are constructed based on the corresponding signals, closely aligning with the system's operating characteristics. Finally, by acquiring and processing the target motion characteristic signals in real time and inputting them into the corresponding models, the combustion pressure reconstruction results and electrical response identification results are obtained synchronously, achieving integrated and accurate online identification. Attached Figure Description

[0009] Figure 1 A flowchart illustrating an online combustion and electrical identification method for an internal combustion power generation system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the global combustion pressure curve reconstruction result provided in an embodiment of the present invention.

[0010] Figure 3 This is a schematic diagram of a combustion pressure peak prediction result provided in an embodiment of the present invention.

[0011] Figure 4This is a schematic diagram of a three-phase voltage curve prediction result provided in an embodiment of the present invention.

[0012] Figure 5 This is a schematic diagram of a three-phase current curve prediction result provided in an embodiment of the present invention.

[0013] Figure 6 This is a schematic diagram of the structure of an online combustion and electrical identification system for an internal combustion power generation system provided in an embodiment of the present invention. Detailed Implementation

[0014] Reference manual attached Figure 1 The diagram shows a flowchart of an online combustion and electrical identification method for an internal combustion power generation system provided by an embodiment of the present invention.

[0015] This invention provides a method for online identification of combustion and electrical systems in an internal combustion power generation system, which may include the following steps: S1: Acquire motion characteristic signals, combustion pressure signals, and electrical response signals of the internal combustion power generation system.

[0016] Among them, the motion characteristic signal is used to reflect the piston running state, the combustion pressure signal is used to characterize the combustion state inside the cylinder, and the electrical response signal is used to reflect the electrical output characteristics of the power generation system.

[0017] In one possible implementation, S1 specifically includes sub-steps S101 to S103: S101: The piston displacement signal is acquired by a displacement sensor, and the piston speed signal is obtained based on the piston displacement signal. The piston displacement signal and the piston speed signal together constitute the motion characteristic signal.

[0018] The displacement sensor is a detection component used to collect changes in the piston position.

[0019] Specifically, the displacement sensor can be any one of the following: Hall effect crankshaft position sensor, photoelectric crankshaft position sensor, electromagnetic crankshaft position sensor, magnetic scale, optical scale, or resolver position sensor, to achieve accurate measurement of piston displacement.

[0020] Specifically, the collected piston displacement signal reflects the change in the piston's position within the cylinder. By differentiating the piston displacement signal, the piston velocity signal can be obtained. The combination of the two forms a complete motion characteristic signal, providing a foundation for subsequent signal processing and model building.

[0021] S102: The combustion pressure signal is acquired by a cylinder pressure sensor installed on the cylinder head.

[0022] Among them, the cylinder pressure sensor is a component placed in the cylinder to detect the pressure inside the cylinder.

[0023] Specifically, the cylinder pressure sensor is fixedly installed on the cylinder head of the internal combustion power generation system. It can directly detect the pressure changes during the combustion process inside the cylinder. The collected combustion pressure signal can directly characterize the combustion state inside the cylinder and is the core input data for building the combustion pressure reconstruction model.

[0024] S103: The motor controller acquires three-phase voltage signals and three-phase current signals, which together constitute the electrical response signal.

[0025] Among them, the motor controller is a control unit used to collect and transmit electrical signals of the motor.

[0026] In this embodiment of the invention, piston displacement, combustion pressure and three-phase electrical signals are collected by multiple sensors to construct a complete system of motion characteristics, combustion pressure and electrical response signals, providing comprehensive and reliable raw data support for subsequent model training and online identification.

[0027] S2: The motion characteristic signal, combustion pressure signal and electrical response signal are denoised to obtain the denoised motion characteristic signal, denoised combustion pressure signal and denoised electrical response signal respectively.

[0028] Among them, noise reduction processing is the process of removing interference components from the signal.

[0029] Optionally, the noise reduction process specifically includes moving average noise reduction, median filtering noise reduction, wavelet noise reduction, and wavelet packet decomposition.

[0030] It should be noted that noise reduction processing is performed separately for motion characteristic signals, combustion pressure signals, and electrical response signals. Different types of signals can use the same or different noise reduction methods. The purpose is to remove interference components in the signal, avoid interference signals affecting the accuracy of subsequent time-frequency analysis and model construction, and ensure that the noise-reduced signal obtained after processing can truly reflect the system's operating status.

[0031] In this embodiment of the invention, multiple noise reduction methods are used to process the three types of signals, effectively eliminating interference components, improving the signal-to-noise ratio and realism of the signal, avoiding the interference of noise on time-frequency analysis and model accuracy, and ensuring the accuracy of subsequent feature extraction and modeling.

[0032] S3: Perform time-frequency analysis on the noise reduction motion characteristic signal to extract the piston acceleration characteristic signal of the internal combustion power generation system.

[0033] Time-frequency analysis is an analytical method that analyzes signals simultaneously in both the time and frequency domains.

[0034] In one possible implementation, S3 specifically includes sub-steps S301 to S303: S301: Based on Fourier transform, frequency analysis is performed on the noise-reduced motion characteristic signal to obtain the main operating frequency of the internal combustion power generation system.

[0035] The Fourier transform is a mathematical method for converting time-domain signals into frequency-domain information.

[0036] It should be noted that the system contains multiple frequency components during operation, and the main operating frequencies can be decomposed through Fourier transform (e.g., (and secondary operating frequency).

[0037] S302: By using time-frequency domain signal processing, the first-order differential signal of the velocity signal in the noise-reduced motion characteristic signal is processed, and the target frequency signal component corresponding to the main operating frequency is obtained.

[0038] Specifically, methods such as short-time Fourier transform, empirical mode decomposition, ensemble empirical mode decomposition, or variational mode decomposition can be used to decompose the first-order differential signal of the velocity signal in the time-frequency domain, thereby obtaining the signal components corresponding to each major operating frequency, such as... corresponding ,and corresponding And so on.

[0039] S303: Reconstruct the target frequency signal components to extract the piston acceleration characteristic signal.

[0040] Among them, the piston acceleration characteristic signal is used to reflect the dynamic changes in piston operation.

[0041] Furthermore, the recombination process involves adding the signal components corresponding to the main operating frequency to obtain the piston acceleration characteristic signal. Its expression is: in, This represents the characteristic signal of piston acceleration. This represents the target frequency signal component corresponding to the first major operating frequency of the internal combustion generator system. This represents the target frequency signal component corresponding to the second primary operating frequency of the internal combustion generator system. This indicates the target frequency signal components corresponding to the third main operating frequency of the internal combustion generator system. Indicates the first [unclear] with internal combustion power generation system The target frequency signal components corresponding to the main operating frequencies. This indicates the number of main operating frequencies of the internal combustion power generation system.

[0042] In this embodiment of the invention, the signal components corresponding to the main operating frequencies are extracted by Fourier transform and time-frequency domain processing, and the piston acceleration feature signal is reconstructed to accurately capture the dynamic characteristics of the piston and provide high-quality time-series feature input for the construction of dual models.

[0043] S4: Construct a combustion pressure reconstruction model based on the noise-reduced motion characteristic signal and the noise-reduced combustion pressure signal.

[0044] Among them, the combustion pressure reconstruction model is used to calculate the in-cylinder combustion pressure of the system.

[0045] Optionally, the combustion pressure reconstruction model is specifically an improved Gaussian process regression model.

[0046] Among them, the improved Gaussian process regression model is a regression modeling method adapted to the phase characteristics of internal combustion systems.

[0047] It should be noted that Gaussian process regression is based on the mean function. Sum of covariance functions By definition, its probability distribution can be expressed as: in, Represents the input vector The combustion pressure signal is a random process with the independent variable being the combustion pressure signal. GPR This indicates an improved Gaussian process regression model. Represents the input vector The mean function, Represents the input vector and The covariance function, This represents the input vector composed of the piston displacement signal and the piston velocity signal. Indicates and The other set of input vectors, ∼, indicates that they follow a certain probability distribution.

[0048] The kernel function of the improved Gaussian process regression model is specifically a phase-adaptive piecewise composite kernel function.

[0049] Specifically, this invention provides a customized improvement to the covariance kernel function for the four stages of the internal combustion engine's operation (intake, compression, combustion expansion, and exhaust), proposing a phase-adaptive piecewise composite kernel function: in, These represent the weighting coefficients of the sub-kernel function during the intake phase. This represents the weight coefficients of the sub-kernel functions in the compression phase. The weighting coefficients of the sub-kernel function representing the combustion expansion stage are shown. This represents the weighting coefficients of the sub-kernel function in the exhaust phase. This represents the quadratic exponential kernel function adapted to the intake phase. This represents the quadratic exponential kernel function adapted for the exhaust phase. Indicates the adaptation compression stage Kernel function, Denotes the rational quadratic kernel function that adapts to the combustion expansion stage. This represents the phase interval indicator function. This indicates the real-time operating phase of the internal combustion engine. This indicates the phase interval corresponding to the intake phase of an internal combustion engine. This indicates the phase interval corresponding to the compression phase of an internal combustion engine. This indicates the phase interval corresponding to the combustion expansion stage of an internal combustion engine. This indicates the phase interval corresponding to the exhaust phase of an internal combustion engine. It indicates that it belongs to.

[0050] In this embodiment of the invention, an improved Gaussian process regression model adapted to the four-stage phase characteristics of an internal combustion engine is adopted, combined with a phase-adaptive piecewise composite kernel function, which can accurately fit the correlation between combustion pressure and motion characteristics, and realize reliable reconstruction of in-cylinder combustion pressure.

[0051] S5: Based on the noise-reduced motion characteristic signal, piston acceleration characteristic signal and noise-reduced electrical response signal, construct an electrical response identification model.

[0052] Among them, the electrical response identification model is used to predict the electrical output signal of the system.

[0053] Optionally, the electrical response identification model is specifically an integrated deep learning model.

[0054] Among them, the ensemble deep learning model is a temporal feature learning model that integrates multiple network structures.

[0055] Furthermore, the model consists of a one-dimensional convolutional neural network, a bidirectional long short-term memory network, a self-attention mechanism module, and a temporally distributed fully connected layer connected sequentially.

[0056] For example, a one-dimensional convolutional neural network consists of two structurally identical modules arranged sequentially. Each module contains a convolutional layer, a batch normalization layer, and a linear rectifier layer, used to extract fused features of multi-channel signals along the time axis. Subsequently, a bidirectional long short-term memory network extracts temporal features and dynamic response characteristics. A self-attention mechanism module captures local features at key instantaneous locations. Finally, a time-distributed fully connected layer outputs the three-phase current. and three-phase voltage The time curve.

[0057] Optionally, the integrated deep learning model specifically includes a one-dimensional convolutional neural network, a bidirectional long short-term memory network, a self-attention mechanism module, and a time-distributed fully connected layer connected in sequence.

[0058] Optionally, the loss function of the integrated deep learning model is specifically a two-branch loss function.

[0059] The loss function is used to measure the difference between the model's prediction and the actual value.

[0060] The bi-branch loss function specifically includes a data-driven loss term and a physical mechanism constraint loss term.

[0061] It should be noted that this loss function incorporates the physical constraints of the linear motor's electrical mechanism equations, and its total loss function... Represented as: in, This represents the total loss function of the ensemble deep learning models. This represents the data-driven loss term. Represents the adaptive balance coefficient. This represents the loss term due to physical mechanism constraints.

[0062] Specifically, the physical mechanism-constrained loss term Based on the voltage balance equation of the linear motor, the expression is as follows: in, This indicates the total number of time steps for signal acquisition. Indicates the first Time of the first The identification value of phase voltage, This represents the stator resistance of the motor. Indicates the first Time of the first Identification value of phase current, Indicates the stator inductance of the motor. This represents the back electromotive force coefficient of the motor. Indicates the first Piston speed signal at any time Indicates the time step number. The phase sequence identifier for a three-phase motor. A , B , C Each represents one phase of the three-phase motor. Represents the differential operator. This represents the squaring operation of the L2 norm.

[0063] In this embodiment of the invention, an integrated deep learning model that incorporates multiple network structures is constructed, and a bi-branch loss function with physical mechanism constraints is introduced. This model learns both temporal features and conforms to the electrical laws of motors, significantly improving the accuracy and robustness of electrical response identification.

[0064] S6: Real-time acquisition of target motion characteristic signals of the internal combustion power generation system, and noise reduction processing of the target motion characteristic signals to obtain noise-reduced target motion characteristic signals.

[0065] Among them, the target motion characteristic signal is the real-time piston operation signal collected during the online identification stage.

[0066] S7: Perform time-frequency analysis on the noise reduction target motion characteristic signal to extract the target piston acceleration characteristic signal.

[0067] S8: Input the noise reduction target motion characteristic signal into the combustion pressure reconstruction model, and input the noise reduction target motion characteristic signal and the target piston acceleration characteristic signal into the electrical response identification model to obtain the combustion pressure reconstruction result and electrical response identification result of the internal combustion power generation system.

[0068] In this embodiment of the invention, real-time signals are input into the combustion pressure reconstruction model and the electrical response identification model respectively, and the combustion state and electrical response results are output synchronously to realize integrated online monitoring of combustion and electrical performance, providing a comprehensive basis for system regulation.

[0069] Reference manual attached Figure 2 The diagram shows a schematic representation of a global combustion pressure curve reconstruction result provided by an embodiment of the present invention.

[0070] Specifically, Figure 2 The horizontal axis represents time (seconds), ranging from 0.00 to 0.25 seconds, representing the operating time of the internal combustion generator system. The vertical axis represents combustion pressure (bars), ranging from 0 to 60 bar, representing the magnitude of combustion pressure inside the cylinder. The graph contains two curves: the dashed line represents the actual combustion pressure value, and the solid line represents the predicted combustion pressure value. Both curves exhibit periodic fluctuations over time, with significant pressure peaks appearing at approximately 0.02, 0.07, 0.12, 0.17, and 0.22 seconds.

[0071] Furthermore, in terms of inclusion relationships, Figure 2 It integrates four core elements: time axis, combustion pressure axis, actual value curve, and predicted value curve. In terms of connectivity, the actual and predicted values ​​of combustion pressure change synchronously with time, and their peak shapes and baseline levels in each cycle highly overlap, with only minor deviations.

[0072] Figure 2 By comparing the global curves of the actual and predicted combustion pressure values, the accuracy of the combustion pressure reconstruction model of this invention in fitting the in-cylinder combustion pressure is intuitively verified. It can effectively capture the peak characteristics and baseline changes of combustion pressure within a complete cycle, providing an intuitive basis for reliable monitoring of the combustion state of the internal combustion power generation system and ensuring the accuracy and completeness of the combustion pressure reconstruction results.

[0073] Reference manual attached Figure 3 The diagram illustrates a peak combustion pressure prediction result provided by an embodiment of the present invention.

[0074] Specifically, Figure 3 The horizontal axis represents the cycle, ranging from 0 to 40, indicating the number of working cycles of the internal combustion generator system. The vertical axis on the left represents the peak combustion pressure (bar), ranging from 40 to 60 bar, indicating the peak value of the combustion pressure in the cylinder. The vertical axis on the right represents the relative error percentage (%), ranging from 0 to 15%, indicating the relative error between the predicted and actual values. The graph contains three curves: the solid line with a triangle represents the actual peak combustion pressure, the dashed line with a circle represents the predicted peak combustion pressure, and the dotted line with a circle represents the relative error.

[0075] Furthermore, in terms of inclusion relationships, Figure 3 It integrates the cycle axis, peak combustion pressure axis, relative error percentage axis, as well as the actual peak combustion pressure curve, predicted peak combustion pressure curve, and relative error curve. In terms of connection, the actual and predicted peak combustion pressure values ​​fluctuate synchronously with the number of cycles, and the numerical difference between the two is intuitively presented through the relative error curve. The relative error is calculated from the actual and predicted values ​​and fluctuates within the range of 0 to 10% with the cycle changes.

[0076] Figure 3 By comparing the actual and predicted values ​​of the peak combustion pressure under multiple cycles, the accuracy of the combustion pressure reconstruction model of this invention in predicting peak pressure is intuitively demonstrated. It can stably capture the trend of peak pressure change in continuous working cycles, and the relative error is generally at a low level. This provides an intuitive basis for long-term reliable monitoring of the combustion state of the internal combustion power generation system and effectively verifies the robustness and prediction accuracy of the model under multiple working conditions.

[0077] Reference manual attached Figure 4 The diagram shows a three-phase voltage curve prediction result provided by an embodiment of the present invention.

[0078] Specifically, Figure 4 The horizontal axis represents time (milliseconds), ranging from 0 to 80 milliseconds, representing the operating time of the internal combustion generator system. The vertical axis represents three-phase voltage (V), ranging from -200V to 200V, representing the voltage magnitude of each phase of the three-phase motor. The graph contains two types of curves: the solid line represents the actual value of the three-phase voltage. The dashed line represents the predicted three-phase voltage value. Each type of curve contains three sine curves of phases A, B, and C, with a phase difference of 120° between them.

[0079] Furthermore, in terms of inclusion relationships, Figure 4It integrates a time axis, a three-phase voltage axis, and sets of actual and predicted three-phase voltage curves. In terms of connection, the actual and predicted three-phase voltage values ​​exhibit periodic sinusoidal fluctuations synchronously over time. The phases of phases A, B, and C differ by 120° sequentially, and the actual and predicted value curves of each phase highly overlap with each other, with only minor deviations.

[0080] Figure 4 By comparing the time-series curves of the actual and predicted three-phase voltage values, the electrical response identification model of this invention is intuitively verified to accurately fit the three-phase voltage. It can effectively capture the phase characteristics and amplitude changes of the three-phase voltage within a complete time period, providing an intuitive basis for the reliable monitoring of the electrical status of the internal combustion power generation system and ensuring the accuracy and completeness of the electrical response identification results.

[0081] Reference manual attached Figure 5 The diagram shows a three-phase current curve prediction result provided by an embodiment of the present invention.

[0082] Specifically, Figure 5 The horizontal axis represents time (milliseconds), ranging from 0 to 80 milliseconds, representing the operating time of the internal combustion generator system. The vertical axis represents three-phase current (A), ranging from -15A to 15A, representing the current magnitude of each phase of the three-phase motor. The graph includes two types of curves: the solid line represents the actual value of the three-phase current. The dashed line represents the predicted three-phase current value. Each type of curve contains three sine curves of phases A, B, and C, with a phase difference of 120° between them.

[0083] Furthermore, in terms of inclusion relationships, Figure 5 It integrates a time axis, a three-phase current axis, and sets of actual and predicted three-phase current curves. In terms of connection, the actual and predicted three-phase current values ​​exhibit periodic sinusoidal fluctuations synchronously over time. The phases of phases A, B, and C differ by 120° sequentially, and the actual and predicted curves of each phase highly overlap with each other, with only minor deviations.

[0084] Figure 5 By comparing the time-series curves of the actual and predicted values ​​of the three-phase current, the electrical response identification model of this invention is intuitively verified to accurately fit the three-phase current. It can effectively capture the phase characteristics and amplitude changes of the three-phase current within a complete time period, providing an intuitive basis for the reliable monitoring of the electrical status of the internal combustion power generation system and ensuring the accuracy and completeness of the electrical response identification results.

[0085] Optionally, after S8, it also includes: S9: Based on the electrical response identification results, calculate the transient output power of the internal combustion power generation system to improve the online electrical identification of the internal combustion power generation system.

[0086] Among them, transient output power is used to characterize the system's real-time power generation capability.

[0087] Specifically, transient output power It can be calculated from the dot product of three-phase current and three-phase voltage: in, This indicates the transient output power of the internal combustion generator system. express A Identification value of phase current, express B Identification value of phase current, express C Identification value of phase current, express A The identification value of phase voltage, express B The identification value of phase voltage, express C The identification value of phase voltage.

[0088] Reference manual attached Figure 6 The diagram shows a schematic representation of an online combustion and electrical identification system for an internal combustion power generation system provided in an embodiment of the present invention.

[0089] This invention provides an online combustion and electrical identification system 20 for an internal combustion power generation system, comprising: a processor 201 and a memory 202; The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-described method for online identification of combustion and electrical systems in an internal combustion power generation system and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.

[0090] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

[0091] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0092] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).

Claims

1. A method for online identification of combustion and electrical systems in an internal combustion power generation system, characterized in that, include: S1: Acquire motion characteristic signals, combustion pressure signals, and electrical response signals of the internal combustion power generation system; S2: Perform noise reduction processing on the motion characteristic signal, the combustion pressure signal, and the electrical response signal to obtain the noise-reduced motion characteristic signal, the noise-reduced combustion pressure signal, and the noise-reduced electrical response signal, respectively. S3: Perform time-frequency analysis on the noise-reduced motion characteristic signal to extract the piston acceleration characteristic signal of the internal combustion power generation system; S4: Based on the noise-reduced motion characteristic signal and the noise-reduced combustion pressure signal, construct a combustion pressure reconstruction model; S5: Based on the noise-reduced motion characteristic signal, the piston acceleration characteristic signal, and the noise-reduced electrical response signal, construct an electrical response identification model; S6: Acquire the target motion characteristic signal of the internal combustion power generation system in real time, and perform noise reduction processing on the target motion characteristic signal to obtain a noise-reduced target motion characteristic signal; S7: Perform time-frequency analysis on the noise reduction target motion feature signal to extract the target piston acceleration feature signal; S8: Input the noise reduction target motion characteristic signal into the combustion pressure reconstruction model, and input the noise reduction target motion characteristic signal and the target piston acceleration characteristic signal into the electrical response identification model to obtain the combustion pressure reconstruction result and electrical response identification result of the internal combustion power generation system.

2. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, S1 specifically includes: S101: Collect piston displacement signal through displacement sensor, and obtain piston speed signal based on piston displacement signal, wherein piston displacement signal and piston speed signal together constitute the motion characteristic signal; S102: The combustion pressure signal is acquired by a cylinder pressure sensor installed on the cylinder head; S103: The motor controller acquires three-phase voltage signals and three-phase current signals, wherein the three-phase voltage signals and the three-phase current signals together constitute the electrical response signal.

3. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, The noise reduction process specifically includes moving average noise reduction, median filtering noise reduction, wavelet noise reduction, and wavelet packet decomposition.

4. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, S3 specifically includes: S301: Based on Fourier transform, perform frequency analysis on the noise-reduced motion characteristic signal to obtain the main operating frequency of the internal combustion power generation system; S302: The first-order differential signal of the velocity signal in the noise-reduced motion characteristic signal is processed by time-frequency domain signal processing, and the target frequency signal component corresponding to the main operating frequency is obtained. S303: Reconstruct the target frequency signal components to extract the piston acceleration feature signal.

5. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, The combustion pressure reconstruction model is specifically an improved Gaussian process regression model; The kernel function of the improved Gaussian process regression model is specifically a phase-adaptive piecewise composite kernel function.

6. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, The electrical response identification model is specifically an integrated deep learning model.

7. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 6, characterized in that, The integrated deep learning model specifically includes a one-dimensional convolutional neural network, a bidirectional long short-term memory network, a self-attention mechanism module, and a time-distributed fully connected layer connected in sequence.

8. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 6, characterized in that, The loss function of the integrated deep learning model is specifically a two-branch loss function; The bi-branch loss function specifically includes a data-driven loss term and a physical mechanism constraint loss term.

9. The method for online identification of combustion and electrical systems in an internal combustion power generation system according to claim 1, characterized in that, Following S8, the following is also included: S9: Based on the electrical response identification results, calculate the transient output power of the internal combustion power generation system to improve the online electrical identification of the internal combustion power generation system.

10. An online combustion and electrical identification system for an internal combustion power generation system, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the online combustion and electrical identification method for an internal combustion power generation system as described in any one of claims 1 to 9.

Images