A natural gas engine control method, related device and natural gas engine

Abstract

This application discloses a natural gas engine control method, related devices, and a natural gas engine, relating to the field of internal combustion engines. By extracting features from the operating parameters of the natural gas engine, characteristic parameters such as knock characteristics, cycle variation indices, and misfire rate are obtained, representing the natural gas combustion state of the engine. After preprocessing these characteristic parameters, a combustion parameter prediction model is used. Taking the combustion performance indicators under the current operating conditions as the target and the characteristic data as a prerequisite, prediction processing is performed to determine the state values ​​of each combustion parameter adapted to the current operating conditions: EGR rate, ignition angle, etc. Then, adjustments are made according to the actuators corresponding to each state value: EGR valve, ignition coil, etc. This method can predict combustion parameters based on real-time changing operating parameters and adjust them in real-time based on the prediction results, ensuring that all combustion performance indicators are within a reasonable range, effectively reducing gas consumption and improving thermal efficiency.

Description

Technical Field

[0001] This application relates to the field of internal combustion engine technology, and in particular to a natural gas engine control method, related devices, and a natural gas engine. Background Technology

[0002] Combustion boundary, a crucial parameter for measuring the operating state of internal combustion engines such as those using natural gas, primarily describes the limiting conditions or critical states of fuel-air combustion within the combustion chamber. It involves constraints related to combustion stability, efficiency, emissions, and safety. In actual vehicle operation, the engine is affected by numerous factors, including changes in the operating boundary, measurement deviations, and engine hardware aging. Currently, most natural gas engines operate using combustion parameters set at their calibration conditions. Under these varying engine operating conditions, achieving high fuel economy is often difficult, leading to reduced combustion efficiency and failure to meet stability requirements. Summary of the Invention

[0003] In view of the above problems, this application provides a natural gas engine control method, related device, and natural gas engine to improve thermal efficiency and gas utilization rate. The specific solution is as follows:

[0004] The first aspect of this application provides a natural gas engine control method, comprising:

[0005] The operating parameters of the natural gas engine are feature extracted to obtain characteristic parameters that characterize the natural gas combustion state of the natural gas engine. The characteristic parameters include at least: knock characteristics, cycle variation index and misfire rate of each cylinder.

[0006] The feature parameters are preprocessed to obtain feature parameters with a uniform format. The feature data, including at least the feature parameters with a uniform format, is then input into the combustion parameter prediction model for prediction processing. The combustion parameter prediction model uses the combustion performance index under the current operating conditions as the target and the feature data as the premise to determine the state value of each combustion parameter.

[0007] The current state of the actuator corresponding to each combustion parameter is adjusted based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle. The actuators include: EGR valve and ignition coil.

[0008] In one possible implementation, the detonation features include: detonation frequency and detonation intensity. The feature extraction of the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine includes:

[0009] After bandpass filtering the acquired detonation signal, the frequency domain features are extracted using fast Fourier transform, and the detonation frequency is determined based on the detonation peak frequency.

[0010] The detonation intensity is determined based on the amplitude of the detonation frequency after bandpass filtering and the amplitude threshold.

[0011] In one possible implementation, the feature extraction of the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine includes:

[0012] Extract pressure index data for each engine cycle, and obtain the inter-cycle variation coefficient based on the pressure index data;

[0013] The cyclic variation coefficient is fused with at least the vortex front pressure fluctuation to obtain the cyclic variation index.

[0014] In one possible implementation, the feature extraction of the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine includes:

[0015] The number of misfires of each cylinder within the engine cycle is determined based on the turbine exhaust pressure, secondary discharge time, and rotational roughness, and the misfire rate of each cylinder is determined based on the number of misfires and the engine cycle.

[0016] In one possible implementation, the process of constructing the combustion parameter prediction model includes:

[0017] Using feature data as input and combustion performance index data as the target, a gradient boosting decision tree regression model is constructed.

[0018] Based on the surrogate model, the mapping relationship between the objective function and the combustion parameters is fitted in the combustion parameter search space to obtain the target combustion parameter combination. The target combustion parameter combination is the combination of combustion parameters that optimizes the function value of the objective function obtained by the surrogate model during the iteration process. The objective function is the target optimization function composed of the combustion performance index.

[0019] The gradient boosting decision tree regression model is trained based on the combination of target combustion parameters and the feature data to obtain the combustion parameter prediction model.

[0020] In one possible implementation, the step of fitting the mapping relationship between the objective function and combustion parameters within the combustion parameter search space based on the surrogate model to obtain the target combustion parameter combination includes:

[0021] Iterative processing is performed within the combustion parameter search space, and the optimal parameter combination and corresponding objective function value of each iteration are recorded. In each iteration, candidate parameter combinations are selected based on the acquisition function, and the candidate parameter combinations are substituted into the engine model to evaluate the objective function value.

[0022] The historical dataset is updated using the optimal parameter combination and corresponding objective function value from each iteration, and the proxy model is retrained based on the updated historical dataset until the iteration termination condition is met.

[0023] The combination of combustion parameters that optimizes the objective function value is selected from the surrogate model that terminates the iteration, and this combination is taken as the target combustion parameter combination.

[0024] In one possible implementation, the preprocessing of the feature parameters to obtain feature parameters with a uniform format includes:

[0025] The feature parameters are sequentially subjected to data cleaning, noise reduction, filtering, and data normalization to obtain the feature parameters with a unified format.

[0026] A second aspect of this application provides a natural gas engine control device, comprising:

[0027] The combustion feature extraction module is used to extract features from the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine. The feature parameters include at least: knock characteristics, cycle variation index and misfire rate of each cylinder.

[0028] A combustion parameter determination module is used to preprocess the feature parameters to obtain feature parameters with a uniform format, and input feature data, including at least the feature parameters with the uniform format, into a combustion parameter prediction model for prediction processing. This allows the combustion parameter prediction model to determine the state values ​​of each combustion parameter based on the combustion performance indicators under the current operating conditions and the feature data as a prerequisite.

[0029] The combustion parameter adjustment module is used to adjust the current state of the actuator corresponding to each combustion parameter based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle, and the actuator includes: EGR valve and ignition coil.

[0030] A third aspect of this application provides a computer program product including computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the natural gas engine control method of the first aspect or any implementation thereof.

[0031] A fourth aspect of this application provides an electronic device, including at least one processor and a memory connected to the processor, wherein:

[0032] The memory is used to store computer programs;

[0033] The processor is used to execute the computer program so that the electronic device can implement the natural gas engine control method of the first aspect or any implementation thereof.

[0034] The fifth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the natural gas engine control method of the first aspect or any implementation thereof.

[0035] The sixth aspect of this application provides a natural gas engine, including the electronic equipment described in the fourth aspect above.

[0036] By employing the aforementioned technical solution, the natural gas engine control method provided in this application extracts features from the operating parameters of the natural gas engine to obtain characteristic parameters such as knock characteristics, cycle variation index, and misfire rate, which characterize the natural gas combustion state of the engine. After preprocessing these characteristic parameters, a combustion parameter prediction model is used, with the combustion performance indicators under the current operating condition as the target and the characteristic data as a prerequisite, to perform prediction processing and determine the state values ​​of each combustion parameter adapted to the current operating condition: EGR rate, ignition angle, etc. Then, adjustments are made according to the actuators corresponding to each state value: EGR valve, ignition coil, etc. This method can predict combustion parameters based on real-time changes in engine operating parameters and adjust the combustion parameters in real time based on the prediction results, ensuring that all combustion performance indicators of the engine are within a reasonable range, effectively reducing gas consumption, improving thermal efficiency, and enhancing operational stability. Attached Figure Description

[0037] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0038] Figure 1 A flowchart of a natural gas engine control method provided in this application;

[0039] Figure 2 A control architecture diagram of a natural gas engine is provided in this application;

[0040] Figure 3A framework diagram of a combustion parameter prediction model provided in this application;

[0041] Figure 4 A cylinder pressure comparison diagram provided for this application;

[0042] Figure 5 A structural diagram of a natural gas engine control device provided in this application;

[0043] Figure 6 A structural diagram of an electronic device provided in this application. Detailed Implementation

[0044] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0045] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0046] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0047] Currently, conventional natural gas engine combustion parameters are too rigid, lacking the ability to control ignition angle and EGR parameters based on combustion boundaries. In actual vehicle operation, numerous factors arise, including changes in operating boundaries, measurement deviations, and engine hardware aging. Traditional solutions often use calibrated combustion parameters, which fail to achieve optimal fuel economy in these scenarios, resulting in decreased combustion efficiency. Furthermore, combustion control using cylinder pressure is prohibitively expensive for commercialization.

[0048] To address the aforementioned problems, this application provides a natural gas engine control method. The natural gas engine control method of this application embodiment will be described in detail below with reference to the accompanying drawings.

[0049] Reference Figure 1 , Figure 1This is a flowchart illustrating a natural gas engine control method provided in an embodiment of this application, as shown below. Figure 1 As shown, the natural gas engine control method may include steps S101 to S103, which are described in detail below.

[0050] Step S101: Extract features from the operating parameters of the natural gas engine to obtain feature parameters that characterize the natural gas combustion state of the natural gas engine. The feature parameters include at least: knock characteristics, cycle variation index and misfire rate of each cylinder.

[0051] Specifically, refer to Figure 2 As shown, in the combustion parameter observation solution stage, existing signals such as the engine's own knock signal, speed roughness (reflecting the fluctuation of the engine crankshaft instantaneous speed and characterizing the non-uniformity of the combustion process in each cylinder; evaluated by calculating the standard deviation or coefficient of variation of crankshaft angular acceleration), turbine exhaust pressure waveform, and ignition coil discharge time are used to extract characteristic parameters characterizing the engine combustion state, replacing the high-cost dependence on cylinder pressure.

[0052] For example, the knock intensity and knock frequency can be obtained by processing the knock signal; the corresponding misfire probability (misfire rate) can be obtained by the rotational roughness, the waveform of the turbine exhaust pressure, and the discharge time of the secondary coil; and the cyclic variation of each cylinder can be obtained by the turbine exhaust pressure parameters.

[0053] Step S102: Preprocess the feature parameters to obtain feature parameters with a uniform format, and input the feature data, including at least the feature parameters with a uniform format, into the combustion parameter prediction model for prediction processing, so that the combustion parameter prediction model determines the state value of each combustion parameter with the combustion performance index under the current working condition as the target and the feature data as the premise.

[0054] Specifically, based on obtaining the above-mentioned characteristic parameters, refer to Figure 2 As shown, in the combustion parameter evaluation stage: various characteristic parameters and real-time combustion parameters are input into the combustion parameter prediction model to predict the optimal combustion parameters under the current operating conditions, thereby obtaining the optimal ignition angle, EGR rate and other control parameters under the current operating conditions.

[0055] Step S103: Adjust the current state of the actuators corresponding to each combustion parameter based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle, and the corresponding actuators include: EGR valve and ignition coil.

[0056] Specifically, based on the combustion parameters mentioned above, and referring to Figure 2, the actuator is adjusted accordingly during the combustion parameter control stage. For example, the ignition timing of the ignition coil is adjusted according to the ignition angle, and the opening of the EGR valve is adjusted according to the EGR rate. This ensures the combustion curve is controlled within a reasonable range, reducing gas consumption and improving thermal efficiency.

[0057] This natural gas engine control method infers the correlation between combustion parameters and knock intensity and frequency by analyzing the performance of knock vibration signals across all operating conditions. It infers the correlation between current combustion parameters and misfire probability by analyzing parameters such as engine speed roughness, turbine exhaust pressure waveform, and secondary coil discharge time. It predicts the cycle variations of combustion parameters in each cylinder based on turbine exhaust pressure parameters. Finally, it adjusts combustion parameters in real time based on current in-cylinder intake air volume, EGR, ignition angle, and a combustion parameter prediction model. Through advanced combustion predictive control technology, the combustion curve is controlled within a reasonable range, reducing fuel consumption and improving thermal efficiency.

[0058] In another embodiment, the aforementioned detonation features include: detonation frequency and detonation intensity. The aforementioned feature extraction of the operating parameters of the natural gas engine to obtain characteristic parameters characterizing the natural gas combustion state of the natural gas engine may specifically include:

[0059] After bandpass filtering the acquired detonation signal, frequency domain features were extracted using fast Fourier transform, and the detonation frequency was determined based on the detonation peak frequency.

[0060] The detonation intensity is determined based on the amplitude of the detonation frequency after bandpass filtering and the amplitude threshold.

[0061] Specifically, the detonation signal is bandpass filtered (typically 5-15 kHz, matching the engine's detonation characteristic frequency) to remove background noise and interference. Then, a fast Fourier transform is used to extract frequency domain features and locate the detonation peak frequency. The root mean square (RMS) or peak amplitude of the filtered signal is taken and compared with a threshold to determine whether detonation has occurred and its severity, thereby determining the detonation intensity.

[0062] The process of obtaining cyclical change indicators includes:

[0063] Pressure index data for each engine cycle is extracted, and the coefficient of variation between cycles is obtained based on the pressure index data.

[0064] The cyclic variation coefficient is fused with at least the vortex front pressure fluctuation to obtain the cyclic variation index.

[0065] Specifically, taking the engine cycle as a unit, key indicators such as indicated mean effective pressure and peak cylinder pressure for each cycle are extracted. The coefficient of variation (COV) between cycles is calculated.

[0066] ;

[0067] in, Standard deviation, The average value is used to quantify the stability of the combustion cycle. Finally, by integrating multi-dimensional information such as turbine exhaust pressure fluctuation and speed fluctuation, a comprehensive cycle variation index is obtained.

[0068] The process of determining the misfire rate of each cylinder includes:

[0069] The number of misfires per cylinder within the engine cycle is determined based on the turbine exhaust pressure, secondary discharge time, and speed roughness. The misfire rate of each cylinder is then determined based on the number of misfires and the engine cycle.

[0070] Specifically, crankshaft speed roughness can be used to identify misfire cycles through instantaneous crankshaft speed fluctuations (time difference between adjacent teeth); a significant drop in speed occurs during a misfire. For secondary discharge time: an abnormally shortened / interrupted spark plug breakdown voltage duration can be used as a criterion for misfire detection. For turbine exhaust pressure: the exhaust pressure during a misfire cycle will be significantly lower than in a normal combustion cycle; this, combined with a threshold, can be used to determine a misfire event. Finally, the misfire rate is calculated by statistically analyzing the number of misfires within a certain number of engine cycles.

[0071] Fire failure rate = (number of fire failure cycles / total number of cycles) × 100%.

[0072] In some embodiments, refer to Figure 3 As shown, the above preprocessing of feature parameters to obtain feature parameters with a uniform format can specifically include:

[0073] The feature parameters are sequentially processed by data cleaning, noise reduction, filtering, and data normalization to obtain feature parameters with a uniform format.

[0074] Specifically, the data cleaning process includes:

[0075] Missing value handling: For missing data caused by sensor failure or communication interruption, linear interpolation, mean filling, or time series prediction filling are used.

[0076] Outlier removal: Use the 3σ principle or box plot method to identify and remove extreme values.

[0077] Time alignment: Align multi-source signals such as knock and speed according to crankshaft angle (°CA) or timestamp to ensure timing consistency.

[0078] The noise reduction and filtering process includes:

[0079] Filters are selected based on the characteristics of different signals. For example, for time-domain signals (such as rotational speed), Butterworth low-pass filters are used to remove high-frequency mechanical noise. For frequency-domain signals (such as detonation signals), band-pass filters are used to retain only the energy of the target frequency band.

[0080] Finally, data normalization is performed: Min-Max normalization or Z-Score normalization can be used for the corresponding normalization process. For example, when using Min-Max, the normalization can be achieved through:

[0081] ;

[0082] Map the data to the interval [0,1], where x′ represents the normalized value, x represents the original value, and x' represents the normalized value. min Let x represent the minimum value in the original dataset. max This represents the maximum value in the original dataset.

[0083] When using Z-Score here, the following is used:

[0084] ;

[0085] The data is standardized by setting the mean to 0 and the variance to 1. Here, μ represents the mean of the original dataset, and σ represents the standard deviation of the original dataset.

[0086] It is understood that those skilled in the art may use other means to normalize the data, and no restrictions are imposed here.

[0087] In other embodiments, the construction process of the combustion parameter prediction model in the above embodiments may specifically include:

[0088] Using feature data as input and combustion performance index data as the target, a gradient boosting decision tree regression model is constructed.

[0089] Based on the surrogate model, the mapping relationship between the objective function and the combustion parameters is fitted in the combustion parameter search space to obtain the target combustion parameter combination. The target combustion parameter combination is the combination of combustion parameters that optimizes the function value of the objective function obtained by the surrogate model during the iteration process. The objective function is the target optimization function composed of combustion performance indicators.

[0090] The gradient boosting decision tree regression model is trained based on the combination of combustion parameters for each target and the feature data to obtain the combustion parameter prediction model.

[0091] The process of fitting the mapping relationship between the objective function and combustion parameters in the combustion parameter search space based on the surrogate model to obtain the target combustion parameter combination can specifically include:

[0092] The process is iterative within the combustion parameter search space, and the optimal parameter combination and corresponding objective function value are recorded for each iteration. In each iteration, candidate parameter combinations are selected based on the acquisition function, and the candidate parameter combinations are substituted into the engine model to evaluate the objective function value.

[0093] The historical dataset is updated using the optimal parameter combination and corresponding objective function value from each iteration, and the surrogate model is retrained based on the updated historical dataset until the iteration termination condition is met.

[0094] The combination of combustion parameters that optimizes the objective function value is selected from the surrogate model that terminates the iteration, and this combination is taken as the target combustion parameter combination.

[0095] Specifically, refer to Figure 3 As shown, in the process of combustion parameter optimization and model training, the parameter search space is first determined: the combustion control parameters to be optimized are identified, such as ignition advance angle (°CA), EGR rate (%), and injection timing. Based on engine bench tests or engineering experience, the physical feasible range of each parameter is set (e.g., ignition angle: -10°~40°ATDC, EGR rate: 0%~30%, etc.). Finally, a discrete or continuous parameter search distribution is constructed as the input space for the subsequent surrogate model.

[0096] Bayesian-GBDT regression model:

[0097] Using preprocessed features (knock intensity, cyclic variation, misfire rate, etc.) as input and combustion performance indicators (thermal efficiency, emissions, gas consumption) as targets, a GBDT regression model (i.e., gradient boosting decision tree regression model) is constructed.

[0098] The surrogate model can employ a Bayesian optimization model, using a Gaussian process model to fit the mapping relationship between the objective function and the parameters. The acquisition function can utilize Expected Improvement (EI) to balance exploring unknown regions with utilizing known optimalities, thus enabling the selection of the next set of parameters to be evaluated.

[0099] Determine the optimal solution of the objective function:

[0100] Multi-objective optimization functions can be defined:

[0101] ;

[0102] Where θ represents the combustion parameters to be optimized, and ω i For the corresponding weights.

[0103] Bayesian optimization iterates within the parameter search space, selecting candidate parameter combinations through a data acquisition function, and then substituting these combinations into the engine model to evaluate the objective function value. The optimal parameter combinations and objective function values ​​for each iteration are recorded.

[0104] Update the dataset:

[0105] The newly evaluated parameter combinations and their corresponding combustion performance indices (objective function values) are added to the historical dataset, and the Bayesian surrogate model is retrained using the updated dataset to provide more accurate predictions for the next round of parameter selection.

[0106] Iterative judgment (reaching the iteration threshold):

[0107] The termination conditions are set: reaching the maximum number of iterations, the change in the objective function being less than a threshold, etc. If the conditions are not met, the process returns to the Bayesian-GBDT regression model to continue iteration. If the conditions are met, the process proceeds to the next step.

[0108] Returns the globally optimal parameters:

[0109] The optimal combination of parameters for the objective function is selected from all iteration records, and after verifying the physical and engineering feasibility of the parameter combination, it can be used as the final optimal combination of combustion parameters.

[0110] Training a combustion parameter prediction model:

[0111] A lightweight GBDT regression model is trained using the optimal parameter combination and corresponding operating condition data as the training set. The input consists of real-time operating conditions (speed, load, intake air temperature, etc.) plus preprocessed features (i.e., the aforementioned feature parameters). The output consists of combustion parameters such as the optimal firing angle and EGR rate, which control various actuators of the engine.

[0112] Combustion parameter control (ignition angle, EGR rate):

[0113] The electronically controlled nitrogen-oxygen system collects real-time characteristic data of the current operating conditions and pre-processed parameters, which are then input into the combustion parameter prediction model. The model outputs the optimal ignition angle and EGR rate, which are precisely controlled through actuators (ignition coil, EGR valve).

[0114] Finally, closed-loop feedback can be implemented: real-time monitoring of indicators such as knock, cyclical changes, and misfire rate. If the deviation from the target range is detected, a new round of parameter optimization iteration is triggered to update the control strategy in order to achieve the best control effect.

[0115] As a comparison of the specific application effects of the above-mentioned natural gas engine control methods, refer to Figure 4 As shown, compared to the previous natural gas control method, this application can control the cylinder pressure within the normal cylinder pressure range, maintain the combustion heat release rate, and keep the pressure rise rate within a safe and efficient range.

[0116] This natural gas engine control method infers the correspondence between combustion parameters and knock boundaries by observing the performance of knock vibration signals across all operating conditions. It also infers the correspondence between current combustion parameters and misfire boundaries by analyzing speed roughness, the waveform of the turbine exhaust pressure, and the discharge time of the secondary coil. Furthermore, it predicts the cycle variations of combustion parameters in each cylinder based on the turbine exhaust pressure parameters. Finally, based on the current intake air volume, EGR, and ignition angle in each cylinder, and combined with knock, misfire, and cycle variation parameters, the combustion parameter predictor is comprehensively adjusted. Combustion parameter regulation is then performed based on the real-time optimized combustion parameter predictor. Through advanced combustion predictive control technology, the combustion curve is controlled within a reasonable range, reducing fuel consumption in multiple scenarios.

[0117] The above describes a natural gas engine control method provided by the embodiments of this application. The following will describe the apparatus for performing the above natural gas engine control method.

[0118] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of a natural gas engine control device provided in an embodiment of this application. Figure 5 As shown, the natural gas engine control device includes:

[0119] Combustion feature extraction module 501 is used to extract features from the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine. The feature parameters include at least: knock characteristics, cycle variation index and misfire rate of each cylinder.

[0120] The combustion parameter determination module 502 is used to preprocess the characteristic parameters to obtain characteristic parameters with a uniform format, and input the characteristic data, including at least the characteristic parameters with the uniform format, into the combustion parameter prediction model for prediction processing. This allows the combustion parameter prediction model to determine the state values ​​of each combustion parameter based on the combustion performance indicators under the current operating conditions and the characteristic data.

[0121] The combustion parameter adjustment module 503 is used to adjust the current state of the actuator corresponding to each combustion parameter based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle, and the corresponding actuator includes: EGR valve and ignition coil.

[0122] In one possible implementation, the detonation features include detonation frequency and detonation intensity. The combustion feature extraction module 501 extracts features from the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine, including:

[0123] After bandpass filtering the acquired detonation signal, the frequency domain features were extracted using fast Fourier transform, and the detonation frequency was determined based on the peak detonation frequency.

[0124] The detonation intensity is determined based on the amplitude of the detonation frequency after bandpass filtering and the amplitude threshold.

[0125] In one possible implementation, the combustion feature extraction module 501 performs feature extraction on the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine, including:

[0126] Extract pressure index data for each engine cycle and obtain the inter-cycle variation coefficient based on the pressure index data;

[0127] The cyclic variation coefficient is fused with at least the vortex front pressure fluctuation to obtain the cyclic variation index.

[0128] In one possible implementation, the combustion feature extraction module 501 performs feature extraction on the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine, including:

[0129] The number of misfires per cylinder within the engine cycle is determined based on the turbine exhaust pressure, secondary discharge time, and speed roughness. The misfire rate of each cylinder is then determined based on the number of misfires and the engine cycle.

[0130] In one possible implementation, the process of constructing the combustion parameter prediction model in the combustion parameter determination module 502 includes:

[0131] Using feature data as input and combustion performance index data as the target, a gradient boosting decision tree regression model is constructed.

[0132] Based on the surrogate model, the mapping relationship between the objective function and the combustion parameters is fitted in the combustion parameter search space to obtain the target combustion parameter combination. The target combustion parameter combination is the combination of combustion parameters that optimizes the function value of the objective function obtained by the surrogate model during the iteration process. The objective function is the target optimization function composed of combustion performance indicators.

[0133] The gradient boosting decision tree regression model is trained based on the combination of combustion parameters for each target and the feature data to obtain the combustion parameter prediction model.

[0134] In one possible implementation, the combustion parameter determination module 502, based on a surrogate model, fits the mapping relationship between the objective function and the combustion parameters within the combustion parameter search space to obtain the target combination of combustion parameters, including:

[0135] Iterative processing is performed within the combustion parameter search space, and the optimal parameter combination and corresponding objective function value of each iteration are recorded. In each iteration, candidate parameter combinations are selected based on the acquisition function, and the candidate parameter combinations are substituted into the engine model to evaluate the objective function value.

[0136] The historical dataset is updated using the optimal parameter combination and corresponding objective function value from each iteration, and the proxy model is retrained based on the updated historical dataset until the iteration termination condition is met.

[0137] The combination of combustion parameters that optimizes the objective function value is selected from the surrogate model that terminates the iteration, and this combination is taken as the target combustion parameter combination.

[0138] In one possible implementation, the combustion parameter determination module 502 preprocesses the characteristic parameters to obtain characteristic parameters with a uniform format, including:

[0139] The feature parameters are sequentially processed by data cleaning, noise reduction, filtering, and data normalization to obtain feature parameters with a uniform format.

[0140] This application also provides an electronic device in its embodiments. (See reference...) Figure 6 The diagram illustrates a structural schematic suitable for implementing the electronic devices in the embodiments of this application. The electronic devices in the embodiments of this application may include, but are not limited to, ECU (Electronic Control Unit), VCU (Vehicle Control Unit), MCU (Micro Controller Unit), HCU (Hybrid Control Unit), etc. Figure 6 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0141] refer to Figure 6 As shown, the electronic device includes at least one processor 601 and a memory 602 connected to the processor 601, wherein: the memory is used to store computer programs; the processor 601 is used to execute the computer programs to enable the electronic device to implement the natural gas engine control method as described in the above embodiment.

[0142] This application also provides a computer program product including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the natural gas engine control methods provided in this application.

[0143] This application also provides a computer-readable storage medium carrying one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the natural gas engine control methods provided in this application.

[0144] This application also provides a natural gas engine, including the electronic equipment described in the above embodiments.

[0145] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0147] In the above embodiments, the implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, in the form of a computer program product.

[0148] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A natural gas engine control method, characterized in that, include: The operating parameters of the natural gas engine are feature extracted to obtain characteristic parameters that characterize the natural gas combustion state of the natural gas engine. The characteristic parameters include at least: knock characteristics, cycle variation index and misfire rate of each cylinder. The knock characteristics include: knock frequency and knock intensity; the cycle variation index is obtained by fusing the inter-cycle variation coefficient with at least the turbine exhaust pressure fluctuation, the inter-cycle variation coefficient is obtained by extracting pressure index data in each engine cycle, the pressure index data includes: cylinder pressure peak and indicated mean effective pressure; The feature parameters are preprocessed to obtain feature parameters with a uniform format. Feature data, including at least the uniformly formatted feature parameters, is then input into a combustion parameter prediction model for prediction processing. This model, using the combustion performance indicators under the current operating conditions as the target and the feature data as a prerequisite, determines the state values ​​of each combustion parameter. The combustion performance indicators include any one of thermal efficiency, emissions, and gas consumption. The construction process of the combustion parameter prediction model includes: Using the aforementioned feature data as input and combustion performance index data as the target, a gradient boosting decision tree regression model is constructed. Based on the Bayesian optimization model, the mapping relationship between the objective function and the combustion parameters is fitted in the combustion parameter search space to obtain the target combustion parameter combination. The target combustion parameter combination is the combination of combustion parameters that makes the function value of the objective function optimal, obtained by the Bayesian optimization model during the iteration process. The objective function is the target optimization function composed of the combustion performance index. The gradient boosting decision tree regression model is trained based on the combination of target combustion parameters and the feature data to obtain the combustion parameter prediction model. The current state of the actuator corresponding to each combustion parameter is adjusted based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle. The actuators include: EGR valve and ignition coil.

2. The natural gas engine control method according to claim 1, characterized in that, The feature extraction of the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine includes: After bandpass filtering the acquired detonation signal, the frequency domain features are extracted using fast Fourier transform, and the detonation frequency is determined based on the detonation peak frequency. The detonation intensity is determined based on the amplitude of the detonation frequency after bandpass filtering and the amplitude threshold.

3. The natural gas engine control method according to claim 1, characterized in that, The feature extraction of the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine includes: The number of misfires of each cylinder within the engine cycle is determined based on the turbine exhaust pressure, secondary discharge time, and rotational roughness, and the misfire rate of each cylinder is determined based on the number of misfires and the engine cycle.

4. The natural gas engine control method according to claim 1, characterized in that, The method of fitting the mapping relationship between the objective function and combustion parameters in the combustion parameter search space based on the Bayesian optimization model to obtain the target combustion parameter combination includes: Iterative processing is performed within the combustion parameter search space, and the optimal parameter combination and corresponding objective function value of each iteration are recorded. In each iteration, candidate parameter combinations are selected based on the acquisition function, and the candidate parameter combinations are substituted into the engine model to evaluate the objective function value. The historical dataset is updated using the optimal parameter combination and corresponding objective function value from each iteration, and the Bayesian optimization model is retrained based on the updated historical dataset until the iteration termination condition is met. The combination of combustion parameters that optimizes the objective function value is selected from the terminated Bayesian optimization model and is taken as the target combustion parameter combination.

5. The natural gas engine control method according to claim 1, characterized in that, The preprocessing of the feature parameters to obtain feature parameters with a uniform format includes: The feature parameters are sequentially subjected to data cleaning, noise reduction, filtering, and data normalization to obtain the feature parameters with a unified format.

6. A natural gas engine control device, characterized in that, include: A combustion feature extraction module is used to extract features from the operating parameters of the natural gas engine to obtain feature parameters characterizing the natural gas combustion state of the natural gas engine. The feature parameters include at least: knock characteristics, cycle variation index, and misfire rate of each cylinder. The knock characteristics include: knock frequency and knock intensity. The cycle variation index is obtained by fusing the inter-cycle variation coefficient with at least the turbine exhaust pressure fluctuation. The inter-cycle variation coefficient is obtained based on the pressure index data extracted in each engine cycle. The pressure index data includes: peak cylinder pressure and indicated mean effective pressure. A combustion parameter determination module is used to preprocess the feature parameters to obtain feature parameters with a uniform format, and input feature data, including at least the feature parameters with the uniform format, into a combustion parameter prediction model for prediction processing. The combustion parameter prediction model, using the combustion performance indicators under the current operating conditions as the target and the feature data as a prerequisite, determines the state values ​​of each combustion parameter. The combustion performance indicators include any one of thermal efficiency, emissions, and gas consumption. The construction process of the combustion parameter prediction model includes: Using the aforementioned feature data as input and combustion performance index data as the target, a gradient boosting decision tree regression model is constructed. Based on the Bayesian optimization model, the mapping relationship between the objective function and the combustion parameters is fitted in the combustion parameter search space to obtain the target combustion parameter combination. The target combustion parameter combination is the combination of combustion parameters that makes the function value of the objective function optimal, obtained by the Bayesian optimization model during the iteration process. The objective function is the target optimization function composed of the combustion performance index. The gradient boosting decision tree regression model is trained based on the aforementioned combinations of target combustion parameters and the aforementioned feature data to obtain the combustion parameter prediction model; and... The combustion parameter adjustment module is used to adjust the current state of the actuator corresponding to each combustion parameter based on the state value. Each combustion parameter includes at least: EGR rate and ignition angle, and the actuator includes: EGR valve and ignition coil.

7. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the electronic device to implement the natural gas engine control method as described in any one of claims 1 to 5.

8. A natural gas engine, characterized in that, include: The electronic device as claimed in claim 7.

Images