Calibration parameter optimization method and device of hybrid electric vehicle and computer program product

Abstract

The invention relates to the technical field of hybrid electric vehicles, in particular to a calibration parameter optimization method and device of a hybrid electric vehicle and a computer program product. Comprising the following steps: acquiring power assembly operation data, and performing dynamic outlier processing on the power assembly operation data; inputting the power assembly operation data after the dynamic outlier processing into a neural network model obtained by pre-training to obtain torque data; working condition distribution calculation is conducted on the vehicle speed, the accelerator pedal opening degree, the rotating speed and the torque data through a Gaussian mixture model, working condition interval probability distribution is obtained, and a working condition interval is determined; and peak value calibration parameter optimization is conducted on the vehicle speed, the accelerator pedal opening degree, the rotating speed and the torque data, working condition interval boundary optimization is conducted on the working condition interval, and optimized calibration parameter data are obtained. High-reliability power assembly calibration parameter data can be provided, and then comprehensive optimization and improvement of vehicle performance and cost are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of hybrid electric vehicles, and in particular to a method, device and computer program product for optimizing calibration parameters of a hybrid electric vehicle. Background Art

[0002] The rapid development of the automotive industry, particularly the continuous advancement of hybrid vehicle technology, has placed higher demands on the intelligent management and optimization of vehicle powertrains. Hybrid vehicles combine traditional internal combustion engines with electric drive systems to achieve higher fuel economy and lower emissions. However, the complexity and variability of powertrains also pose challenges to optimizing vehicle performance.

[0003] Chinese patent application CN119682731A discloses a vehicle operating condition identification method and device. The method obtains multiple characteristic parameters within a preset calculation period; for each characteristic parameter, determines the operating condition corresponding to at least one characteristic parameter interval within which the characteristic parameter falls, and outputs a sub-score for the characteristic parameter under the condition; for each operating condition, determines the characteristic parameter score of the vehicle for that operating condition within the preset calculation period based on the sub-scores of all characteristic parameters under that condition; and, based on the characteristic parameter score, determines the target operating condition and engine start-stop strategy for the vehicle within the preset calculation period. However, this solution struggles to accurately identify transition states in complex operating conditions, and thus the reliability of its powertrain calibration parameter data needs to be improved. Summary of the Invention

[0004] In view of this, the embodiments of the present application provide a method, device and computer program product for optimizing the calibration parameters of a hybrid electric vehicle, which can accurately identify the transition state of complex working conditions, realize calibration parameter optimization, and provide highly reliable powertrain calibration parameter data, thereby achieving comprehensive optimization and improvement of vehicle performance and cost.

[0005] A first aspect of an embodiment of the present application provides a method for optimizing calibration parameters of a hybrid electric vehicle, comprising:

[0006] Acquiring powertrain operating data and performing dynamic outlier processing on the powertrain operating data, wherein the powertrain operating data includes vehicle speed, accelerator pedal opening, and speed data;

[0007] The powertrain operation data after dynamic outlier processing is input into the pre-trained neural network model to obtain torque data;

[0008] Calculating the operating condition distribution of the vehicle speed, accelerator pedal opening, speed, and torque data using a Gaussian mixture model to obtain an operating condition interval probability distribution to determine the operating condition interval;

[0009] Peak calibration parameter optimization is performed on the vehicle speed, accelerator pedal opening, rotational speed and torque data, and operating condition interval boundary optimization is performed on the operating condition interval to obtain optimized parameter data.

[0010] In one embodiment, the performing dynamic outlier processing on the powertrain operating data includes:

[0011] Calculate the mean μ and standard deviation σ of the powertrain operation data using sliding window statistics;

[0012] Performing outlier determination on the vehicle speed, accelerator pedal opening and speed data using an outlier interval [μ-3σ, μ+3σ];

[0013] The outliers are removed and data gaps are compensated.

[0014] In one embodiment, it further includes:

[0015] When the accelerator pedal opening is greater than the preset opening and the gear is in sport mode, the outlier interval is relaxed to [μ-5σ, μ+5σ].

[0016] In one embodiment, it further includes:

[0017] Constructing a neural network model framework, and inputting vehicle speed, accelerator pedal opening, and speed data of the test bench to train the neural network model framework to obtain a neural network model;

[0018] Among them, the neural network model architecture adopts a three-layer fully connected network.

[0019] In one embodiment, the Gaussian mixture model uses K Gaussian distributions to weightedly superimpose to describe the multidimensional feature space. The probability density function of the Gaussian mixture model is:

[0020] ;

[0021] in, is the mixing coefficient, is the mean vector, Calibrate parameters for the covariance matrix.

[0022] In one embodiment, determining the operating condition range includes:

[0023] Determine satisfaction The area is the normal working area;

[0024] Determine satisfaction The area is the general working condition area;

[0025] Determine satisfaction It is an extreme working condition area;

[0026] in, is the probability distribution of the operating condition interval, is the preset high probability boundary, is the preset low probability boundary.

[0027] In one embodiment, the peak calibration parameter optimization of the vehicle speed, accelerator pedal opening, rotational speed, and torque data includes:

[0028] By formula Calculate the first reduction factor ;

[0029] When the first reduction factor When it is less than the preset value, the formula Iterating the basic calibration value, the preset value is less than 1;

[0030] in, is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the first design margin, >1.

[0031] In one embodiment, the optimizing the operating condition interval boundary of the operating condition interval includes:

[0032] For the normal working area, the formula calculate ;

[0033] For general working conditions, ;

[0034] For extreme working conditions, the formula calculate ;

[0035] in, is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second design margin, Less than 100%; is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second reduction coefficient, 0< <1.

[0036] A second aspect of an embodiment of the present application provides a calibration parameter optimization device for a hybrid electric vehicle, comprising:

[0037] an outlier processing unit, configured to obtain powertrain operating data and perform dynamic outlier processing on the powertrain operating data, wherein the powertrain operating data includes vehicle speed, accelerator pedal opening, and speed data;

[0038] A neural network unit is used to input the powertrain operation data after dynamic outlier processing into a pre-trained neural network model to obtain torque data;

[0039] a working condition distribution calculation unit, configured to calculate the working condition distribution of the vehicle speed, accelerator pedal opening, rotational speed, and torque data using a Gaussian mixture model to obtain a working condition interval probability distribution to determine the working condition interval;

[0040] The optimization unit is used to optimize the peak calibration parameters of the vehicle speed, accelerator pedal opening, rotational speed and torque data, and to optimize the operating range boundary of the operating range to obtain optimized calibration parameter data.

[0041] A third aspect of an embodiment of the present application provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the electronic device implements the calibration parameter optimization method for a hybrid vehicle provided in the first aspect of the embodiment of the present application.

[0042] A fourth aspect of the embodiments of the present application provides a computer program product, including a computer program. When the computer program is executed, the method according to the first aspect of the embodiments of the present application is executed.

[0043] A first aspect of an embodiment of the present application provides a method for optimizing calibration parameters for a hybrid electric vehicle. The method comprises acquiring powertrain operating data and performing dynamic outlier processing on the powertrain operating data, wherein the powertrain operating data includes vehicle speed, accelerator pedal opening, and speed data. The method then inputs the processed powertrain operating data into a pre-trained neural network model to obtain torque data. A Gaussian mixture model is used to calculate the operating condition distribution of the speed, accelerator pedal opening, speed, and torque data to obtain a probability distribution of the operating condition intervals, thereby determining the operating condition intervals. Peak calibration parameter optimization is performed on the speed, accelerator pedal opening, speed, and torque data, and operating condition interval boundaries are optimized for the operating condition intervals to obtain optimized calibration parameter data. An efficient closed-loop technology is formed through a framework of data cleaning, torque prediction, operating condition identification, and step-by-step optimization. Dynamic outlier processing is used as a pre-step. Through the dynamic outlier processing technology, outliers in the powertrain operating data can be detected and cleaned in real time, ensuring the reliability and stability of the input data and providing a high-quality data foundation for subsequent analysis. Using neural networks to construct a high-precision nonlinear mapping model of the relationship between multiple parameters and torque during vehicle operation accurately predicts the vehicle's torque output under different operating conditions, providing a scientific basis for operating condition distribution calculation and calibration optimization decisions. Using a Gaussian mixture model to quantify the probability density distribution of multidimensional feature space, the probability distribution of the actual operating condition interval can be accurately identified, and calibration parameters can be optimized by comparing it with the design values. This process not only reduces hardware costs but also significantly improves energy consumption.

[0044] It can be understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 This is a flow chart of a method for optimizing calibration parameters of a hybrid electric vehicle provided in one embodiment of the present application;

[0047] Figure 2 This is a flow chart of a method for optimizing calibration parameters of a hybrid vehicle provided by another embodiment of the present application;

[0048] Figure 3 This is a flow chart of a method for optimizing calibration parameters of a hybrid vehicle provided by another embodiment of the present application;

[0049] Figure 4 is a structural schematic diagram of a hybrid vehicle calibration parameter optimization device provided by an embodiment of the present application.

[0050] Figure 5 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0051] In the following description, for the purposes of explanation and not limitation, specific details are set forth, such as particular sequences of steps, techniques, etc., in order to provide a thorough understanding of the embodiments of the application. However, it will be apparent to those skilled in the art that the embodiments of the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0052] It should be understood that the term "includes" when used in the specification and the appended claims herein, specifies the presence of stated features, integers, steps, operations, elements, and / or components but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0053] It should also be understood that the term "and / or" when used in the specification and the appended claims herein, means any one or more of the associated listed items or a combination thereof. It should be understood that the term "if can be interpreted as meaning "when" or "upon" or "in response to determining" or "in response to detecting," as the context dictates. Similarly, the phrase "if determined" or "if detected [the recited condition or event]" can be interpreted as meaning "upon determining" or "in response to determining" or "upon detecting [the recited condition or event]" or "in response to detecting [the recited condition or event]," as the context dictates.

[0054] In addition, in the description of the specification and the appended claims herein, the terms "first," "second," "third," etc. are used only to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0055] In addition, in the description of the specification and the appended claims herein, the terms "first," "second," "third," etc. are used only to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0056] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0057] like Figure 1 As shown, the method for optimizing calibration parameters of a hybrid electric vehicle provided in an embodiment of the present application includes the following steps S101 to S104:

[0058] Step S101: Acquire powertrain operation data and perform dynamic outlier processing on the powertrain operation data, wherein the powertrain operation data includes vehicle speed, accelerator pedal opening, and speed data.

[0059] In applications, dynamic outlier processing is used to detect and clean outliers in powertrain operating data in real time, ensuring the reliability and stability of input data.

[0060] Step S102: Input the powertrain operation data after dynamic outlier processing into the pre-trained neural network model to obtain torque data.

[0061] In the application, a high-precision nonlinear mapping model of the relationship between multiple parameters and torque during vehicle operation is constructed based on a neural network. Through this neural network model, the real-time torque output of the vehicle can be accurately obtained.

[0062] Step S103: Calculate the operating condition distribution of the vehicle speed, accelerator pedal opening, speed, and torque data using a Gaussian mixture model to obtain a probability distribution of the operating condition interval to determine the operating condition interval.

[0063] In the application, the probability density distribution of the multidimensional feature space is quantified through the Gaussian mixture model to identify the probability distribution of the actual operating condition interval.

[0064] Step S104 : Optimizing peak calibration parameters of vehicle speed, accelerator pedal opening, rotational speed, and torque data, and optimizing operating range boundaries of the operating range to obtain optimized calibration parameter data.

[0065] In application, by comparing the actual working condition distribution with the design value, calibration parameters are optimized to achieve hardware cost reduction and energy consumption optimization.

[0066] The embodiments of the present application form a closed-loop processing flow by integrating data cleaning, torque prediction, working condition recognition and parameter optimization, which significantly improves the response accuracy of the hybrid power system. Dynamic outlier processing eliminates sensor noise interference, neural network model realizes high-precision nonlinear mapping of multiple parameters and torque, Gaussian mixture model quantifies working condition distribution characteristics, and partition optimization strategy synchronously reduces hardware cost and energy consumption. The synergistic effect of each step breaks through the static limitations of traditional calibration methods.

[0067] The embodiments of the present application optimize calibration parameters through peak parameter optimization and working condition interval partition optimization strategy based on real vehicle reported data cleaning analysis, finally realize the reduction of overall hardware cost, optimize energy consumption performance, provide data support and decision basis for calibration parameter optimization of hybrid vehicles. Through dynamic outlier processing technology on real vehicle data, high-quality data is ensured for subsequent analysis, then through the trained neural network multi-parameter and torque mapping model, the real-time torque output of the vehicle can be accurately obtained, then through the Gaussian mixture model quantifying the probability density distribution of the multi-dimensional feature space, the real working condition interval probability distribution can be accurately recognized, finally compared with the design value, through the peak parameter optimization and working condition interval partition optimization strategy, the calibration parameter optimization is carried out, finally the overall hardware cost is reduced, and the energy consumption performance is significantly optimized.

[0068] In one embodiment, a neural network model is also trained to input feature data not containing torque, i.e. powertrain operation data including vehicle speed, accelerator pedal opening degree and speed data, to obtain torque data, and the complete process of this step is shown in Figure 2 . The training of the neural network model specifically includes:

[0069] The neural network model architecture is constructed, and the vehicle speed, accelerator pedal opening degree and speed data of the test bench are input to train the neural network model architecture to obtain the neural network model.

[0070] The neural network model architecture adopts a three-layer fully connected network.

[0071] In application, the hidden layer structure of the neural network model architecture adopts a three-layer fully connected network (neuron number 32-64-32), and the activation function selects ReLU (Rectified Linear Unit) with the formula: The loss function adopts mean absolute error (MAE) to measure the deviation of the predicted value and the actual torque, with the formula: Wherein, is the predicted value, is the actual value, The Adam optimizer was used for optimization, with an initial learning rate of 0.001 and a dynamically adjusted weight update step size to balance convergence speed and stability. The output layer mapping used a data partitioning rule, with training, validation, and test sets divided into a 7:2:1 ratio to ensure model generalization. The batch size of the batch training parameter was set to 64, with 500 iterations. Early stopping was triggered when the validation set error fell below a threshold (e.g., MAE < 5 Nm) after each training round. After constructing the neural network model architecture, bench data, including vehicle speed v, accelerator pedal opening a, and speed n, were input and normalized: the raw values ​​were mapped to the [0, 1] range (e.g., vehicle speed 0-220 km / h was mapped to [0, 1]). The model was trained using the input data, ultimately obtaining a neural network model. Torque data was obtained by inputting the powertrain operating data, which had been processed for dynamic outliers, from step S101, into the model.

[0072] This embodiment of the application constructs a highly robust torque mapping model using a three-layer fully connected neural network architecture trained with bench test data. The ReLU activation function accelerates gradient propagation, the mean absolute error loss function reduces interference from outliers, and an early stopping mechanism prevents overfitting. This model is capable of processing heterogeneous data from multiple sources in real time, accurately reflecting the nonlinear characteristics of the power system.

[0073] In one embodiment, Figure 3 As shown, the vehicle data stream (i.e., powertrain operation data) is input into the sliding window for statistical calculation to determine whether it is an outlier, and the outliers are removed, and then linear interpolation compensation is performed to obtain the output. Step S101 includes:

[0074] Step S1011 : Calculate the mean μ and standard deviation σ of the powertrain operation data using a sliding window.

[0075] In the application, the window period of the sliding window is set to 120 seconds, and the data in the window is strictly aligned by timestamp to avoid sensor delay deviation. The statistical calculation uses the recursive averaging method to update the mean of the data in the window. The formula is:

[0076] ;

[0077] Based on sliding variance iterative update, the formula is: ;

[0078] The final output is the mean updated in real time and standard deviation .

[0079] Step S1012: Perform outlier determination on the vehicle speed, accelerator pedal opening, and speed data using the outlier interval [μ-3σ, μ+3σ].

[0080] In the application, based on the mean value output in step S1011 and standard deviation Perform outlier determination. The determination rules are as follows:

[0081] a. Dynamic interval judgment: Based on the 3σ criterion, the outlier interval is set to [μ-3σ, μ+3σ], covering 99.73% of normal working condition data

[0082] b. Multi-parameter collaborative judgment: The judgment intervals are calculated for the four parameters of vehicle speed, accelerator pedal opening, and speed. If any parameter exceeds the limit, it will be marked as an outlier (e.g., acceleration > 3σ).

[0083] c. Context-aware filtering:

[0084] Combined with working condition calibration: When the accelerator pedal opening is greater than 90% and the gear is in sport mode, the outlier interval is relaxed to [μ-5σ, μ+5σ].

[0085] Timing consistency check: A single point mutation is an abnormality, and three consecutive points exceeding the threshold trigger a special operating condition mark.

[0086] Finally, the marked outliers are output (including timestamp, parameter type, deviation value Δ, such as pedal opening Δa=+3.2σ).

[0087] Step S1013: Eliminate outliers and compensate for data gaps.

[0088] In the application, the data set after the outliers are input and the outliers are removed and their abnormal information is recorded. Then, continuity compensation is performed by filling the data gaps with linear interpolation. The formula is:

[0089] ;

[0090] in, is the compensation value, is the previous value, is the later value, is the current time, For the previous time, The cleaned continuous time series data stream is used as the input of the neural network model.

[0091] This embodiment of the application establishes a dynamic data cleaning mechanism based on real-time statistical calculations using a sliding window combined with the 3σ criterion. Recursive updates of the mean and standard deviation adapt to the temporal characteristics of the data stream, collaborative multi-parameter determination prevents the propagation of single-point errors, and linear interpolation compensation ensures data continuity. This solution maintains input data reliability in complex driving environments, providing a stable foundation for subsequent modeling.

[0092] In one embodiment, it further includes:

[0093] When the accelerator pedal opening is greater than the preset opening and the gear is in sport mode, the outlier interval is relaxed to [μ-5σ, μ+5σ].

[0094] This embodiment of the application designs an adaptive judgment strategy for extreme driving conditions in Sport mode. By linking gear state recognition with the accelerator pedal opening threshold, it dynamically expands the outlier range to a 5σ range. This design retains valid data from aggressive driving conditions, avoids power response lag caused by misjudgments, and enhances the system's adaptability to different driving styles.

[0095] In one embodiment, the Gaussian mixture model uses K Gaussian distributions to weightedly superimpose to describe the multidimensional feature space. The probability density function of the Gaussian mixture model is:

[0096] ;

[0097] in, is the mixing coefficient, is the mean vector, Calibrate parameters for the covariance matrix.

[0098] In the application, step S103 specifically includes multi-dimensional feature space construction and Gaussian mixture model calculation.

[0099] The multidimensional feature space construction includes:

[0100] The input features include four parameters: vehicle speed v, accelerator pedal opening a, speed n, and torque T. Torque T is the output of the neural network by inputting vehicle speed v, accelerator pedal opening a, and speed n into the neural network.

[0101] Feature normalization is performed to perform linear mapping and eliminate dimensional differences.

[0102] Gaussian mixture model calculations include:

[0103] The multidimensional feature space is described by weighted superposition of K Gaussian distributions, and the probability distribution of the operating condition interval is calculated by the following probability density function:

[0104]

[0105] in is the mixing coefficient, is the mean vector, is the covariance matrix.

[0106] In applications, parameter optimization can also be performed using the Expectation Maximization (EM) algorithm and / or Gaussian component selection. The EM algorithm iteratively solves for parameters until the log-likelihood function converges (rate of change < 1e-5). Gaussian component selection automatically determines the optimal K value (typically 5-8) based on the Bayesian Information Criterion (BIC), balancing model complexity with the risk of overfitting.

[0107] This embodiment uses a Gaussian mixture model to quantify the multidimensional feature space distribution using a probability density function. The weighted superposition of K Gaussian components accurately characterizes the operating condition transition state. The mixing coefficient dynamically adjusts the weights of each distribution, and the covariance matrix captures the correlation characteristics between parameters, providing a rigorous mathematical foundation for calculating the operating condition probability distribution.

[0108] In one embodiment, determining the operating condition range includes:

[0109] Determine satisfaction The area is the normal working area;

[0110] Determine satisfaction The area is the general working area;

[0111] Determine satisfaction It is an extreme working condition area;

[0112] in, is the probability distribution of the operating condition interval, is the preset high probability boundary, is the preset low probability boundary.

[0113] In applications, set high probability boundaries , low probability boundary .

[0114] The present embodiment achieves accurate scene recognition through a three-level operating condition partitioning mechanism based on probability density. This hierarchical strategy provides a decision basis for differentiated optimization, balancing system stability and energy efficiency.

[0115] In one embodiment, peak calibration parameter optimization is performed on vehicle speed, accelerator pedal opening, rotational speed, and torque data, including:

[0116] By formula Calculate the first reduction factor ;

[0117] When the first reduction factor When it is less than the preset value, the formula Iterate the basic calibration value, and the preset value is less than 1;

[0118] in, is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the first design margin, >1.

[0119] In application, the preset value can be 0.9, the first design margin 1.1 is recommended.

[0120] The embodiment of the present application establishes a quantitative relationship between the actual peak value and the calibration value by reducing the coefficient γ, and a preset threshold triggers a dynamic iteration mechanism. The design margin k ensures performance redundancy under normal working conditions, and is calculated by the formula Achieve closed-loop self-optimization of calibration parameters to effectively avoid waste of hardware resources.

[0121] In one embodiment, optimizing the operating condition interval boundary includes:

[0122] For the normal working area, the formula calculate ;

[0123] For general working conditions, ;

[0124] For extreme working conditions, the formula calculate ;

[0125] in, is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second design margin, Less than 100%; is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second reduction coefficient, 0< <1.

[0126] In application, a 10% design margin is reserved in the normal operating area to ensure performance stability, which is the second design margin. Take 1.1. Second design margin With the first design margin The values ​​can be the same or different, depending on the actual margin requirements. For extreme operating conditions, the upper limit of the parameter is reduced by 20% to avoid hardware overload and save energy, which is the second reduction coefficient. Take 0.8.

[0127] This embodiment utilizes a partitioned optimization strategy to implement differentiated adjustments for different operating conditions: retaining a safety margin for normal operating conditions, matching general operating conditions to actual requirements, and introducing a reduction factor μ to control hardware load for extreme conditions. This step-by-step parameter adjustment model simultaneously achieves energy optimization and system protection goals, extending the lifespan of key components.

[0128] This application ultimately achieved hardware cost reduction and energy consumption optimization through the above-mentioned calibration optimization decisions (such as adjusting the torque limit value from 500N.m to 448N.m, and reducing the power consumption per 100 kilometers by 5.6% through derating in extreme working conditions).

[0129] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0130] The present application also provides a hybrid electric vehicle calibration parameter optimization device for executing the steps of the hybrid electric vehicle calibration parameter optimization method described above. The hybrid electric vehicle calibration parameter optimization device can be a virtual appliance within an electronic device, executed by a processor within the electronic device, or it can be the electronic device itself.

[0131] like Figure 4 As shown, the hybrid vehicle calibration parameter optimization device 100 provided in the embodiment of the present application includes:

[0132] An outlier processing unit 101 is used to obtain powertrain operation data and perform dynamic outlier processing on the powertrain operation data, wherein the powertrain operation data includes vehicle speed, accelerator pedal opening and speed data;

[0133] A neural network unit 102 is configured to input the powertrain operation data after dynamic outlier processing into a pre-trained neural network model to obtain torque data;

[0134] a working condition distribution calculation unit 103 for performing working condition distribution calculation on the vehicle speed, accelerator pedal opening, speed, and torque data using a Gaussian mixture model to obtain a working condition interval probability distribution to determine the working condition interval;

[0135] The optimization unit 104 is configured to optimize the peak calibration parameters of the vehicle speed, accelerator pedal opening, rotational speed, and torque data, and optimize the operating range boundary of the operating range to obtain optimized calibration parameter data.

[0136] In application, each module in the calibration parameter optimization device for a hybrid electric vehicle may be a software program module, or may be implemented by different logic circuits integrated in a processor, or may be implemented by multiple distributed processors.

[0137] like Figure 5 As shown, the embodiment of the present application further provides an electronic device 200, including: at least one processor 201 ( Figure 5 Only one processor is shown in the figure), a memory 202, and a computer program 203 stored in the memory 202 and executable on at least one processor 201. When the processor 201 executes the computer program 203, the steps in the above-mentioned method embodiments are implemented.

[0138] In applications, electronic devices may include, but are not limited to, processors and memories. Those skilled in the art will appreciate that Figure 5 The electronic device is merely an example and does not limit the electronic device. The electronic device may include more or fewer components than shown in the figure, or may include a combination of certain components or different components.

[0139] In applications, a processor may be a central processing unit (CPU), other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0140] In applications, in some embodiments, memory can be an internal storage unit of an electronic device, such as a hard drive or memory. In other embodiments, memory can also be an external storage device of the electronic device, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, or a flash memory card. Furthermore, memory can include both internal storage units and external storage devices. Memory is used to store operating systems, application programs, boot loaders, data, and other programs, such as computer program code. Memory can also be used to temporarily store data that has been output or is about to be output.

[0141] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment section and will not be repeated here.

[0142] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0143] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0144] An embodiment of the present application provides a computer program product, including a computer program. When the computer program product runs on an electronic device, the electronic device can implement the steps in the above-mentioned various method embodiments when executing the computer program product.

[0145] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms. The computer readable medium at least includes any entity or device capable of carrying the computer program code to the device / equipment, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunication signal.

[0146] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0147] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0148] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the above-described device embodiments are only schematic, and the division of the modules or units is only a logical function division, and there can be another division in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed mutual elements can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0149] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may also be distributed to multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment scheme according to actual needs.

[0150] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for optimizing calibration parameters of a hybrid electric vehicle, characterized in that: include: Acquiring powertrain operating data and performing dynamic outlier processing on the powertrain operating data, wherein the powertrain operating data includes vehicle speed, accelerator pedal opening, and speed data; The powertrain operation data after dynamic outlier processing is input into the pre-trained neural network model to obtain torque data; Calculating the operating condition distribution of the vehicle speed, accelerator pedal opening, speed, and torque data using a Gaussian mixture model to obtain an operating condition interval probability distribution to determine the operating condition interval; Peak calibration parameter optimization is performed on the vehicle speed, accelerator pedal opening, rotational speed and torque data, and operating condition interval boundary optimization is performed on the operating condition interval to obtain optimized calibration parameter data.

2. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: The dynamic outlier processing of the powertrain operation data includes: Calculate the mean μ and standard deviation σ of the powertrain operation data using sliding window statistics; Performing outlier determination on the vehicle speed, accelerator pedal opening and speed data using an outlier interval [μ-3σ, μ+3σ]; The outliers are removed and data gaps are compensated.

3. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 2, wherein: Also includes: When the accelerator pedal opening is greater than the preset opening and the gear is in sport mode, the outlier interval is relaxed to [μ-5σ, μ+5σ].

4. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: Also includes: Constructing a neural network model framework, and inputting vehicle speed, accelerator pedal opening, and speed data of the test bench to train the neural network model framework to obtain a neural network model; Among them, the neural network model architecture adopts a three-layer fully connected network.

5. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: The Gaussian mixture model uses K Gaussian distributions to weightedly superimpose and describe the multidimensional feature space. The probability density function of the Gaussian mixture model is: ； in, is the mixing coefficient, is the mean vector, Calibrate parameters for the covariance matrix.

6. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: Determining the operating condition range includes: Determine satisfaction The area is the normal working area; Determine satisfaction The area is the general working condition area; Determine satisfaction It is an extreme working condition area; in, is the probability distribution of the operating condition interval, is the preset high probability boundary, is the preset low probability boundary.

7. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: The peak calibration parameter optimization of the vehicle speed, accelerator pedal opening, rotational speed and torque data includes: By formula Calculate the first reduction factor ; When the first reduction factor When it is less than the preset value, the formula Iterating the basic calibration value, the preset value is less than 1; in, is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the first design margin, >

1.

8. The method for optimizing calibration parameters of a hybrid electric vehicle according to claim 1, wherein: Optimizing the operating condition interval boundary of the operating condition interval includes: For the normal working area, the formula calculate ; For general working conditions, ; For extreme working conditions, the formula calculate ; in, is the basic calibration value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second design margin, Less than 100%; is the actual peak value of any one of the vehicle speed, accelerator pedal opening, speed and torque data. is the second reduction coefficient, 0< <1.

9. A calibration parameter optimization device for a hybrid electric vehicle, characterized in that: include: an outlier processing unit, configured to obtain powertrain operating data and perform dynamic outlier processing on the powertrain operating data, wherein the powertrain operating data includes vehicle speed, accelerator pedal opening, and speed data; A neural network unit is used to input the powertrain operation data after dynamic outlier processing into a pre-trained neural network model to obtain torque data; a working condition distribution calculation unit, configured to calculate the working condition distribution of the vehicle speed, accelerator pedal opening, rotational speed, and torque data using a Gaussian mixture model to obtain a working condition interval probability distribution to determine the working condition interval; The optimization unit is used to optimize the peak calibration parameters of the vehicle speed, accelerator pedal opening, rotational speed and torque data, and to optimize the operating range boundary of the operating range to obtain optimized calibration parameter data.

10. A computer program product, characterized in that The invention comprises a computer program, which, when executed, causes the method according to any one of claims 1 to 8 to be performed.

Images