Calibration method, device and equipment of carbon loading capacity calculation model, medium and product

The calibration of the GPF carbon load model is optimized through the particle swarm algorithm, and the regenerated measured data and actual weighing data are used to solve the problems of long manual calibration time and low accuracy in the existing technology, achieving efficient model calibration and accuracy improvement.

CN118934288APending Publication Date: 2024-11-12GUANGZHOU AUTOMOBILE GROUP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411078037.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the calibration of the GPF carbon load model requires a lot of manual debugging, which leads to a long time and the global optimal solution cannot be found, resulting in low model accuracy.

Method used

The particle swarm algorithm is used to combine the regeneration measured data and actual weighing data to automatically optimize the parameters to be calibrated in the carbon load calculation model, and the model calibration is performed through the calibration data files of the target virtual space and the electronic control unit.

Benefits of technology

It greatly improves the model accuracy and calibration quality, saves labor costs, shortens the development cycle, and avoids the blindness of manual calibration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118934288A_ABST
    Figure CN118934288A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of engines, in particular to a carbon loading capacity calculation model calibration method, device and equipment, a medium and a product.The method comprises the steps that regeneration actual measurement data of a particulate filter and actual weighing data of carbon loading capacity are obtained; according to the regenerated actual measurement data and the actual weighing data, parameters to be calibrated of a carbon load calculation model in the target virtual space are determined, and the carbon load calculation model in the target virtual space is determined based on a calibration data file of the electronic control unit; and setting a target parameter of the particle swarm algorithm, calibrating the to-be-calibrated parameter based on the particle swarm algorithm to obtain target calibration data, and calibrating the carbon load calculation model in the electronic control unit based on the target calibration data. Therefore, the problems of low model precision and the like caused by long time spent on manual calibration and debugging and incapability of finding a global optimal solution in related technologies are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of engine technology, and in particular to a calibration method, device, equipment, medium and product for a carbon load calculation model. Background Art

[0002] As global environmental regulations become increasingly stringent, especially restrictions on particulate matter in automobile exhaust emissions, GPF (Gasoline Particulate Filter) needs to choose the right time to regenerate it, to avoid too low a regeneration frequency, which causes engine performance degradation, and too high a temperature during regeneration, which damages the carrier, and to avoid too high a regeneration frequency, which causes increased fuel consumption and worse emissions. Therefore, it is necessary to accurately estimate the real-time carbon load in the GPF.

[0003] In the related technology, when calibrating the GPF carbon load model through road tests, dozens of regeneration / accumulation tests are required to accurately weigh the actual carbon load changes. However, since the regeneration and accumulation conditions overlap, a large amount of calibration data is involved, requiring a lot of manual debugging time and experience to meet the model accuracy under each experimental data; secondly, relying solely on the experience of engineers can only obtain feasible solutions, and no further optimization is possible. Summary of the invention

[0004] The present application provides a calibration method, device, equipment, medium and product for a carbon loading calculation model to solve the problems in the related art that manual calibration and debugging takes a long time and cannot find a global optimal solution, resulting in low model accuracy.

[0005] A first aspect embodiment of the present application provides a calibration method for a carbon load calculation model, comprising the following steps: obtaining measured regeneration data of a particulate filter and actual weighing data of the carbon load; determining parameters to be calibrated of the carbon load calculation model in a target virtual space based on the measured regeneration data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on a calibration data file of an electronic control unit; setting target parameters of a particle swarm algorithm, calibrating the parameters to be calibrated based on the particle swarm algorithm to obtain target calibration data, and calibrating the carbon load calculation model in the electronic control unit based on the target calibration data.

[0006] Optionally, the target parameters include the number of individuals in the group, the change dimension of each individual to be optimized, the maximum number of iterations of the algorithm, the individual knowledge learning factor, the group knowledge learning factor, the inertia weight of the particle swarm algorithm, the upper and lower limits of the position of each change dimension of the individual, the upper and lower limits of the speed of each change dimension of the individual, the optimization target of the search constraint search, and at least one of the fitness calculation methods.

[0007] Optionally, the parameters to be calibrated are calibrated based on the particle swarm algorithm to obtain target calibration data, including: initializing the initial position and speed of all particle groups; calculating the historical target fitness of each single particle and the group respectively, iteratively updating the target individual and the target group according to the historical target fitness, and updating the position and speed of each particle after the iterative update; if the updated position and speed exceed the upper and lower boundaries of the position of each change dimension of the corresponding individual and the upper and lower boundaries of the speed of each change dimension of the individual, the corresponding particle is initialized to an arbitrary position and speed within the boundary; if the maximum number of iterations is not reached, continue to calculate the historical target fitness and iteratively update the target individual and the target group, otherwise output the target individual and the historical target fitness value of the target individual, and determine the target calibration data according to the target individual and the historical target fitness value of the target individual.

[0008] Optionally, updating the position and speed of each particle includes: updating the speed according to the current individual position relative to the individual historical target individual and the historical target individual in the group, and its own inertia, and updating the position according to the updated speed and the current individual position.

[0009] Optionally, before calibrating the carbon load calculation model in the electronic control unit based on the target calibration data, it also includes: identifying the target operating condition ratio in the regeneration measured data; and optimizing the target calibration data according to the target operating condition ratio.

[0010] Optionally, the parameters to be calibrated of the carbon load calculation model in the target virtual space are determined based on the regeneration measured data and the actual weighing data, including: inputting the regeneration measured data into the carbon load calculation model of the target virtual space and the carbon load calculation model of the electronic control unit respectively; determining the model fitting level according to the respective output values ​​of the target virtual space and the electronic control unit, and determining the model accuracy according to the output value of the target virtual space and the actual weighing data; determining the parameters to be calibrated according to the model fitting level and the model accuracy.

[0011] A second aspect of the present application provides a calibration device for a carbon load calculation model, including: an acquisition module for acquiring measured regeneration data of a particulate filter and actual weighing data of the carbon load; a determination module for determining parameters to be calibrated of the carbon load calculation model in a target virtual space based on the measured regeneration data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on a calibration data file of an electronic control unit; a calibration module for setting target parameters of a particle swarm algorithm, calibrating the parameters to be calibrated based on the particle swarm algorithm to obtain target calibration data, and calibrating the carbon load calculation model in the electronic control unit based on the target calibration data.

[0012] Optionally, the target parameters include the number of individuals in the group, the change dimension of each individual to be optimized, the maximum number of iterations of the algorithm, the individual knowledge learning factor, the group knowledge learning factor, the inertia weight of the particle swarm algorithm, the upper and lower limits of the position of each change dimension of the individual, the upper and lower limits of the speed of each change dimension of the individual, the optimization target of the search constraint search, and at least one of the fitness calculation methods.

[0013] Optionally, the calibration module is further used to initialize the initial position and speed of all particle groups; calculate the historical target fitness of each single particle and group respectively, iteratively update the target individual and target group according to the historical target fitness, and update the position and speed of each particle after the iterative update; if the updated position and speed exceed the upper and lower boundaries of the position of each change dimension of the corresponding individual and the upper and lower boundaries of the speed of each change dimension of the individual, the corresponding particle is initialized to any position and speed within the boundary; if the maximum number of iterations is not reached, continue to calculate the historical target fitness and iteratively update the target individual and target group, otherwise output the target individual and the historical target fitness value of the target individual, and determine the target calibration data according to the target individual and the historical target fitness value of the target individual

[0014] Optionally, the calibration module is further used to update the position according to the updated speed and the current individual position based on the current individual position relative to the individual historical target individual and the historical target individual in the group, and its own inertia update speed.

[0015] Optionally, the calibration device for the carbon load calculation model also includes: an identification module for identifying the target operating condition proportion in the regeneration measured data before calibrating the carbon load calculation model in the electronic control unit based on the target calibration data; and an optimization module for optimizing the target calibration data according to the target operating condition proportion.

[0016] Optionally, the determination module is further used to input the regeneration measured data into the carbon load calculation model of the target virtual space and the carbon load calculation model of the electronic control unit respectively; determine the model fitting level according to the respective output values ​​of the target virtual space and the electronic control unit, and determine the model accuracy according to the output value of the target virtual space and the actual weighing data; determine the parameters to be calibrated according to the model fitting level and the model accuracy.

[0017] A third aspect of the present application provides a vehicle, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a calibration method for a carbon load calculation model as in the above-mentioned embodiment.

[0018] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement a calibration method for a carbon loading calculation model as in the above-mentioned embodiment.

[0019] The fifth aspect of the present application provides a computer program product, including a computer program or instructions. When the computer program or instructions are executed, they are used to implement the calibration method of the carbon load calculation model as described in the above embodiment.

[0020] Therefore, this application includes the following beneficial effects:

[0021] The embodiment of the present application can determine the parameters to be calibrated of the carbon load calculation model in the target virtual space based on the regenerated measured data and the actual weighing data, and automatically find the target calibration data that meets the actual weighing data by combining the efficient parallel characteristics of the particle swarm algorithm, completely avoiding the blindness of manual calibration, greatly improving the model accuracy and calibration quality, saving labor costs, and shortening the development cycle. Thus, the problems of manual calibration and debugging taking a long time in the related technology, and being unable to find the global optimal solution, resulting in low model accuracy, are solved.

[0022] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0024] Figure 1 A flowchart of a calibration method for a carbon load calculation model provided in an embodiment of the present application;

[0025] Figure 2 A block diagram of a calibration device for a carbon load calculation model provided according to an embodiment of the present application;

[0026] Figure 3 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0028] The following describes the calibration method, device, equipment, medium and product of the carbon load calculation model of the embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides a calibration method for a carbon load calculation model, in which the parameters to be calibrated of the carbon load calculation model in the target virtual space are determined based on the regenerated measured data and the actual weighing data, and combined with the efficient parallel characteristics of the particle swarm algorithm, the target calibration data that meets the actual weighing data is automatically found, the blindness of manual calibration is completely avoided, the model accuracy and calibration quality are greatly improved, the labor cost is saved, and the development cycle is shortened. As a result, the problems of manual calibration and debugging taking a long time and being unable to find the global optimal solution, resulting in low model accuracy, etc. in the related technology are solved.

[0029] Specifically, Figure 1 A schematic flow chart of a method for calibrating a carbon loading calculation model provided in an embodiment of the present application.

[0030] like Figure 1 As shown, the calibration method of the carbon load calculation model includes the following steps:

[0031] In step S101 , actual measured data of regeneration of the particulate trap and actual weighing data of the carbon load are obtained.

[0032] It is understood that the regeneration measured data refers to all key parameters in the regeneration process, including but not limited to the temperature of the particulate trap, pressure difference, exhaust gas composition regeneration start and end time, regeneration efficiency, energy consumption during regeneration, etc. The embodiment of the present application can store the regeneration measured data in a DAT file. After the regeneration experiment is completed, the embodiment of the present application can obtain a certain amount of particulate matter, and the embodiment of the group itself can obtain the actual weighing data of the carbon load by weighing.

[0033] It should be noted that there may be multiple regenerated measured data in the embodiment of the present application, depending on the number of consecutive weighings. When the model error of a single weighing data deviates significantly from the mean, it is eliminated to avoid overfitting. The DCM file is stored in the form of a structure, directly named DCM. Each element in the structure is still nested as a structure, which contains each data label to be calibrated. Any coordinates and data of the label can be obtained through the data label structure.

[0034] In step S102, the parameters to be calibrated of the carbon load calculation model in the target virtual space are determined according to the regeneration measured data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on the calibration data file of the electronic control unit.

[0035] During the actual implementation process, the measured regeneration data are respectively input into the carbon load calculation model of the target virtual space and the carbon load calculation model of the electronic control unit; the model fitting level is determined according to the respective output values ​​of the target virtual space and the electronic control unit, and the model accuracy is determined according to the output value of the target virtual space and the actual weighing data; the parameters to be calibrated are determined according to the model fitting level and the model accuracy.

[0036] Among them, the target virtual space can be set according to the actual situation, and it is not necessary to specify it. The following embodiment of the present application can take the target space as the MATLAB workspace as an example to be elaborated in detail. The DAT files generated in all weighing experiments are read into the MATLAB workspace in the form of a cell array, represented as T_CELL. The cell array is a row vector, and its column dimension is equal to the number of consecutive weighing experiments. Each element is also a cell array data type, represented as a column vector, and its row dimension is equal to the number of single weighings in each continuous weighing experiment. Each row element is in the form of a structure, which represents the time series signal of a single DAT file. As the model input, the time series needs to eliminate blank and repeated time signals, and uniformly perform 10ms interpolation processing. All GPF weighing data is also saved in the form of a cell array. After each continuous weighing DAT file is read, the user is prompted to manually input it in the form of a row vector, which is finally represented as T_SootLoad. In the check of the degree of fit between the Simulink model and the ECU internal model input data, it is sufficient to directly check the model output and the corresponding carbon load time series signal read into the DAT file, and the model regeneration accuracy is expressed as in Indicates the change in carbon load in a single weighing. It indicates the change value of carbon loading of a single model, and the accuracy check result is displayed in a graphical way.

[0037] Specifically, the embodiment of the present application can classify all road test measurement DAT files into corresponding folders according to a continuous regeneration or cumulative carbon storage, and eliminate the weighing data that is obviously outlier or abnormal; read the calibration data file preset in the ECU during the road test to the MATLAB workspace; read the classified continuous regeneration data DAT files in sequence to the MATLAB workspace, and after each batch of continuous regeneration DAT files is read, manually enter the weighing data corresponding to the batch of regeneration DAT files according to the program prompts, the program automatically checks the fitting level of the Simulink model and the ECU internal model output data, and then automatically compares the model regeneration accuracy with the actual road weighing data, and determines the parameters to be calibrated after checking the model fitting degree and regeneration accuracy.

[0038] In step S103, target parameters of the particle swarm algorithm are set, the parameters to be calibrated are calibrated based on the particle swarm algorithm to obtain target calibration data, and the carbon load calculation model in the electronic control unit is calibrated based on the target calibration data.

[0039] Among them, the target parameters include the number of individuals in the group, the change dimension of each individual to be optimized, the maximum number of iterations of the algorithm, the individual knowledge learning factor, the group knowledge learning factor, the inertia weight of the particle swarm algorithm, the upper and lower limits of the position of each change dimension of the individual, the upper and lower limits of the speed of each change dimension of the individual, the optimization target of the search constraint search, and at least one of the fitness calculation methods.

[0040] Specifically, the number of individuals in the group can be expressed as N. The larger its value is, the greater the probability of obtaining the global optimal calibration parameters, but the iteration time becomes longer, and vice versa. It is generally taken between 50 and 100. The dimension of individuals to be optimized is expressed as D, which is equal to the sum of the number of breakpoints of all curves to be optimized, usually between 10 and 30. The maximum number of iterations of the algorithm is expressed as T, and its value is usually between 50 and 200. The more complex the problem to be optimized, the larger the number of iterations should be and the longer the iteration time, and vice versa.

[0041] The individual knowledge learning factor can be expressed as C1, and its value is usually between 1 and 2. The group knowledge learning factor is expressed as C2, and its value is usually between 1 and 2. The self-inertia weight is expressed as w=w max -(w max -w min )*i / T, where w max is the maximum inertia weight, w min is the minimum inertia weight, i is the current iteration number, which means that the inertia weight decreases linearly from the maximum value to the minimum value as the iteration proceeds. The above three coefficients represent the comprehensive effects of individual knowledge, group knowledge and self-inertia of particles in the process of moving to the global optimal position.

[0042] The upper and lower boundaries of the individual change dimension are respectively expressed in the form of row vectors, and their dimensions are equal to the individual dimension D, which limits the search boundary of the breakpoint value in the optimization process. It is generally set according to experience and the actual precision error of a single weighing data. The larger the absolute value of the precision error, the wider the search boundary should be, and vice versa; the upper and lower boundaries of the speed of the individual change dimension are respectively expressed in the form of row vectors, and their dimensions are equal to the individual dimension D, which characterizes the step size of each change of the individual dimension in the optimization process. It is set according to the actual breakpoint value. If it is set too small, it is easy to cause the convergence speed to be too slow, and if it is set too large, it is easy to skip the global optimum.

[0043] In addition, the particle swarm algorithm of the embodiment of the present application also has constraints, which are mainly to eliminate individuals that obviously do not meet the requirements and shorten the algorithm search time. In the calibration of the GPF model, it is not advisable to set too strict constraints. Considering the large amount of weighing data, if the constraints are too strict, it is easy to eliminate suboptimal individuals that meet most of the accuracy requirements, and even it is impossible to find a feasible solution. Therefore, only the breakpoint is set to meet the monotonicity constraint and comply with the physical law. For example, as the carbon load increases, the combustion rate should increase;

[0044] The optimization objective of the particle swarm algorithm is set as Where M represents the total number of weighing times, that is, the optimization target is the sum of the absolute values ​​of the precision of all weighing times, and the optimization target is to minimize this value. The purpose of setting this target is consistent with the purpose of setting the constraint conditions, that is, to give as loose constraints as possible, so that the measurement accuracy of all weighing data is close to zero, and to find as many suboptimal solutions as possible.

[0045] In one embodiment of the present application, the calibration parameters are calibrated based on the particle swarm algorithm to obtain target calibration data, including: initializing the initial position and speed of all particle groups; calculating the historical target fitness of each single particle and the group respectively, iteratively updating the target individual and the target group according to the historical target fitness, and updating the position and speed of each particle after the iterative update; if the updated position and speed exceed the upper and lower boundaries of the position of each change dimension of the corresponding individual and the upper and lower boundaries of the speed of each change dimension of the individual, the corresponding particle is initialized to an arbitrary position and speed within the boundary; if the maximum number of iterations is not reached, the historical target fitness is continued to be calculated and the target individual and the target group are iteratively updated, otherwise the target individual and the historical target fitness value of the target individual are output, and the target calibration data is determined according to the target individual and the historical target fitness value of the target individual.

[0046] On the basis of the above embodiments, the embodiments of the present application select calibration parameters that have a great impact on carbon loading as optimization objects, and use the high sensitivity to changes in carbon loading as the selection criteria (such as selecting a fixed carbon loading combustion rate curve, an original exhaust correction curve, etc.) as the optimization object of the particle swarm algorithm.

[0047] In the actual execution process, the initial positions and velocities of all particle groups are initialized. This initialization process allows each particle to randomly appear at any position within the limits of the position boundary and the velocity boundary. Random values ​​are randomly assigned within the position boundary and the velocity boundary by random number generation. According to the number and dimension of individuals, the initial position and velocity are respectively a matrix of N*D dimensions.

[0048] Furthermore, the embodiments of the present application can respectively calculate the historical best fitness of a single particle relative to itself, and continuously iterate and update the individual historical best individual; calculate the historical best fitness of the group, and continuously iterate and update the historical best individual of the group; wherein, the historical best fitness of a single particle is obtained by comparing it with its own fitness at each iteration, and the fitness is the optimization target, the program automatically runs to write the calibration parameters represented by the individual into the Simulink model, runs the model the same number of times as the weighing number, compares the actual weighing value, calculates and returns the fitness, and directly assigns extremely high fitness to individuals that do not meet the constraint conditions, that is, the breakpoint monotonicity, to automatically eliminate the individual; the historical best fitness of the group and the individual are obtained by comparing the historical best individual fitness of the entire particle group, and are dynamically updated in real time, that is, each time the individual position is updated, the historical best individual and the historical best individual of the group are updated in turn;

[0049] Furthermore, the embodiment of the present application can also update the position and speed of each particle, including: the speed update principle is to compare the current individual position relative to the individual's best individual position in history and the best individual position in the group, and finally combine its own inertia to obtain the updated speed, and the position update can be obtained by summing the current position and the updated speed. Among them, the individual position vector is updated by the following expression in represents the updated velocity vector, represents the old velocity vector, R represents a random number between 0 and 1, Represents the best individual position vector of a single particle itself in history, represents the best individual position vector in the current group history, represents the updated individual position vector. As mentioned above, the updated expressions of the velocity vector represent the inertia part, the individual knowledge part, and the group knowledge part respectively;

[0050] Furthermore, when any element in the updated individual velocity vector and position vector exceeds the predefined upper and lower boundaries, the element is randomly initialized to any value within the boundaries, and it is determined whether the algorithm iteration termination condition is met. If not, the historical target fitness and iterative update target individuals and target groups are continued to be calculated, otherwise the historical best individual and its fitness value are output, and updated to the DCM file; it should be noted that the algorithm iteration termination condition is any of the following conditions: there is no improvement in the historical best fitness of the group for consecutive fixed generations; the number of iterations reaches the maximum predefined generation. After the algorithm ends, the program automatically outputs the historical best fitness of the group, and outputs the model accuracy value of each single weighing data under the optimal optimization target, outputs the historical best individual of the group, that is, the required target calibration data, and writes it directly into the calibration data DCM file.

[0051] In one embodiment of the present application, before calibrating the carbon load calculation model in the electronic control unit based on the target calibration data, it also includes: identifying the target operating condition ratio in the regeneration measured data; and optimizing the target calibration data according to the target operating condition ratio.

[0052] It can be understood that the operating condition distribution ratio in each DAT data file is automatically extracted, such as the exhaust temperature and oxygen flow rate where regeneration is mainly distributed, the speed and load where carbon accumulation is mainly distributed, and the model accuracy and main operating condition distribution of each DAT file are fine-tuned to further improve the overall model accuracy.

[0053] Furthermore, the embodiment of the present application can automatically write the DCM data updated by the particle swarm algorithm and manual fine-tuning into the ECU through the program, thereby completing the GPF desktop calibration based on the particle swarm algorithm.

[0054] According to the calibration method of the carbon load calculation model proposed in the embodiment of the present application, the parameters to be calibrated of the carbon load calculation model in the target virtual space are determined based on the regenerated measured data and the actual weighing data, and the target calibration data that meets the actual weighing data is automatically found in combination with the efficient parallel characteristics of the particle swarm algorithm, completely avoiding the blindness of manual calibration, greatly improving the model accuracy and calibration quality, saving labor costs, and shortening the development cycle. Thus, the problems of manual calibration and debugging taking a long time in the related technology, and being unable to find the global optimal solution, resulting in low model accuracy, are solved.

[0055] Next, a calibration device for a carbon load calculation model proposed in an embodiment of the present application will be described with reference to the accompanying drawings.

[0056] Figure 2 It is a block diagram of a calibration device for a carbon loading calculation model according to an embodiment of the present application.

[0057] like Figure 2 As shown, the calibration device 10 of the carbon load calculation model includes: an acquisition module 100 , a determination module 200 and a calibration module 300 .

[0058] Among them, the acquisition module 100 is used to obtain the measured regeneration data of the particulate filter and the actual weighing data of the carbon load; the determination module 200 is used to determine the parameters to be calibrated of the carbon load calculation model in the target virtual space according to the measured regeneration data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on the calibration data file of the electronic control unit; the calibration module 300 is used to set the target parameters of the particle swarm algorithm, calibrate the parameters to be calibrated based on the particle swarm algorithm to obtain the target calibration data, and calibrate the carbon load calculation model in the electronic control unit based on the target calibration data.

[0059] In one embodiment of the present application, the target parameters include the number of individuals in the group, the change dimension of each individual to be optimized, the maximum number of iterations of the algorithm, the individual knowledge learning factor, the group knowledge learning factor, the inertia weight of the particle swarm algorithm, the upper and lower limits of the position of each change dimension of the individual, the upper and lower limits of the speed of each change dimension of the individual, the optimization target of the search constraint search, and at least one of the fitness calculation methods.

[0060] In one embodiment of the present application, the calibration module 300 is further used to initialize the initial position and speed of all particle groups; calculate the historical target fitness of each single particle and group respectively, iteratively update the target individual and target group according to the historical target fitness, and update the position and speed of each particle after the iterative update; if the updated position and speed exceed the upper and lower boundaries of the position of each change dimension of the corresponding individual and the upper and lower boundaries of the speed of each change dimension of the individual, the corresponding particle is initialized to an arbitrary position and speed within the boundary; if the maximum number of iterations is not reached, continue to calculate the historical target fitness and iteratively update the target individual and target group, otherwise output the historical target fitness value of the target individual and the target individual, and determine the target calibration data according to the historical target fitness value of the target individual and the target individual

[0061] In one embodiment of the present application, the calibration module 300 is further used to update the position according to the updated speed and the current individual position based on the current individual position relative to the individual historical target individual and the historical target individual in the group, as well as its own inertia update speed.

[0062] In one embodiment of the present application, the calibration device 10 for the carbon load calculation model further includes: an identification module and an optimization module.

[0063] Among them, the identification module is used to identify the target operating condition ratio in the regeneration measured data before calibrating the carbon load calculation model in the electronic control unit based on the target calibration data; the optimization module is used to optimize the target calibration data according to the target operating condition ratio.

[0064] In one embodiment of the present application, the determination module 200 is further used to input the regeneration measured data into the carbon load calculation model of the target virtual space and the carbon load calculation model of the electronic control unit respectively; determine the model fitting level according to the respective output values ​​of the target virtual space and the electronic control unit, and determine the model accuracy according to the output value of the target virtual space and the actual weighing data; determine the parameters to be calibrated according to the model fitting level and the model accuracy.

[0065] It should be noted that the above explanation of the embodiment of the calibration method for the carbon loading calculation model is also applicable to the calibration device for the carbon loading calculation model of this embodiment, which will not be repeated here.

[0066] According to the calibration device of the carbon load calculation model proposed in the embodiment of the present application, the parameters to be calibrated of the carbon load calculation model in the target virtual space are determined based on the regenerated measured data and the actual weighing data, and the target calibration data that meets the actual weighing data is automatically found in combination with the efficient parallel characteristics of the particle swarm algorithm, completely avoiding the blindness of manual calibration, greatly improving the model accuracy and calibration quality, saving labor costs, and shortening the development cycle. Thus, the problems of manual calibration and debugging taking a long time in the related technology, and being unable to find the global optimal solution, resulting in low model accuracy, are solved.

[0067] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0068] A memory 301 , a processor 302 , and a computer program stored in the memory 301 and executable on the processor 302 .

[0069] When the processor 302 executes the program, the calibration method of the carbon load calculation model provided in the above embodiment is implemented.

[0070] Furthermore, the electronic device further comprises:

[0071] The communication interface 303 is used for communication between the memory 301 and the processor 302 .

[0072] The memory 301 is used to store computer programs that can be run on the processor 302 .

[0073] The memory 301 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0074] If the memory 301, the processor 302 and the communication interface 303 are implemented independently, the communication interface 303, the memory 301 and the processor 302 can be connected to each other through a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0075] Optionally, in a specific implementation, if the memory 301, the processor 302 and the communication interface 303 are integrated on a chip, the memory 301, the processor 302 and the communication interface 303 can communicate with each other through an internal interface.

[0076] The processor 302 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0077] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the calibration method of the carbon loading calculation model as described above.

[0078] The present application also provides a computer program product, including a computer program or an instruction, which, when executed, is used to implement the calibration method of the carbon load calculation model as described above.

[0079] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0080] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0081] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0082] It should be understood that the various parts of the present application can be implemented in hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, the steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array, a field programmable gate array, etc.

[0083] A person of ordinary skill in the art may understand that all or part of the steps carried by the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the above-mentioned program may be stored in a computer-readable storage medium, which, when executed, includes one of the steps of the method embodiment or a combination thereof.

[0084] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A calibration method for a carbon loading calculation model, It is characterized in that The following steps are involved: Obtain the measured data of the particulate filter regeneration and the actual weighing data of the carbon load; Determining parameters to be calibrated of a carbon load calculation model in a target virtual space according to the regeneration measured data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on a calibration data file of an electronic control unit; The target parameters of the particle swarm algorithm are set, the parameters to be calibrated are calibrated based on the particle swarm algorithm to obtain target calibration data, and the carbon load calculation model in the electronic control unit is calibrated based on the target calibration data.

2. The calibration method of the carbon load calculation model according to claim 1, It is characterized in that The target parameters include the number of individuals in the group, the change dimension of each individual to be optimized, the maximum number of iterations of the algorithm, the individual knowledge learning factor, the group knowledge learning factor, the inertia weight of the particle swarm algorithm, the upper and lower limits of the position of each change dimension of the individual, the upper and lower limits of the speed of each change dimension of the individual, the optimization target of the search constraint search, and at least one of the fitness calculation methods.

3. The calibration method of the carbon load calculation model according to claim 2, It is characterized in that The step of calibrating the parameters to be calibrated based on the particle swarm algorithm to obtain target calibration data includes: Initialize the initial position and velocity of all particle groups; Calculate the historical target fitness of each individual particle and the group respectively, iteratively update the target individual and the target group according to the historical target fitness, and after the iterative update, update the position and speed of each particle; If the updated position and velocity exceed the upper and lower bounds of the position of each change dimension of the corresponding individual and the upper and lower bounds of the velocity of each change dimension of the individual, the corresponding particle will be initialized to any position and velocity within the bounds; If the maximum number of iterations is not reached, continue to calculate the historical target fitness and iteratively update the target individual and the target group; otherwise, output the target individual and the historical target fitness value of the target individual, and determine the target calibration data based on the target individual and the historical target fitness value of the target individual.

4. The calibration method of the carbon load calculation model according to claim 3, It is characterized in that The updating of the position and velocity of each particle includes: According to the current individual position relative to the individual historical target individual and the historical target individual in the group, and the own inertia update speed, the position is updated according to the updated speed and the current individual position.

5. The calibration method of the carbon load calculation model according to claim 1, It is characterized in that Before calibrating the carbon load calculation model in the electronic control unit based on the target calibration data, the method further includes: Identifying a target operating condition ratio in the regeneration measured data; The target calibration data is optimized according to the target operating condition proportion.

6. The calibration method of the carbon load calculation model according to claim 1, It is characterized in that The step of determining the parameters to be calibrated of the carbon load calculation model in the target virtual space according to the regeneration measured data and the actual weighing data includes: inputting the regeneration measured data into the carbon load calculation model of the target virtual space and the carbon load calculation model of the electronic control unit respectively; Determining a model fitting level according to the respective output values ​​of the target virtual space and the electronic control unit, and determining a model accuracy according to the output value of the target virtual space and the actual weighing data; The parameters to be calibrated are determined according to the model fitting level and the model accuracy.

7. A calibration device for a carbon load calculation model, It is characterized in that include: An acquisition module, used to acquire the measured regeneration data of the particulate filter and the actual weighing data of the carbon load; a determination module, configured to determine parameters to be calibrated of a carbon load calculation model in a target virtual space according to the regeneration measured data and the actual weighing data, wherein the carbon load calculation model in the target virtual space is determined based on a calibration data file of an electronic control unit; The calibration module is used to set the target parameters of the particle swarm algorithm, calibrate the parameters to be calibrated based on the particle swarm algorithm to obtain target calibration data, and calibrate the carbon load calculation model in the electronic control unit based on the target calibration data.

8. An electronic device, It is characterized in that include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the calibration method for the carbon load calculation model according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instructions stored thereon, It is characterized in that When the computer program or instruction is executed, the calibration method of the carbon loading calculation model according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program or instructions, It is characterized in that When the computer program or instruction is executed, the calibration method of the carbon loading calculation model according to any one of claims 1 to 6 is implemented.