Training method, identification method, system and equipment of thermal runaway identification model
By screening effective data from vehicle-related data and training a thermal runaway identification model, the problem of thermal runaway risk prediction for electric vehicles and hybrid electric vehicles is solved, and the accuracy and safety of identification are improved.
Patent Information
- Application Number
- CN202411308905.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-19
- Publication Date
- 2025-01-21
AI Technical Summary
The prior art is difficult to accurately predict whether electric vehicles and hybrid vehicles have a risk of thermal runaway, resulting in safety hazards.
By filtering out effective data from vehicle-related data, performing labeling and feature extraction, and training a thermal runaway identification model to predict whether the vehicle has a risk of thermal runaway.
The accuracy and reliability of thermal runaway identification are improved, and the safety of the vehicle is enhanced.
Smart Images
Figure CN119337175A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of vehicle technology, and in particular to a training method, an identification method, a system and a device for a thermal runaway identification model. Background Art
[0002] While the vehicle is running, it can upload vehicle-related data to the server, such as speed, longitude and latitude, vehicle driving status, battery and other information. The battery is an important component of electric vehicles, and most vehicles with thermal runaway start with the battery. Therefore, analyzing thermal runaway from the basic information of the battery is an important research.
[0003] The issue of vehicle thermal runaway is particularly important because it involves the safety of electric and hybrid vehicles. These vehicles usually use lithium-ion batteries as power sources, and lithium-ion batteries may generate a lot of heat during charging or discharging. If the heat cannot be effectively dissipated, it may cause a thermal runaway reaction, causing battery overheating, or even sudden fire or explosion. Therefore, how to accurately predict whether a vehicle has a risk of thermal runaway is particularly important for vehicle safety. Summary of the invention
[0004] Based on this, it is necessary to provide a training method, identification method, system and equipment for a thermal runaway identification model to address the above-mentioned technical problems, which have the advantages of accurate and reliable identification of thermal runaway.
[0005] In a first aspect, a method for training a thermal runaway identification model is provided, comprising:
[0006] Obtaining vehicle-related data, wherein the vehicle-related data includes normal vehicle-related data and abnormal vehicle-related data when thermal runaway occurs;
[0007] Filter out valid vehicle-related data from the vehicle-related data;
[0008] The valid vehicle-related data are labeled to obtain a data set, wherein the label of normal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has not occurred, and the label of abnormal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has occurred or is about to occur;
[0009] The data set is input into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model, wherein the output of the thermal runaway identification model represents a prediction result of thermal runaway.
[0010] In some examples, before filtering out valid vehicle-related data from the vehicle-related data, the method further includes:
[0011] Selecting a predetermined data field from the vehicle-related data, wherein the predetermined data field at least includes part or all of the following:
[0012] General alarm sign, maximum alarm level, maximum single cell temperature, minimum single cell temperature, single cell temperature list, maximum single cell voltage value, minimum single cell voltage value, single cell voltage list, insulation resistance value, total battery voltage, total battery current, battery state of charge, charging state and vehicle status.
[0013] In some examples, filtering out valid vehicle-related data from the vehicle-related data includes:
[0014] Identifying an invalid value corresponding to each data field in the vehicle-related data;
[0015] The invalid values are deleted to filter out valid vehicle-related data from the vehicle-related data.
[0016] In some examples, inputting the data set into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model includes:
[0017] Extracting characteristic data reflecting thermal runaway from the valid vehicle-related data;
[0018] The characteristic data and the corresponding labels are input into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model.
[0019] In some examples, extracting characteristic data reflecting thermal runaway from the valid vehicle-related data includes:
[0020] The characteristic data is extracted according to the data fields of the valid vehicle-related data, wherein the characteristic data includes a plurality of characteristic fields, wherein the characteristic fields at least include part or all of the following characteristic fields:
[0021] General alarm sign, maximum alarm level, maximum single cell temperature, minimum single cell temperature, the range between the maximum single cell temperature and the minimum single cell temperature, the maximum deviation between the single cell temperature and the average single cell temperature, the maximum single cell voltage value, the minimum single cell voltage value, the range between the maximum single cell voltage value and the minimum single cell voltage value, the maximum deviation between the single cell voltage value and the average single cell voltage value, and the insulation resistance value.
[0022] In some examples, inputting the data set into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model includes:
[0023] The data set is input into an initial thermal runaway recognition model, so as to train the initial thermal runaway recognition model according to the loss between the output of the initial thermal runaway recognition model and the label, and obtain a trained thermal runaway recognition model.
[0024] In some examples, after the thermal runaway identification model is trained, the following is also included:
[0025] Obtaining a recall rate according to a prediction result of the thermal runaway identification model;
[0026] The thermal runaway identification model is evaluated according to the recall rate.
[0027] In a second aspect, a thermal runaway identification method is provided, comprising:
[0028] Obtain vehicle-related data;
[0029] The vehicle-related data is input into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model described in the first aspect above.
[0030] In a third aspect, a thermal runaway identification system is provided, comprising:
[0031] An acquisition module, used to obtain vehicle-related data;
[0032] An identification module is used to input the vehicle-related data into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model described in the first aspect above.
[0033] In a fourth aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the thermal runaway identification method according to the second aspect is implemented.
[0034] By adopting the embodiment of the present application, valid vehicle-related data is screened out from the vehicle-related data, the valid vehicle-related data is annotated to obtain a data set, and then the data set is input into the initial thermal runaway identification model to train the initial thermal runaway identification model, thereby obtaining a trained thermal runaway identification model. Before training the model through data, the data is pre-processed and the data is visualized for analysis, thereby being able to identify invalid data from the complex data, effectively remove impurities from the data, and improve the availability and quality of the data. Thus, appropriate data fragments are selected from the vehicle's full life cycle data to improve the accuracy and reliability of model training, so that the model can more accurately identify thermal runaway. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:
[0036] Figure 1 A flowchart of a method for training a thermal runaway identification model provided in an embodiment of the present application;
[0037] Figure 2 A flow chart of a thermal runaway identification method provided in an embodiment of the present application;
[0038] Figure 3 A structural block diagram of a thermal runaway identification system provided in an embodiment of the present application;
[0039] Figure 4 A structural block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0040] The present application is further described in detail below in conjunction with the embodiments and drawings. It is to be understood that the specific embodiments described herein are only used to explain the relevant application, rather than to limit the application. It is also necessary to explain that, for ease of description, only the parts related to the application are shown in the drawings.
[0041] It should be noted that, in the absence of conflict, the embodiments of the present application, that is, the features of the embodiments, can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0042] The following describes in detail the training method, identification method, system and device of the thermal runaway identification model according to the embodiments of the present application in conjunction with the accompanying drawings.
[0043] Figure 1 FIG. 1 is a flow chart of a method for training a thermal runaway identification model according to an embodiment of the present application. Figure 1 As shown, a method for training a thermal runaway identification model according to an embodiment of the present application includes the following steps:
[0044] S101: Obtain vehicle-related data, wherein the vehicle-related data includes normal vehicle-related data and abnormal vehicle-related data when thermal runaway occurs.
[0045] Among them, vehicle-related data includes but is not limited to vehicle data and battery data of power batteries, for example: vehicle-related data includes single-cell voltage, voltage extreme value, temperature extreme value, vehicle driving status, battery fault information, charging and discharging data, etc. of power batteries. These data usually include normal vehicle-related data, i.e., vehicle-related data when there is no risk of thermal runaway. Of course, it is usually necessary to collect vehicle-related data when there is a risk of thermal runaway or when thermal runaway occurs, which is referred to as abnormal vehicle-related data in this application.
[0046] In a specific example, the vehicle-related data may be collected by a server, that is, the vehicle uploads the real-time data of the vehicle and the real-time data of the power battery to the server, so that the server collects the vehicle-related data in real time.
[0047] S102: Filter out valid vehicle-related data from the vehicle-related data.
[0048] In one embodiment of the present application, before filtering out valid vehicle-related data from the vehicle-related data, it includes: selecting predetermined data fields from the vehicle-related data, wherein the predetermined data fields include at least some or all of the following: general alarm flag, maximum alarm level, maximum single cell temperature, minimum single cell temperature, single cell temperature list, maximum single cell voltage value, minimum single cell voltage value, single cell voltage list, insulation resistance value, total battery voltage, total battery current, battery state of charge, charging status and vehicle status.
[0049] That is, vehicle-related data usually includes some data fields that are irrelevant to thermal runaway or have relatively little relevance. These data fields have no effect or little effect on the judgment of thermal runaway. Therefore, before filtering out valid vehicle-related data from the vehicle-related data, data cleaning can be performed to clean up some data fields that are irrelevant to thermal runaway or have relatively little relevance. In this way, the amount of data can be effectively reduced, the data quality can be improved, and the accuracy and efficiency of subsequent thermal runaway judgments can be improved. That is, for vehicle-related data, in a specific example, the data fields retained after cleaning include but are not limited to the several field information shown in Table 1.
[0050] Table 1
[0051]
[0052]
[0053] Among them, the general alarm sign means: when a possible problem is identified, a value is reported. The highest alarm level means: the uploaded value during normal driving time is 0, and when an abnormality is identified, a different value will be prompted, indicating that a fault will occur. The highest temperature probe means: the highest single cell temperature. The lowest temperature probe means: the lowest single cell temperature.
[0054] After cleaning the vehicle-related data, filtering out valid vehicle-related data from the vehicle-related data includes: identifying invalid values corresponding to each data field in the vehicle-related data; and deleting the invalid values to filter out valid vehicle-related data from the vehicle-related data.
[0055] For example: Delete the invalid rows where the maximum temperature probe value and the minimum temperature probe value are both 255.
[0056] Removed invalid value of 255 for charging status.
[0057] Delete the invalid value of the highest alarm level which is null.
[0058] Delete the invalid value of the highest cell voltage value and the lowest voltage value of 65.535.
[0059] Delete the invalid rows in the single cell list where 1 / 4 of the values exceed 65.535. Use the same method to delete invalid values in the temperature probe list.
[0060] In the above description, 255 and 65.535 represent error values.
[0061] S103: Labeling the valid vehicle-related data to obtain a data set, wherein the label of normal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has not occurred, and the label of abnormal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has occurred or is about to occur.
[0062] Specifically, data labeling is the basis of supervised learning algorithms. In supervised learning, model training requires labeled data for learning. Labeled data is the basis for model learning and directly affects the accuracy and robustness of the model. By labeling the data, you can clearly understand the characteristics and labels of the data, thereby evaluating the quality and consistency of the data, and helping to identify possible problems, such as missing values, outliers, or label errors.
[0063] When analyzing based on features, abnormal data points can be obtained through analysis of temperature, voltage, insulation resistance and alarm level in the thermal runaway vehicle data. For example, the temperature of some data reaches 70 degrees in the early stage of thermal runaway, and the highest temperature even reaches 100 degrees. The voltage range of each cell increases from 0.04v to 0.5v. It can be concluded from the data that the current data is valid data for thermal runaway marking.
[0064] Based on the above analysis, the determined thermal runaway data can be extracted from the data as a data set where thermal runaway occurs. When performing label processing, a comprehensive analysis of voltage, temperature, resistance, and alarm level is used as the basis for data labeling, with 1 being used as thermal runaway warning data and 0 being used as data where no thermal runaway warning is required.
[0065] The label represents whether it is normal data or abnormal data indicating that thermal runaway has occurred or is about to occur.
[0066] S104: Inputting the data set into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model, wherein an output of the thermal runaway identification model represents a prediction result of thermal runaway.
[0067] In one embodiment of the present application, the data set is input into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model, including: extracting feature data reflecting thermal runaway from the valid vehicle-related data; inputting the feature data and corresponding labels into the initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model.
[0068] Specifically, the data set is input into the initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model, including: inputting the data set into the initial thermal runaway identification model to train the initial thermal runaway identification model according to the output of the initial thermal runaway identification model and the loss between the label to obtain a trained thermal runaway identification model.
[0069] In this example, characteristic data reflecting thermal runaway is extracted from the valid vehicle-related data, including: extracting the characteristic data according to the data field of the valid vehicle-related data, the characteristic data including multiple characteristic fields, wherein the characteristic field includes at least part or all of the following characteristic fields: general alarm flag, maximum alarm level, maximum single cell temperature, minimum single cell temperature, the range between the maximum single cell temperature and the minimum single cell temperature, the maximum deviation between the single cell temperature and the average single cell temperature, the highest single cell voltage value, the lowest single cell voltage value, the range between the highest single cell voltage value and the lowest single cell voltage value, the maximum deviation between the single cell voltage value and the average single cell voltage value, and insulation resistance value.
[0070] In a specific example, the thermal runaway identification model takes the deep learning LSTM model as an example. Feature engineering during model training is a very important part of data science and machine learning, including steps such as data preprocessing, feature selection, feature extraction, and feature conversion.
[0071] In deep learning, it is determined which features of the model training learning data are used, that is, the input of the model. In the thermal runaway model, it is necessary to find data that can reflect the thermal runaway of the vehicle. The battery temperature, voltage, insulation resistance, and alarm level can be used to directly and effectively reflect the data performance. In terms of voltage change, the battery may experience a rapid voltage drop as the temperature increases. Not only temperature and voltage can be used as direct representations, but the body and battery are insulated, and the insulation value is also a prominent performance. When thermal runaway occurs, high voltage will cause a short circuit. The alarm level reflects the current vehicle status, which is divided into no impact on use, need to reduce power, and serious abnormality. The comprehensive alarm level, temperature and voltage can better reflect the current vehicle status and increase the robustness and real-time performance of the model.
[0072] When analyzing the temperature and voltage list data, the data is processed for range and deviation. The range can reflect the thermal runaway prediction performance of temperature and voltage within the normal range. If the range is too large, it can be partially determined that there is a risk. The deviation processing can reflect the degree of deviation of the data list. If the deviation is too large, it can be determined that a certain battery cell has a problem, and it can be determined that there will be a problem. The formula is as follows:
[0073] Maximum value of cell voltage deviation = max(|cell voltage - cell average voltage|)
[0074] Single cell voltage range = |max(single cell voltage)-min(single cell voltage)|
[0075] Maximum value of monomer temperature deviation = max(|monomer temperature - monomer average temperature|)
[0076] Monomer temperature extreme difference = |max(monomer temperature)-min(monomer temperature)|
[0077] Finally, the features of the thermal runaway identification model are selected as shown in Table 2:
[0078] Table 2
[0079] Field Name Field Type General alarm sign varchar Maximum alarm level varchar Maximum temperature probe value float Minimum temperature probe value float Probe temperature extremes float Probe temperature deviation float Maximum single cell voltage float Minimum single cell voltage float Maximum voltage deviation float Voltage difference float Insulation resistance float
[0080] Before training the data set, the data set can be divided into a training set and a test set in an 8:2 ratio. The training set is used to train the thermal runaway identification model, and the test set is used to verify the accuracy of the model.
[0081] In the selection of the thermal runaway identification model, the LSTM model is selected as the model processing option. This is a deep learning algorithm model that can be used to solve the long short-term memory network problem of the model, making the model more interpretable and accurate.
[0082] Initialization parameters: First, the parameters of the model need to be initialized. You can use some random parameters to initialize the parameters first.
[0083] Select sample data: During each iterative process, select corresponding samples and labels from the training set as input data.
[0084] Gradient estimation: Estimate the gradient of the objective function with respect to the parameters based on the selected samples. The gradient represents the range of variation of the current parameters.
[0085] Update parameters: Use gradient estimation to update model parameters, control the step size of each update, and choose a smaller learning rate to make model updates more stable.
[0086] Repeat iterations: until the predetermined number of iterations is reached or convergence conditions are met.
[0087] Model convergence judgment: In each iteration, the model monitors the value of the objective function or the amplitude of parameter update to determine whether the algorithm has converged.
[0088] Model output: When the final iteration is completed or the target value has reached the expected value, the current model file is output as the result.
[0089] Prediction results: Based on the newly added data, the desired result data is obtained through model processing.
[0090] After the model is trained, that is, after the thermal runaway identification model is trained, it also includes: obtaining the recall rate according to the prediction results of the thermal runaway identification model; and evaluating the thermal runaway identification model according to the recall rate. In other words, when the model is trained, the quality of the model needs to be evaluated by indicators. The currently selected indicator is the recall rate. The goal of the model is not to miss a vehicle with a probability of thermal runaway. Among them: the recall rate is (the correct number of thermal runaway predictions) / (the correct number of thermal runaway predictions + the wrong number of thermal runaway predictions).
[0091] According to the training method of the thermal runaway identification model of the embodiment of the present application, valid vehicle-related data is screened out from the vehicle-related data, the valid vehicle-related data is annotated to obtain a data set, and then the data set is input into the initial thermal runaway identification model to train the initial thermal runaway identification model, thereby obtaining a trained thermal runaway identification model. Before training the model through data, the data is pre-processed and the data is visualized and analyzed, thereby being able to identify invalid data from the complex data, effectively remove impurities from the data, and improve the availability and quality of the data. Thus, appropriate data fragments are selected from the vehicle's full life cycle data to improve the accuracy and reliability of the model training, so that the model can more accurately identify thermal runaway.
[0092] In addition, in specific applications, the model uses the LSTM model, and the interpretability of the LSTM can be achieved by visualizing the internal state of the model. During training and prediction, the weights of the input gate, forget gate, and output gate in the LSTM model can be checked. These weights represent the importance of each feature for the decision at the current time step. By observing the changes in these weights, the time series patterns learned by the LSTM model can be inferred. The cell state represents the internal memory built by the model in the time series, while the output represents the model's prediction for the current time step. By observing the changes in the cell state and output, you can understand how the model makes predictions based on past information.
[0093] In the LSTM model, there are three main gating mechanisms: input gate, forget gate, and output gate, which are responsible for controlling the flow and selection of information. Input gate: determines how much of the information input at the current time step will be passed to the cell state. It uses the sigmoid activation function to compress the input features and generate weights between 0 and 1, thereby determining the importance of the information. Forget gate: determines which information in the cell state should be forgotten at the current time step. The forget gate also uses the sigmoid activation function to generate weights to determine the degree of information retention in the cell state. Output gate: determines which information in the cell state is passed to the output of the current time step. The output gate generates weights through the sigmoid activation function and generates outputs using the tanh activation function, mapping the cell state to between -1 and 1.
[0094] Figure 2 FIG. 1 is a flow chart of a method for identifying thermal runaway according to an embodiment of the present application. Figure 2 As shown, a thermal runaway identification method according to an embodiment of the present application includes:
[0095] S201: Obtain vehicle related data;
[0096] S202: Inputting the vehicle-related data into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model described in any one of the above-mentioned embodiments.
[0097] According to the thermal runaway identification method of the embodiment of the present application, when the model is trained, valid vehicle-related data is screened out from the vehicle-related data, the valid vehicle-related data is annotated, a data set is obtained, and then the data set is input into the initial thermal runaway identification model to train the initial thermal runaway identification model, thereby obtaining a trained thermal runaway identification model. Before the model is trained through data, the data is pre-processed and the data is visualized for analysis, thereby being able to identify invalid data from the complex data, effectively remove impurities from the data, and improve the availability and quality of the data. Thus, appropriate data fragments are selected from the vehicle's full life cycle data, thereby improving the accuracy and reliability of the model training, making the model more accurate in identifying thermal runaway, and thus, the thermal runaway identification method has the advantage of high accuracy in thermal runaway identification.
[0098] Figure 3 is a block diagram of a thermal runaway identification system according to an embodiment of the present application. Figure 3The thermal runaway identification system according to the embodiment of the present application includes: an acquisition module 310 and an identification module 320, wherein:
[0099] An acquisition module 310 is used to obtain vehicle-related data;
[0100] The identification module 320 is used to input the vehicle-related data into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method of the thermal runaway identification model described in any one of the above-mentioned embodiments.
[0101] According to the thermal runaway identification system of the embodiment of the present application, when the model is trained, valid vehicle-related data is screened out from the vehicle-related data, the valid vehicle-related data is annotated, a data set is obtained, and then the data set is input into the initial thermal runaway identification model to train the initial thermal runaway identification model, thereby obtaining a trained thermal runaway identification model. Before the model is trained through data, the data is pre-processed and the data is visualized for analysis, thereby being able to identify invalid data from the complex data, effectively remove impurities from the data, and improve the availability and quality of the data. Thus, appropriate data fragments are selected from the vehicle's full life cycle data, thereby improving the accuracy and reliability of the model training, making the model more accurate in identifying thermal runaway, and thus, the thermal runaway identification method has the advantage of high accuracy in thermal runaway identification.
[0102] For the specific definition of the thermal runaway identification system, please refer to the definition of the thermal runaway identification method above, which will not be repeated here. The various modules of the above-mentioned thermal runaway identification system can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0103] Reference below Figure 4 , Figure 4 A schematic diagram of the structure of a computer device suitable for implementing the embodiments of the present application is shown.
[0104] like Figure 4As shown, the computer system includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1008 into the random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation instructions of the system are also stored. The CPU 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.
[0105] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.
[0106] In particular, according to an embodiment of the present application, the above reference flow chart Figure 2 The described process may be implemented as a computer-readable storage medium. For example, an embodiment of the present application includes a computer-readable storage medium including a computer program, the computer program including program code for executing the method shown in the flowchart, such as executing: obtaining vehicle-related data;
[0107] The vehicle-related data is input into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model described in any one of the above embodiments.
[0108] In particular, according to an embodiment of the present application, the above reference flow chart Figure 2 The described process may be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains program code for executing the method shown in the flowchart.
[0109] In such an embodiment, the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, the above functions defined in the system of the present application are executed.
[0110] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
[0111] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, a program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.
[0112] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.
[0113] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0114] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.
Claims
1. A training method for a thermal runaway identification model, characterized in that: include: Obtaining vehicle-related data, wherein the vehicle-related data includes normal vehicle-related data and abnormal vehicle-related data when thermal runaway occurs; Filter out valid vehicle-related data from the vehicle-related data; The valid vehicle-related data are labeled to obtain a data set, wherein the label of normal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has not occurred, and the label of abnormal vehicle-related data in the valid vehicle-related data indicates that thermal runaway has occurred or is about to occur; The data set is input into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model, wherein the output of the thermal runaway identification model represents a prediction result of thermal runaway.
2. The training method of the thermal runaway identification model according to claim 1, characterized in that: Before filtering out valid vehicle-related data from the vehicle-related data, the method further includes: Selecting a predetermined data field from the vehicle-related data, wherein the predetermined data field at least includes part or all of the following: General alarm sign, maximum alarm level, maximum single cell temperature, minimum single cell temperature, single cell temperature list, maximum single cell voltage value, minimum single cell voltage value, single cell voltage list, insulation resistance value, total battery voltage, total battery current, battery state of charge, charging state and vehicle status.
3. The training method of the thermal runaway identification model according to claim 1, characterized in that: The filtering out valid vehicle-related data from the vehicle-related data comprises: Identifying an invalid value corresponding to each data field in the vehicle-related data; The invalid values are deleted to filter out valid vehicle-related data from the vehicle-related data.
4. The training method of the thermal runaway identification model according to claim 2, characterized in that: The step of inputting the data set into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model includes: Extracting characteristic data reflecting thermal runaway from the valid vehicle-related data; The characteristic data and the corresponding labels are input into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model.
5. The training method for the thermal runaway identification model according to claim 4, characterized in that: The extracting characteristic data reflecting thermal runaway from the valid vehicle-related data includes: The characteristic data is extracted according to the data fields of the valid vehicle-related data, wherein the characteristic data includes a plurality of characteristic fields, wherein the characteristic fields at least include part or all of the following characteristic fields: General alarm sign, maximum alarm level, maximum single cell temperature, minimum single cell temperature, the range between the maximum single cell temperature and the minimum single cell temperature, the maximum deviation between the single cell temperature and the average single cell temperature, the maximum single cell voltage value, the minimum single cell voltage value, the range between the maximum single cell voltage value and the minimum single cell voltage value, the maximum deviation between the single cell voltage value and the average single cell voltage value, and the insulation resistance value.
6. The training method of the thermal runaway identification model according to claim 1, characterized in that: The step of inputting the data set into an initial thermal runaway identification model to train the initial thermal runaway identification model to obtain a trained thermal runaway identification model includes: The data set is input into an initial thermal runaway recognition model, so as to train the initial thermal runaway recognition model according to the loss between the output of the initial thermal runaway recognition model and the label, and obtain a trained thermal runaway recognition model.
7. The method for training a thermal runaway identification model according to any one of claims 1 to 6, characterized in that: After the thermal runaway recognition model is trained, it also includes: Obtaining a recall rate according to a prediction result of the thermal runaway identification model; The thermal runaway identification model is evaluated according to the recall rate.
8. A thermal runaway identification method, characterized in that: include: Obtain vehicle-related data; The vehicle-related data is input into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model according to any one of claims 1 to 7.
9. A thermal runaway identification system, characterized in that: include: An acquisition module, used to obtain vehicle-related data; An identification module is used to input the vehicle-related data into a thermal runaway identification model to obtain a prediction result output by the thermal runaway identification model for characterizing the occurrence of thermal runaway, wherein the thermal runaway identification model is a thermal runaway identification model trained according to the training method for the thermal runaway identification model according to any one of claims 1 to 7.
10. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the thermal runaway identification method according to claim 8 is implemented.