A method, device and storage medium for optimizing the life extension of electric vehicle batteries

Abstract

The present application discloses an optimization method, device and storage medium for extending the battery life of electric vehicles, and relates to the field of electric vehicle batteries. The optimization method of the present application includes: inputting the original data of the vehicle and performing data preprocessing to obtain several charging sequences and discharging sequences of each vehicle on different days; then extracting feature data and extracting the battery life reduction label of the time window; obtaining the optimization result of extending the battery life through the battery life optimization mode optimization model based on the particle swarm algorithm. The present application proposes a new optimization method in the field of extending the battery life of electric vehicles, which expands the search form of the particle swarm algorithm from finding a single optimal solution to finding an optimal range, enhances the global search capability, and improves the robustness and flexibility of the solution; by connecting various sensors to monitor the battery status parameters, and combining artificial intelligence and other means to achieve battery management optimization, the battery performance is improved.

Description

Technical Field

[0001] The present application relates to the field of electric vehicle batteries, and in particular to a method, device and storage medium for optimizing the life extension of electric vehicle batteries. Background Art

[0002] The battery of an electric vehicle is a rechargeable energy storage system that is used to power the electric motor of the vehicle and is an important component of the electric vehicle. Electric vehicle batteries are mainly divided into two categories: batteries and fuel cells. Batteries are suitable for pure electric vehicles, while fuel cells are mostly used in hybrid vehicles. Electric vehicle batteries mainly have the performance characteristics of high energy density, environmental protection, renewable energy, fast charging, low maintenance cost and long life. With the continuous increase in demand for electric vehicles, consumers' requirements for battery performance, weight, cycle capacity and mileage are also continuously increasing. As one of the important indicators for measuring the performance and cost of electric vehicles, the extension of battery life is not only related to user experience, but also directly affects the market competitiveness and sustainable development of electric vehicles.

[0003] The key factors for extending the battery life of electric vehicles usually include: battery material innovation, battery management system (BMS) optimization, production process improvement, quality control system improvement, charging strategy adjustment, and improved operating environment. Among them, the battery management system is the key system responsible for monitoring and managing the battery status in electric vehicles. By optimizing the BMS algorithm, precise control of the battery charging and discharging process can be achieved, thereby extending the battery life.

[0004] The particle swarm algorithm simulates birds in a flock by designing a massless particle. The particle has only two properties: speed and position. The speed represents the speed of movement, and the position represents the direction of movement. In the particle swarm algorithm, birds are abstracted as particles without mass and volume, and extended to n-dimensional space. The position of any particle i in n-dimensional space can be represented as a vector , the flight speed can be expressed as a vector ,in, , are the position and flight speed of particle i in the jth dimension, Each particle has a fitness value determined by the objective function and knows its best position found so far (i.e., the local optimal position p_best) and its current position This can be regarded as the particle's own flight experience. In addition, each particle also knows the best position found by all particles in the entire group so far (i.e. the global optimal position g_best, the best value among all local optimal positions p_best), which can be regarded as the experience of the particle's companions. The particle decides the next movement based on its own experience and the best experience of its companions.

[0005] Therefore, in the face of people's urgent need to extend the battery life of electric vehicles, it is necessary to propose an optimization method for extending the operating life of electric vehicle batteries based on particle swarm algorithm to improve and extend the battery life by optimizing the BMS algorithm. Summary of the invention

[0006] The first aspect of the present application provides a method for optimizing the extension of the battery life of an electric vehicle, the method comprising the following steps:

[0007] Step 1: input the collected raw data of several vehicles and perform data preprocessing, including data cleaning and charge and discharge sequence division, to obtain several charge sequences and discharge sequences of each vehicle on different days;

[0008] The vehicle original data includes at least vehicle driving data and battery status data;

[0009] Step 2, extracting feature data from the original vehicle data after data preprocessing;

[0010] Extract characteristic data of several characteristic items of numerical type from the original data of the vehicle of each charging sequence / discharging sequence, and obtain the cumulative value or / mean value of the characteristic data of each characteristic item of each vehicle on each day based on the cumulative value or / mean value of the characteristic data of several same characteristic items on the same day; then, based on the set time window, take the cumulative value or / mean value of each day as the data point value, obtain the cumulative value or / mean value of the characteristic data of the current time window, and obtain the characteristic data of each characteristic item of each vehicle in each time window;

[0011] Step 3, extracting the battery life reduction label of the time window;

[0012] Based on a preset voltage range, the ampere-hour (ah) integral of each charging sequence / discharging sequence within the voltage range is calculated as a battery life label of each sequence;

[0013] The predicted life of each charging sequence / discharging sequence of each vehicle is extracted based on the trained multi-layer perceptron, and then the value capacity label of each sequence of the vehicle is obtained based on the difference between the battery life label of each sequence and the predicted life, that is, the corrected battery life is obtained;

[0014] The value capacity label of each vehicle on each day is obtained based on the average value of all sequences of the value capacity labels of the vehicle on the same day, and the battery life reduction label of each vehicle in each time window is obtained based on the difference between the value capacity labels of the first and last two days in the time window;

[0015] Step 4, taking the feature data of the time windows of different feature items of each vehicle and the initial value range of each feature item as the input data of the battery life optimization mode optimization model, the battery life optimization mode optimization model obtains the optimization result of the battery life extension of each feature item of each vehicle based on the particle swarm algorithm, and the optimization result includes: the feature item index, the median of the optimal range of the feature data corresponding to the feature item, and the length of the optimal range; and when calculating the global optimum and local optimum, the battery life optimization mode optimization model uses the average value of the battery life reduction label of a single vehicle in all time windows as the calculation basis, the global optimum basis is the average value of the battery life reduction label of all time windows of all vehicles, and the local optimum basis is the average value of the battery life reduction label of all time windows of the vehicle.

[0016] Furthermore, in step 1, the vehicle driving data includes but is not limited to: mileage, vehicle status, and vehicle speed.

[0017] Furthermore, in step 1, the battery status data includes, but is not limited to: charging status, total voltage, total current, maximum / minimum value of each voltage sensor, and maximum / minimum value of each temperature sensor.

[0018] Furthermore, in step 1, data preprocessing also includes: based on preset conversion rules, converting part of the character data in the vehicle's original data into numerical data, that is, converting the character data that can be converted; and in step 2, the feature data of several feature items of the extracted numerical type include feature items that can be converted from character data to numerical data.

[0019] Furthermore, in step 1, data cleaning refers to removing unreasonable data or abnormal data.

[0020] Furthermore, in step 1, the charge and discharge sequence is specifically divided as follows:

[0021] Each sequence is marked with a state mark and a sequence number, and the state mark is used to indicate whether the current sequence is discharging or charging;

[0022] The basis for dividing the charge and discharge sequence includes: charging state, time difference between adjacent data and voltage mutation.

[0023] Furthermore, step 1 also includes abnormal sequence elimination, including: eliminating sequences whose charge / discharge sequence data volume is lower than an expected value, eliminating charging sequences whose voltage stability duration exceeds a threshold, and eliminating charging sequences with voltage mutations.

[0024] Furthermore, in step 4, the optimization mode of the battery life optimization mode optimization model includes the life extension mode and the life reduction mode. In the life extension mode, the optimization goal is to obtain a ratio of battery life reduction labels that is less than the average ratio of all vehicles (the first global ratio) and a ratio of battery life reduction labels that is less than the average ratio of a single vehicle (the first local ratio); in the life reduction mode, the optimization goal is to obtain a ratio of battery life reduction labels that is greater than the average ratio of all vehicles (the second global ratio) and a ratio of battery life reduction labels that is greater than the average ratio of a single vehicle (the second local ratio). That is, the use of the optimization results obtained by the life reduction model is reflected in the user trying to avoid the optimization results of the life reduction model; for the life extension mode, the optimization results obtained by the life extension mode are adopted as much as possible.

[0025] Furthermore, in step 4, the fitness function of the particle swarm algorithm is set to:

[0026] (1) Life extension mode

[0027] Obtaining a first final scale based on a weighted sum of the first global scale and the first local scale;

[0028] Obtain the vehicle coverage rate of the optimal range of the current feature data, map it to between 0 and 1, and then perform a weighted sum with the first final ratio to obtain the adaptation value of the life extension mode;

[0029] (2) Shortened lifespan model

[0030] Obtaining a second final scale based on a weighted sum of the second global scale and the second local scale;

[0031] The vehicle coverage rate of the optimal range of the current feature data is obtained, and it is mapped to between 0 and 1, and then weighted and summed with the second final ratio to obtain the adaptation value of the life-shortening mode.

[0032] Furthermore, in step 4, a logarithmic operation is used to map the vehicle coverage rate to between 0 and 1.

[0033] Furthermore, step 4 also includes a training result data set consisting of optimization results of multiple feature items based on the battery life optimization mode optimization model during training; clustering the optimization results of the training result data set based on the battery life reduction label, and determining a group of feature groups that affect the extension of battery life based on the feature items included in each clustering result, so as to achieve battery life extension by a combination of multiple features.

[0034] A second aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to execute the method for optimizing the extension of the battery life of an electric vehicle as described in the present application.

[0035] A third 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 execute the method for optimizing the extension of the life of an electric vehicle battery as described in the present application.

[0036] The technical solution provided by this application brings at least the following beneficial effects:

[0037] This application proposes a new optimization method in the field of extending the battery life of electric vehicles. The method is based on an improved particle swarm algorithm (PSO) and is applied to the optimization problem of extending the battery operating life. The core of this method is to extract characteristic patterns that can significantly optimize the battery life from the actual usage data of the battery, aiming to extend the battery life through algorithm optimization. This application has made innovative improvements to the traditional particle swarm algorithm, expanding the algorithm's search form from finding a single optimal solution to finding an optimal range, enhancing the global search capability, and improving the robustness and flexibility of the solution, which is of great significance.

[0038] This application emphasizes the importance of data-driven in battery health management. By connecting various sensors to monitor battery status parameters and combining artificial intelligence and other means to process and analyze data, the battery management system (BMS) is optimized to improve battery performance. In the context of increasing attention to carbon emissions in the transportation industry, the optimization method of this application not only helps to reduce dependence on traditional fossil energy, but also effectively reduces the cost of using electric vehicles, achieving both environmental and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] 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:

[0040] Figure 1 A flow chart of a method for optimizing the life extension of an electric vehicle battery provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of the structure of an electronic device provided for an application embodiment. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions of the embodiments of the present application will be described in detail and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the embodiments described with reference to the drawings are exemplary and are intended to be used to explain the present application, and cannot be understood as limiting the present application.

[0043] See also Figure 1 The present application provides a method for optimizing the life extension of an electric vehicle battery, comprising the following steps:

[0044] Step S1, inputting vehicle raw data and performing data preprocessing to obtain a number of charging sequences and discharging sequences for each vehicle on different days;

[0045] In this step, the vehicle original data includes at least vehicle driving data and battery status data;

[0046] Data preprocessing includes data cleaning and charge-discharge sequence division, so as to obtain several charge and discharge sequences for each vehicle on different days;

[0047] In one embodiment, the vehicle raw data includes mileage, charging status, vehicle status, vehicle speed, total voltage, total current, maximum / minimum value of each voltage sensor, maximum / minimum value of each temperature sensor, and the like.

[0048] For another example, the collected vehicle raw data includes multiple categories such as battery initial life, vehicle environment, battery status, user driving behavior, vehicle use status, etc. Each category of data includes several data indicators. For example, the vehicle environment includes charging temperature, discharging temperature, the month of the vehicle, uphill or downhill, urban area or highway, etc.; the vehicle use status may include: vehicle running time, cumulative kilometers, average speed, average mileage, driving acceleration, etc.; user driving behavior includes the number of braking times, the number of acceleration times, the maximum depth of the brake pedal, the average depth of the brake pedal, the maximum depth of the accelerator pedal, the average depth of the accelerator pedal, etc. The battery status may include: the number of charges, the number of discharges, the charging time, the discharging time, the open circuit voltage, the charging current, the discharging current, the charging power, the charging start SOC (State of Charge, battery state of charge), the charging end SOC, the charging SOC range, the discharging start SOC, the discharging end SOC, the discharging SOC range, etc.

[0049] In one embodiment, the charge and discharge sequence division used in step S1 specifically includes:

[0050] Label each sequence with a state and serial number , where state is 1 Indicates that the sequence is a discharge sequence, state 2 Indicates that the sequence is a charging sequence; each time a new sequence appears, the sequence number is incremented by 1.

[0051] The basis for the division of charge and discharge sequences includes the charge state (charge_state), the time difference between adjacent data, and whether there is a voltage mutation. Specifically, in general, the basis for judging whether a new charge and discharge sequence has appeared is the charge state. When the charge state has not changed, but the time interval between adjacent data is greater than one day, it can be considered that there is a new charge and discharge sequence. If it is in the discharge state, and the time interval between adjacent data is less than one day but there is a discharge mutation, it can be considered that there is a new sequence.

[0052] In one embodiment, the data preprocessing in step S1 also includes: for the character data in the vehicle original data that can be converted into numerical data, the character data can be converted into numerical data based on preset conversion rules, so as to enrich the optimization object.

[0053] In addition, in order to further ensure the trainability of the data, the data preprocessing of step S1 also includes the elimination of abnormal sequences. That is, for the charging and discharging sequence, if there is a sequence containing too little data (for example, occasional discharge may occur near the end of charging), this abnormal short sequence needs to be screened out; the situation in which the voltage remains unchanged for a long time or the voltage changes suddenly in the charging sequence should also be regarded as an abnormal sequence and screened out.

[0054] Step S2, extracting feature data from the original vehicle data after data preprocessing;

[0055] Feature data of several numerical feature items are extracted from the original vehicle data of each charging sequence / discharging sequence, and the cumulative value or / mean value of the feature data of several same feature items on the same day is used to obtain the cumulative value or / mean value of the feature data of each feature item of each vehicle on each day; then based on the set time window, the cumulative value or / mean value of each day is used as the data point value to obtain the cumulative value or / mean value of the feature data of the current time window, so as to obtain the feature data of each feature item of each vehicle in each time window.

[0056] Step S2 is based on the needs of the battery life optimization mode optimization model of the present application, extracts various characteristics of the vehicle, generates sequence feature data of all vehicles extracted in units of charging and discharging sequences based on the original vehicle data after data preprocessing, counts multiple sequence features of all vehicles in the specified time window, and generates cumulative features, mean features, variance features, etc. in the specified time window. The scale of the time window can be set based on the actual application scenario, for example, to 50 days, 90 days, etc.

[0057] Step S3, extracting the battery life reduction tag of the time window;

[0058] (1) Based on the preset voltage range , calculate the ampere-hour (ah) integral of each charging sequence / discharging sequence within the voltage range as the battery life label of each sequence, the formula is as follows:

[0059]

[0060] in, For sequence (charge sequence or discharge sequence) battery life label, For the moment The current size when Characterize the starting calculation voltage, Characterize the termination calculation voltage.

[0061] (2) Based on the trained multi-layer perceptron (MLP), the predicted life of each charging sequence / discharging sequence of each vehicle is extracted, and then the value capacity label of each sequence of the vehicle is obtained based on the difference between the battery life label of each sequence and the predicted life;

[0062] (3) The value capacity label of each vehicle on each day is obtained based on the average value of all sequences of the value capacity labels of the vehicle on the same day. Then, the battery life reduction label of each vehicle in each time window is obtained based on the difference between the value capacity labels of the first and last two days in the time window.

[0063] In step S3, the predicted life of each charging sequence / discharging sequence of each vehicle is extracted. A basic multi-layer perceptron neural network model can be used, for example, a model structure including 5 hidden layers. The number of hidden layers can be selected based on actual computing resources, scenarios, etc. The input data of the multi-layer perceptron is the feature data extracted in step S2. During training, the battery life label is the target value, and the purpose of training is to make the predicted life output by the multilayer perceptron as close as possible to the corresponding battery life label, so as to extract the value capacity label that is not affected by short-term temperature and current based on the trained multilayer perceptron.

[0064] Step S4, obtaining an optimization result of extending the battery life through a battery life optimization mode optimization model based on a particle swarm algorithm;

[0065] The characteristic data of the time windows of different characteristic items of each vehicle and the initial value range of each characteristic item are used as input data of the battery life optimization mode optimization model. The battery life optimization mode optimization model obtains the optimization result of the battery life extension of each characteristic item of each vehicle based on the particle swarm algorithm. The optimization result includes: the characteristic item index, the median of the optimal range of the characteristic data corresponding to the characteristic item, and the length of the optimal range; and when calculating the global optimum and local optimum, the battery life optimization mode optimization model uses the average value of the battery life reduction label of a single vehicle in all time windows as the calculation basis. The global optimum basis is the average value of the battery life reduction label of all time windows of all vehicles, and the local optimum basis is the average value of the battery life reduction label of all time windows of the vehicle.

[0066] In one embodiment, in step S4, the optimization mode of the battery life optimization mode optimization model includes a life extension mode and a life reduction mode: in the life extension mode, the optimization goal is to obtain a ratio that is less than the average ratio of battery life reduction of all vehicles (the first global ratio) and a ratio that is less than the average ratio of battery life reduction labels of a single vehicle (the first local ratio); in the life reduction mode, the optimization goal is to obtain a ratio that is greater than the average ratio of battery life reduction labels of all vehicles (the second global ratio) and a ratio that is greater than the average ratio of battery life reduction labels of a single vehicle (the second local ratio). That is, the use of the optimization results obtained by the life reduction model is reflected in that the user tries to avoid the optimization results of the life reduction model; for the life extension mode, the optimization results obtained by the life extension mode are adopted as much as possible.

[0067] In step S4 of the present application, the battery life optimization mode optimization model based on the improved particle swarm algorithm searches for characteristic modes that can optimize battery life from characteristic data. The improvement strategy includes designing an optimal range search method and designing a fitness function.

[0068] 1) Optimal range search method

[0069] The original particle swarm algorithm searches for the optimal value, while the search for the optimization mode of the battery operating life requires finding an optimal range. Therefore, the embodiment of the present application improves the particle swarm algorithm from searching for the optimal value to searching for the optimal range. The traditional search for the optimal range is to search for two values ​​at the same time, one is the left boundary of the optimal range, and the other is the right boundary of the optimal range. However, this will lead to the problem of crossing the boundary, that is, when the particle updates the speed and position, the left boundary of the range found is greater than the right boundary. Therefore, in order to avoid the occurrence of this situation, the present application uses the median value of the range and the length of the range to replace the left boundary and the right boundary, that is, when searching for the optimal range, the median value of the optimal range and the length of the optimal range are searched.

[0070] 2) Fitness function

[0071] The fitness function is used to judge the quality of the found range (determined based on the median value of the optimal range and the length of the optimal range).

[0072] First, the range to be judged (i.e. the currently found range) is passed in, and the data satisfying the range (the characteristic data of the time window) is calculated. After obtaining the data satisfying the range, the range is evaluated from both global and local aspects.

[0073] From a global perspective, the data that meets the range is expected to be smaller than the average value of the battery life reduction value label of all data. This embodiment uses the following formula to calculate the global ratio:

[0074]

[0075] in, Indicates the characteristic data that meets the range, The battery life reduction value label representing the characteristic data that meets the range, Indicates the average value of the battery life reduction value label of all data, Indicates the number of sequences of all vehicles in the data that meet the range. That is, in the calculation formula of the global proportion, the numerator on the right side of the equal sign represents the number of sequences in the data that meet the range that are less than the overall mean, and the denominator on the right side of the equal sign represents the number of sequences in the data that meet the range.

[0076] From a local perspective, for each vehicle, the data that is expected to be found is data that is less than the average value of the battery life degradation value label of the vehicle. This embodiment uses the following formula to calculate the local proportion of each vehicle:

[0077]

[0078] in, Represents vehicles in the data that meet the range , Represents vehicles in the data that meet the range The battery life decreases value label, Indicates vehicle The battery life decreases value label, Indicates the vehicles in the data that meet the range The number of sequences, that is, in the calculation formula of the local proportion, the numerator on the right side of the equal sign represents the number of vehicles in the data that meet the range Smaller than vehicle The number of sequences of the average value of the battery life reduction value label, the denominator on the right side of the equal sign represents the number of vehicles in the data that meet the range The number of sequences.

[0079] Next, the local proportions of all vehicles are calculated according to the weighted average to obtain the final local proportion:

[0080]

[0081] in, Indicates the weights corresponding to different vehicles. The weights can be calculated based on the vehicle The data is obtained by the proportion of all vehicle data.

[0082] Finally, the final ratio can be obtained by combining the global ratio and the local ratio in a certain ratio, as shown in the following formula:

[0083]

[0084] in, , are the weights of the global scale and the local scale, respectively, and satisfy .

[0085] In the embodiment of the present application, in addition to considering the battery life degradation value label, the vehicle coverage also needs to be considered, and the two need to be effectively combined. Since the final ratio is a number between 0 and 1, and the number of vehicles is often relatively large, the two need to be aligned. The operation taken for this is to take the logarithm of the number of vehicles so that it can be taken between 0 and 1, as shown in the following formula:

[0086]

[0087] in, represents the final vehicle coverage, The number of vehicles covered by the table, parameters For the A parameter that is scaled between 1 and 2, depending on the number of vehicles.

[0088] Finally, the final lifespan reduction ratio and vehicle coverage rate are integrated according to a certain ratio to obtain the final adaptation value: ,in, , The final ratios are and vehicle coverage The weight of .

[0089] The above formula is the calculation method for the fitness function of the life extension mode, while for the life reduction mode, it is necessary to replace the greater than in the calculation formula of the global proportion and the local proportion with less than.

[0090] In addition, step S4 may also include: a training result data set composed of optimization results of multiple feature items of the battery life optimization mode optimization model during training; clustering each optimization result of the training result data set based on the battery life reduction amount, and determining a group of feature groups that affect the extension of battery life according to the feature items included in each clustering result, so as to achieve battery life extension by a combination of multiple features.

[0091] In one embodiment, the training and testing of the battery life optimization mode optimization model can be implemented through the following steps:

[0092] (1) Divide the original vehicle data set after data preprocessing into a training set, a validation set, and a test set;

[0093] (2) Initialize the parameters related to particle swarm training, including the initialization time window size, search mode (life extension / life reduction), label, feature name (feature list / feature name / feature list / feature minimum value / feature maximum value);

[0094] (3) Read the original vehicle data set, that is, read the vehicles in it, return the vehicle list index, and obtain vehicle data based on the vehicle index; obtain label data; extract features from several aspects such as battery initial life, vehicle environment, battery status, user driving behavior, and vehicle usage status;

[0095] (4) Calculate the fitness function value, input the optimal range median and range length, calculate the global ratio and local ratio, and analyze from the two perspectives of battery life degradation value label and vehicle coverage rate;

[0096] (5) Train the particle swarm, specify relevant training parameters to train the particle swarm, and screen the data that meets the conditions based on the training results. The input parameters include the number of optimizations for each training, whether there is an initial value, the initial value, the state selection of the training set / validation set / test set, the output optimization range, and the index of the data within the range;

[0097] (6) Verify the particle swarm, perform relevant statistics on the left and right boundaries of the input range on the test set or validation set, and return the index of the data in the corresponding range;

[0098] (7) Statistics of training / validation / test data that meet the optimization range: Input the index of the data that meets the optimization range and the status of the training set / validation set / test set, and count the training / validation / test data that meet the optimization range. If you are looking for a life extension model, you need to count the proportion that is less than the overall mean and the proportion of the average of a single vehicle; if you are looking for a life reduction model, you need to count the proportion that is greater than the overall mean and the proportion of the average of a single vehicle. Output the proportion that meets the conditions and the statistical results of the vehicle coverage;

[0099] (8) Store data with good training set performance for statistical coverage-related results;

[0100] (9) Cluster the results of multiple optimizations to obtain robust performance. Read the feature ranges saved during multiple trainings and cluster them. Similar features will be clustered into the same class, and the center of this class is considered to be more robust. Input the state selection of the training set / validation set / test set, the number of clusters, and output the minimum number of clusters;

[0101] (10) Read the clustering results to verify the clustering results on the validation set, input the minimum number of clusters, and output the clustering result list;

[0102] (11) Filter out good clustering results based on the battery life degradation value labels of each cluster center or the average value of the battery life degradation value labels of the clusters and save them;

[0103] (12) Conduct particle swarm testing and read the best clustering result in the validation set for testing on the test set;

[0104] (13) Store bad feature names. If a feature is difficult to train or performs poorly in the training set, store it in the bad feature category.

[0105] (14) Draw a diagram to verify the surrounding area of ​​the optimization range;

[0106] (15) Perform continuous fitting on the optimization range, and draw and save the fitting results;

[0107] (16) Calculate the overall mean, the mean of the data within the range, and the covered vehicles according to the optimization range;

[0108] (17) Save the test results and check whether the performance of the optimal verification result on the test set can reach the preset threshold. If so, it is regarded as the final result, and relevant statistics are performed on all the data to verify the performance of the method proposed in the embodiment of the present application.

[0109] In an exemplary embodiment, the embodiment of the present application also provides an electronic device, which may include: a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the program, a method for optimizing the extension of the battery life of an electric vehicle provided in the above embodiment is implemented.

[0110] Specifically, Figure 2 As shown, the electronic device includes:

[0111] A processor, used to execute a computer program stored in the memory to implement a method for optimizing the extension of the battery life of an electric vehicle provided in the above embodiment;

[0112] A communication interface for communication between the memory and the processor;

[0113] Memory is used to store computer programs that can be run on the processor.

[0114] The memory may include high-speed random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage. If the memory, processor and communication interface are implemented independently, the communication interface, memory and processor can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0115] Optionally, in a specific implementation, if the memory, processor and communication interface are integrated on a chip, the memory, processor and communication interface can communicate with each other through an internal interface. The processor may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0116] This embodiment also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is loaded and executed by a processor of a computer device to implement the electric vehicle battery life extension optimization method provided in this embodiment.

[0117] In a possible implementation, the computer-readable storage medium may be a read-only memory, a random access memory, a read-only optical disk, a magnetic tape, an optical data storage device, and the like.

[0118] 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 do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more 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.

[0119] In addition, the terms "first", "second", etc. 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. Thus, a feature defined as "first", "second", etc. may explicitly or implicitly include at least one of the features.

[0120] Any process or method description described in a flowchart or otherwise in this specification may be understood to represent a module, fragment or portion of code that includes one or more 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.

[0121] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by 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 for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0122] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing the battery life extension of an electric vehicle, characterized in that: The following steps are involved: Step 1: input the collected raw data of several vehicles and perform data preprocessing, including data cleaning and charge and discharge sequence division, to obtain several charge sequences and discharge sequences of each vehicle on different days; The vehicle original data includes at least vehicle driving data and battery status data; Step 2, extracting feature data from the original vehicle data after data preprocessing; Extract characteristic data of several characteristic items of numerical type from the original data of the vehicle of each charging sequence / discharging sequence, and obtain the cumulative value or / mean value of the characteristic data of each characteristic item of each vehicle on each day based on the cumulative value or / mean value of the characteristic data of several same characteristic items on the same day; then, based on the set time window, take the cumulative value or / mean value of each day as the data point value, obtain the cumulative value or / mean value of the characteristic data of the current time window, and obtain the characteristic data of each characteristic item of each vehicle in each time window; Step 3, extracting the battery life reduction label of the time window; Based on a preset voltage range, the ampere-hour integral of each charging sequence / discharging sequence within the voltage range is calculated as a battery life label of each sequence; The predicted life of each charging sequence / discharging sequence of each vehicle is extracted based on the trained multi-layer perceptron, and then the value capacity label of each sequence of the vehicle is obtained based on the difference between the battery life label of each sequence and the predicted life, that is, the corrected battery life is obtained; The value capacity label of each vehicle on each day is obtained based on the average value of all sequences of the value capacity labels of the vehicle on the same day, and the battery life reduction label of each vehicle in each time window is obtained based on the difference between the value capacity labels of the first and last two days in the time window; Step 4, taking the feature data of the time windows of different feature items of each vehicle and the initial value range of each feature item as the input data of the battery life optimization mode optimization model, the battery life optimization mode optimization model obtains the optimization result of the battery life extension of each feature item of each vehicle based on the particle swarm algorithm, and the optimization result includes: the feature item index, the median of the optimal range of the feature data corresponding to the feature item, and the length of the optimal range; and when calculating the global optimum and local optimum, the battery life optimization mode optimization model uses the average value of the battery life reduction label of a single vehicle in all time windows as the calculation basis, the global optimum basis is the average value of the battery life reduction label of all time windows of all vehicles, and the local optimum basis is the average value of the battery life reduction label of all time windows of the vehicle.

2. The method for optimizing the life extension of an electric vehicle battery as claimed in claim 1, characterized in that: In step 1, the vehicle driving data includes: driving mileage, vehicle status and / or vehicle speed; The battery status data includes: charging status, total voltage, total current, maximum / minimum value of each voltage sensor, and maximum / minimum value of each temperature sensor.

3. The method for optimizing the battery life extension of an electric vehicle as claimed in claim 1, characterized in that: In step 1, data preprocessing also includes: based on preset conversion rules, converting part of the character data in the vehicle original data into numerical data; and in step 2, the feature data of several feature items of the extracted numerical type include feature items that can be converted from character data to numerical data.

4. The method for optimizing the life extension of an electric vehicle battery as claimed in claim 1, characterized in that: In step 1, the charge and discharge sequence is divided into: A state mark and a sequence number are marked for each sequence, and the state mark is used to indicate whether the current sequence is discharging or charging; The basis for dividing the charge and discharge sequence includes: charging state, time difference between adjacent data and voltage mutation.

5. The method for optimizing the life extension of an electric vehicle battery as claimed in claim 1, characterized in that: Step 1 also includes abnormal sequence elimination: eliminating the sequence whose charge / discharge sequence data volume is lower than the expected value, eliminating the charging sequence whose voltage stability duration exceeds the threshold, and eliminating the charging sequence with voltage mutation.

6. The method for optimizing the life extension of an electric vehicle battery as claimed in claim 1, characterized in that: In step 4, the optimization mode of the battery life optimization mode optimization model includes a life extension mode and a life reduction mode; In the life extension mode, the optimization goal is to obtain a ratio that is less than the average value of the battery life reduction labels of all vehicles to obtain a first global ratio, and to obtain a ratio that is less than the average value of the battery life reduction labels of a single vehicle to obtain a first local ratio; In the shortened life mode, the optimization goal is to obtain a ratio greater than the average value of the battery life reduction labels of all vehicles to obtain a second global ratio, and to obtain a ratio greater than the average value of the battery life reduction labels of a single vehicle to obtain a second local ratio.

7. The method for optimizing the extension of the battery life of an electric vehicle as claimed in claim 6, characterized in that: In step 4, the fitness function of the particle swarm algorithm is set as: (1) Life extension mode: Obtaining a first final scale based on a weighted sum of the first global scale and the first local scale; Obtain the vehicle coverage rate of the optimal range of the current feature data, map it to between 0 and 1, and then perform a weighted sum with the first final ratio to obtain the adaptation value of the life extension mode; (2) Shortened lifespan model: Obtaining a second final scale based on a weighted sum of the second global scale and the second local scale; The vehicle coverage rate of the optimal range of the current feature data is obtained, and it is mapped to between 0 and 1, and then weighted and summed with the second final ratio to obtain the adaptation value of the life-shortening mode.

8. The method for optimizing the extension of the battery life of an electric vehicle as claimed in claim 1, characterized in that: Step 4 also includes a training result data set composed of optimization results of multiple feature items of the battery life optimization mode optimization model during training; clustering the optimization results of the training result data set based on the battery life reduction label, and determining a group of feature groups that affect the extension of battery life based on the feature items included in each clustering result.

9. An electronic device, 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 a method for optimizing the extension of the battery life of an electric vehicle as claimed in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement a method for optimizing the extension of the battery life of an electric vehicle as claimed in any one of claims 1 to 8.

Images