Electric vehicle battery failure early warning method and device, storage medium and program product

Abstract

The application relates to an electric vehicle battery failure early warning method and device, a storage medium and a program product. The method comprises the following steps: acquiring target index data of a target electric vehicle under a key battery index in a current period; wherein the key battery index is an index related to vehicle battery failure selected from each candidate battery index according to a cross-entropy loss function and an original data set; each sample data in the original data set comprises sample index data of a sample electric vehicle under each candidate battery index and a sample category label of the sample electric vehicle; determining a battery aging factor of the target electric vehicle according to an initial capacity and a current capacity of the battery of the target electric vehicle; and performing vehicle battery failure early warning on the target electric vehicle based on a Bayesian network model according to the target index data and the battery aging factor. The method can realize accurate early warning of vehicle battery failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle diagnosis, and in particular to an electric vehicle battery fault early warning method and device, a storage medium and a program product. BACKGROUND

[0002] With the increasing popularity of new energy vehicles, promoting the development of new energy vehicles is an important means to promote "carbon neutralization" and "carbon peak". Lithium ion batteries are widely used in various fields such as new energy vehicles due to their high energy density and long service life. However, lithium ion batteries also have some potential safety risks, such as internal short circuit, thermal runaway, and lithium extraction.

[0003] In the prior art, most lithium ion battery safety warning technologies collect battery voltage, current and other data, and use machine learning algorithms to perform safety warning on the battery. However, this method does not take into account factors such as vehicle historical behavior data and battery aging, and therefore cannot accurately predict faults. SUMMARY

[0004] Therefore, it is necessary to provide an electric vehicle battery fault early warning method, device, storage medium and program product to accurately predict vehicle battery faults.

[0005] In a first aspect, the present application provides an electric vehicle battery fault early warning method, which comprises:

[0006] Obtaining target indicator data of a target electric vehicle under a key battery indicator in a current period; wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample class label of the sample electric vehicle;

[0007] Determining a battery aging factor of the target electric vehicle according to an initial capacity of the battery of the target electric vehicle and a current capacity of the battery;

[0008] Based on a Bayesian network model, the target electric vehicle is subjected to vehicle battery fault early warning according to the target indicator data and the battery aging factor.

[0009] In one embodiment, the key battery indicator is determined as follows:

[0010] According to the cross-entropy loss function and the original data set, the correlation between each candidate battery indicator and the vehicle battery fault is determined; and the key battery indicator is selected from each candidate battery indicator according to the correlation between each candidate battery indicator and the vehicle battery fault.

[0011] In one embodiment, the correlation between each candidate battery indicator and the vehicle battery failure is determined according to a cross-entropy loss function and the original data set, including:

[0012] The original data set is divided into a plurality of original sample subsets with the same number of samples; each original sample subset is processed using a cross-entropy loss function to obtain an initial prediction accuracy; for each candidate battery indicator, sample indicator data corresponding to the candidate battery indicator is removed from the sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery indicator; and the correlation between the candidate battery indicator and the vehicle battery failure is determined according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset.

[0013] In one embodiment, each original sample subset is processed using a cross-entropy loss function to obtain an initial prediction accuracy, including:

[0014] For each original sample subset, the original sample subset is processed using a cross-entropy loss function to obtain a first prediction accuracy of the original sample subset; and the average of the first prediction accuracies of each original sample subset is taken as the initial prediction accuracy.

[0015] In one embodiment, the correlation between each candidate battery indicator and the vehicle battery failure is determined according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset, including:

[0016] For each new sample subset, the new sample subset is processed using a cross-entropy loss function to obtain a second prediction accuracy of the new sample subset, and the difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy is taken as the accuracy deviation of the original sample subset corresponding to the candidate battery indicator with respect to the candidate battery indicator; and the correlation between the candidate battery indicator and the vehicle battery failure is determined according to the accuracy deviation of the original sample subset corresponding to the candidate battery indicator with respect to the candidate battery indicator.

[0017] In one embodiment, the target electric vehicle is given a vehicle battery failure warning based on a Bayesian network model according to target indicator data and a battery aging factor, including:

[0018] The target indicator data and the battery aging factor are input into the Bayesian network model to obtain a target failure probability of the target electric vehicle; a target failure level corresponding to the target failure probability is determined based on a corresponding relationship between the failure probability and the failure level; and the target electric vehicle is given a vehicle battery failure warning according to the target failure level.

[0019] In a second aspect, the application further provides an electric vehicle battery failure warning device, which includes:

[0020] The acquisition module is configured to acquire target indicator data of the target electric vehicle under a key battery indicator in a current period; the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle;

[0021] The factor determination module is configured to determine a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery.

[0022] The early warning module is configured to perform vehicle battery failure early warning on the target electric vehicle based on the Bayesian network model, the target indicator data and the battery aging factor.

[0023] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0024] The acquisition module is configured to acquire target indicator data of the target electric vehicle under a key battery indicator in a current period; the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle.

[0025] The factor determination module is configured to determine a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery.

[0026] The early warning module is configured to perform vehicle battery failure early warning on the target electric vehicle based on the Bayesian network model, the target indicator data and the battery aging factor.

[0027] In a fourth aspect, the present application further provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the following steps:

[0028] The acquisition module is configured to acquire target indicator data of the target electric vehicle under a key battery indicator in a current period; the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle.

[0029] The factor determination module is configured to determine a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery.

[0030] The target electric vehicle is warned of battery failure based on the Bayesian network model, the target indicator data and the battery aging factor.

[0031] In a fifth aspect, the present application also provides a computer program product comprising a computer program which, when executed by a processor, implements the following steps:

[0032] Target indicator data of the target electric vehicle in a current period under a key battery indicator is acquired, wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set, and each sample data in the original data set comprises sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle.

[0033] A battery aging factor of the target electric vehicle is determined according to an initial capacity of the battery of the target electric vehicle and a current capacity of the battery.

[0034] The target electric vehicle is warned of battery failure based on the Bayesian network model, the target indicator data and the battery aging factor.

[0035] The electric vehicle battery failure warning method, device, storage medium and program product, by acquiring target indicator data of the target electric vehicle in a current period under a key battery indicator, and determining a battery aging factor of the target electric vehicle according to an initial capacity of the battery of the target electric vehicle and a current capacity of the battery, wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set, and further, the target electric vehicle is warned of battery failure based on the Bayesian network model, the target indicator data and the battery aging factor. The scheme provides a basis for selecting the key battery indicator from each candidate battery indicator by introducing the cross-entropy loss function. Furthermore, by combining the target indicator data and the battery aging factor, the vehicle battery condition is considered comprehensively, and finally the vehicle battery failure is accurately warned. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0037] Figure 1 An application environment diagram of the electric vehicle battery failure warning method provided in the embodiments of the present application;

[0038] Figure 2 This is a schematic diagram of a battery fault early warning process for electric vehicles provided in the embodiments of this application;

[0039] Figure 3 This is a schematic diagram of a process for determining key battery indicators provided in an embodiment of this application;

[0040] Figure 4 This is a schematic diagram of a process for determining relevance provided in an embodiment of this application;

[0041] Figure 5 This is a flowchart illustrating another electric vehicle battery fault early warning method provided in the embodiments of this application;

[0042] Figure 6 This is a structural block diagram of an electric vehicle battery fault early warning device provided in the embodiments of this application;

[0043] Figure 7 This is an internal structural diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0045] The electric vehicle battery fault early warning method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located in the cloud or on other network servers. Terminal 102, installed on the target electric vehicle, monitors various indicators of the target electric vehicle and sends the monitored indicator data to server 104; subsequently, server 104 can provide early warning of vehicle battery failure based on the battery aging factor and indicator data of the target electric vehicle. Terminal 102 can be, but is not limited to, various monitoring devices and sensors. Server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0046] In one exemplary embodiment, such as Figure 2 As shown, a method for early warning of battery faults in electric vehicles is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation may include the following steps:

[0047] S201, obtaining target indicator data of the target electric vehicle under the key battery indicator in a current period.

[0048] The key battery indicator can be selected from candidate battery indicators according to a cross-entropy loss function and an original data set, and each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle. Optionally, the current period can represent a latest charge-discharge period.

[0049] For example, the key battery indicators can include battery indicators, cell indicators, and vehicle indicators; wherein the battery in the target electric vehicle includes a plurality of cells, the battery indicators can include discharge capacity, maximum / minimum current, slow charging times, fast charging times, low-temperature large-current charging times, high-temperature large-current charging times, slow charging ampere-hour number, fast charging ampere-hour number, number of times of charging and discharging period greater than 20% / 40% (SOC (State of Charge) change range of the charging and discharging period greater than 20% / 40%), maximum / minimum voltage at the start / stop of charging, number of times of charging in different SOC change ranges during the charging and discharging period, charging cutoff SOC (number of times of charging in different SOC intervals, number of times of charging with average charging current falling in different current interval ranges, number of times of charging with maximum charging current falling in different current interval ranges, number of times of charging with the highest temperature during the charging process falling in different temperature interval ranges, number of times of charging with the maximum temperature difference during the charging process falling in different temperature interval ranges, number of times of charging with the lowest temperature during the charging process falling in different temperature interval ranges, number of times of charging with the maximum charging current during the charging and the highest temperature corresponding to the maximum charging current falling in different current interval and temperature interval combinations, number of times of charging with the maximum charging current during the charging and the lowest temperature corresponding to the maximum charging current falling in different current interval and temperature interval combinations, number of times of charging with the highest single cell battery voltage at the charging cutoff falling in different voltage interval ranges, number of times of charging with the voltage difference between the highest single cell battery voltage and the lowest single cell battery voltage at the charging cutoff falling in different voltage interval ranges, and number of times of charging with the average charging current and temperature combination falling in different charging current interval ranges and temperature interval range combinations during the charging through different SOC intervals (10% interval range), etc.; the cell indicators can include number of times of charging with the single cell charging cutoff voltage falling in different voltage interval ranges, number of times of charging with the voltage difference (single cell voltage-average voltage) at the single cell charging cutoff falling in different voltage interval ranges, amount of change (initial voltage difference-current voltage difference) of the difference between the single cell voltage and the average voltage in the full life cycle of the battery in different SOC interval ranges, difference between the single cell voltage and the average voltage in the current period in different voltage interval ranges, difference between the capacity (dQ / dV) of the single cell in different voltage ranges and the average battery capacity in the current period, amount of change (initial voltage difference-current voltage difference) of the difference between the single cell voltage and the average voltage in the full life cycle of the battery in different SOC interval ranges, and difference between the capacity (dQ / dV) of the single cell in different voltage ranges and the average battery capacity in the current period; and the vehicle indicators can include total mileage of the vehicle, daily average mileage of the vehicle, whether there is a collision, maximum continuous running time of the vehicle, and average running temperature of the vehicle, etc.

[0050] Exemplarily, the terminal 102 monitors the indicators of the target electric vehicle in real time, and sends the monitored indicator data to the server 104 through the communication module; then, the server 104 can receive the monitoring data of the target electric vehicle, and directly obtain the target indicator data of the target electric vehicle under the key battery indicators in the current period.

[0051] S202, determining a battery aging factor of the target electric vehicle according to the initial battery capacity and the current battery capacity of the target electric vehicle.

[0052] The initial battery capacity can represent the standard battery capacity of the target electric vehicle when it leaves the factory, that is, the rated capacity of the battery in the brand-new state; the current battery capacity can represent the battery capacity of the target electric vehicle in the current period; and the battery aging factor can represent a parameter of the aging degree of the battery in the use process.

[0053] Exemplarily, the battery aging factor λ of the target electric vehicle can be calculated according to the initial battery capacity c0 and the current battery capacity c1 of the target electric vehicle, which can be specifically represented as:

[0054]

[0055] S203, performing vehicle battery fault early warning on the target electric vehicle based on the Bayesian network model according to the target indicator data and the battery aging factor.

[0056] The Bayesian network model is a model for evaluating the vehicle battery fault according to the target indicator data and the battery aging factor based on the Bayesian total probability formula; the fault early warning can include fault level, predicted fault time and fault related position, etc.

[0057] Exemplarily, the battery aging factor λ can be fused into the Bayesian total probability formula of the key battery indicators to the fault, which can be represented as:

[0058]

[0059] Wherein, n can represent the number of key battery indicators; x i may represent the i-th key battery indicator; Y can represent the vehicle label corresponding to the i-th key battery indicator, including two types of normal and fault; is the prior probability of Y, is the prior probability of Y after considering .

[0060] The above electric vehicle battery fault early warning method, by acquiring target index data of the target electric vehicle under the key battery index in the current period, and determining the battery aging factor of the target electric vehicle according to the initial capacity and the current capacity of the battery of the target electric vehicle; wherein the key battery index is an index related to vehicle battery failure selected from each candidate battery index according to the cross-entropy loss function and the original data set; further, based on the Bayesian network model, the target index data and the battery aging factor, the target electric vehicle is subjected to vehicle battery fault early warning. The scheme provides a basis for selecting key battery indicators from various candidate battery indicators by introducing a cross-entropy loss function. Furthermore, by combining the target index data and the battery aging factor, the vehicle battery condition is considered comprehensively, and finally the vehicle battery failure is accurately warned.

[0061] On the basis of the above-mentioned embodiments, the embodiment of the present application explains and describes the above-mentioned embodiment S201 in detail. Specifically, the process of determining the key battery index in the embodiment of the present application is shown in the following steps: Figure 3

[0062] S301, determining the correlation between each candidate battery index and vehicle battery failure according to the cross-entropy loss function and the original data set.

[0063] It should be noted that the original data set can be selected from the total data of each electric vehicle in the whole life cycle; for example, the SOC change range in each charge-discharge period can be judged, and the charge-discharge period whose SOC change range is less than 20% is removed, that is, the original data set can be obtained.

[0064] For example, the original data set can be input into the cross-entropy loss function, and each candidate battery index in the original data set can be analyzed by the cross-entropy loss function to determine the cross-entropy loss corresponding to each candidate battery index. Further, based on the cross-entropy loss corresponding to each candidate battery index, the correlation between each candidate battery index and vehicle battery failure can be determined.

[0065] S302, selecting the key battery index from each candidate battery index according to the correlation between each candidate battery index and vehicle battery failure.

[0066] For example, the correlation between each candidate battery index and vehicle battery failure can be sorted, and the index with high correlation can be selected as the key battery index. For example, each candidate battery index can be screened according to a predetermined proportion, such as 60% of the candidate battery index can be determined as the key battery index; the index with a correlation greater than a correlation threshold can also be determined as the key battery index, and the way of screening the key battery index is not limited in the embodiment of the present application.​

[0067] In the embodiment of the present application, by introducing the cross-entropy loss function, a tool is provided for determining the correlation between each candidate battery index and the vehicle battery failure, thereby laying a foundation for accurately determining the key battery index.

[0068] On the basis of the above-mentioned embodiment, the embodiment of the present application explains and describes the above-mentioned embodiment S301 in detail. Specifically, the embodiment of the present application involves the process of determining the correlation between each candidate battery index and the vehicle battery failure, as shown in Figure 4 , specifically comprising the following steps:

[0069] S401, dividing the original data set into a plurality of original sample subsets with the same number of samples.

[0070] For example, the original data set can be divided into M original sample subsets with the same number of samples; wherein each sample in each original sample subset includes all candidate battery indexes and vehicle labels.

[0071] S402, using a cross-entropy loss function to process each original sample subset to obtain an initial prediction accuracy.

[0072] Wherein, the initial prediction accuracy can represent the loss value of the failure prediction based on all candidate battery indexes.

[0073] One implementation, for each original sample subset, uses a cross-entropy loss function to process the original sample subset to obtain the first prediction accuracy of the original sample subset; the average of the first prediction accuracy of each original sample subset is taken as the initial prediction accuracy.

[0074] For example, for each original sample subset, the original sample subset can be input into the cross-entropy loss function, and the cross-entropy loss function can be used to calculate the loss of the original sample subset and obtain the first prediction accuracy of the original sample subset. Furthermore, based on the first prediction accuracy of each original sample subset, the average is calculated, and the result of the average calculation is determined as the initial prediction accuracy.

[0075] Another implementation, each original sample subset can be directly input into the cross-entropy model, and the cross-entropy model can be used to analyze and process each original sample subset, such as mean processing, removing the maximum value, and removing the minimum value, and finally obtaining the initial prediction accuracy.

[0076] S403, for each candidate battery index, removing the sample index data corresponding to the candidate battery index from the sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery index.

[0077] For example, the candidate battery indicators include three indicators z1, z2 and z3; the original sample subsets include three subsets T1={z1, z2, z3}, T2={z1, z2, z3} and T3={z1, z2, z3}; further, for the candidate battery indicator z1, the sample indicator data corresponding to the candidate battery indicator z1 is removed from the sample data of each original sample subset (T1, T2 and T3), to obtain each new sample subset corresponding to the candidate battery indicator z1, i.e., T 1_1 ={z2, z3}, T 1_2 ={z2, z3} and T 1_3 ={z2, z3}; for the candidate battery indicator z2, the sample indicator data corresponding to the candidate battery indicator z2 is removed from the sample data of each original sample subset (T1, T2 and T3), to obtain each new sample subset corresponding to the candidate battery indicator z2, i.e., T 2_1 ={z1, z3}, T 2_2 ={z1, z3} and T 2_3 ={z1, z3}; for the candidate battery indicator z3, the sample indicator data corresponding to the candidate battery indicator z3 is removed from the sample data of each original sample subset (T1, T2 and T3), to obtain each new sample subset corresponding to the candidate battery indicator z3, i.e., T 3_1 ={z1, z2}, T 3_2 ={z1, z2} and T 3_3 ={z1, z2}.

[0078] S404, according to the cross-entropy loss function, the initial prediction accuracy and each new sample subset, determining the correlation between the candidate battery indicator and the vehicle battery fault.

[0079] In an implementation, for each new sample subset, the cross-entropy loss function is used to process the new sample subset, to obtain the second prediction accuracy of the new sample subset, and the difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy is taken as the accuracy deviation of the original sample subset corresponding to the candidate battery indicator; according to the accuracy deviation of the original sample subset corresponding to the candidate battery indicator for each new sample subset, the correlation between the candidate battery indicator and the vehicle battery fault is determined.

[0080] As an example, based on the candidate battery indicator z1, each new sample subset corresponding to the candidate battery indicator z1 is determined, i.e., T 1_1 ={z2, z3}, T 1_2 ={z2, z3} and T 1_3 ={z2, z3}; further, T 1_1 ={z2, z3}, T 1_2 ={z2, z3} and T 1_3each new sample subset in {z2, z3} is input to the cross-entropy loss function, loss calculation is performed on the new sample subsets (T 1_1 , T 1_2 , and T 1_3 ) by the cross-entropy loss function, and the second prediction accuracy of each new sample subset (T 1_1 , T 1_2 , and T 1_3 ) is obtained; further, for each new sample subset (T 1_1 , T 1_2 , and T 1_3 ), the second prediction accuracy of the new sample subset T 1_i (i = 1, 2, 3) is subtracted from the initial prediction accuracy, and the difference value operation result of the new sample subset T 1_i is determined as the corresponding original sample subset T 1_i of the new sample subset T i , and the accuracy deviation P1 = {w 1_1 , w 1_2 , and w 1_3} of the candidate battery index z1 is obtained.

[0081] Based on the candidate battery index z2, each new sample subset corresponding to the candidate battery index z2 is determined, that is, T 2_1 = {z1, z3}, T 2_2 = {z1, z3}, and T 2_3 = {z1, z3}; further, each new sample subset in {T 2_1 = {z1, z3}, T 2_2 = {z1, z3}, and T 2_3 = {z1, z3} is input to the cross-entropy loss function, loss calculation is performed on the new sample subsets (T 2_1 , T 2_2 , and T 2_3 ) by the cross-entropy loss function, and the second prediction accuracy of each new sample subset (T 2_1 , T 2_2 , and T 2_3 ) is obtained; further, for each new sample subset (T 2_1 , T 2_2 , and T 2_3 ), the second prediction accuracy of the new sample subset T 2_i (i = 1, 2, 3) is subtracted from the initial prediction accuracy, and the difference value operation result of the new sample subset T 2_i is determined as the corresponding original sample subset T 2_i of the new sample subset T i , and the accuracy deviation P2 = {w 2_1 , w 2_2and w 2_3}.

[0082] Based on the candidate battery index z3, each new sample subset corresponding to the candidate battery index z3, i.e., T 3_1 ={z1,z2}、T 3_2 ={z1,z2}and T 3_3 ={z1,z2}is determined; further, each new sample subset in T 3_1 ={z1,z2}、T 3_2 ={z1,z2}and T 3_3 ={z1,z2}may be input to a cross-entropy loss function, and the cross-entropy loss function is used to calculate the loss of each new sample subset (T 3_1 , T 3_2 and T 3_3 ), and the second prediction accuracy of each new sample subset (T 3_1 , T 3_2 and T 3_3 ) is obtained; further, for each new sample subset (T 3_1 , T 3_2 and T 3_3 ), the second prediction accuracy of the new sample subset T 3_i (i=1, 2, 3) is subtracted from the initial prediction accuracy, and the difference value of the new sample subset T 3_i is determined as the accuracy deviation P3={w 3_i , w i and w 3_1} of the original sample subset T 3_2 corresponding to the new sample subset T 3_3 .

[0083] Further, for the candidate battery index z1, w 1_1 , w 1_2 and w 1_3 in the accuracy deviation P1 can be processed by averaging, and if the value of the average processing is greater than 0, it can be determined that the correlation between the candidate battery index z1 and the vehicle battery fault is 0; if the value of the average processing is less than 0, the result of the average processing can be determined as the correlation between the candidate battery index z1 and the vehicle battery fault; for the candidate battery index z2, w 2_1 , w 2_2 and w 2_3 in the accuracy deviation P2 can be processed by averaging, and the correlation between the candidate battery index z2 and the vehicle battery fault is determined; for the candidate battery index z3, w 3_1 , w 3_2 and w 3_3The mean value processing is performed, and the correlation between the candidate battery index z3 and the vehicle battery failure is determined.

[0084] In another implementation manner, the cross-entropy loss function, the initial prediction accuracy and each new sample subset can be directly input into the correlation determination model; the correlation determination model analyzes the same, and the correlation between the candidate battery index and the vehicle battery failure is determined according to the analysis result.

[0085] In the embodiments of the present application, by constructing each new sample subset corresponding to each candidate battery index, a foundation is laid for determining the correlation between each candidate battery index and the vehicle battery failure.

[0086] On the basis of the above-mentioned embodiments, the embodiment of the present application explains and describes the above-mentioned embodiment S203 in detail. Specifically, the process of the target electric vehicle for vehicle battery failure warning in the embodiment of the present application is involved, which specifically includes the following steps: inputting the target index data and the battery aging factor into the Bayesian network model to obtain the target failure probability of the target electric vehicle; determining the target failure level corresponding to the target failure probability based on the corresponding relationship between the failure probability and the failure level; and warning the target electric vehicle for vehicle battery failure according to the target failure level.

[0087] Among them, the target failure probability can represent the probability of predicting that the target electric vehicle fails based on the target index data in the current period; and the target failure level can represent the level of predicting that the target electric vehicle fails based on the target index data in the current period.

[0088] For example, the target index data and the battery aging factor can be input into the Bayesian network model, the Bayesian network model can analyze the target index data and the battery aging factor, and the failure probability corresponding to the target index data and the battery aging factor can be predicted to obtain the target failure probability of the target electric vehicle; further, the target failure level corresponding to the target failure probability can be determined according to the corresponding relationship between the preset failure probability and the failure level; and the vehicle battery failure warning can be performed according to the target failure level.

[0089] Optionally, the corresponding relationship between the failure probability and the failure level can be as shown in Table 1.

[0090]

[0091] Table 1 Corresponding relationship

[0092] In the embodiment of the present application, by introducing the Bayesian network model, the target failure probability of the target electric vehicle is determined, which lays a foundation for accurately making the vehicle battery failure warning.

[0093] It should be understood that although the steps in the flowcharts involved in the embodiments described above are shown in sequence according to the arrows, the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the embodiments described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of the steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0094] Based on the same inventive concept, the embodiments of the present application also provide an electric vehicle battery failure warning device for implementing the above-mentioned electric vehicle battery failure warning method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more electric vehicle battery failure warning device embodiments provided below can refer to the limitations of the electric vehicle battery failure warning method described above, which will not be repeated here.

[0095] In one exemplary embodiment, as shown in Figure 5 An electric vehicle battery failure warning device 1 is provided, comprising: an acquisition module 10, a factor determination module 20 and a warning module 30, wherein:

[0096] The acquisition module 10 is configured to acquire target indicator data of a target electric vehicle under a key battery indicator in a current period; wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle.

[0097] The factor determination module 20 is configured to determine a battery aging factor of the target electric vehicle according to an initial capacity of the battery of the target electric vehicle and a current capacity of the battery.

[0098] The warning module 30 is configured to perform vehicle battery failure warning on the target electric vehicle based on a Bayesian network model according to the target indicator data and the battery aging factor.

[0099] In one embodiment, as shown in Figure 6 The acquisition module 10 specifically further comprises:

[0100] The correlation determination unit 11 is configured to determine the correlation between each candidate battery indicator and vehicle battery failure according to a cross-entropy loss function and an original data set.

[0101] The index screening unit 12 is configured to screen the key battery index from the candidate battery indexes according to the correlation between each candidate battery index and the vehicle battery fault.

[0102] In an embodiment, the correlation determination unit 11 is specifically configured to:

[0103] divide the original data set into a plurality of original sample subsets with the same number of samples; process each original sample subset by using a cross-entropy loss function to obtain an initial prediction accuracy; for each candidate battery index, remove the sample index data corresponding to the candidate battery index from the sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery index; and determine the correlation between the candidate battery index and the vehicle battery fault according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset.

[0104] In an embodiment, the correlation determination unit 11 is specifically configured to:

[0105] for each original sample subset, process the original sample subset by using a cross-entropy loss function to obtain a first prediction accuracy of the original sample subset; and take the average of the first prediction accuracies of each original sample subset as the initial prediction accuracy.

[0106] In an embodiment, the correlation determination unit 11 is specifically configured to:

[0107] for each new sample subset, process the new sample subset by using a cross-entropy loss function to obtain a second prediction accuracy of the new sample subset, and take the difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy as the accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index; and determine the correlation between the candidate battery index and the vehicle battery fault according to the accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index for each new sample subset.

[0108] In an embodiment, the early warning module 30 is specifically configured to:

[0109] input the target index data and the battery aging factor into the Bayesian network model to obtain a target fault probability of the target electric vehicle; determine a target fault level corresponding to the target fault probability based on a corresponding relationship between the fault probability and the fault level; and perform vehicle battery fault early warning on the target electric vehicle according to the target fault level.

[0110] The various modules in the above electric vehicle battery failure early warning device can be implemented wholly or partially by software, hardware, and combinations thereof. The above modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in a computer device in software form, so that the processor invokes and executes the operations corresponding to the above modules.

[0111] In an exemplary embodiment, a computer device, which can be a server, is provided, and an internal structure diagram of the computer device can be as shown in Figure 7 The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store index data. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with external terminals through a network connection. The computer program is executed by the processor to implement an electric vehicle battery failure early warning method.

[0112] Those skilled in the art can understand that Figure 7 The structure shown in the above embodiment is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. Specifically, the computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0113] In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the following steps:

[0114] obtaining target index data of a target electric vehicle under a key battery index in a current period; wherein the key battery index is an index related to vehicle battery failure selected from candidate battery indexes according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample index data of a sample electric vehicle under each candidate battery index and a sample class label of the sample electric vehicle;

[0115] determining a battery aging factor of the target electric vehicle according to an initial capacity of the battery of the target electric vehicle and a current capacity of the battery;

[0116] Based on the Bayesian network model, target electric vehicles are warned of vehicle battery failure according to target index data and battery aging factors.

[0117] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0118] According to the cross-entropy loss function and the original data set, the correlation between each candidate battery index and the vehicle battery failure is determined; and according to the correlation between each candidate battery index and the vehicle battery failure, the key battery index is selected from each candidate battery index.

[0119] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0120] The original data set is divided into a plurality of original sample subsets with the same number of samples; the cross-entropy loss function is used to process each original sample subset to obtain an initial prediction accuracy; for each candidate battery index, the sample index data corresponding to the candidate battery index is removed from the sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery index; and the correlation between the candidate battery index and the vehicle battery failure is determined according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset.

[0121] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0122] For each original sample subset, the cross-entropy loss function is used to process the original sample subset to obtain a first prediction accuracy of the original sample subset; and the average of the first prediction accuracies of each original sample subset is taken as the initial prediction accuracy.

[0123] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0124] For each new sample subset, the cross-entropy loss function is used to process the new sample subset to obtain a second prediction accuracy of the new sample subset, and the difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy is taken as the accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index; and the correlation between the candidate battery index and the vehicle battery failure is determined according to the accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index for each new sample subset.

[0125] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0126] input the target indicator data and the battery aging factor into the Bayesian network model to obtain a target failure probability of the target electric vehicle; determine a target failure level corresponding to the target failure probability based on a corresponding relationship between the failure probability and the failure level; and perform vehicle battery failure early warning on the target electric vehicle according to the target failure level.

[0127] In one embodiment, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0128] obtaining target indicator data of a target electric vehicle under a key battery indicator in a current period; wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle;

[0129] determining a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery;

[0130] performing vehicle battery failure early warning on the target electric vehicle based on the Bayesian network model, the target indicator data and the battery aging factor.

[0131] In one embodiment, the computer program is executed by the processor to implement the following steps:

[0132] determining a correlation between each candidate battery indicator and vehicle battery failure according to a cross-entropy loss function and an original data set; and selecting a key battery indicator from each candidate battery indicator according to the correlation between each candidate battery indicator and vehicle battery failure.

[0133] In one embodiment, the computer program is executed by the processor to implement the following steps:

[0134] dividing the original data set into a plurality of original sample subsets with the same number of samples; processing each original sample subset by using a cross-entropy loss function to obtain an initial prediction accuracy; for each candidate battery indicator, removing sample indicator data corresponding to the candidate battery indicator from sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery indicator; and determining a correlation between the candidate battery indicator and vehicle battery failure according to the cross-entropy loss function, the initial prediction accuracy and each new sample subset.

[0135] In one embodiment, the computer program is executed by the processor to implement the following steps:

[0136] For each original sample subset, the original sample subset is processed by using a cross-entropy loss function to obtain a first prediction accuracy of the original sample subset; and the average of the first prediction accuracies of the original sample subsets is taken as an initial prediction accuracy.

[0137] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0138] For each new sample subset, the new sample subset is processed by using a cross-entropy loss function to obtain a second prediction accuracy of the new sample subset, and a difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy is taken as an accuracy deviation of the original sample subset corresponding to the new sample subset with respect to the candidate battery index; and the correlation between the candidate battery index and the vehicle battery fault is determined according to the accuracy deviation of the original sample subset corresponding to each new sample subset with respect to the candidate battery index.

[0139] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0140] The target index data and the battery aging factor are input into the Bayesian network model to obtain a target fault probability of the target electric vehicle; a target fault level corresponding to the target fault probability is determined based on a corresponding relationship between the fault probability and the fault level; and the target electric vehicle is subjected to vehicle battery fault early warning according to the target fault level.

[0141] In one embodiment, a computer program product is provided, comprising a computer program which, when executed by the processor, implements the following steps:

[0142] Target index data of the target electric vehicle under a key battery index in a current period are acquired; wherein the key battery index is an index related to the vehicle battery fault selected from each candidate battery index according to a cross-entropy loss function and an original data set; each sample data in the original data set comprises sample index data of a sample electric vehicle under each candidate battery index and a sample category label of the sample electric vehicle;

[0143] A battery aging factor of the target electric vehicle is determined according to a battery initial capacity and a battery current capacity of the target electric vehicle;

[0144] The target electric vehicle is subjected to vehicle battery fault early warning based on the Bayesian network model according to the target index data and the battery aging factor.

[0145] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0146] According to the cross-entropy loss function and the original data set, a correlation between each candidate battery index and the vehicle battery fault is determined; and according to the correlation between each candidate battery index and the vehicle battery fault, a key battery index is selected from each candidate battery index.

[0147] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0148] The original data set is divided into a plurality of original sample subsets with the same number of samples; each original sample subset is processed by using a cross-entropy loss function to obtain an initial prediction accuracy; for each candidate battery index, sample index data corresponding to the candidate battery index is removed from sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery index; and a correlation between the candidate battery index and the vehicle battery fault is determined according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset.

[0149] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0150] For each original sample subset, the original sample subset is processed by using a cross-entropy loss function to obtain a first prediction accuracy of the original sample subset; and the average of the first prediction accuracies of each original sample subset is taken as the initial prediction accuracy.

[0151] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0152] For each new sample subset, the new sample subset is processed by using a cross-entropy loss function to obtain a second prediction accuracy of the new sample subset, and a difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy is taken as an accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index; and a correlation between the candidate battery index and the vehicle battery fault is determined according to the accuracy deviation of the original sample subset corresponding to the candidate battery index with respect to the candidate battery index.

[0153] In one embodiment, the computer program, when executed by the processor, implements the following steps:

[0154] The target index data and the battery aging factor are input into the Bayesian network model to obtain a target fault probability of the target electric vehicle; a target fault level corresponding to the target fault probability is determined based on a corresponding relationship between the fault probability and the fault level; and the target electric vehicle is warned of the vehicle battery fault according to the target fault level.

[0155] It should be noted that the information (including but not limited to vehicle information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0156] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0157] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0158] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. An electric vehicle battery failure early warning method, characterized by, The method comprises: obtaining target indicator data of a target electric vehicle in a key battery indicator in a current period; wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set; each sample data in the original data set comprises sample indicator data of a sample electric vehicle in each candidate battery indicator and a sample category label of the sample electric vehicle; determining a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery; based on a Bayesian network model, performing vehicle battery failure warning on the target electric vehicle according to the target indicator data and the battery aging factor; wherein the key battery indicator is determined by the following method: dividing the original data set into a plurality of original sample subsets with the same sample quantity; using a cross-entropy loss function to process each original sample subset to obtain an initial prediction accuracy; for each candidate battery indicator, removing the sample indicator data corresponding to the candidate battery indicator from the sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery indicator, and determining the correlation between the candidate battery indicator and the vehicle battery failure according to the cross-entropy loss function, the initial prediction accuracy and each new sample subset; selecting the key battery indicator from each candidate battery indicator according to the correlation between each candidate battery indicator and the vehicle battery failure.

2. The method of claim 1, wherein, The method comprises: for each original sample subset, using a cross-entropy loss function to process the original sample subset to obtain a first prediction accuracy of the original sample subset; taking the average of the first prediction accuracies of each original sample subset as the initial prediction accuracy.

3. The method of claim 1, wherein, The method comprises: for each new sample subset, using a cross-entropy loss function to process the new sample subset to obtain a second prediction accuracy of the new sample subset, and taking the difference between the second prediction accuracy of the new sample subset and the initial prediction accuracy as the accuracy deviation of the original sample subset corresponding to the new sample subset for the candidate battery indicator; determining the correlation between the candidate battery indicator and the vehicle battery failure according to the accuracy deviation of the original sample subset corresponding to each new sample subset for the candidate battery indicator.

4. The method of claim 1, wherein, The method comprises: inputting the target indicator data and the battery aging factor into the Bayesian network model to obtain a target failure probability of the target electric vehicle; determining a target failure level corresponding to the target failure probability based on the corresponding relationship between the failure probability and the failure level; performing vehicle battery failure warning on the target electric vehicle according to the target failure level.

5. An electric vehicle battery failure early warning device, characterized by, The device comprises: The acquisition module is configured to acquire target indicator data of the target electric vehicle in a current period under a key battery indicator, wherein the key battery indicator is an indicator related to vehicle battery failure selected from each candidate battery indicator according to a cross-entropy loss function and an original data set, and each sample data in the original data set includes sample indicator data of a sample electric vehicle under each candidate battery indicator and a sample category label of the sample electric vehicle. The factor determination module is configured to determine a battery aging factor of the target electric vehicle according to an initial capacity of a battery of the target electric vehicle and a current capacity of the battery. The early warning module is configured to perform vehicle battery failure early warning on the target electric vehicle based on a Bayesian network model according to the target indicator data and the battery aging factor. The acquisition module includes: The correlation determination unit is configured to divide the original data set into a plurality of original sample subsets with the same sample quantity, process each original sample subset by using a cross-entropy loss function to obtain an initial prediction accuracy, eliminate sample indicator data corresponding to each candidate battery indicator from sample data of each original sample subset to obtain each new sample subset corresponding to the candidate battery indicator, and determine a correlation between the candidate battery indicator and vehicle battery failure according to the cross-entropy loss function, the initial prediction accuracy, and each new sample subset. The indicator screening unit is configured to screen the key battery indicator from each candidate battery indicator according to the correlation between each candidate battery indicator and vehicle battery failure. 6.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-5 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 4.

7. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 4.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 4. The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 4.

Images