Method and device for dynamically evaluating safety risk of power battery system, equipment and medium

Abstract

The application discloses a power battery system safety risk dynamic evaluation method, device, equipment and medium, relates to the power battery risk evaluation technical field, and the method comprises the following steps: calculating a target parameter set according to a battery monomer voltage data matrix and a probe temperature data matrix of a target new energy automobile in a current stage; performing parameter distribution difference calculation on each parameter matrix in the target parameter set, and determining N feature intervals corresponding to the parameter matrix; calculating membership values of each feature interval; and determining a safety risk level of the target new energy automobile in the current stage according to the membership values of each feature interval corresponding to each parameter matrix of the target new energy automobile in the current stage and a Bayesian network model. The application can accurately evaluate the safety risk of the power battery system of the new energy automobile.

Description

Technical Field

[0001] This invention relates to the field of power battery risk assessment technology, and in particular to a method, device, equipment and medium for dynamic assessment of the safety risks of power battery systems in new energy vehicles. Background Technology

[0002] Currently, there is a strong push to promote the development of the new energy vehicle industry. As a core component of new energy vehicles, the safety of power batteries is crucial to the overall safety of the vehicle. Therefore, quantifying the safety risks of new energy vehicle power battery systems and achieving accurate risk assessment is essential for improving the safety and reliability of new energy vehicles. Currently, the main method for assessing the safety risks of new energy vehicle power battery systems, both domestically and internationally, is the threshold comparison method. This method assesses safety by triggering alarms when thresholds are exceeded. However, this method cannot accurately assess the safety status of new energy vehicle power battery systems. Summary of the Invention

[0003] The purpose of this invention is to provide a method, device, equipment, and medium for dynamic assessment of the safety risks of power battery systems, which can accurately assess the safety risks of power battery systems for new energy vehicles.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] In a first aspect, the present invention provides a method for dynamic assessment of safety risks of a power battery system, comprising:

[0006] Obtain the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; the row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series; the row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series.

[0007] Based on the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage, a target parameter set is calculated; the target parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix.

[0008] For each parameter matrix in the target parameter set of the current stage target new energy vehicle, the parameter distribution difference is calculated, and the top N feature intervals of the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix.

[0009] Discrete membership processing is performed on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage to determine the membership value of each feature interval;

[0010] Based on the membership values ​​of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model, the safety risk level of the target new energy vehicle in the current stage is determined.

[0011] Optionally, the process for determining the Bayesian network model is as follows:

[0012] Obtain a sample dataset; the sample dataset includes a battery cell voltage data matrix and probe temperature data matrix of new energy vehicles involved in accidents during historical periods, and a battery cell voltage data matrix and probe temperature data matrix of new energy vehicles in normal operation during historical periods.

[0013] Based on the battery cell voltage data matrix and probe temperature data matrix in the sample dataset, a sample parameter set is calculated for each new energy vehicle; the sample parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix.

[0014] For each parameter matrix in the sample parameter set of each new energy vehicle in the historical stage, the parameter distribution difference is calculated, and the top N feature intervals of the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix.

[0015] Discrete membership processing is performed on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval;

[0016] Based on the membership values ​​of each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage, and the attributes of each new energy vehicle, a Bayesian network is trained to obtain a Bayesian network model; the attributes are new energy vehicles involved in accidents or new energy vehicles that are in good condition.

[0017] Optionally, obtain the sample dataset, specifically including:

[0018] Collect the full lifecycle data of new energy vehicles with the same model or the same battery model that have been involved in accidents from the new energy vehicle big data platform, and generate an accident new energy vehicle dataset;

[0019] Collect full lifecycle data of normal new energy vehicles of the same model or the same battery model from the new energy vehicle big data platform to generate a normal new energy vehicle dataset;

[0020] Based on the accident data set of new energy vehicles, the battery cell voltage data matrix and probe temperature data matrix of new energy vehicles involved in accidents in historical periods are obtained.

[0021] Based on the normal new energy vehicle dataset, the battery cell voltage data matrix and probe temperature data matrix of normal new energy vehicles in historical periods are obtained.

[0022] Optionally, the parameter distribution difference is calculated for each parameter matrix in the sample parameter set for each new energy vehicle in the historical period, and the top N feature intervals of the distribution difference values ​​arranged in descending order are selected as the N feature intervals corresponding to the parameter matrix; specifically including:

[0023] Perform a first operation on each parameter matrix in the sample parameter set for each new energy vehicle in the historical stage to determine N feature intervals corresponding to each parameter matrix;

[0024] The first operation is:

[0025] Based on all the data in the parameter matrix, the parameter distribution of the parameter matrix is ​​statistically analyzed, and the differences in parameter distribution are compared to determine multiple feature intervals;

[0026] Calculate the distribution difference value for each feature interval, and determine the feature intervals corresponding to the first N distribution difference values ​​in descending order as the N feature intervals corresponding to the parameter matrix.

[0027] Optionally, discrete membership processing is performed on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval, specifically including:

[0028] The fuzzy logic method is used to perform discrete membership processing on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage, and the membership value of each feature interval is determined.

[0029] Optionally, it also includes:

[0030] The membership values ​​of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, the operating status of the target new energy vehicle in the current stage, and the predicted safety level of the target new energy vehicle in the current stage output by the Bayesian network model are input into the sample dataset of the target new energy vehicle to complete the update of the Bayesian network model.

[0031] Secondly, the present invention provides a dynamic assessment device for the safety risks of a power battery system, comprising:

[0032] The target new energy vehicle operation data acquisition module is used to acquire the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; the row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series; the row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series.

[0033] The target parameter set calculation module is used to calculate the target parameter set based on the battery cell voltage data matrix and the probe temperature data matrix of the target new energy vehicle at the current stage; the target parameter set includes the highest voltage parameter matrix of the power battery system, the lowest voltage parameter matrix of the power battery system, the voltage range parameter matrix of the power battery system, the highest temperature parameter matrix of the power battery system, and the temperature range parameter matrix of the power battery system.

[0034] The feature interval determination module is used to calculate the parameter distribution difference for each parameter matrix in the target parameter set of the target new energy vehicle in the current stage, and select the first N feature intervals with the distribution difference values ​​arranged from largest to smallest to determine the N feature intervals corresponding to the parameter matrix.

[0035] The membership value calculation module is used to perform discrete membership processing on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, and determine the membership value of each feature interval.

[0036] The safety risk level determination module is used to determine the safety risk level of the target new energy vehicle in the current stage based on the membership value of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model.

[0037] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the dynamic assessment method for safety risks of a power battery system according to the first aspect.

[0038] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the dynamic assessment method for safety risks of a power battery system as described in the first aspect.

[0039] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0040] This invention uses the battery cell voltage data and probe temperature data of new energy vehicles to determine the highest voltage parameter matrix, lowest voltage parameter matrix, voltage range parameter matrix, highest temperature parameter matrix, and temperature range parameter matrix of the power battery system. Based on the above parameter matrices and a Bayesian network model, the safety risk level of the new energy vehicle is determined, thereby improving the accuracy of assessing the safety risks of the power battery system of new energy vehicles. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A flowchart illustrating a dynamic safety risk assessment method for a power battery system provided in an embodiment of the present invention;

[0043] Figure 2 This is a distribution difference diagram of the maximum voltage Max-V of the power battery system provided in an embodiment of the present invention;

[0044] Figure 3 A diagram showing the sorting results of data segments for the three characteristic intervals of the maximum voltage Max-V of the power battery system provided in an embodiment of the present invention;

[0045] Figure 4 This is a schematic diagram showing the discrete membership of three characteristic intervals of the maximum voltage Max-V of the power battery system provided in an embodiment of the present invention.

[0046] Figure 5 This is a schematic diagram of a Bayesian network structure provided in an embodiment of the present invention;

[0047] Figure 6 This is a schematic diagram of a dynamic safety risk assessment device for a power battery system provided in an embodiment of the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] Example 1

[0051] like Figure 1 As shown in the figure, an embodiment of the present invention provides a method for dynamic assessment of safety risks of a power battery system, comprising:

[0052] Step 100: Obtain the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; the row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series; the row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series.

[0053] Step 200: Calculate the target parameter set based on the battery cell voltage data matrix and the probe temperature data matrix of the target new energy vehicle at the current stage; the target parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix.

[0054] Step 300: Calculate the parameter distribution difference for each parameter matrix in the target parameter set of the target new energy vehicle in the current stage, and select the first N feature intervals of the distribution difference values ​​arranged from largest to smallest to determine the N feature intervals corresponding to the parameter matrix.

[0055] Step 400: Perform discrete membership processing on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, and determine the membership value of each feature interval.

[0056] Step 500: Determine the safety risk level of the target new energy vehicle in the current stage based on the membership value of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model.

[0057] The process of determining the Bayesian network model is as follows:

[0058] (1) Obtain sample datasets; the sample datasets include battery cell voltage data matrix and probe temperature data matrix of new energy vehicles involved in accidents during historical periods, and battery cell voltage data matrix and probe temperature data matrix of new energy vehicles in normal periods during historical periods.

[0059] Step (1) specifically includes:

[0060] 1) Collect the full lifecycle data of accident-related new energy vehicles of the same model or battery type from the new energy vehicle big data platform to generate an accident-related new energy vehicle dataset; 2) Collect the full lifecycle data of normal new energy vehicles of the same model or battery type from the new energy vehicle big data platform to generate a normal new energy vehicle dataset; 3) Based on the accident-related new energy vehicle dataset, obtain the battery cell voltage data matrix and probe temperature data matrix of accident-related new energy vehicles in historical periods; 4) Based on the normal new energy vehicle dataset, obtain the battery cell voltage data matrix and probe temperature data matrix of normal new energy vehicles in historical periods.

[0061] An example: Collect the full lifecycle data of accident-damaged new energy vehicles (hereinafter referred to as accident vehicles) and normal new energy vehicles (hereinafter referred to as normal vehicles) of the same model or battery type from the new energy vehicle big data platform to form vehicle dataset Q1.

[0062] Based on vehicle dataset Q1, we collected battery cell voltage and probe temperature data for new energy vehicles. We obtained multiple sets of single-frame data for n battery cell voltages and n probe temperatures from the new energy vehicle big data platform. The single-frame battery cell voltage data is as follows: at 2020:09:15 13:00:00, battery cell voltage 1 is 3.21V, battery cell voltage 2 is 3.22V, and so on. The single-frame probe temperature data is as follows: at 2020:09:15 13:00:00, probe temperature 1 is 26℃, probe temperature 2 is 36℃, and so on. Each new energy vehicle forms a battery cell voltage data matrix and a probe temperature data matrix as shown below.

[0063]

[0064]

[0065] Where Vol represents the battery cell voltage data matrix, with rows representing battery cell voltage data and columns representing time series; Tem represents the probe temperature data matrix, with rows representing probe temperature data and columns representing time series.

[0066] (2) Calculate the sample parameter set for each new energy vehicle based on the battery cell voltage data matrix and probe temperature data matrix in the sample dataset; the sample parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix.

[0067] An example: Select five parameters related to the battery cell voltage data and probe temperature data: the maximum voltage of the power battery system Max-V, the minimum voltage of the power battery system Min-V, the voltage range of the power battery system Diff-V, the maximum temperature of the power battery system Max-T, and the temperature range of the power battery system Diff-T.

[0068] Calculate the parameter matrix for each new energy vehicle regarding the above five parameters:

[0069]

[0070]

[0071] Diff-V = Max-V-Min-V.

[0072]

[0073]

[0074] Diff-T = Max-T-Min-T.

[0075] (3) Calculate the parameter distribution difference for each parameter matrix in the sample parameter set for each new energy vehicle in the historical stage, and select the first N feature intervals of the distribution difference values ​​arranged from largest to smallest to determine the N feature intervals corresponding to the parameter matrix.

[0076] Step (3) specifically includes:

[0077] Perform a first operation on each parameter matrix in the sample parameter set for each new energy vehicle in the historical stage to determine N feature intervals corresponding to each parameter matrix;

[0078] The first operation is as follows: 1) Based on all the data in the parameter matrix, statistically analyze the parameter distribution of the parameter matrix and compare the differences in parameter distribution to determine multiple feature intervals; 2) Calculate the distribution difference value of each feature interval, and determine the feature intervals corresponding to the first N distribution difference values ​​as the N feature intervals corresponding to the parameter matrix in descending order.

[0079] An example: Statistically analyze the distribution of the following five parameters for a power battery system: Maximum Voltage (Max-V), Minimum Voltage (Min-V), Voltage Range (Diff-V), Maximum Temperature (Max-T), and Temperature Range (Diff-T). Add up all data from normal vehicles and all data from accident vehicles. Compare the distribution of these five parameters for all normal vehicles and all accident vehicles. Select the three characteristic intervals with the largest distribution differences, designated A1, A2, and A3. Taking the maximum voltage (Max-V) of the power battery system as an example, the interval with the largest distribution difference is as follows: Figure 2 As shown.

[0080] (4) Perform discrete membership processing on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval.

[0081] An example: Since Bayesian networks can only handle discretized data, and each parameter has a continuous density value, it is necessary to discretize the continuous density into two states based on the differences between normal vehicles and accident vehicles. In this embodiment of the invention, a fuzzy logic method is used for data discretization. First, all data segments of normal vehicles and accident vehicles are sorted from smallest to largest within the feature interval. Taking the maximum voltage Max-V of the power battery system as an example, the result diagram of the three feature intervals is shown below. Figure 3 As shown.

[0082] Then, based on the distribution of the data segments of the accident vehicle, a suitable membership function can be selected. This invention selects three membership functions, choosing the appropriate function based on the distribution of the data segments:

[0083]

[0084]

[0085]

[0086] Taking the maximum voltage Max-V of the power battery system as an example, the membership values ​​of the three characteristic intervals calculated according to the above formula are as follows: Figure 4 As shown.

[0087] When the membership degree is greater than 0.5, the value is in the "YES" state; otherwise, it is in the "NO" state. After completing this process, the density values ​​of the above five parameters are divided into two states, where "YES" represents the dangerous state and "NO" represents the safe state, thus completing the parameter discretization process.

[0088] After data discretization, different intervals are labeled with different values. For example: [X1%-X2%]-Yes, [0%-X1%]&[X2%-100%]-No.

[0089] The processing of the above five parameters is shown in Tables 1 and 2.

[0090] Table 1. Hierarchical table of feature intervals in the parameter matrix.

[0091]

[0092]

[0093] Table 2. Classification of Characteristic Intervals of New Energy Vehicles in Historical Stages

[0094]

[0095] (5) Based on the membership values ​​of each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage, and the attributes of each new energy vehicle, a Bayesian network is trained to obtain a Bayesian network model; the attributes are either accident-damaged new energy vehicles or normal new energy vehicles. The Bayesian network structure is as follows: Figure 5 As shown in Table 3, the training results of the Bayesian network are as follows.

[0096] Table 3. Training Results of Bayesian Networks

[0097]

[0098] When a new energy vehicle not in the training set obtains the data features shown in Table 4 through the above process, it is input into the Bayesian network model constructed above to assess the safety risk and obtain the safety risk X1 of the vehicle.

[0099] Table 4. Classification of Target New Energy Vehicle Characteristic Ranges

[0100]

[0101] Furthermore, the method provided in this embodiment of the invention also includes:

[0102] (1) Risk probability update.

[0103] The membership values ​​of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, the operating status of the target new energy vehicle in the current stage, and the predicted safety level of the target new energy vehicle in the current stage output by the Bayesian network model are input into the sample dataset of the target new energy vehicle. The Bayesian network parameters are then retrained to update the Bayesian network model. The operating status includes whether an accident has occurred or not.

[0104] (2) Dynamic assessment of safety risks.

[0105] 1) As historical vehicle data accumulates, regular safety risk assessments can be conducted to obtain the dynamic changes in vehicle safety risks.

[0106] 2) Classify safety risks: [0%-30%] - safe, [30%-60%] - warning, [60%-100%] - dangerous.

[0107] 3) Based on the assessment results, the safety of the vehicle can be assessed in a timely manner, and timely maintenance can be carried out to avoid accidents.

[0108] Example 2

[0109] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, a dynamic assessment device for safety risks of a power battery system is provided below.

[0110] like Figure 6 As shown, the dynamic safety risk assessment device for a power battery system provided in this embodiment of the invention includes:

[0111] The target new energy vehicle operation data acquisition module 1 is used to acquire the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; the row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series; the row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series.

[0112] The target parameter set calculation module 2 is used to calculate the target parameter set based on the battery cell voltage data matrix and the probe temperature data matrix of the target new energy vehicle at the current stage. The target parameter set includes the highest voltage parameter matrix of the power battery system, the lowest voltage parameter matrix of the power battery system, the voltage range parameter matrix of the power battery system, the highest temperature parameter matrix of the power battery system, and the temperature range parameter matrix of the power battery system.

[0113] The feature interval determination module 3 is used to calculate the parameter distribution difference for each parameter matrix in the target parameter set of the target new energy vehicle in the current stage, and select the first N feature intervals with the distribution difference values ​​arranged from largest to smallest to determine the N feature intervals corresponding to the parameter matrix.

[0114] Membership value calculation module 4 is used to perform discrete membership processing on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, and determine the membership value of each feature interval.

[0115] The safety risk level determination module 5 is used to determine the safety risk level of the target new energy vehicle in the current stage based on the membership value of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model.

[0116] Example 3

[0117] This invention provides an electronic device including a memory and a processor. The memory stores a computer program, and the processor runs the computer program to enable the electronic device to perform the dynamic assessment method for safety risks of a power battery system according to Embodiment 1.

[0118] Alternatively, the aforementioned electronic device may be a server.

[0119] In addition, this embodiment of the invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the dynamic assessment method for safety risks of a power battery system according to Embodiment 1.

[0120] This invention utilizes battery cell voltage data and probe temperature data collected according to GB / T 32960-2016 for new energy vehicles. A Bayesian network model of power battery safety risks is established using historical data, and the model is trained to assess the safety risks of new energy vehicle power battery systems. Through the accumulation of historical data, regular inspections of the vehicle's power battery system can be conducted, enabling dynamic assessment of safety risks.

[0121] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0122] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for dynamically evaluating safety risks of a power battery system, characterized in that, include: Obtain a sample dataset; the sample dataset includes a battery cell voltage data matrix and probe temperature data matrix of new energy vehicles involved in accidents during historical periods, and a battery cell voltage data matrix and probe temperature data matrix of new energy vehicles in normal operation during historical periods. Based on the battery cell voltage data matrix and probe temperature data matrix in the sample dataset, a sample parameter set is calculated for each new energy vehicle; the sample parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix. For each parameter matrix in the sample parameter set of each new energy vehicle in the historical stage, the parameter distribution difference is calculated, and the top N feature intervals of the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix. For each parameter matrix in the sample parameter set of each new energy vehicle in the historical period, the parameter distribution difference is calculated, and the top N feature intervals with the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix. Specifically, this includes: performing a first operation on each parameter matrix in the sample parameter set of each new energy vehicle in the historical period to determine the N feature intervals corresponding to each parameter matrix; the first operation is: based on all the data in the parameter matrix, statistically analyzing the parameter distribution of the parameter matrix, comparing the parameter distribution differences, and determining multiple feature intervals; calculating the distribution difference value of each feature interval, and determining the feature intervals corresponding to the top N distribution difference values ​​in descending order as the N feature intervals corresponding to the parameter matrix. Discrete membership processing is performed on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval; Based on the membership values ​​of each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage, and the attributes of each new energy vehicle, a Bayesian network is trained to obtain a Bayesian network model. Obtain the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; the row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series; the row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series. Based on the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage, a target parameter set is calculated; the target parameter set includes the power battery system maximum voltage parameter matrix, the power battery system minimum voltage parameter matrix, the power battery system voltage range parameter matrix, the power battery system maximum temperature parameter matrix, and the power battery system temperature range parameter matrix. For each parameter matrix in the target parameter set of the current stage target new energy vehicle, the parameter distribution difference is calculated, and the top N feature intervals of the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix. Discrete membership processing is performed on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage to determine the membership value of each feature interval; Based on the membership values ​​of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model, the safety risk level of the target new energy vehicle in the current stage is determined.

2. The method of claim 1, wherein, Obtaining the sample dataset specifically includes: Collect the full lifecycle data of new energy vehicles with the same model or the same battery model that have been involved in accidents from the new energy vehicle big data platform, and generate an accident new energy vehicle dataset; Collect full lifecycle data of normal new energy vehicles of the same model or the same battery model from the new energy vehicle big data platform to generate a normal new energy vehicle dataset; Based on the accident data set of new energy vehicles, the battery cell voltage data matrix and probe temperature data matrix of new energy vehicles involved in accidents in historical periods are obtained. Based on the normal new energy vehicle dataset, the battery cell voltage data matrix and probe temperature data matrix of normal new energy vehicles in historical periods are obtained. 3.The method of claim 1, wherein, Discrete membership processing is performed on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval, specifically including: The fuzzy logic method is used to perform discrete membership processing on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage, and the membership value of each feature interval is determined. 4.The method of claim 1, wherein, Also includes: The membership values ​​of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, the operating status of the target new energy vehicle in the current stage, and the predicted safety level of the target new energy vehicle in the current stage output by the Bayesian network model are input into the sample dataset of the target new energy vehicle to complete the update of the Bayesian network model; the operating status includes whether an accident has occurred or not.

5. A device for dynamic assessment of safety risks of a power battery system, characterized in that, include: The target new energy vehicle operation data acquisition module is used to acquire the battery cell voltage data matrix and probe temperature data matrix of the target new energy vehicle at the current stage; The row direction of the battery cell voltage data matrix represents the battery cell voltage data, and the column direction of the battery cell voltage data matrix represents the time series. The row direction of the probe temperature data matrix represents the probe temperature data, and the column direction of the probe temperature data matrix represents the time series. The target parameter set calculation module is used to calculate the target parameter set based on the battery cell voltage data matrix and the probe temperature data matrix of the target new energy vehicle at the current stage. The target parameter set includes the highest voltage parameter matrix of the power battery system, the lowest voltage parameter matrix of the power battery system, the voltage range parameter matrix of the power battery system, the highest temperature parameter matrix of the power battery system, and the temperature range parameter matrix of the power battery system. The feature interval determination module is used to calculate the parameter distribution difference for each parameter matrix in the target parameter set of the target new energy vehicle in the current stage, and select the first N feature intervals with the distribution difference values ​​arranged from largest to smallest to determine the N feature intervals corresponding to the parameter matrix. The membership value calculation module is used to perform discrete membership processing on each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage, and determine the membership value of each feature interval. The safety risk level determination module is used to determine the safety risk level of the target new energy vehicle in the current stage based on the membership value of each feature interval corresponding to each parameter matrix of the target new energy vehicle in the current stage and the Bayesian network model. The dynamic safety risk assessment device for the power battery system is also used to acquire a sample dataset; the sample dataset includes a battery cell voltage data matrix and probe temperature data matrix for new energy vehicles involved in accidents during historical periods, as well as a battery cell voltage data matrix and probe temperature data matrix for new energy vehicles in normal operation during historical periods. Based on the battery cell voltage data matrix and probe temperature data matrix in the sample dataset, a sample parameter set is calculated for each new energy vehicle. The sample parameter set includes the power battery system's highest voltage parameter matrix, power battery system's lowest voltage parameter matrix, power battery system's voltage range parameter matrix, power battery system's highest temperature parameter matrix, and power battery system's temperature range parameter matrix. For each parameter matrix in the sample parameter set of each new energy vehicle in the historical period, the parameter distribution difference is calculated, and the top N feature intervals of the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix. For each parameter matrix in the sample parameter set of each new energy vehicle in the historical stage, the parameter distribution difference is calculated, and the top N feature intervals with the distribution difference values ​​arranged from largest to smallest are selected as the N feature intervals corresponding to the parameter matrix. Specifically, this includes: performing a first operation on each parameter matrix in the sample parameter set of each new energy vehicle in the historical stage to determine the N feature intervals corresponding to each parameter matrix; the first operation is: statistically analyzing the parameter distribution of the parameter matrix based on all data in the parameter matrix, comparing the parameter distribution differences, and determining multiple feature intervals; calculating the distribution difference value of each feature interval, and determining the feature intervals corresponding to the top N distribution difference values ​​in descending order as the N feature intervals corresponding to the parameter matrix; performing discrete membership processing on each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage to determine the membership value of each feature interval; and training a Bayesian network based on the membership value of each feature interval corresponding to each parameter matrix of each new energy vehicle in the historical stage and the attributes of each new energy vehicle to obtain a Bayesian network model.

6. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program and the processor runs the computer program to enable the electronic device to perform the dynamic assessment method for safety risks of a power battery system according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the dynamic assessment method for safety risks of a power battery system as described in any one of claims 1 to 4.

Images