Battery fault detection method and device

Abstract

The invention provides a battery fault detection method and device, and relates to the field of battery technology and fault detection. The method comprises the following steps: sampling to obtain a voltage value sequence corresponding to each battery in a battery system in a target time period, wherein the voltage value sequence is used for representing a voltage value corresponding to each sampling time point of the battery; for any battery module, constructing a voltage value matrix corresponding to the battery module according to the voltage value sequences corresponding to the plurality of batteries included in the battery module, and obtaining a sample entropy calculation result corresponding to the battery module based on the voltage value matrix in combination with a sample entropy calculation method; on the basis of the voltage value sequence, an outlier score corresponding to each battery in the battery system is calculated in combination with an ABOD outlier metric algorithm; and analyzing the outlier score and the sample entropy calculation result to obtain a fault detection result of each battery.

Description

Technical Field

[0001] This application relates to the fields of battery technology and fault detection, and particularly to a method and device for fault detection of a battery. Background Art

[0002] During the operation of an electrochemical large-scale energy storage power station, the battery system, as the core power source, its safety is crucial to the stability and reliability of the entire system. The battery system consists of multiple battery packs, each battery pack contains multiple battery modules, and each battery module is composed of several battery cells connected in series and parallel.

[0003] As the usage time increases, various faults may occur during the long-term charge and discharge cycles of the battery, especially the occurrence of micro-short circuits. A micro-short circuit refers to the abnormal contact between the positive and negative electrodes inside the battery, which may be caused by metal particles, impurities in the electrolyte, or other factors. This micro-short circuit fault will cause local overheating inside the battery, thereby leading to a decline in battery performance, and may even cause the battery to explode or catch fire, forming a serious safety hazard. Therefore, timely and accurately detecting the fault status inside the battery, especially micro-short circuits, is an important task to ensure the safety of the battery system and improve the operation stability of the energy storage power station. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems in the related art to some extent.

[0005] To this end, one object of this application is to propose a method for fault detection of a battery. By sampling, a voltage value sequence corresponding to each battery in the battery system during a target time period is obtained. The voltage value sequence is used to represent the voltage value corresponding to each battery at each sampling time point. Among them, the battery system includes multiple battery packs, each battery pack includes multiple battery modules, and each battery module includes multiple batteries. The battery system is in a charging state, a discharging state, or a stationary state during the target time period; for any battery module, a voltage value matrix corresponding to the battery module is constructed according to the voltage value sequences corresponding to the multiple batteries included in the battery module, and based on the voltage value matrix, combined with the sample entropy calculation method, a sample entropy calculation result corresponding to the battery module is obtained; based on the voltage value sequence, combined with the ABOD outlier metric algorithm, an outlier score corresponding to each battery in the battery system is calculated; the outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module are analyzed to obtain a fault detection result for each battery.

[0006] The second object of this application is to propose a device for fault detection of a battery.

[0007] The third object of this application is to propose an electronic device.

[0008] The fourth object of the present application is to propose a non-transitory computer-readable storage medium.

[0009] The fifth object of the present application is to propose a computer program product.

[0010] To achieve the above object, an embodiment of the first aspect of the present application provides a method for detecting a fault of a battery. Voltage value sequences respectively corresponding to each battery in a battery system in a target time period are obtained through sampling. The voltage value sequences are used to represent the voltage values corresponding to each battery at each sampling time point. The battery system includes a plurality of battery packs, each battery pack includes a plurality of battery modules, and each battery module includes a plurality of batteries. The battery system is in a charging state, a discharging state or a stationary state in the target time period. For any battery module, a voltage value matrix corresponding to the battery module is constructed according to the voltage value sequences respectively corresponding to the plurality of batteries included in the battery module, and based on the voltage value matrix and in combination with the sample entropy calculation method, a sample entropy calculation result corresponding to the battery module is obtained. Based on the voltage value sequences and in combination with the ABOD outlier metric algorithm, an outlier score respectively corresponding to each battery in the battery system is calculated. The outlier scores respectively corresponding to each battery in the battery system and the sample entropy calculation result corresponding to each battery module are analyzed to obtain a fault detection result for each battery.

[0011] According to an embodiment of the present application, each row of the voltage value matrix represents the voltage value sequence of one battery, and each column of the voltage value matrix represents the voltage values respectively corresponding to the plurality of batteries included in the battery module at the same sampling time point. Based on the voltage value matrix and in combination with the sample entropy calculation method, obtaining the sample entropy calculation result corresponding to the battery module includes: for each column of the voltage value matrix, determining the maximum voltage value and the minimum voltage value corresponding to the column, and calculating the sum value of the maximum voltage value and the minimum voltage value as the extreme value corresponding to the column, and calculating the difference value between the maximum voltage value and the minimum voltage value as the minimum value corresponding to the column; sorting all the extreme values in column order to obtain an extreme value sequence, and sorting all the minimum values in column order to obtain a minimum value sequence; sampling the SOC value sequences respectively corresponding to each battery included in the battery module in the target time period according to the same sampling method as the voltage value sequences, and querying the mapping relationship between the candidate SOC intervals and the candidate tolerances based on the SOC value sequences to obtain the tolerance corresponding to the battery module; calculating the sample entropy points corresponding to the battery module according to the extreme value sequence, the minimum value sequence and the tolerance, in combination with the sample entropy calculation method. Each sample entropy point pair includes a maximum value sample entropy and a minimum value sample entropy.

[0012] According to an embodiment of the present application, in response to a tolerance of 1, based on the maximum value sequence, the minimum value sequence, and the tolerance, and in combination with the sample entropy calculation method, the sample entropy point pairs corresponding to the battery module are calculated, including: based on the maximum value sequence and the tolerance, and in combination with the sample entropy calculation method, calculating the maximum value sample entropy corresponding to the battery module; based on the minimum value sequence and the tolerance, and in combination with the sample entropy calculation method, calculating the minimum value sample entropy corresponding to the battery module; establishing a sample entropy point pair based on the maximum value sample entropy and the minimum value sample entropy.

[0013] According to an embodiment of the present application, in response to a tolerance of n, based on the maximum value sequence, the minimum value sequence, and the tolerance, and in combination with the sample entropy calculation method, the sample entropy point pairs corresponding to the battery module are calculated, including: according to the SOC value sequence, dividing the maximum value sequence into n maximum value subsequences, and dividing the minimum value sequence into n minimum value subsequences, where the maximum value subsequences and the minimum value subsequences correspond one by one to form n subsequence groups, and each subsequence group corresponds to a tolerance; for the maximum value subsequence in each subsequence group, based on the maximum value subsequence and the tolerance corresponding to the subsequence group, and in combination with the sample entropy calculation method, calculating the partial maximum value sample entropy corresponding to the battery module; for the minimum value subsequence in each subsequence group, based on the minimum value subsequence and the tolerance corresponding to the subsequence group, and in combination with the sample entropy calculation method, calculating the partial minimum value sample entropy corresponding to the battery module; establishing a sample entropy point pair based on the maximum value sample entropy and the minimum value sample entropy.

[0014] According to an embodiment of the present application, the outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module are analyzed to obtain the fault detection results of each battery, including: screening out the to-be-determined abnormal batteries from the batteries according to the outlier scores corresponding to each battery in the battery system; determining the to-be-determined battery module corresponding to the to-be-determined abnormal battery, and combining the sample entropy point pairs corresponding to the to-be-determined battery module to determine the final fault detection result of the to-be-determined abnormal battery.

[0015] According to an embodiment of the present application, combining the sample entropy point pairs corresponding to the to-be-determined battery module to determine the final fault detection result of the to-be-determined abnormal battery includes: judging the point pair type of each sample entropy point pair corresponding to the to-be-determined battery module, and the point pair type is a normal point pair or an abnormal point pair; in response to any sample entropy point pair corresponding to the to-be-determined battery module having an abnormal point pair type, determining that the to-be-determined abnormal battery has a fault and giving a fault warning; in response to the point pair types of the sample entropy point pairs corresponding to the to-be-determined battery module all being normal point pairs, determining that the to-be-determined abnormal battery has no fault.

[0016] According to an embodiment of the present application, determining the pair type of each sample entropy point pair corresponding to a battery module to be determined includes: in response to the sample entropy point pair satisfying an abnormal condition, determining that the pair type of the sample entropy point pair is an abnormal pair; in response to the sample entropy point pair not satisfying the abnormal condition, determining that the pair type of the sample entropy point pair is a normal pair; wherein, the abnormal condition is that the maximum sample entropy in the sample entropy point pair is greater than the first sample entropy threshold and the minimum sample entropy is greater than the second sample entropy threshold.

[0017] To achieve the above object, an embodiment of the second aspect of the present application provides a fault detection device for a battery, including: a data sampling module, configured to sample voltage value sequences respectively corresponding to each battery in a battery system during a target time period, where the voltage value sequences are used to represent the voltage values corresponding to the battery at each sampling time point, wherein the battery system includes a plurality of battery packs, each battery pack includes a plurality of battery modules, each battery module includes a plurality of batteries, and the battery system is in a charging state, a discharging state or a stationary state during the target time period; a first calculation module, configured to, for any battery module, construct a voltage value matrix corresponding to the battery module according to the voltage value sequences respectively corresponding to the plurality of batteries included in the battery module, and based on the voltage value matrix, in combination with a sample entropy calculation method, obtain a sample entropy calculation result corresponding to the battery module; a second calculation module, configured to calculate an outlier score respectively corresponding to each battery in the battery system based on the voltage value sequences in combination with an ABOD outlier metric algorithm; a combined analysis module, configured to analyze the outlier scores respectively corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module to obtain a fault detection result for each battery.

[0018] According to an embodiment of the present application, the first calculation module is further configured to: for each column of the voltage value matrix, determine the maximum voltage value and the minimum voltage value corresponding to the column, and calculate the sum of the maximum voltage value and the minimum voltage value as the maximum value corresponding to the column, and calculate the difference between the maximum voltage value and the minimum voltage value as the minimum value corresponding to the column; sort all the maximum values in column order to obtain a maximum value sequence, and sort all the minimum values in column order to obtain a minimum value sequence; sample the SOC value sequences respectively corresponding to each battery included in the battery module during the target time period according to the same sampling method as the voltage value sequences, and query the mapping relationship between candidate SOC intervals and candidate tolerances based on the SOC value sequences to obtain the tolerance corresponding to the battery module; calculate the sample entropy point pairs corresponding to the battery module according to the maximum value sequence, the minimum value sequence and the tolerance, in combination with the sample entropy calculation method, and each sample entropy point pair includes a maximum sample entropy and a minimum sample entropy.

[0019] According to an embodiment of the present application, the first calculation module is further configured to: calculate the maximum sample entropy corresponding to the battery module based on the maximum value sequence and the tolerance, in combination with the sample entropy calculation method; calculate the minimum sample entropy corresponding to the battery module based on the minimum value sequence and the tolerance, in combination with the sample entropy calculation method; and establish a sample entropy point pair based on the maximum sample entropy and the minimum sample entropy.

[0020] According to an embodiment of the present application, the first calculation module is further configured to: divide the maximum value sequence into n maximum value subsequences according to the SOC value sequence, and divide the minimum value sequence into n minimum value subsequences, where the maximum value subsequences and the minimum value subsequences correspond one by one to form n subsequence groups, and each subsequence group corresponds to a tolerance; for each maximum value subsequence in each subsequence group, calculate the partial maximum sample entropy corresponding to the battery module based on the maximum value subsequence and the tolerance corresponding to the subsequence group, in combination with the sample entropy calculation method; for each minimum value subsequence in each subsequence group, calculate the partial minimum sample entropy corresponding to the battery module based on the minimum value subsequence and the tolerance corresponding to the subsequence group, in combination with the sample entropy calculation method; and establish a sample entropy point pair based on the maximum sample entropy and the minimum sample entropy.

[0021] According to an embodiment of the present application, the combination analysis module is further configured to: screen out the to-be-determined abnormal batteries from the batteries according to the outlier scores corresponding to each battery in the battery system; determine the to-be-determined battery module corresponding to the to-be-determined abnormal battery, and determine the final fault detection result of the to-be-determined abnormal battery in combination with the sample entropy point pair corresponding to the to-be-determined battery module.

[0022] According to an embodiment of the present application, the combination analysis module is further configured to: determine the point pair type of each sample entropy point pair corresponding to the to-be-determined battery module, where the point pair type is a normal point pair or an abnormal point pair; in response to any sample entropy point pair among the sample entropy point pairs corresponding to the to-be-determined battery module having an abnormal point pair type, determine that the to-be-determined abnormal battery has a fault and issue a fault warning; and in response to the point pair types of the sample entropy point pairs corresponding to the to-be-determined battery module all being normal point pairs, determine that the to-be-determined abnormal battery has no fault.

[0023] According to an embodiment of the present application, the combination analysis module is further configured to: in response to the sample entropy point pair satisfying the abnormal condition, determine that the point pair type of the sample entropy point pair is an abnormal point pair; in response to the sample entropy point pair not satisfying the abnormal condition, determine that the point pair type of the sample entropy point pair is a normal point pair; where the abnormal condition is that the maximum sample entropy in the sample entropy point pair is greater than the first sample entropy threshold and the minimum sample entropy is greater than the second sample entropy threshold.

[0024] To achieve the above object, an embodiment of the third aspect of the present application provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to implement the battery fault detection method as described in the embodiment of the first aspect of the present application.

[0025] To achieve the above object, an embodiment of the fourth aspect of the present application provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to implement the battery fault detection method as described in the embodiment of the first aspect of the present application.

[0026] To achieve the above object, an embodiment of the fifth aspect of the present application provides a computer program product, including a computer program, where the computer program, when executed by a processor, implements the battery fault detection method as described in the embodiment of the first aspect of the present application.

[0027] The present application at least achieves the following beneficial effects: By combining the ABOD algorithm and the sample entropy method, the present application can effectively identify abnormal situations in the battery system from multiple perspectives. The ABOD algorithm can identify outliers in battery data by evaluating the angular differences between voltage value sequences of the battery, especially those abnormal batteries that are different from the performance of most batteries, while the sample entropy can quantify the complexity and regularity of voltage data and capture subtle changes in battery behavior. Therefore, this combined method can not only detect obvious faults of the battery, but also identify potential and subtle abnormal changes, improve the accuracy and robustness of fault detection, and reduce false alarms and missed detections.

[0028] In addition, the present application improves the safety, operating efficiency and lifespan of the battery system; through early diagnosis and real-time monitoring, it can effectively reduce false alarms and missed detections, reduce safety risks, achieve automated real-time monitoring, reduce manual intervention, and improve the automation level and response speed of detection, with broad application prospects and significant value. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of the embodiments in conjunction with the drawings, where:

[0030] Figure 1 is a schematic diagram of an exemplary embodiment of a battery fault detection method shown in an embodiment of the present application.

[0031] Figure 2 is a schematic diagram of the establishment of a voltage value sequence shown in an embodiment of the present application.

[0032] Figure 3It is a schematic diagram of an exemplary implementation of a battery fault detection method shown in an embodiment of the present application.

[0033] Figure 4 It is a schematic diagram of the establishment process of a sample entropy point pair shown in an embodiment of the present application.

[0034] Figure 5 It is a schematic diagram of the establishment process of a sample entropy point pair shown in an embodiment of the present application.

[0035] Figure 6 It is a schematic diagram of a battery fault detection device shown in an embodiment of the present application.

[0036] Figure 7 It is a schematic diagram of an electronic device shown in an embodiment of the present application. Detailed implementation manners

[0037] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals indicate the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present application, but should not be construed as a limitation to the present application.

[0038] Figure 1 It is a schematic diagram of an exemplary implementation of a battery fault detection method shown in the present application. As Figure 1 shown, the battery fault detection method includes the following steps:

[0039] S101, sample to obtain a voltage value sequence corresponding to each battery in the battery system during a target time period, and the voltage value sequence is used to represent the voltage value corresponding to each battery at each sampling time point.

[0040] Among them, the battery system includes a plurality of battery packs, each battery pack includes a plurality of battery modules, and each battery module includes a plurality of batteries.

[0041] Among them, the present application divides the battery system into three states, namely a charging state (current less than 0), a discharging state (current greater than 0), and a static state (current equal to 0). In the present application, when performing fault detection on the battery, generally, it does not cross states. That is, the battery system is in one of the charging state, discharging state, and static state during the target time period.

[0042] Figure 2 It is a schematic diagram of the establishment of a voltage value sequence shown in the present application. As Figure 2 shown, assume that the battery system includes 10 battery packs, each battery pack includes 4 battery modules, and each battery module includes 10 batteries ( Figure 2Only 10 batteries included in the battery module 1 are shown, that is, the battery system includes a total of 400 batteries. Taking the target time period from 00:00:01 to 00:15:00 for 15 minutes as an example (assuming that the battery system is in the charging state from 00:00:01 to 00:15:00), and taking a sampling once every 5 seconds as an example, within 15 minutes, each battery's corresponding voltage value sequence contains 15×12 = 180 voltage values. As Figure 2 shown, the voltage value sampled by battery 1 at sampling time point 1 is denoted as voltage 1-1, the voltage value sampled by battery 1 at sampling time point 2 is denoted as voltage 1-2, and so on. The voltage value sequence corresponding to battery 1 in the target time period can be recorded as {voltage 1-1, voltage 1-2... voltage 1-180}.

[0043] Among them, if the voltage data at a certain sampling time point is not collected, the element corresponding to this sampling time point in the voltage value sequence is filled with the valid voltage value corresponding to the previous time point of this sampling time point.

[0044] It is not difficult to understand that in this application, the sampling method for each battery is the same. Therefore, the number of voltage values included in the voltage value sequence corresponding to each battery in the target time period is the same.

[0045] Among them, in order to achieve timely fault detection, the target time period should not be set too long. For example, it can be set to perform fault detection every 15 minutes. However, if the battery system undergoes a state switch within a target time period of 15 minutes, for example, the battery system switches from the charging state to the discharging state at the 8th minute within these 15 minutes, then this 15-minute period can be divided into 2 target time periods based on the time point of the state switch. That is, the first 7 minutes of these 15 minutes are used as a target time period to execute the method proposed in this application for battery fault detection, and the last 8 minutes of these 15 minutes are also used as a target time period to execute the method proposed in this application for battery fault detection. This can ensure that the charge and discharge state of the battery system is consistent within each target time period.

[0046] S102. For any battery module, construct a voltage value matrix corresponding to the battery module according to the voltage value sequences respectively corresponding to the multiple batteries included in the battery module, and based on the voltage value matrix, combine the sample entropy calculation method to obtain the sample entropy calculation result corresponding to the battery module.

[0047] Next, continue to combine Figure 2 for an example introduction. As Figure 2As shown in the figure, the battery module 1 includes 10 batteries. Each of these 10 batteries corresponds to a voltage value sequence. The voltage value sequence corresponding to battery 1 is recorded as {voltage 1-1, voltage 1-2... voltage 1-180}; the voltage value sequence corresponding to battery 2 is recorded as {voltage 2-1, voltage 2-2... voltage 2-180}; and so on. In this application, based on the voltage value sequences corresponding to these 10 batteries respectively, a voltage value matrix corresponding to the battery module 1 is established. Each row of the voltage value matrix corresponding to the battery module 1 is the voltage value sequence corresponding to the battery in that row. The voltage value matrix corresponding to the battery module 1 is a 10×180 matrix.

[0048] After determining the voltage value matrix corresponding to the battery module 1 as described above, based on this voltage value matrix and combined with the sample entropy calculation method, the sample entropy calculation result corresponding to the battery module 1 is obtained.

[0049] And so on, the sample entropy calculation results corresponding to each battery module included in the battery system are obtained.

[0050] S103, based on the voltage value sequence and combined with the ABOD outlier metric algorithm, the outlier scores corresponding to each battery in the battery system are calculated.

[0051] Among them, ABOD (Angle-based Outlier Detection) is an angle-based method used to measure the outlier degree of data points. Its basic idea is to judge whether a data point is an outlier by calculating the angle difference between the data point and its neighboring data points. If there is a significant difference in the angle distribution between the data point and its neighbors, then this data point is considered an outlier.

[0052] Continue to introduce by way of example below in combination with Figure 2 For example, as Figure 2 shown, this battery system includes a total of 400 batteries. Each battery corresponds to 1 voltage value sequence during the target time period, and each voltage value sequence contains 180 voltage values. In this application, based on the 400 voltage value sequences and combined with the ABOD outlier metric algorithm, the outlier scores corresponding to each battery in the battery system are calculated. Among them, the higher the outlier score of 1 battery, the greater the probability that the battery has an abnormality.

[0053] Exemplarily, for each battery, the weighted cosine similarity between the battery and its neighbors is calculated, and the variance is calculated based on the weighted cosine similarity, and then this variance is used as the outlier score of the battery.

[0054] S104, analyze the outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module to obtain the fault detection results of each battery.

[0055] Screen out the batteries that meet the outlier score anomaly conditions from all the batteries as the to-be-determined abnormal batteries, and then determine the battery module corresponding to the to-be-determined abnormal battery as the to-be-determined battery module. If the sample entropy point pair corresponding to the to-be-determined battery module is also an abnormal sample entropy point pair, it is determined that the battery has truly failed, and a fault alarm is issued.

[0056] Next, continue to combine with Figure 2 For example, after obtaining the outlier scores corresponding to 400 batteries in the battery system, compare each outlier score with the preset outlier score threshold. If an outlier score is greater than the preset outlier score threshold, determine the battery corresponding to the outlier score as the to-be-determined abnormal battery. For example, if after comparison, it is found that the outlier score of battery 3 is greater than the preset outlier score threshold, then determine battery 3 as the to-be-determined abnormal battery, and then determine the module where battery 3 is located as battery module 1. Then, take battery module 1 as the to-be-determined battery module. If the sample entropy point pair corresponding to battery module 1 is determined to be an abnormal sample entropy point pair, it is determined that battery 3 has truly failed, and a fault alarm for battery 3 is issued.

[0057] Among them, if the outlier score corresponding to a certain battery is less than or equal to the preset outlier score threshold, it is determined that the battery has not failed.

[0058] Among them, if the outlier score corresponding to a certain battery is greater than the preset outlier score threshold, but the sample entropy point pair corresponding to the battery module where the battery is located is determined to be a normal sample entropy point pair, it is determined that the battery has not failed.

[0059] In the embodiments of the present application, by combining the ABOD algorithm and the sample entropy method, the abnormal conditions in the battery system can be effectively identified from multiple perspectives. The ABOD algorithm can identify the outliers in the battery data by evaluating the angular differences between the voltage value sequences of the batteries, especially those abnormal batteries that are different from the performance of most batteries. The sample entropy can quantify the complexity and regularity of the voltage data and capture the subtle changes in the battery behavior. Therefore, this combined method can not only detect the obvious failures of the batteries, but also identify potential and subtle abnormal changes, improve the accuracy and robustness of the fault detection, and reduce false alarms and missed alarms.

[0060] In addition, the embodiments of the present application improve the safety, operation efficiency and lifespan of the battery system; through early diagnosis and real-time monitoring, it can effectively reduce false alarms and missed alarms, reduce safety risks, realize automated real-time monitoring, reduce manual intervention, and improve the automation level and response speed of the detection, and has broad application prospects and significant value.

[0061] Figure 3 It is a schematic diagram of an exemplary embodiment of a battery fault detection method shown in the present application, asFigure 3 As shown in Figure 3 , the fault detection method of the battery includes the following steps:

[0062] S301: Sample to obtain the voltage value sequence corresponding to each battery in the battery system during the target time period. The voltage value sequence is used to represent the voltage value corresponding to the battery at each sampling time point.

[0063] Regarding the specific implementation method of step S301, reference can be made to the specific introduction in the relevant part of the above-mentioned embodiment, and details will not be elaborated here.

[0064] S302: For any battery module, construct a voltage value matrix corresponding to the battery module according to the voltage value sequences corresponding to the multiple batteries included in the battery module.

[0065] Among them, each row of the voltage value matrix represents the voltage value sequence of 1 battery, and each column of the voltage value matrix represents the voltage values corresponding to the multiple batteries included in the battery module at the same sampling time point respectively.

[0066] S303: For each column of the voltage value matrix, determine the maximum voltage value and the minimum voltage value corresponding to this column, and calculate the sum value of the maximum voltage value and the minimum voltage value as the extreme value corresponding to this column, and calculate the difference value between the maximum voltage value and the minimum voltage value as the minimum value corresponding to this column.

[0067] S304: For each voltage value matrix, sort all the extreme values corresponding to this voltage value matrix in column order to obtain an extreme value sequence, and sort all the minimum values corresponding to this voltage value matrix in column order to obtain a minimum value sequence.

[0068] Continuing with the voltage value matrix corresponding to battery module 1 as a 10×180 matrix as an example, the extreme value sequence corresponding to this battery module 1 contains 180 extreme values (1 extreme value corresponding to each sampling time point), and the minimum value sequence corresponding to this battery module 1 contains 180 minimum values (1 minimum value corresponding to each sampling time point).

[0069] And so on, each of the other battery modules also corresponds to 1 extreme value sequence (containing 180 extreme values) and 1 minimum value sequence (containing 180 minimum values).

[0070] S305: Sample according to the same sampling method as the voltage value sequence to obtain the SOC value sequence corresponding to each battery included in the battery module during the target time period, and query the mapping relationship between the candidate SOC interval and the candidate tolerance based on the SOC value sequence to obtain the tolerance corresponding to the battery module.

[0071] Among them, SOC, the full name is State of Charge, which is the state of charge of the battery, also called the remaining battery power.

[0072] In this application, exemplarily, the following multiple candidate SOC intervals can be set: 0 ≤ SOC < 25, 25 ≤ SOC < 55, 55 ≤ SOC < 65, 65 ≤ SOC < 90, 90 ≤ SOC < 100.

[0073] Among them, different candidate SOC intervals can correspond to different tolerances r, and the tolerance r is used for subsequent sample entropy calculation.

[0074] Exemplarily, the tolerance r corresponding to the candidate SOC interval 0 ≤ SOC < 25 is 0.045; the tolerance r corresponding to the candidate SOC interval 25 ≤ SOC < 55 is 0.035.

[0075] Exemplarily, the battery module 1 includes 10 batteries. According to the same sampling method as the voltage value sequence, the SOC value sequence corresponding to each battery included in the battery module 1 is sampled during the target time period. For example, the SOC value sequence corresponding to each battery contains 180 SOC values, then the battery module 1 corresponds to a 10×180 SOC value matrix. Compare each value in the 10×180 SOC value matrix with the preset candidate SOC intervals to see how many candidate SOC intervals the 10×180 SOC value matrix involves, that is, it means how many tolerances the battery module 1 corresponds to.

[0076] S306. According to the maximum value sequence, minimum value sequence and tolerance, combined with the sample entropy calculation method, calculate the sample entropy point pairs corresponding to the battery module. Each sample entropy point pair includes a maximum value sample entropy and a minimum value sample entropy.

[0077] Among them, it should be clear first that after a given sequence, the calculation process of calculating the sample entropy corresponding to the sequence based on the sample entropy algorithm is a relatively mature technology. In the following introduction, the formula calculation process of the sample entropy will not be elaborated, but the construction process of the sequence used for sample entropy calculation will be emphasized.

[0078] The following will introduce S306 in two possible situations.

[0079] The first situation, that is, the situation when the tolerance corresponding to the battery module is 1. At this time, according to the maximum value sequence, minimum value sequence and tolerance, combined with the sample entropy calculation method, the steps to calculate the sample entropy point pairs corresponding to the battery module are as follows: Based on the maximum value sequence and tolerance, combined with the sample entropy calculation method, calculate the maximum value sample entropy corresponding to the battery module; Based on the minimum value sequence and tolerance, combined with the sample entropy calculation method, calculate the minimum value sample entropy corresponding to the battery module; Establish a sample entropy point pair based on the maximum value sample entropy and the minimum value sample entropy.

[0080] Combined with Figure 4Understand the first case, Figure 4 which is a schematic diagram of the establishment process of a sample entropy point pair shown in this application. As Figure 4 shown, assume that all values in the 10×180 SOC value matrix corresponding to battery module 1 are in the range of 0≤SOC<25, that is, the tolerance corresponding to battery module 1 is 1. Assume the tolerance r = 0.045. The following two methods can be used to calculate the sample entropy point pair corresponding to battery module 1 ( Figure 4 The example in [ ] shows multiple sample entropy point pairs corresponding to battery module 1, corresponding to the following method 2).

[0081] Method 1: As can be seen from step S304, the maximum value sequence corresponding to battery module 1 contains 180 maximum values, and the minimum value sequence corresponding to battery module 1 contains 180 minimum values. Then, based on the maximum value sequence, combined with the sample entropy calculation method (tolerance r = 0.045, the parameter m involved in the sample entropy calculation method can be set by oneself), 1 maximum value sample entropy can be obtained. Similarly, based on the minimum value sequence, combined with the sample entropy calculation method (tolerance r = 0.045, the parameter m involved in the sample entropy calculation method can be set by oneself), 1 minimum value sample entropy can be obtained. Then, this 1 maximum value sample entropy and this 1 minimum value sample entropy form 1 sample entropy point pair, that is, battery module 1 corresponds to 1 sample entropy point pair.

[0082] Method 2: As can be seen from step S304, the maximum value sequence corresponding to battery module 1 contains 180 maximum values, and the minimum value sequence corresponding to battery module 1 contains 180 minimum values.

[0083] In this application, the maximum value sequence can be intercepted by a sliding window according to a preset step size to obtain multiple subsequences (for example, if the preset step size is 5, 176 subsequences can be intercepted based on the maximum value sequence. The first subsequence is {maximum value 1, maximum value 2, maximum value 3, maximum value 4, maximum value 5}, the second subsequence is {maximum value 2, maximum value 3, maximum value 4, maximum value 5, maximum value 6}, and so on). Then, for each subsequence, combined with the sample entropy calculation method (tolerance r = 0.045, the parameter m involved in the sample entropy calculation method can be set by oneself), 1 maximum value sample entropy can be obtained. If there are 176 subsequences, then battery module 1 corresponds to 176 maximum value sample entropies.

[0084] Then, based on the same processing idea, the same sliding interception and sample entropy calculation processing are performed on the minimum value sequence corresponding to battery module 1 (the preset step size during sliding and the tolerance during sample calculation are the same as those during the processing of the maximum value sequence), and 176 minimum value sample entropies are obtained. Among them, the 176 maximum value sample entropies and the 176 minimum value sample entropies correspond one by one, that is, a total of 176 sample entropy point pairs can be formed.

[0085] Among them, the maximum sample entropy and the minimum sample entropy can be corresponding according to the sliding window. For example, the maximum sample entropy calculated from the first subsequence {maximum value 1, maximum value 2, maximum value 3, maximum value 4, maximum value 5} corresponding to the maximum value sequence corresponds to the minimum sample entropy calculated from the first subsequence {minimum value 1, minimum value 2, minimum value 3, minimum value 4, minimum value 5} corresponding to the minimum value sequence. And so on.

[0086] In the second case, that is, the case where the tolerance is n, at this time, according to the maximum value sequence, the minimum value sequence and the tolerance, combined with the sample entropy calculation method, the steps to calculate the sample entropy point pairs corresponding to the battery module are as follows: According to the SOC value sequence, divide the maximum value sequence into n maximum value subsequences, and divide the minimum value sequence into n minimum value subsequences. Among them, the maximum value subsequences and the minimum value subsequences correspond one by one to form n subsequence groups, and each subsequence group corresponds to a tolerance; for the maximum value subsequence in each subsequence group, based on the maximum value subsequence and the tolerance corresponding to the subsequence group, combined with the sample entropy calculation method, calculate the partial maximum sample entropy corresponding to the battery module; for the minimum value subsequence in each subsequence group, based on the minimum value subsequence and the tolerance corresponding to the subsequence group, combined with the sample entropy calculation method, calculate the partial minimum sample entropy corresponding to the battery module; establish the sample entropy point pairs based on the maximum sample entropy and the minimum sample entropy.

[0087] Combined with Figure 5 For understanding the second case, Figure 5 is a schematic diagram of the establishment process of a sample entropy point pair shown in this application. As Figure 5 shown, assuming that in the 10×180 SOC value matrix corresponding to battery module 1, the SOC values of the columns corresponding to sampling point time 1 to sampling point time 100 are all in the range of 0≤SOC<25, while the SOC values of the columns corresponding to sampling point time 101 to sampling point time 180 are all in the range of 25≤SOC<55. That is, correspondingly, 2 tolerances are obtained based on querying the mapping relationship between the candidate SOC intervals and the candidate tolerances according to the SOC value sequence.

[0088] Then correspondingly, the maximum value sequence can be divided into 2 maximum value subsequences (the first maximum value subsequence is {maximum value 1, maximum value 2... maximum value 100}, and the second maximum value subsequence is {maximum value 101, maximum value 102... maximum value 180}) based on different tolerances, and the minimum value sequence can be divided into 2 minimum value subsequences (the first minimum value subsequence is {minimum value 1, minimum value 2... minimum value 100}, and the second minimum value subsequence is {minimum value 101, minimum value 102... minimum value 180}).

[0089] Among them, the first maximum value subsequence corresponds to the first minimum value subsequence and has the same tolerance, which is assumed to be 0.045.

[0090] Among them, the second maximum value subsequence corresponds to the second minimum value subsequence and has the same tolerance, which is assumed to be 0.035.

[0091] For the first maximum value subsequence and the first minimum value subsequence, the first part of the sample entropy point pairs corresponding to the battery module 1 can be obtained by adopting the idea of any one of Method 1 and Method 2 in the above first case. When calculating based on the sample entropy method by adopting the idea of any one of Method 1 and Method 2, the tolerance r = 0.045, and the parameter m involved in the sample entropy calculation method can be set by oneself.

[0092] For the second maximum value subsequence and the second minimum value subsequence, the second part of the sample entropy point pairs corresponding to the battery module 1 can also be obtained by adopting the idea of any one of Method 1 and Method 2 in the above first case. When calculating based on the sample entropy method by adopting the idea of any one of Method 1 and Method 2, the tolerance r = 0.035, and the parameter m involved in the sample entropy calculation method can be set by oneself.

[0093] Finally, the sample entropy point pairs of the first part and the second part are combined as the sample entropy point pairs of the battery module 1.

[0094] S307. Based on the voltage value sequence and combined with the ABOD outlier metric algorithm, calculate the outlier scores corresponding to each battery in the battery system.

[0095] S308. Analyze the outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module to obtain the fault detection results of each battery.

[0096] First, screen out the to-be-determined abnormal batteries from the batteries according to the outlier scores corresponding to each battery in the battery system. For example, an outlier score threshold can be set, and each outlier score is compared with the preset outlier score threshold. If an outlier score is greater than the preset outlier score threshold, it is determined that the battery corresponding to the outlier score is a to-be-determined abnormal battery.

[0097] Then, determine the to-be-determined battery module corresponding to the to-be-determined abnormal battery, and combine the sample entropy points corresponding to the to-be-determined battery module to determine the final fault detection result of the to-be-determined abnormal battery. Specifically, first, it is necessary to determine the pair type of each sample entropy point pair corresponding to the to-be-determined battery module. The pair type is a normal pair or an abnormal pair. If there is any sample entropy point pair with an abnormal pair type among the sample entropy point pairs corresponding to the to-be-determined battery module, it is determined that the to-be-determined abnormal battery has a fault, and a fault warning is issued; if the pair types of all sample entropy point pairs corresponding to the to-be-determined battery module are normal pairs, it is determined that the to-be-determined abnormal battery has no fault.

[0098] Among them, determining the pair type of each sample entropy point pair corresponding to the to-be-determined battery module includes the following steps: Set the abnormal condition as: the maximum sample entropy in the sample entropy point pair is greater than the first sample entropy threshold (which can be optionally taken as 0) and the minimum sample entropy is greater than the second sample entropy threshold (which can be optionally taken as 0). When determining the pair type of each sample entropy point pair, if a certain sample entropy point pair meets the abnormal condition, it is determined that the pair type of this sample entropy point pair is an abnormal pair; if a certain sample entropy point pair does not meet the abnormal condition, it is determined that the pair type of this sample entropy point pair is a normal pair.

[0099] In the embodiment of the present application, by combining the ABOD algorithm and the sample entropy method, the abnormal conditions in the battery system can be effectively identified from multiple perspectives. The ABOD algorithm can identify the outliers in the battery data by evaluating the angular differences between the battery voltage sequences, especially those abnormal batteries that are different from the performance of most batteries, while the sample entropy can quantify the complexity and regularity of the voltage data and capture the subtle changes in the battery behavior. Therefore, this combined method can not only detect the obvious faults of the battery, but also identify potential and subtle abnormal changes, improve the accuracy and robustness of fault detection, and reduce false alarms and missed detections.

[0100] In addition, the embodiment of the present application improves the safety, operation efficiency and lifespan of the battery system; through early diagnosis and real-time monitoring, it can effectively reduce false alarms and missed detections, reduce safety risks, realize automated real-time monitoring, reduce manual intervention, and improve the automation level and response speed of detection, and has broad application prospects and significant value.

[0101] Figure 6 is a schematic diagram of a fault detection device for a battery shown in the present application, as Figure 6 shown, the fault detection device 600 for the battery includes a data sampling module 601, a first calculation module 602, a second calculation module 603 and a combined analysis module 604, where:

[0102] A data sampling module 601, configured to sample voltage value sequences respectively corresponding to each battery in the battery system during a target time period, where the voltage value sequences are used to represent the voltage values corresponding to each battery at each sampling time point. The battery system includes a plurality of battery packs, each battery pack includes a plurality of battery modules, each battery module includes a plurality of batteries, and the battery system is in a charging state, a discharging state, or a stationary state during the target time period.

[0103] A first calculation module 602, configured to, for any battery module, construct a voltage value matrix corresponding to the battery module according to the voltage value sequences respectively corresponding to the plurality of batteries included in the battery module, and based on the voltage value matrix, in combination with a sample entropy calculation method, obtain a sample entropy calculation result corresponding to the battery module.

[0104] A second calculation module 603, configured to calculate, based on the voltage value sequences and in combination with an ABOD outlier metric algorithm, an outlier score respectively corresponding to each battery in the battery system.

[0105] A combined analysis module 604, configured to analyze the outlier score respectively corresponding to each battery in the battery system and the sample entropy calculation result corresponding to each battery module, to obtain a fault detection result for each battery.

[0106] By using the ABOD algorithm and the sample entropy method in combination, the present device can effectively identify abnormal conditions in the battery system from multiple perspectives. The ABOD algorithm can identify outliers in battery data by evaluating the angular differences between battery voltage sequences, especially those abnormal batteries that behave differently from most batteries, while the sample entropy can quantify the complexity and regularity of voltage data and capture subtle changes in battery behavior. Therefore, this combined method can not only detect obvious faults of the battery, but also identify potential and subtle abnormal changes, improve the accuracy and robustness of fault detection, and reduce false alarms and missed detections.

[0107] Further, the first calculation module 602 is further configured to: for each column of the voltage value matrix, determine the maximum voltage value and the minimum voltage value corresponding to the column, and calculate the sum of the maximum voltage value and the minimum voltage value as the maximum value corresponding to the column, and calculate the difference between the maximum voltage value and the minimum voltage value as the minimum value corresponding to the column; sort all the maximum values in the column order to obtain a maximum value sequence, and sort all the minimum values in the column order to obtain a minimum value sequence; sample the SOC value sequence corresponding to each battery included in the battery module in the target time period according to the same sampling method as the voltage value sequence, and query the mapping relationship between the candidate SOC intervals and the candidate tolerances based on the SOC value sequence to obtain the tolerance corresponding to the battery module; according to the maximum value sequence, the minimum value sequence and the tolerance, and in combination with the sample entropy calculation method, calculate the sample entropy point pairs corresponding to the battery module, and each sample entropy point pair includes a maximum value sample entropy and a minimum value sample entropy.

[0108] Further, the first calculation module 602 is further configured to: based on the maximum value sequence and the tolerance, and in combination with the sample entropy calculation method, calculate the maximum value sample entropy corresponding to the battery module; based on the minimum value sequence and the tolerance, and in combination with the sample entropy calculation method, calculate the minimum value sample entropy corresponding to the battery module; establish a sample entropy point pair based on the maximum value sample entropy and the minimum value sample entropy.

[0109] Further, the first calculation module 602 is further configured to: according to the SOC value sequence, divide the maximum value sequence into n maximum value subsequences, and divide the minimum value sequence into n minimum value subsequences, where the maximum value subsequences and the minimum value subsequences correspond one by one to form n subsequence groups, and each subsequence group corresponds to a tolerance; for the maximum value subsequence in each subsequence group, based on the maximum value subsequence and the tolerance corresponding to the subsequence group, and in combination with the sample entropy calculation method, calculate the partial maximum value sample entropy corresponding to the battery module; for the minimum value subsequence in each subsequence group, based on the minimum value subsequence and the tolerance corresponding to the subsequence group, and in combination with the sample entropy calculation method, calculate the partial minimum value sample entropy corresponding to the battery module; establish a sample entropy point pair based on the maximum value sample entropy and the minimum value sample entropy.

[0110] Further, the combination analysis module 604 is further configured to: screen out the to-be-determined abnormal batteries from the batteries according to the outlier scores corresponding to each battery in the battery system; determine the to-be-determined battery module corresponding to the to-be-determined abnormal battery, and determine the final fault detection result of the to-be-determined abnormal battery in combination with the sample entropy point pair corresponding to the to-be-determined battery module.

[0111] Further, the combination analysis module 604 is further configured to: determine the pair type of each sample entropy point pair corresponding to the battery module to be determined, where the pair type is a normal pair or an abnormal pair; in response to any sample entropy point pair among the sample entropy point pairs corresponding to the battery module to be determined having an abnormal pair type, determine that the battery module to be determined has a fault and issue a fault warning; in response to the pair types of all the sample entropy point pairs corresponding to the battery module to be determined being normal pairs, determine that the battery module to be determined has no fault.

[0112] Further, the combination analysis module 604 is further configured to: in response to the sample entropy point pair satisfying the abnormal condition, determine that the pair type of the sample entropy point pair is an abnormal pair; in response to the sample entropy point pair not satisfying the abnormal condition, determine that the pair type of the sample entropy point pair is a normal pair; where the abnormal condition is that the maximum sample entropy in the sample entropy point pair is greater than the first sample entropy threshold and the minimum sample entropy is greater than the second sample entropy threshold.

[0113] To implement the above embodiments, an electronic device 700 is further proposed in an embodiment of the present application. As Figure 7 shown, the electronic device 700 includes: a processor 701 and a memory 702 communicatively connected to the processor. The memory 702 stores instructions executable by at least one processor. The instructions are executed by at least one processor 701 to implement the battery fault detection method as shown in the above embodiments.

[0114] To implement the above embodiments, a non-transitory computer-readable storage medium storing computer instructions is further proposed in an embodiment of the present application, where the computer instructions are used to cause a computer to implement the battery fault detection method as shown in the above embodiments.

[0115] To implement the above embodiments, a computer program product is further proposed in an embodiment of the present application, including a computer program that implements the battery fault detection method as shown in the above embodiments when executed by a processor.

[0116] In the description of the present application, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present application.

[0117] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality of" means two or more, unless otherwise specifically defined.

[0118] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0119] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A battery fault detection method, characterized in that: include: Sampling to obtain a voltage value sequence corresponding to each battery in the battery system in a target time period, wherein the voltage value sequence is used to represent the voltage value corresponding to the battery at each sampling time point, wherein the battery system includes a plurality of battery packs, each of the battery packs includes a plurality of battery modules, each of the battery modules includes a plurality of batteries, and the battery system is in a charging state, a discharging state, or a static state in the target time period; For any of the battery modules, a voltage value matrix corresponding to the battery module is constructed according to voltage value sequences corresponding to the multiple batteries included in the battery module, and based on the voltage value matrix, a sample entropy calculation result corresponding to the battery module is obtained in combination with a sample entropy calculation method; Based on the voltage value sequence and in combination with the ABOD outlier metric algorithm, the outlier score corresponding to each battery in the battery system is calculated; The outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module are analyzed to obtain a fault detection result for each battery.

2. The method according to claim 1, characterized in that Each row of the voltage value matrix represents a voltage value sequence of a battery, and each column of the voltage value matrix represents voltage values ​​corresponding to a plurality of batteries included in the battery module at the same sampling time point. Based on the voltage value matrix, in combination with the sample entropy calculation method, the sample entropy calculation result corresponding to the battery module is obtained, including: For each column of the voltage value matrix, determine the maximum voltage value and the minimum voltage value corresponding to the column, and calculate the sum of the maximum voltage value and the minimum voltage value as the maximum value corresponding to the column, and calculate the difference between the maximum voltage value and the minimum voltage value as the minimum value corresponding to the column; Sorting all the maximum values ​​in column order to obtain a maximum value sequence, and sorting all the minimum values ​​in column order to obtain a minimum value sequence; Sampling the SOC value sequence corresponding to each battery included in the battery module within the target time period in the same sampling method as the voltage value sequence, and querying the mapping relationship between the candidate SOC interval and the candidate tolerance based on the SOC value sequence to obtain the tolerance corresponding to the battery module; According to the maximum value sequence, the minimum value sequence and the tolerance, combined with the sample entropy calculation method, the sample entropy point pair corresponding to the battery module is calculated, and each of the sample entropy point pairs includes a maximum value sample entropy and a minimum value sample entropy.

3. The method according to claim 2, characterized in that In response to the tolerance being 1, the sample entropy point pair corresponding to the battery module is calculated based on the maximum value sequence, the minimum value sequence and the tolerance in combination with a sample entropy calculation method, including: Based on the maximum value sequence and the tolerance, combined with the sample entropy calculation method, the maximum value sample entropy corresponding to the battery module is calculated; Based on the minimum value sequence and the tolerance, combined with the sample entropy calculation method, the minimum value sample entropy corresponding to the battery module is calculated; A sample entropy point pair is established based on the maximum sample entropy and the minimum sample entropy.

4. The method according to claim 2, characterized in that: In response to the tolerance being n, the sample entropy point pairs corresponding to the battery module are calculated based on the maximum value sequence, the minimum value sequence and the tolerance in combination with a sample entropy calculation method, including: According to the SOC value sequence, the maximum value sequence is divided into n maximum value subsequences, and the minimum value sequence is divided into n minimum value subsequences, wherein the maximum value subsequences and the minimum value subsequences correspond to each other one by one to form n subsequence groups, and each subsequence group corresponds to one tolerance; For each maximum subsequence in the subsequence group, based on the maximum subsequence and the tolerance corresponding to the subsequence group, combined with a sample entropy calculation method, calculate the partial maximum sample entropy corresponding to the battery module; For each minimum value subsequence in the subsequence group, based on the minimum value subsequence and the tolerance corresponding to the subsequence group, combined with the sample entropy calculation method, calculate the partial minimum value sample entropy corresponding to the battery module; A sample entropy point pair is established based on the maximum sample entropy and the minimum sample entropy.

5. The method according to claim 3 or 4, characterized in that: The outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module are analyzed to obtain the fault detection result of each battery, including: Screening out abnormal batteries to be determined from the batteries according to the outlier scores corresponding to each battery in the battery system; The pending battery module corresponding to the pending abnormal battery is determined, and the final fault detection result of the pending abnormal battery is determined in combination with the sample entropy point pair corresponding to the pending abnormal battery module.

6. The method according to claim 5, characterized in that The determining of the final fault detection result of the abnormal battery to be determined by combining the sample entropy point pair corresponding to the battery module to be determined includes: Determine the point pair type of each of the sample entropy point pairs corresponding to the pending battery module, the point pair type being a normal point pair or an abnormal point pair; In response to the point pair type of any sample entropy point pair corresponding to the pending battery module being an abnormal point pair, determining that the pending abnormal battery has a fault, and performing a fault warning; In response to the point pair types of the sample entropy point pairs corresponding to the pending battery module being all normal point pairs, it is determined that the pending abnormal battery has not failed.

7. The method according to claim 6, characterized in that The determining the point pair type of each of the sample entropy point pairs corresponding to the pending battery module includes: In response to the sample entropy point pair satisfying an abnormal condition, determining that the point pair type of the sample entropy point pair is an abnormal point pair; In response to the sample entropy point pair not satisfying an abnormal condition, determining that the point pair type of the sample entropy point pair is a normal point pair; The abnormal condition is that the maximum sample entropy in the sample entropy point pair is greater than a first sample entropy threshold and the minimum sample entropy is greater than a second sample entropy threshold.

8. A battery fault detection device, characterized in that: include: A data sampling module, used for sampling and obtaining a voltage value sequence corresponding to each battery in the battery system in a target time period, wherein the voltage value sequence is used to represent the voltage value corresponding to the battery at each sampling time point, wherein the battery system includes a plurality of battery packs, each of the battery packs includes a plurality of battery modules, each of the battery modules includes a plurality of batteries, and the battery system is in a charging state, a discharging state or a static state in the target time period; A first calculation module is used to construct a voltage value matrix corresponding to any of the battery modules according to voltage value sequences corresponding to a plurality of batteries included in the battery module, and obtain a sample entropy calculation result corresponding to the battery module based on the voltage value matrix and in combination with a sample entropy calculation method; A second calculation module is used to calculate the outlier score corresponding to each battery in the battery system based on the voltage value sequence and in combination with the ABOD outlier measurement algorithm; Combined with the analysis module, it is used to analyze the outlier scores corresponding to each battery in the battery system and the sample entropy calculation results corresponding to each battery module to obtain the fault detection result of each battery.

9. An electronic device, comprising: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-7.

Images