Vehicle lithium battery pack fault diagnosis method and device based on intelligent monitoring

Abstract

The invention discloses a vehicle lithium battery pack fault diagnosis method and device based on intelligent monitoring, and relates to the related technical field of battery fault diagnosis, and the method comprises the steps: collecting a battery operation state parameter set of a target vehicle lithium battery pack and a corresponding battery fault identification data set; establishing a plurality of first identification fault types corresponding to the first battery operation state parameters; analyzing a plurality of first trigger probabilities of the plurality of first identification fault types; constructing a first fault diagnosis channel; and integrating to a fault diagnosis integrated network, analyzing the real-time battery operation state parameters of the target vehicle lithium battery pack, and outputting a trigger fault type sequence with the trigger probability from large to small. The technical problems that in the prior art, the fault type of the vehicle lithium battery pack is difficult to accurately recognize, fault diagnosis under complex working conditions cannot be adapted, the fault diagnosis efficiency is low, and the accuracy is poor are solved, and the technical effect of improving the fault diagnosis efficiency and accuracy of the lithium battery pack is achieved.

Description

Technical Field

[0001] The present application relates to the technical field related to battery fault diagnosis, and specifically to a method and device for diagnosing faults of automotive lithium battery packs based on intelligent monitoring. Background Art

[0002] In the development of modern electric vehicles, the reliability of automotive lithium battery packs is crucial. With the rapid increase in the number of electric vehicles, lithium battery pack failures may lead to problems such as vehicle performance degradation and safety hazards, seriously affecting user experience and traffic safety. Traditional lithium battery pack fault diagnosis methods mostly rely on manual experience or simple threshold judgments, which are difficult to adapt to complex and changeable battery operating conditions. The operating state parameters of automotive lithium battery packs are complex and changeable under different ambient temperatures, charge and discharge rates, and driving conditions, which makes it extremely difficult to accurately capture the early characteristics of faults; when a fault occurs, traditional methods cannot quickly and accurately locate the fault type, delaying the time for maintenance.

[0003] Therefore, in the current relevant technologies, there are technical problems such as difficulty in accurately identifying the fault type of automotive lithium battery packs, inability to adapt to fault diagnosis under complex working conditions, and low fault diagnosis efficiency and poor accuracy. Summary of the invention

[0004] The present application solves the technical problems in the prior art of difficulty in accurately identifying the fault type of a vehicle lithium battery pack, inability to adapt to fault diagnosis under complex working conditions, and low fault diagnosis efficiency and poor accuracy by providing a method and device for fault diagnosis of a vehicle lithium battery pack based on intelligent monitoring, thereby achieving the technical effect of improving the efficiency and accuracy of lithium battery pack fault diagnosis.

[0005] The present application provides a vehicle lithium battery pack fault diagnosis method based on intelligent monitoring, the method comprising: collecting a battery operating status parameter set and a corresponding battery fault identification data set of a target vehicle lithium battery pack; performing consistent clustering of status parameters in combination with the battery operating status parameter set and the battery fault identification data set, and establishing multiple first identification fault types corresponding to a first battery operating status parameter; analyzing multiple first trigger probabilities of the multiple first identification fault types under the first battery operating status parameter; constructing a first fault diagnosis channel with the first battery operating status parameter, the multiple first identification fault types and the multiple first trigger probabilities; integrating the first fault diagnosis channel into a fault diagnosis integrated network, analyzing the real-time battery operating status parameters of the target vehicle lithium battery pack, and outputting a sequence of trigger fault types with trigger probabilities ranging from large to small.

[0006] In a possible implementation, the fault diagnosis method for vehicle-mounted lithium battery packs based on intelligent monitoring further performs the following processing: perform deviation analysis on any two sets of battery operating state parameters in the battery operating state parameter set. If the deviation meets a preset deviation threshold, cluster any two sets of battery operating state parameters to generate multiple clusters of battery operating state parameters; extract the central data of any one cluster of battery operating state parameters in the multiple clusters of battery operating state parameters as the first battery operating state parameter; based on the one-to-one correspondence between the battery operating state parameter set and the battery fault identification data set, determine the battery fault identification data corresponding to any one cluster of battery operating state parameters, and perform distribution analysis of fault types to generate the multiple first identified fault types.

[0007] In a possible implementation, the fault diagnosis method for vehicle-mounted lithium battery packs based on intelligent monitoring further performs the following processing: perform deviation analysis on any two sets of battery operating state parameters in the battery operating state parameter set, where the deviation includes the deviation of each type of parameter corresponding to each battery cell.

[0008] In a possible implementation, the fault diagnosis method for vehicle-mounted lithium battery packs based on intelligent monitoring further performs the following processing: respectively use the multiple first identified fault types as given fault types, count the occurrence probabilities of the first battery operating state parameter to generate multiple first conditional probabilities; calculate the prior probabilities of the multiple first identified fault types to generate multiple first prior probabilities; combine the multiple first conditional probabilities and the multiple first prior probabilities, introduce a trigger probability expression, and calculate the multiple first trigger probabilities of the multiple first identified fault types.

[0009] In a possible implementation, the fault diagnosis method for vehicle-mounted lithium battery packs based on intelligent monitoring further performs the following processing: The trigger probability expression is as follows: ; where, is the first trigger probability of the th first identified fault type under the first battery operating state parameter; is the occurrence probability of the first battery operating state parameter under the th first identified fault type; is the prior probability of the th first identified fault type; is the marginal probability of the first battery operating state parameter, which is used to ensure that the sum of the multiple first trigger probabilities of the multiple first identified fault types is 1.

[0010] In a possible implementation, the method for diagnosing faults in a vehicle-mounted lithium battery pack based on intelligent monitoring further performs the following processing: analyzing the degree of fuzziness of the first battery operating state parameters under the multiple first identified fault types, and establishing a fuzzy membership function; optimizing the trigger probability expression with the fuzzy membership function, and calculating multiple first trigger probabilities of the multiple first identified fault types.

[0011] In a possible implementation, the method for diagnosing faults in a vehicle-mounted lithium battery pack based on intelligent monitoring further performs the following processing: collecting the real-time operating environment information of the target vehicle-mounted lithium battery pack; based on the real-time operating environment information, analyzing the influence relationship between the real-time environment and the trigger probability of each fault type in the trigger fault type sequence, and generating an environment-probability influence relationship; optimizing the trigger probability with the environment-probability influence relationship and then re-arranging them from large to small to generate an updated trigger fault type sequence.

[0012] In a possible implementation, the method for diagnosing faults in a vehicle-mounted lithium battery pack based on intelligent monitoring further performs the following processing: collecting historical fault occurrence environment records based on each fault type; based on the historical fault occurrence environment records, analyzing the influence relationship between the fault occurrence probability of each fault type and the environmental conditions, and constructing the environment-probability influence relationship.

[0013] In a possible implementation, the method for diagnosing faults in a vehicle-mounted lithium battery pack based on intelligent monitoring further performs the following processing: the battery operating state parameter set includes voltage, current, temperature, SOC, and SOH, and the battery operating state parameter set is collected by the battery management system of the target vehicle-mounted lithium battery pack.

[0014] This application also provides a device for diagnosing faults in a vehicle-mounted lithium battery pack based on intelligent monitoring, including: a battery data collection module, configured to collect a battery operating state parameter set and a corresponding battery fault identification data set of a target vehicle-mounted lithium battery pack; a state parameter clustering module, configured to perform consistent clustering of state parameters by combining the battery operating state parameter set and the battery fault identification data set, and establish multiple first identified fault types corresponding to the first battery operating state parameters; a trigger probability determination module, configured to analyze multiple first trigger probabilities of the multiple first identified fault types under the first battery operating state parameters; a fault diagnosis channel construction module, configured to construct a first fault diagnosis channel with the first battery operating state parameters, the multiple first identified fault types, and the multiple first trigger probabilities; an operating state parameter analysis module, configured to integrate the first fault diagnosis channel into a fault diagnosis integration network, analyze the real-time battery operating state parameters of the target vehicle-mounted lithium battery pack, and output a trigger fault type sequence with trigger probabilities arranged from large to small.

[0015] The method and device for fault diagnosis of vehicle lithium battery packs based on intelligent monitoring proposed in this application collect the battery operation state parameter set and the corresponding battery fault identification data set of the target vehicle lithium battery pack; establish multiple first identification fault types corresponding to the first battery operation state parameters; analyze multiple first trigger probabilities of the multiple first identification fault types; construct a first fault diagnosis channel; integrate it into the fault diagnosis integration network, analyze the real-time battery operation state parameters of the target vehicle lithium battery pack, and output a trigger fault type sequence with trigger probabilities from large to small. This solves the technical problems in the prior art, such as the difficulty in accurately identifying the fault types of vehicle lithium battery packs, the inability to adapt to fault diagnosis under complex working conditions, low fault diagnosis efficiency, and poor accuracy, and achieves the technical effect of improving the efficiency and accuracy of lithium battery pack fault diagnosis. Brief Description of the Drawings

[0016] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the devices according to the embodiments of the present application. It should be understood that the operations described above or below do not necessarily need to be executed precisely in sequence. Instead, according to the need, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0017] Figure 1 It is a schematic flowchart of the method for fault diagnosis of vehicle lithium battery packs based on intelligent monitoring provided by the embodiments of the present application.

[0018] Figure 2 It is a schematic structural diagram of the device for fault diagnosis of vehicle lithium battery packs based on intelligent monitoring provided by the embodiments of the present application.

[0019] Description of the reference numerals: Battery data acquisition module 10, state parameter clustering module 20, trigger probability determination module 30, fault diagnosis channel construction module 40, operation state parameter analysis module 50. Detailed Embodiments

[0020] The above description is only an overview of the technical solutions of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically gives the detailed embodiments of this application.

[0021] In order to make the purpose, technical solutions and advantages of this application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0022] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product or server that comprises a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application.

[0023] The embodiments of this application provide a method for fault diagnosis of vehicle-mounted lithium battery packs based on intelligent monitoring, as Figure 1 shown, the method includes: Step S100, collecting a battery operation status parameter set and a corresponding battery fault identification data set of a target vehicle-mounted lithium battery pack.

[0024] Step S100 further includes that the battery operation status parameter set includes voltage, current, temperature, SOC and SOH, and the battery operation status parameter set is collected by the battery management system of the target vehicle-mounted lithium battery pack.

[0025] Preferably, parameters such as voltage, current, temperature, SOC, and SOH of the target vehicle lithium battery pack are collected by its battery management system as a set of battery operating state parameters to reflect the real-time operating condition of the lithium battery pack. Among them, SOC (state of charge) represents the current remaining power of the battery and is an important indicator for evaluating the battery's endurance and charging requirements. SOH (state of health) is used to measure the health of the battery and reflects the performance degradation of the battery relative to a new battery. The battery management system is an electronic system used to manage and monitor the lithium battery pack, which can real-time monitor various parameters of the battery and perform data processing and storage. Specifically, a dedicated voltage acquisition chip or circuit module is used to collect the voltage of each single cell. The chip is connected to the positive and negative electrodes of the battery single cell and uses high-precision analog-to-digital conversion (ADC) technology to convert the analog voltage signal of the battery single cell into a digital signal, thereby accurately measuring the voltage value of each battery single cell. By connecting a voltage dividing resistor network at both ends of the positive and negative electrodes of the battery pack, the high voltage is reduced to a range suitable for the acquisition chip to measure at a certain ratio, and then the acquisition chip performs analog-to-digital conversion to obtain the total voltage data of the battery pack; based on the Hall sensor, using the Hall effect, when current passes through a conductor, a magnetic field will be generated around the conductor. The Hall sensor can detect this magnetic field and convert it into a voltage signal proportional to it, and calculate the magnitude and direction of the battery's charge and discharge current by measuring this voltage signal.

[0026] Preferably, temperature sensors such as thermistors or thermocouples are usually used. The resistance value of the thermistor changes with temperature. The BMS measures the change in the resistance value of the thermistor and converts it into a temperature value through a corresponding algorithm. The thermocouple utilizes the thermoelectric effect of two different metal conductors. When the temperature changes, a thermoelectric potential will be generated at both ends of the thermocouple. The BMS calculates the temperature by measuring the thermoelectric potential. The temperature sensor is generally installed at key positions of the battery pack, such as the surface of the battery single cell, the outer shell of the battery pack, etc., to real-time monitor the temperature distribution of the battery; the SOC of the battery is calculated by integrating the charge and discharge current of the battery. The BMS starts from the initial fully charged state of the battery and real-time records the charge and discharge current and time of the battery, and calculates the change in the battery's power according to the magnitude and direction of the current, thereby obtaining the current SOC value; the SOC is estimated by using the corresponding relationship between the open-circuit voltage of the battery and the SOC. After the battery is static for a period of time, the BMS measures the open-circuit voltage of the battery, and then obtains the corresponding SOC value by looking up the pre-established open-circuit voltage - SOC curve table.

[0027] Preferably, by periodically performing a full charge-discharge cycle test on the battery, measuring the actual capacity of the battery under standard conditions, and comparing it with the initial capacity of the battery, the capacity attenuation rate of the battery is calculated, so as to evaluate the SOH of the battery; the internal resistance of the battery will increase with the aging of the battery. The BMS estimates the SOH of the battery by measuring the AC internal resistance or DC internal resistance of the battery and combining relevant algorithms and empirical formulas. At the same time, a battery fault identification data set corresponding to the battery operating state parameters is collected, which records the identification information such as whether the battery has failed and the type of the fault in the historical data.

[0028] Step S200, perform consistent clustering of the state parameters by combining the battery operating state parameter set and the battery fault identification data set, and establish multiple first identification fault types corresponding to the first battery operating state parameters.

[0029] Preferably, according to the battery operation state parameter set and the battery fault identification data set, consistent clustering of the battery operation state parameters is performed, that is, according to the similarity between the battery operation state parameters, they are divided into different clusters. Among them, consistent clustering is a data analysis method aiming to classify data points with similar characteristics into the same category. Specifically, various clustering algorithms may be used, such as the K-Means algorithm, etc. for clustering. Taking the K-Means algorithm as an example, it randomly selects K initial clustering centers, and then according to the distance between the data points and these centers, assigns the data points to the nearest cluster. Then, it continuously updates the clustering centers and reassigns the data points until the clustering result converges. In this way, the battery operation state parameter set is divided into multiple different clusters, and the data points within each cluster have high similarity in parameter characteristics; after completing the consistent clustering, an association analysis is performed between each cluster and the battery fault identification data set, observing the fault identification situations corresponding to the battery operation state parameters in each cluster. If it is found that the battery operation state parameters in a certain cluster, such as the voltage being within a specific range, the current fluctuation presenting a certain pattern, the temperature being relatively high and the SOC decreasing rapidly, etc., often appear simultaneously with specific fault types such as "battery overheating fault" and "battery aging fault" in the battery fault identification data set, then a corresponding relationship can be established between the first battery operation state parameter represented by this cluster and these first identified fault types; among them, the first battery operation state parameter is a specific parameter combination state obtained through the above-mentioned state parameter consistent clustering. For example, it may be an operation state composed of a specific parameter range such as the voltage being between 3.5V and 3.6V, the current being between 5A and 8A, the temperature being 25°C to 30°C, the SOC being 60% to 70%, and the SOH being 85% to 90%; multiple first identified fault types are various fault types corresponding to the first battery operation state parameter. For example, in the above battery operation state, the possible corresponding fault types may include battery capacity attenuation fault, internal micro-short circuit fault of the battery, overcharge, over-discharge, battery imbalance, external short circuit, and battery management system communication fault, etc.; thereby more clearly understanding the association between different battery operation state parameter combinations and specific fault types, providing data support for the prediction and diagnosis of battery faults.

[0030] Further, step S200 further includes step S210 of performing deviation analysis on any two sets of battery operating state parameters in the battery operating state parameter set. If the deviation meets a preset deviation threshold, clustering any two sets of battery operating state parameters to generate multiple clusters of battery operating state parameters; step S220 of extracting the central data of any one cluster of battery operating state parameters in the multiple clusters of battery operating state parameters as the first battery operating state parameter; step S230 of determining the battery fault identification data corresponding to any one cluster of battery operating state parameters based on the one-to-one correspondence between the battery operating state parameter set and the battery fault identification data set, and performing distribution analysis of the fault types to generate the multiple first identified fault types.

[0031] Step S210 further includes performing deviation analysis on any two sets of battery operating state parameters in the battery operating state parameter set, where the deviation includes the deviation of each type of parameter corresponding to each battery cell.

[0032] Preferably, in the battery operating state parameter set, there are multiple sets of parameters recording the operating states of the battery at different times (such as voltage, current, temperature, SOC, SOH, etc.). Deviation analysis is performed on any two sets of such battery operating state parameters, that is, calculating the degree of difference between their respective parameter values, including comparing the deviation of each type of parameter corresponding to each battery cell. For example, comparing the difference in voltage values, the difference in current values, the difference in temperature values, etc. between the two sets of parameters; then setting a preset deviation threshold. When the deviation between any two sets of battery operating state parameters meets this threshold (indicating that the difference between them is within an acceptable range and they have a certain similarity), these two sets of parameters are classified into one category. By performing such analysis and comparison on all parameter sets, the entire battery operating state parameter set is divided into multiple different clusters, and the parameter sets within each cluster have similar operating state characteristics, finally generating multiple clusters of battery operating state parameters.

[0033] Preferably, among the generated multi-cluster battery operating state parameters, for each cluster, its central data is extracted. The central data can be understood as a set of parameter values that can represent the average or typical characteristics of all parameter groups within the cluster. For example, for a certain cluster of parameters, the average value of all voltage values, the average value of current values, etc. are calculated, and this set of average values can be used as the central data of the cluster. Then, the central data of the operating state parameters of a certain cluster of batteries is defined as the first battery operating state parameter; each set of battery operating state parameters corresponds to a battery fault identification data that records information such as whether the battery has a fault and the type of the fault at that time. For any cluster of battery operating state parameters obtained by clustering, according to this corresponding relationship, the battery fault identification data corresponding to all parameter groups within the cluster is found; then, the battery fault identification data corresponding to the found cluster is analyzed, and information such as the frequency and number of occurrences of different fault types is statistically analyzed to understand the distribution of fault types in the operating state of the cluster. For example, it is statistically analyzed how many times the battery overcharge fault occurs and how many times the battery temperature anomaly fault occurs in a certain cluster operating state; finally, according to the result of the fault type distribution analysis, multiple fault types that may occur under the first battery operating state parameter corresponding to the cluster are determined, and these fault types are defined as multiple first identified fault types. The effective clustering and analysis of the battery operating state parameters are realized, and the corresponding relationship between different operating states and possible fault types is established.

[0034] Step S300: Analyze multiple first trigger probabilities of the multiple first identified fault types under the first battery operating state parameter.

[0035] Preferably, the first trigger probability refers to the likelihood of each first identified fault type occurring under given first battery operating state parameters. Taking the battery capacity attenuation fault as an example, the first trigger probability is the probability of the battery having a capacity attenuation fault within a specific range of voltage, current, temperature, SOC, and SOH parameters. By calculating and analyzing these trigger probabilities, we can gain a deeper understanding of the likelihood of various faults occurring in the battery under different operating states, and thus take corresponding measures in advance. For example, if it is found that the trigger probability of the internal micro-short circuit fault in the battery under a certain operating state is relatively high, we can strengthen the monitoring of the battery, or optimize the design and use of the battery system to reduce the risk of faults. For instance, collect a large amount of battery operation data under the first battery operating state parameters, count the actual occurrence times of each first identified fault type, and then calculate its proportion in the total data volume to estimate the trigger probability. For example, 1000 groups of battery data in a specific operating state are collected, and among them, the battery capacity attenuation fault occurs 50 times. Then the first trigger probability of the battery capacity attenuation fault in this state can be estimated to be 5%. Or, taking the fault type as the top event and the battery operating state parameters as basic events, establish a fault tree. By analyzing the logical relationship between the basic events in the fault tree, calculate the occurrence probability of each first identified fault type under the given first battery operating state parameters. For example, for the internal micro-short circuit fault in the battery, through fault tree analysis, we can determine the influence degree of parameters such as voltage and temperature exceeding a certain range on the occurrence of the micro-short circuit fault, and then calculate the trigger probability of this fault under the current first battery operating state parameters.

[0036] Further, step S300 further includes step S310, respectively taking the multiple first identified fault types as given fault types, counting the occurrence probabilities of the first battery operating state parameters, and generating multiple first conditional probabilities; step S320, calculating the prior probabilities of the multiple first identified fault types, and generating multiple first prior probabilities; step S330, combining the multiple first conditional probabilities and the multiple first prior probabilities, introducing the trigger probability expression, and calculating the multiple first trigger probabilities of the multiple first identified fault types.

[0037] Step S330 further includes that the trigger probability expression is as follows: ; Wherein, is the first trigger probability of the th first identified fault type under the first battery operating state parameters; is the occurrence probability of the first battery operating state parameters under the th first identified fault type; is the prior probability of the th first identified fault type; is the marginal probability of the first battery operating state parameter, which is used to ensure that the sum of the multiple first trigger probabilities of the multiple first identified fault types is 1.

[0038] Preferably, for each first identified fault type, by analyzing historical data and other means, the probability of the current first battery operating state parameter occurring when the fault type has occurred is statistically calculated to generate multiple first conditional probabilities. Then, based on the historical statistical data of battery faults, empirical knowledge, etc., the prior probability of each fault type is estimated, that is, the probability of a certain fault occurring without any observed data, that is, the probability of each first identified fault type occurring without considering the current battery operating state parameter is calculated to obtain multiple first prior probabilities; the obtained multiple first conditional probabilities and multiple first prior probabilities are substituted into the trigger probability expression to calculate the first trigger probability of each first identified fault type under the current first battery operating state parameter. In this way, the likelihood of each fault type occurring under a specific battery operating state is calculated. Among them, the marginal probability of the first battery operating state parameter is the total probability of the observed data D. For example, it can be the sum of the multiple first trigger probabilities of the multiple first identified fault types.

[0039] Further, step S330 further includes step S331 of analyzing the degree of fuzziness of the first battery operating state parameter under the multiple first identified fault types and establishing a fuzzy membership function; step S332 of optimizing the trigger probability expression with the fuzzy membership function and calculating the multiple first trigger probabilities of the multiple first identified fault types.

[0040] Preferably, under multiple first identified fault types, the first battery operating state parameters are not always clearly definable. For example, when the battery temperature is within a certain range, it is difficult to absolutely say that this corresponds to a specific fault type, and there is a certain degree of ambiguity. Analyzing this degree of ambiguity is to study the situation where the boundary between parameter values and fault types is not distinct, and consider the degree of uncertainty of parameter values under different fault type backgrounds. Then, based on expert experience, if there is the experience of domain experts or historical data, the membership function can be artificially set according to the correlation degree between the observed data and the fault mode. For example, if the voltage of a certain battery is close to 3.5V and the SOC is about 80%, experts may judge that the probability of its "aging" fault is relatively high. Then, the membership value of the fault type "battery aging" may be set relatively high to construct a fuzzy membership function. For a given value of the first battery operating state parameter, the degree to which it belongs to a certain first identified fault type can be determined, and the value range is between 0 and 1. For example, when the battery voltage is within a certain interval, through the calculation of the fuzzy membership function, it is obtained that the membership degree of this voltage value belonging to the "internal short circuit fault of the battery" is 0.6, which means that there is a 60% possibility of being related to this fault type.

[0041] Preferably, after introducing the fuzzy membership function, the trigger probability expression is adjusted. For example, the fuzzy membership can be incorporated into the calculation as a weight or correction factor, so that the calculation is not only based on the traditional probability relationship, but also reflects the fuzzy connection between the parameter and the fault type, thus more accurately reflecting the actual situation. After optimization, using the new expression, combining the previously calculated relevant probability data (such as the first conditional probability, the first prior probability, etc.) and the calculation results of the fuzzy membership function, the multiple first trigger probabilities of the multiple first identified fault types are recalculated. The trigger probability can more comprehensively consider the fuzzy characteristics of the battery operating state parameters, provide a more accurate probability basis for battery fault diagnosis, and help to more accurately judge the possible fault types and their likelihoods when the battery is in the current state.

[0042] Step S400, construct a first fault diagnosis channel with the first battery operating state parameters, the multiple first identified fault types, and the multiple first trigger probabilities.

[0043] Preferably, the first battery operating state parameters are input as the basic data for fault diagnosis, which is a quantitative description of the current working state of the battery. For example, when parameters such as abnormal voltage fluctuations, temperature exceeding the normal range, and rapid SOC decline occur in the battery, they may all be signals of potential battery faults; multiple first identified fault types clarify various fault types that the battery may experience under specific operating states, such as battery bulging, reduced charge-discharge efficiency, internal short circuit, etc. Each fault type has its unique manifestation form and possible causes, and is the target object of fault diagnosis, helping to determine what kind of fault state the battery is in; multiple first trigger probabilities reflect the likelihood of each first identified fault type occurring under the given first battery operating state parameters. For example, in a state of high temperature and high SOC, the trigger probability of battery thermal runaway may be relatively high, while the probability of battery capacity attenuation is relatively low, which helps to judge which fault is more likely to occur; specifically, the first battery operating state parameters, multiple first identified fault types, and multiple first trigger probabilities are organically combined, such as storing this information in a database or data structure for subsequent query and analysis, creating a table where one row represents a combination of first battery operating state parameters, and the corresponding columns record the possible first identified fault types and their first trigger probabilities respectively; based on the integrated data, formulate the logic and rules for fault diagnosis. For example, if under a certain first battery operating state parameter, the first trigger probability of a certain first identified fault type exceeds a set threshold, such as exceeding 70%, it can be preliminarily judged that the battery may have this fault, or through more complex logical judgments, comprehensively considering the trigger probabilities of multiple fault types and their relationships to determine the most likely fault situation.

[0044] Step S500: Integrate the first fault diagnosis channel into the fault diagnosis integration network, analyze the real-time battery operating state parameters of the target vehicle lithium battery pack, and output a sequence of trigger fault types with decreasing trigger probabilities.

[0045] Preferably, the fault diagnosis integrated network can integrate multiple different fault diagnosis channels (such as the first fault diagnosis channel here, and possibly other fault diagnosis channels established later), integrate the first fault diagnosis channel into this network, enable it to work in coordination with other parts, share data and resources, thereby improving the overall fault diagnosis ability and accuracy, and then analyze the real-time battery operating state parameters of the target vehicle lithium battery pack, that is, input the collected real-time battery operating state parameters into the fault diagnosis integrated network integrated with the first fault diagnosis channel, and analyze and process these parameters according to the logic and algorithms of the first fault diagnosis channel and other relevant diagnosis modules. For example, match with various previously established battery operating state parameter patterns (corresponding to different fault types and probabilities) to determine which situation the current parameter state belongs to; during the analysis of the real-time battery operating state parameters, the fault diagnosis integrated network will determine the triggering probabilities of each possible fault type in the current operating state according to the calculation results of the first fault diagnosis channel and other relevant modules, indicating the likelihood of each fault type occurring, and finally compare and sort the calculated triggering probabilities of each fault type, arrange these fault types in descending order of the triggering probability, and output the sequence of fault types sorted by probability. For example, the output result may be "thermal runaway fault (triggering probability 80%), battery capacity attenuation fault (triggering probability 30%), battery management system communication fault (triggering probability 10%)". Thus, the real-time monitoring and accurate diagnosis of the faults of the target vehicle lithium battery pack are realized, and the possible fault types and their occurrence probabilities are presented in an intuitive way to ensure the normal operation of the vehicle lithium battery pack.

[0046] Further, step S500 further includes step S510 of collecting the real-time operating environment information of the target vehicle lithium battery pack; step S520 of analyzing the influence relationship between the real-time environment and the triggering probabilities of each fault type in the triggered fault type sequence based on the real-time operating environment information to generate an environment-probability influence relationship; step S530 of re-arranging the triggering probabilities in descending order after optimizing them with the environment-probability influence relationship to generate an updated triggered fault type sequence.

[0047] Preferably, various sensors are used to collect the real-time operating environment information of the target vehicle lithium battery pack, such as environmental temperature, humidity, altitude, vibration intensity, etc. The real-time operating environment information may affect the operating state and failure probability of the lithium battery pack. For example, a high-temperature environment may accelerate the internal chemical reaction of the battery and increase the failure probability of thermal runaway and other failures. A high-humidity environment may cause a decrease in the insulation performance of the battery and an increase in the possibility of short-circuit failures. Based on the collected real-time operating environment information, analyze the relationship between each environmental factor and the triggering probability of each failure type in the triggering failure type sequence. For example, study how the triggering probabilities of failure types such as battery thermal runaway and capacity attenuation change when the environmental temperature changes. Through data analysis and modeling methods, determine the specific influence degree and law of real-time environmental factors on the triggering probability of each failure type, and form an environment-probability influence relationship. For example, obtain a specific relationship such as when the environmental temperature rises by 5°C, the triggering probability of the battery thermal runaway failure increases by 10%.

[0048] Preferably, according to the generated environment-probability influence relationship, adjust the triggering probability of each failure type in the original triggering failure type sequence. If an environmental factor increases the triggering probability of a certain failure type, correspondingly increase the triggering probability value of that failure type, and vice versa; rearrange the optimized triggering probabilities in descending order to obtain an updated triggering failure type sequence, which can allow users to more clearly understand the order of the most likely failure types of the target vehicle lithium battery pack in the current real-time operating environment, so as to conduct targeted monitoring and prevention to ensure the safe and stable operation of the lithium battery pack.

[0049] Furthermore, step S520 further includes step S521, collecting historical failure occurrence environment records based on each failure type; step S522, analyzing the influence relationship between the failure occurrence probability of each failure type and environmental conditions based on the historical failure occurrence environment records, and constructing the environment-probability influence relationship.

[0050] Preferably, for each possible fault type that may occur in the target vehicle lithium battery pack, such as thermal runaway, short circuit, capacity attenuation, etc., collect the environmental-related records of these faults when they occurred in the past through the historical logs of the vehicle fault diagnosis system, battery management system, etc., including but not limited to the environmental conditions at that time, such as ambient temperature, humidity, air pressure, vehicle driving conditions (reflecting vibration conditions), voltage and current during charging status, etc. For example, record the ambient temperature when each battery thermal runaway fault occurs, whether it is in the fast charging state, etc.; conduct in-depth analysis on the collected historical fault occurrence environment records. Specifically, use statistical methods, data analysis algorithms, etc. to study how the changes in each environmental condition affect the occurrence probability of the corresponding fault type. For example, analyze the change trend of the occurrence probability of the battery thermal runaway fault when the ambient temperature rises; or the fluctuation of the occurrence probability of the short circuit fault when the humidity increases, and then quantify the correlation degree between the environmental conditions and the occurrence probability of the fault; finally, organize these correlation relationships into an environment-probability impact relationship. For example, obtain that "when the ambient temperature rises by 10°C, the occurrence probability of the battery thermal runaway fault increases by 20%", so as to conduct targeted monitoring and prevention to ensure the safe and stable operation of the lithium battery pack.

[0051] In the above text, reference is made to Figure 1 A method for diagnosing faults in a vehicle lithium battery pack based on intelligent monitoring according to an embodiment of the present invention is described in detail. Next, reference will be made to Figure 2 Describe a device for diagnosing faults in a vehicle lithium battery pack based on intelligent monitoring according to an embodiment of the present invention.

[0052] The device for diagnosing faults in a vehicle lithium battery pack based on intelligent monitoring according to an embodiment of the present invention is used to solve the technical problems existing in the prior art, such as difficult to accurately identify the fault types of vehicle lithium battery packs, unable to adapt to fault diagnosis under complex working conditions, low fault diagnosis efficiency and poor accuracy, and achieves the technical effect of improving the efficiency and accuracy of lithium battery pack fault diagnosis. As Figure 2 shown, the device for diagnosing faults in a vehicle lithium battery pack based on intelligent monitoring includes: a battery data acquisition module 10, a state parameter clustering module 20, a trigger probability determination module 30, a fault diagnosis channel construction module 40, and an operating state parameter analysis module 50.

[0053] The battery data acquisition module 10 is used to acquire the battery operation state parameter set and the corresponding battery fault identification data set of the target vehicle-mounted lithium battery pack; the state parameter clustering module 20 is used to perform consistent clustering of the state parameters by combining the battery operation state parameter set and the battery fault identification data set, and establish multiple first identification fault types corresponding to the first battery operation state parameters; the trigger probability determination module 30 is used to analyze multiple first trigger probabilities of the multiple first identification fault types under the first battery operation state parameters; the fault diagnosis channel construction module 40 is used to construct a first fault diagnosis channel with the first battery operation state parameters, the multiple first identification fault types and the multiple first trigger probabilities; the operation state parameter analysis module 50 is used to integrate the first fault diagnosis channel into the fault diagnosis integration network, analyze the real-time battery operation state parameters of the target vehicle-mounted lithium battery pack, and output a trigger fault type sequence with the trigger probabilities from large to small.

[0054] Next, the specific configuration of the state parameter clustering module 20 will be described in detail. The state parameter clustering module 20 further includes: performing deviation analysis on any two sets of battery operation state parameters in the battery operation state parameter set, and if the deviation meets a preset deviation threshold, clustering any two sets of battery operation state parameters to generate multiple clusters of battery operation state parameters; extracting the central data of any one cluster of battery operation state parameters in the multiple clusters of battery operation state parameters as the first battery operation state parameters; based on the one-to-one correspondence between the battery operation state parameter set and the battery fault identification data set, determining the battery fault identification data corresponding to any one cluster of battery operation state parameters, and performing distribution analysis of the fault types to generate the multiple first identification fault types.

[0055] Next, the specific configuration of the state parameter clustering module 20 will be continued to be described in detail. The state parameter clustering module 20 further includes: performing deviation analysis on any two sets of battery operation state parameters in the battery operation state parameter set, and the deviation includes the deviation of each type of parameter corresponding to each battery cell.

[0056] Next, the specific configuration of the trigger probability determination module 30 will be described in detail. The trigger probability determination module 30 further includes: respectively using the multiple first identification fault types as the given fault types, statistically calculating the occurrence probability of the first battery operation state parameters to generate multiple first conditional probabilities; calculating the prior probabilities of the multiple first identification fault types to generate multiple first prior probabilities; combining the multiple first conditional probabilities and the multiple first prior probabilities, introducing a trigger probability expression, and calculating the multiple first trigger probabilities of the multiple first identification fault types.

[0057] Next, the specific configuration of the trigger probability determination module 30 will be further described in detail. The trigger probability determination module 30 further includes: The trigger probability expression is as follows: ; Wherein, is the first trigger probability of the th first identified fault type under the first battery operating state parameter; is the occurrence probability of the first battery operating state parameter under the th first identified fault type; is the prior probability of the th first identified fault type; is the marginal probability of the first battery operating state parameter, which is used to ensure that the sum of the first trigger probabilities of multiple first identified fault types is 1.

[0058] Next, the specific configuration of the trigger probability determination module 30 will be further described in detail. The trigger probability determination module 30 further includes: Analyze the degree of fuzziness of the first battery operating state parameter under the multiple first identified fault types, and establish a fuzzy membership function; Optimize the trigger probability expression with the fuzzy membership function, and calculate the first trigger probabilities of the multiple first identified fault types.

[0059] Next, the specific configuration of the operating state parameter analysis module 50 will be described in detail. The operating state parameter analysis module 50 further includes: Collect the real-time operating environment information of the target vehicle lithium battery pack; Based on the real-time operating environment information, analyze the influence relationship between the real-time environment and the trigger probability of each fault type in the trigger fault type sequence, and generate an environment-probability influence relationship; Optimize the trigger probability with the environment-probability influence relationship and then re-arrange it from largest to smallest to generate an updated trigger fault type sequence.

[0060] Next, the specific configuration of the operating state parameter analysis module 50 will be further described in detail. The operating state parameter analysis module 50 further includes: Collect historical fault occurrence environment records based on each fault type; Based on the historical fault occurrence environment records, analyze the influence relationship between the fault occurrence probability of each fault type and the environmental conditions, and construct the environment-probability influence relationship.

[0061] Next, the specific configuration of the battery data acquisition module 10 will be described in detail. The battery data acquisition module 10 further includes: The battery operating state parameter set includes voltage, current, temperature, SOC, and SOH, and the battery operating state parameter set is collected through the battery management system of the target vehicle lithium battery pack.

[0062] The fault diagnosis device for vehicle-mounted lithium battery packs based on intelligent monitoring provided by the embodiments of the present invention can execute the fault diagnosis method for vehicle-mounted lithium battery packs based on intelligent monitoring provided by any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution of the method.

[0063] Although various references are made to certain modules in the device according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0064] The above specific implementation manners do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A method for fault diagnosis of a lithium battery pack for a vehicle based on intelligent monitoring, characterized in that: include: Collect the battery operating status parameter set and the corresponding battery fault identification data set of the target vehicle lithium battery pack; Combining the battery operation state parameter set and the battery fault identification data set to perform consistent clustering of state parameters, and establishing a plurality of first identification fault types corresponding to the first battery operation state parameter; analyzing a plurality of first triggering probabilities of the plurality of first identified fault types under the first battery operating state parameter; constructing a first fault diagnosis channel based on the first battery operating state parameter, the plurality of first identified fault types, and the plurality of first trigger probabilities; The first fault diagnosis channel is integrated into a fault diagnosis integrated network, the real-time battery operating status parameters of the target vehicle lithium battery pack are analyzed, and a sequence of trigger fault types with a trigger probability ranging from large to small is output.

2. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 1, characterized in that: Combining the battery operation state parameter set and the battery fault identification data set to perform consistent clustering of state parameters, and establishing a plurality of first identification fault types corresponding to the first battery operation state parameter, includes: Performing deviation analysis on any two groups of battery operating state parameters in the battery operating state parameter set, and if the deviation meets a preset deviation threshold, clustering the any two groups of battery operating state parameters to generate multiple clusters of battery operating state parameters; Extracting central data of any cluster of battery operating state parameters from the plurality of clusters of battery operating state parameters as a first battery operating state parameter; Based on the one-to-one correspondence between the battery operating state parameter set and the battery fault identification data set, the battery fault identification data corresponding to any cluster of battery operating state parameters is determined, and a distribution analysis of the fault type is performed to generate the multiple first identification fault types.

3. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 2, characterized in that: Deviation analysis is performed on any two groups of battery operating state parameters in the battery operating state parameter set, where the deviation includes the deviation of each type of parameter corresponding to each battery cell.

4. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 1, characterized in that: Analyzing a plurality of first triggering probabilities of the plurality of first identification fault types under the first battery operating state parameter includes: Taking the multiple first identified fault types as given fault types respectively, counting the occurrence probabilities of the first battery operating state parameters, and generating multiple first conditional probabilities; Calculating the priori probabilities of the plurality of first identified fault types to generate a plurality of first priori probabilities; The multiple first conditional probabilities and the multiple first priori probabilities are combined, a trigger probability expression is introduced, and multiple first trigger probabilities of the multiple first identified fault types are calculated.

5. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 4, characterized in that: The trigger probability expression is as follows: ； in, For the a first trigger probability of a first identified fault type under a first battery operating state parameter; For the the probability of occurrence of a first battery operating state parameter under a first identified fault type; For the A priori probability of the first identified fault type; is the marginal probability of the first battery operating state parameter, which is used to ensure that the sum of multiple first triggering probabilities of multiple first identified fault types is 1.

6. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 5, characterized in that: Calculating a plurality of first trigger probabilities of the plurality of first identified fault types further includes: Analyzing the fuzziness degree of the first battery operating state parameter under the multiple first identification fault types, and establishing a fuzzy membership function; The trigger probability expression is optimized by using the fuzzy membership function to calculate a plurality of first trigger probabilities of the plurality of first identified fault types.

7. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 1, characterized in that: Output trigger fault type sequence from large to small trigger probability, including: Collecting real-time operating environment information of the target vehicle lithium battery pack; Based on the real-time operating environment information, analyzing the influence relationship of the real-time environment on the trigger probability of each fault type in the trigger fault type sequence, and generating an environment-probability influence relationship; After optimizing the trigger probability based on the environment-probability influence relationship, the trigger probability is rearranged from large to small to generate an updated trigger fault type sequence.

8. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 7, characterized in that: Based on the real-time operating environment information, analyzing the influence relationship of the real-time environment on the trigger probability of each fault type in the trigger fault type sequence, and generating an environment-probability influence relationship, including: Collect historical fault occurrence environment records based on each fault type; Based on the historical fault occurrence environment records, the influence relationship between the fault occurrence probability of each fault type and the environmental conditions is analyzed to construct the environment-probability influence relationship.

9. The method for diagnosing faults of a lithium battery pack for a vehicle based on intelligent monitoring according to claim 1, characterized in that: The battery operating status parameter set includes voltage, current, temperature, SOC and SOH, and the battery operating status parameter set is collected by a battery management system of the target vehicle lithium battery pack.

10. A vehicle lithium battery pack fault diagnosis device based on intelligent monitoring, characterized in that: The device is used to implement the vehicle lithium battery pack fault diagnosis method based on intelligent monitoring according to any one of claims 1 to 9, and the device comprises: A battery data acquisition module, used to collect a battery operating status parameter set and a corresponding battery fault identification data set of a target vehicle lithium battery pack; A state parameter clustering module, used to perform consistent clustering of state parameters in combination with the battery operation state parameter set and the battery fault identification data set, and establish a plurality of first identification fault types corresponding to the first battery operation state parameter; a trigger probability determination module, configured to analyze a plurality of first trigger probabilities of the plurality of first identified fault types under the first battery operating state parameter; A fault diagnosis channel construction module, configured to construct a first fault diagnosis channel using the first battery operating state parameter, the plurality of first identified fault types and the plurality of first trigger probabilities; The operating state parameter analysis module is used to integrate the first fault diagnosis channel into the fault diagnosis integrated network, analyze the real-time battery operating state parameters of the target vehicle lithium battery pack, and output a sequence of trigger fault types with a trigger probability from large to small.

Images