Lithium battery fault tracing method, apparatus and device, and readable storage medium

Abstract

The invention discloses a lithium battery fault tracing method, apparatus and device, and a readable storage medium. The method comprises the steps of obtaining a fault state parameter and a fault battery type of a fault lithium battery; acquiring a multi-physics field lithium battery simulation model which has the same type as the fault battery, is injected with different fault modes and is labeled as a fault feature vector and a fault value calculation function; selecting a target simulation model based on the fault state parameters, the fault feature vectors and a fault value calculation function; based on the fault state parameters and a target simulation model, fault point cross positioning is carried out on the fault lithium battery, and a target fault point of the fault lithium battery is determined; and based on the target fault point and the target simulation model, constructing a fault propagation path, and determining the fault root cause of the fault lithium battery. It can be seen that the lithium battery fault tracing method and device can integrate the multi-physical field lithium battery simulation model and the electricity-heat-force multi-dimension to conduct fault tracing on the fault lithium battery, the false alarm rate is reduced, and meanwhile accurate fault tracing is conducted on the fault lithium battery.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of lithium batteries, and more particularly to a lithium battery fault tracing method, device, equipment and readable storage medium. BACKGROUND

[0002] Lithium energy storage systems continue to expand in the field of new energy grid connection and smart grid. However, lithium ion batteries are prone to thermal runaway, capacity attenuation, internal short circuit and other problems under complex working conditions, which seriously threaten the safety and economy of the system. In order to quickly and accurately locate the fault cause when the energy storage system fails, fault tracing technology has become a core requirement to ensure the reliable operation of the energy storage system.

[0003] However, the traditional lithium battery fault tracing method relies on threshold comparison and manual experience, and is limited by the rough judgment of fixed thresholds and the influence of human subjective factors, resulting in a high false alarm rate of fault information. SUMMARY

[0004] Therefore, the present application provides a lithium battery fault tracing method, device, equipment and readable storage medium to solve the problem of high false alarm rate of lithium battery faults in the prior art.

[0005] In order to achieve the above purpose, the present scheme is as follows:

[0006] A lithium battery fault tracing method comprises:

[0007] Obtaining the fault state parameters and fault battery type of the fault lithium battery;

[0008] Obtaining a multi-physical field lithium battery simulation model of the same type as the fault battery type and injected with different fault modes, and the label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function;

[0009] Based on the fault state parameters, each fault feature vector and each fault value calculation function, the fault lithium battery is matched with each multi-physical field lithium battery simulation model, and a target simulation model is selected;

[0010] Based on the fault state parameters and the target simulation model, the fault point of the fault lithium battery is cross-located, and the target fault point of the fault lithium battery is determined;

[0011] Based on the target fault point and the target simulation model, a fault propagation path is constructed, and the fault root cause of the fault lithium battery is determined.

[0012] Optionally, the obtaining of the multi-physical field lithium battery simulation model of the same type as the fault battery and injected with different fault modes and the labeled label of each multi-physical field lithium battery simulation model as the fault feature vector and the fault value calculation function comprises:

[0013] Performing multi-physical field simulation on lithium batteries of different battery types to obtain a plurality of simulation models corresponding to each battery type;

[0014] Injecting different fault modes into each simulation model corresponding to each battery type to obtain a plurality of multi-physical field lithium battery simulation models corresponding to each battery type;

[0015] Collecting multi-dimensional fault features of each multi-physical field lithium battery simulation model;

[0016] Processing the multi-dimensional fault features of each multi-physical field lithium battery simulation model to generate at least one fault feature vector;

[0017] Performing weight analysis on each fault feature vector of each multi-physical field lithium battery simulation model to determine the weight value of each fault feature vector;

[0018] Based on each fault feature vector corresponding to the same multi-physical field lithium battery simulation model and the corresponding weight value, a fault value calculation function corresponding to the multi-physical field lithium battery simulation model is generated;

[0019] Each fault feature vector and fault value calculation function of each multi-physical field lithium battery simulation model is used as a labeled label of the corresponding multi-physical field lithium battery simulation model;

[0020] From each multi-physical field lithium battery simulation model, a plurality of multi-physical field lithium battery simulation models of the battery type of the fault battery type are selected.

[0021] Optionally, the multi-physical field simulation on lithium batteries of different battery types to obtain a plurality of simulation models corresponding to each battery type comprises:

[0022] For each battery type, a different level equivalent circuit model is used to simulate the charge and discharge response process of the lithium battery of the battery type, generate a lithium battery model, and increase the battery heat distribution, the battery internal heat source and the battery mechanical deformation process in the lithium battery model to obtain the simulation model of the battery type.

[0023] Optionally, the processing of the multi-dimensional fault features of each multi-physical field lithium battery simulation model to generate at least one fault feature vector comprises:

[0024] Using principal component analysis method, the multi-dimensional fault features are processed to reduce the dimension to generate a plurality of low-dimensional features;

[0025] An importance analysis is performed on each low-dimensional feature, and a fault feature vector is selected from each low-dimensional feature.

[0026] Optionally, the weight analysis on each fault feature vector of each multi-physical field lithium battery simulation model to determine the weight value of each fault feature vector comprises:

[0027] The entropy weight method is used to analyze the weight of each fault feature vector of each multi-physical field lithium battery simulation model to determine the weight value of each fault feature vector.

[0028] Optionally, based on the fault state parameter and the target simulation model, the target fault point of the fault lithium battery is determined by cross positioning the fault point of the fault lithium battery, comprising:

[0029] Based on the fault state parameter, a multi-dimensional time sequence parameter sequence is generated;

[0030] The multi-dimensional time sequence parameter sequence is processed to generate a lightweight parameter sequence;

[0031] The fault point of the target simulation model is determined;

[0032] The lightweight parameter sequence and the fault feature vector of the target simulation model are compared based on the fault point, and the target fault point of the fault lithium battery is determined by cross positioning the fault point of the fault lithium battery in multiple dimensions.

[0033] Optionally, based on the target fault point and the target simulation model, a fault propagation path is constructed to determine the fault root cause of the fault lithium battery, comprising:

[0034] Based on the target fault point and the fault mode of the target simulation model, a fault type feasible solution of the fault lithium battery is determined;

[0035] Based on the target fault point, a propagation path is predicted for each fault type feasible solution to obtain a fault propagation path corresponding to each fault type feasible solution;

[0036] A target path matching the fault state parameter is selected from each fault propagation path;

[0037] Based on the fault type feasible solution corresponding to the target path and the target fault point, a fault root cause of the fault lithium battery is generated.

[0038] A lithium battery fault tracing device, comprising:

[0039] A fault state parameter acquisition module is configured to acquire a fault state parameter and a fault battery type of a fault lithium battery.

[0040] A multi-physics lithium battery simulation model acquisition module is configured to acquire multi-physics lithium battery simulation models of the same type as the fault battery and injected with different fault modes, and each multi-physics lithium battery simulation model is labeled with a fault feature vector and a fault value calculation function;

[0041] A target simulation model selection module is configured to match the fault lithium battery with each multi-physics lithium battery simulation model based on the fault state parameters, each fault feature vector, and each fault value calculation function, and select a target simulation model;

[0042] A target fault point determination module is configured to cross-locate a fault point of the fault lithium battery based on the fault state parameters and the target simulation model, and determine a target fault point of the fault lithium battery.

[0043] A fault root cause determination module is configured to construct a fault propagation path based on the target fault point and the target simulation model, and determine a fault root cause of the fault lithium battery.

[0044] A lithium battery fault tracing device includes a memory and a processor.

[0045] The memory is configured to store a program.

[0046] The processor is configured to execute the program to implement each step of the lithium battery fault tracing method described above.

[0047] A readable storage medium has a computer program stored thereon, and the computer program, when executed by a processor, implements each step of the lithium battery fault tracing method described above.

[0048] It can be seen from the above technical solutions that the lithium battery fault tracing method provided by the present application can obtain the fault state parameters and fault battery type of the faulty lithium battery; based on this, the present application can analyze the actual operating status of the faulty lithium battery through the fault state parameters; the present application can obtain a multi-physics field lithium battery simulation model of the same type as the faulty battery and injected with different fault modes, and the annotation label of each multi-physics field lithium battery simulation model is a fault feature vector and a fault value calculation function; based on the fault state parameters, each fault feature vector and each fault value calculation function, the faulty lithium battery is matched with each multi-physics field lithium battery simulation model to select a target simulation model; based on this, the present application can comprehensively simulate the changes in electro-thermal-mechanical characteristics under different fault modes by adopting a multi-physics field lithium battery simulation model; by comparing the actual operating status of the faulty lithium battery with the fault feature vector and fault value calculation function of the multi-physics field lithium battery simulation model of the same type, the target simulation model that is highly matched with the faulty lithium battery is screened; while narrowing the analysis scope, the matching process is accelerated, the degree of fault is quantified, the direction for fault tracing is indicated, and the false alarm rate is reduced. At the same time, since the fault feature limit is a multi-physics field lithium battery The characteristics of the battery simulation model, therefore, in the comparison process, the present application does not only rely on single-dimensional parameters, but comprehensively considers the factors of multiple physical fields, takes into account the multi-physical field coupling characteristics in the battery failure process, breaks through the limitations of single-dimensional parameters, and thus further reduces the false alarm rate; the present application can cross-locate the fault point of the faulty lithium battery based on the fault state parameters and the target simulation model, and determine the target fault point of the faulty lithium battery; based on the target fault point and the target simulation model, construct a fault propagation path, determine the root cause of the fault of the faulty lithium battery, and by considering the successfully matched target simulation model, the dynamic change characteristics inside the faulty lithium battery can be captured and analyzed in combination with the actual state parameters, so as to realize the early fault tracing analysis of the faulty lithium battery; based on this, the present application comprehensively analyzes the target simulation model and the fault state parameters, accurately locates the target fault point of the faulty lithium battery, and takes the target fault point as the starting point, simulates the fault development process according to the target simulation model, analyzes the changes of the fault inside the battery, sorts out the fault propagation path from the source, and determines the core of the fault problem, thereby generating a reliable fault lithium battery repair plan to prevent similar faults from recurring. It can be seen that this application can integrate multi-physics field lithium battery simulation models, and integrate electrical, thermal, and mechanical dimensions to trace the fault source of faulty lithium batteries. While reducing the false alarm rate, it can accurately trace the fault source of faulty lithium batteries, making it easier for relevant personnel to provide reliable repair solutions based on this. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute a part of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0050] Figure 1 A flow chart of a lithium battery fault tracing method disclosed by the embodiments of the present application;

[0051] Figure 2 A structure block diagram of a lithium battery fault tracing device disclosed by the embodiments of the present application;

[0052] Figure 3 A hardware structure block diagram of a lithium battery fault tracing device disclosed by the embodiments of the present application. DETAILED DESCRIPTION

[0053] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0054] The embodiments of the present application provide a lithium battery fault tracing method, which can be applied in various lithium battery management systems or lithium battery fault repair systems, and can also be applied in various computer terminals or intelligent terminals. The execution subject can be a processor or a server of a computer terminal or an intelligent terminal.

[0055] Next, the lithium battery fault tracing method of the present application will be described in detail, including the following steps: Figure 1

[0056] Step S1, obtaining a fault state parameter of a fault lithium battery and a fault battery type.

[0057] Specifically, the lithium batteries with faults can be selected from various lithium batteries as the fault lithium batteries.

[0058] The voltage, current, internal resistance, temperature and deformation of the fault lithium battery collected by the sensor and other multi-dimensional information can be obtained, and the multi-dimensional information can be initialized to generate the fault state parameter.

[0059] The lithium battery can have various battery types, and the fault battery type corresponding to the fault lithium battery can be determined.

[0060] ​Step S2, obtain a multi-physical field lithium battery simulation model of the same type as the fault battery and injected with different fault modes, and each multi-physical field lithium battery simulation model is labeled with a fault feature vector and a fault value calculation function.

[0061] Specifically, a multi-physical field lithium battery simulation model of different battery types and injected with a fault mode can be constructed in advance.

[0062] Each multi-physical field lithium battery simulation model can be labeled with a corresponding fault feature vector and a fault value calculation function.

[0063] A plurality of multi-physical field lithium battery simulation models of the fault battery type can be selected from each multi-physical field lithium battery cluster.

[0064] Step S3, based on the fault state parameters, each fault feature vector and each fault value calculation function, the fault lithium battery is matched with each multi-physical field lithium battery simulation model, and a target simulation model is selected.

[0065] Specifically, the fault state parameters can be substituted into each fault value calculation function to calculate the fault value of the fault lithium battery.

[0066] The dynamic time warping algorithm DTW or the cosine similarity algorithm can be used to compare the similarity of the fault value of the fault lithium battery and the fault value of the corresponding multi-physical field lithium battery simulation model, and to compare the similarity of the fault state parameters and each multi-physical field lithium battery simulation model fault feature vector, and to select the multi-physical field simulation model with the highest similarity to the fault lithium battery as the target simulation model.

[0067] Step S4, based on the fault state parameters and the target simulation model, the fault point cross positioning of the fault lithium battery is performed to determine the target fault point of the fault lithium battery.

[0068] Specifically, the fault state parameters can be processed to generate an electrical-thermal-mechanical lightweight parameter sequence.

[0069] Based on the electrical-thermal-mechanical lightweight parameter sequence and the target simulation model, the fault point cross positioning of the fault lithium battery is performed to determine the target fault point of the fault lithium battery.

[0070] Step S5, based on the target fault point and the target simulation model, a fault propagation path is constructed to determine the fault root cause of the fault lithium battery.

[0071] Specifically, a graph neural network can be used to construct a fault propagation path based on the target fault point and the target simulation model to determine the fault root cause of the fault lithium battery.

[0072] It can be seen from the above technical solutions that the lithium battery fault tracing method provided by the application can obtain the fault state parameters and the fault battery type of the faulty lithium battery. Based on this, the application can analyze the actual operating conditions of the faulty lithium battery through the fault state parameters. The application can obtain a multi-physical field lithium battery simulation model of the same type as the faulty lithium battery and injected with different fault modes, and the label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function. Based on the fault state parameters, each fault feature vector and each fault value calculation function, the faulty lithium battery is matched with each multi-physical field lithium battery simulation model, and a target simulation model is selected. Based on this, the application can comprehensively simulate the changes in electrical-thermal-mechanical characteristics under different fault modes by using the multi-physical field lithium battery simulation model. By comparing the actual operating conditions of the faulty lithium battery with the fault feature vector and the fault value calculation function of the multi-physical field lithium battery simulation model of the same type, the target simulation model highly matched with the faulty lithium battery is screened. While narrowing the analysis range, the matching process is accelerated, the fault degree is quantified, the direction for fault tracing is indicated, the false positive rate is reduced, and since the fault feature limit is the feature of the multi-physical field lithium battery simulation model, the application relies not only on a single-dimensional parameter during the comparison process, but also considers multiple physical field factors and the multi-physical field coupling characteristics in the battery fault process, thereby breaking through the limitations of a single-dimensional parameter and further reducing the false positive rate. The application can cross-locate the fault point of the faulty lithium battery based on the fault state parameters and the target simulation model, and determine the target fault point of the faulty lithium battery. Based on the target fault point and the target simulation model, a fault propagation path is constructed to determine the fault root cause of the faulty lithium battery. By considering the target simulation model that matches successfully, the internal dynamic change characteristics of the faulty lithium battery can be captured and analyzed in combination with the actual state parameters to realize early fault tracing analysis of the faulty lithium battery. Based on this, the application comprehensively analyzes the target simulation model and the fault state parameters to accurately locate the target fault point of the faulty lithium battery, and takes the target fault point as the starting point to analyze the change of the fault in the battery according to the simulation of the fault development process by the target simulation model, combs the fault propagation path from the source, and determines the core of the fault problem, thereby generating a reliable repair scheme for the faulty lithium battery to prevent similar faults from occurring again. It can be seen that the application can comprehensively analyze the multi-physical field lithium battery simulation model, comprehensively analyze the electrical-thermal-mechanical multi-dimensional fault lithium battery, reduce the false positive rate, accurately trace the fault lithium battery, and facilitate relevant personnel to provide a reliable repair scheme accordingly.

[0073] In some embodiments of the present application, the process of obtaining multiple physical field lithium battery simulation models of the same type as the faulty battery and injected with different fault modes and the labeled label of each multiple physical field lithium battery simulation model as the fault feature vector and the fault value calculation function is described in detail, and the steps are as follows:

[0074] S20, multiple physical field simulation is performed on lithium batteries of different battery types to obtain multiple simulation models corresponding to each battery type.

[0075] Specifically, the electro-thermal-mechanical lithium battery of different battery types can be simulated multiple times to obtain multiple simulation models of each battery type.

[0076] At this time, the simulation models of the same battery type are the same.

[0077] S21, injecting different fault modes into each simulation model corresponding to each battery type to obtain multiple multiple physical field lithium battery simulation models corresponding to each battery type.

[0078] Specifically, a plurality of typical faults of each battery type can be determined as a plurality of fault modes of the corresponding battery type, and different fault modes are injected into different simulation models, the fault evolution process is simulated by parameter perturbation, and the Monte Carlo method is used to generate fault derived data to obtain multiple physical field lithium battery simulation models corresponding to each battery type.

[0079] Among them, the fault mode can be overcharge and overdischarge, overheating, internal gas production, internal short circuit, external short circuit, SEI film thickening, lithium precipitation or thermal runaway, etc.

[0080] The parameter perturbation can include cell resistance mutation, battery cluster internal heat source anomaly, etc.

[0081] S22, collect the multi-dimensional fault features of each multiple physical field lithium battery simulation model.

[0082] Specifically, the electrical, thermal and mechanical characteristics of each multiple physical field lithium battery simulation model can be collected to form multi-dimensional fault features.

[0083] The electrical characteristics can include voltage drop, battery capacity attenuation rate, EIS characteristic frequency point, etc.

[0084] The thermal characteristics can include cell local temperature gradient and battery cluster heat source distribution gradient, system thermal runaway trigger time, system maximum temperature rise rate, etc.

[0085] The mechanical characteristics can include battery shell stress distribution, swelling force change rate, etc.

[0086] S23, process the multi-dimensional fault features of each multi-physical field lithium battery simulation model to generate at least one fault feature vector.

[0087] Specifically, the multi-dimensional fault features can be wavelet transformed, the multi-dimensional transient features are captured, and PCA dimension reduction is performed to generate at least one fault feature vector.

[0088] S24, weight analysis is performed on each fault feature vector of each multi-physical field lithium battery simulation model to determine the weight value of each fault feature vector.

[0089] Specifically, the importance of each fault feature vector in each multi-physical field lithium battery simulation model can be analyzed, and the weight value of each fault feature vector is set.

[0090] S25, based on each fault feature vector corresponding to the same multi-physical field lithium battery simulation model and the corresponding weight value, a fault value calculation function corresponding to the multi-physical field lithium battery simulation model is generated.

[0091] Specifically, based on each fault feature vector of the same multi-physical field lithium battery simulation model and the corresponding normal feature vector and weight value, a fault value calculation function corresponding to the multi-physical field lithium battery simulation model is generated.

[0092] For example, one of the fault value calculation functions of the multi-physical field lithium battery simulation model of the present application is as follows:

[0093]

[0094] In the formula, Loss is the fault value, is the corresponding weight value; is the fault feature vector corresponding to the current; is the fault feature vector corresponding to the voltage; is the fault feature vector corresponding to the temperature; is the fault feature vector corresponding to the stress; I is the current feature vector under normal operating state; U is the voltage feature vector under normal operating state; T is the temperature feature vector under normal operating state; is the stress feature vector under normal operating state.

[0095] S26, each fault feature vector and fault value calculation function of each multi-physical field lithium battery simulation model is used as the labeled label of the corresponding multi-physical field lithium battery simulation model.

[0096] Specifically, each fault feature vector and fault value calculation function of the same physical field lithium battery simulation model can be used as the labeled label of the physical field lithium battery simulation model.

[0097] S27, screening, from the plurality of multi-physical field lithium battery simulation models, a plurality of multi-physical field lithium battery simulation models of which the battery type is the fault battery type.

[0098] Specifically, a plurality of multi-physical field lithium battery simulation models of which the battery type matches the fault battery type can be screened from the plurality of multi-physical field lithium battery simulation models.

[0099] From the above technical solution, it can be seen that the embodiment provides an optional way of obtaining multi-physical field lithium battery simulation models of the same fault battery type and injected with different fault modes, and the annotation label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function. Through the above-mentioned way, a simulation model library can be further constructed in order to cover typical fault scenarios of different battery types, cover fault occurrence probability, and accelerate the fault tracing process.

[0100] In some embodiments of the present application, the process of step S20, multi-physical field simulation of lithium batteries of different battery types, to obtain a plurality of simulation models corresponding to each battery type, is described in detail as follows:

[0101] S200, for each battery type, using different levels of equivalent circuit models to simulate the charge and discharge response process of the lithium battery of the battery type, generate a lithium battery model, and add battery thermal distribution, battery internal heat source and battery mechanical deformation process in the lithium battery model, to obtain a simulation model of the battery type.

[0102] Specifically, different levels of equivalent circuit models can be used to simulate the charge and discharge response process of the lithium battery of the battery type, generate a lithium battery model;

[0103] Specifically, equivalent circuit models can be established for different structural levels of lithium batteries, such as from a single cell to a battery cluster composed of multiple cells, to obtain a lithium battery model.

[0104] Based on fluid dynamics, a battery thermal distribution model combined with heat conduction, convection and radiation can be generated;

[0105] According to the reference point temperature rise rate, the battery internal heat source can be calculated by combining the equivalent thermal resistance network method;

[0106] The finite element method can be used to simulate the mechanical deformation process of the battery due to swelling, aging or external extrusion;

[0107] The lithium battery model, the battery thermal distribution model, the battery internal heat source and the mechanical deformation process are integrated to form a collaborative model;

[0108] The parameterized modeling of the lithium battery parameters and the operation conditions of different battery types is repeated multiple times, and the collaborative model is converted into a simulation model of different battery types.

[0109] It can be seen from the technical solutions that the embodiment provides an optional way of generating a simulation model. The simulation model can be generated by integrating and cooperating in the dimensions of electricity, heat, force, etc., so that the simulation model can represent the multi-dimensional change state.

[0110] In some embodiments of the application, the process of generating at least one fault feature vector by processing the multi-dimensional fault features of each multi-physical field lithium battery simulation model in step S23 is described in detail as follows:

[0111] S230, using principal component analysis method, dimension reduction processing is performed on the multi-dimensional fault features to generate a plurality of low-dimensional features.

[0112] Specifically, the principal component analysis method can be used to perform wavelet transform on the multi-dimensional fault features, capture multi-dimensional transient features, and perform time-frequency analysis and dimension reduction processing on the extracted data to generate a plurality of low-dimensional features.

[0113] S231, important degree analysis is performed on each low-dimensional feature, and a fault feature vector is selected from each low-dimensional feature.

[0114] Specifically, the random forest algorithm can be used to sort the importance of each low-dimensional feature, and the top N fault feature vectors are selected from the sorting result.

[0115] It can be seen from the above technical solutions that the embodiment provides an optional way of screening a fault feature vector. The principal component analysis method and the importance degree analysis can be used to remove redundant feature vectors that do not have much significance in representing the fault operation condition, thereby reducing the matching difficulty of the application.

[0116] In some embodiments of the application, the process of determining the weight value of each fault feature vector by performing weight analysis on each fault feature vector of each multi-physical field lithium battery simulation model in step S24 is described in detail as follows:

[0117] S240, using entropy weight method, weight analysis is performed on each fault feature vector of each multi-physical field lithium battery simulation model to determine the weight value of each fault feature vector.

[0118] Specifically, the statistical characteristics such as mean, variance, and kurtosis can be calculated in combination with the sliding window analysis.

[0119] In combination with the statistical characteristics and the weight method, the information entropy of each fault feature vector in the same multi-physical field lithium battery simulation model is calculated, and the weight value of each fault feature vector is determined.

[0120] From the above technical solution, it can be seen that the embodiment provides an optional way of determining the weight value of each fault feature vector. Through the above way, the representation degree of each fault feature vector to the running state condition can be better determined.

[0121] In some embodiments of the present application, the process of step S4, based on the fault state parameters and the target simulation model, fault point cross positioning of the fault lithium battery is performed to determine the target fault point of the fault lithium battery, which is described in detail as follows:

[0122] S40, based on the fault state parameters, a multi-dimensional time sequence parameter sequence is generated.

[0123] Specifically, the fault state parameters can include multi-dimensional information of multiple collection points.

[0124] According to the sequence of each collection point, the multi-dimensional information can be integrated to generate a multi-dimensional time sequence parameter sequence.

[0125] S41, the multi-dimensional time sequence parameter sequence is processed to generate a lightweight parameter sequence.

[0126] Specifically, the LSTM model can be used to remove noise and redundant detail information, and the multi-dimensional time sequence parameter sequence is processed to reduce the data amount while retaining the key features, thereby generating a lightweight parameter sequence.

[0127] S42, the fault point of the target simulation model is determined.

[0128] Specifically, the fault point of the target simulation model can be determined.

[0129] S43, in combination with the fault point, the lightweight parameter sequence is compared with the fault feature vector of the target simulation model, the fault point of the fault lithium battery is multi-dimensionally cross positioned, and the target fault point of the fault lithium battery is determined.

[0130] Specifically, the fault point of the target simulation model can be fine-tuned in combination with the fault feature vector of the target simulation model and the lightweight parameter sequence. The fault lithium battery is cross positioned in combination with the electrical characteristics, thermal characteristics and mechanical characteristics reflected by the lightweight parameter sequence, and the fault point of the fault lithium battery is determined.

[0131] For example, the fault point of the target simulation model is the top of the battery, the thermal characteristics contained in the light parameter sequence of the fault feature vector show that the heat source of the fault lithium battery is at the bottom of the battery, and the mechanical characteristics contained in the light parameter sequence of the fault feature vector show that the bottom of the fault lithium battery is abnormally expanded, and the cross positioning can determine that the fault point of the fault lithium battery is the bottom of the battery.

[0132] From the above technical solution, it can be seen that the embodiment provides an optional way of cross positioning the fault point of the fault lithium battery based on the fault state parameter and the target simulation model, determining the target fault point of the fault lithium battery. Through the above-mentioned way, the actual fault point of the fault lithium battery can be determined based on the fault point of the target simulation model and the cross verification of multi-dimensional characteristics, rather than directly defining the fault point of the target simulation model as the fault point of the fault lithium battery, thereby further improving the reliability and accuracy of the application.

[0133] In some embodiments of the application, the process of step S5, constructing a fault propagation path based on the target fault point and the target simulation model, determining the fault root cause of the fault lithium battery, is described in detail as follows:

[0134] S50, based on the target fault point and the fault mode of the target simulation model, determining the fault type feasible solution of the fault lithium battery.

[0135] Specifically, each component of the fault lithium battery can be disassembled into a graph node, such as electrical, connecting sheet, cooling pipeline, temperature sensor, etc.

[0136] The same fault mode can contain multiple fault types, for example, internal short circuit can cause local large current, resistance heating, temperature rise, and temperature rise can cause thermal runaway, also can damage the battery separator, promote electrolyte decomposition, electrolyte decomposition can cause internal gas production and temperature rise, accelerate active material degradation, and increase the risk of lithium dendrite growth, i.e. lithium precipitation.

[0137] Therefore, all fault types that can be caused by the fault mode can be used as each fault type feasible solution.

[0138] S51, based on the target fault point, each fault type feasible solution is predicted for the propagation path, and the fault propagation path corresponding to each fault type feasible solution is obtained.

[0139] Specifically, the edge relationship between the graph nodes can be defined according to the physical connection relationship, heat propagation relationship and fault coupling relationship of each component in the fault lithium battery.

[0140] The feasible solutions of each fault type are integrated, and a graph neural network is used to draw a propagation path of each feasible solution of the fault type based on a target fault point.

[0141] S52, screening a target path matching the fault state parameter from each fault propagation path.

[0142] Specifically, the path with the highest matching degree with the fault state parameter can be screened from each fault propagation path as the target path.

[0143] S53, generating a fault root cause of the fault lithium battery based on the fault type feasible solution corresponding to the target path and the target fault point.

[0144] Specifically, the fault root cause can be generated by integrating the fault type feasible solution of the target path and the target fault point.

[0145] As can be seen from the above technical solutions, the embodiment provides an optional way to determine the fault root cause of the fault lithium battery. Through the above way, the fault root cause of the fault lithium battery can be traced through propagation deduction.

[0146] Next, the lithium battery fault tracing device provided in the embodiment will be described in detail. Figure 2 The lithium battery fault tracing device provided in the embodiment will be described in detail.

[0147] Referring to Figure 2 It can be found that the lithium battery fault tracing device can include:

[0148] The fault state parameter acquisition module 10 is configured to acquire the fault state parameter and the fault battery type of the fault lithium battery.

[0149] The multi-physical field lithium battery simulation model acquisition module 20 is configured to acquire multi-physical field lithium battery simulation models of the same type as the fault battery type and injected with different fault modes, and the label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function.

[0150] The target simulation model selection module 30 is configured to match the fault lithium battery with each multi-physical field lithium battery simulation model based on the fault state parameter, each fault feature vector, and each fault value calculation function, and select a target simulation model.

[0151] The target fault point determination module 40 is configured to cross-locate the target fault point of the fault lithium battery based on the fault state parameter and the target simulation model.

[0152] The fault root cause determination module 50 is configured to construct a fault propagation path based on the target fault point and the target simulation model, and determine a fault root cause of the fault lithium battery.

[0153] Further, the multi-physical field lithium battery simulation model acquisition module 20 can include:

[0154] The lithium battery simulation unit is configured to perform multi-physical field simulation on lithium batteries of different battery types, to obtain a plurality of simulation models corresponding to each battery type.

[0155] The fault mode injection unit is configured to inject different fault modes into each simulation model corresponding to each battery type, to obtain a plurality of multi-physical field lithium battery simulation models corresponding to each battery type.

[0156] The multi-dimensional fault feature acquisition unit is configured to acquire multi-dimensional fault features of each multi-physical field lithium battery simulation model.

[0157] The fault feature vector generation unit is configured to process the multi-dimensional fault features of each multi-physical field lithium battery simulation model, to generate at least one fault feature vector.

[0158] The weight value calculation unit is configured to perform weight analysis on each fault feature vector of each multi-physical field lithium battery simulation model, to determine a weight value of each fault feature vector.

[0159] The fault value calculation function generation unit is configured to generate a fault value calculation function of a corresponding multi-physical field lithium battery simulation model based on each fault feature vector corresponding to the same multi-physical field lithium battery simulation model and the corresponding weight value.

[0160] The labeled label determination unit is configured to take each fault feature vector and the fault value calculation function of each multi-physical field lithium battery simulation model as a labeled label of the corresponding multi-physical field lithium battery simulation model.

[0161] The model screening unit is configured to screen a plurality of multi-physical field lithium battery simulation models of the battery type of the fault battery from the plurality of multi-physical field lithium battery simulation models.

[0162] Further, the lithium battery simulation unit can include:

[0163] The first lithium battery simulation subunit is configured to, for each battery type, perform cell level and battery cluster level simulation on lithium batteries of the battery type by using different levels of equivalent circuit models, simulate the charge and discharge response process of the lithium batteries of the battery type, generate a lithium battery model, and add battery heat distribution, battery internal heat source and battery mechanical deformation process in the lithium battery model, to obtain a simulation model of the battery type.

[0164] Further, the fault feature vector generating unit can comprise:

[0165] a first fault feature vector generating subunit, configured to perform dimension reduction processing on the multi-dimensional fault features by using a principal component analysis method, to generate a plurality of low-dimensional features;

[0166] a second fault feature vector generating subunit, configured to perform importance analysis on each low-dimensional feature, and screen a fault feature vector from each low-dimensional feature.

[0167] Further, the weight value calculating unit can comprise:

[0168] a first weight value calculating subunit, configured to perform weight analysis on each fault feature vector of each multi-physical-field lithium battery simulation model by using an entropy weight method, to determine a weight value of each fault feature vector.

[0169] Further, the target fault point determining module 40 can comprise:

[0170] a first target fault point determining subunit, configured to generate a multi-dimensional time sequence parameter sequence based on the fault state parameter;

[0171] a second target fault point determining subunit, configured to perform lightweight processing on the multi-dimensional time sequence parameter sequence, to generate a lightweight parameter sequence;

[0172] a third target fault point determining subunit, configured to determine a fault point of the target simulation model;

[0173] a fourth target fault point determining subunit, configured to combine the fault point, compare the lightweight parameter sequence with a fault feature vector of the target simulation model, perform multi-dimensional cross positioning on a fault point of the fault lithium battery, and determine a target fault point of the fault lithium battery.

[0174] Further, the fault root cause determining module 50 can comprise:

[0175] a first fault root cause determining subunit, configured to determine a fault type feasible solution of the fault lithium battery based on the target fault point and a fault mode of the target simulation model;

[0176] a second fault root cause determining subunit, configured to perform propagation path prediction on each fault type feasible solution based on the target fault point, to obtain a fault propagation path corresponding to each fault type feasible solution;

[0177] a third fault root cause determining subunit, configured to screen a target path matching the fault state parameter from each fault propagation path;

[0178] The fourth fault root cause determination subunit is configured to determine a fault root cause of the faulty lithium battery based on the target path corresponding fault type feasible solution and the target fault point.

[0179] The lithium battery fault tracing device provided by the embodiments of the present application can be applied to a lithium battery fault tracing device, such as a PC terminal, a cloud platform, a server, a server cluster, and the like. Optionally, Figure 3 The hardware structure block diagram of the lithium battery fault tracing device is shown in FIG. 1. Referring to FIG. 1, Figure 3 The hardware structure of the lithium battery fault tracing device can include at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4.

[0180] In the embodiments of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 complete mutual communication through the communication bus 4.

[0181] The processor 1 can be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application, etc.

[0182] The memory 3 can include a high-speed RAM memory, and can also include a non-volatile memory, etc., such as at least one disk memory.

[0183] The memory stores a program, and the processor can call the program stored in the memory, and the program is used to:

[0184] Obtain the fault state parameters and the fault battery type of the faulty lithium battery;

[0185] Obtain a multi-physical field lithium battery simulation model of the same type as the fault battery type and injected with different fault modes, and the label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function;

[0186] Based on the fault state parameters, each fault feature vector, and each fault value calculation function, the faulty lithium battery is matched with each multi-physical field lithium battery simulation model, and a target simulation model is selected;

[0187] Based on the fault state parameters and the target simulation model, the faulty lithium battery is subjected to fault point cross positioning to determine the target fault point of the faulty lithium battery.

[0188] Based on the target fault point and the target simulation model, a fault propagation path is constructed to determine a fault root cause of the fault lithium battery.

[0189] Optionally, the refinement function and the extension function of the program can refer to the above description.

[0190] The embodiment of the application further provides a readable storage medium which can store a program suitable for processor execution, and the program is used for:

[0191] Obtaining a fault state parameter and a fault battery type of the fault lithium battery;

[0192] Obtaining a multi-physical field lithium battery simulation model of the same type as the fault battery type and injected with different fault modes, and a label of each multi-physical field lithium battery simulation model is a fault feature vector and a fault value calculation function;

[0193] Based on the fault state parameter, each fault feature vector and each fault value calculation function, the fault lithium battery is matched with each multi-physical field lithium battery simulation model to select a target simulation model;

[0194] Based on the fault state parameter and the target simulation model, fault point cross positioning is performed on the fault lithium battery to determine a target fault point of the fault lithium battery;

[0195] Based on the target fault point and the target simulation model, a fault propagation path is constructed to determine a fault root cause of the fault lithium battery.

[0196] Optionally, the refinement function and the extension function of the program can refer to the above description.

[0197] Finally, it should be noted that in this document, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitation, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0198] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between various embodiments can be referred to each other.

[0199] The above description of disclosed embodiments allows a skilled person to implement or use the application. Numerous modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. The various embodiments of the application can be combined with each other. Therefore, the application will not be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A lithium battery fault tracing method, characterized in that: include: Obtain the fault status parameters and fault battery type of the faulty lithium battery; Obtain a multi-physics field lithium battery simulation model of the same type as the faulty battery and injected with different fault modes, wherein the label of each multi-physics field lithium battery simulation model is a fault feature vector and a fault value calculation function; Based on the fault state parameters, each fault characteristic vector and each fault value calculation function, matching the faulty lithium battery with each multi-physics field lithium battery simulation model, and selecting a target simulation model; Based on the fault state parameters and the target simulation model, cross-locating the fault point of the faulty lithium battery to determine the target fault point of the faulty lithium battery; Based on the target fault point and the target simulation model, a fault propagation path is constructed to determine the root cause of the fault of the faulty lithium battery.

2. The lithium battery fault tracing method according to claim 1, characterized in that: The method of obtaining a multi-physics field lithium battery simulation model of the same type as the faulty battery and injected with different fault modes, wherein the label of each multi-physics field lithium battery simulation model is a fault feature vector and a fault value calculation function, includes: Perform multi-physics simulation on lithium batteries of different battery types to obtain multiple simulation models corresponding to each battery type; Different failure modes are injected into the simulation models corresponding to each battery type to obtain multiple multi-physics field lithium battery simulation models corresponding to each battery type; Collect multi-dimensional fault characteristics for each multi-physics field lithium battery simulation model; Processing the multi-dimensional fault characteristics of each multi-physics field lithium battery simulation model to generate at least one fault feature vector; Perform weight analysis on each fault feature vector of each multi-physics field lithium battery simulation model to determine the weight value of each fault feature vector; Based on each fault feature vector corresponding to the same multi-physics field lithium battery simulation model and its corresponding weight value, a fault value calculation function corresponding to the multi-physics field lithium battery simulation model is generated; Each fault feature vector and fault value calculation function of each multi-physics field lithium battery simulation model is used as a label of the corresponding multi-physics field lithium battery simulation model; A plurality of multi-physics field lithium battery simulation models whose battery type is the faulty battery type are selected from the various multi-physics field lithium battery simulation models.

3. The lithium battery fault tracing method according to claim 2, characterized in that: The multi-physics field simulation is performed on lithium batteries of different battery types to obtain multiple simulation models corresponding to each battery type, including: For each battery type, equivalent circuit models of different levels are used to perform cell-level and battery cluster-level simulations on the lithium batteries of that battery type, simulate the charge and discharge response process of the lithium batteries of that battery type, generate a lithium battery model, and add battery thermal distribution, internal heat source of the battery, and battery mechanical deformation process to the lithium battery model to obtain a simulation model of the battery type.

4. The lithium battery fault tracing method according to claim 2, characterized in that: The processing of the multi-dimensional fault characteristics of each multi-physics field lithium battery simulation model to generate at least one fault feature vector includes: The principal component analysis method is used to reduce the dimensionality of multi-dimensional fault features and generate multiple low-dimensional features; The importance of each low-dimensional feature is analyzed, and the fault feature vector is filtered from each low-dimensional feature.

5. The lithium battery fault tracing method according to claim 2, characterized in that: The weight analysis of each fault feature vector of each multi-physics field lithium battery simulation model is performed to determine the weight value of each fault feature vector, including: The entropy weight method is used to perform weight analysis on each fault feature vector of each multi-physics field lithium battery simulation model to determine the weight value of each fault feature vector.

6. The lithium battery fault tracing method according to claim 1, characterized in that: The cross-locating of the fault point of the faulty lithium battery based on the fault state parameter and the target simulation model to determine the target fault point of the faulty lithium battery includes: Based on the fault state parameters, generating a multi-dimensional timing parameter sequence; Performing lightweight processing on the multi-dimensional time series parameter sequence to generate a lightweight parameter sequence; Determining a fault point of the target simulation model; In combination with the fault point, the lightweight parameter sequence is compared with the fault feature vector of the target simulation model, the fault point of the faulty lithium battery is multi-dimensionally cross-located, and the target fault point of the faulty lithium battery is determined.

7. The lithium battery fault tracing method according to claim 1, characterized in that: The step of constructing a fault propagation path based on the target fault point and the target simulation model to determine the root cause of the fault of the faulty lithium battery includes: Determining a feasible solution to the fault type of the faulty lithium battery based on the target fault point and the fault mode of the target simulation model; Based on the target fault point, a propagation path prediction is performed for a feasible solution of each fault type to obtain a fault propagation path corresponding to the feasible solution of each fault type; Selecting a target path that matches the fault state parameter from each fault propagation path; Based on the feasible solution of the fault type corresponding to the target path and the target fault point, a root cause of the fault of the faulty lithium battery is generated.

8. A lithium battery fault tracing device, characterized in that: include: A fault status parameter acquisition module is used to obtain the fault status parameters and fault battery type of the faulty lithium battery; A multi-physics field lithium battery simulation model acquisition module is used to obtain a multi-physics field lithium battery simulation model of the same type as the faulty battery and injected with different fault modes, and the label of each multi-physics field lithium battery simulation model is a fault feature vector and a fault value calculation function; a target simulation model selection module, configured to match the faulty lithium battery with each multi-physics field lithium battery simulation model based on the fault state parameters, each fault characteristic vector, and each fault value calculation function, and select a target simulation model; a target fault point determination module, configured to perform fault point cross-location on the faulty lithium battery based on the fault state parameters and the target simulation model, and determine the target fault point of the faulty lithium battery; The fault root cause determination module is used to construct a fault propagation path based on the target fault point and the target simulation model, and determine the fault root cause of the faulty lithium battery.

9. A lithium battery fault tracing device, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the lithium battery fault tracing method according to any one of claims 1 to 7.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the lithium battery fault tracing method according to any one of claims 1 to 7 is implemented.

Images