Lithium battery fault identification method and system

By combining Lasso regression and improving the feature selection and deep belief network of the PSO algorithm, a lithium-ion battery fault diagnosis model is constructed, which solves the response speed and complex fault identification problems of traditional diagnostic methods, and realizes efficient fault identification and classification of lithium-ion batteries. Reduced security risks.

CN118914852APending Publication Date: 2024-11-08STATE GRID ENERGY CONSERVATION SERVICE
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202410958300.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Traditional lithium-ion battery fault diagnosis methods are slow to respond, difficult to identify complex failure modes, and information redundancy and feature selection difficulties when facing a large amount of data, deep learning methods are limited in their application due to the need for a large amount of labeled data.

Method used

Using a pre-trained fault diagnosis model, the voltage, current and temperature data of lithium-ion batteries are obtained, and the features are selected and optimized using Lasso regression method and improved PSO algorithm. The fault diagnosis model is built in combination with the deep belief network to realize real-time fault identification. and classification.

Benefits of technology

It improves the accuracy and robustness of fault diagnosis, reduces redundant information, enhances the generalization ability of the model, can detect potential battery problems in advance, and reduces safety risks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118914852A_ABST
    Figure CN118914852A_ABST
Patent Text Reader

Abstract

The invention provides a lithium battery fault identification method and system, and belongs to the technical field of lithium battery fault identification, and the method comprises the steps: obtaining the voltage, current and temperature data of a lithium ion battery during charging and discharging; and processing the acquired voltage, current and temperature data by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result. According to the method, the advantages of feature selection and deep learning are fully utilized, and the accuracy and robustness of fault diagnosis are improved; feature selection is optimized through a Lasson regression algorithm, redundant information is reduced, and efficiency is improved; high-order feature representation of data can be automatically learned, the method is suitable for complex nonlinear relations, and the generalization ability of the model is improved; network parameters are optimized by using the improved adaptive particle swarm optimization, so that the performance and the convergence speed of the algorithm can be improved; the lithium ion battery fault diagnosis effect is obvious, the potential problems of the battery can be found in advance, and the risk of the fault on equipment and personnel safety is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery fault identification, and in particular to a lithium battery fault identification method and system. Background Art

[0002] Lithium-ion battery energy storage system is an important component of smart microgrid, and its operating status will directly affect the results of smart microgrid energy management and optimization control. With the rapid development and large-scale application of energy storage systems, accidents caused by lithium-ion battery safety issues have occurred frequently, bringing serious social harm and impact. Battery safety performance is affected by many factors, such as overcharge and over-discharge, thermal runaway, battery aging, BMS system failure, etc., which will bring huge safety hazards. Therefore, fault diagnosis of lithium battery packs is a necessary means to maintain the safe and reliable operation of energy storage systems.

[0003] Traditional lithium-ion battery fault diagnosis methods are usually based on thresholds or empirical rules, and determine whether there is a fault by monitoring battery voltage, temperature and other parameters. However, this method has the following problems: first, the response speed to battery status changes is slow, and real-time monitoring and early warning cannot be achieved; second, it is difficult to accurately identify complex fault modes, especially in terms of early warning before the fault occurs; third, when faced with a large amount of data, there is often information redundancy and difficulty in feature selection.

[0004] With the development of deep learning technology, deep neural networks have achieved great success in image recognition, speech recognition and other fields. Deep learning has the advantages of automatically learning feature representation and adapting to complex nonlinear relationships, providing new ideas for lithium-ion battery fault diagnosis. However, traditional deep learning methods often require a large amount of labeled data for training, and it is often difficult to obtain sufficient labeled information for lithium-ion battery fault data, which limits the application of deep learning in this field. Summary of the invention

[0005] The object of the present invention is to provide a lithium battery fault identification method and system to solve at least one technical problem existing in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a lithium battery fault identification method, comprising:

[0008] Obtain voltage, current and temperature data of lithium-ion batteries during charging and discharging;

[0009] The obtained voltage, current and temperature data are processed by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein, the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features by using the Lasso regression method to select the most appropriate feature set; optimizing and selecting the extracted features by using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery status and features, and classifying the fault type.

[0010] Furthermore, the characteristics that can characterize the fault state include: voltage change rate, voltage deviation, voltage extreme difference, temperature change rate, temperature extreme difference,

[0011] Furthermore, Lasso regression achieves feature selection and model simplification by adding L1 regularization to the model coefficients; due to the influence of L1 regularization, the solution of Lasso regression is usually sparse, that is, only a few features have coefficients that are not zero.

[0012] Furthermore, the improved PSO algorithm is used to optimize the extracted features to find the optimal feature combination. The core parameters of the particle swarm algorithm are speed and position, and the update formulas are as follows:

[0013]

[0014]

[0015] in, is the velocity of the ith particle in the kth iteration, pbest id is the maximum value of the fitness of the i-th particle in k iterations, For location information, gbest id is the maximum value among all pbest, c is the cognitive constant, and r is the random function.

[0016] Furthermore, in the particle swarm algorithm, the inertia weight is adaptively optimized; the inertia weight w is used to balance the global and local search capabilities. The inertia weight w is related to the evolution factor f, which allows w to use an S-shaped mapping:

[0017]

[0018] Furthermore, in the particle swarm algorithm, the acceleration coefficient is controlled; the parameter c1 represents self-cognition, which pulls the particle to its own historical best position; the parameter c2 represents social influence, which pushes the group to converge to the current global optimal area and helps to converge quickly; the acceleration coefficient is initialized to 2.0 and adaptively controlled according to the evolutionary state, and the strategy is as follows: increase c1 and decrease c2 in the exploration state.

[0019] In a second aspect, the present invention provides a lithium battery fault identification system, comprising:

[0020] An acquisition module is used to obtain the voltage, current and temperature data of the lithium-ion battery during charging and discharging;

[0021] The diagnosis module is used to process the acquired voltage, current and temperature data using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features using the Lasso regression method to select the most appropriate feature set; optimizing the extracted features using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery state and features, and classifying the fault type.

[0022] In a third aspect, the present invention provides a non-transitory computer-readable storage medium, which is used to store computer instructions. When the computer instructions are executed by a processor, the lithium battery fault identification method as described in the first aspect is implemented.

[0023] In a fourth aspect, the present invention provides a computer device comprising a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the lithium battery fault identification method as described in the first aspect.

[0024] In a fifth aspect, the present invention provides an electronic device, comprising: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes instructions for implementing the lithium battery fault identification method as described in the first aspect.

[0025] Beneficial effects of the present invention: The combination of particle swarm optimization and deep belief network fully utilizes the advantages of feature selection and deep learning, and improves the accuracy and robustness of fault diagnosis; by optimizing feature selection through Lasson regression algorithm, the most representative features can be effectively extracted, redundant information is reduced, and the efficiency of the algorithm is improved; the fault diagnosis model based on deep belief network can automatically learn the high-order feature representation of data, which is suitable for complex nonlinear relationships and improves the generalization ability of the model; the improved adaptive particle swarm algorithm is used to optimize the network parameters, which can improve the performance and convergence speed of the algorithm. This method has a significant effect on the diagnosis of lithium-ion battery faults, can detect potential battery problems in advance, and reduce the risks caused by faults to equipment and personnel safety.

[0026] Additional advantages of the present invention will be more clearly given in the following description or learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0028] Figure 1 This is a fault diagnosis flowchart according to an embodiment of the present invention.

[0029] Figure 2 This is a flow chart of the Lasso regression algorithm described in an embodiment of the present invention.

[0030] Figure 3 This is a diagram of the DBN network structure described in an embodiment of the present invention.

[0031] Figure 4 This is a flow chart of the improved PSO-DBN algorithm described in an embodiment of the present invention.

[0032] Figure 5 The figure is a flow chart of the entire process of diagnostic model training described in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below by the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0034] It should be understood by those skilled in the art that unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs.

[0035] It should also be understood that terms, such as those defined in commonly used dictionaries, should be understood to have a meaning consistent with that in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless as defined herein.

[0036] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements and / or groups thereof.

[0037] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. Different embodiments or examples described in this specification and features of different embodiments or examples may be combined and combined by those skilled in the art without contradiction.

[0038] To facilitate understanding of the present invention, the present invention is further explained below with reference to specific embodiments in conjunction with the accompanying drawings, and the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0039] Those skilled in the art should understand that the drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily necessary for implementing the present invention.

[0040] Example 1

[0041] In this embodiment 1, a lithium battery fault identification system is first provided, including:

[0042] An acquisition module is used to obtain the voltage, current and temperature data of the lithium-ion battery during charging and discharging;

[0043] The diagnosis module is used to process the acquired voltage, current and temperature data using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features using the Lasso regression method to select the most appropriate feature set; optimizing the extracted features using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery state and features, and classifying the fault type.

[0044] In this embodiment, the above-mentioned system can be used to implement a lithium battery fault identification method, including: using an acquisition module to acquire voltage, current, and temperature data of a lithium-ion battery during charging and discharging; using a diagnosis module to process the acquired voltage, current, and temperature data using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein the training of the fault diagnosis model includes: collecting voltage, current, and temperature related data of the lithium-ion battery during charging and discharging, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; using the Lasso regression method to select the extracted features and select the most appropriate feature set; using the improved PSO algorithm to optimize and select the extracted features to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, and determining whether there is a fault by analyzing the battery state and features, and classifying the fault type.

[0045] Specifically, in this embodiment, a lithium-ion battery fault diagnosis model training method based on Lasso regression and improved particle swarm optimization combined with deep belief network is proposed. First, a feature data matrix that can represent battery capacity loss and fault state is extracted during the standard charge and discharge process of the battery. Secondly, feature extraction is performed on the data, and 10 suitable and reliable parameters are selected to describe the fault. Thirdly, in order to improve the accuracy of fault diagnosis, the optimal features are selected for training, and the Lasso regression algorithm is introduced to optimize the features, and 6 optimal features are selected to form a new matrix. Finally, a DBN network is established to perform fault classification diagnosis on the extracted feature matrix. In order to improve the detection accuracy of the network, the improved adaptive particle swarm algorithm is used to optimize the network parameters, which solves the problem that complex multi-peak problems are prone to fall into local optimal solutions compared to traditional particle swarm algorithms. This method makes full use of the advantages of feature selection and deep learning, improves the accuracy of fault diagnosis, and provides a new solution for the safe operation and intelligent management of lithium-ion batteries.

[0046] The specific method is:

[0047] The first step is data collection. The voltage, current, temperature and other relevant data of lithium-ion batteries during charging and discharging are collected through sensors and BMS, and the corresponding data set is established.

[0048] The second step is feature extraction. The following ten features that can characterize the fault status are extracted from the collected data set:

[0049] Voltage change rate:

[0050]

[0051] Voltage deviation:

[0052]

[0053] Voltage difference:

[0054] U i =U max -U min (1.3)

[0055] Temperature change rate:

[0056]

[0057] Temperature extremes:

[0058] W k =W max -W min (1.5)

[0059] SOC (State of charge) battery power state

[0060]

[0061] Among them, T s is the sampling time, C max It is the maximum capacity of the battery. The change of battery capacity with aging affects the estimation accuracy of SOC, and the estimation accuracy of SOC also determines the estimation accuracy of battery capacity.

[0062] Charging constant voltage time T cv

[0063] Normally, batteries are charged using constant current (CC) or constant power (CP) charging, or constant voltage (CV) charging. In theory, a faulty battery takes longer to complete the CV phase, which can be used as a feature to identify battery failure.

[0064] Energy loss E loss

[0065] Compared with the healthy state, when an ISC fault occurs, the total energy input of the battery during the charging phase will be more, the energy output during the discharging phase will be less, and the energy loss will be less, so it can be used as a feature to identify the fault.

[0066]

[0067] SOC relative to V ocv The slope

[0068] In the event of an ISC fault, the OCV during discharge will decrease faster than the estimated SOC, so a high-order polynomial can be drawn between the two to describe this situation. After calculation, it is found that the fourth-order polynomial can fit the data well, and the first-order slope a1 is used as the basis for fault judgment.

[0069] V ocv =a0+a1SOC(k)+a2SOC 2 (k)+a 23 SOC 3 (k)+a4SOC 4 (k)(1.8)

[0070] Battery capacity loss Q loss :

[0071]

[0072] The third step is feature selection. After extracting the above features, the Lasso regression method is used to select the above features and select the 6 most suitable feature sets for subsequent neural network training. Lasso regression is a variant of linear regression, which achieves feature selection and model simplification by adding L1 regularization to the model coefficients.

[0073] Lasso regression tends to reduce the coefficients of unimportant features to zero, thereby achieving feature selection. Due to the influence of L1 regularization, the solution of Lasso regression is usually sparse, that is, only a few features have non-zero coefficients.

[0074] The objective function is as follows:

[0075]

[0076] The fourth step is network training. The selected 6 features form a feature set, which is divided into a training set and a test set in a ratio of 8:2. A deep belief network consisting of multiple restricted Boltzmann machines is constructed. The improved PSO algorithm is used to optimize the extracted features to find the optimal feature combination and improve the accuracy and efficiency of fault diagnosis. The core parameters of the particle swarm algorithm are speed and position, and their update formulas are as follows:

[0077]

[0078]

[0079] in, is the velocity of the ith particle in the kth iteration, pbest id is the maximum value of the fitness of the i-th particle in k iterations, For location information, gbest id is the maximum value among all pbest, c is the cognitive constant, and r is the random function.

[0080] In order to solve the problem that the particle swarm algorithm is prone to fall into the local optimal solution, the following two control strategies are given to improve the algorithm.

[0081] First, the inertia weight is adaptively optimized. The inertia weight w is used to balance the global and local search capabilities. The inertia weight w is related to the evolution factor f, which allows w to use an S-shaped mapping:

[0082]

[0083] Secondly, the acceleration coefficient is controlled. Parameter c1 represents self-cognition, which pulls the particle to its own historical best position; parameter c2 represents social influence, which pushes the group to converge to the current global best area and helps converge quickly. The acceleration coefficient is initialized to 2.0 and adaptively controlled according to the evolutionary state. The strategy is as follows: increase c1 and decrease c2 in the exploration state. This strategy helps particles explore and achieve their own historical best positions individually, rather than crowding around the current best particle that may be associated with the local optimum, to increase the convergence speed of the algorithm.

[0084] The feature set is input into the DBN model after the improved algorithm optimizes the parameters for training, and a deep learning model for lithium-ion battery fault diagnosis is constructed. By analyzing the battery status and characteristics, it is determined whether there is a fault and the fault type is classified.

[0085] Figure 1 The fault diagnosis flow chart provided in this embodiment can ensure the normal progress of fault diagnosis and is the most concise and clear diagnosis process.

[0086] Figure 2 The Lasso regression algorithm flow chart provided for this embodiment uses its sparse properties for feature selection. The input data contains a data set of 10 listed features, and then the data is standardized so that the features have the same scale. Next, the L1 regularization parameter is set, which controls the sparsity of the model for the features. The model is fitted using the Lasso regression algorithm, and the optimal coefficients are found by minimizing the loss function and the L1 regularization term. According to the obtained coefficients, it is determined which features have a greater impact on the target variable, and the features with a coefficient of 0 can be considered to be selected features. Finally, the 6 selected features are output, which constitute a new feature set to be put into the neural network for training classification.

[0087] Figure 3The DBN network structure diagram provided for this embodiment, DBN is composed of many RBM layers, each RBM layer has a visible layer and a hidden layer. The training process of DBN is mainly divided into two stages, pre-training and fine-tuning. In the unsupervised pre-training stage, the weights and biases are first randomly initialized, and the layer-by-layer greedy algorithm is used to train the RBM layer by layer to initialize the parameters. For each layer of RBM: one layer in the DBN is used as a visible layer, and the upper layer is used as a hidden layer, and the RBM is trained to learn the representation between the two layers. Use an unsupervised learning algorithm to maximize the log likelihood or minimize the energy function to update the parameters of the RBM. Finally, use the trained RBM to extract the representation of the visible layer as the input of the next layer. In the supervised fine-tuning stage, the parameters obtained in the unsupervised pre-training stage are used as the initial parameters, and the supervised back propagation algorithm (gradient descent) is used to fine-tune the parameters of the DBN, and then the cross entropy loss function is used to measure the prediction error of the DBN on the training data. According to the gradient of the loss function, the parameters of the DBN (weights and biases) are updated to reduce the loss. Finally, repeat the above steps until convergence or a predetermined number of iterations is reached. This method overcomes the defects of traditional BP network, such as easy to fall into local optimum and time-consuming training due to the randomness of weight parameters.

[0088] Figure 4 The improved PSO-DBN algorithm flow chart provided for this embodiment focuses on optimizing the parameters of the network using an adaptive particle swarm algorithm, improving the accuracy of fault identification, and improving the performance and convergence speed of the algorithm. First, the particle swarm is initialized, and each particle represents a potential DBN parameter solution. Then the fitness of the DBN parameter solution corresponding to each particle is calculated, that is, the performance evaluation index of DBN on the training set, such as error rate or accuracy. According to the rules of the improved PSO algorithm, the speed and position of each particle are updated to find the global optimal solution and the individual optimal solution. According to the updated particle position, the parameters of DBN, such as weights and biases, are updated. Finally, the above steps are repeated until a predetermined number of iterations is reached or the stop condition is met. The DBN network after optimizing the parameters can have better performance and improve the classification accuracy of the network. In summary, the improved adaptive PSO-DBN algorithm combines the global search capability of the PSO algorithm and the feature learning capability of DBN, and solves the problem of local optimal solutions, which can effectively optimize the parameters of DBN, improve the model performance, and has better stability and reliability than traditional optimization methods.

[0089] Example 2

[0090] This embodiment 2 provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the lithium battery fault identification method as described above is implemented. The method includes:

[0091] Obtain voltage, current and temperature data of lithium-ion batteries during charging and discharging;

[0092] The obtained voltage, current and temperature data are processed by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein, the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features by using the Lasso regression method to select the most appropriate feature set; optimizing and selecting the extracted features by using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery status and features, and classifying the fault type.

[0093] Example 3

[0094] This embodiment 3 provides a computer device, including a memory and a processor, the processor and the memory communicate with each other, the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the lithium battery fault identification method as described above, the method comprising:

[0095] Obtain voltage, current and temperature data of lithium-ion batteries during charging and discharging;

[0096] The obtained voltage, current and temperature data are processed by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein, the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features by using the Lasso regression method to select the most appropriate feature set; optimizing and selecting the extracted features by using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery status and features, and classifying the fault type.

[0097] Example 4

[0098] This embodiment 4 provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes instructions for implementing the lithium battery fault identification method as described above, the method comprising:

[0099] Obtain voltage, current and temperature data of lithium-ion batteries during charging and discharging;

[0100] The obtained voltage, current and temperature data are processed by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein, the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features by using the Lasso regression method to select the most appropriate feature set; optimizing and selecting the extracted features by using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery status and features, and classifying the fault type.

[0101] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0102] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0103] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0105] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative work on the basis of the technical solution disclosed in the present invention should be included in the scope of protection of the present invention.

Claims

1. A lithium battery fault identification method, characterized in that: include: Obtain voltage, current and temperature data of lithium-ion batteries during charging and discharging; The obtained voltage, current and temperature data are processed by using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein, the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features by using the Lasso regression method to select the most appropriate feature set; optimizing and selecting the extracted features by using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery status and features, and classifying the fault type.

2. The lithium battery fault identification method according to claim 1, characterized in that: Features that can characterize the fault state include: voltage change rate, voltage deviation, voltage extreme difference, temperature change rate, and temperature extreme difference.

3. The lithium battery fault identification method according to claim 1, characterized in that: Lasso regression achieves feature selection and model simplification by adding L1 regularization to the model coefficients; due to the influence of L1 regularization, the solution of Lasso regression is usually sparse, that is, only a few features have coefficients that are not zero.

4. The lithium battery fault identification method according to claim 1, characterized in that: The improved PSO algorithm is used to optimize the extracted features to find the optimal feature combination. The core parameters of the particle swarm algorithm are speed and position, and the update formulas are as follows: in, is the velocity of the ith particle in the kth iteration, pbest id is the maximum value of the fitness of the i-th particle in k iterations, For location information, gbest id is the maximum value among all pbest, c is the cognitive constant, and r is the random function.

5. The lithium battery fault identification method according to claim 1, characterized in that: In the particle swarm algorithm, the inertia weight is adaptively optimized; the inertia weight w is used to balance the global and local search capabilities. The inertia weight w is related to the evolution factor f, which allows w to use S-shaped mapping:

6. The lithium battery fault identification method according to claim 5, characterized in that: In the particle swarm algorithm, the acceleration coefficient is controlled; parameter c1 represents self-cognition, which pulls the particle to its historical best position; parameter c2 represents social influence, which pushes the group to converge to the current global optimal area and helps to converge quickly; the acceleration coefficient is initialized to 2.0 and adaptively controlled according to the evolutionary state. The strategy is as follows: increase c1 and decrease c2 in the exploration state.

7. A lithium battery fault identification system, characterized in that: include: An acquisition module is used to obtain the voltage, current and temperature data of the lithium-ion battery during charging and discharging; The diagnosis module is used to process the acquired voltage, current and temperature data using a pre-trained fault diagnosis model to obtain a lithium battery fault diagnosis result; wherein the training of the fault diagnosis model includes: collecting voltage, current and temperature related data during charging and discharging of the lithium-ion battery, and establishing a corresponding data set; extracting features that can characterize the fault state from the collected data set; selecting the extracted features using the Lasso regression method to select the most appropriate feature set; optimizing the extracted features using the improved PSO algorithm to find the optimal feature combination, inputting the feature set into the network model after the particle swarm algorithm optimizes the parameters for training, and constructing a deep learning model for lithium-ion battery fault diagnosis, judging whether there is a fault by analyzing the battery state and features, and classifying the fault type.

8. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by the processor, the lithium battery fault identification method according to any one of claims 1 to 6 is implemented.

9. A computer device, characterized in that: It includes a memory and a processor, the processor and the memory communicate with each other, the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the lithium battery fault identification method as described in any one of claims 1-6.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes instructions for implementing the lithium battery fault identification method as described in any one of claims 1-6.

Citation Information

Cited By

  • Lithium battery fault prediction method and system

    CN119322270A

  • Method, medium and equipment for detecting internal structure fault of lead-acid battery

    CN120847642A

  • A method, medium and device for detecting internal structural faults of a lead-acid battery

    CN120847642B