Method and device for establishing lithium battery parallel module identification model

Abstract

The invention discloses a method for establishing a lithium battery parallel module identification model. The method comprises the following steps: measuring an electrochemical impedance spectrum of a lithium battery module to be measured; establishing an equivalent circuit based on the electrochemical impedance spectrum, simulating the equivalent circuit to obtain a simulation image, calculating an error between the simulation image and the electrochemical impedance spectrum, and extracting circuit element parameters of the equivalent circuit meeting the standard; and taking the extracted circuit element parameters as characteristic parameters, constructing a model training data set, performing model training by using a support vector machine algorithm of particle swarm optimization, and verifying a classification effect so as to obtain an optimal support vector machine model of particle swarm optimization for realizing identification of a lithium battery parallel module model. According to the method, battery data are extracted by using an electrochemical impedance spectroscopy method, a lithium battery parallel module identification model is constructed based on an equivalent circuit simulation result and a support vector machine algorithm, and classification and identification of different types of battery parallel modules are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery data drive, and in particular relates to a method and device for establishing a lithium battery parallel module identification model. Background Art

[0002] Lithium-ion batteries are widely used in many fields such as new energy vehicles and energy storage stations due to their high energy density, no memory effect and other advantages. In practical applications, due to the relatively low voltage and limited capacity of single lithium-ion batteries, different topological structures are often used to form modules with single batteries to meet the needs of different application scenarios. However, there are inevitably differences in parameters such as capacity and impedance among the single batteries inside the battery module. When the battery module is composed for charging and discharging, the single batteries with smaller capacity in the lithium-ion battery module often overcharge and over-discharge, which may affect the life of the battery module and the battery performance will drop significantly in a short period of time. In severe cases, it may cause serious accidents such as battery combustion or explosion, which greatly threatens the reliability of equipment and personnel safety. Therefore, it is very important to analyze the impedance characteristics of lithium battery modules under different conditions and to identify different conditions of lithium battery modules.

[0003] Among the existing lithium battery impedance testing methods, commonly used methods include Ohm's law to measure battery voltage and current, DC internal resistance testing, etc. Among them, using Ohm's law to measure voltage and current is relatively simple. In actual engineering applications, this method is often used to test a large number of batteries, but the test results are less accurate and have certain errors; the DC internal resistance test has relatively high accuracy, but it will force the battery to pass a large current during the test, which will cause certain damage to the battery's chemical properties and structure. In addition, the test indicators of the above two methods are relatively single, and it is difficult to identify whether there are differences in single cells inside the battery module. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for establishing a lithium battery parallel module identification model, extracts battery data using electrochemical impedance spectroscopy, collects model data sets based on equivalent circuit simulation results, and uses a support vector machine optimized by particle swarm optimization to identify battery parallel modules.

[0005] A method for establishing a lithium battery parallel module identification model comprises the following steps: Step S100, measuring the electrochemical impedance spectrum of the lithium battery module to be tested; Step S200, constructing an equivalent circuit based on the electrochemical impedance spectrum and simulating the equivalent circuit to obtain a simulation image, calculating the error between the simulation image and the electrochemical impedance spectrum, determining that the equivalent circuit meets the standard when the error is less than a predetermined threshold, and extracting circuit element parameters for the equivalent circuit that meets the standard; Step S300, preliminarily constructing a particle swarm optimized support vector machine model, taking the extracted circuit element parameters as feature parameters, and constructing a data set, wherein the data set includes a training set and a test set; Step S400, training the particle swarm optimized support vector machine model using a training set; Step S500, using a test set to verify the particle swarm optimized support vector machine model to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

[0006] Optionally, in step S200, the predetermined threshold is that the error between the simulation data and the actual data is within 5%.

[0007] Optionally, in step S200, the circuit element parameters include: ohmic internal resistance R s 、SEI film internal resistance R SEI , charge transfer resistance R ct and the Warburg impedance Z w .

[0008] Optionally, step S300 includes: Step S301: Compare the electrochemical impedance spectrum of the lithium battery module, analyze its impedance characteristics, determine the input feature vector and dimension, and form a data set; Step S302: sorting the data set, determining the proportion of the training set to the data set and shuffling the data set; looping to extract samples of different categories, implementing data set division, obtaining training sets and test sets, and implementing data normalization; Step S303: constructing a particle swarm optimized support vector machine model, specifically, selecting a model kernel function and performing parameter design; the parameters include: particle position and velocity, particle swarm size, number of iterations, learning factor, and inertia weight.

[0009] Optionally, in step S500, the particle swarm optimized support vector machine model is verified using a test set, particle swarm optimization is performed according to the accuracy result, and the model inertia weight is adjusted until the test result meets the requirement, and the model optimization is completed to obtain the optimal particle swarm optimized support vector machine model for realizing the identification of the lithium battery parallel module model.

[0010] Optionally, in step S500, the particle swarm optimized support vector machine model is verified using a test set, particle swarm optimization is performed according to the accuracy result, and the model inertia weight is adjusted until the test result meets the requirement, and the model optimization is completed to obtain the optimal particle swarm optimized support vector machine model for realizing the identification of the lithium battery parallel module model.

[0011] Optionally, in step S303, the model kernel function adopts a Gaussian (RBF) kernel function, and uses the Adam gradient descent algorithm to adjust the learning rate, and designs an L2 regularization parameter to control the inertia weight.

[0012] Optionally, in step S303, the speed and position of the particle are set by the following formula:

[0013]

[0014] Among them, ω is the inertia weight; represents the velocity vector of particle i in the dth dimension in the kth iteration; represents the position vector of particle i in the dth dimension in the kth iteration; represents the historical optimal position of the individual particle, represents the historical optimal position of the particle group; c1 is the individual learning factor, c2 is the group learning factor; rand1 and rand2 are two random numbers in the [0,1] interval.

[0015] A device for implementing a method for establishing a lithium battery parallel module identification model, comprising: An acquisition unit, which is used to measure and obtain the electrochemical impedance spectrum of the lithium battery module to be tested; A data unit connected to the acquisition unit, the data unit constructs and simulates an equivalent circuit for the electrochemical impedance spectrum to obtain simulation data of the electrochemical impedance spectrum of the lithium battery module; A model unit, which is used to preliminarily construct a support vector machine model for particle swarm optimization, use the extracted circuit element parameters as feature parameters, and construct a data set, wherein the data set includes a training set and a test set; A training unit, which is used to train the particle swarm optimized support vector machine model using a training set; A verification unit is used to verify the particle swarm optimized support vector machine model using a test set to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

[0016] Compared with the prior art, the present invention has the following advantages: The method for establishing a lithium battery parallel module identification model uses electrochemical impedance spectroscopy to obtain battery module impedance information, realizes accurate simulation of lithium battery parallel modules through equivalent circuits, and systematically summarizes lithium battery module impedance spectrum data under different conditions, which is conducive to comparative research on lithium battery module characteristics and inconsistencies under different conditions, and provides a certain basis for battery system management. Support vector machines have great advantages in solving problems with a large number of samples, a large number of feature vectors, and high data dimensions. Combining particle swarm optimization algorithms to optimize important hyperparameters of support vector machines can achieve fast and accurate classification of data. For lithium battery parallel modules under different conditions, the data differences are large, and the establishment of a battery module identification model requires the support of more sample data. The joint application of support vector machines and electrochemical impedance spectroscopy can realize the classification of battery parallel modules with high classification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings illustrate exemplary embodiments of the present invention and together with the description serve to explain the principles of the present invention. These drawings are included to provide a further understanding of the present invention and are incorporated in and constitute a part of this specification.

[0018] Figure 1 It is a flow chart of a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; Figure 2 is a comparison diagram of equivalent circuit simulation results and actual measurement results of a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; FIG3 (a) is a comparison diagram of electrochemical impedance spectra of different parallel battery modules at 50% SOC in a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; FIG3( b ) is a comparison diagram of the real part of impedance of different parallel battery modules at 50% SOC in a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; FIG3 (c) is a comparison diagram of the imaginary impedance parts of different parallel battery modules at 50% SOC in a method for establishing a lithium battery parallel module identification model in an embodiment of the present invention; Figure 4 It is a flow chart of building a support vector machine model using particle swarm optimization in a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; Figure 5 It is an electrochemical impedance spectrum image obtained by measuring a lithium battery module in a method for establishing a lithium battery parallel module identification model in one embodiment of the present invention; Figure 6 It is a confusion matrix of test results after the lithium battery module identification model of the lithium battery parallel module identification model establishment method in one embodiment of the present invention is completed. DETAILED DESCRIPTION

[0019] The following is combined with Figures 1 to 6 The present invention is further described in detail with reference to the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant contents, rather than to limit the present invention. It is also necessary to explain that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings.

[0020] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0021] Unless otherwise specified, the exemplary embodiments / embodiments shown will be understood as exemplary features that provide various details of some ways in which the technical concept of the present invention can be implemented in practice. Therefore, unless otherwise specified, the features of the various embodiments / embodiments can be combined, separated, interchanged and / or rearranged without departing from the technical concept of the present invention.

[0022] Cross-hatching and / or shading may be used in the drawings, typically to make the boundaries between adjacent components clear. As such, unless otherwise specified, the presence or absence of cross-hatching or shading does not convey or indicate any preference or requirement for the specific materials, material properties, dimensions, proportions, commonalities between the components shown, and / or any other characteristics, attributes, properties, etc. of the components. In addition, in the drawings, the sizes and relative sizes of the components may be exaggerated for clarity and / or descriptive purposes. When the exemplary embodiments may be implemented differently, the specific process sequence may be performed in a different order than the described step sequence. For example, two successively described processes may be performed substantially simultaneously or in an order opposite to the described order. In addition, the same figure numbers represent the same components.

[0023] When a component is referred to as being "on" or "over," "connected to," or "coupled to" another component, the component may be directly on, directly connected to, or directly coupled to the other component, or intervening components may be present. However, when a component is referred to as being "directly on," "directly connected to," or "directly coupled to" another component, there are no intervening components. For this purpose, the term "connected" may refer to a physical connection, an electrical connection, etc., with or without intervening components.

[0024] For descriptive purposes, the present invention may use spatially relative terms such as "below", "beneath", "under", "lower", "above", "upper", "above", "higher", and "side (e.g., as in "sidewall")" to describe the relationship of one component to another (other) component as shown in the accompanying drawings. The spatially relative terms are intended to encompass different orientations of the device in use, operation, and / or manufacture in addition to the orientation depicted in the accompanying drawings. For example, if the device in the accompanying drawings is turned over, components described as "below" or "beneath" other components or features would then be positioned "above" the other components or features. Thus, the exemplary term "below" can encompass both the "above" and "below" orientations. Furthermore, the device may be otherwise oriented (e.g., rotated 90 degrees or at other orientations), and the spatially relative descriptors used herein should be interpreted accordingly.

[0025] The terms used here are for the purpose of describing specific embodiments, and are not intended to be restrictive. As used here, unless the context clearly indicates otherwise, the singular forms "one (kind, person)" and "said (the)" are also intended to include plural forms. In addition, when the terms "comprise" and / or "include" and their variations are used in this specification, it is explained that there are stated features, integral bodies, steps, operations, parts, assemblies and / or their groups, but it is not excluded that there are or add one or more other features, integral bodies, steps, operations, parts, assemblies and / or their groups. It should also be noted that, as used here, the terms "substantially", "approximately" and other similar terms are used as approximate terms and not as degree terms, so that they are used to explain the inherent deviations of the measured values, calculated values ​​and / or the values ​​provided that will be recognized by those of ordinary skill in the art.

[0026] In one embodiment, see Figure 1 The present invention provides a method for establishing a lithium battery parallel module identification model, comprising: Step S100, measuring the electrochemical impedance spectrum of the lithium battery module to be tested; Step S200, constructing an equivalent circuit based on the electrochemical impedance spectrum and simulating the equivalent circuit to obtain a simulation image, calculating the error between the simulation image and the electrochemical impedance spectrum, determining that the equivalent circuit meets the standard when the error is less than a predetermined threshold, and extracting circuit element parameters for the equivalent circuit that meets the standard; Step S300, preliminarily constructing a particle swarm optimized support vector machine model, taking the extracted circuit element parameters as feature parameters, and constructing a data set, wherein the data set includes a training set and a test set; Step S400, training the particle swarm optimized support vector machine model using a training set; Step S500, using a test set to verify the particle swarm optimized support vector machine model to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

[0027] In one embodiment, in step S200, in order to realize accurate simulation of the equivalent circuit model, a higher simulation accuracy is required, and the predetermined threshold is that the error between the simulation data and the actual data is within 5%.

[0028] In one embodiment, in step S200, the circuit element parameters include: ohmic internal resistance R s 、SEI film internal resistance R SEI , charge transfer resistance R ct and the Warburg impedance Z w .

[0029] In one embodiment, step S300 includes: Step S301: Compare the electrochemical impedance spectrum of the lithium battery module, analyze its impedance characteristics, determine the input feature vector and dimension, and form a data set; Step S302: sorting the data set, determining the proportion of the training set to the data set and shuffling the data set; looping to extract samples of different categories, implementing data set division, obtaining training sets and test sets, and implementing data normalization; Step S303: constructing a particle swarm optimized support vector machine model, specifically, selecting a model kernel function and performing parameter design; the parameters include: particle position and velocity, particle swarm size, number of iterations, learning factor, and inertia weight.

[0030] Step S400, training the particle swarm optimized support vector machine model using a training set; Step S500, using the test set to verify the particle swarm optimized support vector machine model, performing particle swarm optimization according to the accuracy result, adjusting the model inertia weight, until the test result meets the requirements, completing the model optimization, and obtaining the optimal particle swarm optimized support vector machine model for realizing the identification of the lithium battery parallel module model.

[0031] In one embodiment, in step S303, the model kernel function adopts a Gaussian (RBF) kernel function, and uses the Adam gradient descent algorithm to adjust the learning rate, and designs an L2 regularization parameter to control the inertia weight.

[0032] In one embodiment, in step S303, the speed and position of the particle are set by the following formula:

[0033]

[0034] Among them, ω is the inertia weight; represents the velocity vector of particle i in the dth dimension in the kth iteration; represents the position vector of particle i in the dth dimension in the kth iteration; represents the historical optimal position of the individual particle, represents the historical optimal position of the particle group; c1 is the individual learning factor, c2 is the group learning factor; rand1 and rand2 are two random numbers in the [0,1] interval.

[0035] In one embodiment, in actual working conditions, the temperature and the SOC and SOH of the battery will affect the electrochemical impedance spectroscopy test of the battery, and the measured electrochemical impedance spectrum will be offset or deformed, causing the simulation results to change. In order to accurately test the electrochemical impedance spectrum of lithium battery parallel modules, it is necessary to accurately determine the state parameters of the battery. This example conducted an electrochemical impedance spectroscopy test on two parallel lithium battery modules at 25°C, 100% SOC, and SOH of 98%, and performed an equivalent circuit simulation of the electrochemical impedance spectrum. Some simulation results are shown in Table 1: Table 1 Simulation data of electrochemical impedance spectroscopy of parallel lithium batteries

[0036] The above table shows some simulation results of the electrochemical impedance spectrum of parallel lithium battery modules at 1kHz-0.1Hz. The accuracy of the simulation results can be judged based on the error of the real part of the impedance and the negative imaginary part of the impedance in the table. It can be seen from the above table that the simulation results of the real part of the impedance are basically consistent with the actual test results, and most of the errors in the test results of the negative imaginary part are also maintained within 5%. Due to the inevitable fluctuations in the system during the test, individual large deviations in the simulation results are acceptable.

[0037] According to the above results, the equivalent circuit simulation results of two parallel lithium battery modules at 25°C, 100% SOC, and 98% SOH are as follows: Figure 2 As shown, the red data points are test results and the black curve is the simulation result.

[0038] In one embodiment, constructing the particle swarm optimized support vector machine model includes: In this embodiment, the support vector machine (SVM) is a generalized linear classifier that classifies data in a supervised learning manner, and its decision boundary is the maximum margin hyperplane solved for the learning samples. Compared with logistic regression and neural networks, support vector machines have better performance in dealing with nonlinear problems: when dealing with linearly inseparable problems, samples in the low-dimensional input space can be mapped to a high-dimensional space by adding slack variables and using nonlinear mapping to make them linearly separable, so that the optimal classification hyperplane can be found in the feature space. The SVM design considers the following issues: 1. Confirmation of kernel function. Nonlinear data is usually indivisible, so it is necessary to use kernel function to map the original data to high-dimensional space, so that the data can be separated in high-dimensional space. At the same time, kernel function can simplify the dot product operation in high-dimensional space into kernel function calculation in low-dimensional space, greatly reducing the complexity of the model.

[0039] In this embodiment, various component parameters are extracted based on the equivalent circuit model of the lithium battery. Combined with the electrochemical impedance spectrum test results of different parallel battery packs under the same SOC state shown in Figures 3 (a) to 3 (c) and the Bode diagrams of the real and imaginary parts of the impedance, it can be judged that the difference in battery modules has a greater impact on the real part of the battery impedance and a smaller impact on the imaginary part. Therefore, the characteristic parameter of the model is defined as: Ohmic internal resistance R s 、SEI film internal resistance R SEI , charge transfer resistance R ct and Warburg impedance Z w , a total of 4 feature parameters.

[0040] In this embodiment, there are 150 groups of experimental data in total, the feature quantity is relatively small, and the model sample data is large. In this case, a support vector machine with a Gaussian kernel function is usually selected.

[0041]

[0042] In the formula, x i 、x j is the battery impedance spectrum data, γ is the parameter of the Gaussian kernel function, and is one of the important indicators for subsequent hyperparameter tuning.

[0043] 2. Slack variable design. Different from the linear separable problem, in most cases the data cannot satisfy the condition of formula (1): (1) In the formula, y i is the index of each training data, ω T To determine the perpendicular normal vector of the hyperplane, b is the distance between the hyperplane and the source point.

[0044] The slack variable allows the distance from some points to the classification plane to not meet the original requirements. A slack term parameter ε is added to the specific constraint conditions. i ≥ 0, becomes: (2) When ε i ≥ 0 is large enough, all training points meet the above conditions, but ε i It is important to avoid taking values ​​that are too large. Therefore, by adding the penalty term C to the objective function, we get the optimization problem:

[0045] Among them, ε∈R n , C is the penalty parameter. The objective function needs to maximize the interval and minimize ε i Part of the penalty term is used to achieve a trade-off between the two.

[0046] 3. Problem solving and optimization. Solving the above optimization problem is often transformed into a quadratic function extreme value problem, where , and the Lagrangian function is used to solve it. According to the test results, the penalty parameter C and the kernel function parameter γ are mainly tuned until the test results meet expectations.

[0047] In this embodiment, the particle swarm optimization (PSO) algorithm regards the problem to be solved as a "particle". Each particle flies in the solution space of the problem and adjusts its flight direction and speed based on its historical search experience and the behavior of surrounding particles. Through multiple iterations, the particle swarm will gradually approach the optimal solution to the problem. Each particle contains two main features: position and speed. Among them, the position represents the current solution, and the speed determines the position of the next step. The algorithm steps of PSO are as follows: 1. Initialize, generate a group of particles, assign the velocity v and position x of each particle, and record the current individual optimal position pBest of each particle and the optimal individual position gBest in the group.

[0048] 2. In each generation of evolution, the fitness function value of each particle is calculated. If the current fitness function value is better than the historical optimal value, pBest is updated; if the current fitness function value is better than the global historical optimal value, gBest is updated. The speed and position of the d-th dimension of each particle i are updated according to formulas (3) and (4) respectively: (3) (4) In formula (3), ω is the inertia weight, which represents the influence of the velocity of the previous generation of particles on the velocity of the current generation of particles. The convergence speed and the globality of the convergence are balanced by dynamically adjusting the size of ω. represents the velocity vector of particle i in the dth dimension in the kth iteration; represents the position vector of particle i in the dth dimension in the kth iteration; c1 is the individual learning factor, c2 is the group learning factor; rand1 and rand2 are two random numbers in the [0,1] interval, which increase the randomness of the particle swarm search.

[0049] 3. When the training reaches the maximum number of iterations or meets other conditions, PSO stops, otherwise repeat the second step to update the optimal position of the particle and the optimal position of the group. At the same time, the hyperparameters can be tuned. The core parameters of the PSO algorithm mainly include the particle group size, inertia weight, and individual and group learning factors. By adjusting the above parameters, the optimization effect of PSO can be improved.

[0050] In summary, the implementation process of the particle swarm optimization support vector machine model (PSO-SVM) is as follows: Figure 4 As shown, the specific design process is: 1. Data acquisition: Use electrochemical impedance spectroscopy to test lithium battery parallel modules, simulate and screen equivalent circuits, and extract component parameters.

[0051] 2. Arrange the data set, determine the number of input feature vectors to be 4 categories, and determine the proportion of the training set and the test set in the data set. The training set accounts for 70% of the data set, and the test set accounts for 30% of the data set. Shuffle the training set and the test set.

[0052] 3. Confirm the kernel function and use the Gaussian kernel function. Parameter setting: design the particle swarm size sizepop to 20, the maximum number of iterations maxgen to 500, the learning rate c1 to 1.5, c2 to 1.6, initialize the penalty parameter C and kernel function parameter γ in the SVM parameters, where the penalty parameter C ranges from 0.1 to 1000, the kernel function parameter γ ranges from 0.01 to 1000, and initialize the particle swarm velocity v.

[0053] 4. Model training: First, generate a random particle swarm, initialize the speed, calculate its fitness, and update the particle speed and position according to formulas (3) and (4), find individual extreme values ​​and global extreme points, and the optimal solution of particles and swarms. Next, perform multiple rounds of iterations according to the parameters of the particle swarm optimization algorithm, calculate the particle fitness and update the optimal position and optimal speed of individuals and swarms once each iteration, until the maximum number of iterations is reached or the optimal position (optimal solution) of the particle swarm that meets the requirements is found in advance.

[0054] 5. Test set verification, perform hyperparameter tuning based on the recognition effect until the expected effect is achieved and output the optimal model.

[0055] In order to facilitate further understanding of the method disclosed in the present invention, a group of parallel lithium battery modules to be tested is taken as an example for explanation below.

[0056] 1. Obtain the electrochemical impedance spectrum of the lithium battery module to be tested, obtain the electrochemical impedance spectrum data of the lithium battery module, and plot the data in the real and imaginary part of the impedance diagram. The resulting image is as follows: Figure 5 As shown; 2. Yes Figure 5 Perform equivalent circuit simulation analysis to extract the ohmic internal resistance R s 、SEI film internal resistance R SEI , charge transfer resistance R ct and the Warburg impedance Z w As raw data; 3. Input the above raw data into the support vector machine model optimized by particle swarm optimization to complete the training and testing of the lithium battery parallel module identification model. The confusion matrix of the classification result is as follows: Figure 6 shown. Figure 6 Among the 45 test set data, only one group was classified incorrectly, with an accuracy rate of up to 97.78%, which meets the classification accuracy requirements and can be used as the output of the lithium battery parallel module identification model.

[0057] In one embodiment, the present invention provides a device for implementing a method for establishing a lithium battery parallel module identification model, comprising: An acquisition unit, which is used to measure and obtain the electrochemical impedance spectrum of the lithium battery module to be tested; A data unit connected to the acquisition unit, the data unit constructs and simulates an equivalent circuit for the electrochemical impedance spectrum to obtain simulation data of the electrochemical impedance spectrum of the lithium battery module; A model unit, which is used to preliminarily construct a support vector machine model for particle swarm optimization, use the extracted circuit element parameters as feature parameters, and construct a data set, wherein the data set includes a training set and a test set; A training unit, which is used to train the particle swarm optimized support vector machine model using a training set; A verification unit is used to verify the particle swarm optimized support vector machine model using a test set to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

[0058] In the description of this specification, the description with reference to the terms "one embodiment / method", "some embodiments / methods", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment / method or example are included in at least one embodiment / method or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment / method or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments / methods or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments / methods or examples described in this specification and the features of the different embodiments / methods or examples, unless they are contradictory.

[0059] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0060] It should be understood by those skilled in the art that the above embodiments are only for the purpose of clearly illustrating the present invention, and are not intended to limit the scope of the present invention. For those skilled in the art, other changes or modifications may be made based on the above disclosure, and these changes or modifications are still within the scope of the present invention.

Claims

1. A method for establishing a lithium battery parallel module identification model, characterized in that: It includes the following steps: Step S100, measuring the electrochemical impedance spectrum of the lithium battery module to be tested; Step S200, constructing an equivalent circuit based on the electrochemical impedance spectrum and simulating the equivalent circuit to obtain a simulation image, calculating the error between the simulation image and the electrochemical impedance spectrum, determining that the equivalent circuit meets the standard when the error is less than a predetermined threshold, and extracting circuit element parameters for the equivalent circuit that meets the standard; Step S300, preliminarily constructing a particle swarm optimized support vector machine model, taking the extracted circuit element parameters as feature parameters, and constructing a data set, wherein the data set includes a training set and a test set; Step S400, training the particle swarm optimized support vector machine model using a training set; Step S500, using a test set to verify the particle swarm optimized support vector machine model to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

2. The method according to claim 1, characterized in that Preferably, in step S200, the predetermined threshold is that the error between the simulation data and the actual data is within 5%.

3. The method according to claim 1, characterized in that: In step S200, the circuit element parameters include: ohmic internal resistance R s 、SEI film internal resistance R SEI , charge transfer resistance R ct and the Warburg impedance Z w .

4. The method according to claim 1, characterized in that Step S300 includes: Step S301: Compare the electrochemical impedance spectrum of the lithium battery module, analyze its impedance characteristics, determine the input feature vector and dimension, and form a data set; Step S302: sorting the data set, determining the proportion of the training set to the data set and shuffling the data set; looping to extract samples of different categories, implementing data set division, obtaining training sets and test sets, and implementing data normalization; Step S303: constructing a particle swarm optimized support vector machine model, specifically, selecting a model kernel function and performing parameter design; the parameters include: particle position and velocity, particle swarm size, number of iterations, learning factor, and inertia weight.

5. The method according to claim 1, characterized in that Step S500, using the test set to verify the particle swarm optimized support vector machine model, performing particle swarm optimization according to the accuracy result, adjusting the model inertia weight, until the test result meets the requirements, completing the model optimization, and obtaining the optimal particle swarm optimized support vector machine model for realizing the identification of the lithium battery parallel module model.

6. The method according to claim 1, characterized in that Step S500, using the test set to verify the particle swarm optimized support vector machine model, performing particle swarm optimization according to the accuracy result, adjusting the model inertia weight, until the test result meets the requirements, completing the model optimization, and obtaining the optimal particle swarm optimized support vector machine model for realizing the identification of the lithium battery parallel module model.

7. The method according to claim 4, characterized in that In step S303, the model kernel function adopts a Gaussian (RBF) kernel function, and uses the Adam gradient descent algorithm to adjust the learning rate, and designs an L2 regularization parameter to control the inertia weight.

8. The method according to claim 4, characterized in that In step S303, the speed and position of the particle are set by the following formula: ， ， Among them, ω is the inertia weight; represents the velocity vector of particle i in the dth dimension in the kth iteration; represents the position vector of particle i in the dth dimension in the kth iteration; represents the historical optimal position of the individual particle, represents the historical optimal position of the particle group; c1 is the individual learning factor, c2 is the group learning factor; rand1 and rand2 are two random numbers in the [0,1] interval.

9. A device for implementing the method for establishing a lithium battery parallel module identification model according to any one of claims 1 to 8, characterized in that: It includes: An acquisition unit, which is used to measure and obtain the electrochemical impedance spectrum of the lithium battery module to be tested; A data unit connected to the acquisition unit, the data unit constructs and simulates an equivalent circuit for the electrochemical impedance spectrum to obtain simulation data of the electrochemical impedance spectrum of the lithium battery module; A model unit, which is used to preliminarily construct a support vector machine model for particle swarm optimization, use the extracted circuit element parameters as feature parameters, and construct a data set, wherein the data set includes a training set and a test set; A training unit, which is used to train the particle swarm optimized support vector machine model using a training set; A verification unit is used to verify the particle swarm optimized support vector machine model using a test set to obtain an optimal particle swarm optimized support vector machine model for realizing the identification of a lithium battery parallel module model.

Images