Method and system for predicting SOH by using battery EIS

Abstract

The invention discloses a method and system for predicting SOH by using a battery EIS and a readable storage medium. The method comprises the following steps: testing and collecting battery EIS data; a PCA-CNN-LSTM model is constructed and trained; and inputting the EIS data into the PCA-CNN-LSTM model, performing battery SOH prediction by using the PCA-CNN-LSTM model, and outputting a prediction result. According to the method, PCA, CNN and LSTM models are combined, and SOH is jointly predicted based on EIS data. EIS is regarded as a characteristic, and a real part and an imaginary part of impedance data corresponding to each frequency are regarded as characteristic values. Firstly, a PCA model is used for carrying out dimensionality reduction on an EIS characteristic value, a small number of refined principal components are extracted, then a CNN model is used for extracting spatial characteristics of an impedance spectrum, an LSTM model is used for extracting a time sequence change rule of the impedance spectrum, and SOH is jointly predicted. According to the method, the SOH prediction precision of the battery can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of electrochemical technology, and more specifically, to a method, system and readable storage medium for predicting SOH using battery EIS. Background Art

[0002] Electrochemical Impedance Spectroscopy (EIS) is a test method used to characterize the electrochemical process inside a battery. EIS can be used to test the impedance spectrum of the battery, including real impedance and imaginary impedance. The attenuation of the battery's state of health (SOH) will lead to an increase in the battery's internal resistance and AC impedance. Therefore, the impedance spectrum obtained by the EIS test can be used to predict the battery's SOH. Impedance spectrum data generally covers dozens of frequency ranges, has a high dimensionality, and contains certain redundant information, which reduces the accuracy of the model.

[0003] Chinese patent CN115097321A, published on November 15, 2022, provides a lithium battery EIS and SOH estimation method based on digital twins. The method includes: constructing a cloud-based digital twin model of lithium batteries based on fractional-order differential equations, continuously transmitting the twin data of the open-circuit voltage of the voltage sensor and the working current collected by the current sensor to the cloud server through the on-board TBOX, and iteratively optimizing the state parameters; constructing a state parameter filter to obtain an estimated electrochemical impedance spectrum; establishing a regression model for the growth of the electrochemical impedance of the lithium battery, introducing it into a non-parametric particle filter framework, estimating the SOH of the lithium battery, and obtaining the SOH of the lithium battery. Although this scheme predicts the battery SOH through the impedance spectrum obtained by the EIS test, it has the defects of data noise and insufficient model accuracy. Summary of the invention

[0004] In view of the above problems, an object of the present invention is to provide a method, system and readable storage medium for predicting SOH using battery EIS.

[0005] A first aspect of the present invention provides a method for predicting SOH using battery EIS, the method comprising:

[0006] Test and collect battery EIS data;

[0007] Build and train the PCA-CNN-LSTM model;

[0008] The EIS data is input into the PCA-CNN-LSTM model, the PCA-CNN-LSTM model is used to predict the battery SOH, and the prediction results are output.

[0009] Preferably, the PCA-CNN-LSTM model includes: an input layer, a PCA model, a CNN model, an LSTM model, and a linear regression layer;

[0010] The input layer is used to input EIS data into the PAC model;

[0011] The PAC model extracts the principal components of the EIS data to obtain the principal component data, and outputs the principal component data to the CNN model;

[0012] The CNN model extracts EIS spatial characteristics from the principal component data to obtain EIS spatial features, and outputs the EIS spatial features to the LSTM model;

[0013] The LSTM model extracts the time series variation law of EIS according to the spatial characteristics of EIS, and outputs the time series variation law of EIS to the linear regression layer;

[0014] The linear regression layer predicts the battery SOH based on the EIS spatial characteristics and the time series variation law of EIS.

[0015] Preferably, the collecting of battery EIS data is specifically:

[0016] Collect battery EIS data at preset frequency intervals;

[0017] When collecting battery EIS data, record the real and imaginary impedance data corresponding to each frequency EIS data; as well as the target value to be predicted, that is, the battery capacity data.

[0018] Preferably, the PAC model performs principal component extraction on the EIS data, specifically:

[0019] The EIS data is input into the PCA model, and the high-dimensional feature space is transformed into several mutually orthogonal principal component spaces using linear transformation;

[0020] Among them, the principal components in the principal component space explain part of the variance of the original data;

[0021] Calculate the percentage of data variance explained by each principal component to the total variance of the original data;

[0022] If the cumulative explained variance rate of the principal component exceeds the preset threshold, it will be determined as the principal component data.

[0023] Preferably, the CNN model includes two convolutional layers and one pooling layer, the activation functions of the convolutional layers are all relu functions, the number of filters is set to 128, the kernel size is 1, and the pooling size of the pooling layer is 1.

[0024] Preferably, the LSTM model includes an LSTM layer and a fully connected layer, the number of neurons in the LSTM layer is 128, the number of neurons in the fully connected layer is 64, and the activation functions are both relu functions.

[0025] Preferably, the activation function of the linear regression layer is a linear function.

[0026] Preferably, the loss function of the PCA-CNN-LSTM model is mean square error, the optimizer is rmsprop, and the evaluation indicator metrics is mean absolute error.

[0027] A second aspect of the present invention provides a system for predicting SOH using battery EIS, characterized in that it includes a memory and a processor, wherein the memory includes a method program for predicting SOH using battery EIS, and when the method program for predicting SOH using battery EIS is executed by the processor, the steps of predicting SOH using battery EIS are implemented.

[0028] A third aspect of the present invention provides a readable storage medium, which includes a method program for predicting SOH using battery EIS. When the method program for predicting SOH using battery EIS is executed by a processor, the steps of the method for predicting SOH using battery EIS are implemented.

[0029] Compared with the prior art, the beneficial effect of the technical solution of the present invention is that the present invention combines PCA, CNN and LSTM models to jointly predict SOH based on EIS data. EIS is regarded as a feature, and the real part and imaginary part of the impedance data corresponding to each frequency are used as eigenvalues. First, the PCA model is used to reduce the dimension of the EIS eigenvalues ​​to extract a few refined principal components, and then the CNN model is used to extract the spatial characteristics of the impedance spectrum, and the LSTM model is used to extract the time series variation law of the impedance spectrum to jointly predict SOH. The present invention can effectively improve the accuracy of battery SOH prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of a method for predicting SOH using battery EIS as described in Example 1.

[0031] Figure 2 Schematic diagram of the PCA-CNN-LSTM model structure.

[0032] Figure 3 Schematic diagram of PCA geometric interpretation.

[0033] Figure 4 Schematic diagram of convolution calculation.

[0034] Figure 5Schematic diagram of pooling calculation.

[0035] Figure 6 This is a schematic diagram of the LSTM model structure.

[0036] Figure 7 This is a sample graph of the prediction results of using battery EIS to predict SOH. DETAILED DESCRIPTION

[0037] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0038] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0039] Example 1

[0040] like Figure 1 As shown, this embodiment discloses a method for predicting SOH using battery EIS, the method comprising:

[0041] Test and collect battery EIS data;

[0042] Build and train the PCA-CNN-LSTM model;

[0043] The EIS data is input into the PCA-CNN-LSTM model, the PCA-CNN-LSTM model is used to predict the battery SOH, and the prediction results are output.

[0044] It should be noted that, since the impedance spectrum data generally covers dozens of frequency ranges when predicting SOH, the dimension is high, and there is a certain amount of redundant information, which leads to reduced model accuracy. Therefore, principal component analysis (Principle Component Analysis, PCA) can be used to reduce the dimension of the impedance spectrum, extract the main information, and reduce the noise of the data. In addition, the convolution neural network (Convolution neural network, CNN) can extract the spatial characteristics of the impedance spectrum, and the long short-term memory network (Long Short-Term Memory, LSTM) can learn the time series change law of the impedance spectrum. This application uses PCA to remove the noise of EIS data, and uses a combination of CNN and LSTM to automatically learn the complex mapping relationship between EIS and SOH to improve the accuracy of battery EIS data in predicting SOH.

[0045] According to an embodiment of the present invention, the PCA-CNN-LSTM model includes: an input layer, a PCA model, a CNN model, an LSTM model, and a linear regression layer;

[0046] The input layer is used to input EIS data into the PAC model;

[0047] The PAC model extracts the principal components of the EIS data to obtain the principal component data, and outputs the principal component data to the CNN model;

[0048] The CNN model extracts EIS spatial characteristics from the principal component data to obtain EIS spatial features, and outputs the EIS spatial features to the LSTM model;

[0049] The LSTM model extracts the time series variation law of EIS according to the spatial characteristics of EIS, and outputs the time series variation law of EIS to the linear regression layer;

[0050] The linear regression layer predicts the battery SOH based on the EIS spatial characteristics and the time series variation law of EIS.

[0051] It should be noted that this application combines PCA, CNN and LSTM models to jointly predict SOH based on EIS data. EIS data is regarded as a feature, and the real and imaginary parts of the impedance data corresponding to each frequency are used as eigenvalues. First, the PCA model is used to reduce the dimension of the EIS eigenvalues ​​and extract a few refined principal components. Then, the CNN model is used to extract the spatial features of the impedance spectrum, and the LSTM model is used to extract the time series variation law of the impedance spectrum to jointly predict SOH.

[0052] The overall structure of the PCA-CNN-LSTM model is as follows Figure 2 As shown, the input part real1 to real60 represents the real data of the impedance data corresponding to high frequency to low frequency, and imaginary1 to imaginary60 represents the imaginary data of the impedance data corresponding to high frequency to low frequency; P1 to P10 represent the principal components obtained after the EIS input data is calculated by PCA; followed by a 2-layer convolution calculation structure with an activation function of relu; then, the convolution calculation result needs to be subjected to a maximum pooling operation; then, the maximum pooling calculation result is flattened, that is, flattened; then, the flattened layer is input into the LSTM structure with an activation function of relu; next, it is connected to a fully connected layer; finally, the SOH is output through a linear regression layer.

[0053] According to an embodiment of the present invention, the collecting of battery EIS data is specifically:

[0054] Collect battery EIS data at preset frequency intervals;

[0055] When collecting battery EIS data, record the real and imaginary impedance data corresponding to each frequency EIS data; as well as the target value to be predicted, that is, the battery capacity data.

[0056] It should be noted that testing and collecting battery electrochemical impedance spectrum data requires professional impedance spectrum testing instruments and needs to cover a certain frequency range. The frequency range can be set from 10-2HZ to 105HZ. Multiple frequency values ​​are selected at certain intervals in the middle of the frequency range. The frequency values ​​are generally selected from 5 to 60. In this embodiment, 60 are selected as a sample.

[0057] When collecting data, it is necessary to record the real and imaginary data of the impedance corresponding to each frequency, as well as the target value that needs to be predicted, which is the corresponding battery capacity data.

[0058] According to an embodiment of the present invention, the PAC model extracts the principal components of the EIS data, specifically:

[0059] The EIS data is input into the PCA model, and the high-dimensional feature space is transformed into several mutually orthogonal principal component spaces using linear transformation;

[0060] Among them, the principal components in the principal component space explain part of the variance of the original data;

[0061] Calculate the percentage of data variance explained by each principal component to the total variance of the original data;

[0062] If the cumulative explained variance rate of the principal component exceeds the preset threshold, it will be determined as the principal component data.

[0063] It should be noted that, since the EIS test data in this embodiment has 60 frequencies, each frequency corresponds to an impedance value, and each impedance value has real data and imaginary data, there are 120 input data in total. Among these input data, most frequencies are not only useless for SOH prediction, but also generate noise data, which will reduce the prediction accuracy. It is necessary to reduce the dimension of the input data and extract the principal component.

[0064] PCA, or principal component analysis, is a linear dimensionality reduction method. It converts a high-dimensional data set into a low-dimensional data set through a standard orthogonal transformation while retaining the information of the original data set to the maximum extent. The geometric interpretation of PCA is as follows Figure 3 shown.

[0065] exist Figure 3In the example, the sample points represented by the circles have a large degree of discreteness along both the X-axis and the Y-axis, and the degree of discreteness can be quantitatively expressed by the variance of the observed variable. Obviously, if only one of X or Y is considered, the economic information contained in the original data will be greatly lost.

[0066] Suppose that the original coordinate axes are transformed to obtain new coordinate axes, denoted as F1 and F2. When the coordinate axis F1 is parallel to the major axis of the ellipse, it means that the variables of the major axis describe the main changes in the data; when the coordinate axis F2 is parallel to the minor axis of the ellipse, it means that the variables of the minor axis describe the minor changes in the data. However, the new coordinate axes are usually not parallel to the major and minor axes of the ellipse. Therefore, it is necessary to find the major and minor axes of the ellipse and transform them so that the new variables are parallel to the major and minor axes of the ellipse. If the major axis variable represents most of the information contained in the data, use this variable to replace the original two variables (discarding the minor dimension), and the dimensionality reduction is completed. The greater the difference between the major and minor axes of the ellipse, the more information the major axis can describe, and the better the dimensionality reduction effect.

[0067] Therefore, PCA is equivalent to a coordinate rotation transformation, which rotates the data covariance matrix, projects it on the new coordinate system, and then selects only those components with higher projection metrics as the final result.

[0068] The main roles of PCA in machine learning are:

[0069] 1) Dimensionality reduction: compress data dimensions while retaining the main information, reducing the complexity of the algorithm and speeding up calculations.

[0070] 2) Denoising: The noise component corresponding to the secondary principal component is removed by projection to improve the prediction accuracy.

[0071] 3) Preprocessing: PCA can be used as data preprocessing for machine learning algorithms to extract key features.

[0072] In this embodiment, the PCA model is used to extract the principal components of the EIS input data to achieve the purpose of dimensionality reduction and denoising. The number of principal components is 10.

[0073] According to an embodiment of the present invention, the CNN model includes two convolutional layers and one pooling layer, the activation functions of the convolutional layers are all relu functions, the number of filters is set to 128, the kernel size is 1, and the pooling size of the pooling layer is 1.

[0074] It should be noted that CNN can automatically learn the spatial correlation of input data and extract high-level abstract features.

[0075] In this embodiment, the CNN structure includes two convolutional layers and one pooling layer. The activation functions of the convolutional layers are all relu functions. The number of filters is set to 128, the kernel size is 1, and the pooling size of the pooling layer is also 1.

[0076] like Figure 4 As shown, in the convolution layer of this embodiment, there is a 1*6*6 rectangle as input data, called the input feature map, which is convolved through a 3*3 filter (also called a convolution kernel), that is, the tensor product is calculated starting from the 3*3 receiving field area in the upper left corner of the input, and moves to the right each time the calculation is completed, with a stride of 1. After completing this calculation, a 1*4*4 output feature map, namely, Result, is obtained.

[0077] The pooling layer divides the input feature map into several local areas and reduces the size of the feature map by counting the characteristics of the data in each local area. The pooling layer performs nonlinear downsampling on the input and takes the maximum value (Max Pooling) or the average value (Avg Pooling) in the local neighborhood as the output. The core of the convolution layer is the convolution kernel, and the core of the pooling layer is the pooling kernel. The size of the pooling kernel is generally 3×3 or 2×2, and the step size is 1 or 2. In the process of pooling the feature map, the pooling kernel moves according to the pooling step size. Commonly used pooling methods include average pooling and maximum pooling. Maximum pooling uses the maximum value to characterize the local area, making the overall feature contrast of the image significant. Average pooling uses the average value to characterize the local area, making the overall feature information of the image smoother. When training the model, the pooling layer can improve the generalization and robustness of the model and avoid network overfitting.

[0078] The calculation logic of maximum pooling and average pooling is as follows Figure 5 As shown in Figure 2, the size of the pooled input feature map is 4×4, the size of the maximum pooling and average pooling convolution kernels are both 2×2, and the size of the pooled output feature map is 2×2.

[0079] Each convolution layer can extract high-level features of the input data, and the pooling layer can downsample to reduce the data dimension. Finally, the three-dimensional features are mapped to a one-dimensional vector through the Flatten layer as the spatial feature expression of EIS.

[0080] According to an embodiment of the present invention, the LSTM model includes an LSTM layer and a fully connected layer, the number of neurons in the LSTM layer is 128, the number of neurons in the fully connected layer is 64, and the activation functions are both relu functions.

[0081] It should be noted that LSTM is a special RNN that can learn long-term dependencies and capture the dynamic changes of time series data. Figure 6 As shown, the structure of the LSTM model contains the following components:

[0082] (1) Input gate: determines which parts of the current input information are allowed to enter the state unit.

[0083] (2) Forget gate: determines which information in the state unit needs to be forgotten or removed.

[0084] (3) Status unit: The core component used to store historical status information.

[0085] (4) Output gate: determines the information output to the next hidden layer.

[0086] (5) Activation function (Sigmoid, Tanh): used for nonlinear transformation.

[0087] (6) Output layer: LSTM outputs data corresponding to the input data.

[0088] LSTM controls the flow of information entry, retention and output by opening and closing the input gate, forget gate and output gate, and realizes the memory and processing of time series information. Compared with simple RNN, LSTM can better capture the long-term dependencies in time series data, process longer time series, and has more powerful modeling capabilities.

[0089] The LSTM model structure of this embodiment includes an LSTM layer and a fully connected layer. The number of neurons in the LSTM layer is 128, the number of neurons in the fully connected layer is 64, and the activation function is the relu function. The LSTM layer can learn the long-term dependencies of the input data and identify the dynamic change pattern of the time series. The fully connected layer further extracts the LSTM output to obtain the high-level feature expression of the time series.

[0090] According to an embodiment of the present invention, it should be noted that the output layer of the PCA-CNN-LSTM model is a linear regression layer, the predicted value is the battery capacity, i.e., SOH, and the activation function is a linear function. The loss function of the PCA-CNN-LSTM model is the mean square error (MSE); the optimizer is rmsprop; and the evaluation metric is the mean absolute error (MAE), i.e., the sum of the absolute values ​​of the difference between the target value and the predicted value.

[0091] As a specific embodiment, the following is described in conjunction with a specific example:

[0092] The present invention tests the effect on a public data set, the data source is https: / / github.com / YunweiZhang / ML-identify-battery-degradation. The impedance data therein is used to predict the battery capacity. The data set is divided into a training set and a test set in a ratio of 8:2, the PCA-CNN-LSTM model is trained on the training set, and the effect is tested on the test set. The test data has a total of 120 features, the feature numbers are 1 to 120, the first 60 features are the real part data of the impedance data corresponding to the frequency from high to low, and the last 60 features are the imaginary part data of the impedance data corresponding to the frequency from high to low. An example of a battery EIS prediction battery SOH data set is shown in Table 1.

[0093] Table 1 Battery EIS prediction battery SOH data set sample (partial)

[0094] f1 f2 f3 …… f119 f120 SOH 0.3847 0.39156 0.39684 …… 0.29026 0.32795 37.2027 0.38886 0.39343 0.40066 …… 0.28064 0.323 36.2230 0.39038 0.396 0.4025 …… 0.28773 0.32955 35.5893 0.39194 0.39643 0.40406 …… 0.29768 0.33673 35.1081 0.39225 0.3976 0.4054 …… 0.28396 0.32562 34.7673

[0095] Among them, f1 to f120 represent the EIS frequency characteristics, and the SOH here represents the maximum capacity of the battery and also reflects the health of the battery.

[0096] PCA is performed on the 120-dimensional data of the input data to obtain 10 main components, as shown below:

[0097] Table 2 Example of PCA calculation results of battery EIS data (partial)

[0098] P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 -0.8257 0.4979 -0.1901 0.1024 0.0352 0.0084 0.0268 -0.0136 0.0046 0.0073 -0.7667 0.4800 -0.1948 0.1019 0.0277 0.0032 0.0192 0.0015 0.0170 0.0051 -0.7396 0.4718 -0.1953 0.0970 0.0339 0.0046 0.0250 -0.0031 0.0122 -0.0086 -0.7062 0.4585 -0.1982 0.0937 0.0432 0.0042 0.0362 -0.0073 0.0042 0.0071 -0.6895 0.4538 -0.2001 0.0948 0.0319 0.0012 0.0215 0.0034 0.0199 0.0059

[0099] Among them, P1 to P10 represent 10 principal components respectively.

[0100] The mean of the test set is 30.5773, the standard deviation is 4.0661, and the mean of the prediction set is 30.4007, the standard deviation is 4.0186, which shows that the data distribution of the prediction results is basically consistent with the data distribution of the real data.

[0101] The root mean square error between the prediction set and the test set is only 0.2025, indicating that the prediction effect of the model is good.

[0102] The prediction results are as follows: Figure 7 As shown in the figure, the solid line above represents the true value, and the dotted line represents the predicted value. The numbers below are the true value, and the numbers above are the predicted value. The dotted line below represents the error value, and the numbers are the error. From the overlap degree of the broken lines and the error value, it can be seen that the prediction effect is good.

[0103] Example 2

[0104] This embodiment discloses a system for predicting SOH using battery EIS, including a memory and a processor, wherein the memory includes a method program for predicting SOH using battery EIS, and when the method program for predicting SOH using battery EIS is executed by the processor, the steps of the method described in Example 1 are implemented.

[0105] Example 3

[0106] This embodiment discloses a readable storage medium, which includes a method program for predicting SOH using battery EIS. When the method program for predicting SOH using battery EIS is executed by a processor, the steps of a method for predicting SOH using battery EIS described in Example 1 are implemented.

[0107] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0108] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0109] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0110] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0111] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

Claims

1. A method for predicting SOH using battery EIS, characterized in that: The method comprises: Test and collect battery EIS data; Build and train the PCA-CNN-LSTM model; The EIS data is input into the PCA-CNN-LSTM model, the PCA-CNN-LSTM model is used to predict the battery SOH, and the prediction results are output.

2. A method for predicting SOH using battery EIS according to claim 1, characterized in that: The PCA-CNN-LSTM model includes: an input layer, a PCA model, a CNN model, a LSTM model, and a linear regression layer; The input layer is used to input EIS data into the PAC model; The PAC model extracts the principal components of the EIS data to obtain the principal component data, and outputs the principal component data to the CNN model; The CNN model extracts EIS spatial characteristics from the principal component data to obtain EIS spatial features, and outputs the EIS spatial features to the LSTM model; The LSTM model extracts the time series variation law of EIS according to the spatial characteristics of EIS, and outputs the time series variation law of EIS to the linear regression layer; The linear regression layer predicts the battery SOH based on the EIS spatial characteristics and the time series variation law of EIS.

3. A method for predicting SOH using battery EIS according to claim 1 or 2, characterized in that: The collecting of battery EIS data is specifically as follows: Collect battery EIS data at preset frequency intervals; When collecting battery EIS data, record the real and imaginary impedance data corresponding to each frequency EIS data; as well as the target value to be predicted, that is, the battery capacity data.

4. A method for predicting SOH using battery EIS according to claim 3, characterized in that: The PAC model extracts the principal components of the EIS data, specifically: The EIS data is input into the PCA model, and the high-dimensional feature space is transformed into several mutually orthogonal principal component spaces using linear transformation; Among them, the principal components in the principal component space explain part of the variance of the original data; Calculate the percentage of data variance explained by each principal component to the total variance of the original data; If the cumulative explained variance rate of the principal component exceeds the preset threshold, it will be determined as the principal component data.

5. A method for predicting SOH using battery EIS according to claim 1 or 4, characterized in that: The CNN model includes two convolutional layers and one pooling layer. The activation functions of the convolutional layers are all relu functions. The number of filters is set to 128, the kernel size is 1, and the pooling size of the pooling layer is 1.

6. A method for predicting SOH using battery EIS according to claim 5, characterized in that: The LSTM model includes an LSTM layer and a fully connected layer. The number of neurons in the LSTM layer is 128, the number of neurons in the fully connected layer is 64, and the activation function is the relu function.

7. A method for predicting SOH using battery EIS according to claim 6, characterized in that: The activation function of the linear regression layer is a linear function.

8. The method for predicting SOH using battery EIS according to claim 7, characterized in that: The loss function of the PCA-CNN-LSTM model is mean square error, the optimizer is rmsprop, and the evaluation metric is mean absolute error.

9. A system for predicting SOH using battery EIS, characterized in that: The method comprises a memory and a processor, wherein the memory comprises a method program for predicting SOH using battery EIS, and the method program for predicting SOH using battery EIS, when executed by the processor, implements the steps of the method according to any one of claims 1 to 8.

10. A readable storage medium, characterized in that: The readable storage medium includes a method program for predicting SOH using battery EIS. When the method program for predicting SOH using battery EIS is executed by a processor, the steps of a method for predicting SOH using battery EIS as described in any one of claims 1 to 8 are implemented.

Images