Battery SOH estimation method and device
By combining linear regression and deep learning models, battery health factors are extracted and predicted and estimated, the accuracy problem of traditional methods when dealing with the nonlinear complex characteristics of lithium-ion batteries is solved, and higher battery SOH estimation accuracy and adaptability are achieved.
Patent Information
- Application Number
- CN202510100425.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-22
- Publication Date
- 2025-03-25
AI Technical Summary
Traditional battery SOH estimation methods are difficult to achieve high accuracy when processing the nonlinear complex characteristics of lithium-ion batteries, and a single deep learning model is also difficult to improve accuracy when processing non-stationary characteristics.
Combining linear regression and deep learning models, the battery charge and discharge process is obtained, health factors are extracted, and the degradation trend part is predicted based on the linear regression model, the capacity regeneration part is estimated based on the deep learning model, and finally the battery SOH estimate value is obtained through weighted fusion.
It effectively improves the accuracy and adaptability of battery SOH estimation. The deep learning model can adaptively learn the nonlinear characteristics and regeneration effects of the battery, enhances the robustness of the model, and the combined model structure has advantages in processing non-stationary data.
Smart Images
Figure CN119689305A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of battery state of health, and in particular to a battery SOH estimation method and device. Background Art
[0002] With the rapid development of electric vehicles and renewable energy systems, lithium-ion batteries play an important role as key energy storage components. However, lithium-ion batteries will gradually degrade during long-term use, and the speed and characteristics of their capacity decay are affected by many factors, such as temperature, number of charge and discharge cycles, and discharge rate. Accurately estimating the state of health (SOH) of the battery is of great significance for the safe use, life prediction, and maintenance optimization of the battery.
[0003] However, traditional battery SOH estimation methods mainly rely on physical models and statistical models. Due to the nonlinear and complex characteristics of lithium-ion batteries, these methods often show large errors in practical applications. In recent years, the development of deep learning technology has provided new ideas for modeling and prediction of complex nonlinear systems. However, a single deep learning model is difficult to achieve high accuracy when dealing with some non-stationary characteristics in the process of battery capacity degradation. Summary of the invention
[0004] To solve the above problems, the present application discloses a battery SOH estimation method, which effectively improves the accuracy and adaptability of battery SOH estimation by combining linear regression and deep learning models; at the same time, a corresponding battery SOH estimation device is proposed to implement the battery SOH estimation method under different situations.
[0005] The first technical solution adopted in this application is: providing a battery SOH estimation method, comprising the following steps:
[0006] Acquire real-time data of the battery charging and discharging process, and extract health factors based on the real-time data;
[0007] The health factor predicts the degradation trend part based on the linear regression model to obtain degradation trend prediction data; the health factor estimates the capacity regeneration part based on the deep learning model to obtain capacity regeneration estimation data;
[0008] The degradation trend prediction data is combined with the capacity regeneration estimation data to obtain a battery SOH estimation value.
[0009] Wherein, the health factor is decomposed into modal components based on variational modal decomposition; the modal components include degradation trend modal components and capacity regeneration modal components; the degradation trend modal components are input into the linear regression model to obtain degradation trend prediction data; the capacity regeneration modal components are input into the deep learning model to obtain capacity regeneration estimation data.
[0010] Among them, the variational constraint model is:
[0011]
[0012] Among them, u k is the modal component, w k is the center frequency, K is the number of modal components, and δ(t) is the pulse signal.
[0013] The variational constraint model can be used to make predictions based on different battery models.
[0014] Among them, the deep learning model is an improved CNN-transformer deep learning model; a linear layer is used instead of a decoder, the model's ability to capture different features is enhanced based on a multi-head attention mechanism, and the nonlinear feature extraction capability is improved based on a feedforward neural network layer.
[0015] Among them, the deep learning model also includes residual connections and layer normalization to accelerate the convergence of the deep learning model and avoid the gradient disappearance problem.
[0016] The degradation trend prediction data and the capacity regeneration estimation data are weightedly fused to obtain a battery SOH estimation value, and the weighted weight can be adjusted based on different real-time data of the battery charging and discharging process.
[0017] The second technical solution adopted by the present application is: providing a battery SOH estimation device using any of the battery SOH estimation methods described above, comprising:
[0018] Data acquisition module, used to collect real-time data of the battery during charging and discharging;
[0019] Feature extraction module, used to extract health factors related to battery SOH from real-time data;
[0020] A linear regression module to model the degradation trend of the battery and output a trend estimate;
[0021] The deep learning model module uses an improved CNN-Transformer deep learning model to predict the capacity regeneration part of the battery;
[0022] The result fusion module is used to fuse the output results of the linear regression module and the deep learning model module to obtain the final SOH estimation value.
[0023] The third technical solution adopted in the present application is: an electronic device is provided, the electronic device comprising: a memory and a processor coupled to each other, the processor being used to execute program instructions stored in the memory to implement the battery SOH estimation method as described in any one of the above items.
[0024] The fourth technical solution adopted in the present application is: providing a computer-readable storage medium, wherein the computer-readable storage medium stores program data, and the program data can be executed by a processor to implement the battery SOH estimation method as described in any one of the above items.
[0025] Due to the adoption of the above technical solution, the present application has at least one of the following beneficial effects compared with the prior art:
[0026] 1. By combining linear regression and deep learning models, the accuracy and adaptability of battery SOH estimation are effectively improved.
[0027] 2. The deep learning model can adaptively learn the nonlinear characteristics and regeneration effects of the battery, adapt to different usage conditions and environmental changes, and enhance the robustness of the model.
[0028] 3. The device of the present invention adopts a modular design, and each module has a clear function, which is convenient for updating and maintenance of the system and can be flexibly applied to batteries of different types and specifications.
[0029] 4. The combined model structure has advantages in processing non-stationary data and can model the capacity regeneration phenomenon more finely, making the estimation results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0031] in:
[0032] Figure 1 A schematic diagram of a flow chart of an embodiment of a battery SOH estimation method provided in the present application;
[0033] Figure 2 Schematic diagram of modal components obtained by the variational constraint model based on different battery models;
[0034] Figure 3 A schematic diagram of the structure of an embodiment of a deep learning model provided in this application;
[0035] Figure 4A schematic diagram of a battery SOH estimation device according to an embodiment of the present invention;
[0036] Figure 5 This is a schematic diagram of the structure of an embodiment of a computer device of the present application;
[0037] Figure 6 This is a schematic diagram of the structure of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION
[0038] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It will be appreciated that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some but not all structures related to the present application are shown in the drawings. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the art without making creative work are within the scope of protection of the present application.
[0039] The terms "first", "second", etc. in this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.
[0040] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0041] The development of deep learning technology has provided new ideas for the modeling and prediction of complex nonlinear systems. However, a single deep learning model is difficult to achieve high accuracy when dealing with some non-stationary characteristics in the process of battery capacity degradation. To this end, the present application provides an adaptive SOH (battery state of health) estimation method that combines deep learning and linear regression models, using linear regression to capture the degradation trend part, and using a deep learning model to predict the capacity regeneration part, thereby achieving a more accurate SOH estimation; it should be clear that the battery SOH estimation method of the present application can be applied to various types of batteries without any limitation, for example, it can be applied to the SOH estimation of lithium batteries; such as Figure 1 As shown, Figure 1 The flowchart of an embodiment of a battery SOH estimation method provided in this application includes the following steps:
[0042] Step S11: Acquire real-time data of the battery charging and discharging process, including but not limited to charging and discharging current, charging and discharging voltage, and temperature; extract health factors based on real-time data, including but not limited to termination voltage, charging time, charging current curve and other factors. For example, specific health indicators can be determined by analyzing the charging curve, such as the time required to reach a certain voltage during the charging process, the voltage value at the end of charging, etc.
[0043] It should be clarified that the battery SOH estimation method of the present application can be applied to different charging and discharging states of the battery, including different charging methods and discharge rates, etc., without any limitation; for example, in one embodiment, the lithium battery is charged to full charge at a constant current and constant voltage, and discharged at a discharge rate of 2C.
[0044] Step S12: The health factor predicts the degradation trend part based on the linear regression model to obtain the degradation trend prediction data; in one embodiment, the input value of the linear regression model is the charge and discharge current, and the output value is the charge and discharge voltage; the charge and discharge voltage predicted by the linear regression model is compared with the actual value of the charge and discharge voltage, and the degradation trend prediction data is output based on the comparison result; in other embodiments, the input value of the linear regression model may also include charging time, etc., without any limitation.
[0045] The health factor estimates the capacity regeneration part based on the deep learning model to obtain the capacity regeneration estimation data. The deep learning model has adaptive learning ability and can automatically adjust parameters to adapt to different battery characteristics and environmental conditions. By using linear regression model and deep learning model to process the degradation trend and capacity regeneration of the battery respectively, the accuracy of SOH estimation can be significantly improved.
[0046] Step S13: The degradation trend prediction data is combined with the capacity regeneration estimation data to obtain the battery SOH estimation value; the influence ratio of the degradation trend prediction data and the capacity regeneration estimation data depends on the working conditions of the battery and the influence of its internal physical and chemical changes; for example, when the battery is shallowly charged and discharged, or intermittently charged or discharged at a high rate, a short-term capacity recovery phenomenon will occur, which is the so-called "capacity regeneration". At this time, the deep learning model's estimation of the capacity regeneration part will be more important. For batteries that are used continuously for a long time, their overall attenuation trend is more obvious, while the fluctuations in the short term are relatively small. The linear regression model is more important for predicting the degradation trend part.
[0047] In summary, the battery SOH estimation method of this embodiment includes the following steps: obtaining real-time data of the battery charging and discharging process, extracting health factors based on real-time data; predicting the degradation trend part of the health factor based on the linear regression model to obtain degradation trend prediction data; estimating the capacity regeneration part of the health factor based on the deep learning model to obtain capacity regeneration estimation data; combining the degradation trend prediction data with the capacity regeneration estimation data to obtain the battery SOH estimation value. By combining linear regression and deep learning models, the accuracy and adaptability of battery SOH estimation are effectively improved, especially when the capacity regeneration effect is obvious, and the actual health status of the battery can be more truly reflected.
[0048] In one embodiment, the health factor is decomposed into modal components (IMF) based on variational mode decomposition (VMD). Variational mode decomposition is a signal processing technology that can effectively decompose complex signals into their components, which helps to more accurately analyze the health status of the battery. Each modal component represents a different frequency component of the health factor.
[0049] The modal components include degradation trend modal components and capacity regeneration modal components; the degradation trend modal components are input into the linear regression model to obtain degradation trend prediction data; the capacity regeneration modal components are input into the deep learning model to obtain capacity regeneration estimation data.
[0050] In one embodiment, variational mode decomposition decomposes the health factor into four mode vectors IMF1, IMF2, IMF3 and IMF4, where IMF1 represents degradation trend, IMF2 and IMF3 represent capacity regeneration, and IMF4 represents noise. Then, IMF1 is input into a linear regression model to obtain degradation trend prediction data, IMF2 and IMF3 are input into a deep learning model to obtain capacity regeneration estimation data, and IMF4 is removed.
[0051] It should be clarified that the present application does not limit the total number of modal vectors of variational modal decomposition and the proportion of modal vectors input into the linear regression model and the deep learning model; in other embodiments, variational modal decomposition decomposes the health factor into two modal vectors IMF5 and IMF6, IMF5 represents the degradation trend, and IMF6 represents capacity regeneration.
[0052] In one embodiment, the variational constraint model is:
[0053]
[0054] Among them, u k is the modal component, w k is the center frequency, K is the number of modal components, δ(t) is the pulse signal, represents the modal component u k(t) Representation in the frequency domain, where F represents Fourier transform; As the objective function, find a set of modal components u k (t) and the corresponding center frequency w k , in order to minimize the sum of the bandwidths of the combined effects of each modal component and the exponential modulation after Hilbert transform; δ(t) represents the pulse signal, j is the imaginary unit, represents the differential with respect to time t, * represents convolution, and Indicates energy; is a constraint condition, which means that the original signal f(t) can be composed of all modal components u k (t) Total reconstruction.
[0055] The calculation principle of this formula is described in detail below:
[0056] Select the modal component u given by the initial value k (t) and center frequency w k ;
[0057] For each modal component u k (t) and center frequency w k Update alternately until the convergence condition is met;
[0058] For each modal component u k (t), estimate its bandwidth by calculating its Hilbert transform and applying exponential modulation, then performing a differentiation operation. According to the objective function, adjust the modal components and center frequencies to minimize the total bandwidth;
[0059] All modal components u k (t) is used to reconstruct the original signal f(t), ensuring that the decomposed modal components can accurately represent the different characteristic parts of the original signal.
[0060] Variational mode decomposition can effectively decompose complex signals into multiple modal components with clear physical meanings, which can be further used to analyze trend prediction and capacity regeneration estimation in battery SOH estimation methods.
[0061] In one embodiment, the variational constraint model can be used for prediction based on different battery models, and customized model training and optimization based on the characteristics and working conditions of different battery models can significantly improve the accuracy of SOH estimation. This is because different types of batteries have differences in material composition, design structure, etc., which leads to different degradation mechanisms.
[0062] By adjusting VMD parameters and model training strategies, the model can be better adapted to the characteristics of different types of batteries. In this way, even in the face of new battery models or special situations, it can quickly adjust and provide reliable prediction results; Figure 2 As shown, Figure 2 This is a schematic diagram of the modal components of the variational constraint model based on different battery models. The battery models used are CS2_35, CS2_36, CS2_37 and CS2_38 in the CALCE dataset of the University of Maryland.
[0063] like Figure 3 As shown, Figure 3 A schematic diagram of a deep learning model embodiment provided in this application; the deep learning model is an improved CNN-transformer deep learning model; the modal component after VMD processing is input into the improved CNN-Transformer model. The model combines the powerful feature extraction capability of convolutional neural network (CNN) and the advantages of Transformer architecture in processing long sequence data.
[0064] A linear layer is used to replace the decoder. In a standard Transformer structure, there are usually two parts: the encoder and the decoder. In this improved model, a simple linear layer is used to replace the decoder part, which simplifies the model structure and is particularly suitable for smaller battery SOH-related datasets, thereby reducing prediction errors.
[0065] The multi-head attention mechanism is used to enhance the model's ability to capture different features. The multi-head attention mechanism enables the model to perform different attention calculations on different "attention heads", more effectively capture the dependencies between different positions in the input data, and enhance the model's ability to recognize complex patterns.
[0066] The feedforward neural network layer is used to improve the nonlinear feature extraction capability. A feedforward neural network layer is added to each basic module of Transformer. By increasing the nonlinear level, the model's ability to learn and represent complex features is further enhanced.
[0067] By combining the local feature extraction capability of CNN and the advantages of Transformer architecture in processing sequence data, especially using the multi-head attention mechanism to better capture the key information in the input data, the accuracy of battery SOH estimation is significantly improved. Due to the use of an adaptive deep learning model, parameters can be automatically adjusted for different types of batteries and their working environments, improving the versatility and robustness of the model.
[0068] The deep learning model also includes residual connections and layer normalization to accelerate the convergence of the deep learning model and avoid the gradient vanishing problem. After completing the multi-head attention calculation, the input is directly added to the output to form a residual connection. That is, assuming that the original input is X, and the output after processing by the multi-head attention mechanism is Y, the result of the residual connection is X+Y. This approach helps alleviate the gradient vanishing problem in deep networks, allowing information to be more effectively transmitted to deeper levels; the results after the residual connection are layer normalized. Layer normalization makes the data more stable by standardizing the input of each layer (adjusting the mean and variance), thereby accelerating the training process and improving model performance.
[0069] The following is a detailed description of the linear transformation and multi-head attention calculation process of the improved CNN-Transformer model:
[0070] Linear transformation: The core process of the attention mechanism in Transformer is to calculate the attention weights using the query matrix Q and the key matrix K, and then apply it to the value vector matrix V to obtain the entire weight and output.
[0071]
[0072] Where W Q , W K and W V is the linear transformation matrix.
[0073] Multi-head attention calculation: First, the dot product of the query and the key is calculated, and then the attention weight is obtained by soft maximization (softmax). The single-head attention value of the i-th subspace can be expressed as:
[0074]
[0075] In the formula, and are the linear transformation matrices of Q, K and V projected onto the i-th subspace, and h is the number of projections (the number of subspaces). Then the outputs of all heads are concatenated and the final output is obtained through a linear transformation:
[0076] MultiHead(Q,K,V)=Concat(head1,head2,..,head h )W O
[0077] Where W O is a linear transformation matrix. The Contact function transforms multiple groups of d k The attention calculation results in the dimensional space are concatenated.
[0078] In one embodiment, the degradation trend prediction data and the capacity regeneration estimation data are weightedly fused to obtain the battery SOH estimation value, and the weighted weight can be adjusted based on different real-time data of the battery charging and discharging process.
[0079] In actual application, the real-time data of the battery's charging and discharging process is continuously monitored; the weighted weights are dynamically adjusted according to the current working conditions (such as temperature, charging and discharging rate, etc.), and the weights can be automatically optimized through machine learning algorithms, so that the SOH estimation can more accurately reflect the actual situation.
[0080] The degradation trend prediction data and capacity regeneration estimation data are weighted and summed according to the real-time adjusted weights to obtain the final battery SOH estimation value; the estimated SOH value is regularly compared with the actual test results to evaluate the accuracy of the model; the weighted weights and model parameters are further fine-tuned based on the feedback information to improve the prediction accuracy.
[0081] like Figure 4 As shown, Figure 4 A schematic diagram of the framework of an embodiment of a battery SOH estimation device provided in the present application includes a data acquisition module, a feature extraction module, a linear regression module, a deep learning model module and a result fusion module.
[0082] The data acquisition module is used to collect real-time data of the battery during the charging and discharging process;
[0083] The feature extraction module is used to extract health factors related to battery SOH from real-time data;
[0084] The linear regression module is used to model the degradation trend of the battery and output the trend estimate;
[0085] The deep learning model module uses an improved CNN-Transformer deep learning model to predict the capacity regeneration part of the battery;
[0086] The result fusion module is used to fuse the output results of the linear regression module and the deep learning model module to obtain the final SOH estimation value.
[0087] For the above embodiment, the present application provides a computer device, see Figure 5 , Figure 5 This is a schematic diagram of the structure of an embodiment of a computer device of the present application. The computer device includes a memory and a processor, wherein the memory and the processor are coupled to each other, the memory stores program data, and the processor is used to execute the program data to implement the steps of any embodiment of the above-mentioned battery SOH estimation method.
[0088] In this embodiment, the processor may also be referred to as a CPU (Central Processing Unit). The processor may be an integrated circuit chip having signal processing capabilities. The processor may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0089] The method of the above embodiment can be implemented in the form of a computer program, so the present application proposes a computer readable storage medium, see Figure 6 , Figure 6 This is a schematic diagram of the structure of an embodiment of a computer-readable storage medium of the present application. The computer-readable storage medium stores program data that can be run by a processor, and the program data can be executed by the processor to implement the steps of any embodiment of the above-mentioned battery SOH estimation method.
[0090] The computer-readable storage medium of this embodiment may be a medium that can store program data, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or may also be a server storing the program data, which may send the stored program data to other devices for execution, or may also execute the stored program data itself.
[0091] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only illustrative, for example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
[0092] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, 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.
[0093] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0094] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.
Claims
1. A battery SOH estimation method, characterized in that: The steps include: Acquire real-time data of the battery charging and discharging process, and extract health factors based on the real-time data; The health factor predicts the degradation trend part based on the linear regression model to obtain degradation trend prediction data; The health factor estimates the capacity regeneration part based on the deep learning model to obtain capacity regeneration estimation data; The degradation trend prediction data is combined with the capacity regeneration estimation data to obtain a battery SOH estimation value.
2. The battery SOH estimation method according to claim 1, characterized in that: Decomposing the health factor into modal components based on variational modal decomposition; the modal components include degradation trend modal components and capacity regeneration modal components; The degradation trend modal component is input into the linear regression model to obtain degradation trend prediction data; The capacity regeneration modal component is input into the deep learning model to obtain capacity regeneration estimation data.
3. The battery SOH estimation method according to claim 2, characterized in that: The variational constraint model is: Among them, u k is the modal component, w k is the center frequency, K is the number of modal components, and δ(t) is the pulse signal.
4. The battery SOH estimation method according to claim 3, characterized in that: The variational constraint model can be used to make predictions based on different battery models.
5. The battery SOH estimation method according to claim 4, characterized in that: The deep learning model is an improved CNN-transformer deep learning model; a linear layer is used instead of a decoder, the ability of the model to capture different features is enhanced based on a multi-head attention mechanism, and the ability to extract nonlinear features is improved based on a feedforward neural network layer.
6. The battery SOH estimation method according to claim 5, characterized in that: The deep learning model also includes residual connections and layer normalization to accelerate the convergence of the deep learning model and avoid the gradient vanishing problem.
7. The battery SOH estimation method according to claim 6, characterized in that: The degradation trend prediction data and the capacity regeneration estimation data are weightedly fused to obtain a battery SOH estimation value, and the weighted weight can be adjusted based on different real-time data of the battery charging and discharging process.
8. A battery SOH estimation device using the battery SOH estimation method according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, used to collect real-time data of the battery during charging and discharging; Feature extraction module, used to extract health factors related to battery SOH from real-time data; A linear regression module to model the degradation trend of the battery and output a trend estimate; The deep learning model module uses an improved CNN-Transformer deep learning model to predict the capacity regeneration part of the battery; The result fusion module is used to fuse the output results of the linear regression module and the deep learning model module to obtain the final SOH estimation value.
9. An electronic device, characterized in that: The electronic device comprises: a memory and a processor coupled to each other, wherein the processor is used to execute program instructions stored in the memory to implement the battery SOH estimation method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program data, and the program data can be executed by a processor to implement the battery SOH estimation method according to any one of claims 1 to 7.