Battery performance prediction method and system based on impedance spectroscopy analysis

Abstract

The present invention relates to a battery performance prediction method and system based on impedance spectrum analysis. The method comprises the following steps: obtaining the frequency, real impedance part, and imaginary impedance part of impedance data of a battery to be tested to form an input array; defining the types and parameters of components in a battery circuit, generating possible impedance values ​​for each component based on the input array, and storing the values ​​in a one-dimensional array Zp; traversing all possible impedance values ​​of components at each position in the array Zp, comparing the calculated real and imaginary total impedance parts with simulated values ​​and sample values ​​of the measured impedance data, obtaining errors corresponding to the simulated values ​​and sample values, summing the obtained total errors, comparing and updating the total errors with the minimum errors, determining a component parameter combination for the current circuit configuration, generating an array M containing a set number of elements based on the component types and parameter values, and generating an equivalent circuit diagram based on the component types and parameter values ​​in the array M; and utilizing the component types and parameter values ​​in the array M in combination with a trained computer model to implement performance prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of battery performance detection, and in particular to a battery performance prediction method and system based on impedance spectroscopy analysis. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Traditional battery performance prediction methods often rely on complex experimental tests and empirical formulas, and have problems such as low efficiency and difficulty in ensuring accuracy. Although there are some attempts to use data analysis to evaluate battery performance, the results are not ideal. There is no mature technology that can effectively convert the battery's impedance spectrum data into equivalent circuit diagram information and combine it with AI small models for accurate performance prediction. Because this technology involves multiple fields, including but not limited to complex functions, circuits, electrochemistry, programming algorithms, and artificial intelligence. Even if there are methods to predict battery performance using impedance spectrum data, they often lead to large errors due to improper use of impedance spectrum data. Considering that when the same impedance spectrum tester measures the same battery at the same time, the results of two measurements may be different, choosing a suitable method to extract features from impedance spectrum data is the key to improving the practicality of the method. Summary of the Invention

[0004] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a battery performance prediction method and system based on impedance spectroscopy analysis, which uses a computer program to realize the rapid and accurate conversion from impedance spectroscopy data to equivalent circuit diagram information. Combined with a computer model to predict battery performance, it can fully explore the potential patterns in the data. Compared with traditional methods, it has higher prediction accuracy and reliability, and can better meet the needs of battery performance evaluation in practical applications.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A first aspect of the present invention provides a battery performance prediction method based on impedance spectroscopy analysis, comprising the following steps:

[0007] Obtain the impedance data of the battery to be tested, extract the frequency, real part of the impedance, and imaginary part of the impedance data to form an input array;

[0008] Define the component types and parameters in the battery circuit, generate the possible impedance values ​​for each component based on the input array, and save them in a one-dimensional array Zp; loop through all possible impedance values ​​of the component at each position in the array Zp and calculate the real and imaginary parts of the total impedance;

[0009] The calculated real and imaginary parts of the total impedance are compared with the simulated and sampled values ​​of the measured impedance data to obtain the corresponding errors. The total error is obtained by summing them up. The component parameter combination of the current circuit configuration is determined by comparing and updating with the minimum error.

[0010] According to the obtained optimal parameter combination, an array M containing a set number of elements is generated by combining the component types and parameter values. According to the component types and parameter values ​​in the array M, the component types are converted into graphic symbols, and the graphic representation of the components is adjusted according to the parameter values ​​to generate an equivalent circuit diagram.

[0011] The performance parameters to be predicted are quantified digitally, and the performance prediction is achieved by combining the component types and parameter values ​​in the array M with the trained computer model.

[0012] Specifically:

[0013] Obtain the impedance data of the battery to be tested, extract the frequency, real part of the impedance, and imaginary part of the impedance data to form an input array;

[0014] Define the types and parameters of resistance elements, capacitance elements, inductance elements, constant phase angle elements, finite length Warburg impedance elements, and Warburg impedance elements in the battery circuit, generate possible impedance values ​​based on the input array, and save them in the one-dimensional array Zp;

[0015] By nesting, looping through all possible impedance values ​​of each position element in array Zp, the total impedance is calculated using the series and parallel calculation formula of impedance, and the real and imaginary parts of the total impedance are calculated separately;

[0016] Compare the calculated real and imaginary parts of the total impedance with the measured analog and sample values ​​of the impedance data to obtain the corresponding errors with the analog and sample values, and sum them to obtain the total error; if the total error of the current circuit configuration is less than the previously recorded minimum error, update the minimum error value and record the component parameter combination of the current circuit configuration;

[0017] According to the obtained optimal parameter combination, an array M containing a set number of elements is generated by combining the component types and parameter values. According to the component types and parameter values ​​in the array M, the component types are converted into graphic symbols, and the graphic representation of the components is adjusted according to the parameter values ​​to generate an equivalent circuit diagram.

[0018] The performance parameters to be predicted are quantified digitally, and the performance prediction is achieved by combining the component types and parameter values ​​in the array M with the trained computer model.

[0019] Furthermore, the types and parameters of the components in the battery circuit are defined, and the possible impedance values ​​of each component are generated according to the input array and stored in the one-dimensional array Zp. Specifically, the types and parameters of the resistance component, capacitance component, inductance component, constant phase angle component, finite length Warburg impedance component and Warburg impedance component in the battery circuit are defined, and the possible impedance values ​​are generated according to the input array and stored in the one-dimensional array Zp.

[0020] Furthermore, possible impedance values ​​are generated based on the input array and stored in a one-dimensional array Zp. Specifically, the parameter value range of each component is determined, and the possible impedance values ​​of each component are converted into a one-dimensional array. Each element in the array contains a character variable frequency. The value of each element of the array is the impedance value obtained by traversing the parameter value range of the component. All elements contained in the array represent all possible impedance values ​​of this component.

[0021] Furthermore, all possible impedance values ​​of each position element in the array Zp are looped through, and the real and imaginary parts of the total impedance are calculated. Specifically, all possible impedance values ​​of each position element in the array Zp are looped through nesting, the total impedance is calculated using the series-parallel calculation formula of the impedance, and the real and imaginary parts of the total impedance are calculated separately.

[0022] Furthermore, the component parameter combination of the current circuit configuration is determined by comparing and updating with the minimum error; specifically: the calculated real and imaginary parts of the total impedance are compared with the measured analog value and sample value of the impedance data to obtain the corresponding errors with the analog value and sample value, and the total error is obtained by summing them; if the total error of the current circuit configuration is less than the previously recorded minimum error, the minimum error value is updated, and the component parameter combination of the current circuit configuration is recorded.

[0023] Furthermore, the performance parameters that need to be predicted are quantified digitally, and the performance corresponding to each battery is obtained through experiments. The values ​​corresponding to these performances are used as standard answers for model training, and the partial M arrays obtained through measurement and calculation and the performance labels of the batteries are input into the model for training.

[0024] A second aspect of the present invention provides a system for implementing the above method, comprising:

[0025] The data acquisition module is configured to: obtain impedance data of the battery to be tested, extract the frequency, the real part of the impedance, and the imaginary part of the impedance from the impedance data to form an input array;

[0026] The real and imaginary impedance processing module is configured to: define the types and parameters of components in the battery circuit, generate possible impedance values ​​for each component based on the input array, and store them in a one-dimensional array Zp; loop through all possible impedance values ​​of the component at each position in the array Zp, and calculate the real and imaginary parts of the total impedance;

[0027] The parameter combination module is configured to: compare the calculated real and imaginary parts of the total impedance with the measured analog value and sample value of the impedance data, obtain the corresponding errors with the analog value and the sample value, sum the errors to obtain the total error, and determine the component parameter combination of the current circuit configuration by comparing and updating the minimum error;

[0028] an equivalent circuit generation module configured to: generate an array M containing a set number of elements based on the obtained optimal parameter combination and in combination with component types and parameter values; convert the component types into graphic symbols based on the component types and parameter values ​​in the array M; and adjust the graphic representation of the components based on the parameter values ​​to generate an equivalent circuit diagram;

[0029] The performance prediction module is configured to: quantify the performance parameters to be predicted digitally, and use the component types and parameter values ​​in the array M in combination with the trained computer model to achieve performance prediction.

[0030] A third aspect of the present invention provides a computer-readable storage medium.

[0031] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the battery performance prediction method based on impedance spectrum analysis as described above.

[0032] A fourth aspect of the present invention provides a computer device.

[0033] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the battery performance prediction method based on impedance spectrum analysis as described above are implemented.

[0034] Compared with the existing technology, one or more of the above technical solutions have the following beneficial effects:

[0035] By pre-defining the component types and parameters in the battery circuit, the frequency, real part of impedance and imaginary part of impedance in the impedance data are formed into an input array. By looping through all possible impedance values ​​of the component at each position in the array, the real and imaginary parts of the total impedance are calculated. By comparing with the simulated value and sample value of the measured impedance data, the component parameter combination of the current circuit configuration is determined using the minimum error. The component types are then converted into graphic symbols, and the graphical representation of the components is adjusted according to the parameter values ​​to generate an equivalent circuit diagram. This achieves rapid and accurate conversion from impedance spectrum data to equivalent circuit diagram information, and combines computer models to predict battery performance, which can fully explore the potential patterns in the data. Compared with traditional methods, it has higher prediction accuracy and reliability, and can better meet the needs of battery performance evaluation in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0037] Figure 1 is a schematic diagram of a battery performance prediction process based on impedance spectroscopy analysis provided by one or more embodiments of the present invention;

[0038] Figure 2 This is a schematic diagram of component types and component parameter information for converting battery impedance spectrum data into an equivalent circuit diagram, provided by one or more embodiments of the present invention. DETAILED DESCRIPTION

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0040] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0041] Explanation of terms:

[0042] Warburg impedance refers to the impedance caused by the diffusion process in an electrochemical system. It is used to describe the obstruction of current transmission due to the diffusion process of reactants or products in redox reactions, especially at the interface between electrolyte and electrode.

[0043] Example 1:

[0044] like Figure 1 As shown, the battery performance prediction method based on impedance spectroscopy analysis includes the following steps:

[0045] 1. Data acquisition: Use professional impedance spectrum measurement equipment to test different types of batteries, obtain their impedance spectrum data (three data arrays: frequency, real impedance part, and imaginary impedance part), and represent these data in an array form for subsequent computer program reading and processing.

[0046] 2. Computer program implementation: Use a programming language (Python as an example) to write a program to implement the algorithm for converting impedance spectrum data into equivalent circuit component information, and organize this information into a format suitable for AI model input.

[0047] 3. Construction and training of small AI models: Select an appropriate neural network architecture (such as a multilayer perceptron, convolutional neural network, etc.) and determine the number of nodes and connection methods for the input layer, hidden layer, and output layer. Utilize a large amount of sample data with known battery performance (including equivalent circuit diagram information converted from corresponding impedance spectrum data and actual measured battery performance indicators) to train the AI ​​model. By continuously adjusting the model's weights and biases, optimize the model's predictive performance so that it can accurately predict battery performance based on the input equivalent circuit diagram information.

[0048] 4. Performance prediction application: In actual applications, the impedance spectrum data of the battery to be predicted is input into the developed computer program to obtain the equivalent circuit component information. This information is then input into the trained AI model to quickly obtain the performance prediction results of the battery, providing a decision-making basis for battery use and management.

[0049] Computer programs enable rapid and accurate conversion of impedance spectrum data into equivalent circuit diagram information, providing key basic data for subsequent performance predictions and greatly improving data processing efficiency and accuracy.

[0050] Using small AI models to predict battery performance can fully tap into the potential patterns in the data. Compared with traditional methods, it has higher prediction accuracy and reliability, can better meet the needs of battery performance evaluation in practical applications, provide strong support for battery research and development, production and use, and help reduce costs and improve battery quality and safety.

[0051] In this embodiment, the impedance spectrum data of the battery is tested and converted into the component types (such as resistance, capacitance, inductance, etc.) and component parameter information (including the specific values ​​of each component) of the equivalent circuit diagram through a specific algorithm. The type and parameter information are combined with the performance of the battery and input into the AI ​​small model for training and adjustment.

[0052] The design idea of ​​the program is as follows Figure 1 As shown, specifically:

[0053] 1. The equivalent circuit is assumed to have a 5×5+1 structure by default. In actual applications, it can also be any other structure such as 4×3+2, 10×100+1000, etc. This embodiment uses 5×5+1 as an example in the first half of the entire algorithm to illustrate that the overall process of the algorithm can improve the accuracy of the simulation. In actual use, the problem of excessive computational complexity is generally considered.

[0054] Considering that the structure of some equivalent circuits is equivalent to turning a certain component into a parallel circuit, the same traversal method can be used. Figure 2 In the figure, each hexagon represents an element, and six types of elements can be used to fit the impedance spectrum:

[0055] (1) Resistance Z = R;

[0056] (2)

[0057] (3) Inductance Z = jωL;

[0058] (4) Constant phase angle element (n is the phase angle, Q is the component parameter);

[0059] (5) Finite-length Warburg impedance

[0060] (6) Warburg impedance

[0061] Constant phase angle element Let n = 0, then Z is a positive real constant that can replace the effect of resistance. Let n = 1, it can replace the effect of capacitance. In this case, Q is equivalent to the capacitance value C. Therefore, the types of components in the model can be divided into four types (3)-(6). In order to simulate the situation where the number of branches of the equivalent circuit is less than 5 or the number of components on a branch is less than 5, the following is introduced:

[0062] (7) R = 0Ω;

[0063] (8) R = infΩ (inf is infinity in programming languages ​​such as Python);

[0064] R = 0Ω is equivalent to the absence of an element in the equivalent circuit (short circuit), which can be used to indicate that the number of elements in a branch is less than 5. R = infΩ is equivalent to an open circuit, which can be used to indicate that the total number of branches is less than 5.

[0065] In summary, when Figure 2 When each of the above elements is selected from element (3)-(8), Figure 2The total impedance of the circuit can represent various situations where the number of equivalent circuit branches ranges from 0 to 5 (0 is an open circuit), and the number of components on a branch can be controlled from 0 to 5 through the number of (7).

[0066] 2. Each element parameter takes a discrete value range, and the possible values ​​of the impedance of various elements are converted into a one-dimensional array (each element contains a character variable frequency, and the frequency is set to pinlv), where the value of each element of the array is the impedance value obtained by traversing the parameter value range of the element (for (4), there are two element parameters: phase angle n and element parameter Q, then nested loops are used to traverse all (n, Q) parameter pairs). Therefore, all elements contained in the array represent all possible impedance values ​​of this type of element. Then merge the corresponding arrays of various elements to obtain Zp. All elements of this array Zp represent all possible impedances of the element at a position. Let the number of elements in Zp be the "iteration number". The multiple sets of [frequency, real impedance, imaginary impedance] data of the battery obtained through measurement are divided into three arrays, namely the frequency array w, the real impedance array real, and the imaginary impedance array ima g.

[0067] 3. Then, a nested loop is performed. Because the circuit has 26 components, the loop is nested 26 times. At each level, all values ​​of Zp are iterated over, thus iterating over all possible impedance values ​​for each component position. Therefore, there are 26 possible iterations.

[0068] 4. When the element at each position takes a certain element value in Zp, use the series-parallel calculation formula of impedance to calculate the total impedance (the series-parallel formula is as follows: The total impedance of the series circuit is the sum of the impedances of each series element: Z 串 =Z1+Z2+…, the total impedance of the parallel circuit is the reciprocal of the sum of the reciprocals of the impedances of each branch: For each of the five branches, the impedances of the five components on the branch are added together to obtain the total impedance of each branch Z1, Z2, Z3, Z4, and Z5. Then the impedances of the five branches are calculated according to the parallel formula to obtain the impedance of the entire parallel circuit. Then add the total impedance of this parallel circuit to the impedance of the main circuit element to get Z 总 =Z 干 +Z 并 ).

[0069] Use the real and imaginary functions in the programming language to calculate the real and imaginary arrays of the total impedance, imag. Then, by replacing the symbolic variable pinlv, which represents the frequencies included in the real and imaginary parts of the total impedance, with the frequency array w, we obtain two arrays of the same length as w (also the same length as the real and imaginary arrays). Each element of these two arrays represents the real and imaginary parts of the total impedance of the circuit at a specific frequency.

[0070] The total impedance real and total impedance imag are functions of frequency w. Therefore, we can use the least squares method to calculate the errors between the total impedance real (simulated value) and real (sample value), and between the total impedance imag (simulated value) and imag (sample value), and then sum them to obtain the total error. Whenever the total error is smaller than the minimum total error, the total error value is assigned to sum_value, and the optimal parameter combination best_choices is replaced by the subscript of the Zp element obtained at that time for each position element.

[0071] Since the calculation of the 5×5+1 equivalent circuit is too large, this embodiment is changed to 2×1+1 in actual use (two branches, each with one component, and one component on the main circuit). Nesting three times is sufficient.

[0072] 5. Process the obtained optimal parameter combination, the best choices array. Each element of the best choices array corresponds to a component type and a component parameter (or parameter pair). In the "2×1+1" circuit, the best choices array is an array with three elements. Since each element x in the best choices array corresponds to two variables, the element type a and the component parameter b, it becomes an array M with six elements after processing. The processing rules are:

[0073] 【1】When the element x corresponds to p in the element types (3)-(8) ((3)-(8) correspond to natural numbers 0-5, respectively) and the element parameter q, the element type a takes the value p (p ranges from 0 to 5), and the element parameter b takes the value q. (For constant elements such as (7) and (8), the element parameter q is set to a constant, such as 0. For elements with two element parameters such as (4), a one-to-one correspondence can be established between each parameter pair (n, Q) and the value of q. For other elements with only one parameter, q is equal to its single parameter.)

[0074] 【2】Through the three x elements of best choices, we get three [a, b] arrays, and merge them to get an array M containing six elements.

[0075] 6. Quantify the battery performance that needs to be predicted with numbers. For example, if batteries are divided into two categories, "good" and "bad", "good" can be set to 1 and "bad" to 0. The corresponding performance of each battery is obtained through experiments, and the values ​​corresponding to these performances are used as the standard answers (i.e., "labels" and "target values") for small model training. A portion of the M array obtained through measurement and calculation and the battery performance labels are input into the AI ​​small model training (the other portion of the M array and labels are used to test the model accuracy). Using the Python language's pytorch library as an example, a convolutional neural network is constructed. After training, the other portion of the M array and labels are used to verify the accuracy. The number of neuron layers, types, learning rate, etc. of the model are adjusted based on the accuracy.

[0076] 7. Once the training and adjustment work is completed, simply input the impedance spectrum data of the battery to be tested, process it through the aforementioned equivalent circuit algorithm, and then input it into the AI ​​small model to predict the performance.

[0077] The above method uses a computer program to achieve rapid and accurate conversion from impedance spectrum data to equivalent circuit diagram information, providing key basic data for subsequent performance prediction and greatly improving data processing efficiency and accuracy.

[0078] Using small AI models to predict battery performance can fully tap into the potential patterns in the data. Compared with traditional methods, it has higher prediction accuracy and reliability, can better meet the needs of battery performance evaluation in practical applications, provide strong support for battery research and development, production and use, and help reduce costs and improve battery quality and safety.

[0079] Example 2:

[0080] A system for implementing the above method includes:

[0081] The data acquisition module is configured to: obtain impedance data of the battery to be tested, extract the frequency, the real part of the impedance, and the imaginary part of the impedance from the impedance data to form an input array;

[0082] The real and imaginary impedance processing module is configured to: define the types and parameters of components in the battery circuit, generate possible impedance values ​​for each component based on the input array, and store them in a one-dimensional array Zp; loop through all possible impedance values ​​of the component at each position in the array Zp, and calculate the real and imaginary parts of the total impedance;

[0083] The parameter combination module is configured to: compare the calculated real and imaginary parts of the total impedance with the measured analog value and sample value of the impedance data, obtain the corresponding errors with the analog value and the sample value, sum the errors to obtain the total error, and determine the component parameter combination of the current circuit configuration by comparing and updating the minimum error;

[0084] an equivalent circuit generation module configured to: generate an array M containing a set number of elements based on the obtained optimal parameter combination and in combination with component types and parameter values; convert the component types into graphic symbols based on the component types and parameter values ​​in the array M; and adjust the graphic representation of the components based on the parameter values ​​to generate an equivalent circuit diagram;

[0085] The performance prediction module is configured to: quantify the performance parameters to be predicted digitally, and use the component types and parameter values ​​in the array M in combination with the trained computer model to achieve performance prediction.

[0086] Example 3:

[0087] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps of the battery performance prediction method based on impedance spectrum analysis as described in the first embodiment above are implemented.

[0088] Example 4:

[0089] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the battery performance prediction method based on impedance spectroscopy analysis as described in the first embodiment are implemented.

[0090] The steps or modules involved in Examples 2 to 4 above correspond to those in Example 1. For detailed implementations, please refer to the relevant description of Example 1. The term "computer-readable storage medium" should be understood to mean a single medium or multiple media that includes one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry an instruction set for execution by a processor and cause the processor to perform any method of the present invention.

[0091] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A battery performance prediction method based on impedance spectroscopy analysis, characterized in that: The following steps are involved: Obtain the impedance data of the battery to be tested, extract the frequency, real part of the impedance, and imaginary part of the impedance data to form an input array; Define the types and parameters of components in the battery circuit, generate possible impedance values ​​for each component based on the input array, and save them in a one-dimensional array Zp; specifically: define the types and parameters of resistance components, capacitance components, inductance components, constant phase angle components, finite length Warburg impedance components, and Warburg impedance components in the battery circuit, generate possible impedance values ​​based on the input array, and save them in a one-dimensional array Zp; determine the parameter value range of each component, and convert the possible impedance values ​​of each component into a one-dimensional array Zp. Each element in the array Zp contains a character variable frequency. The value of each element in the array Zp is the impedance value obtained by traversing the parameter value range of the component. All elements contained in the array Zp represent all possible impedance values ​​of this type of component. The real and imaginary impedances in the input array are used as the sample real array real and the sample imaginary array imag; all possible impedance values ​​of the components at each position in the array Zp are looped through to determine the symbolic expression of the total impedance of the circuit. By replacing the symbolic variable pinlv with the frequency array w, the simulated real array real and the simulated imaginary array imag of the total impedance of the circuit corresponding to the frequency w are obtained; the errors of the simulated real array real and the sample real array real, and the simulated imaginary array imag and the sample imaginary array imag are calculated using the least squares method, and the total error is obtained by summing them. The optimal parameter combination of the current circuit configuration components is determined by comparing and updating with the minimum error; According to the obtained optimal parameter combination, an array M containing a set number of elements is generated by combining the component types and parameter values. According to the component types and parameter values ​​in the array M, the component types are converted into graphic symbols, and the graphic representation of the components is adjusted according to the parameter values ​​to generate an equivalent circuit diagram. The performance parameters to be predicted are quantified digitally, and the performance prediction is achieved by combining the component types and parameter values ​​in the array M with the trained computer model. Specifically, a part of the M array obtained through measurement and calculation and the performance label of the battery are input into the computer model for training. The computer model is the pytorch library of the Python language. After training, the accuracy is verified using another part of the M array and label, and the parameters of the model are adjusted according to the accuracy.

2. The battery performance prediction method based on impedance spectroscopy analysis according to claim 1, characterized in that: The total error is compared with the minimum error and updated as follows: if the total error of the current circuit configuration is less than the previously recorded minimum error, the minimum error value is updated and the component parameter combination of the current circuit configuration is recorded.

3. The battery performance prediction method based on impedance spectroscopy analysis according to claim 1, characterized in that: The performance parameters that need to be predicted are quantified digitally, and the performance corresponding to each battery is obtained through experiments. The values ​​corresponding to these performances are used as the standard answers for model training. The partial M array obtained through measurement and calculation and the performance labels of the batteries are input into the model for training.

4. The battery performance prediction method based on impedance spectroscopy analysis according to claim 3, characterized in that: By selecting a neural network architecture, the number of nodes and the connection method of the input layer, hidden layer, and output layer are determined; the model is trained using the equivalent circuit diagram information converted from sample data of known battery performance and the corresponding impedance spectrum data, as well as the actual measured battery performance indicators. By continuously adjusting the model's weights and biases, the model's predictive performance is optimized to obtain a trained computer model.

5. A battery performance prediction system based on impedance spectroscopy analysis, used to implement the method according to any one of claims 1 to 4, characterized in that: include: The data acquisition module is configured to: obtain impedance data of the battery to be tested, extract the frequency, the real part of the impedance, and the imaginary part of the impedance from the impedance data to form an input array; The real and imaginary impedance processing module is configured to: define the types and parameters of components in the battery circuit, generate possible impedance values ​​for each component based on the input array, and store them in a one-dimensional array Zp; loop through all possible impedance values ​​of the component at each position in the array Zp, and calculate the real and imaginary parts of the total impedance; The parameter combination module is configured to compare the calculated real and imaginary parts of the total impedance with the measured impedance data sample values ​​to obtain an error between the simulated value and the sample value, sum the error to obtain a total error, and determine the component parameter combination for the current circuit configuration by comparing and updating the error with the minimum error; an equivalent circuit generation module configured to: generate an array M containing a set number of elements based on the obtained optimal parameter combination and in combination with component types and parameter values; convert the component types into graphic symbols based on the component types and parameter values ​​in the array M; and adjust the graphic representation of the components based on the parameter values ​​to generate an equivalent circuit diagram; The performance prediction module is configured to: quantify the performance parameters to be predicted digitally, and use the component types and parameter values ​​in the array M in combination with the trained computer model to achieve performance prediction.

6. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the steps of the battery performance prediction method based on impedance spectroscopy analysis according to any one of claims 1 to 4 are implemented.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the battery performance prediction method based on impedance spectroscopy analysis according to any one of claims 1 to 4 are implemented.

Images