Mechanism and data fused motor train unit battery aging evaluation method and system

Abstract

The invention provides a motor train unit battery aging evaluation method and system based on mechanism and data fusion, and relates to the technical field of battery health management, and the method comprises the steps: extracting and processing multiple physical characteristic parameters of a battery in experimental data, and obtaining a processed first structured data parameter and a processed first unstructured data parameter; combining an Arrhenius equation, a quasi-two-dimensional model and an equivalent circuit model to construct a high-speed rail motor train unit battery aging mechanism model with electric-thermal-mechanical multi-physical characteristics, and predicting a first aging state of the battery; establishing a mapping relationship between battery performance and multiple physical characteristics to obtain a second aging state of prediction output; and taking the first aging state and the second aging state as two independent evidence bodies, dynamically adjusting a Dempster combination rule weight based on the conflict factor, and further determining a final aging state through a maximum confidence criterion. According to the method, the aging state prediction precision and reliability are improved, and technical support is provided for service life evaluation and maintenance strategy optimization of the high-speed rail motor train unit battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery health management, and more particularly, to a method and system for evaluating the aging of EMU batteries by integrating mechanism and data. Background Art

[0002] With the development of the intelligence and greenness of high-speed EMUs, the reliability of their on-vehicle power battery systems has become a core factor affecting operation safety. The accurate characterization of the battery aging state is a key technical challenge for optimizing battery life management and preventing sudden failures. However, existing methods still have limitations in aging assessment under complex working conditions. Among them, traditional electrochemical models (such as the single-particle model) are difficult to quantify the impact of multi-physical field coupling on aging, and rely on simplified laboratory conditions, unable to adapt to the actual complex working conditions of long-term floating charge and impact load of EMUs. Pure data-driven methods are limited by the high-dimensional noise and feature redundancy of battery operation data, resulting in poor model generalization ability and a significant increase in prediction errors under extreme temperatures or mechanical vibrations.

[0003] Existing technologies mostly focus on the monitoring of battery electrical performance parameters, ignoring the relevance between mechanical thermodynamics characteristics such as temperature field distribution and stress-strain of electrode materials and aging, resulting in insufficient early aging warning ability. Existing fusion methods do not solve the dynamic conflict problem of the output confidence of mechanism models and data models, and lack a collaborative modeling framework for experimental data and operation data, resulting in limited interpretability and accuracy of the fusion results. Traditional monitoring systems rely on offline laboratory test data and do not integrate edge computing and real-time feedback mechanisms, making it difficult to meet the high-frequency and low-latency online health management requirements of high-speed EMUs. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for evaluating the aging of EMU batteries by integrating mechanism and data to improve the above problems. To achieve the above purpose, the technical solutions adopted by the present invention are as follows: In the first aspect, the present application provides a method for evaluating the aging of EMU batteries by integrating mechanism and data, including: Through the battery condition aging experiment of high-speed EMUs, simulate the working conditions of different temperatures, different loads, and different charge and discharge modes in actual operation, and conduct electrochemical impedance spectroscopy tests and open-circuit voltage measurements on the batteries, record the experimental data during the battery aging process, extract and process the multi-physical characteristic parameters of the batteries in the experimental data to obtain the processed first structured data parameters and first unstructured data parameters; Physically decompose the first unstructured data parameter into a second structured data parameter, fuse the first structured data parameter and the second structured data parameter, and construct an electro-thermal-mechanical multi-physical characteristic high-speed EMU battery aging mechanism model in combination with the Arrhenius equation, quasi-two-dimensional model and equivalent circuit model. Solve the high-speed EMU battery aging mechanism model by numerical simulation method, and then predict the first aging state of the battery; Adopt the locally linear embedding method to decompose and extract the key information of the multi-physical characteristic parameters, use the key information as a sequence to input into the long short-term memory network and self-attention mechanism, establish the mapping relationship between battery performance and multi-physical characteristics, generate a deep learning model of battery aging big data and train it to obtain the predicted output of the second aging state, where the key information includes key features, local structure information and key features after dimensionality reduction; Take the first aging state and the second aging state as two independent evidence bodies, and dynamically adjust the weights of the Dempster combination rule based on the conflict factor to complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

[0005] Preferably, the multi-physical characteristic parameters of the battery in the experimental data are extracted and processed to obtain the processed first structured data parameter and the first unstructured data parameter, where the processing process of the processed first structured data parameter includes: Perform rank transformation on the battery performance parameters and multi-physical characteristic parameters obtained from the high-speed EMU battery condition aging experiment respectively, including: sorting the values of each variable from small to large, and assigning a rank to each value; if there are the same values, assign the average rank of the same values that appear in the variable value sorting process; Divide the parameters after rank transformation into a first rank and a second rank, where the first rank is the rank of the battery performance parameter, and the second rank is the rank of the multi-physical characteristic parameter; Based on the first rank and the second rank, calculate the Spearman correlation coefficient between each pair of performance parameters and multi-physical characteristic parameters, and screen out the multi-physical characteristic parameters greater than 0.8, and record them as the processed first structured data parameters.

[0006] Preferably, the multi-physical characteristic parameters of the battery in the experimental data are extracted and processed to obtain the processed first structured data parameter and the first unstructured data parameter, where the processing process of the processed first unstructured data parameter includes: Obtain the Nyquist diagram in the electrochemical impedance spectroscopy test, perform differential geometry processing on the Nyquist diagram, and extract the set of characteristic frequencies of the curvature mutation points; Construct a dynamic weighted tensor based on the set of characteristic frequencies, decompose the open-circuit voltage curve into time-domain differential intrinsic mode functions, and denote the time-domain differential intrinsic mode functions as the processed first unstructured data parameters.

[0007] Preferably, the physical decomposition of the first unstructured data parameters is transformed into second structured data parameters, and the first structured data parameters and the second structured data parameters are fused, including: Calculate the mean, variance, energy, and time-frequency characteristics of each mode in the first unstructured data parameters, and correlate the battery health state through the aging sensitivity weighting coefficient to generate the second structured data parameters reflecting the microscopic aging mechanism; Use a multi-modal fusion algorithm to fuse the first structured data parameters and the second structured data parameters.

[0008] Preferably, the local linear embedding method is used to decompose and extract the key information of the multi-physical characteristic parameters, including: Based on the data of the multi-physical characteristic parameters, determine the neighborhood size of each data point through the Euclidean distance to obtain the neighborhood information; Combined with the neighborhood information, find the r nearest neighbor points for each data point and calculate the corresponding reconstruction weights; Use the calculated reconstruction weights to construct the local neighborhood matrix of each data point, where each row represents the difference between a neighboring point and the data point; Based on the local neighborhood matrix, extract the low-dimensional embedding result by solving the least squares problem and matrix eigenvalue and eigenvector analysis, and denote the low-dimensional embedding result as the key information.

[0009] Preferably, the first aging state and the second aging state are used as two independent evidence bodies, and the weights of the Dempster combination rule are dynamically adjusted based on the conflict factor to complete the joint confidence assignment of the two independent evidence bodies, and then the final aging state is determined through the maximum confidence criterion, including: Take the first aging state and the second aging state as independent evidence bodies respectively, and let the confidence of the first aging state be and the confidence of the second aging state be ; Use the Dempster combination rule to calculate the conflict factor and dynamically adjust the weights of the Dempster combination rule based on the size of the conflict factor; According to the adjusted Dempster combination rule, calculate the joint confidence of each aging state; Arrange the joint confidence of each aging state from high to low, and select the state with the highest confidence as the final aging state for output.

[0010] In a second aspect, the present application also provides a CRH battery aging assessment system integrating mechanism and data, including: An extraction and processing module: used to simulate the working conditions of different temperatures, different loads, and different charge and discharge modes during actual operation through the CRH battery condition aging experiment, perform electrochemical impedance spectroscopy testing and open-circuit voltage measurement on the battery, record the experimental data during the battery aging process, extract and process the multi-physical characteristic parameters of the battery in the experimental data to obtain the first structured data parameter and the first unstructured data parameter after processing; A first prediction module: used to physically decompose and transform the first unstructured data parameter into a second structured data parameter, fuse the first structured data parameter and the second structured data parameter, and combine the Arrhenius equation, the quasi-two-dimensional model, and the equivalent circuit model to construct a CRH battery aging mechanism model with multi-physical characteristics of electricity-thermal-mechanics, solve the CRH battery aging mechanism model using numerical simulation methods, and then predict the first aging state of the battery; A second prediction module: used to decompose and extract the key information of the multi-physical characteristic parameters by using the locally linear embedding method, input the key information as a sequence into the long short-term memory network and the self-attention mechanism, establish the mapping relationship between the battery performance and the multi-physical characteristics, generate and train a deep learning model of the battery aging big data, and obtain the second aging state of the predicted output, where the key information includes key features, local structure information, and key features after dimensionality reduction; A determination and evaluation module: used to take the first aging state and the second aging state as two independent evidence bodies, dynamically adjust the weights of the Dempster combination rule based on the conflict factor, complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

[0011] In a third aspect, the present application also provides a CRH battery aging assessment device integrating mechanism and data, including: A memory, used to store a computer program; A processor, used to implement the steps of the CRH battery aging assessment method integrating mechanism and data when executing the computer program.

[0012] In a fourth aspect, the present application also provides a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned CRH battery aging assessment method integrating mechanism and data are implemented.

[0013] The beneficial effects of the present invention are: The present invention extracts electro-thermal-mechanical multi-physical characteristic information from the source of experimental data, studies the mapping relationship between the battery performance of high-speed EMUs and electro-thermal-mechanical multi-physical characteristics under actual complex working conditions such as long-term floating charge and impact load, calibrates the performance of the EMU battery through mechanism and data models, and realizes the dynamic monitoring of the battery performance of high-speed EMUs and the characterization of aging performance.

[0014] Other features and advantages of the present invention will be described in the following specification, and in part will be obvious from the specification, or can be understood by implementing the embodiments of the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the written specification, claims, and drawings. Brief Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as a limitation of the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a schematic flow chart of the method for evaluating the aging of EMU batteries by integrating mechanism and data described in the embodiments of the present invention; Figure 2 It is a schematic structural diagram of the system for evaluating the aging of EMU batteries by integrating mechanism and data described in the embodiments of the present invention; Figure 3 It is a schematic structural diagram of the device for evaluating the aging of EMU batteries by integrating mechanism and data described in the embodiments of the present invention.

[0017] In the figure: 701, extraction and processing module; 702, first prediction module; 703, second prediction module; 704, determination and evaluation module; 800, device for evaluating the aging of EMU batteries by integrating mechanism and data; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. Detailed Embodiments

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. Components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0019] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present invention, terms such as "first" and "second" are only used for differential description and cannot be construed as indicating or implying relative importance.

[0020] Embodiment 1:

[0021] This embodiment provides a method for evaluating the aging of EMU batteries by integrating mechanism and data.

[0022] See Figure 1 , which shows that this method includes step S100, step S200, step S300, and step S400.

[0023] S100. Through the aging experiment of the high-speed EMU battery under complex working conditions, simulate the working conditions of different temperatures, different loads, and different charge-discharge modes during actual operation, and conduct electrochemical impedance spectroscopy tests and open-circuit voltage measurements on the battery. Record the experimental data during the battery aging process, extract and process the multi-physical characteristic parameters of the battery in the experimental data to obtain the processed first structured data parameters and first unstructured data parameters.

[0024] It should be noted that the aging experiment of the high-speed EMU battery under complex working conditions also includes performance tests on the battery at different cycle numbers to obtain the aging trend data during the long-term use of the battery, and the physical characteristics include at least three of the battery voltage-current impedance, electrochemical impedance spectroscopy characteristic frequency, charge-discharge curve differential parameters, battery surface temperature field distribution and temperature rise time, and electrode material stress-strain parameters.

[0025] It can be understood that in this step S100, it includes S101, S102, and S103, and the processing process of the processed first structured data parameters includes: S101. Perform rank transformation on the battery performance parameters and multi - physical characteristic parameters obtained from the high - speed EMU battery working condition aging experiment respectively, including: sorting the values of each variable from small to large, and assigning a rank to each value; if there are the same values, assign the average rank of the same values that appear in the sorting process of variable values. S102. Divide the parameters after rank transformation into the first rank and the second rank, where the first rank is the rank of the battery performance parameters, and the second rank is the rank of the multi - physical characteristic parameters. S103. Based on the first rank and the second rank, calculate the Spearman correlation coefficient between each pair of performance parameters and multi - physical characteristic parameters, and screen out the multi - physical characteristic parameters greater than 0.8, which are denoted as the processed first structured data parameters. The calculation formula of the Spearman correlation coefficient is:

[0026] In the formula, is the correlation coefficient, is the battery performance parameter without rank transformation, is the multi - physical characteristic parameter without rank transformation, is the first rank, is the second rank, is the sample size, is the number of battery performance parameters, is the number of physical characteristic parameters.

[0027] It should be noted that through the high - speed EMU battery complex working condition aging experiment, different working conditions such as different temperatures, different loads, and different charge - discharge modes in actual operation are simulated to conduct long - term aging tests on the battery. During the experiment, the performance data of the battery at different cycle numbers are recorded, including but not limited to the following multi - physical characteristic parameters: battery voltage - current impedance, battery surface temperature field distribution and temperature rise time, electrode material stress - strain parameters.

[0028] Data recording and pre - processing Record and organize the above experimental data to form a database, and pre - process the data to ensure the accuracy and consistency of the data, including removing outliers and normalization processing.

[0029] Remove the outliers in the data using the empirical rule, where The calculation formula is:

[0030] Among them, is the standard deviation of the same - group data, is the voltage in the data record, is the voltage in the data record, is the sample size.

[0031] It should be noted that It can also be any one of the original sequence data of current, impedance, temperature rise time, and stress-strain parameters of the electrode material, or any one of the average values of the original sequence data of current, impedance, temperature rise time, and stress-strain parameters of the electrode material.

[0032] The normalization process of the data is carried out by using the maximum-minimum method, and its calculation formula is:

[0033] In the formula, refers to any one of the time series data of voltage, current, impedance, temperature rise time, and stress-strain parameters of the electrode material before normalization after removing outliers, refers to any one of the time series data of voltage, current, impedance, temperature rise time, and stress-strain parameters of the electrode material after normalization after removing outliers, refers to any one of the maximum values of the original sequence data of voltage, current, impedance, temperature rise time, and stress-strain parameters in the data record after removing outliers, refers to any one of the minimum values of the original sequence data of voltage, current, impedance, temperature rise time, and stress-strain parameters in the data record after removing outliers.

[0034] It should be noted that in this step S100, S104 and S105 are also included, and the processing process of the processed first unstructured data parameter includes: S104. Obtain the Nyquist diagram in the electrochemical impedance spectroscopy test, perform differential geometry processing on the Nyquist diagram, and extract the characteristic frequency set of the curvature mutation points. The calculation formula is as follows:

[0035] In the formula, is the angle between the tangent direction and the real axis at the corresponding frequency point, is the characteristic frequency set extracted from the Nyquist diagram of the electrochemical impedance spectroscopy, is the characteristic frequency corresponding to the angle ; S105. Construct a dynamic weighted tensor based on the characteristic frequency set, decompose the open-circuit voltage curve into time-domain differential intrinsic mode functions, and record the time-domain differential intrinsic mode functions as the processed first unstructured data parameter. The calculation formula is as follows:

[0036] In the formula, is the dynamic weighted tensor, For the characteristic frequency corresponding to the angle For the charge-discharge cycle For the charge-discharge cycle For the angle The charge transfer resistance at For the angle The double-layer capacitance at

[0037] The time-domain differential intrinsic mode function is obtained by physically decomposing the unstructured voltage curve into structured characteristic parameters with clear aging correlations, which together with the operating condition parameters screened by Spearman form the "processed multi-physical characteristic parameters". This mechanism and data dual-driven fusion method significantly improve the accuracy and interpretability of the evaluation of the aging state of EMU batteries.

[0038] It should be noted that the experimental data for processing by extracting the multi-physical characteristic parameters of the battery in the experimental data include the following two categories: Structured data: directly quantifiable parameters such as voltage, current, temperature, etc. (used for rank transformation and Spearman analysis); Unstructured data: waveform data such as the Nyquist diagram of electrochemical impedance spectroscopy (EIS), open circuit voltage (OCV) curve, etc. (used for differential geometry processing, dynamic weighted tensor construction, etc.). It can be understood that structured data processing (Spearman) solves the problem of initial feature screening and reduces the dimension; unstructured data processing (differential geometry, modal decomposition) solves the problem of physical interpretability and enhances the feature significance. The differential geometry analysis of the Nyquist diagram captures the impedance spectrum distortion caused by the growth of the SEI film through the curvature mutation points, but the traditional method only relies on frequency domain analysis.

[0039] S200. Physically decompose the first unstructured data parameter into a second structured data parameter, fuse the first structured data parameter and the second structured data parameter, and combine the Arrhenius equation, the quasi-two-dimensional model and the equivalent circuit model to construct an electro-thermal-mechanical multi-physical characteristic aging mechanism model of the high-speed EMU battery, and use the numerical simulation method to solve the aging mechanism model of the high-speed EMU battery, and then predict the first aging state of the battery.

[0040] It can be understood that in this step S200, it includes S201 and S202, where: S201. Calculate the mean, variance, energy and time-frequency characteristics of each mode in the first unstructured data parameter, and correlate the battery health state through the aging sensitivity weighting coefficient to generate a second structured data parameter reflecting the microscopic aging mechanism; S202. Use the multi-modal fusion algorithm to fuse the first structured data parameter and the second structured data parameter, and its calculation formula is as follows:

[0041] In the formula, X fused is the fusion result, and are the first dynamic weight matrix and the second dynamic weight matrix respectively, X struct is the first structured data parameter, X IMF is the second structured parameter, is the Hadamard product.

[0042] It should be noted that the value of the variable is , and its rank is calculated. The rank conversions are respectively performed on the battery performance parameters and multi-physical characteristics. The battery performance parameter is , and the multi-physical characteristic is . Their ranks and are calculated respectively, is the battery performance parameter without rank conversion, is the multi-physical characteristic parameter without rank conversion, is the number of battery performance parameters, and j is the number of physical parameters, where , .

[0043] In this step S200, it also includes constructing an electro-thermal-mechanical multi-physical characteristic high-speed EMU battery aging mechanism model by combining the Arrhenius equation, quasi-two-dimensional model and equivalent circuit model. The calculation formula of the Arrhenius equation is as follows:

[0044] In the formula, is a calculation symbol representing a function, is the reaction rate constant at temperature , is the pre-exponential factor, is the activation energy, is the universal gas constant, is the absolute temperature; Among them, the quasi-two-dimensional model includes the charge conservation equation, mass conservation equation and heat conservation equation. The calculation formulas of the three equations are as follows: Among them, the charge conservation equation , in the formula, is the electric potential in the electrode, is the electrode conductivity, is the current density in the electrode, is the time; Among them, the mass conservation equation , In the formula, is the solid-phase lithium-ion concentration, is the solid-phase diffusion coefficient, is the current density, is the Faraday constant, is the volume fraction of electrode particles, is the solid-phase lithium-ion concentration, is the solid-phase diffusion coefficient is the time, is the absolute temperature; Among them, the heat conservation equation , in the formula, is the density, is the specific heat capacity, is the thermal conductivity, is the heat generated by the electrochemical reaction, is the heat loss, is the time, is the absolute temperature; Among them, the calculation formula of the equivalent circuit model is as follows:

[0045] In the formula, , and are the ohmic internal resistance, the electrochemical polarization resistance and the diffusion polarization resistance respectively, and are the electrochemical polarization capacitance and the diffusion polarization capacitance respectively, is the open-circuit voltage of the battery, is the terminal voltage of the battery, is the electrochemical polarization resistance under the applied voltage, is the diffusion polarization resistance under the applied voltage, is the current flowing through ; Considering the electrochemical reaction, heat transfer and mechanical behavior inside the battery, combining the Arrhenius equation, the quasi-two-dimensional model and the equivalent circuit model, an electro-thermal-mechanical coupling model is established, and its calculation formula is as follows:

[0046] In the formula, is the current density, is the exchange current density, , , and are the liquid-phase surface concentration, the solid-phase surface concentration reference concentration and the maximum concentration respectively, , , and They are the total strain, elastic strain, diffusion-induced strain, and thermal strain, respectively. is the stress. is the elastic modulus.

[0047] It should be noted that by solving the above coupling equations through numerical simulation methods, the prediction results of the mechanism model of the battery aging state are obtained.

[0048] S300. Decompose and extract the key information of multi-physical characteristic parameters by using the locally linear embedding method, use the key information as a sequence to input into the long short-term memory network and the self-attention mechanism, establish the mapping relationship between battery performance and multi-physical characteristics, generate a deep learning model of battery aging big data and train it to obtain the second aging state of the predicted output, where the key information includes key features, local structure information, and key features after dimensionality reduction.

[0049] It can be understood that in this step S300, it includes S301, S302, S303, and S304, where: S301. According to the data of multi-physical characteristic parameters, determine the neighborhood size of each data point through the Euclidean distance to obtain the neighborhood information. S302. Combine the neighborhood information, find its r nearest neighbor points for each data point, and calculate the corresponding reconstruction weights. The calculation formula is as follows:

[0050] In the formula, in the formula, is the reconstruction weight. is the identity matrix. is the weight matrix. is the transpose of the matrix. S303. Use the calculated reconstruction weights to construct the local neighborhood matrix of each data point, where each row represents the difference between a neighboring point and this data point. The calculation formula of its local neighborhood matrix is as follows: For each data point , calculate through the formula :

[0051] In the formula, is the local neighborhood matrix. are all neighboring points. is the point 's reconstruction weight. S304. Based on the local neighborhood matrix, extract the low-dimensional embedding result by solving the least squares problem and matrix eigenvalue and eigenvector analysis, and record the low-dimensional embedding result as the key information. It should be noted that the neighborhood is determined by using a distance metric (such as the Euclidean distance), and the neighborhood size r is determined by balancing the capture of local structure and noise robustness.

[0052] For each data point , find its r nearest neighbor points and calculate their corresponding reconstruction weights . For each data point , construct a local neighborhood matrix , where each row represents the difference between a neighboring point and , and solve the following least squares problem:

[0053] In the formula, is the weight vector, N is the weight matrix. Solve the above least squares problem to obtain the reconstruction weight of each data point , and then after calculating the reconstruction weight of each data point , construct a matrix.

[0054] It can be understood that the long short-term memory network in this step enables the model to effectively transmit information over a long time span by introducing a complex internal structure, including a memory unit, a forget gate, an input gate, and an output gate.

[0055] The long short-term memory network consists of a chain structure of multiple long short-term memory network units. The information output from the previous time step will be used as the operation input for the current time step. As the key information obtained by the local linear embedding method is used as the time series input for the current time step, after being processed by the long short-term memory network units, the aging state of the high-speed train battery pack is finally obtained. The memory unit state is gradually updated with each time step of the sequence. New information can be written, and old information can be cleared by "forgetting". The forget gate determines the removal of information from the memory unit. At each time step, the LSTM reads the current input and the hidden state from the previous time step, and outputs a value between 0 and 1 through the sigmoid function, corresponding to the proportion of forgetting. Information with an output value close to 1 is retained; information close to 0 is forgotten. The calculation formula for the forget gate is:

[0056] In the formula, the output of the forget gate at the current moment is , is the weight matrix of the forget gate, is the hidden state from the previous time step, is the input at the current time step, the bias vector of the forget gate.

[0057] The output gate determines how the state of the memory cell at the current time step affects the output hidden state, and it controls how the content of the memory cell is output at the current time step. After the forget gate is executed, the input gate determines which information in the input at the current time step is added to the memory cell. LSTM uses the sigmoid and tanh functions to process the current input, generating a vector of new information that is added to the memory cell state to ensure that only information relevant to the current context is updated. The calculation process is as follows:

[0058] In the formula, is the value of the input gate, is the weight matrix of the input gate, is the hidden state of the previous time step, is the input of the current time step, is the bias vector of the input gate, is the value of the candidate memory cell, is the weight matrix of the memory cell, is the bias vector of the memory cell, is the updated memory cell state, is the stored state of the memory cell at the previous time step.

[0059] The output gate determines the hidden state at the current time step, which is the output value of the LSTM. After being processed by the output gate, the information of the memory cell state is selectively output to the hidden state and used as the input for the next time step, thus completing the information transfer at the current time step. The calculation formula of the output gate is: The output gate determines the hidden state at the current time step, which is the output value of the LSTM. After being processed by the output gate, the information of the memory cell state is selectively output to the hidden state and used as the input for the next time step, thus completing the information transfer at the current time step. The calculation formula of the output gate is:

[0060] In the formula, is the value of the output gate, is the weight matrix of the output gate, is the hidden state of the previous time step, is the input of the current time step, is the bias vector of the output gate.

[0061] Then the memory cell state is processed by the tanh activation function and multiplied by the value of the output gate to obtain the final hidden state output :

[0062] In the formula, is the hidden state at the current time step, is the value of the output gate, is to update the memory cell state.

[0063] Among them, the attention mechanism has three ternary variables , , , and the above variables are obtained by multiplying the input by the weight matrix. Using , , to complete the mapping of the self-attention layer. The self-attention score of each element is calculated using and values through dot product; the self-attention scores are processed by the softmax function to obtain the self-attention weights of each element; the attention weights are calculated by the dot product with the values to generate the output vector. The output integrated by the attention mechanism is obtained through a fully connected layer, and its calculation formula is as follows:

[0064] In the formula, the matrix , is the length of the input sequence of is the dimension of each position in the sequence; , , represents the th , , corresponding weight matrix; is the corresponding weight vector; is the output, is the th dimension of the key.

[0065] It can be understood that the number of nodes in the input layer of the long short-term memory network-attention mechanism is the same as the number of low-dimensional embeddings after local linear embedding dimensionality reduction, and the output layer is the predicted value of the state of health (SOH) of the battery.

[0066] S400. Take the first aging state and the second aging state as two independent evidence bodies, and dynamically adjust the weights of the Dempster combination rule based on the conflict factor to complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

[0067] It is understandable that steps S401, S402, S403, and S404 are included in this step S400, where: S401: Take the first aging state and the second aging state as independent evidence bodies respectively. Let the confidence level of the first aging state be , and the confidence level of the second aging state be ; S402: Use the Dempster combination rule to calculate the conflict factor , and dynamically adjust the weight of the Dempster combination rule based on the size of the conflict factor. The formula for the conflict factor is as follows:

[0068] In the formula, is the conflict factor, is the empty set, , represent the first aging state and the second aging state respectively, is the confidence level of the first aging state, is the confidence level of the second aging state; S403: According to the adjusted Dempster combination rule, calculate the combined confidence level of each aging state. The formula is as follows:

[0069] In the formula, is the combined confidence level, is the conflict factor, , , represent the final output aging state, the first aging state, and the second aging state respectively, is the confidence level of the first aging state, is the confidence level of the second aging state; S404: Arrange the combined confidence levels of each aging state from high to low, and select the state with the highest confidence level as the final aging state for output.

[0070] It should be noted that the weight of the combination rule is dynamically adjusted according to the size of the conflict factor. When the conflict factor is large, the combination weight is reduced to avoid unreasonable results caused by the conflict.

[0071] The final aging state is determined by the maximum confidence criterion to determine the final aging state output. Calculate the combined confidence level of each aging state and select the state with the highest confidence level as the final result for output.

[0072] It should be noted that in this embodiment, through the aging experiment of the high-speed EMU battery under complex working conditions, different temperatures in actual operation (such as -20°C, 0°C, 25°C, 40°C, etc.), different loads (such as traction loads, auxiliary loads, impact loads with different powers, etc.), and different charge and discharge modes (such as constant current charging, constant voltage charging, long-term floating charging, pulse discharging, etc.) are simulated to conduct long-term aging tests on the battery. During the experiment, the performance data of the battery at different cycle numbers are recorded, including but not limited to the following multi-physical characteristic parameters: battery voltage, current, impedance, battery surface temperature field distribution and temperature rise time, and stress and strain parameters of the electrode material.

[0073] Next, according to the data recording and preprocessing, the above experimental data are recorded and sorted to form a database, and the data are preprocessed to ensure the accuracy and consistency of the data, including removing outliers and normalizing. Through the aging experiment of the high-speed EMU battery under complex working conditions, data such as the electrical (voltage, current, impedance) characteristics of the high-speed EMU lithium-ion battery, internal electrochemical heat generation and heat transfer and battery temperature rise, and reaction decomposition and mechanical deformation of the battery material are obtained.

[0074] After that, voltage-current-time curves, capacity attenuation curves, electrochemical impedance spectroscopy characteristic frequency curves, capacity increment curves, battery surface temperature field distributions, and stress and strain curves of the electrode material are established. On this basis, the extraction of battery characteristics is realized to explore the battery aging law. Among them, the Spearman analysis method is used in the correlation analysis of characteristic parameters to quantify the correlation between battery performance parameters (such as capacity attenuation rate, internal resistance change, etc.) and multi-physical characteristic parameters, and the characteristic parameters with a significance level higher than 0.8 are selected for subsequent model construction.

[0075] Then, an equivalent circuit DP model is constructed, an aging mechanism model is constructed using the Arrhenius equation and the quasi-two-dimensional model theory, equations are constructed, and then the temperature parameters are solved using the experimental data, and the above coupled equations are solved by numerical simulation methods to obtain the prediction results of the mechanism model of the battery aging state.

[0076] Finally, the local linear embedding method is used to extract the value information of multi-physical characteristics, which is used as the input sequence of the long short-term memory network.

[0077] The input of the LSTM model is a sequence , and the output is a sequence . In the attention mechanism, the originally involved in the gating calculation in the LSTM can be set as , is the output of the self-attention mechanism. Among them, let , , and the matrix and the sequence If splicing is performed, a new forget gate is generated:

[0078] The output gate is:

[0079] The input gate is:

[0080] The candidate memory cell value is:

[0081] The new time series output is weighted through the attention mechanism where the weighted sequence and the calculated weight is:

[0082] Then for there is:

[0083] The number of nodes in the input layer of the long short-term memory network-attention mechanism is the same as the number of low-dimensional embeddings after local linear embedding dimensionality reduction, and the output layer is the predicted value of the state of health (SOH) of the battery.

[0084] Taking the aging state confidence degree output by the mechanism model and the aging probability distribution output by the data model as independent evidence bodies, dynamically adjusting the weights of the Dempster combination rule based on the conflict factor, completing the joint confidence degree assignment of the two types of evidence, and determining the final aging state through the maximum confidence criterion.

[0085] Taking the aging state confidence degree output by the mechanism model and the aging probability distribution output by the data model as independent evidence bodies, let the aging state confidence degree output by the mechanism model be and the aging state confidence degree output by the data model be . Using the Dempster combination rule, dynamically adjusting the weights of the Dempster combination rule based on the conflict factor, and completing the joint confidence degree assignment of the two types of evidence.

[0086] Finally, dynamically adjust the weights of the combination rule according to the size of the conflict factor. When the conflict factor is large, reduce the combination weight to avoid unreasonable results caused by conflicts. The final aging state is determined by the maximum confidence criterion to output the final aging state, calculate the joint confidence degree of each aging state, and select the state with the highest confidence degree as the final result.

[0087] Example 2: As Figure 2As shown in the figure, this embodiment provides a high-speed train battery aging assessment system that integrates mechanism and data. Refer to Figure 2 The system includes: The extraction and processing module 701: It is used to simulate the working conditions of different temperatures, different loads, and different charge and discharge modes during actual operation through the high-speed train battery working condition aging experiment, perform electrochemical impedance spectroscopy testing and open-circuit voltage measurement on the battery, record the experimental data during the battery aging process, extract and process the multi-physical characteristic parameters of the battery in the experimental data, and obtain the processed first structured data parameter and first unstructured data parameter. The first prediction module 702: It is used to physically decompose and transform the first unstructured data parameter into a second structured data parameter, fuse the first structured data parameter and the second structured data parameter, and combine the Arrhenius equation, the quasi-two-dimensional model, and the equivalent circuit model to construct a high-speed train battery aging mechanism model with multi-physical characteristics of electricity-thermal-mechanics. Use numerical simulation methods to solve the high-speed train battery aging mechanism model, and then predict the first aging state of the battery. The second prediction module 703: It is used to decompose and extract the key information of the multi-physical characteristic parameters by using the locally linear embedding method, input the key information as a sequence into the long short-term memory network and the self-attention mechanism, establish the mapping relationship between battery performance and multi-physical characteristics, generate and train the deep learning model of battery aging big data, and obtain the predicted second aging state. The key information includes key features, local structure information, and key features after dimensionality reduction. The determination and evaluation module 704: It is used to take the first aging state and the second aging state as two independent evidence bodies, dynamically adjust the weights of the Dempster combination rule based on the conflict factor, complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

[0088] Specifically, for the extraction and processing module 701, the processing process of the processed first structured data parameter includes: The conversion unit: It is used to perform rank transformation on the battery performance parameters and multi-physical characteristic parameters obtained from the high-speed train battery working condition aging experiment respectively. The rank transformation includes: sorting the values of each variable from small to large, and assigning a rank to each value; if there are the same values, assign the average rank of the same values that appear during the sorting of the variable values. The partitioning unit: It is used to partition the parameters after rank transformation into the first rank and the second rank. The first rank is the rank of the battery performance parameters, and the second rank is the rank of the multi-physical characteristic parameters. Calculation and screening unit: Based on the first rank and the second rank, calculate the Spearman correlation coefficient between each pair of performance parameters and multi-physical characteristic parameters, and screen out the multi-physical characteristic parameters greater than 0.8, which are denoted as the processed first structured data parameters. The calculation formula of the Spearman correlation coefficient is:

[0089] In the formula, is the correlation coefficient, is the battery performance parameter without rank transformation, is the multi-physical characteristic parameter without rank transformation, is the first rank, is the second rank, is the number of samples, is the number of battery performance parameters, is the number of physical characteristic parameters.

[0090] Specifically, for the extraction and processing module 701, the processing process of the processed first unstructured data parameter includes: Processing and extraction unit: Used to obtain the Nyquist diagram in the electrochemical impedance spectroscopy test, perform differential geometry processing on the Nyquist diagram, and extract the characteristic frequency set of the curvature mutation points. The calculation formula is as follows:

[0091] In the formula, θ is the angle between the tangent direction at the corresponding frequency point and the real axis, is the characteristic frequency set extracted from the Nyquist diagram of the electrochemical impedance spectroscopy, is related to the angle θ corresponding characteristic frequency; Decomposition unit: Used to construct a dynamic weighted tensor based on the characteristic frequency set, decompose the open circuit voltage curve into time-domain differential eigenmode functions, and denote the time-domain differential eigenmode functions as the processed first unstructured data parameters. The calculation formula is as follows:

[0092] In the formula, is the dynamic weighted tensor, is the characteristic frequency corresponding to the angle θ, τ is the charge-discharge cycle, is the charge transfer resistance at the angle θ, is the double-layer capacitance at the angle θ.

[0093] Specifically, for the first prediction module 702, it includes: Generation unit: It is used to calculate the mean, variance, energy and time-frequency characteristics of each modality in the first unstructured data parameter, and associate the battery health state through the aging sensitivity weighting coefficient to generate the second structured data parameter reflecting the microscopic aging mechanism; Fusion unit: It is used to fuse the first structured data parameter and the second structured data parameter by using a multi-modal fusion algorithm, and its calculation formula is as follows:

[0094] In the formula, X fused is the fusion result, and are the first dynamic weight matrix and the second dynamic weight matrix respectively, X struct is the first structured data parameter, X IMF is the second structured parameter, is the Hadamard product.

[0095] It should be noted that regarding the system in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0096] Embodiment 3:

[0097] Corresponding to the above method embodiment, in this embodiment, a mechanism and data fusion-based EMU battery aging assessment device is also provided. The mechanism and data fusion-based EMU battery aging assessment device described below can be mutually corresponding and referred to the mechanism and data fusion-based EMU battery aging assessment method described above.

[0098] Figure 3 is a block diagram of a mechanism and data fusion-based EMU battery aging assessment device 800 shown according to an exemplary embodiment. As Figure 3 shown, the mechanism and data fusion-based EMU battery aging assessment device 800 includes: a processor 801 and a memory 802. The mechanism and data fusion-based EMU battery aging assessment device 800 further includes one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0099] Among them, the processor 801 is used to control the overall operation of the EMU battery aging evaluation device 800 for mechanism and data fusion to complete all or part of the steps in the above-mentioned EMU battery aging evaluation method for mechanism and data fusion. The memory 802 is used to store various types of data to support the operation of the EMU battery aging evaluation device 800 for mechanism and data fusion. These data may include, for example, instructions for any application program or method operating on the EMU battery aging evaluation device 800 for mechanism and data fusion, as well as application program-related data, such as contact data, sent and received messages, pictures, audio, video, and so on. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen can be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal can be further stored in the memory 802 or sent through the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be a keyboard, a mouse or buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the EMU battery aging evaluation device 800 for mechanism and data fusion and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them. Therefore, the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module or an NFC module.

[0100] In an exemplary embodiment, the evaluation device 800 for EMU battery aging based on mechanism and data fusion can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components, and is used to execute the above-mentioned method for evaluating EMU battery aging based on mechanism and data fusion.

[0101] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When the program instructions are executed by a processor, the steps of the above-mentioned method for evaluating EMU battery aging based on mechanism and data fusion are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above program instructions can be executed by the processor 801 of the evaluation device 800 for EMU battery aging based on mechanism and data fusion to complete the above-mentioned method for evaluating EMU battery aging based on mechanism and data fusion.

[0102] Embodiment 4:

[0103] Corresponding to the above method embodiment, in this embodiment, a readable storage medium is further provided. A readable storage medium described below can be correspondingly referred to with a method for evaluating EMU battery aging based on mechanism and data fusion described above.

[0104] A computer program is stored on the readable storage medium. When the computer program is executed by a processor, the steps of the method for evaluating EMU battery aging based on mechanism and data fusion in the above method embodiment are implemented.

[0105] The readable storage medium can specifically be various readable storage media 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 disc that can store program codes.

[0106] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0107] As described above, this is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or replacements, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for evaluating the aging of EMU batteries by integrating mechanism and data, characterized in that, Including: Through the aging experiment of the high-speed EMU battery under different working conditions, simulate the working conditions of different temperatures, different loads, and different charge and discharge modes during actual operation, and conduct electrochemical impedance spectroscopy tests and open-circuit voltage measurements on the battery, record the experimental data during the battery aging process, extract and process the multi-physical characteristic parameters of the battery in the experimental data to obtain the processed first structured data parameters and first unstructured data parameters; Physically decompose the first unstructured data parameters into second structured data parameters, fuse the first structured data parameters and the second structured data parameters, and combine the Arrhenius equation, quasi-two-dimensional model, and equivalent circuit model to construct an aging mechanism model of the high-speed EMU battery with multi-physical characteristics. Use the numerical simulation method to solve the aging mechanism model of the high-speed EMU battery, and then predict the first aging state of the battery; Adopt the locally linear embedding method to decompose and extract the key information of the multi-physical characteristic parameters, use the key information as a sequence and input it into the long short-term memory network and self-attention mechanism, establish the mapping relationship between the battery performance and multi-physical characteristics, generate and train the deep learning model of the battery aging big data, and obtain the predicted second aging state, where the key information includes key features, local structure information, and key features after dimensionality reduction; Regard the first aging state and the second aging state as two independent evidence bodies, and dynamically adjust the weights of the Dempster combination rule based on the conflict factor to complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

2. The method for evaluating the aging of the EMU battery by mechanism and data fusion according to claim 1, characterized in that The process of extracting and processing the multi-physical characteristic parameters of the battery in the experimental data to obtain the processed first structured data parameters and first unstructured data parameters, where the processing process of the processed first structured data parameters includes: Perform rank transformation on the battery performance parameters and multi-physical characteristic parameters obtained from the aging experiment of the high-speed EMU battery respectively, including: sort the values of each variable from small to large, and assign a rank to each value; if there are the same values, assign the average rank of the same values that appear during the sorting of the variable values; Divide the parameters after rank transformation into the first rank and the second rank, where the first rank is the rank of the battery performance parameters, and the second rank is the rank of the multi-physical characteristic parameters; Based on the first rank and the second rank, calculate the Spearman correlation coefficient between each pair of performance parameters and multi-physical characteristic parameters, and screen out the multi-physical characteristic parameters greater than 0.8, and record them as the processed first structured data parameters, where the calculation formula of the Spearman correlation coefficient is: In the formula, is the correlation coefficient, is the battery performance parameter without rank transformation, is the multi-physical characteristic parameter without rank transformation, is the first rank, is the second rank, is the number of samples, is the number of battery performance parameters, is the number of physical parameters.

3. The method for evaluating the aging of EMU batteries by mechanism and data fusion according to claim 1, characterized in that, The process of extracting and processing the multi-physical characteristic parameters of the battery in the experimental data to obtain the processed first structured data parameters and first unstructured data parameters, where the processing process of the processed first unstructured data parameters includes: Obtain the Nyquist diagram in the electrochemical impedance spectroscopy test, perform differential geometric processing on the Nyquist diagram, and extract the set of characteristic frequencies of the curvature mutation points, and its calculation formula is as follows: In the formula, is the angle between the tangent direction at the corresponding frequency point and the real axis, is the set of characteristic frequencies extracted from the Nyquist plot of the electrochemical impedance spectrum, is the characteristic frequency corresponding to the angle ; Construct a dynamic weighted tensor based on the set of characteristic frequencies, decompose the open-circuit voltage curve into time-domain differential intrinsic mode functions, and denote the time-domain differential intrinsic mode functions as the processed first unstructured data parameters. The calculation formula is as follows: In the formula, is the dynamic weighted tensor, is the characteristic frequency corresponding to the angle , is the charge-discharge cycle, is the charge transfer resistance at the angle , is the double-layer capacitance at the angle .

4. The method for evaluating the aging of the EMU battery by mechanism and data fusion according to claim 1, characterized in that, Physically decompose the first unstructured data parameters into second structured data parameters, and fuse the first structured data parameters and the second structured data parameters, including: Calculate the mean, variance, energy, and time-frequency characteristics of each mode in the first unstructured data parameters, and correlate the battery health state through the aging sensitivity weighting coefficient to generate the second structured data parameters reflecting the microscopic aging mechanism; Use a multi-modal fusion algorithm to fuse the first structured data parameters and the second structured data parameters. The calculation formula is as follows: In the formula, X fused is the fusion result, and are the first dynamic weight matrix and the second dynamic weight matrix respectively, X struct is the first structured data parameter, X IMF is the second structured parameter, is the Hadamard product.

5. The method for evaluating the aging of EMU batteries by mechanism and data fusion according to claim 1, characterized in that, The method of using local linear embedding to decompose and extract the key information of multi-physical characteristic parameters includes: Based on the data of multi-physical characteristic parameters, determine the neighborhood size of each data point through the Euclidean distance to obtain the neighborhood information; Combined with the neighborhood information, find its r nearest neighbor points for each data point and calculate the corresponding reconstruction weights. The calculation formula is as follows: In the formula, is the reconstruction weight, is the identity matrix, is the weight matrix, is the transpose of the matrix; Use the calculated reconstruction weights to construct the local neighborhood matrix of each data point, where each row represents the difference between a neighboring point and the data point. The calculation formula of its local neighborhood matrix is as follows: For each data point , calculate through the formula : In the formula, is the local neighborhood matrix, are all neighboring points, is the point 's reconstruction weight; Based on the local neighborhood matrix, extract the low-dimensional embedding result by solving the least squares problem and matrix eigenvalue eigenvector analysis, and denote the low-dimensional embedding result as the key information.

6. The method for evaluating the aging of the EMU battery by mechanism and data fusion according to claim 1, wherein Regarding the first aging state and the second aging state as two independent evidence bodies, and dynamically adjusting the weights of the Dempster combination rule based on the conflict factor to complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion, including: Take the first aging state and the second aging state as independent evidence bodies respectively, and set the confidence level of the first aging state as , and the confidence level of the second aging state as ; Using the Dempster combination rule, calculate the conflict factor , and dynamically adjust the weight of the Dempster combination rule based on the size of the conflict factor, where the conflict factor is calculated as follows: In the formula, is the conflict factor, is the empty set, and represent the first aging state and the second aging state respectively, is the confidence level of the first aging state, is the confidence level of the second aging state; According to the adjusted Dempster combination rule, calculate the joint confidence of each aging state. The calculation formula is as follows: In the formula, is the combined confidence level, is the conflict factor, , , respectively represent the final output aging state, the first aging state, and the second aging state, is the confidence level of the first aging state, is the confidence level of the second aging state; Arrange the joint confidences of each aging state from high to low, and select the state with the highest confidence as the final aging state for output.

7. A mechanism and data fusion-based EMU battery aging assessment system, based on the mechanism and data fusion-based EMU battery aging assessment method described in claim 1, characterized in that, Including: Extraction processing module: used to simulate the working conditions of different temperatures, different loads, and different charge and discharge modes in actual operation through the battery condition aging experiment of high-speed EMUs, perform electrochemical impedance spectroscopy testing and open-circuit voltage measurement on the battery, record the experimental data during the battery aging process, extract the multi-physical characteristic parameters of the battery in the experimental data for processing, and obtain the processed first structured data parameters and first unstructured data parameters; First prediction module: used to physically decompose the first unstructured data parameters into second structured data parameters, fuse the first structured data parameters and the second structured data parameters, and combine the Arrhenius equation, quasi-two-dimensional model, and equivalent circuit model to construct an electro-thermal-mechanical multi-physical characteristic high-speed EMU battery aging mechanism model, and use the numerical simulation method to solve the high-speed EMU battery aging mechanism model, and then predict the first aging state of the battery; The second prediction module: It is used to decompose and extract the key information of multi-physical characteristic parameters by using the locally linear embedding method, take the key information as a sequence and input it into the long short-term memory network and the self-attention mechanism, establish the mapping relationship between battery performance and multi-physical characteristics, generate a deep learning model for battery aging big data and train it to obtain the second aging state of the predicted output, where the key information includes key features, local structure information and key features after dimensionality reduction; The determination and evaluation module: It is used to take the first aging state and the second aging state as two independent evidence bodies, dynamically adjust the weights of the Dempster combination rule based on the conflict factor, complete the joint confidence assignment of the two independent evidence bodies, and then determine the final aging state through the maximum confidence criterion.

8. The EMU battery aging assessment system based on mechanism and data fusion according to claim 7, characterized in that, The extraction and processing module, where the processing process of the processed first structured data parameter includes: The conversion unit: It is used to perform rank transformation on the battery performance parameters and multi-physical characteristic parameters obtained from the high-speed EMU battery condition aging experiment respectively, including: sorting the values of each variable from small to large, and assigning a rank to each value; if there are the same values, assign the average rank of the same values that appear during the sorting of variable values; The partitioning unit: It is used to partition the parameters after rank transformation into a first rank and a second rank, where the first rank is the rank of the battery performance parameters, and the second rank is the rank of the multi-physical characteristic parameters; The calculation and screening unit: It is used to calculate the Spearman correlation coefficient between each pair of performance parameters and multi-physical characteristic parameters based on the first rank and the second rank, and screen out the multi-physical characteristic parameters greater than 0.8, and record them as the processed first structured data parameters, where the calculation formula of the Spearman correlation coefficient is: In the formula, is the correlation coefficient, is the battery performance parameter without rank transformation, is the multi-physical characteristic parameter without rank transformation, is the first rank, is the second rank, is the sample size, is the number of battery performance parameters, is the number of physical characteristic parameters.

9. The EMU battery aging assessment system based on mechanism and data fusion according to claim 7, characterized in that The extraction and processing module, where the processing process of the processed first unstructured data parameter includes: The processing and extraction unit: It is used to obtain the Nyquist diagram in the electrochemical impedance spectroscopy test, perform differential geometric processing on the Nyquist diagram, and extract the set of characteristic frequencies of the curvature mutation points, and its calculation formula is as follows: In the formula, is the angle between the tangent direction at the corresponding frequency point and the real axis, is the set of characteristic frequencies extracted from the Nyquist plot of the electrochemical impedance spectrum, is related to the angle corresponding characteristic frequency; The decomposition unit: It is used to construct a dynamic weighted tensor based on the set of characteristic frequencies, decompose the open circuit voltage curve into time-domain differential intrinsic mode functions, and record the time-domain differential intrinsic mode functions as the processed first unstructured data parameters, and its calculation formula is as follows: In the formula, is the dynamic weighted tensor, is the characteristic frequency corresponding to the angle , is the charge-discharge cycle, is the charge transfer resistance at the angle , is the double-layer capacitance at the angle .

10. The EMU battery aging assessment system based on mechanism and data fusion according to claim 7, characterized in that The first prediction module, where it includes: The generation unit: It is used to calculate the mean, variance, energy and time-frequency characteristics of each mode in the first unstructured data parameter, and associate the battery health state through the aging sensitivity weighting coefficient to generate the second structured data parameter reflecting the microscopic aging mechanism; The fusion unit: It is used to fuse the first structured data parameter and the second structured data parameter by using a multi-modal fusion algorithm, and its calculation formula is as follows: Wherein, X fused is the fusion result, and are the first dynamic weight matrix and the second dynamic weight matrix respectively, X struct is the first structured data parameter, X IMF is the second structured parameter, is the Hadamard product.

Images