Battery safety performance test system based on artificial intelligence

Abstract

The invention discloses a battery safety performance test system based on artificial intelligence. The system comprises a data acquisition module, an algorithm processing module, a multi-field coupling module, a parameter analysis module, an optimization regulation and control module and a result output module. The data acquisition module acquires multi-dimensional data, transmits the multi-dimensional data to the algorithm processing module, processes the multi-dimensional data through an SE-Res-LSTM algorithm, transmits the processed data to the multi-field coupling module, calculates the processed data through a coupled electrochemical-thermal-mechanical multi-field equation, transmits a result to the parameter analysis module, analyzes a safety performance parameter threshold value, and transmits the analyzed safety performance parameter threshold value to the optimization regulation and control module. And the working condition parameters are adjusted by a self-adaptive fluctuation optimization algorithm and are finally displayed by a result output module. The system overcomes the defect that traditional single physical field analysis and parameter adjustment depend on experience, comprehensive and dynamic evaluation is achieved, the test efficiency and accuracy are improved, and support is provided for battery safety protection.

Description

Technical Field

[0001] The present invention relates to the field of battery safety performance testing, and in particular to a battery safety performance testing system based on artificial intelligence. Background Art

[0002] With the rapid development of the new energy industry, the safety performance of batteries, as core components for energy storage and supply, has attracted considerable attention. In applications such as electric vehicles and energy storage power stations, safety incidents such as thermal runaway and structural damage caused by factors such as overcharging and over-discharging, high and low temperature environments, and mechanical collisions frequently occur, severely restricting the sustainable development of the industry. Furthermore, the complex coupling between electrochemical reactions, heat transfer, and structural stress within batteries makes it difficult for traditional testing methods to fully capture the changing patterns of safety performance under multi-field coupling. Therefore, it is urgent to leverage artificial intelligence technology to build more accurate and efficient testing systems to meet the needs of comprehensive and dynamic battery safety performance assessments.

[0003] Existing technologies have obvious deficiencies in battery safety performance testing. On the one hand, traditional testing systems often use a single physical field model for analysis, failing to fully consider the interactions and coupling mechanisms between electrochemical, thermal, and mechanical fields. This leads to a one-sided identification of potential safety hazards within the battery, making it difficult to accurately predict the occurrence of chain reactions such as thermal runaway and structural failure. On the other hand, the adjustment of operating parameters during the test process often relies on empirical settings or simple algorithms, lacking dynamic optimization capabilities based on real-time safety performance data. This makes the test inefficient and difficult to cover the safety performance boundaries under complex working conditions, making it impossible to provide comprehensive and reliable technical support for battery design optimization and safety protection. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides a battery safety performance testing system based on artificial intelligence.

[0005] The technical solution adopted by the present invention is a battery safety performance testing system based on artificial intelligence, including a data acquisition module, an algorithm processing module, a multi-field coupling module, a parameter analysis module, an optimization and control module, and a result output module; the data acquisition module is connected to the algorithm processing module, and is used to collect the voltage, current, temperature, internal pressure and electrode material deformation data of the battery under different working conditions in real time, and transmit the collected data to the algorithm processing module; the algorithm processing module is respectively connected to the data acquisition module and the multi-field coupling module, and after receiving the data transmitted by the data acquisition module, the data is feature extracted and time series modeled by the SE-Res-LSTM algorithm, and the processed characteristic parameters are transmitted to the multi-field coupling module; the multi-field coupling module is respectively connected to the algorithm processing module and the parameter analysis module, and after receiving the characteristic parameters, a coupled electrochemical-thermal-mechanical multi-field equation is constructed to couple the electrochemical reaction, heat transfer and structural stress distribution inside the battery to perform calculations. , and transmit the calculation results to the parameter analysis module; the parameter analysis module is respectively connected to the multi-field coupling module and the optimization and control module. After receiving the calculation results, it quantitatively analyzes different parameters of battery safety performance, including thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage and battery cycle life attenuation rate, and transmits the parameter thresholds obtained by analysis to the optimization and control module; the optimization and control module is respectively connected to the parameter analysis module and the result output module. After receiving the parameter thresholds, it dynamically adjusts the operating parameters during the battery test, including charge and discharge rate, ambient temperature, external load and test time, through an adaptive fluctuation optimization algorithm, and transmits the adjusted parameters and corresponding safety performance evaluation results to the result output module; the result output module is connected to the optimization and control module. After receiving the transmitted parameters and evaluation results, it presents the battery safety performance test results in the form of data tables and three-dimensional visualization models.

[0006] Furthermore, the feature extraction process of the SE-Res-LSTM algorithm in the algorithm processing module satisfies the following model formula: , ,in, is the feature vector extracted by SE-Res-LSTM algorithm at time t, is the original data input by the data acquisition module at time t, is the weight parameter of the LSTM network, For the squeeze-and-excitation attention mechanism operation, is the Sigmoid activation function, is the global average pooling operation, is the weight parameter of the attention mechanism; the electrochemical reaction sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the conductivity of the electrode material, is the solid phase potential, is the interfacial current density, is the solid phase capacitance, For time, is the electrolyte conductivity, is the liquid phase potential, is the liquid phase capacitor.

[0007] Furthermore, the heat transfer sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module satisfies: ,in, is the battery material density, is the specific heat capacity, is the temperature, is the thermal conductivity, is the heat generation rate of the electrochemical reaction, is the Joule heat generation rate; the parameter adjustment model of the adaptive fluctuation optimization algorithm in the optimization control module is: ,in, For the After the adjustment, the working parameters is the working condition parameter after the kth adjustment, is the coefficient of fluctuation, is a standard normally distributed random number, are the maximum and minimum values ​​of the working condition parameters, is the attenuation coefficient, The number of adjustments.

[0008] Furthermore, the structural stress sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the stress tensor, is the material density, is the body force vector, is the displacement vector, is the elasticity matrix, is the strain tensor; the quantitative analysis of the critical temperature of battery thermal runaway in the parameter analysis module satisfies: ,in, is the thermal runaway critical temperature, is the battery volume, is the initial temperature.

[0009] Furthermore, the quantitative analysis model for the maximum allowable pressure of the battery in the parameter analysis module is: ,in, is the maximum allowable pressure, is the inner radius of the battery shell, is the yield strength of the shell material, is the shell thickness, is the initial pressure inside the battery; the time series modeling loss function of the SE-Res-LSTM algorithm in the algorithm processing module is: ,in, is the loss value, is the sample size, is the true value at time t, is the predicted value at time t, is the regularization coefficient, is the number of network layers, is the network weight of the i-th layer.

[0010] Furthermore, the convergence judgment condition of the adaptive fluctuation optimization algorithm in the optimization and control module is: ,in, is the safety performance evaluation function, is the convergence accuracy threshold; the quantitative analysis of the battery cycle life attenuation rate in the parameter analysis module satisfies: ,in, is the cycle life attenuation rate, is the battery capacity at the nth cycle, is the initial capacity, is the attenuation coefficient, is the number of cycles, is the decay index.

[0011] Furthermore, the multi-field coupling module includes an electrochemical field calculation unit, a temperature field simulation unit, a mechanical stress analysis unit, and a multi-field coupling coordination unit; the electrochemical field calculation unit receives the characteristic parameters transmitted by the algorithm processing module, calculates the lithium ion concentration distribution, electrode potential and interface reaction rate at different positions based on the electrode reaction kinetic equation, meshes the calculation area through the finite element method, and stores the calculation results of each grid node in a temporary data buffer; the temperature field simulation unit obtains the electrochemical field calculation results from the temporary data buffer, calculates the temperature gradient distribution inside and on the surface of the battery according to the heat conduction equation and the convection heat transfer model, and combines the heat generated by the electrochemical reaction, the contact thermal resistance and the environmental heat dissipation Conditions are set, the temperature field is transiently solved and the data buffer is updated; the mechanical stress analysis unit calls the temperature field and electrochemical field data in the data buffer, calculates the expansion / contraction of the electrode material caused by temperature changes and lithium ion insertion / deinsertion based on the thermoelasticity theory, establishes a stress balance equation to solve the normal stress, shear stress and strain distribution inside the battery, and writes the results into the data buffer; the multi-field coupling coordination unit couples and iterates the calculation results of the electrochemical field, temperature field and mechanical stress field in the data buffer, corrects the calculation parameters of each field through the field variable association equation, repeats the iteration until the convergence error of the calculation results of each field meets the preset threshold, and transmits the final coupled calculation results to the parameter analysis module.

[0012] Furthermore, the parameter analysis module includes a thermal safety parameter calculation unit, a mechanical safety parameter evaluation unit, an electrochemical safety parameter analysis unit, and a comprehensive safety performance judgment unit; the thermal safety parameter calculation unit receives the temperature field data transmitted by the multi-field coupling module, calculates the maximum temperature inside the battery, the temperature change rate and the critical temperature of thermal runaway, analyzes the heat accumulation rate and heat diffusion path in different areas, determines the thermal safety warning parameters and stores them in the analysis database; the mechanical safety parameter evaluation unit obtains the stress field data from the multi-field coupling module, calculates the maximum stress, strain value and fatigue damage accumulation of the battery shell and electrode material, evaluates the load-bearing capacity and deformation limit, and write the mechanical safety parameters and their thresholds into the analysis database; the electrochemical safety parameter analysis unit calculates the electrolyte decomposition rate, diaphragm breakdown voltage, electrode polarization degree and lithium ion concentration distribution uniformity based on the electrochemical field data of the multi-field coupling module, analyzes the internal short circuit risk of the battery and the electrochemical performance attenuation law, and stores the electrochemical safety parameters in the analysis database; the comprehensive safety performance judgment unit calls the thermal safety, mechanical safety and electrochemical safety parameters in the analysis database, calculates the comprehensive safety factor of the battery through a multi-parameter weighted analysis method, determines the safety threshold range of different parameters, and transmits the parameter threshold obtained by analysis to the optimization control module.

[0013] Furthermore, the optimization and control module includes an operating condition parameter initialization unit, a fluctuation parameter generation unit, a performance feedback adjustment unit, and an optimization result determination unit; the operating condition parameter initialization unit sets the initial value and value range of the operating condition parameters of the initial charge and discharge rate, ambient temperature, external load and test duration according to the battery type and test requirements, and establishes a mapping relationship model between the operating condition parameters and the safety performance parameters; the fluctuation parameter generation unit generates random fluctuations within the operating condition parameter value range based on an adaptive fluctuation optimization algorithm, adjusts the fluctuation amplitude and direction in combination with the current safety performance evaluation results, and generates new operating condition parameter candidate values; the performance feedback adjustment unit inputs the new operating condition parameter candidate values ​​into the battery testing system, obtains the corresponding safety performance parameter change data, calculates the performance improvement and compares it with the preset threshold. If the requirements are met, the candidate value is retained, otherwise the fluctuation parameter is adjusted to regenerate the candidate value; the optimization result determination unit screens the operating condition parameter candidate values ​​obtained through multiple iterations, selects the operating condition parameter combination that optimizes the safety performance parameters, and transmits the adjusted parameters and the corresponding safety performance evaluation results to the result output module.

[0014] A battery safety performance testing system based on artificial intelligence, the operation of the system includes the following steps: Step S1: Using a data acquisition module, the voltage, current, temperature, internal pressure, and electrode material deformation data of the battery under charge and discharge cycles, high and low temperature shock, and vibration loading conditions are collected at a preset sampling frequency, and the collected data are sorted in time series and abnormal jump data are removed; Step S2: Input the sorted data into the algorithm processing module, perform multi-layer feature extraction on the data using the SE-Res-LSTM algorithm, build a time series prediction model to fit the changing trend of the battery safety performance parameters, and output a feature parameter matrix; Step S3: Input the characteristic parameter matrix into the multi-field coupling module to construct a coupled electrochemical-thermal-mechanical multi-field equation. After setting boundary conditions and initial conditions, numerical solution is performed to obtain the spatiotemporal variation data of the electrochemical reaction rate, temperature distribution, and stress distribution inside the battery. Step S4: input the spatiotemporal variation data into a parameter analysis module, quantitatively calculate the thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage, and battery cycle life attenuation rate, and determine the numerical range and correlation of different parameters; Step S5: Input the parameter value range and correlation relationship into the optimization control module, perform multiple rounds of dynamic adjustment on the charge and discharge rate, ambient temperature, external load and test duration through the adaptive fluctuation optimization algorithm, and record the changes in the safety performance parameters after each adjustment; Step S6: Input the adjusted operating parameters and safety performance evaluation results into the result output module, generate a test report containing the values ​​of various safety performance parameters, change curves and three-dimensional distribution models according to the preset data format, and store them in the form of an editable file.

[0015] Beneficial effects: The present invention proposes a battery safety performance testing system based on artificial intelligence, which comprehensively collects multi-dimensional parameters of the battery through the data acquisition module, extracts features and models them with the help of the SE-Res-LSTM algorithm through the algorithm processing module, and combines the coupled electrochemical-thermal-mechanical multi-field equations constructed by the multi-field coupling module to achieve accurate analysis of multi-field interactions and coupling mechanisms, changing the one-sidedness of traditional single physical field model analysis, and can comprehensively capture the laws of safety performance changes, accurately identify internal safety hazards, and accurately predict chain reactions such as thermal runaway and structural failure. At the same time, the optimization and control module uses an adaptive fluctuation optimization algorithm to dynamically adjust the operating parameters according to real-time safety performance data, replacing the adjustment method that relies on experience or simple algorithms, improving test efficiency, and being able to cover the safety performance boundaries under complex working conditions. It provides comprehensive and reliable technical support for battery design optimization and safety protection, and meets the needs of all-round and dynamic evaluation of battery safety performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a diagram of the system module composition of the present invention; Figure 2 This is a flow chart of the system operation of the present invention. DETAILED DESCRIPTION

[0017] It should be noted that, unless there is a conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The application is further described in detail below with reference to the drawings and specific embodiments.

[0018] like Figure 1 As shown, a battery safety performance testing system based on artificial intelligence includes a data acquisition module, an algorithm processing module, a multi-field coupling module, a parameter analysis module, an optimization and control module, and a result output module; The data acquisition module is connected to the algorithm processing module and is used to collect voltage, current, temperature, internal pressure and electrode material deformation data of the battery under different working conditions in real time, and transmit the collected data to the algorithm processing module; Specifically, the data acquisition module is the foundation for the system to obtain raw battery information, and its technical parameters directly affect the accuracy of subsequent analysis. The module's sampling frequency range is 1kHz-10kHz, adjustable to suit different testing requirements to ensure it captures the battery's rapidly changing electrical signals and physical quantities. The voltage measurement range is 0V-5V with an accuracy of ±0.01V, and the current measurement range is -100A-100A with an accuracy of ±0.1A, covering the charge and discharge voltage and current ranges of common batteries. Temperature acquisition utilizes high-precision thermocouples with a measurement range of -40°C to 150°C and an accuracy of ±0.5°C, enabling real-time monitoring of battery temperature changes under different operating conditions. The internal pressure sensor has a measurement range of 0kPa-500kPa with an accuracy of ±2kPa, capable of sensing subtle fluctuations in internal battery pressure. Electrode material deformation measurement utilizes a laser displacement sensor with a measurement range of 0mm-5mm and an accuracy of ±0.001mm, accurately capturing minute electrode deformations. The setting of these parameters ensures the comprehensiveness and accuracy of the collected data, providing a reliable data basis for subsequent algorithm processing and multi-field coupling analysis. Its significance lies in providing the original basis for the operation of the entire system. Without high-quality data collection, all subsequent analyses will lose accuracy.

[0019] The specific implementation process is as follows: First, determine parameters such as the sampling frequency and measurement range based on the battery type and test purpose. For example, when testing the safety performance of a power battery under fast charging conditions, set the sampling frequency to 8kHz to account for the rapid changes in current and voltage during fast charging. Then, connect the voltage and current sensors to the positive and negative terminals of the battery, ensuring a secure connection and good contact to avoid data errors caused by poor contact. Thermocouples are evenly applied to different locations on the battery surface, including near the positive and negative terminals, and in the middle of the battery. A thermocouple is also implanted inside the battery through a tiny hole to measure the temperature of the core area. An internal pressure sensor is installed at the battery's filling port or a specially reserved pressure measurement port to ensure communication between the sensor and the battery's internal space. A laser displacement sensor is aligned with the battery's electrode tabs and casing surface, and the sensor's position and angle are adjusted so that the laser beam shines perpendicularly on the measurement point. Next, the data acquisition program is activated to simultaneously collect voltage, current, temperature, internal pressure, and electrode material deformation data while the battery undergoes different operating conditions, such as charge and discharge cycles, high and low temperature shock, and vibration testing. During the data collection process, data changes are monitored in real time. If any abnormal data jump occurs, the sensor connection status and battery operating status are immediately checked. If a sensor failure occurs, it is promptly replaced. If a battery anomaly occurs, the time and operating conditions of the anomaly are recorded. The collected data is sorted by timestamp and stored in the system database. A data backup is generated every 5 minutes to prevent data loss. Throughout the entire process, the continuity and stability of data collection are ensured, with no data interruptions from the start of collection to the end of the test. For large-capacity battery testing, segmented storage is used, with each hour's data stored as a separate file to facilitate subsequent data retrieval and processing.

[0020] The algorithm processing module is connected to the data acquisition module and the multi-field coupling module respectively. After receiving the data transmitted by the data acquisition module, the algorithm processing module performs feature extraction and time series modeling on the data through the SE-Res-LSTM algorithm, and transmits the processed feature parameters to the multi-field coupling module; Specifically, the algorithm processing module utilizes the SE-Res-LSTM algorithm, which combines the advantages of the SE attention mechanism and the Res-LSTM network. The SE attention mechanism's compression ratio is set to 16, effectively highlighting key features. The Res-LSTM network has three hidden layers, each with 256 neurons. This configuration enhances the network's processing and feature extraction capabilities for time series data. The input data dimension is determined by the data type of the data acquisition module and includes multiple dimensions, such as voltage, current, temperature, pressure, and deformation. After algorithmic processing, the output feature parameters have a dimension of 128, encompassing key battery status information at different moments. The module's technical parameters are designed to enable the algorithm to more accurately extract valuable features from the raw data. This is crucial for dimensionality reduction and feature extraction of the massive amount of raw data collected, removing redundant information while retaining key features. This provides a more concise and effective input for subsequent multi-field coupling analysis, improving the efficiency and accuracy of the overall system analysis and enabling the multi-field coupling module to perform more in-depth calculations based on these refined feature parameters.

[0021] The specific implementation process is as follows: First, the raw data transmitted by the data acquisition module is received. This data is presented as a time series and contains the voltage, current, temperature, internal pressure, and electrode material deformation values ​​at each moment. The raw data is then preprocessed to remove significant noise. This preprocessing does not include methods such as normalization; it only removes outliers. For example, if the voltage value at a certain moment exceeds the normal measurement range by ±5%, the data at that moment is considered abnormal and removed. Linear interpolation is used to fill the gaps in the removed data to ensure the continuity of the time series. Next, the processed raw data is input into the SE-Res-LSTM algorithm. Initial feature extraction is performed on the data in the first layer of the Res-LSTM network. The neurons in this layer use the tanh activation function to perform a nonlinear transformation on the input data, extracting low-level features. The output of the first layer is then fed into the second layer, which also uses the tanh activation function to extract more complex features. The third layer uses the sigmoid activation function to integrate and optimize the features extracted by the first two layers. After each layer of the Res-LSTM network, the SE attention mechanism is introduced to process the feature maps output by that layer. First, a global average pooling operation is performed to obtain the average feature value for each channel. Then, two fully connected layers and a sigmoid activation function are used to calculate the weight of each channel. These weights are multiplied by the corresponding feature map to enhance the expression of important features and suppress irrelevant features. After processing through the three layers of the Res-LSTM network and the SE attention mechanism, 128-dimensional feature parameters are obtained. These feature parameters contain the dynamic characteristics of the battery's time series changes and key status information. Finally, these feature parameters are arranged in chronological order and transmitted to the multi-field coupling module. Intermediate data and final feature parameters are stored in the system's cache for subsequent analysis and verification. The time delay of the entire processing process is controlled within 100ms to meet the needs of real-time analysis.

[0022] The multi-field coupling module is connected to the algorithm processing module and the parameter analysis module respectively. After receiving the characteristic parameters, it constructs a coupled electrochemical-thermal-mechanical multi-field equation to perform coupled calculations on the electrochemical reaction, heat transfer and structural stress distribution inside the battery, and transmits the calculation results to the parameter analysis module; Specifically, the multi-field coupling module is centered around coupled electrochemical-thermal-mechanical multi-field equations. The reaction rate constant involved in the electrochemical equation is set to 0.001-0.01s⁻¹, which can reflect the speed of electrochemical reactions at different temperatures and concentrations. The thermal conductivity parameter of the heat conduction equation is set according to different battery materials. The thermal conductivity of the positive electrode material is 1-3W / (m・K), the negative electrode material is 5-10W / (m・K), and the electrolyte is 0.5-1W / (m・K). These parameters determine the speed of heat transfer inside the battery. The elastic modulus parameter in the mechanical stress equation is 100-200GPa for the positive electrode material, 50-100GPa for the negative electrode material, and 200-300GPa for the shell material. The Poisson's ratio is set to 0.2-0.3 to describe the deformation characteristics of the material. The technical parameter settings of this module are determined based on the physical and chemical properties of battery materials. Its significance lies in linking the three physical fields of electrochemistry, thermal and mechanical inside the battery by constructing multi-field coupling equations, thereby realizing a comprehensive analysis of the complex interactions within the battery. This overcomes the limitations of traditional single physical field analysis, can more comprehensively reflect the true working state of the battery, and provides key theoretical support for accurately evaluating battery safety performance.

[0023] The specific implementation process is as follows: First, the 128-dimensional feature parameters transmitted by the receiving algorithm processing module contain key information such as the battery's voltage, current, temperature, etc. at different times, and use them as the initial input conditions for multi-field coupling calculations. Next, a coupled electrochemical-thermal-mechanical multi-field equation is constructed. The electrochemical equation considers the insertion and deintercalation of lithium ions into and out of the positive and negative electrode materials. The electrode potential is calculated based on the Nernst equation, and the interfacial reaction rate is calculated based on the Butler-Volmer equation. The reaction rate constant is dynamically adjusted based on the current temperature and lithium ion concentration. When the temperature increases by 10°C, the reaction rate constant increases by approximately 1.5-2 times. The heat conduction equation combines the heat generated by the electrochemical reaction with Joule heating, and calculates the heat conduction within the battery based on Fourier's law. Convective heat transfer between the battery and the external environment is also considered. The convection heat transfer coefficient is set to 5-20 W / (m²・K) and is adjusted according to the ambient wind speed. The higher the wind speed, the greater the convection heat transfer coefficient. The mechanical stress equation considers the thermal expansion caused by temperature changes and the material volume change caused by lithium ion insertion / deintercalation. The stress-strain relationship is established based on Hooke's law to calculate the stress distribution within the battery. The volume expansion rate of the positive electrode material during lithium ion insertion is 2%-5%, and that of the negative electrode material is 10%-20%. These volume changes generate corresponding stresses. Next, the multi-field coupled equations were solved using the finite element method. The battery model was divided into 10,000-50,000 mesh cells. The mesh size was adjusted based on the structural complexity of the battery, with a finer mesh in the electrode active material region and a coarser mesh in simpler structural areas such as the housing. The solution time step was set to 0.1-1 second, dynamically adjusting the time step based on the speed of the battery state changes, using a smaller time step when the battery is in a rapid charge and discharge state and a larger time step when it is in a stable state. During the solution process, multi-field coupled iterations were performed, using the heat generated by the electrochemical calculations as the heat source for the heat conduction equation. The temperature calculation results were fed back into the electrochemical equations to influence the reaction rate constants. The strains generated by temperature changes and electrochemical expansion were used as inputs to the mechanical stress equations. The stress calculation results in turn influenced the electrical and thermal conductivity of the material. Convergence was checked every 10 iterations, and the calculation was considered converged when the difference between the calculation results of two consecutive iterations was less than 1e-5. Finally, the converged calculation results, including data such as lithium ion concentration, temperature distribution, and stress magnitude at different locations, are organized into a structured data format and transmitted to the parameter analysis module. At the same time, information such as the grid model and iteration records during the calculation process are stored for subsequent result verification and model optimization. While ensuring accuracy, the entire calculation process controls the calculation time within 5-10 minutes to meet the real-time requirements of the system.

[0024] The parameter analysis module is connected to the multi-field coupling module and the optimization and control module respectively. After receiving the calculation results, it quantitatively analyzes different parameters of battery safety performance, including thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage and battery cycle life attenuation rate, and transmits the parameter thresholds obtained from the analysis to the optimization and control module; Specifically, the parameter analysis module mainly conducts quantitative analysis on different parameters of battery safety performance. The analysis range of the critical temperature for thermal runaway is set to 60℃-300℃. Different types of batteries have different critical values. For example, the critical temperature for thermal runaway of ternary lithium batteries is relatively low, while that of lithium iron phosphate batteries is relatively high. The analysis range of the maximum allowable pressure is 100kPa-400kPa, which is determined by the battery shell material and structural strength. The analysis range of the fatigue limit of the electrode material is 100MPa-300MPa, which is related to the type of electrode material and preparation process. The analysis range of the electrolyte decomposition rate is 0.001-0.1mg / (s・cm²), which is greatly affected by temperature and voltage. The analysis range of the diaphragm breakdown voltage is 5V-20V, which depends on the material and thickness of the diaphragm. The analysis range of the battery cycle life attenuation rate is 0.1%-5% / 100 cycles, and the attenuation rate varies at different cycle stages. The technical parameter settings of this module cover the key indicators of battery safety performance. Its significance lies in clarifying the boundary conditions and threshold range of battery safety performance through quantitative analysis of these parameters, providing a specific basis for judging whether the battery is in a safe state. It also provides goals and directions for parameter adjustments of the optimization control module, making subsequent optimization adjustments more targeted and effective.

[0025] The specific implementation process is as follows: First, the calculation results transmitted by the multi-field coupling module are received, including detailed data such as temperature, pressure, lithium ion concentration, and stress at various points within the battery. This data is stored in a three-dimensional array containing spatial coordinates and time information. Next, the critical temperature parameter for thermal runaway is analyzed. A time-varying curve of the maximum internal temperature of the battery is extracted from the temperature data. By analyzing the slope and inflection points of the curve, the critical temperature at which thermal runaway occurs is determined. For example, when the temperature exceeds 200°C and the heating rate exceeds 5°C / s, the critical state of thermal runaway is determined and recorded as the threshold for this parameter. For the maximum allowable pressure parameter, the distribution of maximum values ​​in the pressure data is analyzed. Combined with the material strength parameters of the battery housing, the maximum pressure the housing can withstand is calculated. When the internal pressure exceeds 80% of this value, a warning threshold is set; if it exceeds this value, a danger threshold is set. To analyze the fatigue limit of the electrode material, the stress amplitude and cycle number of the electrode material during cycling are calculated based on the stress data. Fatigue cumulative damage theory is used to determine the stress limit at which fatigue failure occurs. A fatigue warning is issued when the calculated stress exceeds 90% of this limit. For the electrolyte decomposition rate parameters, the electrolyte decomposition rates at different locations are calculated based on the temperature and voltage data, combined with the chemical properties of the electrolyte. The area percentage of the decomposition rate exceeding 0.01 mg / (s・cm²) is statistically analyzed. When the percentage exceeds 10%, it is considered an electrolyte decomposition anomaly. When analyzing the diaphragm breakdown voltage parameters, the voltage difference on both sides of the diaphragm is determined based on the voltage distribution data. Combined with the diaphragm breakdown voltage parameters, when the voltage difference exceeds 90% of the breakdown voltage, it is set as the diaphragm breakdown warning threshold. When analyzing the battery cycle life attenuation rate, the capacity attenuation of each cycle is calculated based on the capacity data at different cycle numbers, and a curve of the attenuation rate changing with the cycle number is fitted to determine the attenuation rate threshold at different cycle stages. When the attenuation rate exceeds 3% after 100 cycles, it is determined to be an abnormal life attenuation. Finally, all the parameter thresholds obtained from the analysis are sorted to form a threshold list of different parameters, including the critical temperature for thermal runaway and the maximum allowable pressure. Each parameter includes two levels: warning threshold and danger threshold. This threshold list is transmitted to the optimization and control module. At the same time, the analysis process, calculation basis and original data of different parameters are stored in the system database for subsequent query and tracing. The entire analysis process requires the processing and calculation of a large amount of data. Parallel computing is used to improve efficiency and ensure that a comprehensive parameter analysis is completed within 10 minutes.

[0026] The optimization and control module is connected to the parameter analysis module and the result output module respectively. After receiving the parameter threshold, it dynamically adjusts the operating parameters during the battery test, including charge and discharge rate, ambient temperature, external load and test duration, through an adaptive fluctuation optimization algorithm, and transmits the adjusted parameters and corresponding safety performance evaluation results to the result output module; Specifically, the optimization and control module utilizes an adaptive fluctuation optimization algorithm with a fluctuation coefficient set between 0.1 and 0.5 to control the fluctuation amplitude of operating parameters; a decay coefficient set between 0.01 and 0.1 to determine how quickly the fluctuation amplitude decays with the number of iterations; and a convergence accuracy threshold set between 1e-4 and 1e-3 to determine whether the algorithm has converged. Adjustable operating parameters include charge and discharge rate (ranging from 0.1C to 5C); ambient temperature (ranging from -20°C to 60°C); external load (ranging from 0N to 1000N); and test duration (ranging from 1 to 100 hours). These technical parameters enable the algorithm to adjust operating parameters within a reasonable range. Their significance lies in dynamically optimizing battery test parameters based on threshold information provided by the parameter analysis module. This allows the testing process to fully examine the battery's safety performance under various extreme conditions while avoiding battery damage caused by improper parameter settings. This improves test efficiency and accuracy, ensuring the reliability and validity of test results.

[0027] The specific implementation process is as follows: First, the parameter analysis module receives a list of safety performance parameter thresholds, including warning and danger thresholds for parameters such as thermal runaway critical temperature and maximum allowable pressure. These thresholds are used as constraints for optimization and control. Next, the operating parameters are initialized. Based on the battery type and test standard, the initial charge and discharge rate is set to 1C, the ambient temperature is 25°C, the external load is 100N, and the test duration is 24 hours. The value ranges for each parameter are also determined, for example, with a minimum charge and discharge rate of 0.1C and a maximum of 5C. Next, an adaptive fluctuation optimization algorithm is initiated. The battery safety performance evaluation value under the current operating parameters is calculated. This evaluation value comprehensively considers the distance between each safety parameter and the threshold. When a parameter approaches the warning threshold, the evaluation value decreases; when it approaches the danger threshold, the evaluation value drops significantly. Based on the evaluation value, a fluctuation coefficient for the operating parameter is generated. The fluctuation coefficient is initially set to 0.3. When the evaluation value is high, the fluctuation coefficient is increased to explore more optimal parameter combinations. When the evaluation value is low, the fluctuation coefficient is reduced to fine-tune the parameters. When generating new operating parameters, the charge and discharge rate fluctuations are limited to ±10% of the current value, the ambient temperature fluctuations to ±5°C, the external load fluctuations to ±50N, and the test duration fluctuations to ±2 hours. All parameters must be within the specified range. The new operating parameters are applied to the battery test system. After a period of operation, the parameter analysis module retrieves the updated safety performance parameters, recalculates the evaluation values, and compares them with the previous evaluation values. If the new evaluation values ​​are higher, the parameter adjustment is effective and the parameter combination is retained and used as the basis for the next iteration. If the evaluation values ​​are lower, the parameter combination is discarded and the fluctuations are regenerated for adjustment. As the number of iterations increases, the attenuation coefficient gradually reduces the fluctuation amplitude by 0.05, stabilizing the parameter adjustment. When the change in the evaluation values ​​for five consecutive iterations is less than the convergence accuracy threshold of 1e-3, the algorithm is considered converged, and the current operating parameter combination is determined to be the optimal parameter combination. Finally, the optimized parameters such as charge and discharge rate, ambient temperature, external load, test duration, and the corresponding safety performance evaluation results are transmitted to the result output module. At the same time, the parameter change curve, evaluation value change trend and other information during the entire optimization process are recorded and stored in the system for subsequent analysis. The number of iterations of the entire optimization process is controlled between 20 and 50 times to ensure that the optimization and control are completed within 2 hours.

[0028] The result output module is connected to the optimization and control module, and after receiving the transmitted parameters and evaluation results, presents the battery safety performance test results in the form of data tables and three-dimensional visual models.

[0029] Specifically, the result output module is a key component of system-user interaction, and the settings of its technical parameters directly affect users' understanding and use of battery safety performance test results. The module's data tables use a resolution of 1920×1080 pixels, clearly displaying the specific values, units, and changing trends of different parameters, ensuring that users do not encounter ambiguity or omissions when viewing. The mesh accuracy of the 3D visualization model can be adjusted between 0.1mm and 1mm. For complex battery components such as electrodes and separators, a 0.1mm mesh accuracy is used to accurately present details, while for relatively simple components such as the battery housing, a 1mm mesh accuracy is used to improve rendering efficiency. Supported output file formats include Excel, PDF, and STL. Excel facilitates secondary data analysis and processing, PDF is suitable for reporting and archiving, and STL can be used for printing 3D models and further structural analysis. The data update frequency is consistent with the parameter analysis module, updating every 10 minutes, allowing users to promptly understand the latest status of battery safety performance. The significance of this module is to convert the complex data processed by the system into an intuitive and easy-to-understand form, allowing users to quickly grasp the safety performance of the battery, providing a reliable reference for battery design optimization, production quality control and application scenario planning, while also facilitating data sharing and collaboration between different departments.

[0030] The specific implementation process is as follows: First, the optimized operating parameters and corresponding safety performance evaluation results transmitted by the optimization control module are received. These data include specific values ​​of operating parameters such as charge and discharge rate, ambient temperature, external load, and test duration, as well as thresholds and current values ​​of safety performance parameters such as thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage, and battery cycle life attenuation rate. These data are sorted and classified according to the preset data structure to ensure data integrity and accuracy. Then, a data table is generated. The column headers of the table include parameter name, unit, current value, warning threshold, danger threshold, change trend, etc. The change trend is indicated by arrow symbols or color codes. For example, when the current value of the parameter is within the normal range, it is indicated by green, when it is close to the warning threshold, it is indicated by yellow, and when it exceeds the danger threshold, it is indicated by red. The rows of the table are divided into two categories according to the parameter category, namely operating parameters and safety performance parameters. Under each category, the parameters are arranged in order according to the specific parameter name. During the table generation process, automatic alignment and border settings are used to make the table structure clear and easy to read. Next, a three-dimensional visualization model is constructed. Based on the battery internal structure data transmitted by the multi-field coupling module and the calculation results of the parameter analysis module, a three-dimensional model of the battery is constructed using three-dimensional modeling software. Different components in the model are distinguished by different colors, such as the positive electrode material is represented by red, the negative electrode material is represented by blue, and the electrolyte is represented by transparent color; the safety performance parameter values ​​of key positions, such as the highest temperature point, the maximum stress point, etc., are marked on the model, and cloud maps of different colors are used to display the distribution of parameters such as temperature, pressure, and stress. The color of the cloud map gradually transitions from blue to red, indicating the change of parameter values ​​from low to high; the rendering quality of the model is set to high to ensure that the details of the model are clearly visible. At the same time, according to user needs, the model can be rotated, scaled, and translated to view the parameter distribution inside the battery from different angles. Afterwards, the generated data table and 3D visualization model will be converted and saved in the file format selected by the user. During the conversion process, the data format is ensured to be correct and the clarity of the chart is not affected. For files in Excel format, data validity verification will be set to prevent users from entering incorrect data; for files in PDF format, bookmarks and directories will be added to facilitate users to quickly locate the required content; for model files in STL format, mesh repair and optimization will be performed to ensure the integrity and printability of the model.Finally, the generated file is stored in the system's output folder, and the file's save path and name are displayed on the user interface. File preview and download functions are also provided. Users can view the file's contents by clicking the preview button and save the file to the local device by clicking the download button. Every 10 minutes, the system automatically repeats the above process, updating the data table and 3D visualization model based on the latest parameter analysis results, and overwriting the previously generated files to ensure that users always obtain the latest test results. During the entire process, the system will record the file's generation time and version information for subsequent traceability and management.

[0031] Preferably, the feature extraction process of the SE-Res-LSTM algorithm in the algorithm processing module satisfies the following model formula: , ,in, is the feature vector extracted by SE-Res-LSTM algorithm at time t, is the original data input by the data acquisition module at time t, is the weight parameter of the LSTM network, For the squeeze-and-excitation attention mechanism operation, is the Sigmoid activation function, is the global average pooling operation, is the weight parameter of the attention mechanism; the electrochemical reaction sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the conductivity of the electrode material, is the solid phase potential, is the interfacial current density, is the solid phase capacitance, For time, is the electrolyte conductivity, is the liquid phase potential, is the liquid phase capacitor.

[0032] Specifically, the SE-Res-LSTM algorithm feature extraction process of the algorithm processing module and the electrochemical reaction sub-equation of the multi-field coupling module have clear technical parameters and implementation logic. Their significance lies in laying the foundation for subsequent analysis through precise feature extraction and electrochemical calculations. In specific implementation, the algorithm processing module receives the original data transmitted by the data acquisition module. These data include multi-dimensional information such as voltage, current, and temperature. The weight parameters in the SE-Res-LSTM algorithm will be dynamically adjusted according to the data characteristics. The weight parameters of the attention mechanism are obtained through a specific calculation method, which can highlight the feature vectors that have a significant impact on the safety performance of the battery. The dimension of the feature vector is set between 64-256 according to the complexity of the data to ensure that the extracted features are both comprehensive and targeted. When the multi-field coupling module processes the electrochemical sub-equation, the conductivity of the electrode material is set to 1. 00-500S / m, the electrolyte conductivity is 0.1-1S / m, the solid phase capacitance and liquid phase capacitance are set to 10-50F / m² and 5-20F / m² respectively. Through the precise setting of these parameters, the solid phase potential and liquid phase potential at different positions, as well as the distribution of interfacial current density are calculated in combination with the equations. The finite difference method is used for numerical solution in the calculation process. The time step is controlled at 0.01-0.1 seconds and the space step is 0.1-1 mm to ensure that the dynamic process of the electrochemical reaction is accurately simulated, providing reliable electrochemical basic data for subsequent thermal field and mechanical field coupling calculations.

[0033] Preferably, the heat transfer sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module satisfies: ,in, is the battery material density, is the specific heat capacity, is the temperature, is the thermal conductivity, is the heat generation rate of the electrochemical reaction, is the Joule heat generation rate; the parameter adjustment model of the adaptive fluctuation optimization algorithm in the optimization control module is: ,in, For the After the adjustment, the working parameters is the working condition parameter after the kth adjustment, is the coefficient of fluctuation, is a standard normally distributed random number, are the maximum and minimum values ​​of the working condition parameters, is the attenuation coefficient, The number of adjustments.

[0034] Specifically, the heat transfer sub-equation of the multi-field coupling module and the adaptive fluctuation optimization algorithm parameter adjustment model of the optimization control module achieve accurate simulation of the battery heat transfer process and dynamic optimization of operating parameters by setting reasonable technical parameters. Its significance lies in improving the system's ability to assess battery thermal safety and the optimization efficiency of test parameters. During implementation, the battery material density in the heat transfer sub-equation is set to 2000-3000kg / m³, the specific heat capacity is 800-1200J / (kg・K), and the thermal conductivity varies depending on the material, with electrode materials being 1-5W / (m・K) and electrolytes being 0.5-1W / (m・K). The electrochemical reaction heat generation rate and Joule heat generation rate are calculated based on current density and resistance. The finite element method is used to solve the heat conduction equation during the calculation process, with a mesh division accuracy of 0.5-2 mm, boundary conditions set to natural convection or forced convection, and a convection heat transfer coefficient of 5-20W / (m²・K). This accurately simulates the temperature distribution and changes inside the battery; in the parameter adjustment model of the optimization control module, the fluctuation coefficient is set to 0.1-0.5, the attenuation coefficient is 0.01-0.1, the maximum and minimum values ​​of the operating parameters are determined according to the battery type, the charge and discharge rate range is 0.1-5C, and the ambient temperature is -20-60℃. Each adjustment is made based on the standard normal distribution random number to generate the fluctuation amount, and the attenuation coefficient is combined to gradually reduce the fluctuation amplitude to ensure that the operating parameters can explore the optimal value and maintain stable convergence during the optimization process. The adjusted parameters are applied to the battery test in real time through system feedback to improve the accuracy and efficiency of the test.

[0035] Preferably, the structural stress sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the stress tensor, is the material density, is the body force vector, is the displacement vector, is the elasticity matrix, is the strain tensor; the quantitative analysis of the critical temperature of battery thermal runaway in the parameter analysis module satisfies: ,in, is the thermal runaway critical temperature, is the battery volume, is the initial temperature.

[0036] Specifically, the multi-field coupling module structural stress sub-equation and the parameter analysis module quantitative analysis of the critical temperature of thermal runaway achieve accurate assessment of the critical state of battery structural stress and thermal runaway by setting the material mechanical parameters and temperature calculation parameters. Its significance lies in providing a quantitative basis for judging the mechanical safety and thermal safety of the battery. During the implementation process, the material density in the structural stress sub-equation is set to 2500-3500kg / m³, the volume force vector is determined according to gravity and external loads, the elastic matrix parameters are set according to the material properties, the elastic modulus of the positive electrode material is 100-200GPa, the negative electrode material is 50-100GPa, and the Poisson's ratio is 0.2-0.3. By solving the stress tensor and strain tensor, the normal stress and shear stress distribution inside the battery are obtained. The calculation is performed using finite element software, the unit type selects the solid unit, the mesh size is 0.1-1 mm, and the boundary conditions are set to fixed constraints or displacement constraints to simulate different Battery structural deformation under working conditions; when calculating the critical temperature of thermal runaway, the parameter analysis module sets the battery volume to 100-5000 cm³ according to the actual size, and the initial temperature to 25-30°C. The total heat is obtained by integrating the heat generated by the electrochemical reaction and the Joule heat. The temperature change is calculated based on the density and specific heat capacity of the material. When the temperature rise rate exceeds 5°C / s and an uncontrollable growth trend appears, it is determined that the critical state of thermal runaway has been reached, and the temperature value at this time is recorded as the critical temperature. The time step during the entire calculation process is 1-10 seconds to ensure that the key nodes of temperature mutation can be captured, providing an accurate threshold for battery thermal safety assessment.

[0037] Preferably, the quantitative analysis model for the maximum allowable pressure of the battery in the parameter analysis module is: ,in, is the maximum allowable pressure, is the inner radius of the battery shell, is the yield strength of the shell material, is the shell thickness, is the initial pressure inside the battery; the time series modeling loss function of the SE-Res-LSTM algorithm in the algorithm processing module is: ,in, is the loss value, is the sample size, is the true value at time t, is the predicted value at time t, is the regularization coefficient, is the number of network layers, is the network weight of the i-th layer.

[0038] Specifically, the parameter analysis module's maximum allowable pressure quantitative analysis model and the algorithm processing module's SE-Res-LSTM algorithm time series modeling loss function, by setting structural parameters and loss function parameters, realize the evaluation of the battery's mechanical bearing capacity and the optimization of the algorithm model. Its significance lies in improving the system's judgment accuracy on the battery's mechanical safety and the accuracy of the algorithm's prediction. During implementation, the inner radius of the battery shell in the maximum allowable pressure analysis model is set to 5-50 mm, the shell thickness is 0.5-5 mm, the shell material yield strength is determined according to the material type, 100-300 MPa for aluminum alloy shells, 300-600 MPa for steel shells, and the internal initial pressure is 10-50 kPa. By calculating the shell's bearing capacity and combining the internal initial pressure, the maximum allowable pressure value is determined. When the internal pressure reaches 80% of the maximum allowable pressure, the early warning mechanism is triggered; in the time series modeling loss function of the algorithm processing module, the number of samples is determined according to the test duration. The sampling frequency is determined by the training set, which is usually 1000-10000. The regularization coefficient is set to 0.001-0.01, and the number of network layers is 3-5. The square difference between the predicted value and the true value is calculated, and the weight regularization term is added to avoid overfitting of the model. The gradient descent method is used to optimize the loss function during training. The learning rate is 0.001-0.01 and the number of iterations is 1000-5000 times until the loss value stabilizes in the range of 0.01-0.1, ensuring that the algorithm model can accurately predict the time series changes of battery safety performance parameters and provide reliable prediction results for subsequent analysis.

[0039] Preferably, the convergence judgment condition of the adaptive fluctuation optimization algorithm in the optimization control module is: ,in, is the safety performance evaluation function, is the convergence accuracy threshold; the quantitative analysis of the battery cycle life attenuation rate in the parameter analysis module satisfies: ,in, is the cycle life attenuation rate, is the battery capacity at the nth cycle, is the initial capacity, is the attenuation coefficient, is the number of cycles, is the decay index.

[0040] Specifically, the convergence judgment conditions of the adaptive fluctuation optimization algorithm in the optimization and control module and the quantitative analysis of the battery cycle life attenuation rate in the parameter analysis module, by setting the convergence accuracy and attenuation parameters, can achieve effective control of the optimization process and accurate evaluation of the battery life. Its significance lies in ensuring the efficient convergence of the optimization algorithm and the accurate grasp of the battery life attenuation trend. During the implementation process, the safety performance evaluation function in the convergence judgment condition is calculated based on a comprehensive calculation of multiple safety parameters, with a value range of 0-1. The closer to 1, the better the safety performance. The convergence accuracy threshold is set to 0.001-0.01. When the relative change of the evaluation function value of two consecutive iterations is less than the threshold, the algorithm is judged to have converged and the parameter adjustment is stopped. The operating parameters at this time are the optimal solution; when the parameter analysis module calculates the cycle life decay rate, the initial capacity is set to 10-100Ah according to the battery specifications, the decay coefficient is 0.0001-0.001, and the decay index is 0.5-1.5. By fitting the capacity change curve under different cycle numbers, the capacity decay of each cycle is calculated. When the cycle number reaches 100-1000 times, the decay rate in this range is counted. When the decay rate exceeds 5% / 100 cycles, the battery life decay is judged to be abnormal. During the entire analysis process, the least squares method is used to fit the capacity decay curve, and the fitting error is controlled within 5%, ensuring that the assessment of the battery life decay trend is accurate and reliable, providing a basis for battery service life prediction and maintenance.

[0041] Preferably, the multi-field coupling module includes an electrochemical field calculation unit, a temperature field simulation unit, a mechanical stress analysis unit, and a multi-field coupling coordination unit; the electrochemical field calculation unit receives the characteristic parameters transmitted by the algorithm processing module, calculates the lithium ion concentration distribution, electrode potential and interface reaction rate at different positions based on the electrode reaction kinetic equation, meshes the calculation area through the finite element method, and stores the calculation results of each grid node in a temporary data buffer; the temperature field simulation unit obtains the electrochemical field calculation results from the temporary data buffer, calculates the temperature gradient distribution inside and on the surface of the battery according to the heat conduction equation and the convection heat transfer model, and combines the heat generation of the electrochemical reaction, the contact thermal resistance and the environmental heat dissipation Conditions are set, the temperature field is transiently solved and the data buffer is updated; the mechanical stress analysis unit calls the temperature field and electrochemical field data in the data buffer, calculates the expansion / contraction of the electrode material caused by temperature changes and lithium ion insertion / deinsertion based on the thermoelasticity theory, establishes a stress balance equation to solve the normal stress, shear stress and strain distribution inside the battery, and writes the results into the data buffer; the multi-field coupling coordination unit couples and iterates the calculation results of the electrochemical field, temperature field and mechanical stress field in the data buffer, corrects the calculation parameters of each field through the field variable association equation, repeats the iteration until the convergence error of the calculation results of each field meets the preset threshold, and transmits the final coupled calculation results to the parameter analysis module.

[0042] Specifically, the multi-field coupling module, which includes the electrochemical field calculation unit, temperature field simulation unit, mechanical stress analysis unit and multi-field coupling coordination unit, involves multiple key technical parameters during operation, which is of great significance for the system to accurately analyze the multi-field effects inside the battery. In the electrochemical field calculation unit, the size of the finite element grid is set according to the details of the battery structure. The grid size of the electrode active material area is 0.05-0.2mm, and the current collector area is 0.5-1mm, so as to ensure the balance between calculation accuracy and efficiency. When the unit is calculated by the electrode reaction kinetics equation, the reaction rate constant ranges from 0.001-0.01s⁻¹, and the lithium ion diffusion coefficient is 1e-14-1e-10m² / s. These parameters directly affect the accuracy of the calculation results such as lithium ion concentration distribution and electrode potential. The implementation method is to receive the characteristic parameters of the algorithm processing module, first perform grid division, and then substitute the equation to calculate the data of each grid node and store it; the thermal conductivity parameter of the temperature field simulation unit is set according to different materials, 1-3W / (m・K) for positive electrode materials, 5-10W / (m・K) for negative electrode materials, and 0.5-1W / ( The unit acquires electrochemical data from the buffer zone and then performs a transient solution based on the heat conduction and convection heat transfer models, updating the temperature field data. In the mechanical stress analysis unit, the elastic modulus of the material is set at 100-200 GPa for the positive electrode and 50-100 GPa for the negative electrode, and the thermal expansion coefficient is 1e-6-1e-5 / °C. The deformation and stress distribution caused by temperature and lithium-ion insertion / deinsertion are calculated using thermoelasticity theory and the results are written to the buffer zone. The convergence error threshold of the multi-field coupling coordination unit is set at 1e-5-1e-4. Field variable correlation equations are used to repeatedly correct each field parameter until the error requirements are met, and the final result is transmitted to the parameter analysis module. This entire process achieves precise coupling of multiple field effects, providing comprehensive basic data for subsequent safety performance analysis.

[0043] Preferably, the parameter analysis module includes a thermal safety parameter calculation unit, a mechanical safety parameter evaluation unit, an electrochemical safety parameter analysis unit, and a comprehensive safety performance judgment unit; the thermal safety parameter calculation unit receives the temperature field data transmitted by the multi-field coupling module, calculates the maximum temperature inside the battery, the temperature change rate and the critical temperature of thermal runaway, analyzes the heat accumulation rate and heat diffusion path in different areas, determines the thermal safety warning parameters and stores them in the analysis database; the mechanical safety parameter evaluation unit obtains the stress field data from the multi-field coupling module, calculates the maximum stress, strain value and fatigue damage accumulation of the battery shell and electrode material, evaluates the load-bearing capacity and shape of the structural components The mechanical safety parameters and their thresholds are written into the analysis database; the electrochemical safety parameter analysis unit calculates the electrolyte decomposition rate, diaphragm breakdown voltage, electrode polarization degree and lithium ion concentration distribution uniformity based on the electrochemical field data of the multi-field coupling module, analyzes the internal short circuit risk of the battery and the electrochemical performance attenuation law, and stores the electrochemical safety parameters in the analysis database; the comprehensive safety performance judgment unit calls the thermal safety, mechanical safety and electrochemical safety parameters in the analysis database, calculates the comprehensive safety factor of the battery through a multi-parameter weighted analysis method, determines the safety threshold range of different parameters, and transmits the parameter threshold obtained by analysis to the optimization control module.

[0044] Specifically, the parameter analysis module, its technical parameter settings and implementation methods of the thermal safety parameter calculation unit, mechanical safety parameter evaluation unit, electrochemical safety parameter analysis unit and comprehensive safety performance judgment unit are of key significance to the accurate quantification of the safety performance boundary of the battery. In the thermal safety parameter calculation unit, the spatial resolution of temperature acquisition is 0.5-2mm, and the time interval is 1-5s, so as to ensure that subtle temperature changes can be captured. After obtaining the temperature field data from the multi-field coupling module, the unit calculates the maximum temperature, heating rate and thermal runaway critical temperature, analyzes the heat accumulation rate and diffusion path, and stores the thermal safety warning parameters in the database. The judgment basis of the thermal runaway critical temperature includes conditions such as the temperature exceeds 200°C and the heating rate exceeds 5°C / s; the stress calculation accuracy of the mechanical safety parameter evaluation unit is controlled at ±1MPa, and the fatigue damage accumulation calculation adopts the Miner law. When the damage degree exceeds 0.8, a warning is issued. After obtaining the stress field data, the unit calculates the maximum stress, strain and fatigue damage amount, and evaluates the results. The load-bearing capacity of the structure is determined, and the mechanical safety parameters and thresholds are written into the database; in the electrochemical safety parameter analysis unit, the calculation error of the electrolyte decomposition rate is controlled at ±0.001mg / (s・cm²), and the detection accuracy of the diaphragm breakdown voltage is ±0.1V. Based on the electrochemical field data, this unit calculates parameters such as the electrolyte decomposition rate and diaphragm breakdown voltage, analyzes the short-circuit risk and performance degradation law, and stores them; the comprehensive safety performance judgment unit adopts weighted analysis. The weight of each parameter is set according to the battery type, and the weights of thermal safety, mechanical safety, and electrochemical safety parameters account for 40%, 30%, and 30% respectively. Different parameters in the database are called to calculate the comprehensive safety factor, determine the safety threshold range, and transmit it to the optimization and control module, providing a clear basis for adjusting the operating parameters.

[0045] Preferably, the optimization and control module includes an operating condition parameter initialization unit, a fluctuation parameter generation unit, a performance feedback adjustment unit, and an optimization result determination unit; the operating condition parameter initialization unit sets the initial value and value range of the operating condition parameters of the initial charge and discharge rate, ambient temperature, external load and test time according to the battery type and test requirements, and establishes a mapping relationship model between the operating condition parameters and the safety performance parameters; the fluctuation parameter generation unit generates random fluctuations within the operating condition parameter value range based on an adaptive fluctuation optimization algorithm, adjusts the fluctuation amplitude and direction in combination with the current safety performance evaluation results, and generates new operating condition parameter candidate values; the performance feedback adjustment unit inputs the new operating condition parameter candidate values ​​into the battery testing system, obtains the corresponding safety performance parameter change data, calculates the performance improvement and compares it with the preset threshold. If the requirements are met, the candidate value is retained, otherwise the fluctuation parameter is adjusted to regenerate the candidate value; the optimization result determination unit screens the operating condition parameter candidate values ​​obtained through multiple iterations, selects the operating condition parameter combination that optimizes the safety performance parameters, and transmits the adjusted parameters and the corresponding safety performance evaluation results to the result output module.

[0046] Specifically, the technical parameters and implementation methods of the optimization and control module, its working condition parameter initialization unit, fluctuation parameter generation unit, performance feedback adjustment unit and optimization result determination unit are crucial to achieving dynamic optimization of battery test conditions and improving test efficiency and accuracy. In the working condition parameter initialization unit, the initial value of the charge and discharge rate is set to 0.5-2C, the initial value of the ambient temperature is 25℃±2℃, the initial value of the external load is 0-500N, and the initial value of the test time is 12-48 hours. The value range is determined according to the battery type. For example, the charge and discharge rate range of the power battery is 0.1-5C. The unit sets the initial parameters and range according to the test requirements and establishes a mapping model between the working condition and the safety performance parameters; the initial value of the fluctuation coefficient of the fluctuation parameter generation unit is 0.1-0.5, which gradually decreases with the number of iterations at an attenuation coefficient of 0.01-0.1. The generated working condition parameter candidate value must be within the value range. The unit is based on adaptive fluctuation optimization. The algorithm is used to adjust the fluctuation amplitude and direction in combination with the current safety performance evaluation results to generate new candidate values; the performance improvement threshold of the performance feedback adjustment unit is set to 5%-10%. After the candidate value is input into the test system, the safety performance parameter change data is obtained, the improvement is calculated and compared with the threshold, and if it does not meet the requirements, the candidate value is regenerated; the optimal parameter screening standard of the optimization result determination unit is that the comprehensive safety factor is the highest and the different parameters are all within the safety threshold. The candidate values ​​of multiple iterations are screened, and the optimal combination is selected and transmitted to the result output module. The entire process ensures that the test can fully cover the safety performance boundary by dynamically adjusting the operating parameters, while avoiding unnecessary over-testing.

[0047] The SE-Res-LSTM algorithm in this invention is a deep learning algorithm that combines the SE attention mechanism with the Res-LSTM network. Its core lies in enhancing the expression of key information through multi-layer feature extraction and attention mechanism. The full name of the SE-Res-LSTM algorithm is the "Squeeze-and-Excitation Residual Long Short-Term Memory" algorithm. Among them, "Squeeze-and-Excitation" (SE for short) refers to an attention mechanism used to perform weighted adjustment on feature channels and enhance the expression of important features; "Residual" (Res for short) represents residual connection, which alleviates the gradient vanishing problem during deep network training by introducing cross-layer connections in the network; "Long Short-Term Memory" (LSTM for short) is a special recurrent neural network that can effectively process long-term dependencies in time series data. This algorithm combines these three elements to form a deep learning model suitable for feature extraction from multi-dimensional battery time series data. This model retains the LSTM's ability to capture time series information while improving deep network training through residual connections. It also leverages the SE attention mechanism to enhance key features, thereby more accurately extracting effective safety-related information from battery operating data. Specifically, the algorithm comprises a three-layer Res-LSTM network, each with 256 neurons. It uses tanh and sigmoid activation functions for nonlinear transformations, respectively. It also incorporates a SE attention mechanism with a compression ratio of 16. Global average pooling and fully connected layers are used to calculate the weights of each feature channel, enhancing important features and suppressing irrelevant ones. The implementation process involves receiving raw data from the data acquisition module, removing outliers, and filling gaps to ensure time series continuity. This processed data is then fed into the Res-LSTM network for preliminary feature extraction. The SE attention mechanism then weights the output features of each layer, ultimately generating 128-dimensional feature parameters. This algorithm extracts key features from massive amounts of multi-dimensional battery time-series data, capturing the dynamic patterns and hidden information within the data and providing streamlined and effective input for subsequent multi-field coupling analysis. Its significance lies in overcoming the limitations of traditional feature extraction methods, improving the ability to perceive complex battery state changes, and enabling the system to more accurately grasp the key factors affecting battery safety performance, laying the data analysis foundation for the efficient operation of the entire test system.

[0048] The coupled electrochemical-thermomechanical multi-field equations are the core mathematical model used in this invention to describe the complex physicochemical interactions within the battery. They integrate the coupled relationships between three key physical fields: electrochemical reactions, heat transfer, and structural stresses. The electrochemical component uses the Nernst and Butler-Volmer equations to describe lithium ion migration and interfacial reactions, with the reaction rate constants dynamically adjusting with temperature and concentration. The heat conduction component calculates heat transfer using Fourier's law and a convective heat transfer model, accounting for both electrochemical heat generation and Joule heating. The mechanical stress component uses Hooke's law and thermoelasticity to analyze material deformation and stress distribution caused by temperature changes and lithium ion insertion / deinsertion. The implementation process involves receiving the characteristic parameters from the algorithm processing module, creating a finite element mesh of 10,000-50,000 elements, setting parameters such as thermal conductivity and elastic modulus based on the battery material characteristics, and then solving the multi-field equations using the characteristic parameters as initial conditions. Iterative corrections are performed to achieve coupled feedback between the various fields until the results converge (with a convergence error threshold of 1e-5-1e-4). This equation comprehensively reveals the interplay between electrochemical reactions, heat transfer, and structural stresses within a battery, quantifying the spatiotemporal distribution of each physical field. Its significance lies in transcending the limitations of traditional single-field analysis, enabling comprehensive simulation of complex battery operating conditions. This provides theoretical support for accurately identifying battery safety hazards and assessing safety performance, enabling the system to more realistically reflect the battery's safety status.

[0049] The adaptive fluctuation optimization algorithm is an optimization method used in the present invention to dynamically adjust the battery test operating parameters. It realizes precise control of the test process through adaptive parameter fluctuation and feedback adjustment. The algorithm sets the fluctuation coefficient to 0.1-0.5 to control the parameter adjustment amplitude, the attenuation coefficient to 0.01-0.1 to gradually reduce the fluctuation amplitude with the number of iterations, and the convergence accuracy threshold to 1e-4-1e-3 to determine whether the optimization is completed. The implementation process is as follows: first, the operating parameters such as charge and discharge rate, ambient temperature and the value range are initialized according to the battery type and test requirements, and a mapping model between the operating parameters and safety performance is established; then, based on the current safety performance evaluation results, the parameter fluctuation amount is generated, and new candidate parameters are generated within the value range; then, the candidate parameters are applied to the test system, the safety performance change data is obtained and the performance improvement is calculated. If the threshold of 5%-10% is not reached, the candidate value is regenerated; finally, the parameter combination with the highest comprehensive safety factor is screened out through multiple iterations until the convergence conditions are met. This algorithm dynamically optimizes test conditions based on real-time feedback on battery safety performance, ensuring that the test process fully covers safety performance boundaries while avoiding battery damage caused by improper parameter settings. Its significance lies in improving test efficiency and targeting, overcoming the limitations of traditional empirical parameter settings, ensuring the reliability and effectiveness of test results, and providing more valuable test data for battery design optimization and safety protection.

[0050] like Figure 2 As shown, a battery safety performance testing system based on artificial intelligence includes the following steps: Step S1: Using a data acquisition module, the voltage, current, temperature, internal pressure, and electrode material deformation data of the battery under charge and discharge cycles, high and low temperature shock, and vibration loading conditions are collected at a preset sampling frequency, and the collected data are sorted in time series and abnormal jump data are removed; Step S2: Input the sorted data into the algorithm processing module, perform multi-layer feature extraction on the data using the SE-Res-LSTM algorithm, build a time series prediction model to fit the changing trend of the battery safety performance parameters, and output a feature parameter matrix; Step S3: Input the characteristic parameter matrix into the multi-field coupling module to construct a coupled electrochemical-thermal-mechanical multi-field equation. After setting boundary conditions and initial conditions, numerical solution is performed to obtain the spatiotemporal variation data of the electrochemical reaction rate, temperature distribution, and stress distribution inside the battery. Step S4: input the spatiotemporal variation data into a parameter analysis module, quantitatively calculate the thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage, and battery cycle life attenuation rate, and determine the numerical range and correlation of different parameters; Step S5: Input the parameter value range and correlation relationship into the optimization control module, perform multiple rounds of dynamic adjustment on the charge and discharge rate, ambient temperature, external load and test duration through the adaptive fluctuation optimization algorithm, and record the changes in the safety performance parameters after each adjustment; Step S6: Input the adjusted operating parameters and safety performance evaluation results into the result output module, generate a test report containing the values ​​of various safety performance parameters, change curves and three-dimensional distribution models according to the preset data format, and store them in the form of an editable file.

[0051] An artificial intelligence-based battery safety performance testing system demonstrates significant advantages in multi-field coupling analysis, effectively compensating for the shortcomings of traditional single-physics field models. By constructing electrochemical-thermal-mechanical multi-field coupling equations through the multi-field coupling module and deeply mining multi-dimensional data using the SE-Res-LSTM algorithm in the algorithm processing module, the system can accurately capture the interaction mechanisms between electrochemical reactions, heat transfer, and structural stresses within the battery. Compared to the limitations of existing technologies that can only analyze a single physical field, this system can comprehensively present the patterns of safety performance changes under multi-field coupling, thereby accurately identifying potential safety hazards within the battery and providing more comprehensive technical support for predicting chain reactions such as thermal runaway and structural failure.

[0052] In terms of optimizing and adjusting operating parameters, the system uses an adaptive fluctuation optimization algorithm through the optimization and control module to achieve dynamic adjustment of operating parameters during the test process, overcoming the shortcomings of traditional technologies that rely on experience or simple algorithms to adjust parameters. The system can dynamically optimize operating parameters such as charge and discharge rate, ambient temperature, external load, and test duration based on the real-time safety performance parameters output by the parameter analysis module. This not only improves test efficiency but also covers the safety performance boundaries under complex operating conditions. This dynamic optimization capability makes test results closer to actual application scenarios and provides a more valuable reference basis for battery design optimization and safety protection.

[0053] Through the collaborative work of various modules, the system as a whole forms a complete battery safety performance testing system, far exceeding existing technologies in terms of overall performance. The data acquisition module ensures the comprehensive acquisition of multi-dimensional parameters, the algorithm processing module enables in-depth analysis of data, the parameter analysis module completes the quantitative evaluation of safety performance parameters, and the result output module presents test results in an intuitive form. The close cooperation of various modules not only addresses the shortcomings of traditional testing systems in multi-field analysis and parameter optimization, but also realizes a comprehensive and dynamic evaluation of battery safety performance, meeting the increasingly demanding requirements of the new energy industry for battery safety performance testing and providing strong support for the development of battery safety technology.

[0054] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0055] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A battery safety performance testing system based on artificial intelligence, characterized in that: It includes a data acquisition module, an algorithm processing module, a multi-field coupling module, a parameter analysis module, an optimization and control module and a result output module; the data acquisition module is connected to the algorithm processing module, and is used to collect the voltage, current, temperature, internal pressure and electrode material deformation data of the battery under different working conditions in real time, and transmit the collected data to the algorithm processing module; the algorithm processing module is respectively connected to the data acquisition module and the multi-field coupling module, and after receiving the data transmitted by the data acquisition module, the data is feature extracted and time series modeled by the SE-Res-LSTM algorithm, and the characteristic parameters obtained by processing are transmitted to the multi-field coupling module; the multi-field coupling module is respectively connected to the algorithm processing module and the parameter analysis module, and after receiving the characteristic parameters, a coupled electrochemical-thermal-mechanical multi-field equation is constructed to couple the electrochemical reaction, heat transfer and structural stress distribution inside the battery to perform coupled calculations, and the calculation results are transmitted to the parameter analysis module; The parameter analysis module is respectively connected to the multi-field coupling module and the optimization and control module. After receiving the calculation results, it quantitatively analyzes different parameters of battery safety performance, including the critical temperature of thermal runaway, the maximum allowable pressure, the fatigue limit of the electrode material, the electrolyte decomposition rate, the diaphragm breakdown voltage and the battery cycle life attenuation rate, and transmits the parameter thresholds obtained by analysis to the optimization and control module; the optimization and control module is respectively connected to the parameter analysis module and the result output module. After receiving the parameter thresholds, it dynamically adjusts the operating parameters during the battery test, including the charge and discharge rate, the ambient temperature, the external load and the test duration, through the adaptive fluctuation optimization algorithm, and transmits the adjusted parameters and the corresponding safety performance evaluation results to the result output module; the result output module is connected to the optimization and control module. After receiving the transmitted parameters and evaluation results, it presents the battery safety performance test results in the form of data tables and three-dimensional visualization models.

2. The battery safety performance testing system based on artificial intelligence according to claim 1, characterized in that: The feature extraction process of the SE-Res-LSTM algorithm in the algorithm processing module satisfies the following model formula: , ,in, is the feature vector extracted by SE-Res-LSTM algorithm at time t, is the original data input by the data acquisition module at time t, is the weight parameter of the LSTM network, For the squeeze-and-excitation attention mechanism operation, is the Sigmoid activation function, is the global average pooling operation, is the weight parameter of the attention mechanism; the electrochemical reaction sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the conductivity of the electrode material, is the solid phase potential, is the interfacial current density, is the solid phase capacitance, For time, is the electrolyte conductivity, is the liquid phase potential, is the liquid phase capacitor.

3. The battery safety performance testing system based on artificial intelligence according to claim 1, characterized in that: The heat transfer sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module satisfies: ,in, is the battery material density, is the specific heat capacity, is the temperature, is the thermal conductivity, is the heat generation rate of the electrochemical reaction, is the Joule heat generation rate; the parameter adjustment model of the adaptive fluctuation optimization algorithm in the optimization control module is: ,in, For the After the adjustment, the working parameters is the working condition parameter after the kth adjustment, is the coefficient of fluctuation, is a standard normally distributed random number, are the maximum and minimum values ​​of the working condition parameters, is the attenuation coefficient, The number of adjustments.

4. The battery safety performance testing system based on artificial intelligence according to claim 1, characterized in that: The structural stress sub-equation of the coupled electrochemical-thermal-mechanical multi-field equation in the multi-field coupling module is: ,in, is the stress tensor, is the material density, is the body force vector, is the displacement vector, is the elasticity matrix, is the strain tensor; the quantitative analysis of the critical temperature of battery thermal runaway in the parameter analysis module satisfies: ,in, is the thermal runaway critical temperature, is the battery volume, is the initial temperature.

5. The battery safety performance testing system based on artificial intelligence according to claim 1, characterized in that: The quantitative analysis model for the maximum allowable pressure of the battery in the parameter analysis module is: ,in, is the maximum allowable pressure, is the inner radius of the battery shell, is the yield strength of the shell material, is the shell thickness, is the initial pressure inside the battery; the time series modeling loss function of the SE-Res-LSTM algorithm in the algorithm processing module is: ,in, is the loss value, is the sample size, is the true value at time t, is the predicted value at time t, is the regularization coefficient, is the number of network layers, is the network weight of the i-th layer.

6. The battery safety performance testing system based on artificial intelligence according to claim 1, characterized in that: The convergence judgment condition of the adaptive fluctuation optimization algorithm in the optimization and control module is: ,in, is the safety performance evaluation function, is the convergence accuracy threshold; the quantitative analysis of the battery cycle life attenuation rate in the parameter analysis module satisfies: ,in, is the cycle life attenuation rate, is the battery capacity at the nth cycle, is the initial capacity, is the attenuation coefficient, is the number of cycles, is the decay index.

7. The artificial intelligence-based battery safety performance testing system according to claim 1, characterized in that: The multi-field coupling module includes an electrochemical field calculation unit, a temperature field simulation unit, a mechanical stress analysis unit and a multi-field coupling coordination unit; the electrochemical field calculation unit receives the characteristic parameters transmitted by the algorithm processing module, calculates the lithium ion concentration distribution, electrode potential and interface reaction rate at different positions based on the electrode reaction kinetic equation, meshes the calculation area through the finite element method, and stores the calculation results of each grid node in a temporary data buffer; the temperature field simulation unit obtains the electrochemical field calculation results from the temporary data buffer, calculates the temperature gradient distribution inside and on the surface of the battery according to the heat conduction equation and the convection heat transfer model, and combines the heat generation of the electrochemical reaction, the contact thermal resistance and the environmental heat dissipation conditions , perform transient solution on the temperature field and update the data buffer; the mechanical stress analysis unit calls the temperature field and electrochemical field data in the data buffer, calculates the expansion / contraction of the electrode material caused by temperature change and lithium ion insertion / deinsertion based on the thermoelasticity theory, establishes a stress balance equation to solve the normal stress, shear stress and strain distribution inside the battery, and writes the result into the data buffer; the multi-field coupling coordination unit couples and iterates the calculation results of the electrochemical field, temperature field and mechanical stress field in the data buffer, corrects the calculation parameters of each field through the field variable association equation, repeats the iteration until the convergence error of the calculation results of each field meets the preset threshold, and transmits the final coupling calculation result to the parameter analysis module.

8. The artificial intelligence-based battery safety performance testing system according to claim 1, characterized in that: The parameter analysis module includes a thermal safety parameter calculation unit, a mechanical safety parameter evaluation unit, an electrochemical safety parameter analysis unit, and a comprehensive safety performance determination unit; the thermal safety parameter calculation unit receives the temperature field data transmitted by the multi-field coupling module, calculates the maximum temperature inside the battery, the temperature change rate, and the thermal runaway critical temperature, analyzes the heat accumulation rate and heat diffusion path in different areas, determines the thermal safety warning parameters, and stores them in the analysis database; The mechanical safety parameter evaluation unit obtains stress field data from the multi-field coupling module, calculates the maximum stress, strain value and fatigue damage accumulation of the battery housing and electrode material, evaluates the load-bearing capacity and deformation limit of the structural components, and writes the mechanical safety parameters and their thresholds into the analysis database; The electrochemical safety parameter analysis unit calculates the electrolyte decomposition rate, diaphragm breakdown voltage, electrode polarization degree and lithium ion concentration distribution uniformity based on the electrochemical field data of the multi-field coupling module, analyzes the internal short circuit risk of the battery and the electrochemical performance attenuation law, and stores the electrochemical safety parameters in the analysis database; the comprehensive safety performance judgment unit calls the thermal safety, mechanical safety and electrochemical safety parameters in the analysis database, calculates the battery comprehensive safety factor through a multi-parameter weighted analysis method, determines the safety threshold range of different parameters, and transmits the parameter threshold obtained from the analysis to the optimization and control module.

9. The artificial intelligence-based battery safety performance testing system according to claim 1, characterized in that: The optimization and control module includes an operating parameter initialization unit, a fluctuation parameter generation unit, a performance feedback adjustment unit, and an optimization result determination unit; the operating parameter initialization unit sets the initial values ​​and value ranges of the operating parameters of the initial charge and discharge rate, ambient temperature, external load, and test duration according to the battery type and test requirements, and establishes a mapping relationship model between the operating parameters and the safety performance parameters; The fluctuation parameter generation unit generates random fluctuations within the operating condition parameter value range based on an adaptive fluctuation optimization algorithm, adjusts the fluctuation amplitude and direction based on the current safety performance evaluation results, and generates new candidate values ​​for the operating condition parameters; The performance feedback adjustment unit inputs new operating parameter candidate values ​​into the battery testing system, obtains corresponding safety performance parameter change data, calculates the performance improvement and compares it with a preset threshold. If the requirements are met, the candidate value is retained; otherwise, the fluctuation parameter is adjusted to regenerate the candidate value; the optimization result determination unit screens the operating parameter candidate values ​​obtained through multiple iterations, selects the operating parameter combination that optimizes the safety performance parameters, and transmits the adjusted parameters and the corresponding safety performance evaluation results to the result output module.

10. The artificial intelligence-based battery safety performance testing system according to any one of claims 1 to 9, characterized in that: The system operation includes the following steps: Step S1: Using a data acquisition module, the voltage, current, temperature, internal pressure, and electrode material deformation data of the battery under charge and discharge cycles, high and low temperature shock, and vibration loading conditions are collected at a preset sampling frequency, and the collected data are sorted in time series and abnormal jump data are removed; Step S2: Input the sorted data into the algorithm processing module, perform multi-layer feature extraction on the data using the SE-Res-LSTM algorithm, build a time series prediction model to fit the changing trend of the battery safety performance parameters, and output a feature parameter matrix; Step S3: Input the characteristic parameter matrix into the multi-field coupling module to construct a coupled electrochemical-thermal-mechanical multi-field equation. After setting boundary conditions and initial conditions, numerical solution is performed to obtain the spatiotemporal variation data of the electrochemical reaction rate, temperature distribution, and stress distribution inside the battery. Step S4: input the spatiotemporal variation data into a parameter analysis module, quantitatively calculate the thermal runaway critical temperature, maximum allowable pressure, electrode material fatigue limit, electrolyte decomposition rate, diaphragm breakdown voltage, and battery cycle life attenuation rate, and determine the numerical range and correlation of different parameters; Step S5: Input the numerical range and correlation of the parameters into the optimization control module, perform multiple rounds of dynamic adjustments on the charge and discharge rate, ambient temperature, external load, and test duration through the adaptive fluctuation optimization algorithm, and record the changes in the safety performance parameters after each adjustment; Step S6: Input the adjusted operating parameters and safety performance evaluation results into the result output module, generate a test report containing the values ​​of various safety performance parameters, change curves and three-dimensional distribution models according to the preset data format, and store them in the form of an editable file.

Images