Battery capacity performance test method, device and storage medium

Abstract

The application discloses a battery capacity performance test method and device and a storage medium, and relates to the technical field of battery capacity performance test.The application solves the problems of difficult balance between precision and efficiency, large data acquisition error and fixed test protocol in traditional battery test by constructing a dynamic load model and a multi-load test protocol, combining a high-precision data acquisition and filtering algorithm, adaptive data analysis and a machine learning regression model;by dynamically adjusting test parameters and environmental conditions, the complex working conditions of the battery in actual use are accurately simulated, the dynamic update and optimization of the test protocol are realized, and the test precision, data reliability and test efficiency are significantly improved, so that a comprehensive, efficient and adaptable solution is provided for battery performance evaluation and health management.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of battery capacity performance testing, and particularly relates to a battery capacity performance testing method, device and storage medium. BACKGROUND

[0002] In the battery capacity performance testing process, the existing method has a significant technical contradiction: the balance between testing accuracy and testing efficiency is difficult to achieve; the traditional testing method usually adopts a single load or a fixed load mode, which is simple to operate but cannot accurately simulate the complex working conditions of the battery in actual use, resulting in a large deviation between the test results and the real performance, especially in the evaluation of battery performance decay, the single load test cannot capture the specific performance of the battery at different use stages, such as the discharge characteristics of the battery in a low power state or the instantaneous response capability of the battery in a high load condition.

[0003] Another technical problem is that the correlation between the test environment and the test results is insufficient, and the battery may face various environmental conditions in actual application, such as temperature changes, humidity fluctuations and different charging and discharging rates. However, the traditional testing method is usually carried out in a constant environment, and the testing parameters cannot be dynamically adjusted to adapt to different environmental conditions, resulting in a lack of adaptability of the test results to the actual application scenarios. SUMMARY

[0004] The purpose of the application is to provide a battery capacity performance testing method, device and storage medium, which realizes the balance between battery testing accuracy and efficiency, optimizes the testing protocol and significantly improves the reliability and adaptability of battery performance evaluation through a dynamic load model, high-precision data acquisition and analysis and a machine learning algorithm.

[0005] The purpose of the application can be achieved through the following technical solutions:

[0006] The application provides a battery capacity performance testing method, which comprises the following steps:

[0007] The load demand data of the battery at different use stages, including the discharge characteristics in a low power state and the instantaneous response capability in a high load condition, are acquired, a dynamic load model is constructed, a multi-load testing protocol is designed according to the dynamic load model, the complex working conditions of the battery in actual use are simulated, and different discharge rates and load changes are covered;

[0008] In the testing process, the voltage, current and temperature data of the battery are collected in real time, high-precision sensors and filtering algorithms are used to reduce the data acquisition error, and the temperature, humidity and charging and discharging rate of the test environment are adjusted according to the preset environmental condition parameters to simulate the performance of the battery in different application scenarios;

[0009] When the test environment temperature is lower than a preset threshold, a low-temperature compensation algorithm is started to correct the capacity attenuation data of the battery in a low-temperature environment.

[0010] An adaptive data analysis algorithm is adopted to dynamically adjust test parameters according to real-time collected battery performance data, to optimize the precision and efficiency of the test process, to analyze the battery capacity attenuation trend through a regression model in a machine learning algorithm, and to predict the specific performance of the battery in different use stages.

[0011] According to the prediction result, a battery performance evaluation report is generated, including discharge characteristics in a low-power state, high-load instantaneous response capability and environmental adaptability analysis, and the evaluation report is combined with a dynamic load model to update the test protocol.

[0012] Further, load demand data of the battery in different use stages is acquired, including discharge characteristics in a low-power state and instantaneous response capability in a high-load state, and a dynamic load model is constructed, specifically including: acquiring voltage values and current values of the battery in different states through a battery management system, and recording time values as time stamps;

[0013] For the low-power state, voltage values and current values changes in the discharge process are collected, and discharge capacity data is calculated, and for the high-load state, instantaneous current values and voltage values changes are collected, and response value data is calculated;

[0014] The collected voltage values, current values, discharge capacity, response values and time values are associated to generate a characteristic value data set, and a linear regression algorithm is adopted to fit the characteristic value data set to obtain a discharge characteristic curve and an instantaneous response curve;

[0015] According to the discharge characteristic curve and the instantaneous response curve, a neural network algorithm is adopted to construct a dynamic load model, and when the model precision does not reach a preset threshold, a support vector machine algorithm is adopted to optimize the model.

[0016] Further, according to the dynamic load model, a multi-load test protocol is designed to simulate complex working conditions of the battery in actual use, covering different discharge rates and load changes, specifically including:

[0017] Historical running data of the battery in actual application scenarios is acquired, load change parameters and discharge rate parameters are extracted, a dynamic load model is established, a plurality of test scenarios are set according to the dynamic load model, each scenario includes at least three load change periods, and load change amounts and discharge rate values of each period are determined;

[0018] The random forest algorithm is used to simulate the test scene, adjust the load change parameter and the discharge rate parameter, generate a multi-load test protocol, and execute the multi-load test protocol in the simulated scene. The performance data of the battery under different load changes and discharge rates are collected, including voltage, current and temperature parameters;

[0019] The support vector machine algorithm is used to analyze the collected performance data, judge the performance of the battery under different complex working conditions, identify abnormal working conditions, and optimize the parameter settings in the dynamic load model according to the performance analysis results, adjust the corresponding relationship between the load change amount and the discharge rate value;

[0020] The optimized dynamic load model is fused with the multi-load test protocol to generate a complete test scheme covering the actual use scene.

[0021] Further, in the test process, the voltage, current and temperature data of the battery are collected in real time, high-precision sensors and filtering algorithms are used to reduce data acquisition errors, specifically including:

[0022] In the battery test process, the voltage, current and temperature data are obtained in real time by high-precision sensors, and the collected data are processed according to the preset filtering algorithm to obtain denoised voltage, current and temperature data;

[0023] The Kalman filtering algorithm is used to further process the denoised voltage, current and temperature data to obtain smoothed voltage, current and temperature data. The power data of the battery are calculated according to the smoothed voltage and current data, and the working state of the battery is judged in combination with the temperature data;

[0024] Wherein, when the working state of the battery is in an abnormal range, an alarm mechanism is triggered and abnormal data are recorded;

[0025] According to the recorded abnormal data and historical data, the health status of the battery is evaluated and predicted using the support vector machine algorithm, and the remaining life of the battery is predicted through the linear regression algorithm to obtain the remaining service life of the battery. The working state, health status and remaining service life data of the battery are stored in the database.

[0026] Further, by adjusting the temperature, humidity and charge / discharge rate of the test environment through preset environmental condition parameters, the performance of the battery in different application scenarios is simulated, specifically including:

[0027] The preset temperature value, humidity value, charge rate and discharge rate are obtained as the basic parameters of the environmental conditions, and a dynamic adjustment scheme of the test environment is generated according to the temperature value, humidity value, charge rate and discharge rate.

[0028] The dynamic adjustment scheme is adopted to adjust the environment value of the test environment to obtain a target test condition, and the charge-discharge test of the battery is performed under the target test condition to obtain a test value of the battery.

[0029] According to the test value, the performance value of the battery is calculated in combination with a preset scene value, and when the performance value meets a preset threshold value, the applicability of the battery in a specific application scene is determined.

[0030] Through the mapping relationship between the performance value and the scene value, a simulation result of the battery in different application scenes is generated.

[0031] Further, an adaptive data analysis algorithm is adopted to dynamically adjust the test parameters according to the real-time collected battery performance data, so as to optimize the precision and efficiency of the test process, specifically including:

[0032] The battery performance data is transmitted to a data processing module for feature extraction, the adaptive algorithm is used to analyze the battery performance characteristics, a performance evaluation model is established, and a data fluctuation trend is determined.

[0033] When the data fluctuation exceeds a preset threshold value, a dynamic adjustment mechanism is started, the optimal test parameters are calculated, the test equipment operating state is adjusted according to the optimal test parameters, and new battery performance data is obtained.

[0034] The new data is input into the performance evaluation model to determine whether the test precision meets a preset standard, and when the test precision does not meet the standard, the adaptive algorithm parameters are updated, the optimal test parameters are recalculated, and the test efficiency and precision are balanced until the optimal balance is achieved.

[0035] Further, through a regression model in a machine learning algorithm, the battery capacity attenuation trend is analyzed, and the specific performance of the battery in different use stages is predicted, specifically including:

[0036] The historical capacity data of the battery in different use stages is obtained, data cleaning and preprocessing are performed, and a feature extraction method is used to extract key feature parameters of battery attenuation from the preprocessed data.

[0037] According to the extracted feature parameters, a linear regression model and a random forest regression model are constructed, and when the battery capacity data has a nonlinear feature, the random forest regression model is used for training, otherwise the linear regression model is used for training.

[0038] The trained regression model is used to predict the battery capacity to obtain a capacity prediction value of a future use stage, the remaining service life of the battery is determined according to the capacity prediction value, and a battery performance evaluation report is output.

[0039] Further, according to the prediction result, a battery performance evaluation report is generated, including low power state discharge characteristics, high load transient response capability and environmental adaptability analysis, and the evaluation report is combined with the dynamic load model to update the test protocol, specifically including:

[0040] Real-time operation data of the battery is acquired, voltage and current parameters in the low power state are extracted, and discharge characteristic indexes are calculated through a preset discharge model;

[0041] Based on the current change curve under the high load condition, the stability and fluctuation range of the transient response capability are judged by using the time series analysis method, a multi-dimensional data matrix is established for different environmental temperature and humidity conditions, and the comprehensive score of environmental adaptability is evaluated by using the principal component analysis method;

[0042] The discharge characteristics, response capability and adaptability evaluation results are integrated to generate a battery performance evaluation report, structured data storage is formed, load change characteristic parameters are extracted from the dynamic load model, and matching degree calculation is performed on the performance indexes in the evaluation report;

[0043] When the matching degree is lower than a preset threshold, a regression algorithm is used to optimize the parameter setting of the test protocol, the matching degree of the model and the measured data is improved, after updating the test protocol, the new parameter configuration is written into the test system, and the test protocol updating process is completed.

[0044] A battery capacity performance test device, comprising a processor, a memory and computer program instructions stored in the memory, when the computer program instructions are executed by the processor, the above-mentioned battery capacity performance test method is realized.

[0045] A storage medium having computer program instructions stored thereon, when the computer program instructions are executed by a processor, the above-mentioned battery capacity performance test method is realized.

[0046] The beneficial effects of the present application are:

[0047] The present application solves the technical problem that the test precision and efficiency are difficult to balance in the traditional test method by designing a dynamic load model and a multi-load test protocol, acquires load demand data of the battery in different use stages, including discharge characteristics in the low power state and transient response capability under high load conditions, constructs a dynamic load model, and designs a multi-load test protocol to simulate complex working conditions of the battery in actual use; the random forest algorithm and the neural network optimization model are combined to improve the model precision, so that the test protocol can accurately reflect the real performance of the battery under complex working conditions; not only the test precision is improved, but also the test efficiency is improved by dynamically adjusting the test parameters, and the optimal balance of precision and efficiency is realized.

[0048] By high-precision sensors and filtering algorithms, combined with adaptive data analysis algorithms, the problems of large data acquisition errors and insufficient monitoring of abnormal working conditions in traditional tests are solved. In the testing process, high-precision sensors are used to collect voltage, current and temperature data in real time, and Kalman filtering algorithm is used to denoise and smooth the data. Combined with support vector machine and linear regression algorithm, the battery health status and remaining life are evaluated. By dynamically adjusting the temperature, humidity and charge-discharge rate of the test environment, the performance of the battery in different application scenarios is simulated to ensure the reliability and adaptability of the test results. The reliability of data acquisition and analysis is significantly improved, providing more accurate data support for battery performance evaluation.

[0049] By combining the battery performance evaluation report with the dynamic load model, dynamic updating and optimization of the test protocol are realized, solving the problem of fixed test protocol in traditional testing methods that is difficult to adapt to the actual working conditions of the battery. According to the prediction results, a battery performance evaluation report is generated, including low power discharge characteristics, high load instantaneous response capability and environmental adaptability analysis. The regression algorithm is used to optimize the parameter settings of the test protocol to ensure that the test protocol is highly consistent with the actual working conditions. When the matching degree of the evaluation results and the model is lower than the preset threshold, the test protocol is automatically updated and written into the test system, forming a closed-loop optimization process. This not only improves the adaptability and accuracy of the test protocol, but also provides a dynamically optimized test scheme for battery health management. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to better understand and implement, the technical solutions of the present application are described in detail below in conjunction with the drawings.

[0051] Fig. 1 A flowchart of a battery capacity performance test method provided by the present application;

[0052] Fig. 2 A flowchart of a battery capacity performance test method provided by the present application for obtaining load demand data of the battery in different use stages;

[0053] Fig. 3 A flowchart of a battery capacity performance test method provided by the present application for reducing data acquisition errors. DETAILED DESCRIPTION

[0054] For further illustrating the technical means and effects taken by the present application to achieve the predetermined inventive purpose, hereinafter exemplary embodiments will be described in detail, which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they only describe methods and systems consistent with some aspects of the present application, as detailed in the appended claims.

[0055] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in this application and the appended claims, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0056] The specific embodiments according to the present application, features and effects thereof will be described in detail below with reference to the accompanying drawings and preferred embodiments.

[0057] Please refer to Figs. 1-3 The embodiment provides a battery capacity performance test method, comprising the following steps:

[0058] S1, acquiring load demand data of the battery in different use stages, including discharge characteristics in a low power state and instantaneous response capability in a high load condition, constructing a dynamic load model, designing a multi-load test protocol according to the dynamic load model, simulating complex working conditions of the battery in actual use, covering different discharge rates and load changes;

[0059] Further, the load demand data of the battery in different use stages is acquired, including the discharge characteristics in the low power state and the instantaneous response capability in the high load condition, and the dynamic load model is constructed, specifically comprising: S11, acquiring voltage values and current values of the battery in different states through a battery management system, and recording time values as time stamps;

[0060] S12, for the low power state, collecting voltage value and current value changes in the discharge process, and calculating discharge capacity data, for the high load state, collecting instantaneous current value and voltage value changes, and calculating response value data;

[0061] S13, associating the collected voltage value, current value, discharge capacity, response value and time value to generate a characteristic value data set, and then using a linear regression algorithm to fit the characteristic value data set to obtain a discharge characteristic curve and an instantaneous response curve;

[0062] S14, according to the discharge characteristic curve and the transient response curve, a dynamic load model is constructed by using a neural network algorithm, and when the model accuracy does not reach a preset threshold, a support vector machine algorithm is used to optimize the model.

[0063] The construction process of the dynamic load model includes: based on the voltage, current, discharge amount and response value data of the battery in different use stages, a linear regression algorithm is used to fit the discharge characteristic curve and the transient response curve, and the initial model parameters are determined; when the dynamic load model is constructed by the neural network algorithm, a three-layer fully connected network structure is used, the input layer is the voltage, current, temperature and timestamp data, the number of hidden layer nodes is 64, the activation function is ReLU, and the output layer is the load prediction value; finally, when the model accuracy does not reach a preset threshold (such as 95%), a support vector machine algorithm is used for optimization, the kernel function is selected as radial basis function (RBF), the regularization parameter C is set to 1.0, the gamma value is 0.1, the hyperparameters are adjusted through cross-validation to ensure that the model accuracy meets the requirements.

[0064] Specifically, by constructing a dynamic load model, the load demand data of the battery in different use stages is accurately obtained, including low power discharge characteristics and high load transient response capability. The characteristic curve is fitted by using linear regression and neural network algorithm and the model is established, and the model accuracy is optimized by support vector machine, which effectively solves the technical problems of low accuracy, inability to simulate complex working conditions and lack of environmental adaptability of traditional test methods, significantly improves the accuracy and practicality of battery performance test, and realizes the balance between test accuracy and efficiency.

[0065] Further, according to the dynamic load model, a multi-load test protocol is designed to simulate the complex working conditions of the battery in actual use, covering different discharge rates and load changes, specifically including:

[0066] The historical running data of the battery in the actual application scene is obtained, the load change parameters and the discharge rate parameters are extracted, the dynamic load model is established, and according to the dynamic load model, a plurality of test scenes are set, each scene at least contains three load change periods, the load change amount and the discharge rate value of each period are determined;

[0067] A random forest algorithm is used to simulate the test scene, adjust the load change parameters and the discharge rate parameters, generate a multi-load test protocol, and execute the multi-load test protocol in the simulation scene, collect the performance data of the battery under different load changes and discharge rates, including voltage, current and temperature parameters;

[0068] In the random forest algorithm, the number of trees is set to 100, the maximum depth is set to 10, and the feature selection criterion is the Gini coefficient. The hyperparameters are optimized through grid search to determine the best model parameters. Then, in the test scenario, the load variation parameters and discharge rate parameters are adjusted to generate a multi-load test protocol. Next, these protocols are executed in a simulated environment to collect performance data of the battery under different load variations and discharge rates, including voltage, current, and temperature parameters.

[0069] Using the support vector machine algorithm, the collected performance data is analyzed to determine the performance of the battery under different complex working conditions, identify abnormal working conditions, and optimize the parameter settings in the dynamic load model according to the performance analysis results, adjust the corresponding relationship between the load variation and the discharge rate value.

[0070] The optimized dynamic load model is fused with the multi-load test protocol to generate a complete test scheme that covers the actual use scenario, ensuring that the test protocol can accurately reflect the true performance of the battery under complex working conditions.

[0071] Specifically, through the multi-load test protocol design based on the dynamic load model, the technical problem that the traditional battery test method cannot accurately simulate complex working conditions is solved. The random forest algorithm is used to generate a multi-load test protocol, adjust the load variation and discharge rate parameters, and simulate the performance of the battery under different working conditions. The support vector machine algorithm is used to analyze the collected performance data, identify abnormal working conditions, and optimize the model parameters. Finally, the optimized dynamic load model is fused with the test protocol to generate a complete test scheme that covers the actual use scenario, ensuring that the test results can accurately reflect the true performance of the battery under complex working conditions. The test accuracy and efficiency are significantly improved, and the adaptability of the test results to actual applications is enhanced.

[0072] S2, during the test process, real-time collection of battery voltage, current and temperature data, using high-precision sensors and filtering algorithms to reduce data acquisition errors; through the pre-set environmental condition parameters, adjusting the temperature, humidity and charge-discharge rate of the test environment to simulate the performance of the battery in different application scenarios;

[0073] When the test environment temperature is lower than the pre-set threshold, the low-temperature compensation algorithm is started to correct the capacity decay data of the battery in a low-temperature environment.

[0074] It should be explained that the low-temperature compensation algorithm is a key technology to ensure the performance and accuracy of the battery in low-temperature environment. When the test environment temperature is lower than the preset threshold, the low-temperature compensation algorithm can monitor the current temperature of the battery in real time through the temperature sensor, and correct the capacity attenuation data of the battery according to the preset compensation model. Specifically, the algorithm can establish the mapping relationship between temperature and battery performance through polynomial fitting or neural network, so as to calculate the compensation parameters that need to be adjusted. For example, the BP neural network algorithm can be used to predict the performance change of the battery in low-temperature environment by training the neural network, and adjust the test parameters in real time. In addition, in order to improve the accuracy and response speed of the compensation, the PID control system can be combined to ensure that the heating element can quickly respond and maintain the battery in the appropriate working temperature range. Through these measures, the low-temperature compensation algorithm can effectively correct the capacity attenuation data of the battery in low-temperature environment, and ensure the accuracy and reliability of the test results.

[0075] Further, in the test process, the voltage, current and temperature data of the battery are collected in real time, high-precision sensors and filtering algorithms are used to reduce data acquisition errors, specifically including:

[0076] S21, in the battery test process, the voltage, current and temperature data are acquired in real time by high-precision sensors, and the collected data are processed according to the preset filtering algorithm to obtain the denoised voltage, current and temperature data;

[0077] S22, the denoised voltage, current and temperature data are further processed by using Kalman filtering algorithm to obtain the smoothed voltage, current and temperature data, the power data of the battery are calculated according to the smoothed voltage and current data, and the working state of the battery is judged in combination with the temperature data;

[0078] Wherein, when the working state of the battery is in the abnormal range, the alarm mechanism is triggered and the abnormal data are recorded;

[0079] S23, according to the recorded abnormal data and historical data, the health state of the battery is evaluated and predicted by using support vector machine algorithm, and the remaining life of the battery is predicted by using linear regression algorithm, the remaining service life of the battery is obtained, and the working state, health state and remaining service life data of the battery are stored in the database for subsequent analysis and decision.

[0080] It should be explained that in the Kalman filtering algorithm, the state vector includes voltage, current and temperature, the observation vector is the original data collected by the sensor, and the state transition matrix and observation matrix are calibrated by experimental data; in the linear regression algorithm, the least square method is used to fit the discharge characteristic curve, and the objective function is to minimize the sum of squares of errors.

[0081] Further, by presetting environmental condition parameters, adjusting the temperature, humidity and charge / discharge rate of the test environment, simulating the performance of the battery in different application scenarios, specifically including:

[0082] Obtain the preset temperature value, humidity value, charge rate and discharge rate as the basic parameters of the environmental conditions, and generate a dynamic adjustment scheme for the test environment according to the temperature value, humidity value, charge rate and discharge rate.

[0083] Using the dynamic adjustment scheme, adjust the environmental values of the test environment to obtain the target test conditions, and run the charge / discharge test of the battery under the target test conditions to obtain the test values of the battery;

[0084] According to the test values, combined with the preset scene values, calculate the performance values of the battery, and when the performance values meet the preset threshold, determine the applicability of the battery in the specific application scenario;

[0085] Through the mapping relationship between the performance value and the scene value, the simulation results of the battery in different application scenarios are generated.

[0086] Specifically, by using high-precision sensors and filtering algorithms to collect and process the voltage, current and temperature data of the battery in real time, the data acquisition error is effectively reduced and the data quality is improved. Kalman filtering algorithm is used to further smooth the data and monitor the battery working state in real time. Once an anomaly is detected, an alarm is triggered and the data is recorded. Based on the abnormal data and historical data, combined with support vector machine and linear regression algorithm, the battery health status and remaining life are evaluated and predicted, and the results are stored in the database to provide support for subsequent analysis and decision-making. In addition, by dynamically adjusting the temperature, humidity and charge / discharge rate of the test environment, the performance of the battery in different application scenarios is simulated to ensure the accuracy and applicability of the test results. The problems of low data acquisition accuracy, inability to monitor anomalies in real time and lack of environmental adaptability in traditional testing are solved, and the reliability and practicality of battery testing are significantly improved.

[0087] S3, using adaptive data analysis algorithm, according to the real-time collected battery performance data, dynamically adjusting the test parameters, optimizing the precision and efficiency of the test process, through the regression model in machine learning algorithm, analyzing the battery capacity attenuation trend, predicting the specific performance of the battery in different use stages.

[0088] Further, using adaptive data analysis algorithm, according to the real-time collected battery performance data, dynamically adjusting the test parameters, optimizing the precision and efficiency of the test process, specifically including:

[0089] Obtain the battery performance data and transmit it to the data processing module for feature extraction, use adaptive algorithm to analyze the battery performance characteristics, establish a performance evaluation model, and judge the data fluctuation trend;

[0090] When the data fluctuation exceeds the preset threshold, a dynamic adjustment mechanism is started, the optimal test parameters are calculated, the test equipment operating state is adjusted according to the optimal test parameters, and new battery performance data are obtained.

[0091] The new data are input into the performance evaluation model, whether the test precision reaches the preset standard is judged, when the test precision does not reach the standard, the adaptive algorithm parameters are updated, the optimal test parameters are recalculated, and the test efficiency and precision reach the optimal balance.

[0092] Specifically, through the adaptive data analysis algorithm and the machine learning regression model, the technical problems that the accuracy and efficiency are difficult to balance in traditional battery testing and the battery capacity attenuation prediction is inaccurate are solved. The adaptive algorithm can dynamically adjust the test parameters according to the real-time collected battery performance data, the data fluctuation is monitored in real time through the feature extraction and performance evaluation model, when the fluctuation exceeds the threshold, the optimal test parameters are automatically calculated and the test equipment operating state is adjusted, which not only improves the test precision, but also maximizes the test efficiency through continuous optimization.

[0093] Further, through the regression model in the machine learning algorithm, the battery capacity attenuation trend is analyzed, and the specific performance of the battery at different use stages is predicted, specifically including:

[0094] The historical capacity data of the battery at different use stages are obtained, data cleaning and preprocessing are performed, and the key feature parameters of battery attenuation are extracted from the preprocessed data by using a feature extraction method;

[0095] According to the extracted feature parameters, a linear regression model and a random forest regression model are constructed, when the battery capacity data has nonlinear characteristics, the random forest regression model is used for training, otherwise the linear regression model is used for training;

[0096] The trained regression model is used to predict the battery capacity, the capacity prediction value of the future use stage is obtained, the remaining service life of the battery is determined according to the capacity prediction value, and a battery performance evaluation report is output.

[0097] Specifically, through the linear regression and random forest regression model in machine learning, the battery capacity attenuation trend is analyzed and predicted, which can accurately predict the performance of the battery at different use stages. The key feature parameters are extracted according to the historical capacity data of the battery, and the appropriate regression model is selected for training according to the linear or nonlinear characteristics of the data, so as to accurately predict the remaining service life of the battery and output the performance evaluation report. This data-driven prediction method significantly improves the accuracy and reliability of battery performance evaluation, and provides strong support for the optimal design and health management of the battery,

[0098] S4, generating a battery performance evaluation report according to the prediction result, including discharge characteristics in a low power state, high load transient response capability and environmental adaptability analysis, and combining the evaluation report with the dynamic load model to update the test protocol, thereby further improving the accuracy and reliability of the battery capacity performance test.

[0099] Further, according to the prediction result, a battery performance evaluation report is generated, including discharge characteristics in a low power state, high load transient response capability and environmental adaptability analysis, and the evaluation report is combined with the dynamic load model to update the test protocol, specifically including:

[0100] Real-time operation data of the battery is obtained, voltage and current parameters in a low power state are extracted, and discharge characteristic indicators are calculated through a preset discharge model;

[0101] Based on the current change curve under high load conditions, a time series analysis method is used to judge the stability and fluctuation range of the transient response capability, a multi-dimensional data matrix is established for different environmental temperature and humidity conditions, and a principal component analysis method is used to evaluate the comprehensive score of environmental adaptability;

[0102] The discharge characteristics, response capability and adaptability evaluation results are integrated to generate a battery performance evaluation report, structured data is stored, load change characteristic parameters are extracted from the dynamic load model, and matching degree calculation is performed with the performance indicators in the evaluation report;

[0103] When the matching degree is lower than a preset threshold, a regression algorithm is used to optimize the parameter settings of the test protocol, the fitting degree of the model and the measured data is improved, the new parameter configuration is written into the test system after updating the test protocol, and the test protocol updating process is completed.

[0104] Specifically, by integrating the battery performance evaluation report and the dynamic load model, comprehensive analysis of the battery performance and dynamic updating of the test protocol are realized, and the technical problems of incomplete performance evaluation and difficulty in adapting the test protocol to the actual use conditions of the battery in the traditional test method are solved.

[0105] A battery capacity performance test device, comprising a processor, a memory and computer program instructions stored in the memory, which, when executed by the processor, implement the above-mentioned battery capacity performance test method.

[0106] A storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the above-mentioned battery capacity performance test method.

[0107] The above merely describes the preferred embodiments of the present application, and is not intended to limit the present application in any form. Although the present application has been disclosed with the preferred embodiments as above, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content to obtain equivalent embodiments with equivalent changes, as long as the changes or modifications do not deviate from the technical solution of the present application. Any modification, change, equivalent change and modification of the above embodiments made according to the technical essence of the present application still belong to the scope of the technical solution of the present application.

Claims

1. A method for testing battery capacity performance, characterized in that: Includes the following steps: Acquire battery load demand data at different usage stages, including discharge characteristics under low charge and instantaneous response capability under high load, construct a dynamic load model, and design a multi-load test protocol based on the dynamic load model to simulate the complex working conditions of the battery in actual use, covering different discharge rates and load changes. During the test, the battery's voltage, current, and temperature data are collected in real time. High-precision sensors and filtering algorithms are used to reduce data acquisition errors. By setting preset environmental condition parameters, the temperature, humidity, and charge / discharge rate of the test environment are adjusted to simulate the battery's performance in different application scenarios. Among them, when the test environment temperature is lower than the preset threshold, the low temperature compensation algorithm is activated to correct the capacity decay data of the battery in the low temperature environment. An adaptive data analysis algorithm is used to dynamically adjust test parameters based on real-time battery performance data, thereby optimizing the accuracy and efficiency of the test process. Through regression models in machine learning algorithms, the battery capacity degradation trend is analyzed, and the specific performance of the battery at different stages of use is predicted. Based on the prediction results, a battery performance evaluation report is generated, which includes low-charge state discharge characteristics, high-load instantaneous response capability, and environmental adaptability analysis. The evaluation report is then combined with the dynamic load model to update the test protocol.

2. The battery capacity performance testing method according to claim 1, characterized in that: Obtain battery load demand data at different usage stages, including discharge characteristics under low charge and instantaneous response capability under high load, and build a dynamic load model. Specifically, this includes: obtaining the battery voltage and current values ​​under different states through the battery management system and recording the time values ​​as timestamps. For low-charge conditions, the voltage and current values ​​during the discharge process are collected to calculate the discharge amount data. For high-load conditions, the instantaneous current and voltage values ​​are collected to calculate the response value data. The collected voltage, current, discharge, response, and time values ​​are correlated to generate a characteristic value dataset. Then, a linear regression algorithm is used to fit the characteristic value dataset to obtain the discharge characteristic curve and the instantaneous response curve. Based on the discharge characteristic curve and instantaneous response curve, a dynamic load model is constructed using a neural network algorithm. If the model accuracy does not reach the preset threshold, a support vector machine algorithm is used to optimize the model.

3. The battery capacity performance testing method according to claim 1, characterized in that: Based on the dynamic load model, a multi-load test protocol is designed to simulate the complex operating conditions of batteries in actual use, covering different discharge rates and load variations, specifically including: Obtain historical operating data of the battery in actual application scenarios, extract load change parameters and discharge rate parameters, establish a dynamic load model, set multiple test scenarios based on the dynamic load model, each scenario contains at least three load change cycles, and determine the load change amount and discharge rate value of each cycle. The random forest algorithm is used to simulate the test scenario, adjust the load change parameters and discharge rate parameters, generate a multi-load test protocol, and execute the multi-load test protocol in the simulated scenario to collect battery performance data under different load changes and discharge rates, including voltage, current and temperature parameters. The support vector machine algorithm is used to analyze the collected performance data, determine the battery's performance under different complex operating conditions, identify abnormal operating conditions, and optimize the parameter settings in the dynamic load model based on the performance analysis results, adjusting the correspondence between load change and discharge rate values. The optimized dynamic load model is integrated with multiple load testing protocols to generate a complete test plan covering real-world usage scenarios.

4. The battery capacity performance testing method according to claim 1, characterized in that: During the test, the battery's voltage, current, and temperature data are collected in real time. High-precision sensors and filtering algorithms are used to reduce data acquisition errors, specifically including: During battery testing, voltage, current and temperature data are acquired in real time through high-precision sensors. The acquired data are then processed according to a preset filtering algorithm to obtain denoised voltage, current and temperature data. The Kalman filter algorithm is used to further process the denoised voltage, current and temperature data to obtain smoothed voltage, current and temperature data. The power data of the battery is calculated based on the smoothed voltage and current data, and the working status of the battery is determined by combining the temperature data. Specifically, when the battery's operating state is outside the normal range, an alarm mechanism is triggered and abnormal data is recorded. Based on the recorded abnormal and historical data, the support vector machine algorithm is used to assess and predict the battery's health status, and then the linear regression algorithm is used to predict the battery's remaining lifespan. The remaining lifespan of the battery is then obtained, and the battery's operating status, health status, and remaining lifespan are stored in the database.

5. The battery capacity performance testing method according to claim 1, characterized in that: By adjusting the temperature, humidity, and charge / discharge rate of the test environment using preset environmental parameters, the battery performance in different application scenarios is simulated, specifically including: The system acquires preset temperature, humidity, charge rate, and discharge rate as basic parameters of the environmental conditions, and generates a dynamic adjustment scheme for the test environment based on these parameters. A dynamic adjustment scheme is adopted to adjust the environmental values ​​of the test environment to obtain the target test conditions. Under the target test conditions, the battery charge and discharge test is run to obtain the battery test values. Based on the test values ​​and the preset scenario values, the battery performance value is calculated. When the performance value meets the preset threshold, the applicability of the battery in a specific application scenario is determined. By mapping performance values ​​to scenario values, simulation results of batteries in different application scenarios are generated.

6. The battery capacity performance testing method according to claim 1, characterized in that: An adaptive data analysis algorithm is employed to dynamically adjust test parameters based on real-time battery performance data, optimizing the accuracy and efficiency of the testing process. Specifically, this includes: The acquired battery performance data is transmitted to the data processing module for feature extraction. An adaptive algorithm is used to analyze the battery performance characteristics, establish a performance evaluation model, and determine the data fluctuation trend. When data fluctuations exceed a preset threshold, a dynamic adjustment mechanism is activated to calculate the optimal test parameters, adjust the operating status of the test equipment based on the optimal test parameters, and obtain new battery performance data. New data is input into the performance evaluation model to determine whether the test accuracy meets the preset standard. If the test accuracy does not meet the standard, the adaptive algorithm parameters are updated and the optimal test parameters are recalculated until the test efficiency and accuracy reach the optimal balance.

7. The battery capacity performance testing method according to claim 1, characterized in that: By using regression models in machine learning algorithms, the battery capacity degradation trend can be analyzed, and the specific performance of the battery at different stages of use can be predicted, including: Historical capacity data of batteries at different usage stages are obtained, data is cleaned and preprocessed, and key characteristic parameters of battery degradation are extracted from the preprocessed data using feature extraction methods. Based on the extracted feature parameters, two machine learning algorithms, linear regression model and random forest regression model, are constructed. When the battery capacity data has non-linear characteristics, the random forest regression model is used for training; otherwise, the linear regression model is used for training. The battery capacity is predicted by a trained regression model, and the predicted capacity value for the future use stage is obtained. Based on the predicted capacity value, the remaining service life of the battery is determined, and a battery performance evaluation report is output.

8. The battery capacity performance testing method according to claim 1, characterized in that: Based on the prediction results, a battery performance evaluation report is generated, including analysis of low-charge state discharge characteristics, high-load instantaneous response capability, and environmental adaptability. The evaluation report is then combined with a dynamic load model to update the test protocol, specifically including: Acquire real-time operating data of the battery, extract voltage and current parameters under low charge conditions, and calculate discharge characteristic indicators through a preset discharge model; Based on the current change curve under high load conditions, time series analysis is used to determine the stability and fluctuation range of instantaneous response capability. For different ambient temperature and humidity conditions, a multi-dimensional data matrix is ​​established, and the comprehensive score of environmental adaptability is evaluated by principal component analysis. The discharge characteristics, response capability, and adaptability evaluation results are integrated to generate a battery performance evaluation report, forming structured data storage. Load change characteristic parameters are extracted from the dynamic load model and their matching degree is calculated with the performance indicators in the evaluation report. When the matching degree is lower than the preset threshold, the regression algorithm is used to optimize the parameter settings of the test protocol, improve the fit between the model and the test data, update the test protocol, write the new parameter configuration into the test system, and complete the test protocol update process.

9. A battery capacity performance testing device, characterized in that: It includes a processor, a memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the battery capacity performance testing method of claims 1-8.

10. A storage medium storing computer program instructions thereon, characterized in that, When computer program instructions are executed by a processor, a battery capacity performance testing method as described in claims 1-8 is implemented.

Images