A new energy vehicle power battery pack performance testing method and system

Abstract

The present invention relates to the field of battery performance testing, and in particular to a new energy vehicle power battery pack performance testing method and system. The method comprises the following steps: obtaining a real-time monitoring log of a power battery pack; performing multi-period feature analysis on the real-time monitoring log of the power battery pack, and performing operation scenario inference to generate battery pack operation scenarios in different time periods; extracting multi-dimensional battery pack operation status parameters based on abnormal optimization of the real-time monitoring log; performing scene state matching on battery pack operation scenarios in different time periods based on the multi-dimensional battery pack operation status parameters, thereby obtaining the operation status characteristics of each operation scenario; performing real-time operation condition analysis on the operation status characteristics of each operation scenario, and performing virtual operation condition modeling to construct a battery pack virtual operation condition model. The present invention realizes efficient and accurate power battery pack performance testing.

Description

Technical Field

[0001] The present invention relates to the technical field of battery performance testing, and in particular to a new energy vehicle power battery pack performance testing method and system. Background Art

[0002] New energy vehicles are the main direction of future automobile development, and power battery packs are their core power systems. Their performance directly affects the key performance indicators of the entire vehicle, such as cruising range, acceleration performance, safety, etc. With the rapid development of the new energy vehicle industry, how to accurately evaluate and optimize the performance of power battery packs has become an important challenge facing the industry.

[0003] Traditional power battery pack performance testing methods usually rely on manual collection of relevant working parameters, such as voltage, current, temperature, etc., and then offline analysis and evaluation. This method not only has low testing efficiency and inaccurate performance test results, but also makes it difficult to promptly discover battery pack performance risks, which brings many difficulties to subsequent maintenance and optimization work. As the new energy vehicle industry develops towards intelligence and networking, it is urgent to establish a new type of intelligent battery pack performance testing method to meet the needs of industry development. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a new energy vehicle power battery pack performance testing method and system to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides a new energy vehicle power battery pack performance testing method, comprising the following steps:

[0006] Step S1: obtaining a real-time monitoring log of a power battery pack; performing multi-period feature analysis on the real-time monitoring log of the power battery pack, and performing operation scenario inference to generate battery pack operation scenarios in different time periods;

[0007] Step S2: extracting multi-dimensional battery pack operating status parameters based on the abnormal optimization real-time monitoring log; performing scene state matching on battery pack operating scenarios in different time periods based on the multi-dimensional battery pack operating status parameters, thereby obtaining operating status characteristics of each operating scenario;

[0008] Step S3: Perform real-time operating condition analysis on the operating status characteristics of each operating scenario, perform virtual operating condition modeling, and construct a virtual operating condition model of the battery pack;

[0009] Step S4: using the battery pack virtual operating condition model to simulate the advanced accelerated evolution of the multi-dimensional battery pack operating state parameters, and mining the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend;

[0010] Step S5: performing residual quantity attenuation rate trend analysis on the acceleration evolution state change trend, and performing battery life prediction to obtain normalized battery pack life prediction data;

[0011] Step S6: Perform a probability prediction of the failure time point for the accelerated evolution state change trend, thereby generating a predicted time point for the failure probability; obtain a comprehensive performance test evaluation result of the battery pack based on the predicted time point for the failure probability and the normalized battery pack life prediction data.

[0012] The present invention obtains the real-time monitoring log of the power battery pack to understand the operating status and performance of the battery pack in real time. The acquisition of the real-time monitoring log is helpful to timely discover the problems and abnormal conditions of the battery pack for further analysis and processing. The real-time monitoring log is subjected to multi-time period feature analysis and operation scenario inference, and the battery pack operation scenarios of different time periods are generated, which is helpful to deeply understand the operation of the battery pack, extract multi-dimensional battery pack operation status parameters and perform scenario status matching, and more comprehensively describe the operation status and performance characteristics of the battery pack, so as to provide data support for subsequent analysis, and construct a virtual operation condition model of the battery pack to help simulate the operation status of the battery pack under different working conditions, and provide a virtual experimental environment for battery pack performance analysis. The establishment of the virtual operation condition model is helpful to deeply understand the response characteristics of the battery pack under different working conditions, and provide support for performance optimization and fault prediction. Conduct advanced accelerated evolution simulation and state change trend analysis on multi-dimensional battery pack operating status parameters to reveal the performance evolution law of the battery pack under different working conditions. Analyzing the battery evolution state change trend is helpful to evaluate the life and performance degradation of the battery pack, and provide a basis for battery maintenance and management. Through residual attenuation rate trend analysis and battery life prediction, the life and performance degradation of the battery pack can be inferred to provide decision support for maintenance and update. Normalized battery pack life prediction data is helpful to evaluate the long-term performance of the battery pack and help formulate reasonable maintenance plans and replacement strategies. Predicting the probability of failure time point helps to identify the timing of failure in advance, so as to take preventive measures and reduce the impact of failure on the system. Combined with the prediction of failure probability and normalized battery pack life prediction data, the performance of the battery pack is comprehensively evaluated to provide comprehensive performance test evaluation results for maintenance and update decisions.

[0013] Preferably, step S1 comprises the following steps:

[0014] Step S11: Obtaining a real-time monitoring log of a power battery pack;

[0015] Step S12: removing abnormal values ​​from the real-time monitoring log of the power battery pack to obtain an abnormal optimized real-time monitoring log;

[0016] Step S13: Calculate the real-time output power of the power battery pack based on the abnormal optimization real-time monitoring log to obtain the real-time output power;

[0017] Step S14: Perform multi-period feature analysis on the real-time output power to extract the output power in multiple time periods;

[0018] Step S15: identifying the operating status of the battery pack based on the output power in multiple time periods to obtain the operating status of the battery in different time periods;

[0019] Step S16: Inferring the operating scenarios based on the battery operating status in different time periods, and generating the battery pack operating scenarios in different time periods.

[0020] The present invention helps to understand the operating status and performance of the battery pack in real time by obtaining the real-time monitoring log of the power battery pack. The real-time monitoring log provides a data basis for analyzing the real-time operation of the battery pack. The real-time monitoring log is eliminated for outliers to improve data quality and accuracy, and a more reliable monitoring log is obtained. The abnormal optimized real-time monitoring log is helpful to remove interference data and ensure the effectiveness of subsequent analysis. Calculating the real-time output power of the power battery pack is helpful to evaluate the actual performance and output capacity of the battery pack. The real-time output power is helpful to monitor the operating status and performance indicators of the battery pack. Multi-time period feature analysis of the real-time output power reveals the output power variation law of the battery pack in different time periods. Extracting the output power of multiple time periods is helpful to understand the performance and change trend of the battery pack under different working conditions. Battery pack operating status identification of the output power of multiple time periods is helpful to judge the working status and performance of the battery pack. The battery operating status of different time periods is helpful to monitor the health status and operating stability of the battery pack. Based on the battery operating status of different time periods, the operating scenario is inferred to generate the battery pack operating scenarios of different time periods. The generated battery pack operating scenarios are helpful to understand the working mode and performance characteristics of the battery pack under different conditions, and provide a basis for performance testing and optimization.

[0021] Preferably, step S2 comprises the following steps:

[0022] Step S21: extracting multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs;

[0023] Step S22: performing battery pack operating state evolution on the multi-dimensional battery pack operating state parameters, thereby obtaining battery pack operating state characteristics in different time windows;

[0024] Step S23: extracting the operating scenario timestamp of each time period based on the operating scenarios of the battery pack in different time periods;

[0025] Step S24: Based on the operating scenario timestamp of each time period, scene state matching is performed on the operating state characteristics of the battery packs in different time windows, so as to obtain the operating state characteristics of each operating scenario.

[0026] The present invention extracts multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs, which helps to fully understand the operating status of the battery pack. The multi-dimensional parameter extraction provides richer data information, supports subsequent analysis and performance evaluation, and performs state evolution on the multi-dimensional battery pack operating status parameters to reveal the changing trend of the battery pack operating status over time. Obtaining the battery pack operating status characteristics in different time windows helps to analyze the long-term performance and evolution law of the battery pack. Extracting the operating scenario timestamp of each time period based on the battery pack operating scenarios in different time periods helps to mark the time periods of different operating scenarios. The operating scenario timestamp extraction helps to correspond the operating status of the battery pack to the specific scenario. Based on the operating scenario timestamp of each time period, the battery pack operating status characteristics of different time windows are matched with the scenario status to establish the corresponding relationship between the operating status and the scenario. Obtaining the operating status characteristics of each operating scenario helps to deeply understand the performance and operating characteristics of the battery pack under different scenarios.

[0027] Preferably, the specific steps of step S22 are:

[0028] The multi-dimensional battery pack operating state parameters include battery pack temperature parameters, load parameters and current output frequency;

[0029] Calculate the temperature rise change rate of the battery pack temperature parameters to generate the battery pack temperature rise change rate;

[0030] Perform time series fluctuation fitting on the temperature rise change rate of the battery pack and construct the temperature rise change curve of the battery pack;

[0031] Perform time series load trend analysis on load parameters and extract time series load trends;

[0032] Divide the abnormal optimization real-time monitoring log into time windows to obtain monitoring logs of multiple time windows;

[0033] Based on the monitoring logs of multiple time windows, the current output frequency is counted for each time window, so as to obtain the current output frequency of each time window;

[0034] The battery pack operating state evolution is performed on the battery pack temperature rise change curve, time-series load trend and current output frequency in each time window, so as to obtain the battery pack operating state characteristics in different time windows.

[0035] The present invention helps evaluate the thermal management performance and stability of the battery pack by calculating the battery pack temperature rise change rate, generates the battery pack temperature rise change rate to help monitor the thermal characteristics of the battery pack under different loads and working conditions, constructs the battery pack temperature rise change curve through time series fluctuation fitting to more clearly show the temperature rise change trend, constructs the change curve to help analyze the thermal response characteristics of the battery pack and predict the temperature change trend, extracts the time series load trend to help understand the load change of the battery pack in different time periods, analyzes the load trend to reveal the working state and load characteristics of the battery pack, and provides a basis for performance evaluation, divides the time window to help segment the monitoring data for analysis and comparison for different time periods, obtains monitoring logs of multiple time windows to refine the evaluation of the battery pack performance and operating state, counts the current output frequency of each time window to help understand the current output characteristics and stability of the battery pack, obtains the current output frequency of each time window to help monitor the current output of the battery pack under different working conditions, combines the temperature rise change curve, load trend and current output frequency to perform battery pack operating state evolution analysis to comprehensively evaluate the performance of the battery pack, and obtains the battery pack operating state characteristics of different time windows to help discover performance change trends and problems, and provide guidance for further optimization.

[0036] Preferably, step S3 specifically comprises the following steps:

[0037] Step S31: Identify the scene type of the running state characteristics of each running scene to obtain a running scene label;

[0038] Step S32: adaptively encoding the running state characteristics of each running scene based on the running scene label, thereby generating a plurality of running scene state codes;

[0039] Step S33: performing deep evolutionary learning on multiple operation scenario state encodings to generate a potential feature space for each operation scenario;

[0040] Step S34: performing real-time operating condition analysis on the potential feature space of each operating scenario to generate real-time operating condition parameters for each scenario;

[0041] Step S35: Perform virtual operating condition modeling on the real-time operating condition parameters of each scenario to construct a virtual operating condition model of the battery pack.

[0042] The present invention helps to distinguish the performance of battery packs under different working conditions by identifying the operating status characteristics of each operating scenario. The operating scenario labels obtained are helpful to classify and mark the performance of the battery pack in different scenarios. Adaptive encoding is performed based on the operating scenario labels to convert the operating status characteristics into a machine-readable form. Generating multiple operating scenario status codes helps to vectorize the data in different scenarios and provide input for subsequent deep learning. The potential feature space of each operating scenario is generated through deep evolutionary learning to extract high-level features in the data. The potential feature space obtained helps to discover the similarities and differences between different scenarios and provide deeper information for pattern recognition and classification. Real-time operating condition analysis of the potential feature space of each scenario helps to monitor the real-time performance of the battery pack in different scenarios. Real-time operating condition parameters of each scenario are generated to provide real-time performance data and support real-time monitoring and control of the battery pack status. Virtual condition modeling is performed on the real-time operating condition parameters of each scenario to help simulate the performance of the battery pack under different conditions. Constructing a virtual operating condition model of the battery pack helps to predict the performance of the battery pack under future conditions and provide a reference for performance optimization and fault diagnosis.

[0043] Preferably, the specific steps of step S4 are:

[0044] Step S41: defining an acceleration simulation time step, and using a virtual operating condition model of a battery pack to perform an advance acceleration evolution simulation on a multi-dimensional battery pack operating state parameter according to the acceleration simulation time step, to generate acceleration operation evolution simulation data;

[0045] Step S42: performing time series evolution division on the accelerated operation evolution simulation data to obtain operation evolution data sequences of multiple time steps;

[0046] Step S43: mining the battery evolution state segment by segment on the operation evolution data sequence of multiple time steps to generate the battery evolution state of each time step;

[0047] Step S44: analyzing the state change trend of the battery evolution state at each time step, thereby obtaining the accelerated evolution state change trend.

[0048] The present invention shortens the simulation time by defining an accelerated simulation time step, speeds up the data generation and analysis process, and uses a virtual operating condition model of the battery pack to perform advanced accelerated evolution simulation of multi-dimensional battery pack operating status parameters, which helps to simulate the behavior of the battery pack at different time scales. The accelerated operating evolution simulation data is divided into time series evolution, and the data is sorted and analyzed by time period. The operation evolution data sequence obtained for multiple time steps is helpful for a more detailed analysis of the evolution of battery pack performance in different time periods. The battery evolution state mining segment by segment reveals the state changes and characteristics of the battery pack in different time periods. The battery evolution state generated for each time step is helpful to understand the performance evolution process of the battery pack under different working conditions. The state change trend analysis of the battery evolution state for each time step helps to monitor the changing trend of the battery pack state. The accelerated evolution state change trend is helpful to predict the future state evolution trend of the battery pack, providing a basis for performance optimization and fault prediction.

[0049] Preferably, the specific steps of step S5 are:

[0050] Step S51: Calculating the battery residual capacity change according to the acceleration evolution state change trend to generate battery residual capacity change data;

[0051] Step S52: performing a residual capacity attenuation rate trend analysis on the battery residual capacity change data, and constructing a battery pack attenuation rate trend curve;

[0052] Step S53: predicting the battery life based on the battery pack attenuation rate trend curve, thereby obtaining accelerated simulation battery life prediction data;

[0053] Step S54: performing conventional operation life deduction calculation on the accelerated simulation battery life prediction data to obtain normalized battery pack life prediction data.

[0054] The present invention reflects the life consumption of the battery pack under different working conditions by calculating the battery residual capacity change for the accelerated evolution state change trend, and the generated battery residual capacity change data is helpful to understand the change of the battery pack life over time, and the residual capacity decay rate trend analysis of the battery residual capacity change data reveals the rate and law of the battery pack life consumption, and constructing the battery pack decay rate trend curve is helpful to visualize the battery life consumption trend and provide a quantitative basis for life prediction, and the battery life prediction based on the battery pack decay rate trend curve is used to estimate the life consumption of the battery pack, and the accelerated simulation battery life prediction data is obtained, which is helpful to understand the life status of the battery pack in advance and provide support for maintenance and update plans, and the conventional operation life deduction calculation is performed on the accelerated simulation battery life prediction data to map the accelerated simulation results to the actual operation conditions, and the normalized battery pack life prediction data is obtained, which is helpful to predict the life performance of the battery pack in the actual operation process and provide a reference for operation and maintenance.

[0055] Preferably, the specific steps of step S6 are:

[0056] Step S61: identifying the battery pack state deviation based on the acceleration evolution state change trend, and extracting the battery pack state deviation feature;

[0057] Step S62: performing a deviation characteristic distribution analysis on the battery pack state deviation characteristic to generate state deviation characteristic distribution data;

[0058] Step S63: performing a fault time point probability prediction on the state deviation characteristic distribution data, thereby generating a fault occurrence probability prediction time point;

[0059] Step S64: Perform a comprehensive battery performance test on the failure probability prediction time point and the normalized battery life prediction data to obtain a comprehensive battery performance test evaluation result.

[0060] The present invention helps to capture abnormal conditions of battery pack state changes by identifying battery pack state offsets based on the trend of accelerated evolution state changes, extracts battery pack state offset features to help identify potential fault risks and problems, performs distribution analysis on battery pack state offset features to understand the distribution law and change trend of the state offset features, generates state offset feature distribution data to help provide basic data support for subsequent fault prediction and maintenance, performs fault time point probability prediction on state offset feature distribution data to predict the time point when a battery pack fault will occur, and generates a predicted time point for the probability of fault occurrence to help formulate maintenance strategies and reduce the impact of faults on system performance, performs a comprehensive battery pack performance test on the predicted time point for the probability of fault occurrence and normalized battery pack life prediction data to evaluate the performance of the battery pack in actual operation, and obtains a comprehensive battery pack performance test evaluation result to help identify potential problems, improve the design, and improve the reliability and performance of the battery pack.

[0061] In this specification, a new energy vehicle power battery pack performance test system is provided, which is used to perform the new energy vehicle power battery pack performance test method as described above, including:

[0062] The operation scenario inference module is used to obtain the real-time monitoring log of the power battery pack; perform multi-period feature analysis on the real-time monitoring log of the power battery pack, and perform operation scenario inference to generate battery pack operation scenarios in different time periods;

[0063] The scenario status module is used to extract multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs; based on the multi-dimensional battery pack operating status parameters, the scenario status is matched for the battery pack operating scenarios in different time periods, so as to obtain the operating status characteristics of each operating scenario;

[0064] The virtual operating condition module is used to perform real-time operating condition analysis on the operating status characteristics of each operating scenario, and to perform virtual operating condition modeling to build a virtual operating condition model of the battery pack;

[0065] The accelerated evolution simulation module is used to use the battery pack virtual operating condition model to perform advanced accelerated evolution simulation on the multi-dimensional battery pack operating state parameters, and to mine the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend;

[0066] The life prediction module is used to analyze the residual attenuation rate trend of the accelerated evolution state change trend and predict the battery life to obtain normalized battery pack life prediction data;

[0067] The comprehensive performance test module is used to predict the probability of failure time point based on the trend of accelerated evolution state change, so as to generate the predicted time point of failure probability; based on the predicted time point of failure probability and the normalized battery pack life prediction data, the comprehensive performance test evaluation result of the battery pack is obtained.

[0068] The present invention obtains the real-time monitoring log of the power battery pack to provide a data basis for subsequent analysis, performs multi-period feature analysis on the monitoring log, and infers the battery pack operation scenarios in different time periods, which is helpful to understand the performance of the battery pack in different operating states, extracts multi-dimensional battery pack operation status parameters, and provides detailed data support for subsequent scene state matching. According to the abnormal optimization, the real-time monitoring log is matched to the battery pack operation scene status in different time periods, so as to obtain the operation state characteristics of each scene, which is helpful to deeply understand the state of the battery pack in various operation scenarios, and performs real-time working condition analysis on the operation state characteristics of each operation scenario, which provides data support for subsequent model construction, and constructs a virtual operation condition model of the battery pack. The subsequent accelerated evolution simulation provides a simulation basis, uses the virtual operating condition model to perform advanced accelerated evolution simulation on the multi-dimensional battery pack operating status parameters, explores the trend of battery evolution status, provides data support for life prediction, and performs residual attenuation rate trend analysis on the accelerated evolution state change trend to predict the battery life, obtain normalized battery pack life prediction data, provide support for maintenance planning, predict the probability of failure time point for the state change trend, generate the predicted time point of failure probability, help preventive maintenance and fault handling, and conduct comprehensive performance testing of the battery pack based on the predicted time point of failure probability and life prediction data to evaluate the performance of the battery pack in actual operation and provide guidance for performance improvement and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 A schematic diagram of the steps of a new energy vehicle power battery pack performance testing method of the present invention;

[0070] Figure 2Detailed implementation flow chart of step S1;

[0071] Figure 3 Detailed implementation flow chart of step S2;

[0072] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0073] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0074] The present application example provides a new energy vehicle power battery pack performance test method and system. The execution subject of the new energy vehicle power battery pack performance test method and system includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload equipment, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0075] See also Figures 1 to 4 The present invention provides a new energy vehicle power battery pack performance test method, the new energy vehicle power battery pack performance test method comprising the following steps:

[0076] Step S1: obtaining a real-time monitoring log of a power battery pack; performing multi-period feature analysis on the real-time monitoring log of the power battery pack, and performing operation scenario inference to generate battery pack operation scenarios in different time periods;

[0077] Step S2: extracting multi-dimensional battery pack operating status parameters based on the abnormal optimization real-time monitoring log; performing scene state matching on battery pack operating scenarios in different time periods based on the multi-dimensional battery pack operating status parameters, thereby obtaining operating status characteristics of each operating scenario;

[0078] Step S3: Perform real-time operating condition analysis on the operating status characteristics of each operating scenario, perform virtual operating condition modeling, and construct a virtual operating condition model of the battery pack;

[0079] Step S4: using the battery pack virtual operating condition model to simulate the advanced accelerated evolution of the multi-dimensional battery pack operating state parameters, and mining the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend;

[0080] Step S5: performing residual quantity attenuation rate trend analysis on the acceleration evolution state change trend, and performing battery life prediction to obtain normalized battery pack life prediction data;

[0081] Step S6: Perform a probability prediction of the failure time point for the accelerated evolution state change trend, thereby generating a predicted time point for the failure probability; obtain a comprehensive performance test evaluation result of the battery pack based on the predicted time point for the failure probability and the normalized battery pack life prediction data.

[0082] The present invention obtains the real-time monitoring log of the power battery pack to understand the operating status and performance of the battery pack in real time. The acquisition of the real-time monitoring log is helpful to timely discover the problems and abnormal conditions of the battery pack for further analysis and processing. The real-time monitoring log is subjected to multi-time period feature analysis and operation scenario inference, and the battery pack operation scenarios of different time periods are generated, which is helpful to deeply understand the operation of the battery pack, extract multi-dimensional battery pack operation status parameters and perform scenario status matching, and more comprehensively describe the operation status and performance characteristics of the battery pack, so as to provide data support for subsequent analysis, and construct a virtual operation condition model of the battery pack to help simulate the operation status of the battery pack under different working conditions, and provide a virtual experimental environment for battery pack performance analysis. The establishment of the virtual operation condition model is helpful to deeply understand the response characteristics of the battery pack under different working conditions, and provide support for performance optimization and fault prediction. Conduct advanced accelerated evolution simulation and state change trend analysis on multi-dimensional battery pack operating status parameters to reveal the performance evolution law of the battery pack under different working conditions. Analyzing the battery evolution state change trend is helpful to evaluate the life and performance degradation of the battery pack, and provide a basis for battery maintenance and management. Through residual attenuation rate trend analysis and battery life prediction, the life and performance degradation of the battery pack can be inferred to provide decision support for maintenance and update. Normalized battery pack life prediction data is helpful to evaluate the long-term performance of the battery pack and help formulate reasonable maintenance plans and replacement strategies. Predicting the probability of failure time point helps to identify the timing of failure in advance, so as to take preventive measures and reduce the impact of failure on the system. Combined with the prediction of failure probability and normalized battery pack life prediction data, the performance of the battery pack is comprehensively evaluated to provide comprehensive performance test evaluation results for maintenance and update decisions.

[0083] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a new energy vehicle power battery pack performance test method of the present invention. In this example, the steps of the method include:

[0084] Step S1: obtaining a real-time monitoring log of a power battery pack; performing multi-period feature analysis on the real-time monitoring log of the power battery pack, and performing operation scenario inference to generate battery pack operation scenarios in different time periods;

[0085] In this embodiment, various sensors, such as voltage, current, temperature, vibration and other sensors, are deployed on the battery pack to build a comprehensive status monitoring network. The data acquisition module records and stores the power battery pack operation status data collected by these sensors in real time to form a real-time monitoring log. The acquired real-time monitoring log is segmented according to the time dimension to form data subsets of different time periods. For each time period data subset, feature analysis and extraction are performed separately, including: statistical features: mean value, standard deviation, skewness, kurtosis, etc. Frequency domain features: spectrum features obtained by Fourier transform, wavelet transform, etc. Time series features: self- Correlation coefficient, cross-correlation coefficient, etc., through principal component analysis and other methods, the most representative characteristic indicators of the battery pack status are extracted from the above multidimensional features, and the characteristic pattern library of different operating scenarios is established in combination with prior knowledge of the battery pack usage environment, working conditions, etc., and supervised learning algorithms, such as support vector machines and decision trees, are used to compare the battery pack feature vector extracted in step S1-2 with the scene pattern library. According to the similarity between the feature vector and each scene pattern, the operating scene of the battery pack in the current time period is inferred, and the inference result is recorded in the battery pack operation log to form the battery pack operation scene information of different time periods.

[0086] Step S2: extracting multi-dimensional battery pack operating status parameters based on the abnormal optimization real-time monitoring log; performing scene state matching on battery pack operating scenarios in different time periods based on the multi-dimensional battery pack operating status parameters, thereby obtaining operating status characteristics of each operating scenario;

[0087] In this embodiment, an anomaly detection algorithm, such as an isolation forest, a class of support vector machines, etc., is used to analyze the real-time monitoring log data obtained in step S1-1 to identify abnormal points or abnormal patterns in the monitoring data. These abnormalities are caused by abnormal changes in the internal state of the battery pack. For these abnormal points or patterns, multidimensional features rich in battery pack operating status information are further extracted, including: statistical features: mean, variance, skewness, kurtosis, etc., frequency domain features: spectrum features obtained by Fourier transform and wavelet transform, time series features: autocorrelation coefficient, cross-correlation coefficient, etc., and these features are reduced in dimension by methods such as principal component analysis , obtain the most representative multi-dimensional operating status parameters of the battery pack. Match the obtained battery pack operating scenario information in different time periods with the extracted multi-dimensional operating status parameters. For each operating scenario, the corresponding multi-dimensional state parameter distribution characteristics are counted separately, including the mean value, standard deviation, correlation coefficient, etc. Through this scenario-state matching method, the typical operating status characteristics of the battery pack under different operating scenarios are obtained. The obtained battery pack multi-dimensional operating status parameters and the status characteristics under different scenarios are recorded in the battery pack operating status database, providing detailed battery pack operating status data support for subsequent performance analysis, abnormal diagnosis, etc.

[0088] Step S3: Perform real-time operating condition analysis on the operating status characteristics of each operating scenario, perform virtual operating condition modeling, and construct a virtual operating condition model of the battery pack;

[0089] In this embodiment, the state characteristic data of the battery pack under different operating scenarios are used to analyze the distribution characteristics of key operating parameters such as temperature, current, SOC, etc. corresponding to each operating scenario. The probability density distribution of each operating condition parameter is characterized by statistical analysis methods such as histogram and kernel density estimation. Based on the distribution characteristics of each operating condition parameter, a large number of virtual operating condition samples are generated by Monte Carlo simulation and other methods. These virtual operating condition samples should be able to well reflect the various complex operating conditions encountered by the battery pack in actual use. Mathematical models such as probability distribution function and probability density function are used to construct a virtual operating condition model of the battery pack. The constructed virtual operating condition model is compared and verified with the actual monitored battery pack operating condition data, the differences in statistical characteristics between the two are analyzed, and the parameters of the virtual operating condition model are further optimized to make it closer to reality. Through continuous iterative optimization, a virtual battery pack operating model that can be highly close to the actual use conditions is finally constructed. The optimized battery pack virtual operating condition model is output in a standard data format to facilitate subsequent applications. This model can provide highly realistic simulation conditions for battery pack performance testing, accelerated life testing, performance optimization, etc. It can also be used for the development and verification of battery pack control strategies to improve actual use effects.

[0090] Step S4: using the battery pack virtual operating condition model to simulate the advanced accelerated evolution of the multi-dimensional battery pack operating state parameters, and mining the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend;

[0091] In this embodiment, the constructed virtual operating condition model of the battery pack is used to simulate the operating condition parameters of different time periods, and these operating condition parameters are input into the performance model of the battery pack to simulate the changes in the multi-dimensional operating state parameters of the battery pack under the virtual working condition. These operating state parameters include key performance indicators such as battery capacity, internal resistance, and power. The change curve of the simulated multi-dimensional state parameters over time is analyzed, and the evolution trend of the battery pack performance indicators under accelerated working conditions is extracted by using methods such as curve fitting and time series analysis. By comparing the evolution trends under different working conditions, the rapid degradation stage of the key performance indicators of the battery pack is identified. In this section, combined with the mechanism models of chemical reactions and heat conduction inside the battery pack, the root causes of rapid performance degradation are analyzed. For different degradation mechanisms, corresponding prevention and optimization measures are proposed, such as adjusting operating parameters and optimizing thermal management. Through in-depth analysis of the internal state and performance of the battery pack, targeted optimization strategies are provided to extend the battery life. The obtained analysis results of the accelerated evolution state of the battery pack are output in a visual form. These analysis results can provide valuable references for battery pack design optimization, control strategy formulation, fault diagnosis, etc., and can also be fed back to the virtual operating condition modeling link to further improve the battery pack simulation model.

[0092] Step S5: performing residual quantity attenuation rate trend analysis on the acceleration evolution state change trend, and performing battery life prediction to obtain normalized battery pack life prediction data;

[0093] In this embodiment, based on the evolution trend of the battery pack performance indicators under accelerated conditions, the decay curves of key performance indicators, such as capacity and internal resistance, over time are analyzed. The exponential decay model, Vickers model, etc. are used to fit the variation law of the residual decay rate of the performance indicators over time. The obtained residual decay rate of the performance indicators is compared with the defined life endpoint threshold of the battery pack. According to the time point when the performance indicator drops to the life threshold, the remaining service life of the battery pack under normal conditions is predicted. For different performance indicators, multiple life prediction models are constructed respectively, and a more reliable life prediction result is obtained through comprehensive analysis. The obtained life prediction result is compared and verified with the degradation of the battery pack in actual use, and the difference between the two is analyzed. The parameters of the life prediction model are further optimized to improve the prediction accuracy. Through continuous iterative optimization, a model that can accurately predict the normalized life of the battery pack is finally established. The optimized battery pack life prediction model is output in a standard data format for subsequent application. The model can provide valuable reference for health management, maintenance, and use planning of the battery pack, and can also be used for optimizing the battery pack control strategy to maximize the service life of the battery pack.

[0094] Step S6: Perform a probability prediction of the failure time point for the accelerated evolution state change trend, thereby generating a predicted time point for the failure probability; obtain a comprehensive performance test evaluation result of the battery pack based on the predicted time point for the failure probability and the normalized battery pack life prediction data.

[0095] In this embodiment, based on the evolution trend of the battery pack performance indicators under accelerated conditions, the statistical distribution characteristics of the degradation curves of the key performance indicators in the rapid degradation stage are analyzed, and the failure time prediction model, such as Weibull distribution, lognormal distribution, etc., is used to predict the law of the probability of the performance indicator reaching the failure threshold over time, and a reasonable threshold, such as 80%, is set for the predicted failure probability. According to the probability threshold, the predicted time point when the performance indicator reaches the failure level is inferred. For different performance indicators, multiple failure probability prediction time points are obtained, and the obtained battery pack normal life prediction results are combined and compared with the failure probability prediction time point. If the failure probability prediction time point is significantly ahead of the normal life prediction result, it means that the battery pack has serious reliability problems. According to the time difference between the two, the evaluation results of the comprehensive performance test of the battery pack are given, such as excellent / good / average / poor. The obtained battery pack comprehensive performance test evaluation results are output in the form of a standard report. The report can provide valuable basis for design optimization, production process improvement, use strategy adjustment, etc. of the battery pack, and can also provide reference for battery pack procurement, repair and maintenance, etc., which is helpful to improve the overall use efficiency and safety of the battery pack.

[0096] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0097] Step S11: Obtaining a real-time monitoring log of a power battery pack;

[0098] Step S12: removing abnormal values ​​from the real-time monitoring log of the power battery pack to obtain an abnormal optimized real-time monitoring log;

[0099] Step S13: Calculate the real-time output power of the power battery pack based on the abnormal optimization real-time monitoring log to obtain the real-time output power;

[0100] Step S14: Perform multi-period feature analysis on the real-time output power to extract the output power in multiple time periods;

[0101] Step S15: identifying the operating status of the battery pack based on the output power in multiple time periods to obtain the operating status of the battery in different time periods;

[0102] Step S16: Inferring the operating scenarios based on the battery operating status in different time periods, and generating the battery pack operating scenarios in different time periods.

[0103] In this embodiment, the monitoring system of the power battery pack is confirmed, the source of the real-time monitoring data, such as a sensor, a controller or a data acquisition system, is identified, and a real-time monitoring log is obtained from the monitoring system using a data interface (such as an API, an MQTT protocol), and information including voltage, current, temperature, and charging status is recorded. The standard of abnormal values ​​is determined, such as exceeding the normal voltage range, abnormal temperature, sudden change in charging current, etc., and statistical methods (such as Z-score, IQR method) or machine learning algorithms (such as isolation forest) are used to detect abnormal values. The selected abnormal value detection method is applied to the real-time monitoring log, and the identified abnormal data is removed to obtain an abnormally optimized real-time monitoring log. The voltage and current data are extracted from the abnormally optimized real-time monitoring log for power calculation. The real-time output power (P) = voltage (V) × Current (I), power calculation is performed for each monitoring log record to obtain real-time output power. Multiple time periods are defined according to demand, such as by hour, day or specific events (such as charging, discharging, etc.). Data analysis tools (such as Pandas, NumPy) are used to perform multi-period feature analysis. According to the defined time period, the output power characteristics of each time period are counted, such as average power, peak power, power fluctuation, etc., to determine the operating status standard of the battery pack, such as charging status, discharging status, standby status, fault status, etc. Machine learning classification algorithms (such as decision trees, support vector machines) or rule engines are used for operating status identification. The output power characteristics of multiple time periods are analyzed to identify the battery operating status of each time period. Different operating scenarios are defined according to the battery operating status, such as high-load charging, low-load discharging, fault emergency, etc. The operating scenario inference is performed using rule engines, logical inference or model-based methods. Based on the battery operating status of different time periods, the corresponding battery pack operating scenarios are inferred, and the characteristics of each scenario are recorded. The generated battery pack operating scenarios of different time periods are stored in the database for subsequent analysis and decision support.

[0104] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0105] Step S21: extracting multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs;

[0106] Step S22: performing battery pack operating state evolution on the multi-dimensional battery pack operating state parameters, thereby obtaining battery pack operating state characteristics in different time windows;

[0107] Step S23: extracting the operating scenario timestamp of each time period based on the operating scenarios of the battery pack in different time periods;

[0108] Step S24: Based on the operating scenario timestamp of each time period, scene state matching is performed on the operating state characteristics of the battery packs in different time windows, so as to obtain the operating state characteristics of each operating scenario.

[0109] In this embodiment, relevant data including voltage, current, temperature, charging state, SOC (state of charge), etc. are extracted from the abnormal optimization real-time monitoring log, and data analysis tools (such as Pandas and NumPy) are used to calculate the extracted data to generate multi-dimensional battery pack operating state parameters, determine the time window for analysis, such as dividing by hours, days, or specific events (such as charging, discharging, etc.), use time series analysis, state space model or machine learning method (such as LSTM) to analyze the evolution of the battery pack operating state, analyze the multi-dimensional battery pack operating state parameters, obtain the battery pack operating state characteristics of different time windows, store the evolved state characteristics in the database, and form a time window. A battery pack operating status feature dataset is generated, relevant data is extracted from the previously generated battery pack operating scenarios, and preparations are made for subsequent timestamp extraction. The criteria for extracting timestamps are determined. For example, according to operating status changes, event triggers (such as the start of charging, the end of discharging), etc., in different operating scenarios, the operating scenario timestamps of each time period are extracted, and relevant status information is recorded. The operating scenario timestamps of each time period and the battery pack operating status features of different time windows are extracted from the database, and the criteria for scene status matching are determined, such as the proximity of timestamps, the similarity of status features, etc. The battery pack operating status features of different time windows are matched with the operating scenario timestamps, and the operating status features corresponding to each operating scenario are identified.

[0110] In this embodiment, the specific steps of step S22 are:

[0111] The multi-dimensional battery pack operating state parameters include battery pack temperature parameters, load parameters and current output frequency;

[0112] Calculate the temperature rise change rate of the battery pack temperature parameters to generate the battery pack temperature rise change rate;

[0113] Perform time series fluctuation fitting on the temperature rise change rate of the battery pack and construct the temperature rise change curve of the battery pack;

[0114] Perform time series load trend analysis on load parameters and extract time series load trends;

[0115] Divide the abnormal optimization real-time monitoring log into time windows to obtain monitoring logs of multiple time windows;

[0116] Based on the monitoring logs of multiple time windows, the current output frequency is counted for each time window, so as to obtain the current output frequency of each time window;

[0117] The battery pack operating state evolution is performed on the battery pack temperature rise change curve, time-series load trend and current output frequency in each time window, so as to obtain the battery pack operating state characteristics in different time windows.

[0118] In this embodiment, the temperature parameter data of the battery pack is extracted from the abnormal optimization real-time monitoring log, including the timestamp and the corresponding temperature value, the temperature rise change rate (ΔT / Δt) = (T1 - T0) / (t1 - t0), where T1 and T0 are the temperatures at two time points, and t1 and t0 are the corresponding times. The extracted temperature data is calculated for each time period to obtain the battery pack temperature rise change rate. The battery pack temperature rise change rate data is extracted from the database. In preparation for fitting analysis, linear regression, polynomial fitting, or spline interpolation methods are used to perform time series fluctuation fitting on the temperature rise change rate. The battery pack temperature rise change curve is generated using the selected fitting method to show the trend of the temperature rise change. The load parameters are extracted from the monitoring log to ensure that the timestamp and load value are included. The time series analysis method (such as moving average, exponential smoothing) is used to analyze the trend of the load parameters. The time series load trend of the load parameters is calculated and plotted, and key features (such as peak values, fluctuations) are extracted. The abnormal optimization real-time monitoring log is extracted from the database. The monitoring log data is collected and the length of the time window is defined according to the demand (such as every hour, every 5 minutes). The monitoring log is divided according to the set time window to generate monitoring logs of multiple time windows. The current output data is extracted from the abnormal optimization real-time monitoring log. The counting method or frequency analysis tool is used to count the current output frequency of each time window. Each time window is traversed to count the frequency of current output in the window. The temperature rise change curve, time series load trend and current output frequency of each time window are extracted from the database. Multivariate time series analysis or state space model is used to analyze the evolution of the operating status of the battery pack. According to the integrated data, the operating status of the battery pack in different time windows is analyzed to identify the status characteristics.

[0119] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0120] Step S31: Identify the scene type of the running state characteristics of each running scene to obtain a running scene label;

[0121] Step S32: adaptively encoding the running state characteristics of each running scene based on the running scene label, thereby generating a plurality of running scene state codes;

[0122] Step S33: performing deep evolutionary learning on multiple operation scenario state encodings to generate a potential feature space for each operation scenario;

[0123] Step S34: performing real-time operating condition analysis on the potential feature space of each operating scenario to generate real-time operating condition parameters for each scenario;

[0124] Step S35: Perform virtual operating condition modeling on the real-time operating condition parameters of each scenario to construct a virtual operating condition model of the battery pack.

[0125] In this embodiment, the operating status characteristics of each operating scenario are extracted from the database, including temperature rise changes, load trends, and current output frequency, etc. According to the operating characteristics of the battery pack, the scenario type (such as normal charging and discharging, overload, fault, etc.) is defined, and the scene type is identified using a machine learning classification algorithm (such as KNN, decision tree, support vector machine), and the classification model is trained using labeled historical data. The state characteristics of each operating scenario are combined with the corresponding scene label to form a data set, and the state characteristics are encoded using an adaptive coding technology (such as autoencoder, feature embedding), and the adaptive coding model is trained to convert the operating state characteristics into a low-dimensional coding representation, and the battery pack is extracted from the database. Take the generated state codes of multiple operating scenarios, use deep learning models (such as deep neural networks, long short-term memory networks LSTM) for evolutionary learning, use state code data to train the deep learning model to capture the potential feature space between scenarios, extract the information of each operating scenario from the potential feature space generated in the previous step, use real-time analysis tools (such as stream data processing platforms Apache Kafka, Storm) or data analysis libraries (such as Pandas, NumPy) to perform condition analysis, perform real-time condition analysis on each scenario based on the potential feature space, generate real-time operating condition parameters (such as efficiency, power, temperature, etc.), use virtual modeling tools (such as MATLAB Simulink, AnyLogic) or simulation platforms (such as Unity) to build the model, build a virtual operating condition model of the battery pack according to the real-time operating condition parameters, show the performance of the battery pack in different scenarios, and verify the model through historical data or real-time data to ensure the accuracy and reliability of the model.

[0126] In this embodiment, step S4 includes the following steps:

[0127] Step S41: defining an acceleration simulation time step, and using a virtual operating condition model of a battery pack to perform an advance acceleration evolution simulation on a multi-dimensional battery pack operating state parameter according to the acceleration simulation time step, to generate acceleration operation evolution simulation data;

[0128] Step S42: performing time series evolution division on the accelerated operation evolution simulation data to obtain operation evolution data sequences of multiple time steps;

[0129] Step S43: mining the battery evolution state segment by segment on the operation evolution data sequence of multiple time steps to generate the battery evolution state of each time step;

[0130] Step S44: analyzing the state change trend of the battery evolution state at each time step, thereby obtaining the accelerated evolution state change trend.

[0131] In this embodiment, the time step of the accelerated simulation is defined according to the requirements and model characteristics, for example, it is set to 1 minute, 5 minutes or other suitable time intervals. Ensure that the virtual operating condition model of the battery pack has been constructed and can be used for simulation. Use the defined acceleration time step to simulate the virtual operating condition model of the battery pack to generate advanced acceleration evolution simulation data. During the simulation process, multi-dimensional battery pack operating state parameters (such as temperature, SOC, load, etc.) are recorded to form an accelerated operation evolution simulation data set. According to the characteristics of the accelerated simulation, multiple time steps are set (for example, 2 years, 5 years). The accelerated operation evolution simulation data is divided according to the set time period to generate an operation evolution data sequence of multiple time steps. Use data mining algorithms (such as cluster analysis, decision trees, support vector machines) to analyze each segment of data to identify the evolution state of the battery. The operation evolution data sequence of each time period is analyzed segment by segment to identify the battery evolution state of each time step. Use time series analysis methods (such as moving average, trend line fitting) or statistical analysis tools (such as regression analysis) to analyze state change trends. Perform trend analysis on the battery evolution state at each time step to identify the changing trend of the accelerated evolution state (such as increase, decrease, fluctuation, etc.).

[0132] In this embodiment, the specific steps of step S5 are:

[0133] Step S51: Calculating the battery residual capacity change according to the acceleration evolution state change trend to generate battery residual capacity change data;

[0134] Step S52: performing a residual capacity attenuation rate trend analysis on the battery residual capacity change data, and constructing a battery pack attenuation rate trend curve;

[0135] Step S53: predicting the battery life based on the battery pack attenuation rate trend curve, thereby obtaining accelerated simulation battery life prediction data;

[0136] Step S54: performing conventional operation life deduction calculation on the accelerated simulation battery life prediction data to obtain normalized battery pack life prediction data.

[0137] In this embodiment, the accelerated evolution state change trend data is extracted from the database, including the battery operating state parameters (such as SOC, temperature, etc.), the battery residual capacity (R) = initial capacity (C) - used capacity (U), where the used capacity is calculated by current and time, and the battery residual capacity change is calculated for each time period according to the accelerated evolution state data, and the residual capacity data for each time period is recorded.

[0138] initial_capacity = 100 # Initial capacity, unit: Ah

[0139] simulated_data = [

[0140] {"time": 0, "capacity": 100},

[0141] {"time": 10, "capacity": 98},

[0142] {"time": 20, "capacity": 95},

[0143] {"time": 30, "capacity": 92} ]

[0145] for entry in simulated_data:

[0146] entry['remaining_capacity'] = initial_capacity - entry['capacity']

[0147] # Result example;

[0148] print(simulated_data).

[0149] Residual quantity decay rate (D) = (R1 - R2) / (t1 - t2), where R1 and R2 are the residual quantities at two time points, and t1 and t2 are the corresponding times. The residual quantities in each time period are calculated segment by segment to obtain the residual quantity decay rate data. A decay rate trend curve is drawn using a data visualization tool (such as Matplotlib, Seaborn) to show the change in the decay rate of the battery pack. A life prediction method (such as linear regression, exponential decay model) is used to predict the battery life according to the decay rate trend curve. According to the decay rate trend curve, the remaining service life of the battery is calculated to generate accelerated simulation battery life prediction data. The accelerated simulation battery life prediction data is extracted from the database. The battery life deduction model under a normal operating environment (such as a cumulative working time model, a historical data regression model) is used for analysis. According to the accelerated simulation battery life prediction data, the battery life prediction data under normal operating conditions is deduced. Each deduction result is recorded to form a normalized battery life prediction data set. The normalized battery life prediction data is saved. The normalized battery life prediction data is stored in a database for subsequent analysis and application.

[0150] from sklearn.linear_model import LinearRegression

[0151] import numpy as np

[0152] # Assume decay_rates is a 2D array, X is the time step, y is the decay rate

[0153] X = np.array(range(1, len(decay_rates) + 1)).reshape(-1, 1) # time step

[0154] y = np.array(decay_rates).reshape(-1, 1) # decay rate

[0155] model = LinearRegression()

[0156] model.fit(X, y)

[0157] # Predict future decay rates

[0158] future_steps = np.array([[len(decay_rates) + 1]])

[0159] predicted_decay_rate = model.predict(future_steps)

[0160] print(predicted_decay_rate)

[0161] In this embodiment, the specific steps of step S6 are:

[0162] Step S61: performing battery pack state deviation identification on the acceleration evolution state change trend to extract battery pack state deviation features;

[0163] Step S62: performing a deviation characteristic distribution analysis on the battery pack state deviation characteristic to generate state deviation characteristic distribution data;

[0164] Step S63: performing a fault time point probability prediction on the state deviation characteristic distribution data, thereby generating a fault occurrence probability prediction time point;

[0165] Step S64: Perform a comprehensive battery performance test on the failure probability prediction time point and the normalized battery life prediction data to obtain a comprehensive battery performance test evaluation result.

[0166] In this embodiment, the standard for state shift is determined, such as defining a shift threshold (such as a change in SOC exceeding a certain percentage), comparing the state parameter of each time step with its expected value, identifying the shift situation, extracting relevant shift features, such as the shift amplitude, shift direction, and shift duration, and storing the extracted battery pack state shift features in a database for subsequent analysis.

[0167] import numpy as np

[0168] def identify_state_shift(simulated_data):

[0169] # Assume the normal range is a certain threshold

[0170] normal_range = (3.2, 4.2) # Normal voltage range

[0171] shifts = []

[0172] for entry in simulated_data:

[0173] if entry['voltage']<normal_range[0] or entry['voltage']> normal_range[1]:

[0174] shifts.append({

[0175] "time": entry['time'],

[0176] "voltage": entry['voltage'],

[0177] "shift": "Out of Range"

[0178] })

[0179] return shifts

[0180] state_shifts = identify_state_shift(simulated_data)

[0181] Use statistical analysis methods (such as histograms, kernel density estimation) or machine learning methods (such as cluster analysis) to analyze the distribution of offset features. Based on the extracted offset feature data, calculate its distribution characteristics and generate a corresponding distribution graph. Record the analysis results to form a distribution data set of state offset features. Use a probabilistic prediction model (such as logistic regression, Bayesian network, time series prediction model) to predict the probability of fault occurrence. Based on the state offset feature distribution data, calculate the probability of fault occurrence at each time point and generate a predicted time point for the probability of fault occurrence.

[0182] from sklearn.linear_model import LogisticRegression

[0183] # Assume there is a dataset containing labels of time points and failure occurrences

[0184] X = np.array([[t] for t in range(len(distribution))]) # time point

[0185] y = np.array([1 if shift == 'Out of Range' else 0 for shift indistribution['shift_type']])

[0186] model = LogisticRegression()

[0187] model.fit(X, y)

[0188] # Predict the probability of failure at a future time point

[0189] future_time_points = np.array([[len(distribution) + i] for i in range(1, 6)]) # 5 future time points

[0190] predicted_probabilities = model.predict_proba(future_time_points)[:, 1] # Predicted failure probability

[0191] Determine the standards and indicators for comprehensive performance testing, such as capacity retention rate, internal resistance change, cycle life, etc. Based on the extracted data, simulate or actually test the performance of the battery pack at the time point when the failure probability occurs, record the results of each test, form a comprehensive performance test evaluation result, and store the comprehensive performance test evaluation result in a database for subsequent analysis and decision support.

[0192] def performance_testing(predicted_probabilities, life_expectancy):

[0193] results = []

[0194] for prob in predicted_probabilities:

[0195] # Assume that the failure probability is used to evaluate the overall performance

[0196] performance_score = (1 - prob) * life_expectancy # Comprehensive performance evaluation

[0197] results.append(performance_score)

[0198] return results

[0199] life_expectancy = 200 # Expected lifespan, unit: hours

[0200] performance_results = performance_testing(predicted_probabilities, life_expectancy).

[0201] In this embodiment, a new energy vehicle power battery pack performance test system is provided, which is used to execute the new energy vehicle power battery pack performance test method as described above, including:

[0202] The operation scenario inference module is used to obtain the real-time monitoring log of the power battery pack; perform multi-period feature analysis on the real-time monitoring log of the power battery pack, and perform operation scenario inference to generate battery pack operation scenarios in different time periods;

[0203] The scenario status module is used to extract multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs; based on the multi-dimensional battery pack operating status parameters, the scenario status is matched for the battery pack operating scenarios in different time periods, so as to obtain the operating status characteristics of each operating scenario;

[0204] The virtual operating condition module is used to perform real-time operating condition analysis on the operating status characteristics of each operating scenario, and to perform virtual operating condition modeling to build a virtual operating condition model of the battery pack;

[0205] The accelerated evolution simulation module is used to use the battery pack virtual operating condition model to perform advanced accelerated evolution simulation on the multi-dimensional battery pack operating state parameters, and to mine the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend;

[0206] The life prediction module is used to analyze the residual attenuation rate trend of the accelerated evolution state change trend and predict the battery life to obtain normalized battery pack life prediction data;

[0207] The comprehensive performance test module is used to predict the probability of failure time point based on the trend of accelerated evolution state change, so as to generate the predicted time point of failure probability; based on the predicted time point of failure probability and the normalized battery pack life prediction data, the comprehensive performance test evaluation result of the battery pack is obtained.

[0208] The present invention obtains the real-time monitoring log of the power battery pack to provide a data basis for subsequent analysis, performs multi-period feature analysis on the monitoring log, and infers the battery pack operation scenarios in different time periods, which is helpful to understand the performance of the battery pack in different operating states, extracts multi-dimensional battery pack operation status parameters, and provides detailed data support for subsequent scene state matching. According to the abnormal optimization, the real-time monitoring log is matched to the battery pack operation scene status in different time periods, so as to obtain the operation state characteristics of each scene, which is helpful to deeply understand the state of the battery pack in various operation scenarios, and performs real-time working condition analysis on the operation state characteristics of each operation scenario, which provides data support for subsequent model construction, and constructs a virtual operation condition model of the battery pack. The subsequent accelerated evolution simulation provides a simulation basis, uses the virtual operating condition model to perform advanced accelerated evolution simulation on the multi-dimensional battery pack operating status parameters, explores the trend of battery evolution status, provides data support for life prediction, and performs residual attenuation rate trend analysis on the accelerated evolution state change trend to predict the battery life, obtain normalized battery pack life prediction data, provide support for maintenance planning, predict the probability of failure time point for the state change trend, generate the predicted time point of failure probability, help preventive maintenance and fault handling, and conduct comprehensive performance testing of the battery pack based on the predicted time point of failure probability and life prediction data to evaluate the performance of the battery pack in actual operation and provide guidance for performance improvement and optimization.

[0209] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0210] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A new energy vehicle power battery pack performance testing method, characterized in that: The following steps are involved: Step S1: obtaining a real-time monitoring log of a power battery pack; performing multi-period feature analysis on the real-time monitoring log of the power battery pack, and performing operation scenario inference to generate battery pack operation scenarios in different time periods; Step S2: extracting multi-dimensional battery pack operating status parameters based on the abnormal optimization real-time monitoring log; performing scene state matching on battery pack operating scenarios in different time periods based on the multi-dimensional battery pack operating status parameters, thereby obtaining operating status characteristics of each operating scenario; Step S3: Perform real-time operating condition analysis on the operating status characteristics of each operating scenario, perform virtual operating condition modeling, and construct a virtual operating condition model of the battery pack; Step S4: using the battery pack virtual operating condition model to simulate the advanced accelerated evolution of the multi-dimensional battery pack operating state parameters, and mining the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend; Step S5: performing residual quantity attenuation rate trend analysis on the acceleration evolution state change trend, and performing battery life prediction to obtain normalized battery pack life prediction data; Step S6: predicting the probability of a fault time point based on the changing trend of the accelerated evolution state, thereby generating a predicted time point of the fault occurrence probability; The comprehensive performance test and evaluation results of the battery pack are obtained based on the predicted time point of failure probability and normalized battery pack life prediction data.

2. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: Obtaining a real-time monitoring log of a power battery pack; Step S12: removing abnormal values ​​from the real-time monitoring log of the power battery pack to obtain an abnormal optimized real-time monitoring log; Step S13: Calculate the real-time output power of the power battery pack based on the abnormal optimization real-time monitoring log to obtain the real-time output power; Step S14: Perform multi-period feature analysis on the real-time output power to extract the output power in multiple time periods; Step S15: identifying the operating status of the battery pack based on the output power in multiple time periods to obtain the operating status of the battery in different time periods; Step S16: Inferring the operating scenarios based on the battery operating status in different time periods, and generating the battery pack operating scenarios in different time periods.

3. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: extracting multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs; Step S22: performing battery pack operating state evolution on the multi-dimensional battery pack operating state parameters, thereby obtaining battery pack operating state characteristics in different time windows; Step S23: extracting the operating scenario timestamp of each time period based on the operating scenarios of the battery pack in different time periods; Step S24: Based on the operating scenario timestamp of each time period, scene state matching is performed on the operating state characteristics of the battery packs in different time windows, so as to obtain the operating state characteristics of each operating scenario.

4. The new energy vehicle power battery pack performance testing method according to claim 3, characterized in that: The specific steps of step S22 are: The multi-dimensional battery pack operating state parameters include battery pack temperature parameters, load parameters and current output frequency; Calculate the temperature rise change rate of the battery pack temperature parameters to generate the battery pack temperature rise change rate; Perform time series fluctuation fitting on the temperature rise change rate of the battery pack and construct the temperature rise change curve of the battery pack; Perform time series load trend analysis on load parameters and extract time series load trends; Divide the abnormal optimization real-time monitoring log into time windows to obtain monitoring logs of multiple time windows; Based on the monitoring logs of multiple time windows, the current output frequency is counted for each time window, so as to obtain the current output frequency of each time window; The battery pack operating state evolution is performed on the battery pack temperature rise change curve, time-series load trend and current output frequency in each time window, so as to obtain the battery pack operating state characteristics in different time windows.

5. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: Identify the scene type of the running state characteristics of each running scene to obtain a running scene label; Step S32: adaptively encoding the running state characteristics of each running scene based on the running scene label, thereby generating a plurality of running scene state codes; Step S33: performing deep evolutionary learning on multiple operation scenario state encodings to generate a potential feature space for each operation scenario; Step S34: performing real-time operating condition analysis on the potential feature space of each operating scenario to generate real-time operating condition parameters for each scenario; Step S35: Perform virtual operating condition modeling on the real-time operating condition parameters of each scenario to construct a virtual operating condition model of the battery pack.

6. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: defining an acceleration simulation time step, and using a virtual operating condition model of a battery pack to perform an advance acceleration evolution simulation on a multi-dimensional battery pack operating state parameter according to the acceleration simulation time step, to generate acceleration operation evolution simulation data; Step S42: performing time series evolution division on the accelerated operation evolution simulation data to obtain operation evolution data sequences of multiple time steps; Step S43: mining the battery evolution state segment by segment on the operation evolution data sequence of multiple time steps to generate the battery evolution state of each time step; Step S44: analyzing the state change trend of the battery evolution state at each time step, thereby obtaining the accelerated evolution state change trend.

7. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: Calculating the battery residual capacity change according to the acceleration evolution state change trend to generate battery residual capacity change data; Step S52: performing a residual capacity attenuation rate trend analysis on the battery residual capacity change data, and constructing a battery pack attenuation rate trend curve; Step S53: predicting the battery life based on the battery pack attenuation rate trend curve, thereby obtaining accelerated simulation battery life prediction data; Step S54: performing conventional operation life deduction calculation on the accelerated simulation battery life prediction data to obtain normalized battery pack life prediction data.

8. The new energy vehicle power battery pack performance testing method according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: identifying the battery pack state deviation based on the acceleration evolution state change trend, and extracting the battery pack state deviation feature; Step S62: performing a deviation characteristic distribution analysis on the battery pack state deviation characteristic to generate state deviation characteristic distribution data; Step S63: performing a fault time point probability prediction on the state deviation characteristic distribution data, thereby generating a fault occurrence probability prediction time point; Step S64: Perform a comprehensive battery performance test on the failure probability prediction time point and the normalized battery life prediction data to obtain a comprehensive battery performance test evaluation result.

9. A new energy vehicle power battery pack performance test system, characterized in that: Used to perform the new energy vehicle power battery pack performance testing method as claimed in claim 1, comprising: The operation scenario inference module is used to obtain the real-time monitoring log of the power battery pack; perform multi-period feature analysis on the real-time monitoring log of the power battery pack, and perform operation scenario inference to generate battery pack operation scenarios in different time periods; The scenario status module is used to extract multi-dimensional battery pack operating status parameters based on abnormal optimization real-time monitoring logs; based on the multi-dimensional battery pack operating status parameters, the scenario status is matched for the battery pack operating scenarios in different time periods, so as to obtain the operating status characteristics of each operating scenario; The virtual operating condition module is used to perform real-time operating condition analysis on the operating status characteristics of each operating scenario, and to perform virtual operating condition modeling to build a virtual operating condition model of the battery pack; The accelerated evolution simulation module is used to use the battery pack virtual operating condition model to perform advanced accelerated evolution simulation on the multi-dimensional battery pack operating state parameters, and to mine the battery evolution state segment by segment, so as to obtain the accelerated evolution state change trend; The life prediction module is used to analyze the residual attenuation rate trend of the accelerated evolution state change trend and predict the battery life to obtain normalized battery pack life prediction data; The comprehensive performance test module is used to predict the probability of failure time point based on the trend of accelerated evolution state change, so as to generate the predicted time point of failure probability; based on the predicted time point of failure probability and the normalized battery pack life prediction data, the comprehensive performance test evaluation result of the battery pack is obtained.

Images