Method and device for analyzing and testing lithium polymer battery

Abstract

The invention provides an analysis test method and device for a lithium polymer battery, and relates to the technical field of lithium polymer batteries, and the method comprises the steps: carrying out the working condition extraction and parameter analysis of the application scene information of a target lithium polymer battery; according to the battery working condition test parameter table, placing the target lithium polymer battery on a tool test platform for working condition simulation test, and monitoring to obtain a battery multi-dimensional data stream; extracting environment features and discharge features in parallel; introducing an attention mechanism to carry out weighted fusion, and determining a discharge performance-environment interactive fusion feature set; and constructing a battery performance analyzer, evaluating and analyzing the discharge performance-environment interactive fusion feature set, and generating a lithium polymer battery test analysis result. According to the method and the device, the technical problem that the test precision is insufficient due to the fact that the complexity of an application scene is not fully considered in a test method in the prior art is solved, and the precision and the efficiency of battery analysis test are improved by simulating a real dynamic working condition.

Description

Technical Field

[0001] This application relates to the field of lithium polymer battery technology, and in particular to an analytical testing method and apparatus for lithium polymer batteries. Background Technology

[0002] As a high-energy-density, lightweight energy storage device, lithium polymer batteries are currently tested using methods including constant current charge-discharge testing, cycle life testing, rate performance testing, and safety performance testing. These methods can, to a certain extent, evaluate the battery's basic performance indicators. However, in real-world applications, batteries are often affected by dynamic current curves, temperature changes, and load fluctuations. Testing under idealized, static, or simply dynamic conditions makes it difficult to simulate the complex factors of dynamic current curves, temperature fluctuations, and load changes encountered in actual use. Consequently, the test results cannot accurately reflect the battery's performance in real-world applications, severely impacting test accuracy.

[0003] In summary, existing technologies suffer from insufficient testing accuracy due to inadequate consideration of the complexity of application scenarios in testing methods, which further affects the safety of lithium polymer batteries. Summary of the Invention

[0004] The purpose of this application is to provide an analytical testing method and apparatus for lithium polymer batteries, in order to solve the technical problem in the prior art where the testing method does not fully consider the complexity of the application scenario, resulting in insufficient testing accuracy and further affecting the safety of lithium polymer batteries.

[0005] In view of the above problems, this application provides an analytical testing method and apparatus for lithium polymer batteries.

[0006] Firstly, this application provides an analytical testing method for lithium polymer batteries. This method is implemented using an analytical testing device for lithium polymer batteries. The method includes: extracting operating conditions and parsing parameters from the application scenario information of the target lithium polymer battery to generate a battery operating condition test parameter table; placing the target lithium polymer battery on a tooling test platform according to the battery operating condition test parameter table for operating condition simulation testing, and monitoring and acquiring multi-dimensional battery data streams; extracting environmental features and discharge features from the multi-dimensional battery data streams in parallel to obtain a battery environmental time-series feature set and a battery discharge waveform feature set; introducing an attention mechanism to weightedly fuse the battery environmental time-series feature set and the battery discharge waveform feature set to determine a discharge performance-environment interaction fusion feature set; constructing a battery performance analyzer based on a battery performance index system, and evaluating and analyzing the discharge performance-environment interaction fusion feature set based on the battery performance analyzer to generate lithium polymer battery test analysis results.

[0007] Optionally, the application scenario information of the target lithium polymer battery is classified, extracted, and analyzed to obtain a set of battery application scenario types and a set of scenario application requirement parameters; operating condition elements are extracted from the set of battery application scenario types according to the set of scenario application requirement parameters to determine a set of battery scenario operating condition elements; historical data mining is performed on the set of battery scenario operating condition elements to obtain a historical operating dataset of battery operating condition elements; and parameter parsing is performed on the set of battery scenario operating condition elements based on the historical operating dataset of battery operating condition elements to obtain a battery operating condition test parameter table.

[0008] Optionally, the battery scenario operating condition element set is analyzed for parameter distribution based on the historical operating data set of the battery operating condition elements to obtain the battery operating condition parameter distribution threshold; the amount of operating condition parameter selection data is determined according to the battery testing accuracy requirements; the operating condition parameter selection data is evenly divided and selected within the battery operating condition parameter distribution threshold according to the amount of operating condition parameter selection data to obtain multiple battery operating condition test parameter sets; the multiple battery operating condition test parameter sets are arranged and integrated based on the battery scenario operating condition element set to obtain the battery operating condition test parameter table.

[0009] Optionally, a data filter is initialized based on the noise characteristics of the battery multi-dimensional data stream; the data filter is used to preprocess the battery multi-dimensional data stream to generate a usable battery multi-dimensional data stream; environmental features and discharge features are extracted from the usable battery multi-dimensional data stream in parallel to obtain a set of associated battery environmental features and a set of associated battery discharge features; the set of associated battery environmental features and the set of associated battery discharge features are then subjected to dimensionality reduction processing to obtain a set of battery environmental time-series features and a set of battery discharge waveform features.

[0010] Optionally, the associated battery environmental feature set and the associated battery discharge feature set are respectively decentralized to obtain the environmental feature covariance matrix and the discharge feature covariance matrix; using the environmental feature covariance matrix and the discharge feature covariance matrix, principal component analysis is performed on the associated battery environmental feature set and the associated battery discharge feature set to obtain the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set; key feature screening is performed on the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set to obtain the battery environmental time series feature set and the battery discharge waveform feature set.

[0011] Optionally, the battery environment time-series feature set and the battery discharge waveform feature set are synchronized, aligned, and normalized to obtain a standard battery environment time-series feature set and a standard battery discharge waveform feature set; an attention mechanism is introduced to calculate the similarity matrix of the standard battery environment time-series feature set and the standard battery discharge waveform feature set; and the standard battery environment time-series feature set and the standard battery discharge waveform feature set are weighted and fused based on the similarity matrix to obtain a discharge performance-environment interaction fusion feature set.

[0012] Optionally, a Softmax operation is performed on each row of the similarity matrix to obtain the environment-discharge attention weight and the discharge-environment attention weight; the standard battery environment time-series feature set and the standard battery discharge waveform feature set are weighted and fused based on the environment-discharge attention weight and the discharge-environment attention weight to obtain the discharge performance-environment interaction fusion feature set.

[0013] Optionally, a battery condition label is introduced as a condition input to determine the condition gating factor; the environment-discharge attention weight and the discharge-environment attention weight are dynamically corrected based on the condition gating factor.

[0014] Optionally, based on the battery performance index system, a set of battery performance evaluation indicators is extracted; based on the set of battery performance evaluation indicators, feature data association and analysis training fitting are performed to generate a set of sub-performance index analyzers; the set of sub-performance index analyzers is merged and integrated to construct the battery performance analyzer.

[0015] Secondly, this application also provides an analysis and testing apparatus for lithium polymer batteries, used to execute an analysis and testing method for lithium polymer batteries as described in the first aspect, wherein the analysis and testing apparatus for lithium polymer batteries includes: a scenario analysis module, used to extract operating conditions and analyze parameters of the application scenario information of the target lithium polymer battery to generate a battery operating condition test parameter table; an operating condition simulation test module, used to place the target lithium polymer battery on a tooling test platform according to the battery operating condition test parameter table to perform operating condition simulation testing and monitor and acquire multi-dimensional data streams of the battery; a feature extraction module, used to extract environmental features and discharge features from the multi-dimensional data streams of the battery in parallel to obtain a battery environmental time-series feature set and a battery discharge waveform feature set; a weighted fusion module, used to introduce an attention mechanism to weightedly fuse the battery environmental time-series feature set and the battery discharge waveform feature set to determine a discharge performance-environment interaction fusion feature set; and a performance evaluation module, used to construct a battery performance analyzer according to the battery performance index system, and to evaluate and analyze the discharge performance-environment interaction fusion feature set based on the battery performance analyzer to generate lithium polymer battery test analysis results.

[0016] One or more technical solutions provided in this application have at least the following beneficial effects: By extracting operating conditions and parsing parameters from the application scenario information of the target lithium polymer battery, a battery operating condition test parameter table is generated. The target lithium polymer battery is then placed on a tooling test platform for operating condition simulation testing according to the battery operating condition test parameter table, and multi-dimensional data streams of the battery are monitored and acquired. Environmental features and discharge features are extracted in parallel from the multi-dimensional data streams of the battery, resulting in a battery environmental time-series feature set and a battery discharge waveform feature set. An attention mechanism is introduced to weightedly fuse the battery environmental time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set. A battery performance analyzer is constructed based on the battery performance index system, and the discharge performance-environment interaction fusion feature set is evaluated and analyzed based on the battery performance analyzer to generate lithium polymer battery test analysis results. In other words, by analyzing the application scenario of the target lithium polymer battery, extracting real operating conditions for simulation testing, extracting environmental and discharge features of the battery in parallel, introducing an attention mechanism for weighted fusion, and evaluating and analyzing based on the battery performance analyzer, lithium polymer battery test analysis results are obtained, improving the accuracy and efficiency of battery testing, thereby improving the safety and reliability of lithium polymer batteries.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of an analytical testing method for lithium polymer batteries according to this application.

[0020] Figure 2 This is a schematic diagram of the structure of an analytical testing device for lithium polymer batteries according to this application.

[0021] Figure labeling: Scene analysis module 11, working condition simulation test module 12, feature extraction module 13, weighted fusion module 14, performance evaluation module 15. Detailed Implementation

[0022] This application provides an analytical testing method and apparatus for lithium polymer batteries, solving the technical problem in existing technologies where insufficient testing accuracy is caused by inadequate consideration of the complexity of application scenarios, thus affecting the safety of lithium polymer batteries. By analyzing the application scenarios of the target lithium polymer battery, simulating real-world operating conditions, and extracting environmental and discharge characteristics of the battery in parallel, an attention mechanism is introduced for weighted fusion, and evaluation and analysis are performed based on a battery performance analyzer to obtain the lithium polymer battery test analysis results. This improves the accuracy and efficiency of battery testing, thereby enhancing the safety and reliability of lithium polymer batteries.

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0024] Example 1, please refer to the appendix. Figure 1 This application provides an analytical testing method for lithium polymer batteries, wherein the analytical testing method for lithium polymer batteries is executed by an analytical testing apparatus for lithium polymer batteries, and the analytical testing method for lithium polymer batteries specifically includes the following steps: S100: Extracts operating conditions and analyzes parameters for the application scenario information of the target lithium polymer battery, generating a battery operating condition test parameter table.

[0025] Furthermore, this application S100 includes: The application scenario information of the target lithium polymer battery is classified, extracted, and analyzed to obtain a set of battery application scenario types and a set of scenario application requirement parameters. Operating condition elements are extracted from the battery application scenario type set according to the set of scenario application requirement parameters to determine the battery scenario operating condition element set. Historical data mining is performed on the battery scenario operating condition element set to obtain a historical operating data set of battery operating condition elements. Parameter parsing is performed on the battery scenario operating condition element set based on the historical operating data set of battery operating condition elements to obtain a battery operating condition test parameter table.

[0026] Furthermore, this application also includes the following steps: Based on the historical operating dataset of the battery operating condition elements, the parameter distribution of the battery scenario operating condition element set is analyzed to obtain the battery operating condition parameter distribution threshold; according to the battery testing accuracy requirements, the amount of data selected for operating condition parameters is determined; the data selected for operating condition parameters is evenly divided and selected within the battery operating condition parameter distribution threshold according to the amount of data selected for operating condition parameters to obtain multiple battery operating condition test parameter sets; based on the battery scenario operating condition element set, the multiple battery operating condition test parameter sets are arranged and integrated to obtain the battery operating condition test parameter table.

[0027] Specifically, this involves acquiring application scenario information for the target lithium polymer battery, i.e., the environment and usage patterns of the target lithium polymer battery in actual use, such as temperature, humidity, and usage mode (continuous or intermittent use). A detailed analysis is conducted on each application scenario of the target lithium polymer battery, identifying and classifying these scenarios, for example, categorizing application scenarios into high-temperature environments, low-temperature environments, and fast-charging scenarios. For each scenario, specific requirement parameters that the battery must meet are determined, resulting in a battery application scenario type set and a scenario application requirement parameter set. The battery application scenario type set, obtained through classification and extraction, includes typical application scenarios of different types that the target lithium polymer battery may face, such as high-temperature environments, low-temperature environments, fast charging, high-rate discharge, and standby. The scenario application requirement parameter set, obtained through requirement analysis, includes specific quantitative requirements for battery performance for each application scenario type, such as discharge / charge current range, temperature range, duration, capacity threshold, and upper limit of internal resistance. For example, the application requirements for high-temperature environments include maximum discharge current (e.g., 1500mA), upper temperature tolerance limit (e.g., 60℃), capacity retention rate threshold (e.g., ≥85%), and thermal runaway initiation temperature (e.g., ≥75℃). The application requirements for low-temperature environments include low-temperature start-up capability (e.g., discharge efficiency ≥50% at -30℃), capacity retention rate (e.g., capacity ≥30% of nominal capacity at -40℃), and charging limitations (e.g., charging current must be reduced to below 0.1C at <0℃).

[0028] Based on the parameter set of application scenario requirements, the battery application scenario type set is processed by extracting operating condition elements. From the parameter set, key variables representing the battery's operating state under each application scenario are identified and extracted, resulting in a battery scenario operating condition element set. This set includes all key operating condition elements corresponding to each application scenario, such as temperature, current mode (including peak and base values), cycle time, and number of cycles. For example, from the parameters of 25.6A to 34.8A, voltage drop 0.18V±0.05V, and temperature −25.0℃ to −5.3℃ in the low-temperature start-up scenario, two levels {28A, 32A} and {−20℃, −10℃} are extracted, along with three voltage drop elements after start-up: 0.13V / 0.18V / 0.23V. All scenarios and all extracted discrete element points are organized and summarized to form the battery scenario operating condition element set.

[0029] Historical data mining is performed based on a set of battery scenario operating conditions. This involves batch querying, cleaning, and feature extraction of existing operating logs and experimental records (such as laboratory test databases) to reconstruct the actual operating performance of the battery under each combination of operating conditions. The historical operating dataset of battery operating conditions includes detailed performance data recorded by the target lithium polymer battery during operation in different battery application scenarios, including current, voltage, temperature, SOC, internal resistance, and discharge capacity for each operation.

[0030] Based on the historical operating dataset of battery operating conditions, parameter distribution analysis is performed on the battery scenario operating condition element set. This involves statistically analyzing all values ​​of each key parameter in the historical operating dataset, calculating the statistics (mean, maximum, minimum, standard deviation, percentile, etc.) for each key parameter, thereby establishing the statistical boundaries of its numerical changes. Based on the parameter distribution analysis results, a reasonable value range is determined for each operating condition parameter. For example, the mean of the discharge current is calculated to be 20A, with a standard deviation of 1.2A, resulting in a distribution threshold of [18.5A, 21.5A]; the mean of the ambient temperature is 52℃, with a standard deviation of 1.8℃, resulting in a distribution threshold of [48℃, 56℃]; the mean of the duration is 27min, with a standard deviation of 3.1min, resulting in a distribution threshold of [20min, 34min]; and the initial SOC is 70% to 90%, with an ending SOC of 30% to 50%.

[0031] Based on the specific application and testing objectives of the target lithium polymer battery (such as quality control, life assessment, extreme performance testing, etc.), specific standards are proposed for the accuracy, reliability, and representativeness of the test results, resulting in battery testing accuracy requirements, including the accuracy of each parameter. Based on these accuracy requirements, the amount of data selected for operating condition parameters is determined, i.e., the number of test parameters selected within the parameter distribution threshold, to ensure the accuracy and representativeness of the test. Within the battery operating condition parameter distribution threshold, test parameters are selected by uniformly dividing the data according to the amount of data selected for the operating condition parameters. In other words, within the battery operating condition parameter distribution threshold, each parameter is equally compensated and divided to form multiple sets of battery operating condition test parameters, ensuring that the test conditions cover the entire parameter distribution range and improving the comprehensiveness and representativeness of the test. For example, assuming the current distribution threshold is [18.5A, 21.5A], its total range is 3. Based on the accuracy requirements, 3 data points are determined, with a calculation interval of 1.5A. Therefore, the points after uniform selection are 18.5A, 20A, and 21.5A. Through uniform division and selection, a set of specific test values ​​is obtained for each operating condition parameter (such as current, temperature, etc.) that needs to be tested. These parameter values ​​are combined to form multiple sets of battery operating condition test parameters.

[0032] Based on the battery scenario operating condition element set, multiple battery operating condition test parameter sets are arranged and integrated. In other words, the previously obtained multiple battery operating condition test parameter sets are systematically combined and arranged according to the operating condition elements they represent (such as current and temperature) to form a structured list or table. The battery operating condition test parameter table refers to a parameter table that can be used for testing, formed by arranging and combining all representative test parameter sets in a specified format, containing all test parameters and their values.

[0033] Starting from practical application scenarios and combining the accuracy requirements of battery testing, a battery operating condition test parameter table is obtained through operating condition element extraction and historical data mining. This overcomes the shortcomings of traditional testing methods that may be too idealistic or detached from reality. Through a data-driven approach, it ensures that the test operating conditions can truly reflect the battery's performance in actual use, thereby improving the accuracy and effectiveness of the test.

[0034] S200: The target lithium polymer battery is placed on a tooling test platform for operating condition simulation testing according to the battery operating condition test parameter table, and the multi-dimensional data stream of the battery is monitored and acquired.

[0035] Specifically, a tooling test platform is an experimental platform that simulates various complex operating conditions. It typically includes a constant temperature chamber, electronic load, battery tester (such as Neware or Arbin), thermocouple acquisition module, DAQ module, CAN communication interface, and other devices. It is used to simulate various operating conditions of the target lithium polymer battery in a controlled environment and to conduct multi-dimensional and multi-stage testing and monitoring of battery performance. The constant temperature chamber is a sealed space that maintains the internal environment within a set temperature range (e.g., -40℃ to +80℃) through a temperature control system, simulating the battery's operating state under high and low temperature environments. The electronic load is a programmable load device used to simulate the battery's discharge process, performing discharge tests by setting constant current, constant voltage, constant power, or dynamic operating conditions (e.g., pulse current). The battery tester has multi-channel charge and discharge functions, capable of performing charging, discharging, and cycle life tests, and collecting key data such as voltage, current, capacity, and energy. The thermocouple acquisition module is responsible for collecting temperature data from the battery surface or interior, typically using K-type or T-type thermocouples with an accuracy of ±0.1℃ and a sampling frequency supporting 1Hz to 10Hz. The DAQ module is a data acquisition system used to integrate and synchronize sensor signals such as voltage, current, temperature, and displacement, and can be connected to a host computer for real-time curve display and data recording. The CAN communication interface is used for data exchange between the battery management system (BMS) and the test platform, reading status data such as SOC, SOH, alarm status, and temperature of the battery pack.

[0036] The target lithium polymer battery is connected to the battery tester, and the discharge conditions are jointly controlled with the electronic load. An embedded constant temperature chamber is used to simulate the temperature conditions. The temperature is recorded by attaching the battery cell with a thermocouple and adding a DAQ module. The BMS is connected via a CAN bus to obtain the SOC / SOH status in real time. The host computer software integrates the data from all modules and records and visualizes it in a unified manner.

[0037] For example, if testing the performance of a target lithium polymer battery at a low temperature of -20°C and a high rate of discharge, the temperature of the constant temperature chamber is set to -20°C. After the battery is placed in the chamber, it is left to stand for 2 hours to equalize the temperature. The battery is then fully charged to 100% SOC, and the electronic load is started to perform a 10A (2C) constant current discharge. Thermocouples record the surface temperature, and DAQ records the data per second. The battery tester records the voltage, current, and SOC changes. The internal fault status and cell consistency data are synchronized in real time through the CAN interface with the BMS. The discharge is terminated at 3.0V, and the discharge capacity, maximum voltage drop, and maximum temperature rise are recorded. The following are some sample data obtained from the recording: At 0s, the voltage is 4.2V, the current is 0A, the temperature is -19.8℃, the SOC is 100%, and the internal resistance is 4.2mΩ; at 60s, the voltage is 3.88V, the current is 10A, the temperature is -18.0℃, the SOC is 100%, and the internal resistance is 4.2mΩ; at 300s, the voltage is 3.51V, the current is 10A, the temperature is -13.4℃, the SOC is 76.3%, and the internal resistance is 5.1mΩ; At 600s, the voltage was 3.23V, the current was 10A, the temperature was -10℃, the SOC was 52.5%, and the internal resistance was 5.4mΩ; at 900s, the voltage was 3.01V, the current was 10A, the temperature was -8.2℃, the SOC was 28.9%, and the internal resistance was 5.6mΩ. Data analysis revealed that the discharge capacity decreased to 4180mAh (approximately 16% less than at room temperature); the temperature rise was significantly slower, indicating a low chemical reaction rate; and the internal resistance increased over time, affecting the discharge capacity.

[0038] According to each test condition in the battery operating condition test parameter table, the target lithium polymer battery is precisely installed onto the tooling test platform. The tooling test platform operates by precisely controlling the test environment (e.g., temperature control via a temperature chamber) and applying loads (e.g., controlling charging / discharging current and voltage via a power supply / load unit). It also records various key performance indicators of the battery in real time through its configured sensors (e.g., voltage, current, and temperature sensors), forming a multi-dimensional battery data stream. This multi-dimensional data stream refers to the various data types of battery performance collected during the test, such as voltage, current, temperature, and pressure. Through operating condition simulation testing, the performance of the battery in real-world application scenarios is realistically simulated, comprehensively reflecting the performance changes of the battery during the testing process.

[0039] S300: Perform parallel extraction of environmental features and discharge features from the multi-dimensional data stream of the battery to obtain a battery environmental time-series feature set and a battery discharge waveform feature set.

[0040] Furthermore, this application S300 includes: Based on the noise characteristics of the battery multi-dimensional data stream, a data filter is initialized; the data filter is used to preprocess the battery multi-dimensional data stream to generate a usable battery multi-dimensional data stream; environmental features and discharge features are extracted from the usable battery multi-dimensional data stream in parallel to obtain a set of associated battery environmental features and a set of associated battery discharge features; the associated battery environmental features and the associated battery discharge features are then subjected to dimensionality reduction processing to obtain a set of battery environmental time-series features and a set of battery discharge waveform features.

[0041] Furthermore, this application also includes the following steps: The associated battery environmental feature set and associated battery discharge feature set are respectively decentered to obtain the environmental feature covariance matrix and the discharge feature covariance matrix. Using the environmental feature covariance matrix and the discharge feature covariance matrix, principal component analysis is performed on the associated battery environmental feature set and associated battery discharge feature set to obtain the dimensionality-reduced environmental feature set and dimensionality-reduced discharge feature set. Key feature screening is performed on the dimensionality-reduced environmental feature set and dimensionality-reduced discharge feature set to obtain the battery environmental time series feature set and the battery discharge waveform feature set.

[0042] Specifically, the multi-dimensional data stream of a battery typically contains various noises during acquisition, such as sensor noise, electromagnetic interference, and environmental fluctuations, which can interfere with the judgment of the battery's true performance. Analyzing the noise characteristics of the multi-dimensional battery data stream, including noise type, noise amplitude, and noise frequency distribution, is crucial. This involves determining whether there are periodic abrupt changes or abnormal high-frequency disturbances; calculating the standard deviation, kurtosis, and skewness of each dimension to identify the error distribution; and comparing the data stream to equipment accuracy, such as the allowable error of ±0.5℃ for temperature sensors and ±0.1A for current sampling errors.

[0043] Select an appropriate filter type based on the noise characteristics. For example, if the noise is mainly high-frequency, choose a low-pass filter; if the noise amplitude is large, choose a filter with strong noise suppression capabilities. Based on the noise characteristics and the selected filter type, set the filter parameters, such as cutoff frequency, filter order, and window function. In other words, configure suitable filters according to the data channel and noise type, and initialize the data filter.

[0044] The battery's multi-dimensional data stream is preprocessed using an initialized data filter to remove noise components and smooth the data curves. The processed data stream retains its main characteristics while significantly reducing noise levels. For example, the original temperature data segment (unit: °C) obtained at a sampling frequency of 1 Hz: [25.3, 26.0, 27.9, 26.2, 28.5, 26.5] shows a clear jump. After median filtering, the temperature data becomes: [25.3, 26.0, 26.0, 26.2, 26.2, 26.5] (removing abrupt changes).

[0045] Parallel extraction of environmental and discharge features is performed from the filtered multi-dimensional data stream of available batteries. This involves simultaneously extracting environmental and discharge features from the data stream to form a set of associated battery environmental features and a set of associated battery discharge features. Features related to the battery's operating environment, such as ambient temperature, temperature change rate, air pressure, and test chamber temperature control stability, are extracted from the filtered multi-dimensional data stream. Simultaneously, features directly related to the battery discharge process are extracted from the filtered multi-dimensional data stream, such as discharge capacity, average discharge voltage, slope of the discharge curve, discharge time, discharge power, and current change patterns.

[0046] Decentralization is performed on the associated battery environmental feature set and the associated battery discharge feature set separately. This involves subtracting the mean of each feature from each set to construct a covariance matrix. Decentralization, which sets the mean of the data to zero, is a step in data standardization preprocessing and helps eliminate the influence of bias. The covariance matrix is ​​a square matrix used to describe the covariance relationship between multiple variables. The diagonal elements of the covariance matrix are the variances of the individual variables, and the off-diagonal elements are the covariances between different variables.

[0047] Principal component analysis (PCA) was performed on the environmental feature set and the discharge feature set of the correlated batteries using the environmental feature covariance matrix and the discharge feature covariance matrix, resulting in dimensionality-reduced environmental feature set and dimensionality-reduced discharge feature set. PCA is a commonly used data dimensionality reduction technique that transforms a set of potentially correlated variables into a set of linearly uncorrelated variables through orthogonal transformation; these new variables are called principal components. Principal components are linear combinations of the original variables and are arranged in descending order of variance. Eigenvalue decomposition was performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvectors corresponding to the k largest eigenvalues ​​were selected to construct a new feature space. The original data was projected onto this new feature space composed of the selected eigenvectors to obtain the dimensionality-reduced feature set. The dimensionality-reduced feature set retains the main variance information of the original data while reducing the dimensionality of the features.

[0048] Dimensionality reduction of the environmental feature set refers to transforming the associated battery environmental feature set into a new set of linearly uncorrelated features through principal component analysis (PCA), preserving the variance information of the original data to the greatest extent. Dimensionality reduction of the discharge feature set refers to transforming the associated battery discharge feature set into a new set of linearly uncorrelated features through PCA, preserving the variance information of the original data to the greatest extent. Principal component analysis (PCA) is a commonly used dimensionality reduction technique. It uses the covariance matrix to find a new set of orthogonal coordinate axes (principal components), which are the result of linear combinations of the original features, and can preserve the variance (information content) of the original data to the greatest extent. By projecting the original features onto these principal components, the original data can be approximated with fewer principal components, thus achieving dimensionality reduction. PCA is performed on the associated battery environmental feature set and the associated battery discharge feature set respectively to obtain their dimensionality-reduced representations, namely the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set.

[0049] Although the dimensionality of the reduced feature set is lower, it still contains some features that do not contribute much to the final evaluation objective (such as battery health status and performance prediction). By setting thresholds (such as the contribution rate of principal components), key features are screened from the reduced environmental and discharge feature sets to select the most representative and important features for the objective. The selected key features will constitute the battery environmental time-series feature set and the battery discharge waveform feature set. In principal component analysis, the eigenvalue corresponding to each principal component represents the proportion of variance explained by that principal component. The larger the eigenvalue, the more original information the principal component contains. The correlation between each principal component and the key battery states (such as remaining capacity and health status) is calculated; principal components with high correlation are more critical. Screening criteria are set to retain principal components whose cumulative variance contribution rate reaches a certain threshold (such as 85% or 95%). Screening is performed according to the screening criteria to obtain the battery environmental time-series feature set and the battery discharge waveform feature set. For example, suppose the reduced environmental feature set has 3 principal components and the discharge feature set has 4 principal components. Variance contribution rate analysis shows that the variance contribution rates of the environmental principal components are 40%, 35%, and 25% (cumulative 100%), while those of the discharge principal components are 50%, 30%, 15%, and 5% (cumulative 100%). If the cumulative variance contribution rate threshold is set to 75%, the first two principal components of the environmental feature set are selected (cumulative 75%), and the first two principal components of the discharge feature set are selected (cumulative 80%).

[0050] Starting from the raw, noisy, multi-dimensional data stream, through filtering, parallel feature extraction, decentralization, PCA dimensionality reduction, and key feature screening, two highly condensed and information-rich feature sets were finally generated: the battery environment time series feature set and the battery discharge waveform feature set. These features capture the key performance characteristics and environmental impact of the battery under actual working conditions, focusing on the features that have the greatest impact on battery performance evaluation, thereby improving the accuracy and efficiency of performance evaluation.

[0051] S400: An attention mechanism is introduced to perform weighted fusion of the battery environment time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set.

[0052] Furthermore, this application S400 includes: The battery environment time-series feature set and the battery discharge waveform feature set are synchronized, aligned, and normalized to obtain a standard battery environment time-series feature set and a standard battery discharge waveform feature set. An attention mechanism is introduced to calculate the similarity matrix of the standard battery environment time-series feature set and the standard battery discharge waveform feature set. Based on the similarity matrix, the standard battery environment time-series feature set and the standard battery discharge waveform feature set are weighted and fused to obtain a discharge performance-environment interaction fusion feature set.

[0053] Furthermore, this application also includes the following steps: A Softmax operation is performed on each row of the similarity matrix to obtain the environment-discharge attention weight and the discharge-environment attention weight. Based on the environment-discharge attention weight and the discharge-environment attention weight, the standard battery environment time series feature set and the standard battery discharge waveform feature set are weighted and fused to obtain the discharge performance-environment interaction fusion feature set.

[0054] Specifically, the battery environmental time-series feature set and the battery discharge waveform feature set are synchronized and aligned. This involves aligning each data point in the two feature sets according to its timestamp, ensuring that each time point has a corresponding environmental and discharge feature. Synchronous alignment refers to aligning the two datasets by timestamp or other common dimensions, ensuring that environmental and discharge features correspond under the same time points or conditions. If the temporal resolution of a feature set is low, interpolation methods (such as linear interpolation, spline interpolation, etc.) are used to estimate the feature values ​​at missing time points. The two feature sets are then resampled to have the same temporal resolution. The aligned feature sets are then normalized, as the numerical ranges of environmental features (such as temperature, in degrees Celsius) and discharge features (such as voltage, in volts) may differ significantly. Normalization (such as scaling the data to a range of 0 to 1 or a mean of 0 and a standard deviation of 1) eliminates the influence of dimensions and numerical ranges, ensuring that the values ​​of different features are on similar scales. After normalization, a standard battery environmental time-series feature set and a standard battery discharge waveform feature set are obtained. For example, normalizing the temperature of 25.3℃ yields 0.585, and normalizing the voltage of 3.9V yields 0.643.

[0055] An attention mechanism is introduced to calculate the similarity between the standard battery environmental time-series feature set and the standard battery discharge waveform feature set. The similarity between each feature vector in the standard battery environmental time-series feature set and each feature vector in the standard battery discharge waveform feature set is calculated, resulting in a similarity matrix. The number of rows equals the number of features in the standard battery environmental time-series feature set, and the number of columns equals the number of features in the standard battery discharge waveform feature set. Each element represents the correlation strength between a specific feature in the standard battery environmental time-series feature set and a specific feature in the standard battery discharge waveform feature set. For example, the similarity between temperature and voltage is 0.873, temperature and current is 0.654, temperature and power is 0.781, humidity and voltage is 0.632, humidity and current is 0.479, humidity and power is 0.521, wind speed and voltage is 0.285, wind speed and current is 0.192, and wind speed and power is 0.243, etc. These elements are then combined to obtain the overall similarity matrix. A similarity matrix is ​​used to quantify the relationship between two feature sets, with each element representing the degree of similarity between a pair of features.

[0056] A Softmax operation is performed on each row of the similarity matrix, transforming the vector into a probability distribution where the sum of all elements is 1 and each element is between (0,1). Softmax is a normalization function commonly used in deep learning, converting a set of real numbers into a set of probability values ​​that sum to 1, focusing on high values ​​and suppressing low values. Applying Softmax to each row of the similarity matrix yields the environment-discharge attention weights, representing the degree of attention each environment feature receives from all discharge features; these are attention weights obtained from the row dimension (environment to discharge). The discharge-environment attention weights represent the degree of attention each discharge feature receives from all environment features; these are attention weights obtained from the column dimension (discharge to environment). For example, applying Softmax to the temperature row calculates the exponential term e. 0.873 =2.394, e 0.654 =1.923, e 0.781 =2.184, summing them up gives 6.501, thus the weights are calculated to be 0.368, 0.296, and 0.336.

[0057] The standard battery environmental time-series feature set and the standard battery discharge waveform feature set are weighted and fused based on environment-discharge attention weights and discharge-environment attention weights. This means that different features are combined according to attention weights to obtain a discharge performance-environment interaction fused feature set. The standard battery discharge waveform feature set is weighted using environment-discharge attention weights, and the standard battery environmental time-series feature set is weighted using discharge-environment attention weights. The two weighted results are combined to obtain the discharge performance-environment interaction fused feature set. Through data preprocessing (alignment, normalization), attention mechanisms (similarity calculation, softmax to obtain weights), and weighted fusion, the originally independent environmental time-series features and discharge waveform features are transformed into a fused feature set that can effectively represent the dynamic interaction between the two, improving the accuracy of test analysis. By introducing an attention mechanism, a deep coupling modeling between the battery environment and discharge behavior is achieved. Compared with traditional simple splicing or static weighting methods, this approach can better capture the correlation features of performance degradation in complex scenarios.

[0058] Furthermore, this application also includes the following steps: Battery operating condition labels are introduced as conditional inputs to determine the operating condition gating factor; the environment-discharge attention weight and discharge-environment attention weight are dynamically corrected based on the operating condition gating factor.

[0059] Specifically, the battery condition label describes the current or expected operating state of the target lithium polymer battery. It can be a category label (e.g., fast charging, high temperature environment, light load), or one or more numerical parameters (e.g., load current range, ambient temperature range, SOC range), representing the specific usage scenario or testing conditions of the target lithium polymer battery. Using the battery condition label as input, one or more values ​​are calculated using a predefined function (e.g., a simple lookup table), i.e., the condition gating factor. The condition gating factor is a parameter or set of parameters driven by the battery condition label, similar to a switch or regulator, adjusting or gating the values ​​of other parameters (i.e., attention weights) according to the current operating condition. If the condition label is discrete (e.g., normal, high load, high temperature), a vector is used as the gating factor, where each element of the vector corresponds to a condition; when a specific condition is activated, its corresponding element is 1, and others are 0. If the condition label contains a numerical value (e.g., load current > 10A), the gating factor is a function value based on that value.

[0060] The environmental-discharge attention weights and discharge-environment attention weights are dynamically adjusted using a condition-gating factor. This means adjusting the attention weight allocation based on the current condition label to highlight the features most relevant to the current condition. In other words, the condition-gating factor is combined with the original attention weights to adjust their values. The gating factor is multiplied by the original weights. For example, if the environment has a greater impact on discharge under high-load conditions, the gating factor can be a value greater than 1 to amplify the environmental-discharge attention weights. If there are multiple gating factors (corresponding to different sensitivity levels under different conditions), they are weighted and summed with the baseline weights for each condition, thus dynamically adjusting the environmental-discharge attention weights and discharge-environment attention weights. Under high-temperature conditions, the environmental-discharge attention weights are increased because temperature has a greater impact on discharge performance; while under low-temperature conditions, the discharge-environment attention weights are increased because the battery's discharge behavior (such as voltage drops) more sensitively reflects the effects of low temperatures. For example, after determining the operating condition gating factor, the average attention increased to 0.42, an improvement of 44.8%. After correcting the weights of the temperature row from 0.368, 0.296, and 0.336, the values ​​became 0.41, 0.58, and 0.65.

[0061] Gating mechanisms allow attention to dynamically adapt to specific operating conditions, effectively representing the complex response relationships of batteries under nonlinear conditions such as high temperature, low temperature, and high rate. Dynamically adjusting attention weights enables the model to automatically adjust the importance of features based on the current operating conditions, helping to improve the accuracy of performance evaluation under different conditions. By introducing operating condition gating factors, the model can flexibly adapt to various complex application scenarios, improving its generalization ability and practicality.

[0062] S500: Construct a battery performance analyzer based on the battery performance index system, and evaluate and analyze the discharge performance-environment interaction fusion feature set based on the battery performance analyzer to generate lithium polymer battery test analysis results.

[0063] Furthermore, this application S500 includes: Based on the battery performance index system, a set of battery performance evaluation indices is extracted; based on the set of battery performance evaluation indices, feature data association and analysis training and fitting are performed to generate a set of sub-performance index analyzers; the set of sub-performance index analyzers is merged and integrated to construct the battery performance analyzer.

[0064] Specifically, obtaining a battery performance index system includes metrics used to measure various battery performance characteristics, such as capacity, energy density, power, internal resistance, cycle life, rate performance, high and low temperature performance, and safety. This system is typically a predefined set of indicators built around the key performance aspects of the target lithium polymer battery. Extracting a set of battery performance evaluation indicators from this system involves selecting the most relevant and critical set of indicators based on specific application requirements or testing objectives. For example, if used to evaluate the battery's state of health (SOH), indicators such as capacity decay rate, internal resistance growth, and cycle life would be extracted.

[0065] Based on the battery performance evaluation metric set, historical battery multi-dimensional test data (environmental characteristics, discharge characteristics, etc.) is collected. For each performance metric, a training dataset is constructed, divided into training, validation, and test sets. A separate model is trained for each performance metric, allowing each model to focus more on that metric and potentially achieve higher evaluation accuracy. If the metric is a continuous value (e.g., capacity, internal resistance, decay rate), a regression model is selected, such as linear regression, ridge regression, support vector regression, random forest regression, etc. If the metric is a discrete category (e.g., state of health (SOH) level classification, fault type), a classification model is selected, such as logistic regression, support vector machine, decision tree, random forest classification, neural network, etc.

[0066] For each metric in the battery performance evaluation metric set, initialize an empty model object. Extract the column corresponding to the metric from the target label data. Train the empty model object using the input features from the training set and the corresponding target labels. During training, use the validation set to monitor model performance (e.g., mean squared error (MSE), R² score, etc.) and perform hyperparameter tuning (e.g., learning rate, number of trees, regularization parameters, etc.) to prevent overfitting and select the optimal model configuration. Use the test set to evaluate the final performance of the trained model, save the trained model, and obtain the performance metric analyzer corresponding to each metric, i.e., generate a set of sub-performance metric analyzers. The set of sub-performance metric analyzers is a set of individual performance analysis models trained for each evaluation metric.

[0067] A battery performance analyzer is constructed by merging and integrating a set of sub-performance index analyzers. A unified model architecture comprising multiple sub-modules is designed, with each sub-module corresponding to the function of a sub-analyzer, and these are integrated at the top level. The battery performance analyzer can receive multi-dimensional data from the battery and simultaneously or sequentially call / merge the results of various sub-analyzers to evaluate and analyze the overall performance of the battery. Through the battery performance analyzer, the discharge performance-environment interaction fusion feature set is evaluated and analyzed. The input features are mapped to the feature space corresponding to each sub-performance index analyzer within the battery performance analyzer. Each sub-performance index analyzer evaluates the features in parallel and outputs its own evaluation results. The evaluation results of all sub-performance index analyzers are then merged to form a comprehensive lithium polymer battery test analysis result, including performance index evaluation, comprehensive score, health status judgment, state trend prediction, and recommended measures.

[0068] By employing a predefined battery performance index system, the evaluation of battery performance is ensured to be based on a unified and standardized set of rules, avoiding subjectivity and arbitrariness. All evaluations are conducted around established indicators, guaranteeing consistency. The construction of a multi-dimensional performance evaluation model improves the accuracy of battery health status assessment and remaining life prediction. Particularly when the analyzer can integrate multi-dimensional information (such as discharge performance and environmental factors), the evaluation results better reflect the battery's true performance under actual operating conditions, reducing the bias that may arise from single-index evaluations. By comprehensively considering the impact of environmental factors (such as temperature and humidity) on battery performance, the evaluation results not only reflect the battery's intrinsic state but also its performance under specific environmental conditions, making them more valuable for practical applications.

[0069] In summary, the analytical testing method for lithium polymer batteries provided in this application has the following beneficial effects: By extracting operating conditions and parsing parameters from the application scenario information of the target lithium polymer battery, a battery operating condition test parameter table is generated. The target lithium polymer battery is then placed on a tooling test platform for operating condition simulation testing according to the battery operating condition test parameter table, and multi-dimensional data streams of the battery are monitored and acquired. Environmental features and discharge features are extracted in parallel from the multi-dimensional data streams of the battery, resulting in a battery environmental time-series feature set and a battery discharge waveform feature set. An attention mechanism is introduced to weightedly fuse the battery environmental time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set. A battery performance analyzer is constructed based on the battery performance index system, and the discharge performance-environment interaction fusion feature set is evaluated and analyzed based on the battery performance analyzer to generate lithium polymer battery test analysis results. In other words, by analyzing the application scenario of the target lithium polymer battery, extracting real operating conditions for simulation testing, extracting environmental and discharge features of the battery in parallel, introducing an attention mechanism for weighted fusion, and evaluating and analyzing based on the battery performance analyzer, lithium polymer battery test analysis results are obtained, improving the accuracy and efficiency of battery testing, thereby improving the safety and reliability of lithium polymer batteries.

[0070] Example 2: Based on the same inventive concept as the analytical testing method for lithium polymer batteries in Example 1, this application also provides an analytical testing apparatus for lithium polymer batteries. Please refer to the appendix. Figure 2 The analytical testing apparatus for lithium polymer batteries includes: The scenario analysis module 11 is used to extract operating conditions and analyze parameters of the application scenario information of the target lithium polymer battery to generate a battery operating condition test parameter table; the operating condition simulation test module 12 is used to place the target lithium polymer battery on the tooling test platform according to the battery operating condition test parameter table to conduct operating condition simulation tests and monitor and acquire multi-dimensional data streams of the battery; the feature extraction module 13 is used to extract environmental features and discharge features from the multi-dimensional data streams of the battery in parallel to obtain a battery environmental time-series feature set and a battery discharge waveform feature set; the weighted fusion module 14 is used to introduce an attention mechanism to perform weighted fusion of the battery environmental time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set; the performance evaluation module 15 is used to construct a battery performance analyzer according to the battery performance index system, and evaluate and analyze the discharge performance-environment interaction fusion feature set based on the battery performance analyzer to generate lithium polymer battery test analysis results.

[0071] Furthermore, the scenario analysis module 11 in the analytical testing device for lithium polymer batteries is also used for: classifying and extracting application scenario information of the target lithium polymer battery and performing demand analysis to obtain a set of battery application scenario types and a set of scenario application demand parameters; extracting operating condition elements from the set of battery application scenario types according to the set of scenario application demand parameters to determine a set of battery scenario operating condition elements; performing historical data mining based on the set of battery scenario operating condition elements to obtain a historical operating dataset of battery operating condition elements; and performing parameter analysis on the set of battery scenario operating condition elements based on the historical operating dataset of battery operating condition elements to obtain a battery operating condition test parameter table.

[0072] Furthermore, the scenario analysis module 11 in the analytical testing device for lithium polymer batteries is also used for: performing parameter distribution analysis on the battery scenario operating condition element set based on the historical operating data set of the battery operating condition elements to obtain the battery operating condition parameter distribution threshold; determining the amount of operating condition parameter selection data according to the battery testing accuracy requirements; uniformly dividing and selecting the operating condition parameter selection data within the battery operating condition parameter distribution threshold according to the amount of operating condition parameter selection data to obtain multiple battery operating condition test parameter sets; and arranging and integrating the multiple battery operating condition test parameter sets based on the battery scenario operating condition element set to obtain the battery operating condition test parameter table.

[0073] Furthermore, the feature extraction module 13 in the analytical testing device for lithium polymer batteries is also used to: initialize a data filter based on the noise characteristics of the battery multi-dimensional data stream; perform filtering preprocessing on the battery multi-dimensional data stream using the data filter to generate a usable battery multi-dimensional data stream; extract environmental features and discharge features from the usable battery multi-dimensional data stream in parallel to obtain a set of associated battery environmental features and a set of associated battery discharge features; and perform dimensionality reduction processing on the set of associated battery environmental features and the set of associated battery discharge features to obtain a set of battery environmental time-series features and a set of battery discharge waveform features.

[0074] Furthermore, the feature extraction module 13 in the analytical testing device for lithium polymer batteries is also used to: perform decentralization processing on the associated battery environmental feature set and the associated battery discharge feature set respectively to obtain the environmental feature covariance matrix and the discharge feature covariance matrix; use the environmental feature covariance matrix and the discharge feature covariance matrix to perform principal component analysis on the associated battery environmental feature set and the associated battery discharge feature set to obtain the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set; and perform key feature screening on the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set to obtain the battery environmental time series feature set and the battery discharge waveform feature set.

[0075] Furthermore, the weighted fusion module 14 in the analytical testing device for lithium polymer batteries is also used to: synchronously align and normalize the battery environment time-series feature set and the battery discharge waveform feature set to obtain a standard battery environment time-series feature set and a standard battery discharge waveform feature set; introduce an attention mechanism to calculate the similarity matrix of the standard battery environment time-series feature set and the standard battery discharge waveform feature set; and perform weighted fusion of the standard battery environment time-series feature set and the standard battery discharge waveform feature set based on the similarity matrix to obtain a discharge performance-environment interaction fusion feature set.

[0076] Furthermore, the weighted fusion module 14 in the analysis and testing device for lithium polymer batteries is also used to: perform a Softmax operation on each row of the similarity matrix to obtain environment-discharge attention weights and discharge-environment attention weights; and perform weighted fusion on the standard battery environment time series feature set and the standard battery discharge waveform feature set based on the environment-discharge attention weights and discharge-environment attention weights to obtain the discharge performance-environment interaction fusion feature set.

[0077] Furthermore, the weighted fusion module 14 in the analytical testing device for lithium polymer batteries is also used to: introduce battery operating condition labels as condition inputs to determine operating condition gating factors; and dynamically correct the environment-discharge attention weights and discharge-environment attention weights based on the operating condition gating factors.

[0078] Furthermore, the performance evaluation module 15 in the analytical testing device for lithium polymer batteries is also used for: extracting a set of battery performance evaluation indicators according to the battery performance indicator system; performing feature data association and analysis training fitting based on the set of battery performance evaluation indicators to generate a set of sub-performance indicator analyzers; and merging and integrating the set of sub-performance indicator analyzers to construct the battery performance analyzer.

[0079] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Figure 1 The analytical testing method and specific example for lithium polymer batteries in Example 1 are also applicable to the analytical testing apparatus for lithium polymer batteries in this embodiment. Through the foregoing detailed description of the analytical testing method for lithium polymer batteries, those skilled in the art can clearly understand the analytical testing apparatus for lithium polymer batteries in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0080] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily 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 this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0081] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. An analytical testing method for lithium polymer batteries, characterized in that, include: The application scenario information of the target lithium polymer battery is extracted and the parameters are analyzed to generate a battery operating condition test parameter table. According to the battery operating condition test parameter table, the target lithium polymer battery is placed on the tooling test platform for operating condition simulation test, and the multi-dimensional data stream of the battery is monitored and acquired. In parallel, environmental and discharge features are extracted from the multi-dimensional data stream of the battery to obtain a battery environmental time-series feature set and a battery discharge waveform feature set. An attention mechanism is introduced to weightedly fuse the battery environment time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set; A battery performance analyzer is constructed based on the battery performance index system, and the discharge performance-environment interaction fusion feature set is evaluated and analyzed based on the battery performance analyzer to generate lithium polymer battery test analysis results.

2. The analytical testing method for lithium polymer batteries as described in claim 1, characterized in that, The generated battery operating condition test parameter table includes: The application scenario information of the target lithium polymer battery is classified, extracted, and analyzed to obtain a set of battery application scenario types and a set of scenario application requirement parameters. Based on the application requirement parameter set of the scenario, the working condition element set of the battery application scenario type set is extracted to determine the battery scenario working condition element set. Historical data mining is performed on the battery scenario operating condition element set to obtain a historical operating data set of battery operating condition elements. Based on the historical operation dataset of the battery operating condition elements, the parameters of the battery scenario operating condition element set are parsed to obtain the battery operating condition test parameter table.

3. The analytical testing method for lithium polymer batteries as described in claim 2, characterized in that, The obtained battery operating condition test parameter table includes: Based on the historical operation dataset of the battery operating condition elements, the parameter distribution of the battery scenario operating condition element set is analyzed to obtain the battery operating condition parameter distribution threshold. Based on the battery testing accuracy requirements, determine the amount of data to be selected for the operating condition parameters; According to the operating condition parameters, the data volume is evenly divided within the battery operating condition parameter distribution threshold to obtain multiple battery operating condition test parameter sets. Based on the battery scenario operating condition element set, the multiple battery operating condition test parameter sets are arranged and integrated to obtain the battery operating condition test parameter table.

4. The analytical testing method for lithium polymer batteries as described in claim 1, characterized in that, The obtained battery environment time-series feature set and battery discharge waveform feature set include: Initialize the data filter based on the noise characteristics of the battery's multi-dimensional data stream; The data filter is used to preprocess the multi-dimensional data stream of the battery to generate a usable multi-dimensional data stream of the battery. In parallel, environmental features and discharge features are extracted from the multi-dimensional data stream of the available batteries to obtain the associated battery environmental feature set and the associated battery discharge feature set. The associated battery environment feature set and associated battery discharge feature set are subjected to dimensionality reduction processing to obtain the battery environment time series feature set and the battery discharge waveform feature set.

5. The analytical testing method for lithium polymer batteries as described in claim 4, characterized in that, The process of obtaining the battery environment time-series feature set and the battery discharge waveform feature set includes: The associated battery environmental feature set and associated battery discharge feature set are respectively decentralized to obtain the environmental feature covariance matrix and the discharge feature covariance matrix; Using the environmental feature covariance matrix and the discharge feature covariance matrix, principal component analysis is performed on the associated battery environmental feature set and the associated battery discharge feature set to obtain the dimensionality-reduced environmental feature set and the dimensionality-reduced discharge feature set. Key features are filtered from the reduced-dimensional environment feature set and the reduced-dimensional discharge feature set to obtain the battery environment time-series feature set and the battery discharge waveform feature set.

6. The analytical testing method for lithium polymer batteries as described in claim 1, characterized in that, The determined discharge performance-environment interaction fusion feature set includes: The battery environment time series feature set and the battery discharge waveform feature set are synchronized, aligned and normalized to obtain the standard battery environment time series feature set and the standard battery discharge waveform feature set. An attention mechanism is introduced to calculate the similarity matrix between the standard battery environment time-series feature set and the standard battery discharge waveform feature set; Based on the similarity matrix, the standard battery environment time series feature set and the standard battery discharge waveform feature set are weighted and fused to obtain the discharge performance-environment interaction fusion feature set.

7. The analytical testing method for lithium polymer batteries as described in claim 6, characterized in that, The obtained discharge performance-environment interaction fusion feature set includes: Perform a Softmax operation on each row of the similarity matrix to obtain the environment-discharge attention weight and the discharge-environment attention weight; The standard battery environment time-series feature set and the standard battery discharge waveform feature set are weighted and fused based on the environment-discharge attention weight and the discharge-environment attention weight to obtain the discharge performance-environment interactive fusion feature set.

8. The analytical testing method for lithium polymer batteries as described in claim 7, characterized in that, The analytical testing method for lithium polymer batteries further includes: Battery operating condition labels are introduced as conditional inputs to determine the operating condition gating factor; The environment-discharge attention weight and discharge-environment attention weight are dynamically corrected based on the operating condition gating factor.

9. The analytical testing method for lithium polymer batteries as described in claim 1, characterized in that, The construction of the battery performance analyzer based on the battery performance index system includes: Based on the battery performance index system, a set of battery performance evaluation indicators is extracted and obtained. Based on the battery performance evaluation index set, feature data association and analysis training and fitting are performed to generate a set of sub-performance index analyzers. The set of sub-performance index analyzers is merged and integrated to construct the battery performance analyzer.

10. An analytical testing apparatus for lithium polymer batteries, characterized in that, The step of implementing the analytical testing method for lithium polymer batteries according to any one of claims 1 to 9, wherein the analytical testing apparatus for lithium polymer batteries comprises: The scenario analysis module is used to extract operating conditions and parse parameters of the target lithium polymer battery's application scenario information, and generate a battery operating condition test parameter table. The working condition simulation test module is used to place the target lithium polymer battery on the tooling test platform for working condition simulation test according to the battery working condition test parameter table, and monitor and acquire multi-dimensional data streams of the battery. The feature extraction module is used to extract environmental features and discharge features from the multi-dimensional data stream of the battery in parallel, so as to obtain the battery environmental time series feature set and the battery discharge waveform feature set. The weighted fusion module is used to introduce an attention mechanism to weight and fuse the battery environment time-series feature set and the battery discharge waveform feature set to determine the discharge performance-environment interaction fusion feature set; The performance evaluation module is used to construct a battery performance analyzer based on the battery performance index system, and to evaluate and analyze the discharge performance-environment interaction fusion feature set based on the battery performance analyzer, thereby generating lithium polymer battery test analysis results.

Images