Lithium battery detection method and device based on charge and discharge curve analysis

Abstract

The invention relates to the technical field of battery management, in particular to a lithium battery detection method and device based on charge and discharge curve analysis. The method comprises the following steps: carrying out a charge-discharge test on a target object based on a preset standard test signal to obtain first charge-discharge monitoring data; analyzing the first charge and discharge monitoring data based on a pre-constructed analysis model to obtain a quality detection result associated with the first charge and discharge monitoring data; in response to a predicted fault of which the probability index exceeds a preset fault probability threshold in the quality detection result, generating a characteristic excitation signal matched with the predicted fault; performing an excitation test on the target object based on the characteristic excitation signal to obtain second charging and discharging monitoring data; and performing verification analysis on the second charging and discharging monitoring data based on the analysis model, and updating the quality detection result according to the verification analysis result to obtain the verified and updated quality detection result. By adopting the method, the accuracy of battery detection can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery management, and in particular to a lithium battery detection method and device based on charge-discharge curve analysis, a computer device, a storage medium and a computer program product. BACKGROUND

[0002] Lithium-ion batteries (LIBs) have become the core energy devices for modern portable electronic devices, electric vehicles and energy storage systems due to their high energy density, long cycle life, low self-discharge rate and other advantages. With the growing global demand for clean energy and sustainable development, the market size of lithium batteries continues to expand and is widely used in smartphones, laptops, power tools, new energy vehicles and grid energy storage. The performance and safety of lithium batteries are highly dependent on their manufacturing process, which mainly includes electrode preparation, cell assembly, liquid injection, formation, aging and capacity distribution. Among them, electrode preparation mainly mixes active materials with conductive agents and binders to form a slurry, which is uniformly coated on a metal current collector, dried, rolled and then cut into electrode sheets; cell assembly mainly stacks or winds the positive and negative electrode sheets with the separator, and then assembles them into an aluminum plastic film or a metal shell, welds the tabs and packages them; liquid injection mainly injects electrolyte into an organic solvent to fully soak the electrode material in a dry environment; the formation process mainly performs the first charge and discharge of the battery to activate the battery and form a stable solid electrolyte interface (SEI) film; the aging and capacity distribution process is to stabilize the performance of the battery and then select the battery with qualified capacity and internal resistance through charge and discharge test.

[0003] In related technologies, in order to ensure the product quality of lithium batteries, quality detection such as electrochemical test and physical characterization test is usually performed on lithium batteries, for example, by analyzing the charge-discharge curve to obtain a quality analysis result.

[0004] However, the current battery detection method has the following technical problems:

[0005] The traditional detection method relies on charge-discharge curve analysis, but the charge-discharge curve is affected by environmental factors and other factors, and there are many noises such as oscillation, which leads to the accuracy of the detection result to be optimized. SUMMARY

[0006] Therefore, it is necessary to provide a lithium battery detection method, device, computer device, computer readable storage medium and computer program product based on charge-discharge curve analysis, which can improve the accuracy of battery detection.

[0007] In a first aspect, the present application provides a lithium battery detection method based on charge-discharge curve analysis. The method comprises:

[0008] performing a charge-discharge test on the target object based on a preset standard test signal to obtain first charge-discharge monitoring data;

[0009] analyzing the first charge-discharge monitoring data based on a pre-constructed analysis model to obtain a quality detection result associated with the first charge-discharge monitoring data;

[0010] in response to a predicted fault in which a probability indicator in the quality detection result exceeds a preset fault probability threshold, generating a characteristic excitation signal matched with the predicted fault;

[0011] performing an excitation test on the target object based on the characteristic excitation signal to obtain second charge-discharge monitoring data;

[0012] performing verification analysis on the second charge-discharge monitoring data based on the analysis model, and updating the quality detection result according to a result of the verification analysis to obtain a verification-updated quality detection result.

[0013] In one of the embodiments, the response to the predicted fault in which the probability indicator in the quality detection result exceeds the preset fault probability threshold includes:

[0014] determining a plurality of fault types and corresponding probability indicators according to the quality detection result, and screening out the predicted fault based on the preset fault probability threshold;

[0015] matching the predicted fault according to a pre-constructed test rule library to determine excitation signal parameters and test environment parameters associated with the predicted fault, the excitation signal parameters including a time-domain dynamic signal sequence in a plurality of battery cycle periods, and the test environment parameters including an environment temperature and a state of charge during the target object test;

[0016] generating the characteristic excitation signal based on the excitation signal parameters, and constructing a test environment for the excitation test based on the test environment parameters.

[0017] In one of the embodiments, the excitation test on the target object based on the characteristic excitation signal to obtain the second charge-discharge monitoring data includes:

[0018] if there are a plurality of predicted faults to be verified, sorting the predicted faults based on a preset test priority logic to obtain a sorted fault sequence;

[0019] adjusting the fault sequence based on a preset signal superposition constraint condition, and performing the excitation test step by step based on the adjusted fault sequence, and the signal superposition constraint condition is used to constrain the superposition avoidance of the excitation signals that exist in the excitation test and have interference influence.

[0020] In one of the embodiments, before the analyzing the first charge-discharge monitoring data based on the pre-constructed analysis model to obtain the quality detection result associated with the first charge-discharge monitoring data, the method further comprises:

[0021] obtaining a pair of historical charge-discharge monitoring data and a historical quality detection result, fitting the pre-processed historical charge-discharge monitoring data into a unified preset scale charge-discharge curve;

[0022] constructing a sample set based on the charge-discharge curve and the historical quality detection result, the sample set comprising a training set, a validation set and a test set distributed in a preset proportion;

[0023] constructing an initial model, training the initial model to convergence based on the sample set to obtain the analysis model, the analysis model being configured to input the charge-discharge curve, analyze the charge-discharge curve, and output the fault type existing in the target object and the corresponding probability index.

[0024] In one of the embodiments, the constructing a sample set based on the charge-discharge curve and the historical quality detection result, the sample set comprising a training set, a validation set and a test set distributed in a preset proportion comprises:

[0025] constructing the sample set based on a preset generalization constraint condition, the generalization constraint condition being a constraint condition for constraining the distribution proportion of various samples in different classification dimensions in the sample set, the classification dimensions comprising true-false sample classification, fault type classification, test environment classification, battery health state classification, and load state classification.

[0026] In one of the embodiments, after the verifying and analyzing the second charge-discharge monitoring data based on the analysis model, updating the quality detection result according to the result of the verifying and analyzing to obtain the quality detection result updated by verification, the method further comprises:

[0027] determining feedback data of the quality detection result according to the result of the verifying and analyzing, the feedback data comprising positive samples, negative samples and weak samples, the positive samples being fault samples correctly predicted by verification, the negative samples being fault samples misjudged by verification, and the weak samples being newly generated fault samples in the verification test;

[0028] re-weighting a loss function of the analysis model based on a predefined dynamic weight according to the feedback data to obtain the analysis model updated by feedback, the dynamic weight having different penalty weights for the negative samples and the weak samples.

[0029] In a second aspect, the application further provides a lithium battery detection device based on charge-discharge curve analysis.

[0030] The device comprises:

[0031] a standard test module configured to perform charge-discharge test on a target object based on a preset standard test signal, and obtain first charge-discharge monitoring data;

[0032] a fault prediction module configured to analyze the first charge-discharge monitoring data based on a pre-constructed analysis model, and obtain a quality detection result associated with the first charge-discharge monitoring data;

[0033] a feature excitation module configured to generate a feature excitation signal matched with a predicted fault existing in the quality detection result, in response to a probability index of the predicted fault exceeding a preset fault probability threshold;

[0034] a verification test module configured to perform excitation test on the target object based on the feature excitation signal, and obtain second charge-discharge monitoring data;

[0035] a verification update module configured to perform verification analysis on the second charge-discharge monitoring data based on the analysis model, and update the quality detection result according to a result of the verification analysis, to obtain a verification-updated quality detection result.

[0036] In one of the embodiments, the feature excitation module comprises:

[0037] a probability screening module configured to determine a plurality of fault types and corresponding probability indexes according to the quality detection result, and screen out the predicted fault based on the preset fault probability threshold;

[0038] a test parameter module configured to match the predicted fault according to a pre-constructed test rule library, determine excitation signal parameters and test environment parameters associated with the predicted fault, wherein the excitation signal parameters comprise a time-domain dynamic signal sequence in a plurality of battery cycle periods, and the test environment parameters comprise an environmental temperature and a state of charge during target object test;

[0039] a verification environment module configured to generate the feature excitation signal based on the excitation signal parameters, and construct a test environment of the excitation test based on the test environment parameters.

[0040] In one of the embodiments, the verification test module comprises:

[0041] a fault sorting module configured to sort the predicted faults based on a preset test priority logic, to obtain a sorted fault sequence, if there are a plurality of predicted faults to be verified;

[0042] The superposition avoidance module is configured to adjust the fault sequence based on a preset signal superposition constraint condition, and to perform the excitation test step by step based on the adjusted fault sequence, wherein the signal superposition constraint condition is configured to constrain the superposition avoidance of the excitation signals that have interference effects in the excitation test.

[0043] In one of the embodiments, the fault prediction module further comprises, before the step of obtaining the historical charging and discharging monitoring data and the historical quality detection result:

[0044] The curve fitting module is configured to obtain the paired historical charging and discharging monitoring data and the historical quality detection result, and fit the preprocessed historical charging and discharging monitoring data into a charging and discharging curve of a unified preset scale.

[0045] The sample set module is configured to construct a sample set based on the charging and discharging curve and the historical quality detection result, wherein the sample set comprises a training set, a validation set and a test set distributed in a preset proportion.

[0046] The model training module is configured to construct an initial model, train the initial model to convergence based on the sample set, and obtain the analysis model, wherein the analysis model is configured to take the charging and discharging curve as input, perform feature analysis on the charging and discharging curve, and output the fault type existing in the target object and the corresponding probability index.

[0047] In one of the embodiments, the sample set module comprises:

[0048] The sample construction module is configured to construct the sample set based on a preset generalization constraint condition, wherein the generalization constraint condition is a constraint condition for constraining the distribution proportion of various samples in different classification dimensions in the sample set, and the classification dimensions comprise true and false sample classification, fault type classification, test environment classification, battery health state classification and load state classification.

[0049] In one of the embodiments, the validation updating module further comprises, after the step of obtaining the feedback data:

[0050] The feedback data module is configured to determine feedback data of the quality detection result according to the result of the validation analysis, wherein the feedback data comprises positive samples, negative samples and weak samples, the positive samples are fault samples correctly predicted after validation, the negative samples are fault samples misjudged after validation, and the weak samples are newly generated fault samples in the validation test.

[0051] The feedback updating module is configured to reweight a loss function of the analysis model based on a predefined dynamic weight according to the feedback data, and obtain the analysis model after feedback updating, wherein different penalty weights are set for the negative samples and the weak samples in the dynamic weight.

[0052] In a third aspect, the present application also provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the lithium battery detection method based on charge-discharge curve analysis according to any one of the embodiments of the first aspect when executing the computer program.

[0053] In a fourth aspect, the present application also provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the lithium battery detection method based on charge-discharge curve analysis according to any one of the embodiments of the first aspect.

[0054] In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program, and the computer program is executed by a processor to implement the steps of the lithium battery detection method based on charge-discharge curve analysis according to any one of the embodiments of the first aspect.

[0055] The lithium battery detection method, device, computer device, storage medium and computer program product based on charge-discharge curve analysis can achieve the following beneficial effects in solving the technical problems in the corresponding background art through the technical features in the solutions:

[0056] The present application provides a lithium battery detection method based on charge-discharge curve analysis, characterized in that the method comprises: performing charge-discharge testing on a target object based on a pre-set standard test signal to obtain first charge-discharge monitoring data; analyzing the first charge-discharge monitoring data based on a pre-constructed analysis model to obtain a quality detection result associated with the first charge-discharge monitoring data; in response to a predicted fault existing in the quality detection result, generating a characteristic excitation signal matched with the predicted fault, wherein the predicted fault is indicated by a probability index exceeding a pre-set fault probability threshold; performing excitation testing on the target object based on the characteristic excitation signal to obtain second charge-discharge monitoring data; and performing verification analysis on the second charge-discharge monitoring data based on the analysis model, updating the quality detection result according to the result of the verification analysis to obtain a verification-updated quality detection result. In implementation, through detection and analysis of the first charge-discharge data, the high-risk fault type can be quickly locked, thereby realizing coarse positioning of the fault type. Subsequently, a directional characteristic excitation signal is generated for the high-risk fault type and verification testing is performed, which helps to realize specific targeted excitation for verification testing, saves other testing procedures with lower relevance, helps to improve detection accuracy and efficiency of verification testing. In addition, by setting the excitation signal scheme for the fault type, battery loss caused by testing can be reduced, thereby realizing low-loss detection excitation. Finally, the efficiency and accuracy of battery detection are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 1 is a schematic diagram of a first flow chart of a lithium battery detection method based on charge and discharge curve analysis in one embodiment;

[0059] Figure 2 1 is a schematic diagram of a second flow chart of a lithium battery detection method based on charge and discharge curve analysis in another embodiment;

[0060] Figure 3 1 is a schematic diagram of a third flow chart of a lithium battery detection method based on charge and discharge curve analysis in another embodiment;

[0061] Figure 4 4 is a schematic diagram of a fourth flow chart of a lithium battery detection method based on charge and discharge curve analysis in another embodiment;

[0062] Figure 5 is a fifth flow chart of a lithium battery detection method based on charge and discharge curve analysis in another embodiment;

[0063] Figure 6 6 is a schematic diagram of a sixth flow chart of a lithium battery detection method based on charge and discharge curve analysis in another embodiment;

[0064] Figure 7 is a structural block diagram of a lithium battery detection device based on charge and discharge curve analysis in one embodiment;

[0065] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0067] In related technologies, in order to ensure the product quality of lithium batteries, lithium batteries are usually subjected to quality tests such as electrochemical tests and physical characterization tests. For example, quality analysis results are obtained by analyzing the charge and discharge curves.

[0068] However, the current battery detection method has the following technical problems:

[0069] The traditional detection method relies on charge-discharge curve analysis, and the charge-discharge curve is affected by environmental factors and other factors, and there are many oscillations and other noises, which leads to the accuracy of the detection result to be optimized.

[0070] In one embodiment, as shown in Figure 1 A lithium battery detection method based on charge-discharge curve analysis is provided. This embodiment is used in terminals, and it can be understood that the method can also be applied to servers, and can also be applied to systems including terminals and servers, and can be realized through the interaction of terminals and servers. In this embodiment, the method includes the following steps:

[0071] Step 102: Perform charge-discharge test on the target object based on the preset standard test signal, and obtain first charge-discharge monitoring data.

[0072] Step 104: Analyze the first charge-discharge monitoring data based on a pre-constructed analysis model, and obtain a quality detection result associated with the first charge-discharge monitoring data.

[0073] Step 106: In response to a predicted failure in which a probability indicator in the quality detection result exceeds a preset failure probability threshold, generate a characteristic excitation signal matched with the predicted failure.

[0074] Step 108: Perform excitation test on the target object based on the characteristic excitation signal, and obtain second charge-discharge monitoring data.

[0075] Step 1010: Based on the analysis model, perform verification analysis on the second charge-discharge monitoring data, and update the quality detection result according to the result of the verification analysis to obtain a verification-updated quality detection result.

[0076] In the above lithium battery detection method based on charge-discharge curve analysis, the technical features in the embodiments are reasonably deduced to achieve the beneficial effects of solving the technical problems in the background art:

[0077] The application provides a lithium battery detection method based on charge-discharge curve analysis, characterized in that the method comprises the following steps: performing charge-discharge test on a target object based on a preset standard test signal, and obtaining first charge-discharge monitoring data; analyzing the first charge-discharge monitoring data based on a pre-constructed analysis model, and obtaining a quality detection result associated with the first charge-discharge monitoring data; in response to a predicted fault existing in the quality detection result and having a probability index exceeding a preset fault probability threshold, generating a characteristic excitation signal matched with the predicted fault; performing excitation test on the target object based on the characteristic excitation signal, and obtaining second charge-discharge monitoring data; performing verification analysis on the second charge-discharge monitoring data based on the analysis model, updating the quality detection result according to the result of the verification analysis, and obtaining the quality detection result after verification update. In implementation, through detection and analysis of the first charge-discharge data, a high-risk fault type can be quickly locked, so as to realize coarse positioning of the fault type. Then, a directional characteristic excitation signal is generated for the high-risk fault type, and verification test is performed, which helps to realize specific targeted excitation for verification test, save other test procedures with low correlation, improve detection accuracy, and improve the efficiency of verification detection. In addition, by setting an excitation signal scheme for the fault type, battery loss caused by test can be reduced, so as to realize low-loss detection excitation. Finally, the efficiency and accuracy of battery detection are improved.

[0078] In one embodiment, as shown in FIG. 1, Figure 2 The step 106 comprises:

[0079] Step 202: According to the quality detection result, determine a plurality of fault types and corresponding probability indexes, and screen out the predicted fault based on the preset fault probability threshold.

[0080] Step 204: According to a pre-constructed test rule library, match the predicted fault, and determine excitation signal parameters and test environment parameters associated with the predicted fault.

[0081] The excitation signal parameters comprise time-domain dynamic signal sequences in a plurality of battery cycle periods, and the test environment parameters comprise an environmental temperature and a state of charge during target object test.

[0082] For example, the voltage value of the excitation signal may be dynamically adjusted in the same cycle period and between different cycle periods, so the excitation signal is set as a time-domain dynamic signal sequence following the time change of the test procedure.

[0083] Step 206: Generate the characteristic excitation signal based on the excitation signal parameters, and construct a test environment for the excitation test based on the test environment parameters.

[0084] In this embodiment, by configuring dynamic excitation signals and test environment, it is helpful to build a targeted test process for fault types and improve the accuracy of verification.

[0085] In one of the embodiments, as shown in Figure 3 The step 108 includes:

[0086] Step 302: If there are multiple predicted faults to be verified, the predicted faults are sorted based on a preset test priority logic to obtain a sorted fault sequence.

[0087] Step 304: The fault sequence is adjusted based on a preset signal superposition constraint condition, and the excitation test is performed step by step based on the adjusted fault sequence.

[0088] The signal superposition constraint condition is used to constrain the excitation signals with interference influence to be superposed and avoided in the excitation test.

[0089] In this embodiment, by priority sorting, it is helpful to test test items with higher priority first, improve the efficiency of fault verification, and by avoiding design, it is helpful to avoid simultaneous testing of test items with interference, and improve the stability of the test process.

[0090] In one of the embodiments, as shown in Figure 4 Before the step 104, it further includes:

[0091] Step 402: Obtain paired historical charge and discharge monitoring data and historical quality detection results, and fit the preprocessed historical charge and discharge monitoring data into a unified preset scale charge and discharge curve.

[0092] For example, the terminal can clean the collected historical charge and discharge monitoring data, remove outliers (such as obvious measurement errors, data deviation caused by equipment failure, etc.). Standardize or normalize the data to make the voltage, current, time and other data of different magnitudes in the same scale, and avoid interference with model training. In the process of feature engineering processing of historical charge and discharge monitoring data, in addition to traditional slope, platform voltage, capacity and other features, Fourier transform, wavelet transform and other methods can be used to extract features from the charge and discharge curve to obtain frequency domain and time domain features, and enrich the feature dimension of the input model.

[0093] Step 404: Construct a sample set based on the charge and discharge curve and the historical quality detection result, the sample set including a training set, a validation set and a test set distributed in a preset proportion.

[0094] Exemplarily, the terminal can divide the processed data set into a training set, a validation set and a test set in a certain proportion (such as 7:2:1). Among them, the training set is used for parameter learning of the model, the validation set is used for adjusting the hyperparameters of the model, such as learning rate, network layer number, node number, etc., and the test set is used for evaluating the generalization ability of the model.

[0095] Step 406: constructing an initial model, training the initial model to convergence based on the sample set, obtaining the analysis model, and using the analysis model to input the charge-discharge curve to perform feature analysis on the charge-discharge curve, and output the fault type existing in the target object and the corresponding probability index.

[0096] Among them, the initial model can be a convolutional neural network model, a recurrent neural network model and a gradient boosting tree model, etc., which can be a single model or a combination of multiple models. Specifically, the convolutional neural network (CNN) is suitable for processing data with spatial structure. The charge-discharge curve can be regarded as a kind of one-dimensional "image". CNN can extract local features and overall features in the curve, such as mutation points and fluctuation rules of the curve, which is suitable for battery fault detection, performance evaluation and other tasks. The recurrent neural network (RNN), especially the long short-term memory network (LSTM), is suitable for processing time series data. Since the charge-discharge curve is time series data that changes over time, LSTM can capture the dependence relationship of the curve in the time dimension, such as the trend of capacity attenuation and the dynamic characteristics of voltage change, which is suitable for battery life prediction, state estimation and other tasks. Gradient boosting tree (GBDT) can process nonlinear data and has good interpretability. The feature parameters (such as slope, platform voltage, capacity, etc.) of the charge-discharge curve can be used as input, and classification or regression can be performed through the GBDT model to realize the evaluation of battery performance.

[0097] Exemplarily, the terminal can use the training set to train the selected initial model, and optimize the model parameters through a back propagation algorithm or the like, so that the error between the prediction result of the model and the actual label (such as the health status, life, fault type of the battery, etc.) is minimized. During the training process, the performance of the model can be monitored through the validation set to prevent overfitting.

[0098] In this embodiment, by constructing a sample to train an artificial intelligence model, it is helpful to use the analysis model trained by the specified training to perform fault prediction, and to improve the efficiency and accuracy of fault prediction.

[0099] In one of the embodiments, as shown in Figure 5 The step 404 includes:

[0100] Step 502: constructing the sample set based on a preset generalization constraint condition.

[0101] The generalization constraint condition is a constraint condition for constraining the distribution proportion of each type of sample in different classification dimensions in the sample set, and the classification dimensions include true or false sample classification, fault type classification, test environment classification, battery health state classification, and load state classification.

[0102] For example, the terminal can perform tests on lithium batteries of different models, different aging degrees (such as brand-new batteries, batteries with different cycle times), and different fault types (such as internal short circuit, bulging, liquid leakage, and capacity attenuation) according to the implementation process of the charge-discharge test, collect the charge-discharge curve data, and record the relevant information (such as model, cycle time, fault type, health state, etc.) of the battery as labels to obtain real historical samples. In order to ensure the generalization ability of the model, the number of samples can be kept at a sufficient data amount. Generally, the number of samples in each category should not be too small (such as the number of samples in each fault type should not be less than 100). At the same time, in order to ensure the diversity of the samples, lithium batteries of different models, different production batches, different use conditions, and different fault types can also be covered, so that the model can learn the general rule instead of being limited to specific sample features.

[0103] In this embodiment, by setting the generalization constraint condition, the distribution of the samples can be controlled, so as to avoid the expected deviation of the trained model, and the efficiency and accuracy of the fault analysis can be improved.

[0104] In one of the embodiments, as shown in FIG. 10, after the step 1010, the method further includes: Figure 6

[0105] Step 602: determining feedback data of the quality detection result according to the result of the verification analysis.

[0106] The feedback data includes positive samples, negative samples, and weak samples. The positive samples are fault samples that are correctly predicted after verification, the negative samples are fault samples that are misjudged after verification, and the weak samples are newly generated fault samples in the verification test.

[0107] Step 604: reweighting a loss function of the analysis model based on a predefined dynamic weight according to the feedback data, to obtain the analysis model after feedback update.

[0108] In the dynamic weight, different penalty weights are set for the negative samples and the weak samples.

[0109] For example, the negative samples corresponding to the misjudgment condition can be set with a higher penalty weight than the weak samples corresponding to the missed judgment condition.

[0110] ​In this embodiment, the analysis model is updated by feedback data, which helps to adjust the feature analysis weight of each module in the analysis model in real time, thereby enhancing the analysis performance of the model.

[0111] It should be understood that, although each step in the flowchart involved in each embodiment as described above is shown in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be alternately executed with at least part of other steps or steps or stages in other steps.

[0112] Based on the same inventive concept, the embodiments of the present application also provide a lithium battery detection device based on charge-discharge curve analysis for implementing the above-mentioned lithium battery detection method based on charge-discharge curve analysis. The problem-solving implementation scheme provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more lithium battery detection device embodiments based on charge-discharge curve analysis provided below can refer to the limitations of the lithium battery detection method based on charge-discharge curve analysis in the above text, and will not be repeated here.

[0113] In one embodiment, as shown in Figure 7 a lithium battery detection device based on charge-discharge curve analysis is provided, comprising: a standard test module, a fault prediction module, a feature excitation module, a feature excitation module and a verification update module, wherein:

[0114] The standard test module is configured to perform charge-discharge testing on the target object based on a preset standard test signal, and obtain first charge-discharge monitoring data.

[0115] The fault prediction module is configured to analyze the first charge-discharge monitoring data based on a pre-constructed analysis model, and obtain a quality detection result associated with the first charge-discharge monitoring data.

[0116] The feature excitation module is configured to generate a feature excitation signal matched with the predicted fault when the probability index in the quality detection result exceeds a preset fault probability threshold.

[0117] The verification test module is configured to perform excitation testing on the target object based on the feature excitation signal, and obtain second charge-discharge monitoring data.

[0118] The verification updating module is configured to perform verification analysis on the second charge-discharge monitoring data based on the analysis model, and update the quality detection result according to a result of the verification analysis to obtain a verification-updated quality detection result.

[0119] In one of the embodiments, the feature excitation module comprises:

[0120] The probability screening module is configured to determine a plurality of fault types and corresponding probability indicators according to the quality detection result, and screen out the predicted fault based on the preset fault probability threshold.

[0121] The test parameter module is configured to match the predicted fault according to a pre-constructed test rule library, determine excitation signal parameters and test environment parameters associated with the predicted fault, the excitation signal parameters comprising a time-domain dynamic signal sequence in a plurality of battery cycle periods, and the test environment parameters comprising an ambient temperature and a state of charge during the test of the target object.

[0122] The verification environment module is configured to generate the feature excitation signal based on the excitation signal parameters, and construct a test environment for the excitation test based on the test environment parameters.

[0123] In one of the embodiments, the verification test module comprises:

[0124] The fault sorting module is configured to sort the predicted faults based on a preset test priority logic if there are a plurality of predicted faults to be verified, to obtain a sorted fault sequence.

[0125] The superposition avoidance module is configured to adjust the fault sequence based on a preset signal superposition constraint condition, and perform the excitation test step by step based on the adjusted fault sequence, the signal superposition constraint condition being used to constrain the excitation signals that exist in the excitation test and have interference influence to perform superposition avoidance.

[0126] In one of the embodiments, the fault prediction module further comprises:

[0127] The curve fitting module is configured to obtain paired historical charge-discharge monitoring data and historical quality detection results, and fit the pre-processed historical charge-discharge monitoring data into a unified charge-discharge curve of a preset scale.

[0128] The sample set module is configured to construct a sample set based on the charge-discharge curve and the historical quality detection results, the sample set comprising a training set, a verification set and a test set distributed in a preset proportion.

[0129] The model training module is configured to construct an initial model, train the initial model to convergence based on the sample set, and obtain the analysis model, which is configured to take the charge-discharge curve as input, perform feature analysis on the charge-discharge curve, and output the fault type existing in the target object and the corresponding probability index.

[0130] In one of the embodiments, the sample set module comprises:

[0131] The sample construction module is configured to construct the sample set based on a preset generalization constraint condition, which is a constraint condition for constraining the distribution proportion of various types of samples in different classification dimensions in the sample set, the classification dimensions including true-false sample classification, fault type classification, test environment classification, battery health state classification, and load state classification.

[0132] In one of the embodiments, the verification and update module further comprises:

[0133] The feedback data module is configured to determine feedback data of the quality detection result according to the result of the verification analysis, the feedback data comprising positive samples, negative samples, and weak samples, the positive samples being fault samples correctly predicted after verification, the negative samples being fault samples misjudged after verification, and the weak samples being newly generated fault samples in the verification test.

[0134] The feedback update module is configured to reweight a loss function of the analysis model based on a predefined dynamic weight according to the feedback data, and obtain the analysis model after feedback update, the dynamic weight having different penalty weights for the negative samples and the weak samples.

[0135] The above-described various modules in the lithium battery detection device based on the charge-discharge curve analysis can be realized by software, hardware, or a combination thereof, in whole or in part. The above-described various modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform the operations corresponding to the above-described various modules.

[0136] In one of the embodiments, a computer device is provided, which can be a terminal, and an internal structure diagram of the computer device can be as shown in Figure 8As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control ability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with the external terminal in a wired or wireless manner. The wireless manner can be realized through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to realize a lithium battery detection method based on charge-discharge curve analysis. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0137] Those skilled in the art can understand that, Figure 8 The skilled in the art can understand that,

[0138] In one embodiment, a computer device is also provided, including a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the steps in the above method embodiments.

[0139] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by the processor to realize the steps in the above method embodiments.

[0140] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by the processor to realize the steps in the above method embodiments.

[0141] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of the country and region.

[0142] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of each method can be included. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetic variable memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0143] The technical features of the above embodiments can be combined in any way. In order to make the description simple, not all possible combinations of the technical features in the above embodiments are described, but as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0144] The above-described embodiments are merely illustrative of several embodiments of the present application, which are described in more detail and in a specific manner, but should not be construed as limiting the scope of the patent of the present application. It should be noted that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A lithium battery detection method based on charge-discharge curve analysis, characterized in that, The method comprises: performing charge-discharge testing on the target object based on a preset standard test signal to obtain first charge-discharge monitoring data; analyzing the first charge-discharge monitoring data based on a pre-constructed analysis model to obtain a quality detection result associated with the first charge-discharge monitoring data; in response to a predicted fault existing in the quality detection result and having a probability index exceeding a preset fault probability threshold, generating a characteristic excitation signal matched with the predicted fault; performing excitation testing on the target object based on the characteristic excitation signal to obtain second charge-discharge monitoring data; performing verification analysis on the second charge-discharge monitoring data based on the analysis model, and updating the quality detection result according to a result of the verification analysis to obtain a verification-updated quality detection result.

2. The method of claim 1, wherein, The response to the predicted fault existing in the quality detection result and having a probability index exceeding a preset fault probability threshold includes: determining a plurality of fault types and corresponding probability indexes according to the quality detection result, and screening out the predicted fault based on the preset fault probability threshold; matching the predicted fault according to a pre-constructed test rule library to determine excitation signal parameters and test environment parameters associated with the predicted fault, the excitation signal parameters including time-domain dynamic signal sequences in a plurality of battery cycle periods, and the test environment parameters including an environmental temperature and a state of charge during the target object testing; generating the characteristic excitation signal based on the excitation signal parameters, and constructing a test environment for the excitation testing based on the test environment parameters.

3. The method of claim 2, wherein, The excitation testing on the target object based on the characteristic excitation signal to obtain second charge-discharge monitoring data includes: if there are a plurality of predicted faults to be verified, sorting the predicted faults based on a preset test priority logic to obtain a sorted fault sequence; adjusting the fault sequence based on a preset signal superposition constraint condition, and performing the excitation testing step by step based on the adjusted fault sequence, the signal superposition constraint condition being used to constrain the excitation signals having interference influence to be superposed and avoided.

4. The method of claim 1, wherein, Before the analysis of the first charge-discharge monitoring data based on the pre-constructed analysis model to obtain the quality detection result associated with the first charge-discharge monitoring data, the method further comprises: obtaining paired historical charge-discharge monitoring data and historical quality detection results, and fitting the pre-processed historical charge-discharge monitoring data into a unified preset scale charge-discharge curve; constructing a sample set based on the charge-discharge curve and the historical quality detection results, the sample set including a training set, a validation set and a test set distributed at a preset proportion; constructing an initial model, training the initial model to convergence based on the sample set to obtain the analysis model, the analysis model being used to input the charge-discharge curve to perform feature analysis on the charge-discharge curve, and output a fault type existing in the target object and a corresponding probability index.

5. The method of claim 4, wherein, The sample set is constructed based on the charge-discharge curve and the historical quality detection result, and the sample set includes a training set, a validation set and a test set distributed in a preset proportion. The sample set is constructed based on a preset generalization constraint condition, the generalization constraint condition is a constraint condition for constraining the distribution proportion of each type of sample in different classification dimensions in the sample set, and the classification dimensions include true-false sample classification, fault type classification, test environment classification, battery health state classification and load state classification.

6. The method of claim 4, wherein, After the quality detection result is updated according to the result of the verification analysis, the second charge-discharge monitoring data is verified and analyzed based on the analysis model, and the quality detection result is updated after verification. According to the result of the verification analysis, feedback data of the quality detection result is determined, the feedback data includes positive samples, negative samples and weak samples, the positive samples are fault samples correctly predicted after verification, the negative samples are fault samples misjudged after verification, and the weak samples are new fault samples generated in the verification test. According to the feedback data, the loss function of the analysis model is reweighted based on a predefined dynamic weight, and the analysis model is updated after feedback, and the dynamic weight has different penalty weights for the negative samples and the weak samples.

7. A lithium battery detection device based on charge-discharge curve analysis, characterized in that, The device comprises: a standard test module configured to perform charge-discharge testing on a target object based on a preset standard test signal and obtain first charge-discharge monitoring data; a fault prediction module configured to analyze the first charge-discharge monitoring data based on a pre-constructed analysis model and obtain a quality detection result associated with the first charge-discharge monitoring data; a feature excitation module configured to generate a feature excitation signal matched with a predicted fault when the probability index of the predicted fault in the quality detection result exceeds a preset fault probability threshold; a verification test module configured to perform excitation testing on the target object based on the feature excitation signal and obtain second charge-discharge monitoring data; a verification update module configured to perform verification analysis on the second charge-discharge monitoring data based on the analysis model and update the quality detection result according to the result of the verification analysis to obtain the quality detection result updated after verification.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 6.

Images