Battery anomaly detection method based on Chebyshev polynomial

Abstract

The invention relates to a Chebyshev polynomial-based battery anomaly detection method, and belongs to the technical field of battery management. The method comprises the following steps: S1, collecting operation data of the electric vehicle, including voltage data and other parameters of each single battery in a battery pack, and carrying out data preprocessing; s2, carrying out Kalman filtering processing on the preprocessed voltage data of the single battery; dividing the filtered data into a training set and a verification set; s3, performing Chebyshev polynomial feature extraction based on the filtered single battery voltage data, and extracting a polynomial coefficient as a feature; s4, reconstructing an approximate value of the original data by using a Chebyshev polynomial coefficient, and calculating a reconstruction error; and S5, setting an abnormal detection threshold value, and judging whether each single battery has a fault or not according to whether the reconstruction error exceeds the set threshold value or not. According to the method, the abnormality detection of the power battery of the electric vehicle can be effectively realized, and the abnormal single body can be accurately positioned.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of battery management, and relates to a battery abnormality detection method based on Chebyshev polynomials. BACKGROUND

[0002] With the popularity of electric vehicles (EVs), the performance, safety and life of power battery packs, as core energy storage units of electric vehicles, directly determine the endurance, reliability and use cost of the whole vehicle. Therefore, establishing a precise, fast and reliable battery fault diagnosis system is of great significance for ensuring the safe operation of electric vehicles, preventing catastrophic accidents such as thermal runaway, and prolonging the service life of power batteries.

[0003] At present, the mainstream battery fault detection technologies in the industry can be mainly divided into two categories: threshold judgment method based on direct measurement parameters and analysis method based on battery model.

[0004] The first method directly monitors the key physical quantities such as voltage, current and temperature of the battery monomer, and compares them with the preset fixed threshold value. When the measured value exceeds the safe range, the system determines that there is a fault. Although this method is simple and intuitive and easy to implement, its limitations are also very significant. It can only identify faults that have developed to a serious degree, and it is difficult to effectively warn potential faults in the early stage with weak characteristics. In addition, the fixed threshold value cannot adapt to the dynamic changes of the normal working window of the battery at different life cycle stages, different working temperatures and different charge-discharge rates, and is prone to false positives or false negatives.

[0005] The second method is to establish a mathematical model of the battery, and to diagnose faults by comparing the residual error between the model prediction value and the actual measured value of the sensor. Among them, the analysis methods based on equivalent circuit model (ECM) and electrochemical model are the research hotspots. However, these traditional methods based on models face many challenges in practical application. For example, when using numerical methods such as finite difference method (FDM) to solve complex nonlinear partial differential equations to construct high-precision electrochemical models, the calculation process is very tedious. Especially when dealing with complex boundary conditions and geometric configurations, the amount of calculation is huge, and it is difficult to meet the stringent real-time requirements of the vehicle management system (VMS). More importantly, numerical methods such as finite difference method are sensitive to data noise. In the complex electromagnetic and vibration environment of actual driving of electric vehicles, noise will inevitably be mixed into the signals collected by the sensors, which will seriously interfere with the calculation process of the model, significantly reducing the accuracy and reliability of fault detection.

[0006] In addition, most existing battery management systems (BMS) often use a unified, ideal battery model to represent all individual cells in the entire battery pack when performing fault diagnosis. This "one-size-fits-all" approach seriously ignores the inherent inconsistencies between individual cells in the battery pack. Due to minor differences in manufacturing processes, uneven temperature distribution over long-term use, and different aging rates, the electrochemical characteristics (such as internal resistance, capacity, open-circuit voltage, etc.) of each battery cell will exhibit unique, dynamic differences over time. Applying a standard model to all cells will inevitably mask subtle abnormalities caused by early faults that are similar to normal individual differences, making it difficult to adapt to the diverse diagnostic needs of different battery types and different aging levels, ultimately resulting in insufficient accuracy and sensitivity of diagnosis.

[0007] Finally, traditional fault detection methods generally lack robustness when faced with complex and variable actual working conditions of electric vehicles (such as sudden acceleration, sudden deceleration, bumpy road conditions) and strong noise interference. These external disturbances can cause dramatic fluctuations in battery parameters, and traditional methods have difficulty effectively distinguishing them from real internal battery faults (such as internal short circuits, loose connections, electrolyte leaks, etc.), thus frequently causing false positives and false negatives, reducing system reliability and user trust.

[0008] In summary, existing battery fault detection methods have obvious bottlenecks in terms of diagnostic accuracy, real-time performance, adaptability to individual differences, and robustness under complex working conditions. To overcome the above technical defects, there is an urgent need in the field to develop a new battery fault detection method that can effectively handle the highly nonlinear characteristics of battery systems, intelligently adapt to individual differences between battery cells, and still maintain high stability and reliability in complex noise environments, to meet the higher safety and reliability requirements of the next generation of electric vehicles. SUMMARY

[0009] Therefore, the purpose of the present application is to provide a battery anomaly detection method based on Chebyshev polynomials, which extracts Chebyshev polynomial coefficients as features by fitting battery voltage data, and uses these features to reconstruct data and calculate reconstruction error, thereby achieving anomaly detection. It not only reduces noise impact and improves detection sensitivity, but also balances detection accuracy and computational complexity by adjusting the polynomial order, adapts to different battery characteristics and working condition changes, and significantly improves the accuracy and robustness of battery fault detection.

[0010] To achieve the above purpose, the present application provides the following technical solutions:

[0011] A battery anomaly detection method based on Chebyshev polynomials, specifically comprising the following steps:

[0012] S1: Collect electric vehicle operation data, including battery pack voltage data and other parameters of each battery monomer, and perform data preprocessing;

[0013] S2: Kalman filter processing is performed on the preprocessed battery monomer voltage data to reduce noise interference; and the filtered data is divided into training set and validation set to provide data basis for subsequent feature extraction and anomaly detection;

[0014] S3: Based on the filtered battery monomer voltage data, Chebyshev polynomial feature extraction is performed, and the polynomial coefficients are extracted as features;

[0015] S4: The approximate value of the original data is reconstructed using the Chebyshev polynomial coefficients, and the reconstruction error is calculated;

[0016] S5: Set the anomaly detection threshold, and judge whether each monomer battery has failed according to whether the reconstruction error exceeds the set threshold.

[0017] Further, step S1 specifically comprises the following steps:

[0018] S11: Collect normal operation data of a certain electric vehicle, including total voltage, each battery monomer voltage and time parameters;

[0019] S12: Data filtering and data cleaning are performed on the extracted battery monomer voltage to fill in the missing values of the data;

[0020] S13: Based on the collected operation data of the power battery pack and each battery monomer, a power battery operation database is established.

[0021] Further, in step S2, the preprocessed battery monomer voltage data is subjected to Kalman filter processing, specifically comprising the following steps:

[0022] S21: Determine the initial state estimate and estimate error covariance according to the preprocessed monomer voltage data;

[0023] S22: Use the mean and variance of the first few data points as the initial state estimate, use the state estimate at the last time to predict the state at the current time, and update the estimate error covariance;

[0024] S23: Calculate the Kalman gain according to the current measurement value, correct the predicted state to obtain a more accurate estimate, and update the estimate error covariance;

[0025] S24: filtering processing is performed on each battery monomer voltage time sequence, and the filtering process outputs a smooth and stable voltage data sequence at each time point.

[0026] Further, in step S2, the filtered data is divided into a training set and a validation set, specifically including: dividing the filtered time-ordered data into the first 70% normal data and the last 30% abnormal data; based on the divided data, defining the first 70% normal data as the training set, and defining the last 30% abnormal data as the validation set, to provide a data basis for subsequent feature extraction and anomaly detection.

[0027] The validation set in the data segmentation method can be based on normal data, and the validation set can also be based on abnormal data, and the segmentation rule is optional.

[0028] Further, step S3 specifically includes the following steps:

[0029] S31: according to experience or through cross-validation, a suitable Chebyshev polynomial order m is selected to balance the accuracy and computational complexity of feature extraction; selecting a higher Chebyshev polynomial order can capture more detailed features, but may also lead to overfitting.

[0030] In order to reduce the influence of abnormal battery monomer on the average voltage, the normalized median voltage value is compared with the normalized monomer voltage value to realize battery fault detection.

[0031] S32: construct a Chebyshev polynomial basis function;

[0032] Define m+1 Chebyshev polynomial basis functions T j (x), where j=0,1,…,m;

[0033] Satisfy the recursive relationship:

[0034] T0(x)=1

[0035] T1(x)=x

[0036] T j+1 (x)=2xT j (x)-T j-1 (x), j≥1

[0037] Where x represents the independent variable, i.e. the corresponding number of mapping battery monomer voltage data to the zero point of Chebyshev polynomial, which is an equidistant point on the interval [-1,1]; m represents the order of Chebyshev polynomial; j represents the order index of Chebyshev polynomial, which is a non-negative integer (j≥0), used to identify different Chebyshev polynomials; T j(x) represents the j-th order Chebyshev polynomial, which is a function of x, and is used to represent Chebyshev polynomials of different orders, T0(x) is the 0-th order Chebyshev polynomial, T1(x) is the 1-st order Chebyshev polynomial, T j+1 (x) is the j+1-th order Chebyshev polynomial, T j-1 (x) is the j-1-th order Chebyshev polynomial;

[0038] S33: According to the Chebyshev polynomial basis function T j (x), for the voltage data of each single battery in the training set and the validation set, the Chebyshev polynomial is fitted using the least square method, and the polynomial coefficient c j is obtained, so that the error sum of squares of the fitted curve and the original data is minimized; the extracted coefficient c j is taken as the feature of the battery single voltage data, reflecting the main trend and change mode of the data.

[0039] Further, step S4 specifically comprises the following steps:

[0040] S41: Reconstruct the approximation value j of the original data using the extracted Chebyshev polynomial coefficient c j and the basis function T

[0041]

[0042] wherein, represents the approximation value of the reconstructed original data of the i-th data point; x i represents the i-th independent variable; T j (x i ) represents the value of the j-th order Chebyshev polynomial at x i ;

[0043] S42: Calculate the reconstruction error i between the original data y and the reconstructed data

[0044] Further, step S5 specifically comprises the following steps:

[0045] S51: Analyze the reconstruction error distribution of the validation set, select a suitable threshold setting method such as fixed threshold method, dynamic threshold method or threshold method based on statistical distribution, etc., and determine the anomaly detection threshold threshold;

[0046] S52: Compare the reconstruction error e i of the validation set with the threshold threshold, if e i > threshold, it is determined that the battery single data at this moment is an abnormal point;

[0047] S53: If the battery fails, locate the failed cell according to the failed cell voltage.

[0048] The beneficial effects of the present application are:

[0049] 1) The present application uses Chebyshev polynomial feature extraction, which utilizes its orthogonality to effectively separate different frequency components of the data, extracting more representative and independent features. This not only enhances the discrimination ability of the features, but also significantly improves the accuracy of anomaly detection, enabling the method to more sensitively capture subtle changes in battery voltage data.

[0050] 2) The excellent convergence speed of the method of the present application enables it to achieve high-precision data approximation at a lower order, thereby reducing the number of features required and reducing the consumption of computing resources, which is particularly important in large-scale battery monitoring scenarios, as it can shorten the detection time and improve the system response speed.

[0051] 3) The present application has good robustness to noise and can stably extract key features of battery voltage in noisy data environments, which ensures that the method can still provide reliable results when facing complex working conditions and noise interference in practical applications, reducing false positives and false negatives, and enhancing the overall stability of the system.

[0052] 4) The battery anomaly detection method proposed in the present application has high flexibility and wide applicability, suitable for various battery types and complex working conditions, which makes the method have good applicability in many fields such as electric vehicles and energy storage systems. At the same time, its high sensitivity and robustness enable it to detect battery anomalies in a timely and accurate manner, providing strong support for the safe management and maintenance of batteries.

[0053] Other advantages, objects, and features of the present application will be set forth in part in the following specification, and in part will become apparent to those skilled in the art from a consideration of the following specification, or can be learned from practice of the present application. The objects and other advantages of the present application can be realized and attained by means of the instrumentalities and combinations pointed out in the following specification. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to make the purposes, technical solutions and advantages of the present application clearer, the preferred detailed description of the present application will be combined with the drawings as follows, wherein:

[0055] Figure 1 The overall flowchart of the battery anomaly detection method based on Chebyshev polynomials provided for Example 1 is shown in Figure 1.

[0056] Figure 2 The algorithm framework diagram for Example 1 is shown in Figure 2.

[0057] Figure 3 The individual cell voltage curves for normal and faulty cells in Example 2;

[0058] Figure 4 The Kalman filter curves of the voltage of the normal and faulty cells in Example 2 are shown.

[0059] Figure 5 The feature extraction results based on Chebyshev polynomials in Example 2;

[0060] Figure 6 This is a comparison chart of the original training data and the reconstructed data in Example 2;

[0061] Figure 7 This is a comparison chart of the original and reconstructed data of the validation set in Example 2;

[0062] Figure 8 This is the training set reconstruction error map in Example 2;

[0063] Figure 9 This is the reconstruction error map of the validation set in Example 2;

[0064] Figure 10 This is the result of the validation set anomaly detection in Example 2. Detailed Implementation

[0065] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0066] Example 1:

[0067] This embodiment provides a battery anomaly detection method based on Chebyshev polynomials, which can determine the health status of each cell in the power battery pack based on on-site data of electric vehicles and perform real-time fault detection on each battery cell.

[0068] Please see Figures 1-2 The method specifically consists of the following steps:

[0069] Step S1: Collect electric vehicle operating data, including the voltage data of each cell in the battery pack and other characteristics, and perform data preprocessing, specifically including the following steps:

[0070] Step S11: Collect normal operation data of a certain electric vehicle, including total voltage, individual cell voltage, time and other parameters;

[0071] Step S12: Data filtering and data cleaning are performed on the extracted individual cell voltage, and missing values of the data are filled.

[0072] Among them, the data cleaning removes the rows with total voltage of 0, and the missing value filling can use the adjacent cell average interpolation method.

[0073] Step S13: According to the collected operation data of the power battery pack and each battery cell, a power battery operation database is established.

[0074] Step S2: Kalman filtering is performed on the battery cell voltage data to reduce noise interference, including the following steps:

[0075] Step S21: According to the processed individual cell voltage data, the initial state estimation and the estimation error covariance are determined, and the initial state estimation value is usually calculated by taking the mean of the first few data points, and the error covariance P0 is usually calculated by taking the variance of the first few data points.

[0076] Step S22: The mean and variance of the first few data points are used as the initial state estimation, and the state estimation value at the last time k-1 is used to predict the prior state estimation value at the current time k, and the prior estimation error covariance P is updated. k|k-1 ;

[0077] The prior state estimation value at the current time k is calculated as follows:

[0078]

[0079] The updated prior estimation error covariance P k|k-1 is calculated as follows:

[0080] P k|k-1 = P k-1|k-1 + Q

[0081] Where P k-1|k-1 represents the estimation error covariance at the last time k-1; Q is the process noise covariance.

[0082] Step S23: According to the current measurement value z k , the Kalman gain K k is calculated, the predicted state is corrected, and the posterior state estimation value is obtained, and the posterior estimation error covariance P k|k is updated.

[0083] Kalman gain K k The calculation formula is as follows:

[0084]

[0085] Where R is the measurement noise covariance.

[0086] Posterior state estimate The calculation formula is as follows:

[0087]

[0088] Posterior estimation error covariance P k|k The calculation formula is as follows:

[0089] P k|k = (1-K) k )P k|k-1

[0090] Step S24: Perform steps S21 to S23 sequentially for the voltage time series of each battery cell, and output a smooth and stable voltage data sequence at each time point.

[0091] Step S3: Based on the filtered unit voltage data, perform Chebyshev polynomial feature extraction, extracting the polynomial coefficients as features. This specifically includes the following steps:

[0092] Step S31: Divide the filtered data sorted by time into the first 70% of normal data and the last 30% of abnormal data;

[0093] Step S32: Based on the segmented data, define the first 70% of normal data as the training set and the last 30% of abnormal data as the validation set, providing a data foundation for subsequent feature extraction and anomaly detection.

[0094] In the data partitioning method, the validation set can be based on normal data or abnormal data, and the partitioning rules are optional.

[0095] Step S4: Based on the filtered unit voltage data, Chebyshev polynomial feature extraction is performed, and the polynomial coefficients are extracted as features. This specifically includes the following steps:

[0096] Step S41: Based on experience or through cross-validation, select an appropriate Chebyshev polynomial order m to balance the accuracy and computational complexity of feature extraction.

[0097] To reduce the impact of abnormal individual cells on the average voltage, the normalized median voltage value is compared with the normalized individual cell voltage values ​​to achieve battery fault detection.

[0098] Step S42: Constructing Chebyshev orthogonal polynomial basis functions.

[0099] Define m+1 Chebyshev polynomial basis functions T j (x), where j=0,1,…,m;

[0100] Satisfy the recursive relationship:

[0101] T0(x)=1

[0102] T1(x)=x

[0103] T j+1 (x)=2xT j (x)-T j-1 (x) (j≥1)

[0104] Where x represents the independent variable, i.e. the corresponding number of mapping the battery voltage data to the zero point of the Chebyshev polynomial, is an equidistant point on the interval [-1,1]; m represents the order of the Chebyshev polynomial; j represents the order index of the Chebyshev polynomial, which is a non-negative integer (j≥0), used to identify different Chebyshev polynomials; T j (x) represents the j-th order Chebyshev polynomial, which is a function of x, used to represent Chebyshev polynomials of different orders, T0(x) is the 0-th order Chebyshev polynomial, T1(x) is the 1-st order Chebyshev polynomial, T j+1 (x) is the j+1-th order Chebyshev polynomial, T j-1 (x) is the j-1-th order Chebyshev polynomial.

[0105] Step S43: According to the Chebyshev orthogonal polynomial basis function T j (x), for the voltage data of each single battery in the training set and the validation set, use the least squares method to fit the Chebyshev polynomial, get the polynomial coefficients c j , so that the error sum of squares of the fitted curve and the original data is minimized. The extracted coefficients c j are used as features of the battery voltage data, reflecting the main trend and change pattern of the data.

[0106] Step S5: Reconstruct the approximate value of the original data using the Chebyshev polynomial coefficients and calculate the reconstruction error, specifically including the following steps:

[0107] Step S51: Reconstruct the approximate value of the original data using the extracted Chebyshev polynomial coefficients c j and the basis function T j (x).

[0108]

[0109] wherein, represents the approximation of the reconstructed raw data of the i-th data point; x i represents the i-th independent variable x; T j (x i ) represents the value of the j-th order Chebyshev polynomial at x i .

[0110] Step S52: compare the raw data y i and the reconstructed data Calculate the reconstruction error

[0111] Step S6: set the anomaly detection threshold, determine whether each single battery has failed according to whether the reconstruction error exceeds the set threshold, specifically including the following steps:

[0112] Step S61: analyze the reconstruction error distribution of the verification set, select a suitable threshold setting method such as fixed threshold method, dynamic threshold method or threshold method based on statistical distribution, and determine the anomaly detection threshold threshold;

[0113] Step S62: compare the reconstruction error e i of the verification set with the threshold threshold, if e i > threshold, then determine that the battery monomer data at this moment is an abnormal point.

[0114] Step S63: if the battery fails, then locate the failed monomer according to the monomer voltage of the failed monomer.

[0115] Wherein, locating the failed monomer includes locating the failed monomer number and the abnormal point position.

[0116] Example 2:

[0117] In order to illustrate the effectiveness of the present application, this embodiment prepares a running data of a fault vehicle, wherein one battery monomer fails and is alarmed by the BMS. Figure 3 is the monomer voltage curve of the normal monomer and the failed monomer of the fault vehicle battery. The data is subjected to Kalman filtering and normalization processing. Figure 4 is the monomer voltage Kalman filtering curve of the normal monomer and the failed monomer of the fault vehicle battery. According to the division of the first 70% normal data and the last 30% abnormal data, the first 70% normal data is defined as the training set, and the last 30% abnormal data is defined as the verification set. The Chebyshev polynomial order m = 10 is selected, Figure 5 is the feature extraction result, the low-order feature reflects the main trend of the data, and the high-order feature may capture the subtle changes in the data. The Chebyshev polynomial and the saved coefficients are used to reconstruct the approximation of the original data. Figure 6The original data of the training set is compared with the reconstructed data, Figure 7 The original data of the verification set is compared with the reconstructed data. The reconstruction error of the verification set is used to detect the anomaly, the reconstruction error of the training set is observed, and a fixed threshold of 0.335 is set. Figure 8 The reconstruction error of the training set is Figure 9 The reconstruction error of the verification set is. The mean square error (MSE) of the training set and the verification set is less than 0.2, and the reconstruction error of the training set is less than 0.335. Figure 10 The anomaly detection result of the verification set has no false positives for normal monomers and can effectively identify faulty monomers and abnormal points. Therefore, the application can effectively extract the main features of the data, reduce noise interference, and capture subtle changes in the data, and can better achieve real-time fault detection of the power battery of the electric vehicle and accurately locate the faulty monomer.

[0118] Finally, it should be pointed out that the above examples are only used to illustrate the technical solutions of the application and are not limiting. Although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions, and they should be covered in the scope of the claims of the application.

Claims

1.A battery abnormality detection method based on Chebyshev polynomials, characterized by, The method specifically comprises the following steps: S1: Collecting electric vehicle operation data, including battery pack voltage data and other parameters of each battery monomer, and performing data preprocessing; S2: Kalman filtering processing is performed on the preprocessed battery monomer voltage data; and the filtered data is divided into a training set and a validation set; S3: Based on the filtered battery monomer voltage data, Chebyshev polynomial feature extraction is performed, and polynomial coefficients are extracted as features; S4: The Chebyshev polynomial coefficients are used to reconstruct the approximate value of the original data, and the reconstruction error is calculated; S5: Setting an anomaly detection threshold, judging whether each monomer battery fails according to whether the reconstruction error exceeds the set threshold. 2.The Chebyshev polynomial based battery anomaly detection method of claim 1, wherein, Step S1 specifically comprises the following steps: S11: Collecting normal operation data of a certain electric vehicle, including total voltage, each battery monomer voltage and time; S12: Filtering and cleaning the extracted battery monomer voltage data, and filling the missing values of the data; S13: Based on the collected operation data of the power battery pack and each battery monomer, a power battery operation database is established. 3.The Chebyshev polynomial based battery anomaly detection method of claim 1, wherein, In step S2, the preprocessed battery monomer voltage data is subjected to Kalman filtering processing, which specifically comprises the following steps: S21: Based on the preprocessed monomer voltage data, the initial state estimation and the estimation error covariance are determined; S22: The mean and variance of the first few data points are used as the initial state estimation, the state at the current time is predicted using the state estimation at the last time, and the estimation error covariance is updated; S23: The Kalman gain is calculated according to the current measurement value, the predicted state is corrected to obtain a more accurate estimation, and the estimation error covariance is updated; S24: Each battery monomer voltage time series is subjected to filtering processing, and the filtering process outputs a smooth and stable voltage data sequence at each time point. 4.The Chebyshev polynomial based battery anomaly detection method of claim 1, wherein, In step S2, the filtered data is divided into a training set and a validation set, specifically: the filtered time-ordered data is divided into the first 70% normal data and the last 30% abnormal data; based on the divided data, the first 70% normal data is defined as the training set, and the last 30% abnormal data is defined as the validation set. 5.The Chebyshev polynomial based battery anomaly detection method of claim 1, wherein, Step S3 specifically comprises the following steps: S31: According to experience or through cross-validation, the Chebyshev polynomial order m is selected to balance the accuracy and computational complexity of feature extraction; S32: Constructing a Chebyshev polynomial basis function; define m+1 Chebyshev polynomial basis functions T j (x), where j = 0, 1,..., m; Satisfying the recursive relationship: T0(x) = 1 T1(x) = x T j+1 (x) = 2xT j (x) - T j-1 (x), j > 1 wherein x represents an independent variable, i.e., a corresponding number of mapping the battery monomer voltage data to the zero point of the Chebyshev polynomial, is an equidistant point on the interval [-1, 1]; m represents the order of the Chebyshev polynomial; j represents the order index of the Chebyshev polynomial, which is a non-negative integer, used to identify different Chebyshev polynomials; T j (x) represents the j-th order Chebyshev polynomial, which is a function of x, used to represent Chebyshev polynomials of different orders, T0(x) is the 0-th order Chebyshev polynomial, T1(x) is the 1-st order Chebyshev polynomial, T j+1 (x) is the j+1-th order Chebyshev polynomial, T j-1 (x) is the j-1-th order Chebyshev polynomial; S33: According to the Chebyshev polynomial basis function j(x), for the voltage data of each single battery in the training set and the validation set, the Chebyshev polynomial is fitted using the least square method, and the polynomial coefficient c is obtained j , so that the error sum of squares of the fitted curve and the original data is minimized; the extracted coefficient c j As a feature of the battery monomer voltage data, it reflects the trend and change pattern of the data. 6.The Chebyshev polynomial-based battery abnormality detection method according to claim 5, wherein Step S4 Specifically comprises the following steps: S41: Reconstructing an approximation value of the original data using the extracted Chebyshev polynomial coefficients c j and the basis function T j (x), reconstructing an approximation value of the original data wherein, represents an approximation of the reconstructed raw data of the i-th data point; x i represents the i-th independent variable; T j (x i ) represents the value of the j-th Chebyshev polynomial at x i ; S42: compare the original data y i and the reconstructed data calculate the reconstruction error 7. The Chebyshev polynomial based battery anomaly detection method of claim 1, wherein, Step S5 specifically comprises the following steps: S51: Analyzing the reconstruction error distribution of the validation set, selecting a threshold setting method to determine the anomaly detection threshold threshold; S52: The reconstruction error e of the validation set i Compare with the threshold; if e i If the threshold is greater than 1, then the data of a single battery cell at that moment is considered an outlier. S53: If the battery fails, the failed monomer battery is located according to the monomer battery voltage. 8.The Chebyshev polynomial based battery anomaly detection method of claim 7, wherein, In step S51, the selected threshold setting method includes a fixed threshold method, a dynamic threshold method or a threshold method based on statistical distribution.

Images