A method for extracting envelope features of lithium batteries for random usage data

Abstract

This invention belongs to the field of lithium battery health management and state assessment technology, and particularly relates to a method for extracting envelope features of lithium batteries for random usage data, including the following steps: S1: Obtain the original sampling sequence of various key features of the target battery; the key features include cumulative usage features, health status features, operating condition features, and time series features; the original sampling sequence includes the cumulative usage time series, health status time series, operating condition time series, and corresponding time index sequences, the time index sequences being used to identify or record the order or moment of occurrence of the original sampling sequence in time, and to record the time point of data collection; S2: Construct a multidimensional feature vector F using the obtained original sampling sequence… This method can effectively improve the accuracy of lithium battery remaining life prediction and state assessment by collecting, analyzing, and processing multidimensional data such as cumulative usage, health status, and operating condition of lithium batteries.

Description

Technical Field

[0001] This invention belongs to the field of lithium battery health management and status assessment technology, and particularly relates to a method for extracting lithium battery envelope features for random usage data. Background Technology

[0002] With the widespread application of lithium batteries in electric vehicles, energy storage systems, and other fields, accurately assessing the battery's State of Health (SOH) and predicting its Remaining Useful Life (RUL) has become a critical issue. Traditional battery state assessment methods mostly rely on regular sampling data or data acquisition under ideal operating conditions, making it difficult to cope with complex and irregular real-world usage environments. That is, existing envelope methods are typically mainly applied to processing time-frequency signals, such as speech, vibration, and biological signals. These methods help analyze the frequency components and trends of signals by extracting their envelope features. However, in applications such as electric vehicles and energy storage systems, lithium batteries often operate under random conditions, making assessment methods based on fixed conditions unable to accurately reflect the battery's true state. Therefore, developing a feature extraction method that can adapt to complex operating conditions and utilize randomly available data is of great significance.

[0003] In particular, lithium batteries, as the most commonly used energy storage solution, directly impact the driving range of electric vehicles, the operating efficiency of energy storage devices, and their safety. Therefore, effective methods for health management and lifespan prediction of lithium batteries are crucial. A technology is needed that can randomly sample data under random operating conditions to ensure accurate assessment of the state of lithium batteries, thereby improving their safety and cost-effectiveness. Summary of the Invention

[0004] This invention provides a method for extracting envelope features of lithium batteries based on random usage data. This method can effectively improve the accuracy of lithium battery remaining life prediction and condition assessment by collecting, analyzing and processing multi-dimensional data such as cumulative usage, health status and operating condition of lithium batteries.

[0005] To solve the above-mentioned technical problems, this application provides the following technical solution:

[0006] A method for extracting the envelope features of lithium batteries for randomly used data includes the following steps:

[0007] S1: Obtain the original sampling sequence of each key feature of the target battery; the key features include cumulative usage features, health status features, operating condition features, and time series features; the original sampling sequence includes the cumulative usage time series, health status time series, operating condition time series, and corresponding time index sequence, the time index sequence is used to identify or record the order or moment of the original sampling sequence in time, and record the time point of data collection;

[0008] S2: Construct a multidimensional feature vector F using the acquired original sampling sequence; calculate the point distance between the corresponding values ​​of the multidimensional feature vector F at each time step and the origin of the coordinate system, and generate the distance feature vector D;

[0009] S3: Perform envelope processing on the obtained distance feature vector D to obtain discrete envelope distance values, and determine the corresponding discrete envelope time.

[0010] S4: Obtain the order of discrete envelope moments in the time index sequence, sample from the original sampling sequences of each key feature, and construct the envelope sampling sequence of each key feature.

[0011] S5: Based on the relationship between the envelope sampling sequence of each key feature and the discrete envelope time, construct an interpolation function; calculate the feature value of each key feature on the original time index through the interpolation function, and generate a continuous interpolation repair sequence of each key feature.

[0012] The basic principles and beneficial effects of this scheme are as follows: First, key characteristic data, including cumulative usage, health status, operating condition, and time-series characteristics, are collected from the operation of the lithium battery to form time-series data. A multi-dimensional feature vector is constructed using the collected data, and then the point distance between the value of this vector at each time point and the origin is calculated to generate a distance feature vector. Envelope processing is performed on the distance feature vector to extract discrete envelope distance values ​​and corresponding discrete envelope times. Based on the order of the discrete envelope times in the time index sequence, key feature values ​​are extracted from the original sampling sequence to construct an envelope sampling sequence. An interpolation function is constructed using the relationship between the envelope sampling sequence and the discrete envelope times. This function is used to calculate the feature values ​​at the original time index, generating a continuous interpolation repair sequence.

[0013] This solution improves accuracy: by using envelope processing and interpolation functions, key features of lithium batteries can be extracted and repaired more accurately, thus improving the accuracy of feature extraction.

[0014] Enhanced Robustness: This method can handle randomly used data, making feature extraction unrestricted by specific operating conditions or usage patterns, thus enhancing its robustness. Real-time Monitoring and Prediction: Through continuous interpolation and sequence repair, real-time monitoring and prediction of lithium battery health status can be achieved, which is of great significance for battery maintenance and lifespan prediction. Improved Data Utilization: This method can effectively utilize the collected data, improving data utilization and the completeness of information extraction by constructing multi-dimensional feature vectors and envelope sampling sequences. Enhanced Safety: Real-time monitoring of battery health status is essential for the safe operation of battery systems. This method helps to promptly identify potential battery problems and prevent safety accidents. Reduced Maintenance Costs: Accurate assessment of battery health status allows for more rational scheduling of battery maintenance and replacement, reducing maintenance costs and avoiding unnecessary replacements.

[0015] The first difference between this solution and existing technologies lies in its ability to process random data. Existing lithium battery life prediction methods are primarily based on regular sampling or fixed operating conditions, failing to effectively handle irregular or random data. In contrast, this method innovatively targets random operating condition data, accurately extracting multidimensional features of lithium batteries (such as cumulative usage, health status, and operating condition) in complex usage environments and constructing feature vectors—something impossible in existing technologies. This enhanced random data processing capability frees battery state assessment from ideal conditions, thereby improving the practical applicability of battery management. This is crucial for applications such as electric vehicles and energy storage systems.

[0016] The second difference between this case and existing technologies lies in the combination of multidimensional feature vectors and distance features. Traditional methods typically focus only on a single key feature of the battery, neglecting its complex, multidimensional characteristics. This patent provides a comprehensive description of the lithium battery's state of being through the construction of multidimensional feature vectors and the generation of distance feature vectors. This allows the method to more accurately assess the overall health of the battery when faced with complex data. This multidimensional analysis approach provides a deeper understanding of the overall battery evaluation, helping to identify potential performance problems in a timely manner under varying usage environments.

[0017] The third difference between this case and existing technologies lies in the application of envelope processing technology. Existing technologies lack effective processing methods when faced with irregular data, while this patent is the first to introduce envelope processing technology (such as Hilbert transform and extreme point interpolation) into the field of lithium battery data processing, enabling the extraction of key information from complex data and improving evaluation accuracy. The application of envelope processing technology opens up new directions for battery data analysis, allowing for the extraction of effective information even under irregular data conditions. This innovation improves the reliability of battery evaluation and promotes the development of related fields.

[0018] The fourth difference between this case and existing technologies lies in the innovative application of interpolation repair. Existing data repair methods often employ simple linear interpolation, which cannot handle complex discrete data. This method innovatively uses interpolation repair techniques following envelope sampling to generate continuous feature sequences from discrete data, overcoming the problem of data discontinuity and improving the accuracy of the evaluation and the flexibility of data processing. The application of interpolation repair technology enhances the continuity and integrity of the data, providing a more reliable data foundation for subsequent analysis. This innovation ensures that the evaluation results still have high credibility even when facing incomplete data.

[0019] The fifth difference between this case and existing technologies lies in the flexible selection of multiple distance calculation methods. Traditional methods often employ a single distance calculation method, while this method introduces multiple distance calculation methods, including Euclidean distance and Mahalanobis distance, providing flexible options for feature extraction and thus achieving optimal evaluation results in different application scenarios. The flexible application of multiple distance calculation methods allows this method to select the most suitable analysis method for specific data features. This innovation ensures improved accuracy in battery state assessment and guarantees the accuracy of the assessment results.

[0020] The fifth difference between this method and existing technologies lies in its ability to predict battery life under complex application scenarios. Existing battery life prediction models typically assume stable operating conditions and cannot adapt to complex and variable real-world usage environments. This method, through dynamic analysis of multi-dimensional features and the integration of historical data, can provide more accurate predictions under uneven usage conditions. This ability to adapt to complex application scenarios allows the method to better meet the needs of battery management, thereby improving battery safety and reliability. This adaptability enhances the flexibility and practicality of batteries in different applications.

[0021] This solution is adaptable to complex operating conditions. This innovation can extract key characteristics of lithium batteries from irregular and random data, adapting to complex usage scenarios and effectively improving the accuracy of battery state assessment. Through adaptive processing of complex operating conditions, this method effectively enhances the battery state assessment capability, ensuring reliable data acquisition under different usage environments. This is crucial for ensuring the safe operation of electric vehicles and energy storage systems.

[0022] This solution integrates multi-dimensional data. By constructing and analyzing multi-dimensional feature vectors, it enables a comprehensive assessment of factors such as cumulative battery usage, health status, and operating condition, providing a more comprehensive basis for subsequent lifespan prediction. The integration of multi-dimensional data not only improves the comprehensiveness of the assessment but also provides a more scientific basis for subsequent maintenance and management decisions. Through comprehensive analysis, potential problems can be identified in a timely manner, and corresponding measures can be taken.

[0023] This solution achieves a seamless integration of envelope processing and interpolation repair. Envelope processing technology is used to extract key information from the data, and interpolation repair methods are used to reconstruct the data sequence. This enables the generation of continuous feature sequences even when the original data is incomplete, improving the robustness and accuracy of data processing. The innovation lies in enhancing the continuity and integrity of the data through the combination of envelope processing and interpolation repair. This lays a more solid foundation for subsequent data analysis and battery state assessment, making the assessment results more reliable.

[0024] This solution enables flexible distance calculation. By combining multiple distance calculation methods, such as Euclidean distance and Manhattan distance, it flexibly adapts to different data characteristics, improving the accuracy of feature extraction. The flexible distance calculation method allows the invention to select the most suitable calculation method based on the data conditions of different features. This flexibility ensures the accuracy of feature extraction and helps to further improve the effect of battery evaluation. Furthermore, it includes step S6: using at least one of the envelope sampling sequence of each key feature, the discrete sampling sequence of each key feature, and the sampling repair sequence of each key feature; using historical data as input and remaining lifetime as output, to train the model.

[0025] Furthermore, the health status time series includes changes in battery capacity, changes in battery power capacity, changes in internal resistance, and changes in charge / discharge efficiency.

[0026] Furthermore, the cumulative usage characteristics include cumulative charging power, cumulative discharging power, total cumulative absolute value charging and discharging power, cumulative charging time, cumulative discharging time, total cumulative charging and discharging time, cumulative charging amount, cumulative discharging amount, total cumulative absolute value charging and discharging amount, cumulative number of charging cycles, cumulative number of discharging cycles, total cumulative number of charging and discharging cycles, cumulative number of idle cycles, cumulative idle time, cumulative calendar service time, cumulative charging ratio, cumulative discharging ratio, total cumulative absolute value of charging ratio and discharging ratio, cumulative charging power ratio, cumulative discharging power ratio, total cumulative absolute value of charging power ratio and discharging power ratio, cumulative actual workload generated by the battery module supplying power to the operating equipment, cumulative actual work done by the battery module supplying power to the operating equipment, and cumulative actual mileage generated by the battery module supplying power to the vehicle.

[0027] Furthermore, the health status characteristics include: actual maximum energy storage capacity, actual maximum energy storage capacity attenuation, actual maximum power storage capacity, actual maximum power storage capacity attenuation, relative energy storage capacity, relative energy storage capacity attenuation, relative power storage capacity attenuation, actual internal resistance, actual internal resistance attenuation, single actual discharge duration, single actual discharge duration attenuation, relative discharge duration, single actual charging duration, single actual charging duration attenuation, relative charging duration, actual work done by the battery's actual maximum energy storage capacity for the operation of power-consuming equipment, and the mileage generated by the battery's actual maximum energy storage capacity for the vehicle's driving range.

[0028] Furthermore, the operating condition characteristics include: various operating conditions during battery operation; during actual battery operation, different operating condition settings will affect battery performance. The selectable types of operating conditions include: specific changes or averages of the battery's output current, output voltage, and output power during battery operation; air humidity, heat dissipation conditions, air pressure conditions, equipment operating power, equipment production efficiency, and vehicle speed.

[0029] Furthermore, the point distance calculation in step S2 includes any one of Manhattan distance, Euclidean distance, Hamming distance, standardized Euclidean distance, Mahalanobis distance, Chebyshev distance, and Minkowski distance.

[0030] Furthermore, the sampling of the discrete envelope time in step S3 is selected from either uniform sampling or non-uniform sampling.

[0031] Envelope processing methods include extreme point interpolation, Hilbert transform, filtering, and moving average.

[0032] Furthermore, the sampling step for the interpolation repair sequence in step S5 is as follows: First, spline interpolation is performed on the envelope sampling sequence of each key feature to obtain the interpolation function for the cumulative usage as S. x (·), to achieve interpolation of cumulative usage and time series; the interpolation function for health status is S. y (·), to achieve interpolation of health status and time series; the interpolation function for operating conditions is S. c (·), to achieve interpolation of operating conditions and time series.

[0033] This invention addresses scientific problems such as random operating condition data processing, comprehensive analysis of multi-dimensional features, and data envelopment processing and interpolation repair.

[0034] This solution addresses the problem of processing random operating condition data, specifically how to extract representative features from irregular or random usage data to assess the health status of lithium batteries and predict their remaining lifespan. Processing random operating condition data is a significant challenge in the current field of lithium battery evaluation. Through an effective feature extraction method, this invention provides a solution that enables the acquisition of true state information of lithium batteries even under irregular data conditions. Emphasizing the unique and complex nature of lithium batteries, this invention offers greater flexibility and adaptability when processing such data.

[0035] This case study aims to address the challenge of comprehensive multi-dimensional feature analysis, specifically how to combine information such as cumulative usage, health status, and operating condition of lithium batteries to generate a feature vector reflecting the overall condition of the battery. Comprehensive analysis of multi-dimensional features is crucial for improving the accuracy of battery condition assessment. By integrating data from different dimensions, a more comprehensive understanding of the lithium battery's operating status and potential failure risks can be achieved, thereby optimizing maintenance strategies. This process emphasizes the performance of lithium batteries in variable environments, providing a holistic perspective for their health management.

[0036] This case aims to address the problem of how to implement data envelopment processing (DEF) and interpolation repair, specifically how to extract key information from irregular data using DEF techniques and generate continuous feature sequences through interpolation repair to compensate for data incompleteness. The combination of DEF and interpolation repair helps improve the robustness of the evaluation model. In the case of missing data, interpolation repair can ensure the model's accuracy, thus providing a reliable data foundation for lithium battery life prediction. This process reflects an emphasis on the characteristics of lithium battery data, making the feature extraction process more meticulous.

[0037] This solution enables the construction of multi-dimensional feature vectors. By extracting multi-dimensional data such as the cumulative usage, health status, and operating condition of lithium batteries, multi-dimensional feature vectors are generated. This comprehensive feature extraction method surpasses the single-feature analysis methods in existing technologies, significantly improving the comprehensiveness and accuracy of battery status assessment. The construction of multi-dimensional feature vectors allows battery evaluation to go beyond a single performance indicator, integrating information from multiple aspects. This innovation provides crucial support for a comprehensive understanding of battery status.

[0038] This solution achieves an innovative application of envelope processing. By introducing envelope processing techniques (such as Hilbert transform and extreme point interpolation) into the field of lithium battery state assessment, it is possible to extract feature values ​​from random data and generate envelope sequences—an innovative application not previously used in existing technologies. This innovative application of envelope processing provides a new perspective for battery data analysis, enabling the extraction of key information even in cases of data irregularity. This innovation enhances the flexibility and applicability of the assessment.

[0039] This scheme combines interpolation and restoration techniques. After extracting discrete feature values, the method generates a continuous feature sequence through interpolation and restoration, and can combine multiple interpolation methods (such as spline interpolation) to enhance the continuity and accuracy of the feature values. The combination of interpolation and restoration techniques not only compensates for missing data but also provides greater continuity for the analysis of the feature sequence. This innovation makes the evaluation results more reliable and helps to better reflect the true state of the battery.

[0040] This solution can flexibly adapt to random data. Most existing technologies are designed for regular data or specific operating conditions, while this method, through adaptive processing of random data and multi-dimensional feature extraction, can flexibly cope with complex usage scenarios. This ability to flexibly adapt to random data allows the method to maintain high evaluation accuracy even when facing complex and ever-changing real-world usage conditions. This characteristic makes this invention have broad application prospects. Attached Figure Description

[0041] Figure 1 This is a line chart of the original data;

[0042] Figure 2 This is a scatter plot of the original data;

[0043] Figure 3 A schematic diagram showing the results after applying the envelope feature extraction method from this case to the original data;

[0044] Figure 4 This is a schematic diagram of the final envelope curve result;

[0045] Figure 5 This is a schematic diagram illustrating the steps of a lithium battery envelope feature extraction method for randomly used data. Detailed Implementation

[0046] The following detailed description illustrates the specific implementation method:

[0047] Example 1 is attached. Figure 5 As shown,

[0048] S1: Obtain the original sampling sequence of each key feature of the rechargeable battery. The key features of the lithium battery include cumulative usage, health status, operating condition, and time series characteristics.

[0049] The original sampling sequences of key features of lithium batteries include cumulative usage time series, health status time series, operating condition time series, and corresponding time index sequences.

[0050] Time-indexed sequences can identify or record the order or timing of certain events or data. They are typically used to record the time points of data collection for subsequent analysis and comparison of data changes over different time periods; subsequently, {t1, t2, t3, ... t...} are used. k} refers to the time index sequence, where each time point is t1, t2, t3, ... t k The spacing between them may not be equal;

[0051] Operating condition time series includes feature values ​​of various operating conditions recorded at different time points, which can be used to describe the state changes of a system or equipment under different operating conditions; subsequently, {c1,c2,c3,…c k} refers to {t1,t2,t3,…t} k The corresponding operating condition time series;

[0052] A health state time series describes a set of data depicting the changes in the health state of a rechargeable battery at different points in time. These characteristics can include changes in battery capacity, power capacity, internal resistance, and charge / discharge efficiency. By recording and analyzing this time series data, the battery's health status can be assessed, its future performance predicted, and appropriate maintenance or operation strategies formulated; subsequently, {y1, y2, y3, ... y...} k} refers to {t1,t2,t3,…t} k The corresponding health status time series;

[0053] A health state time series describes a set of data depicting the changes in the health state of a rechargeable battery at different points in time. These characteristics can include changes in battery capacity, power capacity, internal resistance, and charge / discharge efficiency. By recording and analyzing this time series data, the battery's health status can be assessed, its future performance predicted, and appropriate maintenance or operation strategies formulated; subsequently, {y1, y2, y3, ... y...} k} refers to {t1,t2,t3,…t} k The corresponding health status time series;

[0054] The cumulative usage characteristics include: cumulative charging power, cumulative discharging power, total cumulative absolute value of charging and discharging power, cumulative charging time, cumulative discharging time, total cumulative charging and discharging time, cumulative charging amount, cumulative discharging amount, total cumulative absolute value of charging and discharging amount, cumulative number of charging cycles, cumulative number of discharging cycles, total cumulative number of charging and discharging cycles, cumulative number of idle cycles, cumulative idle time, cumulative calendar service time, cumulative charging ratio, cumulative discharging ratio, total cumulative absolute value of charging ratio and discharging ratio, cumulative charging power ratio, cumulative discharging power ratio, total cumulative absolute value of charging power ratio and discharging power ratio, cumulative actual workload generated by the battery module for power-consuming equipment operation, cumulative actual work done by the battery module for power-consuming equipment operation, and cumulative actual mileage generated by the battery module for vehicle driving.

[0055] Health status characteristics include: actual maximum energy storage capacity, actual maximum energy storage capacity decay, actual maximum power storage capacity, actual maximum power storage capacity decay, relative energy storage capacity, relative energy storage capacity decay, relative power storage capacity decay, relative power storage capacity decay, actual internal resistance, actual internal resistance decay, single actual discharge duration, single actual discharge duration decay, relative discharge duration, single actual charging duration, single actual charging duration decay, relative charging duration, actual work done by the battery's actual maximum energy storage capacity for power-consuming equipment operation, and the mileage that the battery's actual maximum energy storage capacity can provide for vehicle driving.

[0056] Operating condition characteristics include various operating conditions during battery operation; different operating condition settings will affect battery performance during actual battery operation. Optional operating condition types include: specific changes or averages of battery output current, output voltage, and output power during battery operation; air humidity, heat dissipation conditions, air pressure conditions, equipment operating power, equipment production efficiency, vehicle speed, etc.

[0057] In this step and any subsequent steps, the original sampled sequence can be normalized or denormalized as needed. Normalization and denormalization are two key operations in data processing, especially when dealing with feature data of rechargeable batteries. Normalization maps data to a fixed range, usually [0,1] or [-1,1], making the scale of each feature consistent, which is beneficial for model learning and prediction. Denormalization, on the other hand, restores the normalized data to its original scale so that it can be interpreted and applied to actual values. The following is a detailed explanation of these two processes. The purpose of normalization is to transform data features of different scales into the same range to facilitate subsequent processing. Usually, different features in the original data (such as time, cumulative usage, health status, etc.) differ greatly in magnitude, which can affect the model training effect. Commonly used normalization methods include min-max normalization and Z-score standardization.

[0058] Min-Max Normalization: Min-Max Normalization maps data to the range [0,1]. For a feature sequence, such as the cumulative usage sequence X = {x1, x2, x3, ..., x...}, ... k The normalized value x' i The calculation formula is as follows:

[0059] x' i =(x i -x min ) / (x max -x min )

[0060] in:

[0061] ·x i It is a value in the original data.

[0062] ·x min and x max These are the minimum and maximum values ​​in sequence X, respectively.

[0063] After min-max normalization, all data will be scaled to between 0 and 1, giving the data a uniform scale. This method can be applied to cumulative usage, health status, and time-series data.

[0064] Z-score standardization: Z-score standardization adjusts the mean of the data to 0 and the variance to 1. The calculation formula is:

[0065] x' i =(x i -μ) / σ

[0066] in:

[0067] μ is the average value of sequence X.

[0068] σ is the standard deviation.

[0069] Z-score standardization has lower requirements for data distribution, making it suitable for situations where the original data may not fall within a fixed interval. This method makes the data more closely resemble a normal distribution, which is more compatible with some machine learning models.

[0070] Denormalization is the process of restoring data to its original scale after model processing. It enables normalized predicted data to be restored to meaningful original values, facilitating interpretation and application. For example, if health status data is normalized to between 0 and 1 during min-max normalization, denormalization can restore the predicted health status values ​​to the initial range.

[0071] Min-Max Inverse Normalization: For data x' that has already undergone min-max normalization i The original scale can be restored using the following formula:

[0072] x i =x' i *(x max -x min )+x min

[0073] This will restore the normalized value x'i to its original numerical range.

[0074] Z-score denormalization: For data that has already been standardized by Z-score, the original data can be restored using the following formula:

[0075] x i =x' i *σ+μ

[0076] This method can also restore the data from the standardized distribution to the original distribution, so that the predicted data can truly reflect the physical meaning of the original data.

[0077] The raw data sequence obtained in the feature acquisition step, such as the cumulative usage feature X = {x1, x2, x3, ..., x k Health status characteristics Y = {y1, y2, y3, ... y k These sequences may have variations in numerical range and distribution. Normalization can be performed on these sequences as needed so that the features can function effectively in subsequent processing or models.

[0078] After the model calculations are complete, to make the results meaningful, the processed data can be denormalized. For example, if the predicted values ​​of health status or cumulative usage have been normalized, they need to be denormalized back to the original scale for easier subsequent interpretation and application.

[0079] By normalizing and denormalizing, we can process and apply data with different characteristics more flexibly, ensuring consistency in data calculation and interpretation.

[0080] S2: Construct a multidimensional feature vector F using the acquired original sampling sequence; calculate the point distance between the corresponding values ​​of the multidimensional feature vector F at each time step and the origin of the coordinate system, and generate the distance feature vector D;

[0081] The distance types that can be adopted in the point distance calculation process include any one of Manhattan distance, Euclidean distance, Hamming distance, standardized Euclidean distance, Mahalanobis distance, Chebyshev distance, and Minkowski distance.

[0082] For example, the time index sequence, the cumulative usage time series, and the health status time series can be represented as T = {t1, t2, t3, ... t}. k}、X={x1,x2,x3,…,x k} and Y = {y1, y2, y3, ... y k Then, these three time series can be merged into a multidimensional feature vector F = {(t1,x1,y1),(t2,x2,y2),(t3,x3,y3),…(t…}. k ,x k ,y k Alternatively, simply combine the cumulative usage and health status time series to obtain a multidimensional feature vector F = {(x1,y1),(x2,y2),(x3,y3),…(x k ,y k )}.

[0083] To quantify the multidimensional data in the feature vector, a distance feature vector D can be constructed to represent the distance between each feature point and the origin. This can be represented as D = {d1, d2, d3, ... dn}. k} represents the distance feature vector corresponding to the multidimensional feature vector F = {(x1,y1),(x2,y2),(x3,y3),…(x4,y4)}, where d1 = distance((x1,y1),(0,0)) and d2 = distance((x2,y2),(0,0)), that is, the distances of each one-dimensional data point d1, d2, d3,…d in D. k The value of is equal to the values ​​of each multidimensional data point (x1, y1), (x2, y2), (x3, y3), ... (x) in F. k ,y k The distance between d and the multidimensional origin (0,0). Or, for each d... i (where i = 1, 2, ..., k), define d i =distance((x i ,yi ),(0,0)), that is, calculate (x i ,y i The Euclidean distance between the feature vector D and the multidimensional origin (0,0). In some cases, other distances can also be used. Thus, each element d in the feature vector D is... i This corresponds to the distance between each data point in the feature vector F and the origin.

[0084] S3: Perform envelope processing on the obtained distance feature vector to obtain discrete envelope distance values, and determine the corresponding discrete envelope time.

[0085] Depending on the specific distribution of the discrete envelope time intervals, the discrete envelope time intervals can be uniformly sampled or non-uniformly sampled.

[0086] For example, the discrete envelope distance values ​​after envelope processing are d2, d5, d 10 ,d 23 ,…d k-2 The order in the number distance feature vector is 2. nd 5 th 10 th ,twenty three th ,…k-2 th Therefore, the corresponding discrete envelope times are t2, t5, t 10 ,t 23 ,…t k-2 It is non-uniform;

[0087] Envelope processing is commonly used in signal processing to extract the envelope information of a signal, i.e., the shape of the signal. Optional envelope processing methods include: extreme point interpolation, Hilbert transform, filtering, and moving average.

[0088] The Hilbert transform is a commonly used method in signal processing to calculate the analytic signal of a signal and extract its envelope. The specific steps are to calculate the Hilbert transform of the signal to obtain the analytic signal, and then the magnitude of the analytic signal is the envelope.

[0089] The extreme point interpolation method first finds all local maxima in the signal, and then uses these maxima to perform interpolation (such as spline interpolation) to obtain the envelope.

[0090] The filtering method uses a low-pass filter to remove high-frequency components from the signal, thereby obtaining the envelope; first, a low-pass filter is generated, and then the signal is passed through the low-pass filter to obtain the envelope.

[0091] The moving average method smooths a signal by calculating its moving average value to obtain its envelope. First, a suitable window size is selected, and then the moving average value of the signal is calculated.

[0092] S4: Obtain the order of discrete envelope moments in the time index sequence, sample from the original sampling sequences of each key feature, and construct the envelope sampling sequence of each key feature.

[0093] S5: Based on the relationship between the envelope sampling sequence of each key feature and the discrete envelope time, construct an interpolation function to describe this correlation; calculate the feature value of each key feature on the original time index through the interpolation function, and generate a continuous interpolation repair sequence of each key feature.

[0094] For example, after envelope processing, the obtained discrete envelope distance values ​​are d2, d5, d 10 ,d 23 ,…d k-2 The order in the number distance feature vector is 2. nd 5 th 10 th ,twenty three th ,…k-2 th Then, the same order is used to sample the original sampling sequences of each key feature to obtain the envelope sampling sequence {x2, x5, x...} of the cumulative usage. 10 ,x 23 ,…x k-2}, Health status envelope sampling sequence {y2, y5, y 10 ,y 23 ,…y k-2}, Operating condition envelope sampling sequence {c2,c5,c 10 ,c 23 ,…c k-2};

[0095] When obtaining the sampling repair sequence, spline interpolation is first performed on the envelope sampling sequence of each key feature to obtain the interpolation function. For example, the interpolation function for the cumulative usage is S. x (·) can be used to interpolate cumulative usage and time series; the interpolation function for health status is S. y (·) can be used to interpolate health status and time series; the interpolation function for operating conditions is S. c (·) can be used to interpolate operating conditions and time series;

[0096] After obtaining the interpolation function, the original time index sequence {t1,t2,t3,…t} can be used. k Using} as the time base, an interpolation function is used to generate a cumulative usage sampling repair sequence {X1,X2,X3,…X}. k}, Health status sampling repair sequence {Y1,Y2,Y3,…Y} k}, Operating condition sampling repair sequence {C1,C2,C3,…C k};

[0097] S6: Use at least one of the following: envelope sampling sequence of each key feature, discrete sampling sequence of each key feature, and sampling repair sequence of each key feature.

[0098] The model is trained using historical data as input and remaining lifespan as output.

[0099] Figure 1 Line charts of raw data are used to display the changing trends of continuous data. Data points are connected by lines, making them suitable for representing changes in variables over time or sequentially. Figure 1 The distribution of raw capacity data relative to cumulative computation for three different lithium batteries (A, B, and C) is shown. The curves exhibit significant fluctuations, indicating substantial noise and irregular fluctuations in the raw data, making it unsuitable for accurate condition assessment or lifetime prediction.

[0100] Figure 2 Scatter plots of raw data are used to illustrate the relationship between two variables. Each point independently represents a data pair, and there are no connections between the points, making them suitable for observing the correlation between variables. Figure 2 The original data was converted into a scatter plot, showing the discrete distribution of capacity as a function of cumulative computational cost. Despite... Figure 1 While visualizations have become clearer, the data still exhibits significant dispersion and discontinuity, making it difficult to process directly using traditional methods. The charts at this stage highlight the challenges posed by data discontinuity.

[0101] Figure 3The results are obtained after applying the envelope feature extraction method described in this case. After processing, the original capacity data yielded three envelope curves: A, B, and C. By enveloping the data, noise was significantly suppressed, data fluctuations were smoothed, and the overall trend between capacity and cumulative computational cost was revealed. This step is crucial; through envelope feature extraction technology, smooth feature curves are generated from discrete data, providing a more stable foundation for subsequent analysis. The difference between the new method in this case and existing methods lies in the fact that traditional methods, such as Hilbert transform, extreme point interpolation, and filtering, are mainly used to process periodic and regular time-frequency signals. These methods extract the envelope features of the signal by capturing its frequency and time components, performing well in time-frequency domains such as vibration and speech signals. However, these methods face significant challenges when processing lithium battery data because the operating environment and usage of lithium batteries are often random, and the data lacks obvious periodicity. In contrast, the new envelope feature extraction method is more suitable for the complex and random usage data environment of lithium batteries. By combining multidimensional feature analysis and advanced envelope algorithms, this method effectively eliminates noise and extracts the core health characteristics and usage patterns of lithium batteries. This groundbreaking extraction technology overcomes the limitations of traditional methods, making it applicable not only to periodic signals but also to random and nonlinear operating conditions in battery systems. The new envelope feature extraction method in this case... Figures 2 to 3 It plays a decisive role in the conversion process. Compared with traditional methods, it can more effectively process random usage data of lithium batteries, reduce noise interference, and improve the accuracy and stability of feature extraction through multidimensional data analysis, making it suitable for more complex usage scenarios and lifespan prediction tasks.

[0102] Figure 4 The final envelope curve results are presented, further simplifying and removing unnecessary fluctuations, and showing the continuous change in capacity with cumulative computation. This step demonstrates the powerful capability of the new envelope extraction method in handling complex data, capable of eliminating noise while preserving the core features of the data, generating highly accurate battery characteristic curves, and providing strong support for battery life prediction and condition monitoring.

[0103] The above are merely embodiments of the present invention. The invention is not limited to the fields covered by these embodiments. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are able to access all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for extracting the envelope features of lithium batteries for randomly used data, characterized in that, Includes the following steps: S1: Obtain the original sampling sequence of each key feature of the target battery; the key features include cumulative usage features, health status features, operating condition features, and time series features; the original sampling sequence includes the cumulative usage time series, health status time series, operating condition time series, and corresponding time index sequence, the time index sequence is used to identify or record the order or moment of the original sampling sequence in time, and record the time point of data collection; S2: Construct a multidimensional feature vector F using the acquired original sampling sequence; calculate the point distance between the corresponding values ​​of the multidimensional feature vector F at each time step and the origin of the coordinate system, and generate the distance feature vector D; S3: Perform envelope processing on the obtained distance feature vector D to obtain the discrete envelope distance value, and determine the corresponding discrete envelope time. S4: Obtain the order of discrete envelope moments in the time index sequence, sample from the original sampling sequences of each key feature, and construct the envelope sampling sequence of each key feature. S5: Based on the relationship between the envelope sampling sequence of each key feature and the discrete envelope time, construct an interpolation function; calculate the feature value of each key feature on the original time index through the interpolation function, and generate a continuous interpolation repair sequence of each key feature.

2. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: It also includes step S6: using at least one of the envelope sampling sequence of each key feature, the discrete sampling sequence of each key feature, and the sampling repair sequence of each key feature; using historical data as input and remaining lifetime as output, train the model.

3. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The health status time series includes changes in battery capacity, changes in battery power capacity, changes in internal resistance, and changes in charge / discharge efficiency.

4. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The cumulative usage characteristics include cumulative charging power, cumulative discharging power, total cumulative absolute value charging and discharging power, cumulative charging time, cumulative discharging time, total cumulative charging and discharging time, cumulative charging amount, cumulative discharging amount, total cumulative absolute value charging and discharging amount, cumulative number of charging cycles, cumulative number of discharging cycles, total cumulative number of charging and discharging cycles, cumulative number of idle cycles, cumulative idle time, cumulative calendar service time, cumulative charging ratio, cumulative discharging ratio, total cumulative absolute value of charging ratio and discharging ratio, cumulative charging power ratio, cumulative discharging power ratio, total cumulative absolute value of charging power ratio and discharging power ratio, cumulative actual workload generated by the battery module for power-consuming equipment operation, cumulative actual work done by the battery module for power-consuming equipment operation, and cumulative actual mileage generated by the battery module for vehicle driving.

5. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The health status characteristics include: actual maximum energy storage capacity, actual maximum energy storage capacity attenuation, actual maximum power storage capacity, actual maximum power storage capacity attenuation, relative energy storage capacity, relative energy storage capacity attenuation, relative power storage capacity attenuation, actual internal resistance, actual internal resistance attenuation, single actual discharge duration, single actual discharge duration attenuation, relative discharge duration, single actual charging duration, single actual charging duration attenuation, relative charging duration, actual work done by the battery's actual maximum energy storage capacity for the operation of power-consuming equipment, and the mileage generated by the battery's actual maximum energy storage capacity for the vehicle's driving range.

6. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The operating condition characteristics include: various operating conditions during battery operation; different operating condition settings will affect battery performance during actual battery operation; and the selectable types of operating conditions include: the specific changes or average values ​​of battery output current, output voltage, and output power during battery operation, air humidity, heat dissipation conditions, air pressure conditions, equipment operating power, equipment production efficiency, and vehicle speed.

7. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The types of point distance calculations in step S2 include any one of Manhattan distance, Euclidean distance, Hamming distance, standardized Euclidean distance, Mahalanobis distance, Chebyshev distance, and Minkowski distance.

8. The lithium battery envelope feature extraction method for randomly used data according to claim 1, characterized in that: The sampling of the discrete envelope time in step S3 is either uniform sampling or non-uniform sampling. Envelope processing methods include extreme point interpolation, Hilbert transform, filtering, and moving average.

9. A method for extracting lithium battery envelope features for randomly used data according to claim 8, characterized in that: The sampling steps for the interpolation repair sequence in step S5 are as follows: First, spline interpolation is performed on the envelope sampling sequence of each key feature to obtain the interpolation function for the cumulative usage as S. x (·), to achieve interpolation of cumulative usage and time series; the interpolation function for health status is S. y (·), to achieve interpolation of health status and time series; the interpolation function for operating conditions is S. c (·), to achieve interpolation of operating conditions and time series.

Images