A lithium battery endurance data monitoring method and system

Abstract

The present application relates to the technical field of electric digital data processing, in particular to a kind of lithium battery endurance data monitoring method and system, comprising: obtaining the multidimensional parameter sequence data of lithium battery in each set time window in history, the deviation degree of each dimension parameter sequence data and the fluctuation degree of each parameter data in each dimension parameter sequence data are calculated, and the product of the two is used as correction factor;With the absolute value of the difference between parameter data and its predicted value and its standard deviation, calculate the first abnormality degree, then multiply the normalized correction factor to obtain the second abnormality degree;In response to the parameter data of second abnormality degree greater than set threshold value is abnormal data;If the proportion of abnormal data is greater than set proportion threshold, it is determined that there is an abnormality in lithium battery endurance.This application solves the problem of misjudgment or omission in the process of lithium battery endurance data monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric digital data processing. More particularly, the present application relates to a lithium battery endurance data monitoring method and system. BACKGROUND

[0002] With the promotion of new energy vehicles, electric tools, portable electronic devices and distributed energy storage systems, lithium batteries, as the current mainstream energy storage devices, have become the key foundation to support intelligent mobility and green energy development. Lithium batteries have the advantages of high energy density, long cycle life and low self-discharge rate, and are therefore applied in various scenarios. However, corresponding to the increasing application scale, users' requirements for battery endurance are also continuously improving, especially in complex working conditions, real-time monitoring, state evaluation and abnormal warning capabilities are more challenging. The battery endurance not only determines the continuous running time of the equipment, but also is closely related to the vehicle energy efficiency, safety guarantee and cost control. Therefore, building a scientific, efficient and reliable lithium battery endurance monitoring mechanism has become an important research direction in battery management systems.

[0003] At present, the mainstream lithium battery state monitoring method is mostly based on basic physical quantities such as voltage, current and temperature, and identifies the running abnormalities or performance degradation of the battery by setting fixed thresholds, constructing empirical models or using simple fitting strategies. Such methods have certain practicality in theory, but still have many deficiencies in actual application. The working state of lithium batteries presents typical nonlinearity, time-varying and volatility, and the parameter changes may be the result of normal response to external load changes, or may imply systematic degradation trends or potential fault characteristics.

[0004] Traditional abnormality identification methods often focus on static deviation at a certain moment, lack in-depth modeling of data change trends, and cannot distinguish between sudden abnormalities and normal transition states, limiting their application capabilities in intelligent state identification and accurate endurance warning. At the same time, since some monitoring methods do not combine historical data for periodic analysis, it is difficult to discover slowly evolving abnormal patterns, such as capacity degradation, active material loss, electrolyte decomposition, etc. These problems often accumulate without being detected, eventually leading to a sudden drop in battery performance or failure, increasing the safety risk, and thus leading to misjudgment or missed judgment. SUMMARY

[0005] To solve the misjudgment or missed judgment problem raised in the background art, the present application provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for monitoring lithium battery endurance data, comprising: acquiring multi-dimensional parameter sequence data of the lithium battery within historical set time windows; acquiring the deviation degree of the parameter sequence data of each dimension and the fluctuation degree of each parameter data in the parameter sequence data of each dimension; using the product of the deviation degree and the fluctuation degree as a correction factor for each parameter data; acquiring the absolute value of the difference between each parameter data and its corresponding predicted value in historical set time windows, and the standard deviation of the sequence formed by the absolute values, using the ratio between the absolute value and the standard deviation as a first degree of anomaly of each parameter data, and using the product of the first degree of anomaly and the correction factor after standardization as a second degree of anomaly; in response to parameter data whose second degree of anomaly is greater than a set threshold being considered abnormal data, if the proportion of the abnormal data is greater than a set proportion threshold, then determining that the lithium battery endurance is abnormal.

[0007] The above technical solution not only improves the accuracy and robustness of anomaly detection, but also effectively identifies hidden anomalies in highly fluctuating or multi-parameter coupled environments. It is particularly suitable for complex scenarios in lithium battery operating data where nonlinearity, time-varying nature, and multi-source noise coexist. Finally, by determining the proportion of anomalies, it achieves a quantitative assessment of the overall health status of lithium battery endurance performance, solving the problem of misjudgment or missed judgment.

[0008] Furthermore, the degree of deviation for: , , The first In the sequence data of the nth dimension parameter, the first The, the The fitted values ​​corresponding to the parameter data, , The first In the sequence data of the nth dimension parameter, the first The, the One parameter data, For the first The total number of data contained in the sequence of parameters for each dimension, and the fitted value is obtained by curve fitting of the sequence of parameters for each dimension.

[0009] The above technical solution significantly enhances the ability to perceive and express complex abnormal patterns by incorporating both the overall fitting error and local trend deviation into the deviation measurement system, providing a solid foundation for the accurate diagnosis of the subsequent lithium battery range health status.

[0010] Furthermore, the degree of fluctuation for: , For the natural constant An exponential function with base 0. , The first In the sequence data of the nth dimension parameter, the first The fitted value corresponding to the parameter data, the first parameter data One parameter data, For the first The total number of data points contained in the sequence of parameters for each dimension, wherein the fitted value is obtained by curve fitting to the sequence of parameters for each dimension. , For the first The sequence number of the parameter data in the dimensional parameter sequence data.

[0011] The aforementioned technical solution centers on each target data point, comprehensively considering its temporal distance to surrounding data points. By assigning higher weights to adjacent locations, it makes the calculation of fluctuation degree focus more on local changes and reduces the interference of distant points on fluctuation assessment. Compared to traditional global fluctuation indicators, this structure is more sensitive to the identification of short-term drastic fluctuations or small-scale anomalies, and can more accurately capture abnormal disturbances in the data while maintaining the stability of overall trend judgment. Especially in scenarios where lithium battery operating parameters have strong time correlation and dynamic change characteristics, it helps to improve the response capability to high-frequency minor anomalies, providing a more timely and reliable basic indicator for subsequent anomaly correction and status early warning.

[0012] Furthermore, the multi-dimensional parameter sequence data includes: lithium battery voltage data, current data, temperature data, and internal resistance data.

[0013] Furthermore, the multi-dimensional parameter sequence data is subjected to data cleaning processing.

[0014] The above technical solutions can effectively remove missing values, duplicate values ​​and obvious erroneous values, and smooth or impute abnormal noise, making the original data more continuous, reliable and consistent.

[0015] Furthermore, the standardization process is Z-score standardization.

[0016] Furthermore, the curve fitting is a polynomial fitting or an exponential fitting.

[0017] Furthermore, the set percentage threshold is 0.1.

[0018] Furthermore, the predicted value is obtained using exponential smoothing.

[0019] The above technical solution generates predicted values ​​by using the exponential smoothing method, which can fully extract the changing trends of historical data and give higher weight to the latest data changes while maintaining computational efficiency, so that the prediction results have good timeliness and responsiveness.

[0020] In a second aspect, the present application provides a lithium battery endurance data monitoring system comprising a memory and a processor, wherein the memory stores computer program instructions which, when executed by the processor, implement the lithium battery endurance data monitoring method of any one of the above.

[0021] The present application has the following beneficial effects:

[0022] The present application constructs an abnormality recognition and dynamic monitoring mechanism based on multi-dimensional parameters, integrates multiple key factors such as data fitting deviation, change trend consistency, local volatility, and prediction deviation, forms a set of layer-by-layer progressive and mutually corroborative abnormality recognition system, significantly improves the recognition accuracy and response sensitivity of lithium battery endurance abnormality. Not only realizes the dual perception of abnormality amplitude and trend change at the parameter level, but also realizes the quantitative evaluation and hierarchical response of abnormality degree by introducing the standardization processing and threshold mechanism, has stronger adaptability and stability. At the same time, combined with the exponential smoothing prediction method, the ability to capture the change trend of time series is improved, so that the system can identify the signs of performance degradation or potential failure of lithium battery in the early stage, and enhance the reliability of endurance prediction and the forward-looking of early warning. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a flow chart of a lithium battery endurance data monitoring method according to an embodiment of the present application;

[0024] Figure 2 is a structural block diagram of a lithium battery endurance data monitoring system according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] A lithium battery endurance data monitoring method embodiment.

[0026] As Figure 1 shown, the flow chart of the lithium battery endurance data monitoring method of the embodiment of the present application comprises the following steps:

[0027] S1: Obtain multi-dimensional parameter sequence data of the lithium battery in each set time window in history.

[0028] In a preferred embodiment, the set time window is one day, so as to accurately reflect the state change of the lithium battery in the conventional charge and discharge cycle. Of course, it can also be set according to the actual situation to adapt to different granularity analysis requirements.

[0029] The multi-dimensional parameter sequence data includes lithium battery voltage data, current data, temperature data, and internal resistance data. These parameters can comprehensively depict the operating status and performance characteristics of the lithium battery from multiple perspectives. Among them, voltage and current reflect the battery's energy output capability, temperature indicates thermal management status, and internal resistance reveals the degree of aging or potential fault characteristics inside the battery.

[0030] To ensure the accuracy of subsequent anomaly detection and status assessment, the collected multi-dimensional parameter sequence data needs to be cleaned. This process includes removing missing values, eliminating obvious erroneous values, interpolating to complete discontinuous data points, and combining moving average or median filtering methods to eliminate occasional fluctuations, thereby effectively improving the stability and reliability of the data. Implementing this step can significantly reduce the false positive and false negative rates, laying a solid data foundation for subsequent feature extraction and anomaly identification.

[0031] S2: Obtain the deviation of the parameter sequence data for each dimension and the fluctuation of each parameter data in the parameter sequence data for each dimension, and use the product of the two as a correction factor.

[0032] In one embodiment, the degree of deviation for: , , The first In the sequence data of the nth dimension parameter, the first The, the The fitted values ​​corresponding to the parameter data, , The first In the sequence data of the nth dimension parameter, the first The, the One parameter data, For the first The total number of data contained in the parameter sequence data of each dimension, the fitting value is obtained by curve fitting of the parameter sequence data of each dimension, the curve fitting is a polynomial fitting or an exponential fitting.

[0033] By integrating the dual indicators of residual magnitude and consistency of trend, a more discriminative method for calculating deviation is constructed. This method not only examines the absolute difference between the original parameter data and its curve-fitted values ​​to measure the degree of data deviation, but also further introduces the ratio of the rate of change between adjacent data points to evaluate the consistency between the trend and the fitted trend, thus effectively avoiding misjudgments that may arise from relying solely on point-to-point errors. This composite calculation mode, which couples residuals and trends, significantly improves the sensitivity and identification ability of abnormal patterns, enhancing the model's accuracy in depicting complex abnormal behaviors in lithium battery operation.

[0034] In one embodiment, the fluctuation degree is: , is an exponential function with a natural constant as the base, , are respectively the fitting value corresponding to the i-th parameter data in the i-th dimensional parameter sequence data, the i-th parameter data, is the total number of data contained in the i-th dimensional parameter sequence data, and the fitting value is obtained by curve fitting each dimensional parameter sequence data, , is the serial number of the parameter data in the i-th dimensional parameter sequence data.

[0035] By introducing a weighted residual average method based on exponential decay weight, the sensitivity and stability of fluctuation degree calculation to local abnormal changes are effectively improved. The above method considers the relationship between each target point and other time points, and gives higher weight to adjacent data points through an exponential function, thereby highlighting the fitting deviation of the point in the local range while maintaining the accuracy of overall trend judgment. Compared with the traditional global or equal weight processing method, this design with position decay mechanism can more accurately capture the influence of abnormal disturbance in local time series, reduce the interference of distant irrelevant points on fluctuation judgment, improve the model's ability to describe abnormal fluctuations, and strengthen the response ability to key parameter disturbance characteristics, which is helpful to realize early warning and dynamic tracking of small abnormalities in lithium battery operating state.

[0036] The product of the deviation degree and the fluctuation degree is taken as the correction factor of each parameter data. By multiplying the deviation degree and the fluctuation degree to construct the correction factor, the comprehensive trade-off and quantification of parameter data abnormality and instability are realized, so that the identification result not only reflects the degree of data deviation from the fitting trend, but also reflects the fluctuation intensity in the local sequence, fuses the amplitude abnormality and dynamic disturbance two types of information, effectively enhances the robustness and accuracy of abnormal detection, avoids misjudgment or missed detection caused by a single indicator, and thus improves the comprehensive evaluation ability and discrimination precision of lithium battery state abnormality.

[0037] S3: Obtain the absolute value of the difference between each parameter data and its corresponding predicted value, the standard deviation of the sequence formed by the absolute value, take the ratio of the two as the first abnormality degree, and take the product of the first abnormality degree and the normalized correction factor as the second abnormality degree.

[0038] ​​​​​In one embodiment, to further improve the accuracy and discrimination ability of lithium battery operating parameter anomaly detection, the absolute value of the difference between each parameter data and its corresponding predicted value is first obtained, and the absolute value sequence is constructed. The absolute value reflects the deviation between the actual observation value and the predicted trend at each time, and the standard deviation of the constructed absolute value sequence describes the fluctuation amplitude and distribution dispersion of the deviation in the entire time window. By taking the ratio of the two as the first abnormality degree, the normalization measurement of single-point deviation in the overall fluctuation background can be realized, that is, if the deviation of a certain data point is much higher than the overall fluctuation standard, it will be identified as a potential abnormal point. This normalization processing avoids the deviation caused by the inconsistency of dimensions or fluctuation benchmarks between different parameters or different time periods, thereby enhancing the universality and stability of anomaly detection.

[0039] In a preferred embodiment, the predicted value is obtained by exponential smoothing method. Further, the correction factor obtained by multiplying the deviation degree and the fluctuation degree is standardized, and the standardization method adopts Z-score standardization.

[0040] Finally, the first abnormality degree is multiplied by the standardized correction factor to obtain the second abnormality degree. The product integrates two levels of abnormality information: on the one hand, the relative significance of local deviation is measured by ratio, and on the other hand, the weight correction of abnormal trend and fluctuation background is introduced by correction factor, so as to realize more accurate and objective comprehensive evaluation of abnormal points. Through this fusion strategy, abnormal deviations that occur within the normal fluctuation range or small abnormalities that are more risky in high fluctuation background can be effectively identified, which improves the early warning ability and judgment credibility of lithium battery state monitoring system for potential faults, performance degradation and other implicit problems, and provides strong data support for subsequent battery management strategy.

[0041] S4: The parameter data is abnormal data in response to the second abnormality degree being greater than a set threshold, and if the proportion of the abnormal data is greater than a set proportion threshold, it is determined that the lithium battery endurance is abnormal.

[0042] In one embodiment, to realize accurate judgment of potential endurance anomaly in lithium battery operating state, further based on the calculated second abnormality degree, each parameter data point is identified for anomaly. Specifically, a global abnormality judgment threshold is set, and when the second abnormality degree of a certain data point exceeds the threshold, it is considered that the deviation of the point from the normal operating state has reached a significant level, and therefore the point can be identified as an abnormal data point.

[0043] To avoid misjudgments caused by individual abnormal data points triggering the warning mechanism, the system further introduces a global proportion control mechanism. This mechanism calculates the proportion of data points identified as abnormal within a set historical time window and compares it to a preset abnormality proportion threshold. If the proportion of abnormal data exceeds this threshold, the system determines that the current lithium battery's operating state has a systematic deviation, potentially indicating a degradation in range, performance deterioration, or an early risk of failure. This method expands the assessment from point-based anomalies to surface-based anomalies, extending from the identification of individual data anomalies to a comprehensive evaluation of the overall operating state, thus avoiding erroneous warnings based solely on fluctuations in individual data points.

[0044] The aforementioned threshold ratio mechanism has significant technical effects: on the one hand, it enhances the system's adaptability to complex operating conditions such as persistent anomalies, periodic deviations, or multidimensional anomaly synergistic changes, and can identify non-random structural problems; on the other hand, it effectively filters out isolated deviations caused by instantaneous disturbances, measurement errors, or random environmental factors, thereby improving the system's fault tolerance and stability.

[0045] The set percentage threshold is 0.1, but it can also be set according to the actual situation.

[0046] The present invention constructs a multi-dimensional parameter fusion-based endurance monitoring mechanism, comprehensively considering the deviation and fluctuation characteristics of various key parameters of lithium batteries within a historical time window. It utilizes curve fitting residuals and trend deviations to construct highly sensitive anomaly indicators, and employs exponential decay weighted calculations to accurately characterize local fluctuations, effectively improving the ability to identify minute anomalies under complex nonlinear operating conditions. By introducing the ratio of predicted values ​​to residual standard deviations, it enhances the normalized perception of actual parameter deviations. Combined with a standardized correction factor to constitute a second anomaly level, it ensures that anomaly judgment considers both the intensity of numerical deviation and the background of fluctuations, significantly improving the accuracy and robustness of anomaly detection. Furthermore, by incorporating an anomaly data proportion threshold judgment mechanism, it achieves a shift from single-point anomaly identification to overall state assessment, avoiding false alarms and enhancing the reliability of intelligent early warnings.

[0047] An embodiment of a lithium battery endurance data monitoring system:

[0048] like Figure 2 As shown in the figure, a structural block diagram of a lithium battery endurance data monitoring system according to an embodiment of the present invention includes a processor and a memory.

[0049] This invention also provides a lithium battery endurance data monitoring system. For example... Figure 2 As shown, the system includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement a lithium battery life data monitoring method according to the present invention.

[0050] The lithium battery endurance data monitoring system also includes other components such as communication interfaces, which are well known to those skilled in the art, and their settings and functions are known in the art, so they are not described here.

[0051] In the present application, the aforementioned memory can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any appropriate magnetic storage medium or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random-Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store the desired information and can be accessed by an application, module, or both. Any such computer storage media can be part of a device or accessible or connectable to a device. Any application or module described in the present application can be implemented by computer-readable instructions stored or otherwise held by such computer-readable media.

[0052] In the description of the present application, the meaning of "a plurality of", "several" is at least two, for example, two, three or more, etc., unless otherwise explicitly specified.

[0053] Although the present application has shown and described the preferred embodiments of the present application, it will be apparent to those skilled in the art that many modifications, changes and substitutions can be made thereto without departing from the spirit and scope of the present application. It should be understood that various alternatives to the embodiments of the present application described herein can be employed in practicing the present application.

Claims

1. A method for monitoring lithium battery endurance data, characterized in that, The method comprises: acquiring multi-dimensional parameter sequence data of a lithium battery in each historical set time window; acquiring a deviation degree of each dimension parameter sequence data and a fluctuation degree of each parameter data in each dimension parameter sequence data; multiplying the deviation degree and the fluctuation degree as a correction factor of each parameter data; deviation degree is: , , are the fitting values corresponding to the first , the first , and the first parameter data in the first , are the first , the first , and the first parameter data in the first dimensional parameter sequence data, and the fitting value is obtained by curve fitting on each dimensional parameter sequence data; total number of data contained in the first dimensional parameter sequence data. acquiring an absolute value of a difference between each parameter data and a corresponding predicted value in each historical set time window, a standard deviation of a sequence constituted by the absolute value, a first abnormality degree of each parameter data as a ratio between the absolute value and the standard deviation, and a second abnormality degree as a product of the first abnormality degree and a normalized correction factor; in response to the second abnormality degree of a parameter data being greater than a set threshold value, the parameter data being abnormal data, and a proportion of the abnormal data being greater than a set proportion threshold value, determining that there is an abnormality in the endurance of the lithium battery.

2. The method of claim 1, wherein, The fluctuation degree Is: , Is an exponential function with a natural constant As the base, , The fitting value corresponding to the first The first parameter data in the first Dimensional parameter sequence data, Parameter data, The total number of data contained in the first Dimensional parameter sequence data, the fitting value is obtained by curve fitting each dimensional parameter sequence data, , The serial number of the parameter data in the first Dimensional parameter sequence data.

3. The method of claim 1, wherein, The multi-dimensional parameter sequence data comprises voltage data, current data, temperature data and internal resistance data of the lithium battery.

4. The lithium battery endurance data monitoring method of claim 1, wherein, The multi-dimensional parameter sequence data is subjected to data cleaning processing.

5. The method of claim 1, wherein, The standardization processing is Z-score standardization.

6. The lithium battery endurance data monitoring method according to claim 1 or 2, characterized in that, The curve fitting is polynomial fitting or exponential fitting.

7. The method of claim 1, wherein, The set proportion threshold value is 0.

1.

8. The method of claim 1, wherein, The predicted value is obtained by an exponential smoothing method.

9. A lithium battery endurance data monitoring system characterized by, The method comprises a memory and a processor, and the memory stores computer program instructions, which, when executed by the processor, implement the method for monitoring the endurance data of the lithium battery according to any one of claims 1-8.

Images