BMS Status Detection Method Based on Big Data

Abstract

The present invention discloses a BMS state detection method based on big data. The method includes data acquisition, data optimization, alleviating battery area differences, preliminary abnormal point monitoring, and abnormal identification. The present invention belongs to the field of battery detection, specifically referring to a BMS state detection method based on big data. In this solution, the capacity increment is feedback-corrected by introducing environmental parameters; the second derivative of the battery health characteristics and the environmental optimization coefficient are introduced to adjust the battery health state estimation according to different environmental changes; the temperature sensitivity coefficient and the environmental optimization coefficient are introduced to accurately reflect the battery health condition; thereby improving the accuracy of BMS state detection; by calculating the local abnormal score and introducing a local non-uniformity enhancement mechanism; and through a dynamic temperature compensation mechanism, the detection accuracy is improved; combining the dynamic load factor and the local abnormal score to ensure the accurate marking of abnormal points and improve the BMS state detection effect.

Description

Technical Field

[0001] The present invention relates to the technical field of battery detection, and specifically refers to a method for detecting the state of a BMS based on big data. Background Art

[0002] The BMS state detection method refers to the technologies and methods used to monitor and evaluate the health status and performance of batteries in a battery management system. These methods monitor the working state of the battery by collecting various data of the battery, and judge the health status, charge and discharge efficiency, aging degree, etc. of the battery, so as to optimize the use of the battery, extend its life and prevent failures. However, the general BMS state detection method has the problems that the collected data cannot reflect the actual health status of the battery, the lack of environmental sensitivity leads to insufficient detection accuracy, and thus the accuracy of BMS state detection is poor; the general BMS state detection method cannot accurately capture abnormal conditions in a local area when the battery state is unevenly distributed, and fails to timely identify some abnormalities or misidentify normal data as abnormal. Summary of the Invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method for detecting the state of a BMS based on big data. Aiming at the problems that the general BMS state detection method has the problems that the collected data cannot reflect the actual health status of the battery, the lack of environmental sensitivity leads to insufficient detection accuracy, and thus the accuracy of BMS state detection is poor, this solution converts the voltage curve of the battery into a capacity increment curve, and introduces environmental parameters to feedback and correct the capacity increment; by introducing the second derivative of the battery health characteristics and the environmental optimization coefficient, the health state estimation of the battery can be adjusted according to different environmental changes; by introducing the temperature sensitivity coefficient and the environmental optimization coefficient, the battery management system can dynamically adjust according to the actual environmental changes and accurately reflect the health status of the battery; thereby improving the accuracy of BMS state detection; aiming at the problems that the general BMS state detection method cannot accurately capture abnormal conditions in a local area when the battery state is unevenly distributed, and fails to timely identify some abnormalities or misidentify normal data as abnormal, this solution calculates the local anomaly score and introduces a local non-uniformity enhancement mechanism, which can accurately identify the abnormal points in the low non-uniformity area; and through the dynamic temperature compensation mechanism, considering the influence of environmental temperature on battery performance, adjusts the abnormal detection score of the battery according to the external temperature change, so that the BMS can adapt to the change of battery performance under different temperature conditions and improve the detection accuracy; combining the dynamic load factor and the local anomaly score, realizing a dynamically adjusted anomaly recognition process, and thus ensuring the accurate marking of abnormal points and improving the BMS state detection effect.

[0004] The technical solution adopted by the present invention is as follows: The method for detecting the state of a BMS based on big data provided by the present invention includes the following steps:

[0005] Step S1: Data collection;

[0006] Step S2: Data optimization;

[0007] Step S3: Alleviate battery area differences;

[0008] Step S4: Preliminary abnormal point monitoring;

[0009] Step S5: Abnormality identification.

[0010] Furthermore, in step S1, the data collection is to collect BMS status data; the BMS status data includes voltage collection, current collection, temperature collection, state of charge health collection, Coulomb count collection, environmental parameter collection, battery module status information, and the health status of battery cells; and the collected data is converted into a vector form and normalized.

[0011] Furthermore, in step S2, the data optimization is to calculate the capacity increment Z and introduce environmental parameter feedback , and correct the capacity increment and voltage change of the battery, expressed as: ; ; where is the change in battery voltage; s is the time series index, and u and r are index parameters; is within the time period to the current value; T, H, and P are temperature, humidity, and external pressure respectively; , and are standard environmental values; , and are environmental sensitivity coefficients; , and are influence parameters on the capacity increment; based on the change of the capacity increment curve, health characteristics related to battery aging are extracted, including peak position, peak height, valley position, and valley height; calculate the second derivative ER of the battery health characteristics and introduce the environmental optimization coefficient , expressed as: ; ; where l is the starting point index; is the change in battery charge under voltage change; Q is the battery charge; , and are parameters used to adjust the influence of battery health decline; , and It is a parameter for adjusting the sensitivity of battery health degradation; replace the voltage acquisition collected in step S1 with the capacity increment, the extracted peak position, peak height, valley position, valley height, and the second derivative of the battery health characteristics; and then obtain the final BMS status dataset.

[0012] Further, in step S3, the alleviation of battery regional differences is to introduce a thermal balance factor , define two battery cells and the inhomogeneity change between them , expressed as: ; ; where and are the weight parameters of the battery cells and are the feature vectors of the battery cells; is the joint feature vector of the two battery cells; is the temperature sensitivity coefficient; is the external environmental temperature; is the standard operating temperature of the battery system; perform K-means clustering on the BMS status dataset, and based on the inhomogeneity change between battery cells, merge adjacent clusters with inhomogeneity change lower than the inhomogeneity threshold to obtain the BMS status data grouping; pre-divide the cases of different grouping numbers; calculate the module index M for different grouping numbers respectively, expressed as: ; where H is the number of groups and h is the group index; is the number of data points in the h-th group; is the within-group variance; N is the total number of samples; is the variance of the entire dataset; select the grouping case corresponding to the maximum module index; and then achieve the alleviation of battery regional differences.

[0013] Further, in step S4, the preliminary outlier monitoring is based on calculating the outlier score within each data point group and adjusting the outlier score through local inhomogeneity enhancement; specifically: calculate the local outlier score, expressed as: ; add a dynamic temperature compensation mechanism, and the local inhomogeneity enhancement is expressed as: ; ; ; where is the local outlier score of the i-th data point; and are the data point feature values, and i, j, and k are the data point indices; y is the mean of the data point features; is the variance of the i-th data point within the local neighborhood; n is the total number of data points within the group; is the normalized distance between data points; is the local non-uniformity enhancement value; is the local anomaly score of the k-th data point; is the temperature compensation factor; is the influence coefficient; is the amplification coefficient of the influence of temperature on performance; is the adjusted anomaly score; is the mean value of the local non-uniformity enhancement values of all data points.

[0014] Furthermore, in step S5, the anomaly recognition adopts dynamic threshold setting to cope with the dynamic distribution of BMS data, thereby realizing anomaly recognition; introducing a load factor , the dynamic threshold Td is expressed as: : where is the mode of the adjusted anomaly scores; is the standard deviation of the adjusted anomaly scores; is the adjustment factor; is the absolute value of the charging current; is the absolute value of the discharging current; is the standard load current of the battery; mark the BMS status data corresponding to the data points with local anomaly scores lower than the dynamic threshold as anomalies.

[0015] The beneficial effects achieved by the present invention using the above solution are as follows:

[0016] (1) Aiming at the problems existing in the general BMS status detection method that the collected data cannot reflect the actual health status of the battery, the lack of environmental sensitivity leads to insufficient detection accuracy, and further leads to poor accuracy of BMS status detection. This solution converts the voltage curve of the battery into a capacity increment curve and introduces environmental parameters to feedback and correct the capacity increment; by introducing the second derivative of the battery health characteristics and the environmental optimization coefficient, the health status estimation of the battery can be adjusted according to different environmental changes; by introducing the temperature sensitivity coefficient and the environmental optimization coefficient, the battery management system can be dynamically adjusted according to the actual environmental changes, accurately reflecting the health status of the battery; thereby improving the accuracy of BMS status detection.

[0017] (2)Regarding the problem that the general BMS status detection method cannot accurately capture abnormal conditions in local areas when the battery status is unevenly distributed, and fails to timely identify some abnormalities or misidentify normal data as abnormal, this solution can accurately identify abnormal points in low-uniformity areas by calculating local anomaly scores and introducing a local non-uniformity enhancement mechanism; and by considering the influence of ambient temperature on battery performance through a dynamic temperature compensation mechanism and adjusting the abnormal detection score of the battery according to external temperature changes, the BMS can adapt to the changes in battery performance under different temperature conditions, improving the accuracy of detection; combining the dynamic load factor and local anomaly scores, a dynamically adjusted abnormal recognition process is realized, thereby ensuring the accurate marking of abnormal points and improving the BMS status detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic flowchart of the BMS status detection method based on big data provided by the present invention.

[0019] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0021] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the system or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0022] Embodiment 1, refer to Figure 1 , the BMS status detection method based on big data provided by the present invention, the method includes the following steps:

[0023] Step S1: Data collection; collect BMS status data;

[0024] Step S2: Data optimization; convert the battery voltage curve into a capacity increment curve and perform feedback correction in combination with environmental parameters; obtain the final BMS status data set;

[0025] Step S3: Alleviate battery area differences; group and optimize the BMS status data based on the inhomogeneity changes among battery cells through the K-means clustering method;

[0026] Step S4: Preliminary outlier monitoring; realize preliminary outlier monitoring by calculating the local outlier score, introducing a local inhomogeneity enhancement mechanism, and combining dynamic temperature compensation to adjust the outlier score;

[0027] Step S5: Outlier identification; identify and mark the outliers in the BMS status data through dynamic threshold setting according to the battery load and the adjusted outlier score.

[0028] Embodiment 2, refer to Figure 1 , this embodiment is based on the above embodiment. In step S1, the BMS status data includes voltage acquisition, current acquisition, temperature acquisition, charging health status acquisition, Coulomb count acquisition, environmental parameter acquisition, battery module status information, and the health status of battery cells; and the collected data is converted into a vector form and standardized; the voltage acquisition includes the voltage of single battery cells, the total voltage of the battery pack, and the voltage difference between battery cells; the current acquisition includes discharge current, charging current, and battery charge-discharge rate; the temperature acquisition includes the temperature of single battery cells, the overall temperature of the battery pack, and local temperature differences; the charging health status acquisition includes SOC acquisition and SOH acquisition; the Coulomb count acquisition includes charge-discharge power and charge-discharge times; the environmental parameter acquisition includes external pressure, external temperature, and humidity; the battery module status information includes the charge-discharge status of the battery module and the voltage, temperature, and current of the module; the health status of the battery cells includes the capacity of the battery cells and the self-discharge situation of the battery.

[0029] Embodiment 3, refer to Figure 1 , this embodiment is based on the above embodiment. In step S2, data optimization is to reveal the changes in the electrochemical characteristics of the battery by converting the voltage curve of the battery into a capacity increment curve; calculate the capacity increment Z, and introduce environmental parameter feedback , and correct the capacity increment and voltage change of the battery, expressed as: ; ; where is the change in battery voltage; s is the time series index, and u and r are index parameters; is within the time period to the current value; T, H, and P are temperature, humidity, and external pressure respectively; , and are the standard environmental values; , and is the environmental sensitivity coefficient; 、 and are the influence parameters on the capacity increment; based on the change of the capacity increment curve, the health characteristics related to battery aging are extracted, including the peak position, peak height, valley position and valley height; furthermore, it helps to judge the degree of battery capacity attenuation; calculate the second derivative ER of the battery health characteristics, and introduce the environmental optimization coefficient , expressed as: ; ; where l is the starting point index; is the change in battery charge under voltage change; Q is the battery charge; 、 and are the parameters used to adjust the influence of battery health decline; 、 and are the parameters used to adjust the sensitivity of battery health decline; replace the voltage acquisition collected in step S1 with the capacity increment, the extracted peak position, peak height, valley position and valley height, and the second derivative of the battery health characteristics; thus, the final BMS state data set is obtained.

[0030] By performing the above operations, aiming at the problem that the data collected by the general BMS state detection method cannot reflect the actual health status of the battery, and the lack of environmental sensitivity leads to insufficient detection accuracy, which in turn leads to poor accuracy of BMS state detection. This solution converts the voltage curve of the battery into a capacity increment curve, and introduces environmental parameters to perform feedback correction on the capacity increment; by introducing the second derivative of the battery health characteristics and the environmental optimization coefficient, the health state estimation of the battery can be adjusted according to different environmental changes; by introducing the temperature sensitivity coefficient and the environmental optimization coefficient, the battery management system can be dynamically adjusted according to the actual environmental changes, accurately reflecting the health status of the battery; thus improving the accuracy of BMS state detection.

[0031] Example 4, refer to Figure 1 , based on the above example, in step S3, alleviating the battery regional difference is to perform K-means clustering on the BMS state data set and group them, and optimize the number of groups based on the module index to ensure high homogeneity within the region and high non-uniformity between regions; specifically: introduce the thermal balance factor , define the non-uniformity change and between two battery cells , expressed as: ; ; where, and are the weight parameters of the battery cells and is the eigenvector of the battery cell; is the combined eigenvector of two battery cells; is the temperature sensitivity coefficient; is the external environmental temperature; is the standard operating temperature of the battery system; Perform K-means clustering on the BMS status data set. Based on the non-uniformity change between battery cells, merge adjacent clusters with non-uniformity changes lower than the non-uniformity threshold to obtain the BMS status data grouping; Pre-divide the cases with different numbers of groupings; Calculate the module index M for different numbers of groupings respectively, expressed as: ; where H is the number of groupings and h is the grouping index; is the number of data points in the h-th group; is the within-group variance; N is the total number of samples; is the variance of the entire data set; Select the grouping case corresponding to the maximum module index; Further achieve alleviating the battery area difference.

[0032] Example Five, refer to Figure 1 , this example is based on the above example. In step S4, the preliminary outlier monitoring is based on calculating the outlier scores within each data point group, and adjusting the outlier scores through local non-uniformity enhancement to make the outliers in the low non-uniformity area more prominent; Specifically: Calculate the local outlier score, expressed as: ; Add a dynamic temperature compensation mechanism, and the local non-uniformity enhancement is expressed as: ; ; ; where is the local outlier score of the i-th data point; and are the data point eigenvalue, and i, j, and k are the data point indices; y is the mean of the data point eigenvalues; is the variance of the i-th data point within the local neighborhood; n is the total number of data points within the group; is the normalized distance between data points; is the local non-uniformity enhancement value; is the local outlier score of the k-th data point; is the temperature compensation factor; is the influence coefficient; is the amplification coefficient of the influence of temperature on performance; is the adjusted outlier score; is the mean of the local non-uniformity enhancement values of all data points.

[0033] Example Six, refer to Figure 1, based on the above embodiment, in step S5, anomaly recognition adopts dynamic threshold setting to cope with the dynamic distribution of BMS data, thereby realizing anomaly recognition; a load factor is introduced , the dynamic threshold Td is expressed as: : where is the mode of the adjusted anomaly score; is the standard deviation of the adjusted anomaly score; is the adjustment factor; is the absolute value of the charging current; is the absolute value of the discharging current; is the standard load current of the battery; the BMS status data corresponding to the data points with local anomaly scores lower than the dynamic threshold is marked as abnormal.

[0034] By performing the above operations, for the problem that the general BMS status detection method cannot accurately capture anomalies in local areas when the battery status distribution is uneven, fails to identify some anomalies in time or misidentifies normal data as anomalies, this solution can accurately identify anomaly points in low non-uniformity areas by calculating local anomaly scores and introducing a local non-uniformity enhancement mechanism; and through a dynamic temperature compensation mechanism, it considers the influence of environmental temperature on battery performance, adjusts the anomaly detection score of the battery according to external temperature changes, so that the BMS can adapt to changes in battery performance under different temperature conditions and improve the accuracy of detection; combined with the dynamic load factor and local anomaly scores, a dynamically adjusted anomaly recognition process is realized, thereby ensuring the accurate marking of anomaly points and improving the BMS status detection effect.

[0035] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0036] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention.

[0037] The above describes the present invention and its implementation manners. Such description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. In general, if those of ordinary skill in the art are inspired by it and, without departing from the gist of the present invention, design similar structural manners and embodiments to this technical solution without creative efforts, they shall fall within the protection scope of the present invention.

Claims

1. A method for detecting the state of a BMS based on big data, characterized in that: The method includes the following steps: Step S1: Data collection; collect BMS status data; Step S2: Data optimization; Step S3: Alleviate battery area differences; Step S4: Preliminary abnormal point monitoring; Step S5: Abnormality identification; identify and mark the abnormal points in the BMS status data; In step S2, the data optimization is to calculate the capacity increment Z and introduce environmental parameter feedback , and correct the capacity increment and voltage change of the battery, expressed as: ; ; where is the change in battery voltage; s is the time series index, and u and r are index parameters; is the current value within the time period to ; T, H, and P are temperature, humidity, and external pressure respectively; , and are the standard environmental values; , and are the environmental sensitivity coefficients; , and are the influence parameters on the capacity increment; Based on the change of the capacity increment curve, health characteristics related to battery aging are extracted, including peak position, peak height, valley position, and valley height; Calculate the second derivative ER of the battery health characteristics and introduce the environmental optimization coefficient , expressed as: ; ; where l is the starting point index; is the change in battery charge under voltage change; Q is the battery charge; , and are the parameters used to adjust the impact of battery health decline; , and are the parameters used to adjust the sensitivity of battery health decline; Replace the voltage acquisition collected in step S1 with the capacity increment, the extracted peak position, peak height, valley position, and valley height, and the second derivative of the battery health characteristics; Then obtain the final BMS status dataset.

2. The BMS status detection method based on big data according to claim 1, characterized in that: In step S3, the mitigation of battery regional differences is to introduce a thermal balance factor , define two battery cells and The uneven variation between , expressed as: ; ;in, and is the weight parameter of the battery cell; and is the characteristic vector of the battery cell; is the joint eigenvector of the two battery cells; is the temperature sensitivity coefficient; is the external ambient temperature; is the standard operating temperature of the battery system; perform K-means clustering on the BMS status data set, merge adjacent clusters whose heterogeneity changes are lower than the heterogeneity threshold based on the heterogeneity changes between battery cells, and obtain BMS status data grouping; pre-divide the situation of different number of groups; calculate the module index M for different number of groups respectively, expressed as: ; Where H is the number of groups and h is the group index; is the number of data points in the hth group; is the within-group variance; N is the total number of samples; is the variance of the entire data set; select the grouping corresponding to the maximum module indicator; and thus alleviate the regional differences in batteries.

3. The BMS status detection method based on big data according to claim 2, characterized in that: In step S4, the preliminary anomaly point monitoring is based on calculating the anomaly scores within each data point group and adjusting the anomaly scores through local non-uniformity enhancement; specifically: calculating the local anomaly score, expressed as: ; adding a dynamic temperature compensation mechanism, the local non-uniformity enhancement is expressed as: ; ; ; where, is the local anomaly score of the i-th data point; and are the data point eigenvalue, i, j, and k are data point indices; y is the mean of the data point eigenvalues; is the variance of the i-th data point within the local neighborhood; n is the total number of data points within the group; is the normalized distance between data points; is the local non-uniformity enhancement value; is the local anomaly score of the k-th data point; is the temperature compensation factor; is the influence coefficient; is the amplification coefficient of the influence of temperature on performance; is the adjusted anomaly score; is the mean of the local non-uniformity enhancement values of all data points.

4. The method for detecting the BMS state based on big data according to claim 3, wherein: In step S5, the anomaly recognition adopts dynamic threshold setting to cope with the dynamic distribution of BMS data, so as to realize anomaly recognition; a load factor is introduced , and the dynamic threshold Td is expressed as: : where is the mode of the adjusted anomaly score; is the standard deviation of the adjusted anomaly score; is the adjustment factor; is the absolute value of the charging current; is the absolute value of the discharging current; is the standard load current of the battery; the BMS status data corresponding to the data points with local anomaly scores lower than the dynamic threshold are marked as anomalies.

5. The method for detecting the state of the BMS based on big data according to claim 4, wherein: In step S1, the BMS status data includes voltage collection, current collection, temperature collection, charging health status collection, Coulomb count collection, environmental parameter collection, battery module status information, and the health status of battery cells; and the collected data is converted into a vector form and standardized.

Images