SOC anomaly detection method for battery pack

Abstract

The invention provides an SOC anomaly detection method for a battery pack. The method comprises the steps that S1, a plurality of charging and discharging operation data fragments are acquired based on historical actual operation data of the battery pack; s2, according to the plurality of charging and discharging operation data segments, calculating to obtain a sequence corresponding to the time and the actual available capacity value under each charging and discharging operation data segment; and S3, inputting a plurality of sequences corresponding to the time and the actual available capacity values into a mutation detection algorithm and a cumulative change detection algorithm in a sliding window mode according to a time sequence, obtaining an SOC abnormal time point through the mutation detection algorithm, and obtaining an SOC abnormal window fragment through the cumulative change detection algorithm.

Description

Technical Field

[0001] The present application relates to the field of energy storage, and in particular to a method for detecting SOC anomalies in a battery pack. Background Art

[0002] In the field of power electronics, battery packs are widely used as energy storage elements in various scenarios, such as electric vehicles, portable electronic devices, and energy storage.

[0003] The battery pack consists of multiple single cells connected in series, parallel, or series-parallel. The battery pack also includes a Battery Management System (BMS), which manages and monitors the battery pack. For example, it collects information such as the total battery pack voltage, maximum and minimum single cell voltages, battery pack current, maximum and minimum temperatures, cumulative battery pack charge, and battery pack SOC at regular sampling intervals.

[0004] The battery's state of charge (SOC) refers to the ratio of the battery's remaining capacity to its fully charged capacity. SOC is crucial for battery energy management, safety monitoring, and health maintenance, so estimating SOC is very necessary.

[0005] Current methods all focus on how to accurately estimate SOC. However, there are few methods for detecting anomalies in the estimated SOC. Even if there are, there are many problems, such as strong parameter dependence, computational complexity, high resource consumption, and poor effect. Summary of the Invention

[0006] According to one embodiment, the present application provides a method for detecting SOC anomalies in a battery pack, including: S1: obtaining multiple charging and discharging operation data fragments based on the historical actual operation data of the battery pack; S2: calculating and obtaining a sequence corresponding to the time and actual available capacity value under each charging and discharging operation data fragment based on the multiple charging and discharging operation data fragments; S3: inputting the multiple sequences corresponding to the time and actual available capacity values into a mutation detection algorithm and a cumulative change detection algorithm in chronological order and in a sliding window manner, obtaining the SOC anomaly time point through the mutation detection algorithm, and obtaining the SOC anomaly window fragment through the cumulative change detection algorithm.

[0007] Furthermore, step S1 also includes: S11: the cloud computing platform obtains the historical actual operation data of the battery pack from the battery management system of the battery pack to form a database to be processed; S12: obtaining the charging and discharging operation data of the same battery pack in continuous operation from the database to be processed; S13: dividing the charging and discharging operation data according to the time column of the charging and discharging operation data to form multiple charging and discharging operation data segments.

[0008] Furthermore, the historical actual operation data in step S1 is a database corresponding to the sampling time, the battery pack SOC, and the battery pack throughput.

[0009] Furthermore, the database to be processed in the cloud computing platform in step S11 is a database corresponding to time, battery pack SOC, and battery pack throughput.

[0010] Furthermore, in step S13, the time column of the charge and discharge operation data is divided into multiple charge and discharge operation data segments, which is: identifying the time position in the time column of the charge and discharge operation data where the time interval exceeds the threshold time, and dividing the charge and discharge operation data into multiple charge and discharge operation data segments with the time position as the cutting point.

[0011] Furthermore, step S2 also includes: S21: obtaining each time in the time column of each charging and discharging operation data segment and the SOC and throughput corresponding to the time, calculating the difference in SOC values and the difference in throughput between adjacent times in each charging and discharging operation data segment, and forming a sequence corresponding to the difference ΔSOC between time and SOC and a sequence corresponding to the difference ΔET between time and throughput in each charging and discharging operation data segment; S22: calculating the actual available capacity corresponding to each time in each charging and discharging operation data segment according to the formula Cr = ΔET / ΔSOC, and forming a sequence corresponding to the time and the actual available capacity value in each charging and discharging operation data segment.

[0012] Furthermore, between step S21 and step S22, the following is also included: step S211: finding whether there is a ΔSOC less than a preset value in the sequence corresponding to the difference ΔSOC between time and SOC in each charging and discharging operation data segment; if so, eliminating the ΔSOC corresponding to the time, and simultaneously eliminating the ΔET corresponding to the time in the sequence corresponding to the difference ΔET between time and throughput.

[0013] Furthermore, step S3 also includes: S31: selecting a window length; S32: inputting multiple sequences corresponding to the time and actual available capacity values into the mutation detection algorithm and the cumulative change detection algorithm in chronological order according to the window length; S33: the mutation detection algorithm outputs the SOC abnormal time point, and the cumulative change detection algorithm outputs the window segment of the SOC abnormality.

[0014] Furthermore, step S32 is: dividing the data from the first time to the nth time in the sequence corresponding to the multiple times and the actual available capacity values into a window segment, where n is the window length; dividing the data from the second time to the n+1th time into a window segment; dividing the data from the third time to the n+2th time into a window segment; ...; dividing the data from the mth time to the n+m-1th time into a window segment, where m is a natural number greater than 1, and then inputting it into the mutation detection algorithm and the cumulative change detection algorithm.

[0015] Furthermore, in step S33, the mutation detection algorithm is executed: the maximum and minimum values in each window segment are eliminated, and then the mean and standard deviation of the window segment are calculated, and the three times standard deviation interval of the window segment is obtained based on the mean and standard deviation. It is determined whether the actual available capacity value corresponding to the next time of the window segment is within the three times standard deviation interval. If not, it is considered that the SOC corresponding to the next time of the window segment is abnormal, and the mutation detection algorithm outputs the time point of the SOC abnormality.

[0016] Furthermore, in step S32, the cumulative change detection algorithm is executed: the maximum and minimum values in each window segment are eliminated, and then the average value of the window segment is calculated, and the negative differences between all the actual available capacity values and the average value in the window segment are accumulated to obtain the cumulative deviation sum; it is determined whether the cumulative deviation sum is greater than or equal to the set threshold. If so, it is considered that there is an abnormal SOC change in the window segment, and the cumulative change detection algorithm outputs the window segment with the abnormal SOC change.

[0017] The features and technical advantages of the present disclosure have been summarized quite broadly above so that the detailed description disclosed below may be better understood. Additional features and advantages of the present disclosure will be described below, which form the subject matter of the claims of the present disclosure. It will be understood by those skilled in the art that the concepts and specific embodiments disclosed herein may be readily used as a basis for modifying or designing other structures or processes for achieving the same purposes of the present disclosure. It will also be appreciated by those skilled in the art that such equivalent structures do not depart from the spirit and scope of the present disclosure as set forth in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] For a more complete understanding of the present disclosure and its advantages, reference is now made to the following description in conjunction with the accompanying drawings, in which:

[0019] Figure 1 A schematic flow chart of a method for detecting SOC abnormality of a battery pack according to an embodiment of the present application is shown;

[0020] Figure 2 A schematic diagram of a process for obtaining multiple charging and discharging operation data segments based on historical actual operation data of a battery pack according to an embodiment of the present application is shown;

[0021] Figure 3 A schematic diagram of a process for calculating a sequence corresponding to a time and an actual available capacity value in each charging and discharging operation data segment based on the multiple charging and discharging operation data segments according to an embodiment of the present application is shown;

[0022] Figure 4 A flow chart of an embodiment of the present application is shown, in which a plurality of sequences corresponding to the time and actual available capacity values are input into a mutation detection algorithm and a cumulative change detection algorithm in chronological order and in a sliding window manner, the SOC abnormal time point is obtained by the mutation detection algorithm, and the window segment of the SOC abnormality is obtained by the cumulative change detection algorithm.

[0023] Corresponding numerals and symbols in the different figures generally refer to corresponding parts unless otherwise indicated. These figures are drawn to clearly illustrate the relevant aspects of the various embodiments and are not necessarily drawn to scale. DETAILED DESCRIPTION

[0024] The following is a clear and complete description of the technical solutions in this application in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0025] One embodiment of the present application is to provide a method for detecting abnormal SOC of a battery pack. Figure 1 The flowchart of the SOC abnormality detection method of a battery pack according to one embodiment of the present application is shown. The SOC abnormality detection method of a battery pack proposed in the present application includes:

[0026] S1: Acquire multiple charging and discharging operation data segments based on historical actual operation data of the battery pack;

[0027] S2: Calculating, based on the plurality of charging and discharging operation data segments, a sequence corresponding to a time and an actual available capacity value in each charging and discharging operation data segment;

[0028] S3: Multiple sequences corresponding to the time and actual available capacity values are input into the mutation detection algorithm and the cumulative change detection algorithm in chronological order and in a sliding window manner. The SOC abnormal time point is obtained by the mutation detection algorithm, and the window segment of the SOC abnormality is obtained by the cumulative change detection algorithm.

[0029] See also Figure 2The flowchart shown is a flowchart of obtaining multiple charging and discharging operation data segments based on the historical actual operation data of the battery pack according to an embodiment of the present application. Specifically, step S1 includes: S11: the cloud computing platform obtains the historical actual operation data of the battery pack from the battery management system of the battery pack to form a database to be processed; S12: obtaining the charging and discharging operation data of the same battery pack in continuous operation from the database to be processed; S13: dividing the charging and discharging operation data into multiple charging and discharging operation data segments by dividing the charging and discharging operation data according to the time column.

[0030] Regarding step S11, specifically, during the actual operation of the battery pack, the battery management system of the battery pack will obtain parameters including but not limited to the battery pack voltage, current, temperature, cumulative charge capacity and SOC at a certain sampling interval (such as 10s, but not limited to this, and can be determined according to different application scenarios), that is, the historical actual operation data of the battery pack. Therefore, the historical actual operation data in step S1 is the actual data of the battery pack in the real world, generated during the historical operation process. In one embodiment of the present application, the historical actual operation data of the battery pack is a database corresponding to the sampling time, the battery pack SOC, and the battery pack throughput.

[0031] In actual applications, the cloud computing platform obtains and stores historical actual operation data from the battery management system. Due to factors such as network transmission and cloud computing platform storage, the time intervals in the time column of the to-be-processed database in the cloud computing platform may no longer be the sampling interval of the battery management system. In this way, the to-be-processed database in the cloud computing platform is a database corresponding to time, battery pack SOC, and battery pack throughput.

[0032] Specifically, in actual applications, affected by factors such as network transmission and cloud computing platform storage, the adjacent time in the database to be processed obtained and stored by the cloud computing platform from the battery management system may be relatively long, such as exceeding the threshold time (such as ten minutes). If there is no corresponding data to be processed during such a long period of time, it is impossible to know whether the battery pack is in a charging state, a discharging state, or a state of charging and then discharging, etc. In this way, the throughput of the battery pack may change, but the SOC of the battery pack may not change, which will affect the effect of the battery pack SOC abnormality detection. Therefore, it is necessary to process the fragments with long intervals and not include them in subsequent calculations, that is, it is necessary to process the charging and discharging operation data obtained in step S12.

[0033] Specifically, step S13 involves dividing the charging and discharging operation data into multiple charging and discharging operation data segments based on the time column of the charging and discharging operation data. This involves identifying time locations within the time column of the charging and discharging operation data where the time interval exceeds a threshold time, and using these time locations as cut points to divide the charging and discharging operation data into multiple charging and discharging operation data segments. Thus, the charging and discharging operation data segments are data segments corresponding to the time, battery pack SOC, and battery pack throughput, with the time intervals not exceeding the threshold time. In one embodiment, the threshold time is ten minutes, but this application does not limit this specific value and can be set based on experience.

[0034] In one embodiment of the present application, in step S13, Pandas is used to identify the time positions in the time column of the charge and discharge operation data where the time interval exceeds the threshold time. The charge and discharge operation data is divided into multiple charge and discharge operation data segments using the time position as the cut point. However, this application is not limited to this specific method, and any applicable method can be applied to this application.

[0035] For example, if the charging and discharging operation data in the cloud computing platform (wherein only the time column is displayed, and the battery pack SOC and battery pack throughput corresponding to the time column are not displayed) is as shown in Table 1.

[0036] Table 1

[0037] Time=[

[0038] 2025 / 4 / 8 0:02

[0039] 2025 / 4 / 8 0:07

[0040] 2025 / 4 / 8 0:11

[0041] 2025 / 4 / 8 0:16

[0042] 2025 / 4 / 8 0:20

[0043] 2025 / 4 / 8 0:25

[0044] 2025 / 4 / 8 0:38

[0045] 2025 / 4 / 8 0:47

[0046] 2025 / 4 / 8 0:56

[0047] 2025 / 4 / 8 1:05

[0048] 2025 / 4 / 8 1:10

[0049] 2025 / 4 / 8 1:14

[0050] 2025 / 4 / 8 1:19 ]

[0052] If the threshold time is ten minutes, it can be identified that the time interval between the time 2025 / 4 / 8 0:25 and the time 2025 / 4 / 8 0:38 in the time column is thirteen minutes, which is greater than the threshold time (e.g., ten minutes). Then, using this position as the cutting point, the charge and discharge operation data shown in Table 1 (where only the time column is displayed, and the battery pack SOC and battery pack throughput corresponding to the time column are not displayed) can be divided into two charge and discharge operation data segments (where only the time column is displayed, and the battery pack SOC and battery pack throughput corresponding to the time column are not displayed), which are Table 2 and Table 3, respectively.

[0053] Table 2

[0054] Time=[

[0055] 2025 / 4 / 8 0:02

[0056] 2025 / 4 / 8 0:07

[0057] 2025 / 4 / 8 0:11

[0058] 2025 / 4 / 8 0:16

[0059] 2025 / 4 / 8 0:20

[0060] 2025 / 4 / 8 0:25 ]

[0062] Table 3

[0063] Time=[

[0064] 2025 / 4 / 8 0:38

[0065] 2025 / 4 / 8 0:47

[0066] 2025 / 4 / 8 0:56

[0067] 2025 / 4 / 8 1:05

[0068] 2025 / 4 / 8 1:10

[0069] 2025 / 4 / 8 1:14

[0070] 2025 / 4 / 8 1:19 ]

[0072] In this way, uncertain data fragments are processed and the effect of battery pack SOC abnormality detection is improved.

[0073] See also Figure 3 The flowchart shown in FIG1 is a flowchart of an embodiment of the present application for calculating a sequence corresponding to the time and the actual available capacity value in each charging and discharging operation data segment based on the multiple charging and discharging operation data segments. Specifically, step S2 includes: S21: obtaining each time in the time column of each charging and discharging operation data segment and the SOC and throughput corresponding to the time, calculating the difference between the SOC values and the throughput value of adjacent times in each charging and discharging operation data segment, and forming a sequence corresponding to the time and SOC difference ΔSOC and a sequence corresponding to the time and throughput difference ΔET in each charging and discharging operation data segment; S22: calculating the actual available capacity (realCapacity) corresponding to each time in each charging and discharging operation data segment according to the formula Cr = ΔET / ΔSOC, and forming a sequence corresponding to the time and the actual available capacity value in each charging and discharging operation data segment.

[0074] For step S21, for example, the SOC and throughput ET (Energy Throughput) corresponding to each time in the time column in one of the charging and discharging operation data segments are: SOC = [0, 10, 20, 21, 40, 50, 60, 70, 80, 90, 100] (percentage); ET = [0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0] (unit: kWh), through the first-order difference formula: Δx t =x t -x {t-1} The sequence corresponding to the difference ΔSOC between time and SOC in the charging and discharging operation data segment can be calculated as ΔSOC = [10, 10, 1, 19, 10, 10, 10, 10, 10], and the sequence corresponding to the difference ΔET between time and throughput in the charging and discharging operation data segment can be calculated as ΔET = [0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5] (unit: kWh).

[0075] In step S22 , the actual available capacity corresponding to each time in the charge and discharge operation data segment can be calculated according to the formula Cr = ΔET / ΔSOC, forming a sequence Cr = [5, 5, 50, 2.6, 5, 5, 5, 5, 5] corresponding to the time in the charge and discharge operation data segment and the actual available capacity value.

[0076] At this point, the sequence corresponding to the time and actual available capacity value under each charging and discharging operation data segment is obtained and stored in the cloud computing platform.

[0077] In another embodiment of the present application, step S211 is further included between step S21 and step S22: searching for a ΔSOC smaller than a preset value in a sequence corresponding to a time-SOC difference ΔSOC in each charging and discharging operation data segment; if so, removing the ΔSOC corresponding to the time, and simultaneously removing the ΔET corresponding to the time in a sequence corresponding to a time-throughput difference ΔET.

[0078] In one embodiment, the preset value may be 2 (percentage), but this application does not limit the specific value.

[0079] As described above, in step S22, the actual available capacity corresponding to each time period in each charging and discharging operation data segment is calculated according to the formula Cr = ΔET / ΔSOC. Since ΔSOC serves as the denominator for calculating the actual available capacity, and both ΔET and ΔSOC are affected by precision, when precision is insufficient, the smaller the ΔSOC as the denominator, the greater the impact on the accuracy of the calculated actual available capacity Cr. Furthermore, due to the influence of SOC precision, the cloud computing platform only retains integer digits, so the SOC changes very little when the charging and discharging currents are very low, which can easily affect the calculation results. Therefore, to improve the accuracy of the calculated actual available capacity Cr, it is necessary to eliminate ΔSOC values that are too small and the corresponding ΔET values at the same time.

[0080] Thus, a sequence corresponding to multiple times and actual available capacity values under multiple charging and discharging operation data segments is obtained.

[0081] See also Figure 4 The flowchart of an embodiment of the present application shown is that multiple sequences corresponding to the time and actual available capacity values are input into the mutation detection algorithm and the cumulative change detection algorithm in a sliding window manner in chronological order, the SOC abnormal time point is obtained by the mutation detection algorithm, and the window fragment of the SOC abnormality is obtained by the cumulative change detection algorithm. Specifically, step S3 includes: S31: selecting the window length; S32: inputting multiple sequences corresponding to the time and actual available capacity values into the mutation detection algorithm and the cumulative change detection algorithm in chronological order according to the window length; S33: the mutation detection algorithm outputs the SOC abnormal time point, and the cumulative change detection algorithm outputs the window fragment of the SOC abnormality.

[0082] For step S31, the window length can be set according to actual needs, such as dividing the ten time points in the sequence corresponding to the time and the actual available capacity value and their corresponding actual available capacity values into a window to form data within a window segment.

[0083] For step S32, specifically: divide the data from the first time to the nth time in the sequence corresponding to the multiple times and the actual available capacity values into a window segment, where n is the window length; divide the data from the second time to the n+1th time into a window segment; divide the data from the third time to the n+2th time into a window segment; ...; divide the data from the mth time to the n+m-1th time into a window segment, where m is a natural number greater than 1, and then input it into the mutation detection algorithm and the cumulative change detection algorithm.

[0084] Specifically, for step S33, the mutation detection algorithm executes: eliminating the maximum and minimum values in each window segment, then calculating the average value and standard deviation of the window segment, obtaining the three times standard deviation interval of the window segment based on the average value and standard deviation, and judging whether the actual available capacity value corresponding to the next time of the window segment is within the three times standard deviation interval. If not, it is considered that the SOC corresponding to the next time of the window segment is abnormal. In this way, the SOC corresponding to each time point after the first window segment is detected in turn to determine whether it is abnormal, and then the mutation detection algorithm outputs the time point of the SOC abnormality.

[0085] For example, the actual available capacity value sequence corresponding to a window segment of length ten is [10,9,8,7,8,9,9,9,8,10] (unit: kwh), and the actual available capacity value corresponding to the time point after the window segment is 2 kwh. The mutation detection algorithm removes the maximum value 10 and the minimum value 7 in the window segment, and then calculates that the average value of the remaining data in the window segment is approximately 8.57, and the standard deviation is approximately 0.5. The three times standard deviation interval is: [7.07,10.07]. The actual available capacity value of 2 kwh corresponding to the next time point in the window segment is not within the three times standard deviation interval.

[0086] [7.07,10.07], so the SOC corresponding to the time point after this window segment is considered abnormal, and the abnormal time point is output.

[0087] Specifically, for step S32, the cumulative change detection algorithm executes: eliminating the maximum and minimum values in each window segment, then calculating the average value of the window segment, accumulating the negative differences between all actual available capacity values and the average value in the window segment, and obtaining the cumulative deviation sum; judging whether the cumulative deviation sum is greater than or equal to the set threshold value, if so, it is considered that there is an abnormal SOC change in the window segment, and in this way, detecting whether there is an abnormal SOC change in each divided window segment in turn, and then the cumulative change detection algorithm outputs the window segment with the abnormal SOC change.

[0088] Furthermore, the cumulative change detection algorithm also outputs the time point corresponding to the actual available capacity value less than the average value in each window segment as the SOC abnormal time point.

[0089] For example, a window segment of length ten corresponds to the sequence of actual available capacity values of [10, 9, 8, 7, 6, 5, 4, 3, 2, 2] (unit: kwh), and the threshold is set to 8 kwh. The cumulative change detection algorithm removes the maximum value of 10 and the minimum value of 2 within the window segment, and then calculates the average value of the remaining data in the window segment to be approximately 5.5. The negative differences between all the actual available capacity values in the window segment and the average value of 5.5 are then accumulated, resulting in a cumulative deviation sum of 11.5. At this time, the cumulative deviation sum of 11.5 is greater than the set threshold of 8, indicating that an abnormal SOC change has occurred within the window segment, and the window segment is output. Furthermore, the actual available capacity value within the window segment is determined to be less than the average value in chronological order. The time point corresponding to the actual available capacity value less than the average value within the window segment is used as the SOC abnormality start time point. Once an actual available capacity value greater than or equal to the average value appears, the time point corresponding to the actual available capacity value greater than or equal to the average value is used as the SOC abnormality end time point. If an actual available capacity value greater than or equal to the average value still does not appear by the end of the window segment, the time point at the end of the window segment is used as the SOC abnormality end time point, and the time from the SOC abnormality start time to the SOC abnormality end time point is output as the SOC abnormality time segment. As described above, the time point corresponding to the actual available capacity value of 5 is used as the SOC abnormality start time. The actual available capacity values corresponding to the time points after the time corresponding to the actual available capacity value of 5 until the end of the window segment are all less than the average value of 5. Therefore, the time point at the end of the window segment is used as the SOC abnormality end time point. In other words, the segment formed by the time point corresponding to the actual available capacity value of 5 to the time point at the end of the window segment is output as the SOC abnormality time segment.

[0090] The above-mentioned threshold value can be set manually based on experience, and can be dynamically bound to the actual available capacity as described above. According to the accuracy of the SOC abnormality detection method of the above-mentioned battery pack, the jump judgment setting threshold value under different battery cells and different capacity nominals is corrected, which solves the misjudgment problem caused by the high coupling between ΔSOC and rated capacity, improves the accuracy of SOC abnormality detection, and has strong applicability.

[0091] In this way, the SOC anomaly detection method for the battery pack provided by this application can locate SOC abnormal points and abnormal fragments. It can be seen that this application does not judge whether it is abnormal based on the SOC accuracy, but dynamically binds ΔSOC to the actual available capacity. This is because the actual available capacity is dynamically changing. Affected by the actual application scenario, type, and pressure difference of the battery pack, the SOC is also changing, and the abnormal change of SOC will be reflected in the calculated actual available capacity. Then, when the actual available capacity fluctuates abnormally, it can be determined that the SOC here has changed abnormally. Therefore, it can be determined whether there is an SOC abnormality based on the actual available capacity. It can be seen that this application makes anomaly detection more accurate by processing the SOC calculated by any algorithm and judging whether the SOC accurately follows the dynamic changes of the actual available capacity.

[0092] In actual applications, the output results of the mutation detection algorithm and the cumulative change detection algorithm are saved to the cloud computing platform to provide data support for subsequent monitoring, alarming, and in-depth analysis.

[0093] Furthermore, the battery pack SOC anomaly detection method described in this application is adaptable to different battery product applications, including micro-storage, household energy storage, industrial and commercial energy storage, and power batteries. Regardless of the type and specifications of the battery cells, it can be seamlessly deployed without configuring additional parameters. Furthermore, it is applicable to various charging and discharging modes in battery packs, such as shallow charge and shallow discharge, deep charge and deep discharge, and other complex operating conditions, thus providing strong adaptability.

[0094] In practical applications, the battery pack SOC anomaly detection method provided in this application is executed in real time or offline by a cloud computing platform, which sends the anomaly analysis results. The cloud computing platform has relatively high memory and computing efficiency, making it more suitable for model training of large amounts of data. In practical applications, the battery management system collects and obtains the battery pack's operating data during the operation of the battery pack. Therefore, the cloud computing platform obtains the above-mentioned historical actual operating data from the battery management system of the battery pack.

[0095] Although the embodiments of the present disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the disclosure as defined by the appended claims.

[0096] Furthermore, the scope of the present application is not intended to be limited to the particular embodiments of the processes, machines, manufacture, compositions of matter, means, methods, and steps described in the specification. As will be readily understood by one of ordinary skill in the art from the disclosure herein, processes, machines, manufacture, compositions of matter, means, methods, or steps that perform substantially the same functions now exist or will later be developed or that achieve substantially the same results as the corresponding embodiments described herein may be utilized in accordance with the present disclosure. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

Claims

1. A method for detecting SOC abnormality of a battery pack, characterized in that: include: S1: Acquire multiple charging and discharging operation data segments based on historical actual operation data of the battery pack; S2: Calculating, based on the plurality of charging and discharging operation data segments, a sequence corresponding to a time and an actual available capacity value in each charging and discharging operation data segment; S3: Multiple sequences corresponding to the time and actual available capacity values are input into the mutation detection algorithm and the cumulative change detection algorithm in chronological order and in a sliding window manner. The SOC abnormal time point is obtained by the mutation detection algorithm, and the window segment of the SOC abnormality is obtained by the cumulative change detection algorithm.

2. The SOC abnormality detection method of a battery pack according to claim 1, characterized in that: Step S1 further includes: S11: The cloud computing platform obtains historical actual operation data of the battery pack from the battery management system of the battery pack to form a database to be processed; S12: Acquire the charging and discharging operation data of the same battery pack in continuous operation from the database to be processed; S13: dividing the charging and discharging operation data into a plurality of charging and discharging operation data segments based on the time column of the charging and discharging operation data.

3. The method for detecting SOC abnormality of a battery pack according to claim 2, wherein: The historical actual operation data in step S1 is a database corresponding to the sampling time, the battery pack SOC, and the battery pack throughput.

4. The method for detecting SOC abnormality of a battery pack according to claim 2, wherein: The database to be processed in the cloud computing platform in step S11 is a database corresponding to time, battery pack SOC, and battery pack throughput.

5. The method for detecting SOC abnormality of a battery pack according to claim 2, wherein: In step S13, the charging and discharging operation data is divided into a plurality of charging and discharging operation data segments based on the time column of the charging and discharging operation data, which are: Identify the time position in the time column of the charge and discharge operation data where the time interval exceeds the threshold time, and use the time position as a cutting point to divide the charge and discharge operation data into multiple charge and discharge operation data segments.

6. The method for detecting SOC abnormality of a battery pack according to claim 2, wherein: Step S2 further includes: S21: Obtain each time in the time column of each charging and discharging operation data segment and the SOC and throughput corresponding to the time, calculate the difference between the SOC values and the throughput values of adjacent times in each charging and discharging operation data segment, and form a sequence corresponding to the difference between time and SOC ΔSOC and a sequence corresponding to the difference between time and throughput ΔET in each charging and discharging operation data segment; S22: Calculate the actual available capacity corresponding to each time in each charging and discharging operation data segment according to the formula Cr=ΔET / ΔSOC, and form a sequence corresponding to the time and the actual available capacity value in each charging and discharging operation data segment.

7. The method for detecting SOC abnormality of a battery pack according to claim 6, wherein: Also included between step S21 and step S22: Step S211: Find whether there is a ΔSOC less than a preset value in the sequence corresponding to the difference ΔSOC between time and SOC in each charging and discharging operation data segment. If so, remove the ΔSOC corresponding to the time, and simultaneously remove the ΔET corresponding to the time in the sequence corresponding to the difference ΔET between time and throughput.

8. The method for detecting SOC abnormality of a battery pack according to claim 6, wherein: Step S3 further includes: S31: Select window length; S32: Inputting the plurality of sequences corresponding to the time and the actual available capacity values into the mutation detection algorithm and the cumulative change detection algorithm in chronological order according to the window length; S33: The mutation detection algorithm outputs the SOC abnormal time point, and the cumulative change detection algorithm outputs the SOC abnormal window segment.

9. The method for detecting SOC abnormality of a battery pack according to claim 8, wherein: Step S32 is: divide the data from the first time to the nth time in the sequence corresponding to the multiple times and the actual available capacity values into a window segment, where n is the window length; divide the data from the second time to the n+1th time into a window segment; divide the data from the third time to the n+2th time into a window segment; ...; divide the data from the mth time to the n+m-1th time into a window segment, where m is a natural number greater than 1, and then input it into the mutation detection algorithm and the cumulative change detection algorithm.

10. The method for detecting SOC abnormality of a battery pack according to claim 8, wherein: In step S33, the mutation detection algorithm executes: Eliminate the maximum and minimum values within each window segment, then calculate the mean and standard deviation of the window segment, and obtain the three times standard deviation interval of the window segment based on the mean and standard deviation. Determine whether the actual available capacity value corresponding to the next time in the window segment is within the three times standard deviation interval. If not, it is considered that the SOC corresponding to the next time in the window segment is abnormal, and the mutation detection algorithm outputs the time point of the SOC abnormality.

11. The method for detecting SOC abnormality of a battery pack according to claim 8, wherein: In step S32, the cumulative change detection algorithm is executed: The maximum and minimum values within each window segment are eliminated, and then the average value of the window segment is calculated. The negative differences between all actual available capacity values and the average value within the window segment are accumulated to obtain a cumulative deviation sum. It is determined whether the cumulative deviation sum is greater than or equal to a set threshold. If so, it is considered that an abnormal SOC change exists within the window segment, and the cumulative change detection algorithm outputs the window segment with the abnormal SOC change.

Images