Method and device for detecting energy storage equipment and electronic equipment

Abstract

The invention provides an energy storage equipment detection method and device, and electronic equipment, and relates to the technical field of power system energy storage. The method comprises the following steps: acquiring charging and discharging data of each single battery in the energy storage equipment; constructing a multi-dimensional feature vector based on the charging and discharging data of each single battery; processing the multi-dimensional feature vector to obtain a local outlier factor value of each single battery; and determining whether each single battery is abnormal or not based on the local outlier factor value of each single battery. The method solves the technical problems that an existing energy storage equipment abnormal behavior detection method is insufficient in mass data processing capacity and limited in detection algorithm efficiency.

Description

Technical Field

[0001] This disclosure relates to the field of power system energy storage technology, and more specifically, to a testing method and apparatus for energy storage devices, and electronic equipment. Background Technology

[0002] With the large-scale grid connection of renewable energy, energy storage power stations, as important regulating units of the power system, directly impact the safe operation of the power grid through effective monitoring of their charging and discharging behavior. Current technologies for detecting abnormal behavior of energy storage devices mainly face the following technical bottlenecks: 1. Insufficient capacity for processing massive amounts of data; 2. The efficiency of anomaly detection algorithms is limited. Summary of the Invention

[0003] This disclosure provides a detection method, apparatus, and electronic device for energy storage devices, which addresses the technical problems of insufficient massive data processing capabilities and limited detection algorithm efficiency in existing methods for detecting abnormal behavior of energy storage devices.

[0004] According to one aspect of the present disclosure, a detection method for an energy storage device is provided. The method includes: acquiring charge and discharge data of each individual battery cell in the energy storage device; constructing a multidimensional feature vector based on the charge and discharge data of each individual battery cell; processing the multidimensional feature vector to obtain the local outlier factor value of each individual battery cell; and determining whether each individual battery cell has an anomaly based on the local outlier factor value of each individual battery cell.

[0005] Optionally, a multi-dimensional feature vector is constructed based on the charge and discharge data of each individual battery cell, including: cleaning the charge and discharge data of each individual battery cell; dividing the cleaned charge and discharge data into charging state data, discharging state data, and resting state data; constructing a two-dimensional feature vector based on the voltage data in the charging state data; the two-dimensional feature vector represents the voltage of different individual batteries at the same time and the voltage of the same individual battery at different times in two dimensions.

[0006] Optionally, the multidimensional feature vector is processed to obtain the local outlier value of each individual cell, including: determining the average value of any row feature vector or any column feature vector representing the voltage of the same individual cell at different times in the multidimensional feature vector; determining the average value as point P; determining the local outlier value of point P; and determining the local outlier value of point P as the local outlier value of the individual cell.

[0007] Optionally, determining the local outlier factor value of point P includes: determining the k-nearest neighbor distance of point P; the k-nearest neighbor distance is the distance from point P to its k-th nearest neighbor, where k is a positive integer; determining the k-distance neighborhood of point P based on the k-nearest neighbor distance; the k-distance neighborhood is a circular region centered on point P and with the k-nearest neighbor distance of point P as the radius; determining the reachability distance of point P relative to each point within the k-distance neighborhood of point P based on the k-nearest neighbor distance of point P; determining the local reachability density of point P based on the reachability distance of point P relative to each point within the k-distance neighborhood of point P; and determining the local outlier factor value of point P based on the local reachability density of point P.

[0008] Optionally, the reachable distance of point P relative to each point within the k-distance range of point P is determined based on the k-nearest neighbor distance of point P, including: determining the first distance between point P and each point within the k-distance range of point P; determining the maximum distance between the k-nearest neighbor distance of point P and the first distance corresponding to each point; and determining the maximum value as the reachable distance of point P relative to each point within the k-distance range of point P.

[0009] Optionally, the local reachability density of point P is determined based on the reachability distance of point P relative to each point within the k-distance range of point P, including: determining the average reachability distance of point P relative to each point within the k-distance range of point P; and determining the reciprocal of the average reachability distance as the local reachability density of point P.

[0010] Optionally, determining the local outlier factor value of point P based on the local reachability density of point P includes: determining the local reachability density of each point within the k-distance neighborhood of point P; determining the sum of the ratios of the local reachability density of each point within the k-distance neighborhood of point P to the local reachability density of point P; and determining the ratio of the sum to the number of points within the k-distance neighborhood of point P as the local outlier factor value of point P.

[0011] Optionally, determining whether each individual cell is abnormal based on its local outlier factor value includes: comparing the local outlier factor value of each individual cell with the local outlier factor values ​​of other individual cells in the energy storage device; and determining that an individual cell is abnormal if the difference between the local outlier factor value of an individual cell and the local outlier factor value of other individual cells exceeds a preset threshold.

[0012] According to another aspect of the present disclosure, a detection device for an energy storage device is provided. The device includes: an acquisition module for acquiring charge and discharge data of each individual battery cell in the energy storage device; a construction module for constructing a multi-dimensional feature vector based on the charge and discharge data of each individual battery cell; a processing module for processing the multi-dimensional feature vector to obtain the local outlier factor value of each individual battery cell; and a determination module for determining whether each individual battery cell has an anomaly based on the local outlier factor value of each individual battery cell.

[0013] According to another aspect of the present disclosure, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the above-described method.

[0014] According to another aspect of the present disclosure, a computer-readable storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described method.

[0015] The beneficial effects of the technical solution provided in this disclosure are as follows: by dividing battery data into charge and discharge states, the voltage information of the battery in the time and space dimensions is obtained, and combined with anomaly detection algorithms, abnormal batteries can be accurately located at the end of each charge and discharge cycle. This improves the accuracy and efficiency of anomaly detection in energy storage devices and enables rapid loading and efficient processing of large amounts of data from energy storage devices. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments of this disclosure will be briefly introduced below.

[0017] Figure 1 This is a flowchart of a detection method for an energy storage device provided in this embodiment of the disclosure; Figure 2 This is a schematic diagram illustrating the calculation of the k-nearest neighbor distance in an LOF algorithm provided in this embodiment of the disclosure; Figure 3 A heat map of the LOF value of a single battery cell in an energy storage power station is provided as an embodiment of this disclosure; Figure 4 This is a flowchart of another detection method for an energy storage device provided in the embodiments of this disclosure; Figure 5 This is a schematic diagram of the structure of a detection device for an energy storage device provided in an embodiment of the present disclosure; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0018] The embodiments of this disclosure are described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of this disclosure, and do not constitute a limitation on the technical solutions of the embodiments of this disclosure.

[0019] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the terms “comprising” and “including” as used in embodiments of this disclosure mean that the corresponding feature can be implemented as the presented feature, information, data, step, operation, element, and / or component, but do not exclude implementation as other features, information, data, step, operation, element, component, and / or combinations thereof supported by the art. It should be understood that when we say that an element is “connected” or “coupled” to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element are connected through an intermediate element. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The term “and / or” as used herein indicates at least one of the items defined by the term, for example, “A and / or B” or “A, B” indicates implementation as “A,” or implementation as “B,” or implementation as “A and B.”

[0020] To make the objectives, technical solutions, and advantages of this disclosure clearer, the embodiments of this disclosure will be described in further detail below with reference to the accompanying drawings.

[0021] First, the technical terms used in this disclosure will be introduced and explained: Individual batteries are the basic units that make up the entire energy storage system. Individual batteries are connected in series, parallel or a combination of both to form modules, clusters and finally battery containers to meet different voltage and capacity requirements.

[0022] SOC, short for State of Charge, refers to the ratio of a battery's current stored electrical energy to its maximum electrical energy, usually expressed as a percentage. For example, if a battery has a maximum capacity of 100 kilowatt-hours (kWh) and currently stores 50 kWh of energy, then its SOC is 50%.

[0023] LOF, short for Local Outlier Factor, is an algorithm used to identify outliers in a dataset. It determines whether a point is an outlier by comparing the density of a point with that of its neighbors.

[0024] In existing technologies, the detection of abnormal behavior in energy storage devices mainly faces the following technical bottlenecks: 1. Insufficient capacity for processing massive amounts of data: A single energy storage power station can generate TB (1TB equals 1024GB) of operational data (including multi-dimensional parameters such as voltage, current, SOC status, and temperature) every day. Traditional single-machine algorithms are difficult to perform real-time analysis, resulting in problems such as high computational latency and the risk of memory overflow.

[0025] 2. Limited efficiency of anomaly detection algorithms: Existing LOF-based outlier detection methods employ single-machine serial computation, which leads to high time complexity when dealing with high-dimensional, multi-time-series energy storage data. The scale is such that when energy storage power stations charge and discharge in large quantities, the detection time increases exponentially, which cannot meet the operation and maintenance requirements of minute-level response.

[0026] The testing methods, apparatus, and electronic equipment for energy storage devices disclosed herein are intended to solve at least one of the above-mentioned technical problems in the prior art.

[0027] The technical solution provided in this disclosure achieves accurate and efficient anomaly identification through the synergy of three technologies: 1. Distributed computing framework: Relying on Spark's in-memory computing, batch-stream integrated architecture and dynamic resource scheduling mechanism, it realizes parallel feature extraction of battery data from multiple power stations and multiple compartments, breaking through the bottlenecks of traditional single-machine systems in terms of horizontal scalability and real-time processing efficiency; 2. Fast and accurate anomaly detection: Construct a spatiotemporal two-dimensional feature vector (for example, data from different individual cells at the same time are horizontal features, and historical time-series data of the same individual cell are vertical features), input it into the LOF algorithm model to calculate density deviation, and accurately locate abnormal individual cells (such as voltage drop, high charge and low discharge, etc.) through dual-dimensional density comparison. 3. High-performance data storage: The Doris database is used to achieve efficient end-to-end data management. Its Massively Parallel Processing (MPP) architecture and columnar storage engine support millisecond-level response for TB-level charging and discharging data. The spatiotemporal partition index (collection timestamp and cluster ID) accelerates the query, and the Spark-DorisConnector enables real-time storage of feature data and dynamic visualization of detection results, forming a closed loop of "data computation, storage and feedback".

[0028] The technical solution proposed in this disclosure combines big data distributed computing technology with unsupervised machine learning algorithms to achieve real-time processing, feature extraction, and battery anomaly analysis of operational data from large-scale energy storage systems. This method is applicable to the status monitoring and anomaly detection of energy storage equipment in scenarios such as new energy power plants.

[0029] The following description of several exemplary embodiments illustrates the technical solutions of this disclosure and the technical effects produced by these solutions. It should be noted that the following embodiments can be referenced, learned from, or combined with each other. Identical terms, similar features, and similar implementation steps in different embodiments will not be repeated.

[0030] This disclosure provides a method for detecting energy storage devices, such as... Figure 1 As shown, the method includes: Step S101: Obtain the charging and discharging data of each individual battery cell in the energy storage device.

[0031] According to some optional embodiments of this disclosure, the above-mentioned energy storage devices include, but are not limited to, energy storage power stations.

[0032] The above charging and discharging data includes, but is not limited to, voltage data and current data.

[0033] Step S102: Construct a multi-dimensional feature vector based on the charge and discharge data of each individual battery cell.

[0034] This step includes two parts: data cleaning and feature extraction. The data cleaning part determines whether the data needs cleaning based on the data quality of the energy storage power station. The feature extraction part relies on the Spark distributed computing engine to quickly parse the battery data during the charging and discharging process of individual batteries in the energy storage power station and classify the charging and discharging states based on parameters such as current direction. Then, two-dimensional feature vectors are constructed horizontally and vertically.

[0035] Step S103: Process the multidimensional feature vector to obtain the local outlier factor value of each individual cell.

[0036] Step S104: Determine whether there is an anomaly in each individual cell based on the local outlier factor value of each individual cell.

[0037] In steps S103-S104, the two-dimensional feature vector is input into the LOF algorithm model for processing. LOF is an unsupervised anomaly detection algorithm that determines whether a data frame is an anomaly by calculating its relative density compared to its neighboring data frames. Specifically, density refers to the number of data points within a certain radius, while relative density is the relative value obtained by comparing the density of a point with the density of its neighbors. If the relative density of a point is smaller than the density of its neighbors, it indicates that the point is located in a low-density region and may be an anomaly.

[0038] Compared to other anomaly detection algorithms, the LOF algorithm considers local data, analyzing the data within the neighborhood of an object point. This method can more accurately detect anomalies when multiple regions of varying densities exist in the dataset.

[0039] The proposed solution, by dividing battery data into charge and discharge states, obtains voltage information of the battery in both time and space dimensions. Combined with anomaly detection algorithms, it can accurately locate abnormal batteries at the end of each charge and discharge cycle. This improves the accuracy and efficiency of anomaly detection in energy storage devices and enables rapid loading and efficient processing of large volumes of data from energy storage devices.

[0040] This disclosure provides a possible implementation method, in which step S102 constructs a multi-dimensional feature vector based on the charge and discharge data of each individual battery cell, including the following steps: Step S1021: Perform data cleaning on the charge and discharge data of each individual battery cell.

[0041] In this step, abnormal battery data is removed, such as abnormal current, abnormal cell voltage array, abnormal SOC, etc.

[0042] Step S1022: Divide the cleaned charge and discharge data into charging state data, discharging state data, and static state data.

[0043] In this step, the charge and discharge data after cleaning are divided into charging state data, discharging state data, and static state data according to the direction of current and the direction of SOC change.

[0044] In embodiments of this disclosure, anomaly detection is performed on individual battery cells based on state-of-charge data.

[0045] Step S1023: Construct a two-dimensional feature vector based on the voltage data in the state of charge data; the two-dimensional feature vector represents the voltage of different individual cells at the same time and the voltage of the same individual cell at different times in two dimensions.

[0046] In embodiments of this disclosure, voltage data for each individual cell is recorded at every moment from the start of charging, while filtering out voltage data from abnormal individual cells during the recording process, such as invalid voltage values ​​(type overflow), zero voltage, etc. Then, two-dimensional feature vectors, horizontal and vertical, are constructed based on the recorded voltage data for each individual cell at each moment.

[0047] In an optional embodiment, the horizontal dimension of the above-mentioned two-dimensional feature vector focuses on the difference between the average voltage of different individual cells at the same time, while the vertical dimension focuses on the voltage change of the same individual cell at different times. The two-dimensional feature vector is shown in the following formula (1):

[0048] In another alternative embodiment, the horizontal dimension of the above-mentioned two-dimensional feature vector focuses on the voltage changes of the same single cell at different times, while the vertical dimension focuses on the differences between the voltage of different single cells and the average voltage at the same time. The two-dimensional feature vector is shown in the following formula (2):

[0049] The technical solution proposed in this disclosure constructs a spatiotemporal two-dimensional feature vector, and then inputs the two-dimensional feature vector into the LOF algorithm for density deviation calculation. Through dual-dimensional density comparison, the accuracy and efficiency of abnormal single cell detection can be improved.

[0050] According to some optional embodiments of this disclosure, step S103 processes the multidimensional feature vector to obtain the local outlier factor value of each individual cell, including the following steps: Step S1031: Determine the average value of any row feature vector or any column feature vector representing the voltage of the same single cell at different times in the multidimensional feature vector.

[0051] Step S1032: Determine the average value as point P.

[0052] As mentioned above, the two-dimensional feature vector is input into the LOF algorithm model for processing. The first step is to calculate the k-nearest neighbor distance. Figure 2 This is a schematic diagram of calculating the k-nearest neighbor distance in an LOF algorithm provided in this disclosure embodiment. The k-nearest neighbor distance of point P refers to the distance from point P to its k-th nearest neighbor.

[0053] Before calculating the k-nearest neighbor distance, point P needs to be determined first. In the embodiments of this disclosure, the average value of any row feature vector or any column feature vector representing the voltage of the same single cell at different times in the multidimensional feature vector is taken as point P.

[0054] If the two-dimensional feature vector is as shown in formula (1) above, in this step, the average value of any column feature vector in formula (1) is taken as point P.

[0055] If the two-dimensional feature vector is as shown in formula (2) above, in this step, the average value of any row feature vector in formula (2) is taken as point P.

[0056] Step S1033: Determine the local outlier factor value at point P.

[0057] According to some optional embodiments of this disclosure, step S1033, which determines the local outlier factor value at point P, includes the following steps: Step S10331: Determine the k-nearest neighbor distance of point P; the k-nearest neighbor distance is the distance from point P to its k-th nearest neighbor, where k is a positive integer.

[0058] In the dataset, the distance between point P and point O is denoted as... For any positive integer k, the k-nearest neighbor distance of point P is... .

[0059] It can be considered that the following conditions are met. equal There must be at least k points. ,satisfy There are at most k-1 ,satisfy That is, each point The distances to point P are sorted from largest to smallest, and the k-th distance is the distance numbered k. Figure 2 The figure shows the nearest neighbor distance of point P when k=7.

[0060] Step S10332: Determine the k-distance neighborhood of point P based on the k-nearest neighbor distance; the k-distance neighborhood is a circular region with point P as the center and the k-nearest neighbor distance of point P as the radius.

[0061] The region formed by a circle with point P as its center and the k-nearest neighbor distance of point P as its radius is called the k-distance neighborhood of point P. All points within this neighborhood are k-nearest neighbors of point P.

[0062] In this step, the set of points q and P whose distance does not exceed the k nearest neighbors of point P is defined as follows, as shown in formula (3):

[0063] In formula (3), point q represents all points within the k-distance neighborhood of point P.

[0064] Step S10333: Determine the reachable distance of each point in the k-distance range of point P relative to point P based on the k-nearest neighbor distance of point P.

[0065] The reachability distance is the maximum value between the actual distance from one point to another and the distance to that point's k-th nearest neighbor.

[0066] According to an optional embodiment of this disclosure, step S10333 determines the reachable distance of point P relative to each point within the k-distance range of point P based on the k-nearest neighbor distance of point P, including the following steps: determining the first distance between point P and each point within the k-distance range of point P; determining the maximum distance between the k-nearest neighbor distance of point P and the first distance corresponding to each point; and determining the maximum value as the reachable distance of point P relative to each point within the k-distance range of point P.

[0067] The reachable distance of point P relative to point O (any point within the k-distance neighborhood of point P) is shown in formula (4):

[0068] in, It is the reachable distance of point P relative to point O. It is the k-nearest neighbor distance of point P; It is the distance between point P and point O (i.e., the first distance mentioned above).

[0069] Step S10334: Determine the local reachability density of point P based on the reachability distance of each point in the k-distance neighborhood of point P relative to point P.

[0070] According to another optional embodiment of this disclosure, step S10334 determines the local reachability density of point P based on the reachability distance of each point in the k-distance range of point P relative to point P, including the following steps: determining the average reachability distance of point P relative to each point in the k-distance range of point P; and determining the reciprocal of the average reachability distance as the local reachability density of point P.

[0071] Assuming k is a positive integer, the local reachability density of point P is the average reachability distance between point P and every point in its k-neighborhood, and then taking the reciprocal, as shown in formula (5):

[0072] in, It is the locally accessible density at point P; It is the reachable distance of point P relative to point O; It is the number of all points other than P within the k-distance neighborhood of point P.

[0073] Step S10335: Determine the local outlier factor value of point P based on the local reachability density of point P.

[0074] According to some optional embodiments of this disclosure, the execution step S10335, which determines the local outlier factor value of point P based on the local reachability density of point P, includes the following steps: determining the local reachability density of each point within the k-distance neighborhood of point P; determining the sum of the ratios of the local reachability density of each point within the k-distance neighborhood of point P to the local reachability density of point P; and determining the ratio of the sum to the number of points within the k-distance neighborhood of point P as the local outlier factor value of point P.

[0075] In this step, the LOF value of point P is calculated. The LOF value is calculated as shown in formula (6):

[0076] in, It is the locally reachable density at point O. It is the locally reachable density at point P; It is the number of all points other than P within the k-distance neighborhood of point P.

[0077] Step S1034: Determine the local outlier value at point P as the local outlier value of the single cell.

[0078] The LOF value of P-cell obtained by the above steps is the LOF value of a single cell.

[0079] This disclosure provides an optional implementation method in which step S104 determines whether each individual battery has an anomaly based on the local outlier factor value of each individual battery. This is achieved by the following method: comparing the local outlier factor value of each individual battery with the local outlier factor values ​​of other individual batteries in the energy storage device; and determining that an individual battery has an anomaly if the difference between the local outlier factor value of an individual battery and the local outlier factor value of other individual batteries exceeds a preset threshold.

[0080] The data of each individual battery in the energy storage power station is processed using the algorithm in steps S1031-S1034, and the LOF value of each individual battery can be obtained. The larger the LOF value of each individual battery compared to the LOF values ​​of other individual batteries, the further its LOF value is from the group, and the more abnormal the individual battery is. Figure 3 The maximum value shown in the heatmap is the abnormal single cell.

[0081] In some optional embodiments of this disclosure, the calculated local outlier values ​​of each battery cell are stored in the Doris database.

[0082] The Doris database is used to achieve efficient end-to-end data management. Its MPP architecture and columnar storage engine support millisecond-level response of TB-level charging and discharging data. The query is accelerated by spatiotemporal partition index (collection timestamp and cluster ID). Combined with Spark-Doris Connector, feature data is put into the database in real time and detection results are dynamically visualized, forming a closed loop of "data computation, storage and feedback".

[0083] In some optional embodiments of this disclosure, reporting tools are used to visualize the detection indicators, such as... Figure 3 As shown in the image. Furthermore, the visualization results can be used to detect anomalies in energy storage power stations and provide feedback.

[0084] In the embodiments disclosed herein, the system deeply applies the Spark distributed computing framework throughout the entire lifecycle of data loading, computation, and storage, supporting the parallel acquisition of batch data from multiple power plants, significantly improving data processing efficiency, and laying a solid data foundation for subsequent feature extraction and anomaly detection.

[0085] Spark is an open-source unified analytics engine designed for large-scale data processing. It provides an efficient, general-purpose, and easy-to-use programming model, supporting a variety of workloads including batch processing, stream processing, machine learning, and graph computation. Spark's core strengths lie in its in-memory computing capabilities and high-level abstractions, enabling it to outperform traditional disk-based computing frameworks (such as Hadoop MapReduce).

[0086] In the embodiments of this disclosure, at the feature engineering level, the system innovatively designs a two-dimensional battery voltage feature vector. For example, the horizontal dimension quantifies the dispersion of individual cells by analyzing the difference between the voltage of different individual cells and the average voltage at the same time; the vertical dimension characterizes the dynamic change trend of individual cell voltage based on the voltage difference of the same individual cell at different times. In the algorithm calculation stage, this two-dimensional feature is combined with the LOF anomaly detection algorithm to achieve real-time and accurate identification of abnormal individual cells in the energy storage power station.

[0087] In the embodiments of this disclosure, a hybrid storage strategy is adopted at the data storage layer, storing real-time detection results in the Doris storage engine to ensure the efficiency of data query and analysis. Simultaneously, an automatic archiving channel is established with the Hadoop Distributed File System (HDFS) to complete cold data backup, meeting the needs of long-term data backtracking while reducing storage space usage.

[0088] Figure 4 This is a flowchart of another detection method for an energy storage device provided in this disclosure embodiment, such as... Figure 4 As shown, this method mainly includes five steps: data loading, feature extraction, anomaly identification, data storage, and indicator display. During the data loading phase, the system completes the import of charging and discharging data from the energy storage power station.

[0089] The feature extraction stage includes data cleaning and feature extraction. The data cleaning part determines whether the data needs cleaning based on the data quality of the energy storage power station. The feature extraction part relies on the Spark distributed computing engine to quickly parse the battery data of the charging and discharging process of individual batteries in the energy storage power station and classify the charging and discharging states based on parameters such as current direction. Then, a two-dimensional feature vector is constructed.

[0090] In the anomaly detection phase, a two-dimensional feature vector is input into the LOF algorithm model. LOF is an unsupervised anomaly detection method that determines whether a data frame is an anomalous by calculating its relative density compared to its neighboring data frames. Specifically, density refers to the number of data points within a certain radius, while relative density is the relative value obtained by comparing the density of a point with the density of its neighbors. If the relative density of a point is lower than that of its neighbors, it indicates that the point is located in a low-density region and may be an anomalous point. Compared to other anomaly detection algorithms, the LOF algorithm considers local data, analyzing the data within the neighborhood of the object point. This method can more accurately detect anomalous points when multiple regions of different densities exist in the dataset.

[0091] During the data storage phase, the calculated local outlier factor values ​​of each individual cell are stored in the Doris database.

[0092] During the indicator display phase, reporting tools are used to visualize the detection indicators, such as... Figure 3 As shown. Furthermore, the visualization results can be used to detect abnormal situations in the power plant and provide feedback.

[0093] Before executing the above steps, firstly, you need to set up the JDK, Spark, and Doris environment on the server, configuring the database address according to the set port and host IP. Then, develop a Spark program. The Spark program will read files from the specified directory according to the configuration and load and process the data from the energy storage power station according to the corresponding code logic. Finally, store the processing results in Doris and write corresponding SQL statements to display the calculated data. In addition, users can evaluate the indicators through the feedback system and update the indicator database. This completes the entire electricity consumption behavior analysis process.

[0094] By leveraging the technical solutions proposed in this disclosure and utilizing technologies such as Spark, Doris, and SQL, complex data becomes easier to read and clearer, enabling easy access to crucial information such as real-time anomalies and historical data related to the charging and discharging of energy storage power stations. This not only enhances the user experience but also provides strong data support for relevant decision-making.

[0095] The advantages of the technical solution proposed in this disclosure compared to the prior art are as follows: 1. Enabled rapid loading and efficient processing of large volumes of data from multiple energy storage power stations. Leveraging the powerful Spark computing framework, it efficiently processes individual battery data, effectively ensuring the rapid loading of large volumes of data.

[0096] 2. It can accurately and quickly detect abnormal batteries. By segmenting battery data according to charge and discharge states based on current, it obtains the battery's voltage information in both time and space dimensions. Combined with the LOF anomaly detection algorithm, it can accurately locate abnormal batteries at the end of each charge and discharge cycle.

[0097] 3. It can intuitively display abnormal indicators. The calculation results are persistently stored in the Doris database. Utilizing its efficient query performance, abnormal individual indicators can be displayed in a timely manner, enabling users to grasp the battery status of the power station in a timely and easy manner, and providing strong support for the long-term and safe operation of the power station.

[0098] This disclosure provides a detection device for energy storage equipment, such as... Figure 5 As shown, the detection device 50 of the energy storage device may include: an acquisition module 501, a construction module 502, a processing module 503, and a determination module 504, wherein, The acquisition module 501 is used to acquire the charging and discharging data of each individual battery in the energy storage device.

[0099] Module 502 is used to construct multidimensional feature vectors based on the charge and discharge data of each individual battery cell.

[0100] The processing module 503 is used to process the multidimensional feature vector to obtain the local outlier factor value of each individual cell.

[0101] The determination module 504 is used to determine whether there is an anomaly in each individual cell based on the local outlier factor value of each individual cell.

[0102] According to some optional embodiments of this disclosure, the construction module 502 is further configured to perform data cleaning on the charge and discharge data of each individual battery; divide the cleaned charge and discharge data into charging state data, discharging state data and resting state data; construct a two-dimensional feature vector based on the voltage data in the charging state data; the two-dimensional feature vector represents the voltage of different individual batteries at the same time and the voltage of the same individual battery at different times in two dimensions.

[0103] According to some optional embodiments of this disclosure, the processing module 503 includes: a first determining unit, configured to determine the average value of any row feature vector or any column feature vector representing the voltage of the same single cell at different times in the multidimensional feature vector; a second determining unit, configured to determine the average value as point P; a third determining unit, configured to determine the local outlier factor value of point P; and a fourth determining unit, configured to determine the local outlier factor value of point P as the local outlier factor value of the single cell.

[0104] According to some optional embodiments of this disclosure, the third determining unit is further configured to determine the k-nearest neighbor distance of point P; the k-nearest neighbor distance is the distance from point P to its k-th nearest neighbor, where k is a positive integer; determine the k-distance neighborhood of point P based on the k-nearest neighbor distance; the k-distance neighborhood is a circular region centered on point P and with the k-nearest neighbor distance of point P as the radius; determine the reachability distance of point P relative to each point within the k-distance neighborhood of point P based on the k-nearest neighbor distance of point P; determine the local reachability density of point P based on the reachability distance of point P relative to each point within the k-distance neighborhood of point P; and determine the local outlier factor value of point P based on the local reachability density of point P.

[0105] According to some optional embodiments of this disclosure, the third determining unit is further configured to determine the first distance between point P and each point within the k-distance range of point P; determine the maximum distance between the k-nearest neighbor distance of point P and the first distance corresponding to each point; and determine the maximum value as the reachable distance of point P relative to each point within the k-distance range of point P.

[0106] According to some optional embodiments of this disclosure, the third determining unit is further configured to determine the average reachability distance of point P relative to each point within a k-distance range of point P; and to determine the reciprocal of the average reachability distance as the local reachability density of point P.

[0107] According to some optional embodiments of this disclosure, the third determining unit is further configured to determine the local reachability density of each point within the k-distance neighborhood of point P; determine the sum of the ratios of the local reachability density of each point within the k-distance neighborhood of point P to the local reachability density of point P; and determine the ratio of the sum to the number of points within the k-distance neighborhood of point P as the local outlier factor value of point P.

[0108] According to some optional embodiments of this disclosure, the determining module 504 is further configured to compare the local outlier factor value of each individual cell with the local outlier factor value of other individual cells in the energy storage device; if the difference between the local outlier factor value of the individual cell and the local outlier factor value of other individual cells exceeds a preset threshold, it is determined that the individual cell is abnormal.

[0109] The apparatus of this disclosure embodiment can execute the method provided in this disclosure embodiment, and its implementation principle is similar, and it has corresponding technical effects. The actions performed by each module in the apparatus of each embodiment of this disclosure correspond to the steps in the method of each embodiment of this disclosure. For a detailed functional description of each module of the apparatus, please refer to the description in the corresponding method shown above, and it will not be repeated here.

[0110] This disclosure provides an electronic device including a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps of the method provided in any optional embodiment of this disclosure. Compared with the prior art, it can achieve the following: by dividing battery data into charge and discharge states, obtaining battery voltage information in the time and spatial dimensions, and combining it with anomaly detection algorithms, it can accurately locate abnormal batteries at the end of each charge and discharge cycle. This improves the accuracy and efficiency of anomaly detection in energy storage devices and enables rapid loading and efficient processing of large amounts of data from energy storage devices. In one alternative embodiment, an electronic device is provided, such as Figure 6 As shown, Figure 6 The illustrated electronic device 6000 includes a processor 6001 and a memory 6003. The processor 6001 and the memory 6003 are connected, for example, via a bus 6002. Optionally, the electronic device 6000 may further include a transceiver 6004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 6004 is not limited to one type, and the structure of the electronic device 6000 does not constitute a limitation on the embodiments of this disclosure.

[0111] Processor 6001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with this disclosure. Processor 6001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0112] Bus 6002 may include a pathway for transmitting information between the aforementioned components. Bus 6002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 6002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0113] The memory 6003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium capable of carrying or storing computer programs and capable of being read by a computer, without limitation herein.

[0114] The memory 6003 stores computer programs that execute embodiments of the present disclosure and is controlled by the processor 6001. The processor 6001 executes the computer programs stored in the memory 6003 to implement the steps shown in the foregoing method embodiments.

[0115] Electronic devices include, but are not limited to, computer equipment.

[0116] This disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0117] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0118] It should be understood that although arrows indicate various operation steps in the flowcharts of the embodiments of this disclosure, the order in which these steps are implemented is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of the embodiments of this disclosure, the implementation steps in each flowchart can be executed in other orders as required. Furthermore, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages based on the actual implementation scenario. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage can also be executed at different times. In scenarios where execution times differ, the execution order of these sub-steps or stages can be flexibly configured as required, and the embodiments of this disclosure do not limit this.

[0119] The above description is only an optional implementation method for some implementation scenarios of this disclosure. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this disclosure without departing from the technical concept of this disclosure also fall within the protection scope of the embodiments of this disclosure.

Claims

1. A method for testing an energy storage device, characterized in that, include: Acquire charging and discharging data of each individual battery cell in the energy storage device; A multidimensional feature vector is constructed based on the charge and discharge data of each individual battery cell; The multidimensional feature vector is processed to obtain the local outlier factor value of each individual battery cell; The presence or absence of anomalies in each individual cell is determined based on the local outlier factor value of each cell.

2. The method according to claim 1, characterized in that, The construction of a multi-dimensional feature vector based on the charge and discharge data of each individual battery cell includes: The charge and discharge data of each individual battery cell are cleaned. The cleaning-processed charge and discharge data are divided into charging status data, discharging status data, and static status data. A two-dimensional feature vector is constructed based on the voltage data in the charging state data; the two-dimensional feature vector represents the voltage of different individual cells at the same time and the voltage of the same individual cell at different times in two dimensions.

3. The method according to claim 1, characterized in that, The multidimensional feature vector is processed to obtain the local outlier factor value of each individual battery cell, including: Determine the average value of any row feature vector or any column feature vector representing the voltage of the same single cell at different times in the multidimensional feature vector; The average value is defined as point P; Determine the local outlier value at point P; The local outlier value at point P is determined as the local outlier value of the single cell.

4. The method according to claim 3, characterized in that, Determining the local outlier factor value at point P includes: Determine the k-nearest neighbor distance of point P; the k-nearest neighbor distance is the distance from point P to its k-th nearest neighbor, where k is a positive integer; The k-distance region of point P is determined based on the k-nearest neighbor distance; the k-distance region is a circular area with point P as the center and the k-nearest neighbor distance of point P as the radius; Based on the k-nearest neighbor distances of point P, determine the reachable distance of each point within the k-distance range of point P relative to point P. Based on the reachability distance of point P relative to each point in the k-distance neighborhood of point P, determine the local reachability density of point P; The local outlier factor value of point P is determined based on the local reachability density of point P.

5. The method according to claim 4, characterized in that, The step of determining the reachable distance of point P relative to each point within the k-distance neighborhood of point P based on the k-nearest neighbor distances of point P includes: Determine the first distance between point P and each point within the k-distance range of point P; Determine the maximum distance between the k-nearest neighbor distances of point P and the first distances corresponding to each point; The maximum value is determined as the reachable distance of point P relative to each point within the k-distance range of point P.

6. The method according to claim 4, characterized in that, Determining the local reachability density of point P based on the reachability distance of each point within the k-distance neighborhood of point P includes: Determine the average reachable distance of each point within the k-distance range of point P relative to point P; The reciprocal of the average reachable distance is determined as the local reachability density of point P.

7. The method according to claim 4, characterized in that, Determining the local outlier factor value of point P based on the local reachability density of point P includes: Determine the local reachability density of each point within the k-distance neighborhood of point P; Determine the sum of the ratios of the local reachability density of each point within the k-distance neighborhood of point P to the local reachability density of point P; The ratio of the number of points within the k-distance neighborhood of point P is determined as the local outlier factor value of point P.

8. The method according to any one of claims 1 to 7, characterized in that, The determination of whether each individual cell is abnormal based on its local outlier factor value includes: The local outlier factor value of each individual battery cell is compared with the local outlier factor values ​​of other individual batteries cells in the energy storage device. If the difference between the local outlier factor value of a single cell and the local outlier factor values ​​of other single cells exceeds a preset threshold, the single cell is determined to be abnormal.

9. A detection device for an energy storage device, characterized in that, include: The acquisition module is used to acquire the charging and discharging data of each individual battery in the energy storage device; A construction module is used to construct a multi-dimensional feature vector based on the charge and discharge data of each individual battery cell; The processing module is used to process the multidimensional feature vector to obtain the local outlier factor value of each individual battery cell; The determination module is used to determine whether each individual cell has an anomaly based on its local outlier factor value.

10. An electronic device, comprising: A memory, a processor, and a computer program stored on the memory, characterized in that the processor executes the computer program to implement the method of any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 8.

Images