System and method for detecting abnormal charging events

Abstract

Embodiments of this application provide a system and method for detecting abnormal charging events. The system (100) may include a communication interface (202) configured to receive multivariate charging data for at least two charging events (702, 703, 704, 705), each charging event (702, 703, 704, 705) including at least two variables, each variable corresponding to a charging characteristic. The system (100) may further include at least one processor (204). The at least one processor may be configured to determine the differences between each pair of charging events (702, 703, 704, 705) based on the multivariate charging data of the two charging events (702, 703, 704, 705). The at least one processor (204) may be further configured to cluster the at least two charging events (702, 703, 704, 705) based on the determined differences. The at least one processor (204) may also be configured to detect abnormal charging events based on the clustering results.

Description

Technical Field

[0001] This invention relates to systems and methods for detecting abnormal charging events, and more specifically, to systems and methods for detecting abnormal charging events based on clustering at least two charging events each having multivariable time series characteristics. Background Technology

[0002] Electrical energy can be provided by sources such as batteries. A battery is a device consisting of one or more electrochemical units with external connections, used to provide electrical energy to electronic devices such as mobile phones, flashlights, and electric vehicles. When a battery is used to provide electrical energy, the electrochemical units generate electrical energy through chemical reactions.

[0003] To provide sufficient power to drive large electrical equipment, such as electric vehicles, many battery cells are connected in series and / or parallel to form a battery pack. For safety purposes, the battery management system (BMS) of an electric vehicle can monitor the characteristics of the battery pack and / or battery cells (e.g., voltage, current, and temperature) in real time. If battery characteristic values ​​reach predetermined safety thresholds during operation or charging, the BMS can send a safety alert to the user. However, real-time monitoring of battery performance may not be sufficient to prevent serious battery hazards (e.g., fire and / or explosion). Other methods use big data technologies to analyze battery consistency (e.g., voltage inconsistencies between battery cells). However, these methods cannot predict battery hazards not caused by battery inconsistency issues.

[0004] Embodiments of this application address the aforementioned problems by providing a system and method for detecting abnormal charging events based on clustering at least two charging events, in order to detect abnormal charging events. Summary of the Invention

[0005] Embodiments of this application provide a system for detecting abnormal charging events. The system may include a communication interface configured to receive multivariate charging data for at least two charging events, each charging event including at least two variables, each variable corresponding to a charging characteristic. The system may further include at least one processor. The at least one processor may be configured to determine the difference between every two charging events based on the multivariate charging data of the two charging events. The at least one processor may be further configured to cluster the at least two charging events based on the determined differences. The at least one processor may also be configured to detect abnormal charging events based on the clustering results.

[0006] Embodiments of this application also provide a method for detecting abnormal charging events. The method may include receiving multivariate charging data for at least two charging events via a communication interface, wherein the charging data for each charging event includes at least two variables, each variable corresponding to a charging characteristic. The method may further include determining the differences between every two charging events based on the multivariate charging data of the two charging events using at least one processor. The method may also include clustering the at least two charging events based on the determined differences using at least one processor. The method may further include detecting abnormal charging events based on the clustering results using at least one processor.

[0007] Embodiments of this application further provide a non-transitory computer-readable storage medium storing computer instructions that, when executed by one or more processors, cause the processors to perform a method for detecting abnormal charging events. The method may include receiving multivariate charging data for at least two charging events, each charging event including at least two variables, each variable corresponding to a charging characteristic. The method may further include determining the difference between each pair of charging events based on the multivariate charging data. The method may also include clustering the at least two charging events based on the determined differences. The method may further include detecting abnormal charging events using at least one processor based on the clustering results.

[0008] It should be understood that, as stated, the above general description and the following detailed description are exemplary and explanatory only, and not restrictive of the invention. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of an exemplary system for detecting abnormal charging events, according to some embodiments of this application.

[0010] Figure 2 This is a block diagram of an exemplary server for detecting abnormal charging events, according to some embodiments of this application.

[0011] Figure 3 This is a flowchart illustrating an exemplary method for detecting abnormal charging events according to some embodiments of this application.

[0012] Figure 4 This is a flowchart illustrating an exemplary method for determining the distance between every two charging events, according to some embodiments of this application.

[0013] Figure 5 This is a flowchart illustrating an exemplary method for clustering charging events according to some embodiments of this application.

[0014] Figures 6A-6B These are two example time series and corresponding alignment matrices shown according to some embodiments of this application.

[0015] Figures 7A-7E This is an exemplary clustering method shown according to some embodiments of this application. Detailed Implementation

[0016] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. Where possible, the same reference numerals will be used throughout the drawings to refer to the same or similar parts.

[0017] Embodiments of this application provide systems and methods for detecting anomalous charging events based on clustering at least two charging events. A charging event can be the process by which any electrical device (e.g., an electric vehicle) charges its rechargeable battery pack from an external power source. Compared to existing solutions, the disclosed systems and methods do not rely on analyzing battery consistency data. Instead, the disclosed systems and methods can cluster charging data from multiple charging events (e.g., battery state of charge (SoC), charging current, and / or charging voltage) to detect one or more anomalous charging events exhibiting behavior different from other charging events. Detecting these anomalous charging events can help monitor the condition of electrical devices (e.g., electric vehicles) and prevent serious battery damage.

[0018] Figure 1 This is a schematic diagram of an exemplary system 100 (hereinafter referred to as "System 100") for detecting abnormal charging events, according to some embodiments of this application. Figure 1 As shown, system 100 can monitor charging events of vehicle 110 and / or charging station 120 and detect abnormal charging events of the vehicle. In some embodiments, system 100 may include database 130, server 140, and display device 150. In some embodiments, server 140 can request / download charging data (not shown) from database 130 via a network. Battery charging data can be obtained from one or more vehicles 110 and / or charging stations 120 describing the characteristics of their respective charging events. Server 140 can detect abnormal charging events from the charging data and transmit the detection result 103 to display device 150 for display. In some embodiments, server 140 can determine whether further action is required, such as the need to replace at least one cell of the battery pack and / or the need for a service recommendation for further diagnostics of the battery pack, and display the detection result 103 including the recommendation on display device 150.

[0019] Consistent with some embodiments, each vehicle 110 may be an electric vehicle with an electric motor or a hybrid vehicle including an internal combustion engine and at least one electric motor. Vehicle 110 may have a battery pack to power the electric motor. The battery pack may have multiple battery cells connected in series and / or in parallel to provide a greater power output.

[0020] In some embodiments, vehicle 110 may be charged at charging station 120. In some embodiments, vehicle 110 may be equipped with sensors (not shown) for detecting / measuring charging data 102 that reflects the characteristics of the charging process of the battery pack and / or battery cells. In some embodiments, the sensors may include electrical sensor units, such as voltage sensors, current and / or temperature sensors for the battery pack and / or battery cells.

[0021] In some embodiments, charging data 102 may be indicative data of the charging characteristics of the batteries within the battery pack when the vehicle 110 is in a charging state. For example, charging data 102 may include partial and / or complete data acquired under the control of a battery management system (BMS) that manages rechargeable batteries (e.g., battery packs and / or battery cells) for purposes such as protecting the battery system from operating outside its safe operating area, monitoring its status, calculating auxiliary data, etc. It should be understood that charging data 102 may contain certain metadata (e.g., vehicle VIN information) to map charging events to the vehicle that uploaded the charging data.

[0022] In some embodiments, charging data 102 may include a multivariate time series (MTS) with multiple time-varying variables, each corresponding to a charging characteristic (e.g., the charging current of the battery pack). For example, time-varying variables may include, but are not limited to, the battery SoC of the battery pack, the charging current of the battery pack, the current required by the battery pack's BMS, the total charging voltage of the battery pack, the maximum charging voltage of the battery cells, and the highest temperature of the battery cells. The battery SoC represents the battery's charge level relative to its capacity. SoC is measured as a percentage (0% = empty; 100% = full).

[0023] Figure 1Table 115 illustrates an exemplary charging data segment that vehicle 110 may generate during a charging event. As shown in Table 115, the "Voltage 1" column includes a series of charging voltage values. Current 1 can be the charging current value of the battery pack corresponding to the charging voltage value under Voltage 1. Both charging voltage and charging current can be indexed in chronological order. The "Voltage 2" and "Voltage 3" columns can include the charging voltage values ​​of two individual battery cells, which can be used to calculate the maximum charging voltage of the battery cells at at least two time points. In some alternative embodiments, sensors on vehicle 110 can calculate the maximum voltage of the battery cells so that the raw charging data of a single battery cell is not included in charging data 102. Similarly, sensors on vehicle 110 can calculate the highest temperature of the battery cells at at least two time points.

[0024] Consistent with some embodiments, such as Figure 1 As shown, charging data 102 can be obtained from charging station 120. In some embodiments, at least one of the charging stations 120 may be an infrastructure that provides electrical power for charging plug-in electric vehicles (including electric vehicles, neighborhood electric vehicles, and plug-in hybrid electric vehicles). The charging station can be used by multiple electric vehicles and has an additional current or connection sensing mechanism that can disconnect the power supply when an electric vehicle is not charging.

[0025] Similar to vehicle 110, each charging station 120 may be additionally equipped with sensors (not shown) for detecting / measuring charging data 102 of the battery pack and / or battery cells. For example, a current sensor coupled to charging station 120 can monitor power consumption and maintain connection only when demand is within a predetermined range. These sensors react faster, have fewer faulty parts, and may have lower design and implementation costs. These sensors can use standard connectors and can help suppliers monitor or charge actual power consumption. The sensors can record the charging current of the battery pack (e.g., actual current and current required by the BMS) and charging voltage at at least two time points.

[0026] In some embodiments, some charging stations may not be equipped with voltage and / or temperature sensors. These charging stations may, with permission, download charging voltage and / or temperature data from sensors on the plug-in vehicle. In some alternative embodiments, the plug-in vehicle may not allow the charging station to download charging data from vehicle sensors. Therefore, the charging data obtained from the charging station may differ from the data obtained from the vehicle. Since both vehicle 110 and charging station 120 can generate charging data for the charging event, duplicate data may exist in the charging data 102. The charging data 102 may be cleaned / filtered during data preprocessing (also known as data cleaning). The data cleaning process will be disclosed in more detail below.

[0027] In some embodiments, charging data 102 may be stored in memory and / or storage coupled to the sensor. For example, charging data 102 may be stored in formats such as *.xls, *.xlsx, *.csv, etc. It is understood that the format in which charging data 102 is stored is not limited to the format disclosed herein and may be modified for other charging purposes.

[0028] In some embodiments, charging data 102 may be uploaded to database 130 in real time (e.g., via a stream from a sensor to database 130), or uploaded to database 130 via a network (not shown) after a period of time. In some embodiments, the network may be a wireless local area network (WLAN), a wide area network (WAN), a wireless network (e.g., radio waves), a cellular network, a satellite communication network, and / or a local or short-range wireless network (e.g., Bluetooth). TM (Or near-field communication) is used to transmit charging-related information of the vehicle 110 battery. In some other embodiments, charging data 102 can also be uploaded to database 130 via a direct link (e.g., via a communication cable). For example, the user of vehicle 110 (i.e., the driver / operator) can periodically drive / guide vehicle 110 to the terminal where database 130 is located to upload data.

[0029] In some embodiments, server 140 can download charging data 102 from database 130 in real time via the same and / or different networks that upload charging data 102 to database 130, or via a communication cable used for centralized downloading (e.g., every few seconds, every few minutes, etc.) of charging data 102. In some embodiments, server 140 can process charging data 102 and generate a detection result 103 of an abnormal charging event based on the processed charging data 102. Server 140 can monitor the battery pack and determine measures to be taken to maintain the safety and high performance of the battery pack. In some embodiments, recommended measures may be included in the detection result 103. In some embodiments, system 100 may optionally include a display device 150 for displaying the detection result 103, for example, to the manager of charging station 120 and / or the user of vehicle 110. It is conceivable that, with Figure 1 Compared to the components shown, system 100 may include more or fewer components.

[0030] Figure 2 This is a block diagram of an exemplary server (hereinafter referred to as "server 140") for detecting abnormal charging events, according to some embodiments of this application. Consistent with this application, server 140 can receive charging data 102 and generate a detection result 103 indicating an abnormal charging event based on the charging data 102. Although as Figure 2As shown, server 140 is a physical backup device, but it is conceivable that in some embodiments, server 140 may be implemented as a cloud software, an application on database 130 and / or display device 150, a virtual server, or a distributed server for multiple devices. For example, in some embodiments, charging data cleaning and preprocessing may be implemented by a database management system equipped with database 130, while the remaining functions may be implemented by display device 150. Consistent with the present invention, server 140 may be a general-purpose server or a proprietary device specifically designed for detecting abnormal charging events.

[0031] like Figure 2 As shown, in some embodiments, server 140 may include communication interface 202 and processor 204. In some embodiments, server 140 may also include memory 206 and storage 208. In some embodiments, server 140 may have different modules in a single device, such as integrated circuit (IC) chips (implemented as application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs)) or stand-alone devices with dedicated functions. In some embodiments, one or more components of server 140 may be located in a cloud computing environment, or alternatively, may be located in a single location or distributed locations. Components of server 140 may be in an integrated device or distributed in different locations, but communicate with each other via a network (not shown).

[0032] Communication interface 202 can receive data (e.g., charging data 102) from database 130 and can do so via communication cable, wireless local area network (WLAN), wide area network (WAN), wireless network (e.g., radio waves), cellular network, satellite communication link and / or local or short-range wireless network (e.g., Bluetooth). TM Data (e.g., detection result 103) can be sent to the display device 150 via other communication methods. In some embodiments, the communication interface 202 may be an Integrated Services Digital Network (ISDN) card, a cable modem, a satellite modem, or a modem that provides a data communication connection. As another example, the communication interface 202 may be a Local Area Network (LAN) card to provide a data communication connection with a compatible LAN. Wireless links can also be implemented through the communication interface 202. In such implementations, the communication interface 202 can send and receive electrical, electromagnetic, or optical signals carrying digital data streams representing various types of information over a network.

[0033] Consistent with some embodiments, the communication interface 202 may further provide the received data to the memory 208 for storage or to the processor 204 for processing. The communication interface 202 may also receive the detection result 103 generated by the processor 204 and provide the detection result 103 to the display device 150.

[0034] Processor 204 may include any suitable type of general-purpose or special-purpose microprocessor, digital signal processor, or microcontroller. Processor 204 may be configured as a separate processor module dedicated to processing charging data 102. Alternatively, processor 204 may be configured as a shared processor module for performing other functions unrelated to detecting abnormal charging events (e.g., processor 204 may be a shared processor module on database 130 and / or a shared processor module on display device 150).

[0035] like Figure 2 As shown, processor 204 may include multiple modules, such as a data cleaning unit 210, a difference determination unit 212, a data clustering unit 214, an anomaly determination module 216, etc. These modules (and any corresponding sub-modules or sub-units) may be hardware units (e.g., part of an integrated circuit) of processor 204 designed for use with other components, or software units implemented by processor 204 by executing at least a portion of a program. This program may be stored on a computer-readable medium and may perform one or more functions when executed by processor 204. Although Figure 2 Units 210-216, all located within a single processor 204, are shown, but it is conceivable that these units could be distributed among multiple processors that are adjacent to or remote from each other.

[0036] After receiving charging data 102 from database 130, data cleaning unit 210 can clean and preprocess (e.g., filter) the data. In some embodiments, damaged, incorrect, and / or inaccurate data (i.e., data with errors) can be discarded. For example, data with predetermined types of errors, such as typographical errors and / or formatting errors (e.g., duplicates of one or more data packets, misalignment of one or more data packets, incorrect payload format, payload with null values, etc.), can be cleaned by data cleaning unit 210.

[0037] In some embodiments, the data cleaning unit 210 can identify data segments of the charging data 102 obtained from the vehicle 110 and / or the charging station 120, and further filter the charging data 102 based on the data segments. Consistent with some embodiments, the charging data 102 may include duplicate charging data segments. This may occur when the vehicle and the charging station upload charging data for the same charging event. Once duplicate data segments are detected, the data cleaning unit 210 can delete the duplicate data segments.

[0038] In some embodiments, the data cleaning unit 210 can process the charging data 102 to obtain a multivariate time series, which can be used to detect abnormal charging events performed by other units of the processor 204. For example, the charging data 102 may include measured charging information, such as the temperature of the battery cell at at least two time points, and the data cleaning unit 210 can calculate the highest temperature of the battery cell at each time point based on the measured temperature information.

[0039] In some embodiments, the difference determination unit 212 may be configured to measure the difference (or conversely, similarity) between every two charging events. Consistent with some embodiments, the charging data for each charging event may include a multivariate time series (e.g., Figure 1 Table 115 (as shown) includes time-varying data for multiple variables. For example, a charging event can be described using six time series of charging characteristics. These charging characteristics could be the battery SoC, the charging current of the battery pack, the current required by the battery pack's BMS, the total voltage of the battery pack, the maximum voltage of the battery cell, and the highest temperature of the battery cell. It is possible to consider using time series of more or less charging characteristics to describe a charging event. Similarly, a charging event can be described using charging characteristics other than those mentioned above, and / or some of the charging characteristics mentioned above together.

[0040] In some embodiments, the difference determination unit 212 may determine the distance between every two charging events. "Distance" measures the similarity or difference between two charging events in their respective charging characteristics. In some embodiments, the distance is smaller when the charging characteristics are similar to each other, and larger when the charging characteristics are different from each other. In some embodiments, distance is a set measure of the overall difference between multiple variables indicating multiple charging characteristics.

[0041] In some embodiments, vehicle 110 and / or charging station 120 may be equipped with sensors manufactured by different manufacturers. For example, the sensors may operate in different ways, such as at different sampling frequencies. Therefore, charging data 102 acquired from different vehicles and / or charging stations may differ and cannot be directly mapped to each other. Furthermore, some data values ​​may be lost or erroneous during data transmission. To compensate for these inconsistencies in the charging data 102, the difference determination unit 212 may use Dynamic Time Warping (DTW) to calculate the distance between every two charging events in the charging data space. Typically, DTW is an algorithm used to measure the difference / similarity between two time series (e.g., time series). The calculation of the distance between every two charging events will be performed in conjunction with... Figure 4 and Figures 6A-6B More details are described in the related section.

[0042] In some embodiments, the data clustering unit 214 can be configured to cluster charging events based on the distance determined by the difference determination unit 212. In some embodiments, the data clustering unit 214 can assign each charging event to two clusters, one corresponding to normal charging and the other to abnormal charging. It is conceivable that the number of clusters is not limited to two, but can be greater than two, in which case one or more clusters can correspond to normal charging, and the remaining clusters correspond to abnormal charging. The data clustering unit 214 can implement any suitable clustering method to cluster charging events.

[0043] Data clustering unit 214 can initiate the clustering process by initializing cluster centers. In some embodiments, when using two clusters, two charging events can be selected as initial cluster centers. For example, these two charging events can have the largest distance among all pairs of charging events in at least two charging events. In another example, a charging event can be randomly selected as the first initial cluster center, and a second cluster center can be selected as another charging event, for example, based on its distance from the first initial cluster center. For example, the second cluster center can be the charging event with the largest distance from the charging event selected as the first initial cluster center.

[0044] After initializing the cluster centers, data clustering unit 214 can be configured to associate the remaining charging events with one of the two initial cluster centers. For example, this assignment can be based on a determined distance calculated by difference determination unit 212. In some embodiments, data clustering unit 214 can dynamically recalculate the cluster centers based on the charging events associated with each cluster. Data clustering unit 214 can repeat the steps of assigning the remaining charging events and recalculating the cluster centers until all charging events are clustered. (This will be combined with...) Figure 5 and Figures 7A-7E The clustering of charging events is disclosed in more detail.

[0045] In some embodiments, an anomaly determination unit 216 can be configured to detect abnormal charging events based on the clustering results obtained by the data clustering unit 214. In some embodiments, the anomaly determination unit 216 can label the clusters obtained by the data clustering unit 214 as normal charging or abnormal charging. In some embodiments, when two clusters are obtained, the cluster with more charging events can be labeled as normal, while the other cluster with fewer charging events can be labeled as abnormal. In practice, a normal cluster can contain more charging events than an abnormal cluster. For example, in 18 charging events, 16 can be associated with the first cluster, and the remaining 2 are associated with the second cluster; in this case, the first cluster is labeled as normal, and the second cluster is labeled as abnormal. In some embodiments, when more than two clusters are obtained, the anomaly determination unit 216 can label clusters based on their cluster centers. For example, the cluster whose center is furthest from other clusters can be identified as abnormal. The data clustering unit 214 can determine charging events associated with abnormal clusters as abnormal charging events.

[0046] After identifying an abnormal charging event, the anomaly identification unit 216 can send the detection result (e.g., detection result 103) to a display device (e.g., display device 150) via the communication interface 202. In some embodiments, the display device can be an output device that presents information in a visual form. For example, the display device 150 can be installed in the vehicle 110 and / or charging station 120 for vehicle users or charging station managers to view the detection result 103. The detection result 103 may include cleaned charging data 102 and a corresponding detection tag. The detection tag may be a binary value that indicates whether the corresponding charging data comes from an abnormal charging event. In some embodiments, the display device 150 can graphically display abnormal charging events and normal charging in the same chart. The display device 150 can also graphically display a single dimension of the charging data. For example, the display device 150 can use different colors or markers to show the battery SoC of abnormal and normal charging events in a time series. In some alternative embodiments, only the identification of the abnormal charging event and the corresponding vehicle 110 and / or charging station 120 can be provided to the display device 150 for display.

[0047] In some embodiments, server 140 may further include memory 206 and storage 208. Memory 206 and storage 208 may include any suitable type of mass storage for storing any type of information that processor 204 may need to process. Memory 206 and storage 208 may be volatile or non-volatile, magnetic, semiconductor, magnetic tape, optical, removable, non-removable, or other types of storage devices or tangible (i.e., non-transitory) computer-readable media, including but not limited to ROM, flash memory, dynamic RAM, and static RAM. Memory 206 and / or storage 208 may be configured to store one or more computer programs executable by processor 204 to detect abnormal charging events disclosed in this application. For example, memory 206 and / or storage 208 may be configured to store programs executable by processor 204 to clean charging data, determine the distance between cleaned charging data, and / or cluster the charging data based on distance.

[0048] Memory 206 and / or storage 208 can be further configured to store information and data used by processor 204. For example, memory 206 and / or storage 208 can be configured to store various types of data (e.g., raw charging data received from database 130, detection results 103, etc.). Memory 206 and / or storage 208 can also store intermediate data, such as cleaned / filtered charging data, distance, cluster labels, etc. Various types of data can be permanently stored, periodically deleted, or ignored immediately after processing certain data segments.

[0049] Figure 3 This is a flowchart illustrating an exemplary method 300 for detecting abnormal charging events according to some embodiments of this application. In some embodiments, method 300 can be implemented by system 100. Method 300 may include steps S302-S316 as described below. It should be understood that some steps may be optional to perform the disclosure provided in this application. Furthermore, some steps may be performed simultaneously or in conjunction with... Figure 3 The different execution orders shown.

[0050] In step S302, database 130 can receive charging data (e.g., charging data 102) from vehicle 110 and / or charging station 120. For example, database 130 can receive charging data 102 from vehicle 110 in real time (e.g., via a stream from a sensor to database 130) or centrally via a network (not shown) after a period of time. In another example, database 130 can receive charging data 102 directly from vehicle 110 or charging station 120 via a direct link (e.g., via a communication cable). For example, a user of vehicle 110 (i.e., driver / operator) can periodically drive / guide one of the vehicles 110 to the terminal where database 130 is located to upload charging data 102. In step S304, database 130 can store the received charging data 102.

[0051] In some embodiments, in step S306, server 140 can download charging data from database 130. For example, server 140 can download / receive charging data in real time or centrally (e.g., every few seconds, every few minutes, etc.) via communication cable or network (e.g., uploading charging data 102 to database 130 via the same and / or different networks).

[0052] In step S308, server 140 may preprocess (e.g., clean / filter) data (e.g., charging data 102), implemented by data cleaning unit 210 of processor 204. In some embodiments, server 140 may discard / filter corrupt, incorrect, and / or inaccurate data (i.e., erroneous data). In some embodiments of step S308, duplicate charging data may also be filtered by data cleaning unit 210. Consistent with some embodiments, intermediate data, such as the highest temperature of the battery cell, may be calculated based on the temperature data of the battery cell at each time point.

[0053] In step S310, the distance between every two charging events can be calculated by the difference determination unit 212 of the processor 204. In some embodiments, step S310 may include three sub-steps, such as... Figure 4 As shown. In sub-step S412, standardization can be applied to each charging characteristic data sequence individually to eliminate inconsistencies between the acquired charging data. For example, battery SoC data can be rescaled to a mean of 0 and a standard deviation of 1 (unit variance). Charging current data can also be rescaled to a mean of 0 and a standard deviation of 1 (unit variance). In some alternative embodiments, a normalization method can be used instead of a standardization method to rescale the data.

[0054] Consistent with some embodiments, in sub-step S414, the difference determination unit 212 can use DTW to construct an alignment matrix for every two charging events. For example, as Figure 6A As shown, lines 610 and 620 are data plots of two charging events (variable value v.time). DTW can determine the best match of data points between lines 610 and 620 (e.g., ...). Figure 6A (As shown by the dashed lines). The best match is associated with the lowest cost, where the cost is calculated as the sum of the absolute differences between the variable values ​​in each matched pair of data points. In some embodiments, the lowest cost becomes the distance between lines 610 and 620. Figure 6B An alignment matrix 630 is shown, which indicates the best matching path (gray square) for lines 610 and 620.

[0055] In substep S416, the distance between every two charging events can be calculated based on the alignment matrix. For example, the distance between charging events t and r at times i and j can be defined by formula (1):

[0056]

[0057] in 1≤i≤L t and 1≤j≤L r k represents the number of charging features included in charging events t and r. L t L represents the number of time points in the charging event t. r This represents the number of time points in the charging event 'r'. 'w' is the predetermined weight for each charging characteristic, indicating the importance of each characteristic in the clustering. For example, a user can preset the weight of charging current to 1.2 and the weight of other charging characteristics to 1.0 to reflect that charging current is a more important characteristic in identifying abnormal charging events.

[0058] Using formula (1), the distance D(i,j) between two charging events can be defined by formula (2):

[0059]

[0060] Where D(1,1)=d(1,1), D(1,0)=D(2,0)=…=D(i,0)=0, D(0,1)=D(0,2)=…=D(0,j)=0. These two charging events have time points i and j, respectively. The distance between any two charging events can be calculated using formulas (1) and (2).

[0061] return Figure 3 In step S312, the data clustering unit 214 of the processor 204 can be configured to cluster charging events based on the distance obtained in step S310. Consistent with some embodiments, the details of the clustering are as follows: Figure 5 The sub-steps S512-S522 are described in the text. Figures 7A-7E This is an exemplary clustering method shown in some embodiments of this application.

[0062] In sub-step S512, data clustering unit 214 can randomly select a charging event as the first initial cluster center, such as... Figure 7A As shown. For example, it is necessary to cluster six charging events (e.g., events 701-706) in the charging space into two clusters. Figure 7A As shown, event 706 (a solid square) can be randomly selected as the first cluster center. Data clustering unit 214 can be further configured to select a second charging event based on its distance to the first cluster center. In some embodiments, the charging event with the largest distance to the first cluster center can be selected as the second cluster center. Figure 7B As shown, the dashed lines represent the distances between event 706 and other events. Since event 701 (the solid square) is the event with the largest distance from event 706, it was selected as the second cluster center.

[0063] As another example, data clustering unit 214 can be configured to select two charging events as initial cluster centers. In some embodiments, the two charging events may have the largest distance among all distances calculated between any two charging events. For example, data clustering unit 214 can rank the differences calculated in step S310, determine the largest distance among all distances, and then select the two charging events associated with that distance. For example, as Figure 7B As shown, the two charging events with the maximum distance are events 701 and 706.

[0064] In sub-step S514, data clustering unit 214 can be configured to cluster remaining charging events (e.g., Figure 7C Events 702-705 in the data are associated with the nearest cluster center (e.g., Figure 7C (Event 701 or Event 706 in the context). For example, such as... Figure 7C As shown, event 702 is assigned to cluster center 701 because its distance to cluster center 701 (dashed line) is shorter than its distance to cluster center 706. Figure 7C As shown, events 702 and 703 are associated with cluster center 701, and events 704 and 705 are associated with cluster center 706. For illustrative purposes only, solid lines are used for separation. Figure 7C Two clusters in the data.

[0065] In sub-step S516, data clustering unit 214 can be configured to calculate the sum of distances from each charging event to the remaining events in the same cluster. For example, as Figure 7DAs shown, the distance between events 704 and 705 is D45 (dashed line). Similarly, the distance between events 704 and 706 is D46, and the distance between events 705 and 706 is D56. The sum of the distances between event 704 and the remaining events (i.e., events 705 and 706) equals the value of (D45 + D46). The sum of the distances between event 705 and the remaining events equals the value of (D45 + D56). The sum of the distances between event 706 and the remaining events equals the value of (D46 + D56).

[0066] In sub-step S518, data clustering unit 214 can be configured to set the charging event with the minimum total distance as the new cluster center. For example, as Figure 7D As shown, since the value of (D45+D56) is less than the value of (D45+D46) or the value of (D46+D56), event 705 is selected as the new cluster center. Figure 7D As shown, events 702 and 705 (solid squares) become the new cluster centers of the two clusters.

[0067] In sub-step S520, data clustering unit 214 can be configured to update clusters by associating remaining charging events with the nearest new cluster centers. For example, as Figure 7E As shown, event 701 is assigned to the new cluster center 702; events 703, 704, and 706 are associated with the new cluster center 705. Figure 7D Compared to the clustering assignment in [previous context], event 703 in [previous context] Figure 7E The cluster label was changed in the event, while other events did not change the clusters they were associated with.

[0068] In sub-step S522, data clustering unit 214 may be configured to determine whether any charging events have changed their cluster labels. If, compared to after step S514, no charging event has changed its cluster label after step S520 (sub-step S522: No), then data clustering unit 214 can complete the clustering. If the cluster label of one or more charging events changes (e.g., event 703 changes its cluster label from the previous one), then the data clustering unit 214 can complete the clustering. Figure 7D The first cluster in the association is changed to be related to Figure 7E If the second cluster is associated (sub-step S522: yes), then sub-steps S516-S522 can be repeated until no charging event changes its cluster label after associating with the most recent new cluster center.

[0069] return Figure 3In step S314, the anomaly determination unit 206 of the processor 204 can be configured to detect abnormal charging events based on the clustering results obtained by the data clustering unit 214. Consistent with some embodiments, the anomaly determination unit 206 can label the clusters obtained by the data clustering unit 214 as normal charging or abnormal charging. For example, a cluster with more charging events can be labeled as normal, while another cluster with fewer charging events can be labeled as abnormal. In some alternative embodiments, the anomaly determination unit 206 can label clusters based on the variance of the two clusters. The variance of the clusters describes the similarity between charging events. In practice, charging events in normal clusters can have more similar charging behaviors than charging events in abnormal clusters.

[0070] In step S316, the detection result (e.g., detection result 103) can be displayed on a display device (e.g., display device 150) for further review. Consistent with some embodiments, display device 150 can graphically display normal and abnormal clusters. Consistent with some embodiments, only abnormal charging events and the corresponding vehicles 110 and / or charging stations 120 are provided to display device 150 for display.

[0071] In some embodiments, the detection results 103 obtained by method 300 can be used to train a machine learning model for detecting abnormal charging events. For example, charging data 102 of detected abnormal charging events and remaining normal charging events can be used as training data to train the machine learning model. The learned model can be used to detect anomalies from subsequently acquired charging data. For example, the trained machine learning model can be pre-programmed in a vehicle or charging station to detect abnormal charging events in real time.

[0072] Another aspect of the invention relates to a non-transitory computer-readable medium storing instructions that, when executed, cause one or more processors to perform the methods described above. The computer-readable medium may include volatile or non-volatile, magnetic, semiconductor, magnetic tape, optical, removable, non-removable, or other types of computer-readable media or computer-readable storage devices. For example, as disclosed, a computer-readable medium may be a storage device or storage module having computer instructions stored thereon. In some embodiments, the computer-readable medium may be a disk or flash drive having computer instructions stored thereon.

[0073] It will be apparent to those skilled in the art that various modifications and variations can be made to the disclosed system and related methods. Other embodiments will be apparent to those skilled in the art in light of the specifications and practices of the disclosed system and related methods.

[0074] This specification and examples are to be considered exemplary only, and the true scope is indicated by the following claims and their equivalents.

Claims

1. A system for detecting abnormal charging events, comprising: The communication interface is configured to receive multivariate charging data for at least two charging events, each charging event including at least two variables, each variable corresponding to a charging characteristic, wherein the charging event is the process of an electrical device charging its rechargeable battery pack from an external power source. as well as At least one processor is coupled to the communication interface and configured to: Based on the multivariate charging data of two charging events, the distance between each pair of charging events is determined to determine the difference between each pair of charging events; Cluster the at least two charging events based on the identified differences; as well as Based on the clustering results, the clusters are labeled as normal charging or abnormal charging in order to detect abnormal charging events.

2. The system as described in claim 1, characterized in that, The multivariate charging data for each charging event includes a multivariate time series, which includes the values ​​of at least two variables at at least two time points.

3. The system as described in claim 1, characterized in that, The at least two variables correspond to at least two of the following: the battery's state of charge (SoC), charging current, required current, total battery pack voltage, maximum cell voltage, or maximum cell temperature.

4. The system as described in claim 1, characterized in that, To determine the difference between every two charging events, the at least one processor is further configured to: Based on the multivariate charging data of the two charging events, an alignment matrix is ​​constructed for the two charging events; as well as Based on the alignment matrix, the dynamic time warping (DTW) distance between the two charging events is calculated.

5. The system as described in claim 4, characterized in that, To calculate the DTW distance between the two charging events, the at least one processor is further configured to: Calculate each multivariate difference based on one variable of the multivariate charging data; and The DTW distance is calculated as a weighted sum of the multivariate differences, with each multivariate difference weighted by a predetermined weight.

6. The system as described in claim 1, characterized in that, In order to cluster the charging events, the at least one processor is configured to: Two charging events were selected as the initial cluster centers; Based on the identified differences, the remaining charging events are associated with the nearest cluster center; as well as The cluster centers are recalculated using the charging events associated with each cluster.

7. The system as described in claim 6, characterized in that, To recalculate the cluster centers, the at least one processor is further configured to: Calculate the sum of the differences between each charging event associated with a cluster and the remaining charging events associated with the same cluster; as well as The charging event with the smallest sum of differences is set as the new cluster center of the cluster.

8. The system as described in claim 6, characterized in that, The two charging events include a first charging event and a second charging event, and the difference between the first charging event and the second charging event is the largest among all differences between any two of the at least two charging events.

9. The system as described in claim 1, characterized in that, The multivariate charging data comes from an electric vehicle that is being charged or a charging station that is charging the electric vehicle.

10. A method for detecting abnormal charging events, comprising: Receive multivariate charging data for at least two charging events, wherein the charging data for each charging event includes at least two variables, each variable corresponding to a charging characteristic, and the charging event is the process of an electrical device charging its rechargeable battery pack from an external power source; Based on the multivariate charging data of two charging events, the distance between each pair of charging events is determined to determine the difference between each pair of charging events; Cluster the at least two charging events based on the identified differences; as well as Based on the clustering results, the clusters are labeled as normal charging or abnormal charging in order to detect abnormal charging events.

11. The method as described in claim 10, characterized in that, The multivariate charging data for each charging event includes a multivariate time series, which includes the values ​​of at least two variables at at least two time points.

12. The method as described in claim 10, characterized in that, The at least two variables correspond to at least two of the following: the battery's state of charge (SoC), charging current, required current, total battery pack voltage, maximum cell voltage, or maximum cell temperature.

13. The method as described in claim 10, characterized in that, Determining the difference between every two charging events further includes: Based on the multivariate charging data of the two charging events, an alignment matrix is ​​constructed for the two charging events; and Based on the alignment matrix, the dynamic time warping (DTW) distance between the two charging events is calculated.

14. The method as described in claim 13, characterized in that, Calculating the DTW distance between the two charging events further includes: Calculate each multivariate difference based on one variable of the multivariate charging data; and The DTW distance is calculated as a weighted sum of the multivariate differences, with each multivariate difference weighted by a predetermined weight.

15. The method as described in claim 10, characterized in that, The clustering of charging events further includes: Two charging events were selected as the initial cluster centers; Based on the identified differences, the remaining charging events are associated with their nearest cluster centers; and The cluster centers are recalculated using the charging events associated with each cluster.

16. The method as described in claim 15, characterized in that, Recalculating the cluster centers further includes: Calculate the sum of the differences between each charging event associated with a cluster and the remaining charging events associated with the same cluster; and The charging event with the smallest sum of differences is set as the new cluster center of the cluster.

17. The method as described in claim 15, characterized in that, The two charging events include a first charging event and a second charging event, and the difference between the first charging event and the second charging event is the largest among all differences between any two of the at least two charging events.

18. The method as described in claim 10, characterized in that, The multivariate charging data comes from an electric vehicle that is being charged or a charging station that is charging the electric vehicle.

19. A non-transitory computer-readable storage medium storing computer instructions that, when executed by one or more processors, cause the one or more processors to perform a method for detecting an abnormal charging event, the method comprising: Receive multivariate charging data for at least two charging events, wherein the charging data for each charging event includes at least two variables, each variable corresponding to a charging characteristic, and the charging event is a process in which an electrical device charges its rechargeable battery pack from an external power source. Based on the multivariate charging data of two charging events, the distance between each pair of charging events is determined to determine the difference between each pair of charging events; Cluster the at least two charging events based on the identified differences; as well as Based on the clustering results, the clusters are labeled as normal charging or abnormal charging in order to detect abnormal charging events.

20. The non-transitory computer-readable storage medium as claimed in claim 19, characterized in that, The multivariate charging data for each charging event includes a multivariate time series, which includes the values ​​of at least two variables at at least two time points.

Images