A charging abnormality detection method and device

Abstract

The present invention relates to the field of abnormal diagnosis technology, and discloses a charging anomaly detection method and device. The method comprises acquiring charging status data in real time; preprocessing the charging status data to obtain normalized data; inputting the normalized data into a pre-trained abnormal warning model to determine whether the current charging status of the electric vehicle and / or charging pile is abnormal; if the current charging status of the electric vehicle and / or charging pile is abnormal, a first warning signal is output, wherein the abnormal warning model performs anomaly detection according to a pre-defined abnormal type and a corresponding alarm threshold; the first warning signal is sent to a central management system, and the normalized data is sent to the cloud, so that the cloud outputs a second warning signal through a pre-trained machine learning model and sends it to the central management system. The present invention can shorten the response time of the entire abnormal detection process and improve the response speed of safety issues.

Description

Technical Field

[0001] The present invention belongs to the technical field of abnormality diagnosis, and in particular relates to a charging abnormality detection method and device. Background Art

[0002] With the popularization of new energy vehicles, the demand for charging piles in the market is growing rapidly, so more and more attention is paid to the safety of the charging process of charging piles.

[0003] Existing charging piles generate a large amount of data during the charging process, which usually needs to be processed and analyzed by a central server. However, this will lead to data transmission delays and slow response speeds on the one hand; on the other hand, it will increase the risk of relying on the central server. For example, if the central server suddenly crashes, the data generated by the charging pile cannot be processed at the moment, resulting in the inability to respond to charging pile circuit failures, equipment damage and safety hazards in a timely manner, further leading to low charging efficiency and increased safety risks. Summary of the invention

[0004] In order to solve at least one technical problem in the background technology, the present invention proposes a charging anomaly detection method and device, which integrate edge computing and big data analysis technology into the existing charging pile infrastructure, so that the data generated by the charging pile can be directly processed and analyzed in the charging pile, reducing the data transmission delay and bandwidth consumption, shortening the response time of the entire anomaly detection process, and being able to quickly respond to various safety issues.

[0005] According to a first aspect of the present invention, the present invention first provides a charging anomaly detection method, which is applied to a charging anomaly detection device, and the charging anomaly detection device operates in a charging pile. The method includes:

[0006] When the electric vehicle is in a charging state, the charging state data is acquired in real time, wherein the charging state data includes the charging pile state data during the charging process and the vehicle state data during the vehicle-pile interaction process;

[0007] Preprocessing the charging state data to obtain normalized data;

[0008] Input the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, output a first warning signal, the first warning signal includes the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and the corresponding alarm threshold;

[0009] The first warning signal is sent to the central management system for alarm, and the normalized data is sent to the cloud so that the cloud can detect the current charging status of the electric vehicle and / or charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system.

[0010] Optionally, the real-time acquisition of charging status data includes:

[0011] Start the temperature sensor, voltage sensor and current sensor installed in the charging pile to obtain the pile end status data in real time, wherein the pile end status data includes the temperature, total output voltage, total output current, power and insulation signal of the charging pile;

[0012] The BMS component installed in the charging pile is started to obtain the vehicle-side status data in the vehicle-pile interaction process in real time, wherein the vehicle-side status data includes a timestamp and a rated single cell voltage, a rated total voltage, a single cell voltage, a single cell temperature, a current, and a remaining power.

[0013] Optionally, the preprocessing of the charging status data specifically includes:

[0014] Performing interpolation or deletion operations on repeated values ​​and missing values ​​in the charging state data to obtain cleaned data;

[0015] Deleting outliers in the cleaned data to obtain normal data;

[0016] The normal data is normalized to obtain normalized data.

[0017] Optionally, the training process of the abnormal warning model includes:

[0018] Construct an initial abnormal warning model, which is used to issue warnings for different abnormal types. Each abnormal type determines whether an abnormality occurs based on a preset warning threshold.

[0019] Acquire historical data and use it as a training set, wherein the training set includes historical pile-end state data and historical vehicle-end state data during the vehicle-pile interaction process;

[0020] The preset alarm threshold is adjusted according to the change trend of the historical data to obtain a trained abnormal warning model.

[0021] Optionally, the normalized data is input into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal, including:

[0022] Based on a predefined abnormal type, obtaining corresponding type data from the normalized data;

[0023] Performing corresponding type of calculation on the type of data to obtain data to be tested;

[0024] The data to be tested is compared with the alarm threshold corresponding to the abnormal type. If the data to be tested exceeds the set alarm threshold, it is determined that the current charging state of the electric vehicle and / or the charging pile is abnormal.

[0025] Optionally, if the current charging state of the electric vehicle and / or the charging pile is abnormal, outputting a first warning signal includes:

[0026] Further determining the threshold interval of the data to be tested according to the multiple alarm thresholds, wherein two adjacent alarm thresholds form a threshold interval, and different threshold intervals represent different risk levels of the same abnormality type;

[0027] Determining the risk level of the abnormal type according to the threshold interval of the data to be tested;

[0028] According to the pre-stored correspondence between the risk level of the abnormal type and the protection scheme, the risk level of the current abnormal type and the corresponding protection scheme are output.

[0029] Optionally, the predefined abnormality types include charging pile charging abnormality types and vehicle charging abnormality types, wherein:

[0030] The vehicle charging abnormality types include: too high temperature difference, high battery temperature, too high single cell voltage, too low single cell voltage, too high voltage difference, too high total voltage, soc overcharge, soc jump, too high temperature rate, and too high voltage rate;

[0031] The charging pile charging abnormality types include: over-temperature and over-current.

[0032] Optionally, the method further comprises:

[0033] In response to a first control instruction and a second control instruction of a central management system, the operating parameters of the charging pile during the charging process are automatically adjusted; wherein the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal.

[0034] According to a second aspect of the present invention, the present invention further provides a charging anomaly detection device, which is installed in a charging pile, and the device includes:

[0035] A data acquisition module, used to acquire charging status data in real time when the electric vehicle is in a charging state, wherein the charging status data includes first status data of the charging pile during the charging process and second status data of the vehicle during the vehicle-pile interaction process;

[0036] A data processing module, used for preprocessing the charging state data to obtain normalized data;

[0037] An abnormal warning module is used to input the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, a first warning signal is output, and the first warning signal includes the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and alarm threshold;

[0038] A data sending module is used to send the first warning signal to the central management system, and send the normalized data to the cloud, so that the cloud can detect the current charging status of the electric vehicle and / or charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system.

[0039] Optionally, the device further includes a parameter adjustment module, which is used to automatically adjust the operating parameters of the charging pile during the charging process in response to the first control instruction and the second control instruction of the central management system; wherein the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal.

[0040] Through the technical solution of the present invention, the following technical effects can be obtained:

[0041] (1) The charging anomaly detection method and device provided by the present invention greatly shortens the response time of the entire detection process because data processing and analysis occur at the edge of the charging pile, and can respond to various situations more quickly. At the same time, edge computing also reduces dependence on central servers and reduces the overall operation and maintenance costs of the system.

[0042] (2) The present invention can detect and prevent potential safety hazards in real time, such as circuit overheating, excessive current, etc., to ensure the safety of the charging process and enhance the safety of charging; predictive equipment maintenance based on predicted abnormal risks can reduce sudden equipment failures and related maintenance costs;

[0043] (3) The present invention makes the charging process of electric vehicles smoother through real-time data analysis and intelligent detection and adjustment, avoiding charging interruptions caused by problems such as battery thermal runaway or equipment failure, thereby improving charging efficiency; further providing electric vehicles with a more efficient and safe charging solution, while bringing significant economic benefits to operators;

[0044] (4) The abnormal warning model / machine learning model provided by the present invention can self-learn and optimize according to new data, so that after collecting more data, its performance can be continuously improved and the prediction accuracy can be improved;

[0045] (5) By sending charging status data and analysis results to a central management system, the operation or maintenance team can use these real-time feedback to make decisions and improve operational efficiency and equipment utilization;

[0046] (6) The present invention saves all relevant data of each charging process, providing rich data for future analysis and learning, which helps to further improve the performance of the entire charging system and optimize the operation strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 A flow chart of a charging abnormality detection method provided by an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of the installation principle of the charging abnormality detection device provided by an embodiment of the present invention;

[0050] Figure 3 A diagram showing the indicator light status after the abnormal charging detection device provided by an embodiment of the present invention is correctly installed and deployed;

[0051] Figure 4 A flow chart of a training method for the abnormal warning model provided in an embodiment of the present invention;

[0052] Figure 5 A flow chart of a method for determining whether an abnormality occurs in the current charging state of the electric vehicle and / or the charging pile provided in an embodiment of the present invention;

[0053] Figure 6 A flow chart of a method for outputting a first warning signal if an abnormality occurs in the current charging state of the electric vehicle and / or the charging pile provided in an embodiment of the present invention;

[0054] Figure 7 A schematic diagram of the structure of a charging abnormality detection device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] It should be noted that the terms "first", "second", "third" and "fourth" in the specification and claims of this application are used to distinguish different objects rather than to describe a specific order. The terms "including" and "having" in the embodiments of this application and any variations thereof are intended to cover non-exclusive inclusions.

[0057] The optional embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0058] According to one embodiment of the present invention, Figure 1 A flow chart of a charging abnormality detection method provided by an embodiment of the present invention is as follows: Figure 1 As shown, an embodiment of the present invention provides a charging anomaly detection method, which is applied to a charging anomaly detection device, and the charging anomaly detection device runs in a charging pile, and the method includes:

[0059] S200, when the electric vehicle is in a charging state, acquiring charging state data in real time, wherein the charging state data includes pile end state data during the charging process and vehicle end state data during the vehicle-pile interaction process;

[0060] As an optional embodiment, before step S200, the method further includes:

[0061] S100, checking whether the power supply status (POWER), network connection status (NET), and system operation status (RUN) of the charging abnormality detection device are normal.

[0062] Specifically, before using the charging anomaly detection device to perform charging anomaly detection, the charging anomaly detection device needs to be installed in the charging pile to enable it to work normally. Figure 2 The schematic diagram of the installation principle of the charging abnormality detection device provided by the embodiment of the present invention is shown; Figure 2 As shown, in this embodiment, the specific installation method of the charging abnormality detection device is as follows:

[0063] (1) Connect the control power supply +12V / GND inside the charging pile to the +12V / GND of the abnormal charging detection device in parallel;

[0064] (2) Find the connection ports of S+ and S- of the charging gun lines A and B inside the charging pile, and lead out two communication lines from S+ and S- in parallel to the CAN-H and CAN-L of the charging abnormality detection device. The S+ of gun A is connected to CAN-H1, and the S- is connected to CAN-L1. The S+ of gun B is connected to CAN-H2, and the S- is connected to CAN-L2.

[0065] (3) Connect the 4G antenna to the antenna interface of the charging abnormality detection device and lead it out of the charging pile.

[0066] (4) After completing the above wiring, use the beep function of the multimeter to confirm the correctness of the wiring. Do not make mistakes or short circuits.

[0067] (5) After completing the above steps (1) to (4), restore power to the charging piles in turn. Figure 3 FIG. 2 shows the indicator light status diagram after the charging abnormality detection device provided by the embodiment of the present invention is correctly installed and deployed. Figure 3 As shown, after power-on is completed, if the POWER light of the charging abnormality detection device is always on, the RUN light is always flashing, and the NET light is flashing just after power-on, and turns to always on after communication is normal (about 2 minutes), it means that the charging abnormality detection device is installed normally.

[0068] (6) After installing the module, check whether the tools and wire connectors are left inside the charging pile.

[0069] In step S200, different types of sensors can be set inside the charging pile to collect the required data, and then a low-latency communication protocol, such as MQTT and CoAP, is used to transmit the data to an edge computing node, which is a small computing device, such as a microserver or a high-performance embedded system, capable of processing a large amount of real-time data. In other words, the charging detection device includes a sensor and an edge computing node, the sensor is used to execute step S200, and the edge computing node is used to execute steps S400-S800.

[0070] As an optional embodiment, the real-time acquisition of charging status data includes:

[0071] Start the temperature sensor, voltage sensor and current sensor installed in the charging pile to obtain the pile-end status data in real time, and the pile-end status data includes the temperature, total output voltage, total output current, power, and insulation signal of the charging pile; and start the BMS component installed in the charging pile to obtain the vehicle-end status data in real time, and the vehicle-end status data includes the current timestamp and the rated single cell voltage, rated total voltage, single cell voltage, single cell temperature, current, and remaining capacity (soc).

[0072] In this embodiment, BMS components and various sensors, such as current sensors, voltage sensors and temperature sensors, are installed in the charging pile before detection. The temperature, voltage and current of the charging pile are collected in real time through various sensors, and the power, insulation signal and the like of the charging pile can be further obtained. The type of data obtained can be set according to actual conditions; at the same time, the interaction data between the vehicle end and the pile end can be collected and uploaded to the edge computing node through BMS components in accordance with GB / T 27930. A fast-response data storage solution is installed on the edge computing node, such as Time-series databases (such as InfluxDB), which can store the obtained data.

[0073] S400, preprocessing the charging state data to obtain normalized data;

[0074] In this step, since the acquired charging status data formats are different and there are null values, invalid values, etc., the charging status data is preprocessed to unify the format to facilitate subsequent calculations. As an optional embodiment, the preprocessing of the charging status data specifically includes:

[0075] Performing interpolation or deletion operations on repeated values ​​and missing values ​​in the charging state data to obtain cleaned data;

[0076] Deleting outliers in the cleaned data to obtain normal data;

[0077] The normal data is normalized to obtain normalized data.

[0078] S600, inputting the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, outputting a first warning signal, the first warning signal including the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and the corresponding alarm threshold;

[0079] In this step, the abnormal warning model calculates the input real-time data according to the algorithm, so as to determine the charging status of the vehicle and issue an alarm. The present invention believes that when most faults occur, parameters such as voltage, temperature and their rate of change can be reflected. This idea is centered on thresholds and rates, and provides real-time warnings to vehicles at three levels. Data is transmitted through sensors, with two data pieces as input each time, and the instantaneous rate of change of the data is calculated using timestamps, as well as voltage differences, temperature differences, overvoltage and undervoltage risks, etc. calculated using other information. The abnormal warning model mainly considers warnings based on three aspects: consistency between single cells, basic vehicle data, and short-term transient nature of data, as follows:

[0080] (1) Measure the consistency between single cells and use data to warn about temperature, voltage difference, voltage stability, etc. between cells;

[0081] (2) Measure the basic data of the vehicle, mainly to warn about the vehicle's single cell temperature and total voltage;

[0082] (3) Measures the short-term transient nature of the data, mainly alerting the data from the temperature change rate and voltage change rate between the previous frame data and the current frame data.

[0083] Before using the abnormal warning model for warning, the abnormal warning model can be "diagnosed" by combining historical data and information on the fault triggering mechanism of the faulty battery pack with expert systems, various diagnostic methods and fault knowledge graphs, and further calibrated by adjusting the threshold. As an optional embodiment, Figure 4 The flow chart of the training method of the abnormal warning model is shown as follows: Figure 4 As shown, the training process includes:

[0084] S610, constructing an initial abnormality warning model, where the initial abnormality warning model is used to issue alarms for different abnormality types, and for each abnormality type, determining whether an abnormality currently occurs based on a preset alarm threshold;

[0085] S620, acquiring historical data and using it as a training set, wherein the training set includes historical pile-end state data and historical vehicle-end state data during the vehicle-pile interaction process;

[0086] S630, adjusting the preset alarm threshold according to the change trend of the historical data to obtain a trained abnormal warning model. For example, a change trend curve of the historical data is obtained based on the data collected from December 1 to December 3, and the preset alarm threshold is adjusted according to the curve to obtain an alarm threshold with higher prediction accuracy, thereby obtaining a trained abnormal warning model.

[0087] As an optional embodiment, Figure 5 A flow chart of a method for determining whether an abnormality occurs in the current charging state of the electric vehicle and / or the charging pile is shown. Figure 5 As shown, the normalized data is input into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal, including:

[0088] S640, based on predefined abnormality types, obtaining corresponding type data from the normalized data; the predefined abnormality types include charging pile abnormality types and vehicle charging abnormality types, wherein:

[0089] The vehicle charging abnormality types include: too high temperature difference, high battery temperature, too high single cell voltage, too low single cell voltage, too high voltage difference, too high total voltage, soc overcharge, soc jump, too high temperature rate, too high voltage rate; the charging pile charging abnormality types include: too high temperature, too large current, etc. It should be noted that the above-mentioned various charging abnormality types are just common abnormalities listed, and are not limited to the above-mentioned abnormality types.

[0090] In this embodiment, the abnormal warning model includes multiple real-time warning modules corresponding to the abnormal types, and each warning module is responsible for different alarm tasks, for example: temperature_diff module, temperature_high module, temperature_rise module, Voltage_all module, Voltage_rise module, Voltage_cell_high module, Voltage_cell_low module, Voltage_diff module, Soc_jump module, and Soc_longhigh module.

[0091] S650, performing a corresponding type of calculation on the type of data to obtain data to be tested;

[0092] In this step, after the normalized data is input into the abnormal warning model, the real-time data stream is processed based on a streaming data processing framework, such as Apache Kafka and Apache Flink, so as to divert the real-time data required for different abnormal types. For example, the excessive temperature difference abnormal type requires the highest temperature and the lowest temperature in the single cell; the single cell voltage only requires the current real-time voltage of each single cell.

[0093] Since some abnormal types require the calculation of parameters such as temperature difference, pressure difference, and rate, the acquired charging status data needs to be further processed. For example, the difference between the highest temperature and the lowest temperature in the single battery is calculated to obtain the current vehicle temperature difference value; or the current temperature rate is calculated based on the temperature corresponding to the current timestamp and the previous timestamp.

[0094] S660: Compare the data to be tested with the alarm threshold corresponding to the abnormal type. If the data to be tested exceeds the set alarm threshold, it is determined that the current charging state of the electric vehicle and / or the charging pile is abnormal.

[0095] In this step, different alarm thresholds are set based on different abnormality types. After the measured data is calculated in step S650, such as the temperature rate, the real-time temperature rate is compared with multiple alarm thresholds corresponding to the abnormal type of excessive temperature rate. If the real-time temperature rate is less than the set alarm threshold (equivalent to the subsequent first alarm threshold), the current real-time temperature rate is considered normal, that is, no abnormality has occurred; if the real-time temperature rate is greater than the set alarm threshold (equivalent to the subsequent first alarm threshold), the current real-time temperature rate is considered to be too high, that is, an abnormality has occurred.

[0096] It should be noted that during the comparison process, data comparison can be performed based on the same battery type and / or the same type of vehicle, which can achieve higher prediction accuracy. For example, the vehicle's unique identification VIN is similar to a personal ID card, and historical orders, specific parameters, etc. can be identified through the VIN.

[0097] Further, after determining that an abnormality has occurred, the risk level of the abnormality is determined according to the alarm threshold. Specifically, each abnormality type may correspond to multiple alarm thresholds. In this embodiment, the alarm threshold includes a first alarm threshold, a second alarm threshold, and a third alarm threshold. The first alarm threshold is less than the second alarm threshold, and the second alarm threshold is less than the third alarm threshold. Figure 6 A flowchart of a method for outputting a first warning signal if an abnormality occurs in the current charging state of the electric vehicle and / or the charging pile is shown. As shown in 6, if an abnormality occurs in the current charging state of the electric vehicle and / or the charging pile, outputting a first warning signal includes:

[0098] S670, further determine the threshold interval of the data to be tested based on the multiple alarm thresholds, wherein two adjacent alarm thresholds generate a threshold interval, and different threshold intervals represent different risk levels for the same abnormality type. In this embodiment, the risk level includes three levels, such as low risk, medium risk and high risk.

[0099] S680, determine the risk level of the abnormal type according to the threshold range of the data to be tested; when the data to be tested is between the first alarm threshold and the second alarm threshold, it is a low risk; when the data to be tested is between the second alarm threshold and the third alarm threshold, it is a medium risk; when the data to be tested is greater than the third alarm threshold, it is a high risk.

[0100] S690: Output the risk level of the current abnormality type and the corresponding protection plan according to the pre-stored correspondence between the risk level of the abnormality type and the protection plan.

[0101] The data preprocessing and analysis of the above steps S640-S690 are all completed on the edge computing node, reducing the delay and bandwidth consumption of data transmission, allowing the system to respond to security issues more quickly and accurately.

[0102] In practical applications, the input and output parameters of the abnormal warning model are as follows:

[0103] (1) Enter the specific parameters of the two data:

[0104] Timestamp, time (2021-8-2 15:16:40), rated single cell voltage, rated total voltage, charging start time, single cell voltage, single cell temperature, current, vin, Soc;

[0105] (2) The output parameters are shown in Table 1 below:

[0106] Table 1

[0107]

[0108]

[0109] The basic values ​​of the above output parameters, except for the alarm, are used to be sent to the central management system to facilitate the charging pile management personnel to view the data. Of course, the basic values ​​may not be output, depending on the actual situation.

[0110] S800, sending the first warning signal to the central management system for alarm, and sending the normalized data to the cloud, so that the cloud can detect the current charging status of the electric vehicle and / or charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system.

[0111] Before step S800, the method further includes: when the output times of the first warning signal reaches a threshold, the first warning signal is sent to the central management system. In other words, the result of each time is stored, and an alarm is sent to the central management system only when the number of consecutive output times exceeds N, so as to improve the accuracy of the warning.

[0112] In step S800, the machine learning model is generated based on any algorithm such as decision tree, random forest or neural network. After acquiring real-time data, the embodiment of the present invention not only performs anomaly detection locally on the charging pile, but also performs big data analysis in the cloud, so as to avoid being unable to respond to security issues when the cloud server crashes. Even if the cloud crashes, the charging pile can respond locally, and the response time is greatly shortened, so that various situations can be dealt with more quickly.

[0113] As an optional embodiment, the method further comprises:

[0114] S900, in response to the first control instruction and the second control instruction of the central management system, automatically adjust the operating parameters of the charging pile during the charging process; wherein, the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal.

[0115] In this embodiment, the operating parameters include the output voltage, output current and temperature of the charging pile. By adjusting the operating parameters, including making the charging pile continue to charge normally, eliminating abnormalities of the charging pile, or stopping charging, shutting down, etc., potential safety hazards such as circuit overheating, excessive current, etc. are prevented to ensure a safe charging process.

[0116] In practical applications, the present invention can design an intelligent safety detection circuit to implement the above steps S800-S900. The functions of the intelligent safety detection circuit include: integrating current, voltage and temperature protection mechanisms to ensure the safety of the charging process; adding an early warning system, which can automatically stop charging and issue an alarm when any potential problem or fault is detected.

[0117] As an optional embodiment, the method further includes: sending the charging status data and the first warning signal to the cloud so as to optimize the machine learning model and improve the prediction accuracy of the machine learning model.

[0118] In actual application, when charging is completed or the user terminates charging, the central management system saves all relevant data of the charging process for future analysis and learning; then the charging pile enters standby mode, waiting for the next charging task.

[0119] Example 1:

[0120] The charging anomaly detection process of electric vehicle charging stations: install charging anomaly detection devices at charging stations to collect real-time charging data; deploy edge computing nodes to preprocess data and transmit it to the central management system through a secure network; apply abnormal warning models to timely identify potential faults or safety risks; automatically disconnect abnormal charging piles based on intelligent safety detection circuits to prevent accidents; and continuously optimize the performance of the entire detection process through user feedback.

[0121] Example 2:

[0122] Charging anomaly detection process at public charging stations: In a public electric vehicle charging station, each charging pile is equipped with sensors and edge computing nodes; the sensors monitor parameters such as current, voltage and temperature in real time during the charging process, and send the data to the edge computing nodes; the edge computing nodes first perform preprocessing operations such as data cleaning, outlier detection and data standardization, and then input the preprocessed data into a pre-trained anomaly warning model, and use the streaming data processing framework to perform real-time analysis on the preprocessed data; when an anomaly or potential circuit problem is detected, the intelligent safety detection circuit automatically adjusts the charging parameters or stops charging, and automatically triggers an early warning.

[0123] At the same time, the edge computing nodes send key data and analysis results to the central management system for the operation or maintenance team to make decisions; the system also self-learns based on new data to optimize the performance of the machine learning model.

[0124] Example 3:

[0125] The process of detecting charging anomalies in a home charging pile: In a home electric vehicle charging pile, the charging pile is equipped with sensors and edge computing nodes. The sensors are responsible for real-time monitoring of key parameters in the charging process, such as current, voltage, and temperature, and sending the data to the edge computing node. The edge computing node first cleans the data, detects anomalies, and normalizes the data. After that, the data is input into the anomaly warning model, and the pre-processed data is analyzed in real time using the streaming data processing framework. If an anomaly or potential circuit problem is detected, the intelligent safety detection circuit will automatically adjust the charging parameters or stop charging, and trigger an early warning.

[0126] In addition, the edge computing node also sends key data and analysis results to the user's mobile phone application or home central management system so that the user can monitor the charging status in real time and make decisions when necessary. Similarly, the central management system will self-learn and optimize based on the newly collected data to improve prediction accuracy and user experience.

[0127] The charging anomaly detection method provided by the present invention greatly shortens the response time of the entire detection process because data processing and analysis occur at the edge of the charging pile, and can respond to various situations more quickly; at the same time, edge computing also reduces dependence on central servers and reduces the overall operation and maintenance costs of the system; the present invention can detect and prevent potential safety hazards in real time, such as circuit overheating, excessive current and other problems, to ensure the safety of the charging process and enhance the safety of charging; predictive equipment maintenance based on predicted abnormal risks can reduce sudden equipment failures and related maintenance costs; the present invention makes the charging process of electric vehicles smoother through real-time data analysis and intelligent detection and adjustment, avoiding accidents caused by problems such as battery thermal runaway or equipment failure. Charging interruption can improve charging efficiency; further provide electric vehicles with more efficient and safe charging solutions, while bringing significant economic benefits to operators; the abnormal warning model / machine learning model provided by the present invention can self-learn and optimize according to new data, so that after collecting more data, it can continuously improve its performance and improve prediction accuracy; by sending charging status data and analysis results to the central management system, the operation or maintenance team can use these real-time feedback to make decisions, improve operational efficiency and equipment utilization; the present invention saves all relevant data of each charging process, provides rich data for future analysis and learning, and helps to further improve the performance of the entire charging system and optimize operational strategies.

[0128] Based on a charging anomaly detection method provided in the above embodiment, this embodiment provides a device corresponding to the charging anomaly detection method. For the same content, please refer to the above method embodiment, which will not be described in detail in this embodiment. Figure 7 A schematic diagram of the structure of a charging abnormality detection device is shown. Figure 7 As shown, the charging abnormality detection device is installed in the charging pile, and the charging abnormality detection device includes:

[0129] The data acquisition module 100 is used to acquire charging status data in real time when the electric vehicle is in a charging state, wherein the charging status data includes first status data of the charging pile during the charging process and second status data of the vehicle during the vehicle-pile interaction process;

[0130] The data processing module 200 is used to pre-process the charging state data to obtain normalized data;

[0131] The abnormal warning module 300 is used to input the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, a first warning signal is output, and the first warning signal includes the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and alarm threshold;

[0132] The data sending module 400 is used to send the first warning signal to the central management system, and send the normalized data to the cloud, so that the cloud can detect the current charging status of the electric vehicle and / or charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system.

[0133] As an optional embodiment, the abnormal warning module 300 includes:

[0134] The data distribution unit 310 is used to obtain corresponding type data from the normalized data based on a predefined abnormal type;

[0135] A data calculation unit 320, configured to perform a corresponding type of calculation on the type of data to obtain data to be tested;

[0136] The abnormality detection unit 330 is used to compare the data to be tested with the alarm threshold corresponding to the abnormality type. If the data to be tested exceeds the set alarm threshold, it is determined that the current charging state of the electric vehicle and / or charging pile is abnormal.

[0137] The anomaly detection unit 330 is also used to further determine the threshold interval of the data to be tested based on the multiple alarm thresholds, wherein two adjacent alarm thresholds generate a threshold interval, and different threshold intervals represent different risk levels of the same anomaly type. In this embodiment, the risk level includes three levels, such as low risk, medium risk and high risk.

[0138] The anomaly detection unit 330 is also used to determine the risk level of the anomaly type according to the threshold range of the data to be tested; when the data to be tested is between the first alarm threshold and the second alarm threshold, it is a low risk; when the data to be tested is between the second alarm threshold and the third alarm threshold, it is a medium risk; when the data to be tested is greater than the third alarm threshold, it is a high risk.

[0139] The alarm unit 340 is used to output the risk level of the current abnormal type and the corresponding protection plan according to the pre-stored correspondence between the risk level of the abnormal type and the protection plan.

[0140] As an optional embodiment, the charging anomaly detection device further includes a parameter adjustment module 500, which is used to automatically adjust the operating parameters of the charging pile during the charging process in response to the first control instruction and the second control instruction of the central management system; wherein the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal.

[0141] As an optional embodiment, the charging abnormality detection device further includes a feedback optimization module 600, which is used to send the charging status data and the first warning signal to the cloud so as to optimize the machine learning model.

[0142] The charging anomaly detection device provided by the present invention greatly shortens the response time of the entire detection process because data processing and analysis occur at the edge of the charging pile, and can respond to various situations more quickly; at the same time, edge computing also reduces dependence on central servers and reduces the overall operation and maintenance costs of the system; the present invention can detect and prevent potential safety hazards in real time, such as circuit overheating, excessive current and other problems, to ensure the safety of the charging process and enhance the safety of charging; predictive equipment maintenance based on predicted abnormal risks can reduce sudden equipment failures and related maintenance costs; the present invention makes the charging process of electric vehicles smoother through real-time data analysis and intelligent detection and adjustment, avoiding accidents caused by problems such as battery thermal runaway or equipment failure. Charging interruption can improve charging efficiency; further provide electric vehicles with more efficient and safe charging solutions, while bringing significant economic benefits to operators; the abnormal warning model / machine learning model provided by the present invention can self-learn and optimize according to new data, so that after collecting more data, it can continuously improve its performance and improve prediction accuracy; by sending charging status data and analysis results to the central management system, the operation or maintenance team can use these real-time feedback to make decisions, improve operational efficiency and equipment utilization; the present invention saves all relevant data of each charging process, provides rich data for future analysis and learning, and helps to further improve the performance of the entire charging system and optimize operational strategies.

[0143] Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described device and equipment can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0144] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative.

[0145] Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.

[0146] It should be understood that the size of the serial numbers of each step in the content of the invention and the embodiments of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

Claims

1. A charging abnormality detection method, characterized in that: Applied to a charging anomaly detection device, the charging anomaly detection device operates in a charging pile, and the method includes: When the electric vehicle is in a charging state, the charging state data is acquired in real time, wherein the charging state data includes the charging pile state data during the charging process and the vehicle state data during the vehicle-pile interaction process; Preprocessing the charging state data to obtain normalized data; Input the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, output a first warning signal, the first warning signal includes the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and the corresponding alarm threshold; Sending the first warning signal to a central management system for alarm, and sending the normalized data to the cloud, so that the cloud can detect the current charging status of the electric vehicle and / or charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system; In response to the first control instruction and the second control instruction of the central management system, automatically adjusting the operating parameters of the charging pile during the charging process; Wherein, the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal; The predefined abnormality types include charging pile charging abnormality types and vehicle charging abnormality types, wherein: The vehicle charging abnormality types include: too high temperature difference, high battery temperature, too high single cell voltage, too low single cell voltage, too high voltage difference, too high total voltage, soc overcharge, soc jump, too high temperature rate, and too high voltage rate; The charging pile charging abnormality types include: over-temperature and over-current.

2. The charging abnormality detection method according to claim 1, characterized in that: The real-time acquisition of charging status data includes: Start the temperature sensor, voltage sensor and current sensor installed in the charging pile to obtain the pile end status data in real time, wherein the pile end status data includes the temperature, total output voltage, total output current, power and insulation signal of the charging pile; The BMS component installed in the charging pile is started to obtain the vehicle-side status data in the vehicle-pile interaction process in real time, wherein the vehicle-side status data includes a timestamp and a rated single cell voltage, a rated total voltage, a single cell voltage, a single cell temperature, a current, and a remaining power.

3. The charging abnormality detection method according to claim 1, characterized in that: The preprocessing of the charging status data specifically includes: Performing interpolation or deletion operations on repeated values ​​and missing values ​​in the charging state data to obtain cleaned data; Deleting outliers in the cleaned data to obtain normal data; The normal data is normalized to obtain normalized data.

4. The charging abnormality detection method according to claim 1, characterized in that: The training process of the abnormal warning model includes: Construct an initial abnormal warning model, which is used to issue alarms for different abnormal types. Each abnormal type determines whether an abnormality occurs based on a preset alarm threshold. Acquire historical data and use it as a training set, wherein the training set includes historical pile-end state data and historical vehicle-end state data during the vehicle-pile interaction process; The preset alarm threshold is adjusted according to the change trend of the historical data to obtain a trained abnormal warning model.

5. The charging abnormality detection method according to claim 1, characterized in that: Inputting the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal includes: Based on a predefined abnormal type, obtaining corresponding type data from the normalized data; Performing corresponding type of calculation on the type of data to obtain data to be tested; The data to be tested is compared with the alarm threshold corresponding to the abnormal type. If the data to be tested exceeds the set alarm threshold, it is determined that the current charging state of the electric vehicle and / or the charging pile is abnormal.

6. The charging abnormality detection method according to claim 5, characterized in that: If the current charging state of the electric vehicle and / or the charging pile is abnormal, outputting a first warning signal includes: Further determining the threshold interval of the data to be tested according to the multiple alarm thresholds, wherein two adjacent alarm thresholds form a threshold interval, and different threshold intervals represent different risk levels of the same abnormality type; Determining the risk level of the abnormal type according to the threshold interval of the data to be tested; According to the pre-stored correspondence between the risk level of the abnormal type and the protection scheme, the risk level of the current abnormal type and the corresponding protection scheme are output.

7. A charging abnormality detection device, characterized in that: The charging abnormality detection device is installed in the charging pile, and the device includes: A data acquisition module, used to acquire charging status data in real time when the electric vehicle is in a charging state, wherein the charging status data includes first status data of the charging pile during the charging process and second status data of the vehicle during the vehicle-pile interaction process; A data processing module, used for preprocessing the charging state data to obtain normalized data; An abnormal warning module is used to input the normalized data into a pre-trained abnormal warning model to determine whether the current charging state of the electric vehicle and / or the charging pile is abnormal; if the current charging state of the electric vehicle and / or the charging pile is abnormal, a first warning signal is output, and the first warning signal includes the risk level of the current abnormal type and the corresponding protection plan; wherein the abnormal warning model performs abnormality detection according to the pre-defined abnormal type and alarm threshold; A data sending module, used to send the first warning signal to a central management system, and send the normalized data to a cloud, so that the cloud can detect the current charging status of the electric vehicle and / or the charging pile through a pre-trained machine learning model, output a second warning signal and send it to the central management system; A parameter adjustment module, configured to automatically adjust the operating parameters of the charging pile during charging in response to a first control instruction and a second control instruction of the central management system; wherein the first control instruction is generated according to the first warning signal, and the second control instruction is generated according to the second warning signal; The abnormality type classification module is used for the pre-defined abnormality types including charging pile charging abnormality types and vehicle charging abnormality types, wherein the vehicle charging abnormality types include: temperature difference is too high, battery temperature is too high, single cell voltage is too high, single cell voltage is too low, voltage difference is too high, total voltage is too high, soc overcharge, soc jump, temperature rate is too high, voltage rate is too high; the charging pile charging abnormality types include: temperature is too high, current is too high.

Images