Intelligent remote regulation and control method and system applied to alternating current charging pile

Abstract

The invention discloses an intelligent remote regulation and control method and system applied to an alternating current charging pile, and the method comprises the steps: monitoring the operation condition of a target alternating current charging pile during working, and constructing a real-time operation condition feature sequence; constructing an anomaly detection model, and inputting a real-time operation condition feature sequence to detect abnormal operation data under the current operation condition; a risk threshold value is preset, whether an operation risk exists or not is analyzed, and risk grade evaluation and management and control strategy making are carried out; risk evolution characteristics of various risk events are analyzed, a risk prediction model is constructed, whether the abnormal operation condition of the target alternating current charging pile evolves to the risk events or not is analyzed, and maintenance early warning prompt is carried out; and scene monitoring video stream data during charging is acquired, malicious behavior detection and analysis are performed on the charging scene of the target alternating current charging pile, and if malicious behaviors exist, a regulation and control strategy is acquired to control the target alternating current charging pile to warn a malicious behavior target. The operation safety of the charging pile is improved, and charging safety accidents are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of remote control of AC charging piles, and in particular to an intelligent remote control method and system applied to AC charging piles. Background Art

[0002] With the rapid development of the electric vehicle industry, the demand for AC charging piles, as an important part of electric vehicle charging facilities, has been rising year by year. Existing charging piles mainly adopt a single-machine local control method, lack a unified intelligent management platform, and cannot effectively realize remote monitoring and control of the charging process. Traditional charging piles have many shortcomings in user experience, grid energy scheduling, and equipment operation efficiency. For the operating status and potential faults of the equipment, traditional charging piles mainly rely on regular manual inspections, and cannot achieve real-time status perception and fault warnings, which can easily cause the charging pile to shut down for a long time or operate abnormally, affecting the overall operational efficiency. At the same time, when users use charging piles for charging, there may be other non-users who perform malicious acts on the charging piles, such as maliciously unplugging the charging gun, etc., which may lead to charging accidents.

[0003] Therefore, how to better monitor the operation of AC charging piles and control possible malicious behaviors to improve the operation safety of charging piles is an urgent problem to be solved. Summary of the invention

[0004] The present invention overcomes the defects of the prior art and provides an intelligent remote control method and system for an AC charging pile, an important purpose of which is to improve the operating safety of the charging pile and avoid charging safety accidents.

[0005] To achieve the above-mentioned purpose, the first aspect of the present invention provides an intelligent remote control method applied to an AC charging pile, comprising:

[0006] Monitor the operating status of the target AC charging pile when it is working, obtain operating status monitoring information, pre-process the operating status monitoring information, and construct a real-time operating status feature sequence;

[0007] Building an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information;

[0008] Preset risk thresholds, compare the abnormal detection information with the preset risk thresholds, analyze whether there are operational risks, and conduct risk level assessment and control strategy formulation;

[0009] Obtain historical risk examples of AC charging piles, analyze the risk evolution characteristics of various risk events, build a risk prediction model and analyze whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provide maintenance warning prompts;

[0010] The scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target.

[0011] In this solution, the operation status of the target AC charging pile is monitored, the operation status monitoring information is obtained, the operation status monitoring information is preprocessed, and a real-time operation status feature sequence is constructed, which specifically includes:

[0012] Arrange a sensor array in the target AC charging pile, monitor the operating status of the target AC charging pile when it is working based on the sensor network, and obtain operating status monitoring information;

[0013] Calculate the data standard deviation according to the acquired operation status monitoring information, compare the calculated data standard deviation with the preset standard deviation range, and define the data corresponding to the data standard deviation that is not within the preset standard deviation range as an abnormal value;

[0014] The operation status monitoring data defined as outliers are deleted, and a linear interpolation algorithm is introduced to supplement the missing values ​​of the operation status monitoring data after outliers are deleted, so as to obtain preprocessing result information;

[0015] Performing data dimension reduction on the preprocessing result information based on the PCA algorithm, calculating the principal component score of each operating status monitoring data, and selecting the operating status monitoring data corresponding to the principal component score greater than a preset threshold as the principal component feature;

[0016] The principal component analysis results are obtained by using one-hot encoding to encode the principal component features, and the encoded principal component features are processed in time series to generate a time series feature sequence to obtain a real-time operating status feature sequence.

[0017] In this solution, the anomaly detection model is constructed based on the generative adversarial network, the real-time operating status feature sequence is input, the abnormal operating data under the current operating status is detected, and the anomaly detection information is obtained, which specifically includes:

[0018] Acquire historical normal operation status monitoring information of the target AC charging pile under different operation modes based on data retrieval, classify the historical normal operation status monitoring information according to the operation mode, and obtain first classification result information;

[0019] Performing principal component analysis based on the first classification result information using a PCA algorithm to obtain principal component characteristics of each operating mode category, and calculating the operating state transition probability based on the principal component characteristics of each operating mode category;

[0020] Generate a state transfer matrix according to the calculated operating state transition probability, construct a state transfer space in combination with the first classification result information, build an anomaly detection model through a generative adversarial network, and perform model training through the state transfer space;

[0021] Obtain the real-time operating status feature sequence, input it into the trained anomaly detection model for analysis, match the state transfer matrix through the state transfer space, and the generator reconstructs the sequence according to the matched state transfer matrix to obtain the reconstructed sequence;

[0022] The reconstructed sequence is input into the discriminator, and combined with the matching state transfer matrix, it is determined whether the current reconstructed sequence is accepted. If the determination result is not accepted, the next reconstructed sequence of the generator is obtained for determination until it meets the state transfer rule of the matching state transfer matrix, and the final reconstructed sequence is output;

[0023] A reconstruction error between the final reconstruction sequence and the real-time operation status feature sequence is calculated, and abnormal operation status data is detected according to the calculated reconstruction error to obtain abnormality detection information.

[0024] In this solution, the preset risk threshold is used to judge the abnormal detection information with the preset risk threshold, analyze whether there is an operational risk, and conduct risk level assessment and control strategy formulation, specifically including:

[0025] A preset risk threshold is set, and the abnormal detection information is compared with the preset risk threshold. If it is greater than the preset threshold, it indicates that there is a risk in the real-time operation state of the target AC charging pile, and risk control is performed;

[0026] Acquire abnormal detection information and perform feature extraction on the abnormal detection information, extract abnormal operation status features of the target AC charging pile in real time, and obtain abnormal operation status feature information;

[0027] Based on the abnormal operation condition characteristic information, a normal operation benchmark corresponding to the abnormal data is obtained, deviation calculation is performed to obtain the deviation, and a risk level is evaluated according to the calculated deviation value;

[0028] According to the risk level assessment result and the characteristic information of the abnormal operating condition, the corresponding risk management and control strategy is obtained through big data retrieval, and the risk management and control strategy with the highest frequency of use is selected to generate remote control instructions to control the target AC charging pile.

[0029] In this solution, the historical risk examples of AC charging piles are obtained, the risk evolution characteristics of various risk events are analyzed, a risk prediction model is constructed, and analysis is performed on whether the abnormal operation of the target AC charging pile will evolve into a risk event, and maintenance warning prompts are provided, specifically including:

[0030] Obtain historical risk instances of AC charging piles by means of big data, extract features of the historical risk instances to obtain feature information of the historical risk instances, and generate a time series feature sequence of each historical risk instance according to the feature information of the historical risk instances;

[0031] Calculate the Euclidean distance between the time series feature sequences of each historical risk instance, perform clustering based on the calculated Euclidean distance to obtain several clusters, and input the obtained clusters into the LSTM network for analysis;

[0032] The risk evolution characteristics are obtained by extracting features from various clusters through the LSTM network, and the risk evolution characteristics are reconstructed by feature encoding using a stacked autoencoder. The attention mechanism is introduced to weight the features to represent the importance of the risk evolution characteristics.

[0033] Obtain the reconstructed risk evolution characteristics, conduct risk level assessment on the reconstructed risk evolution characteristics, form a number of risk evolution characteristic data sets classified by risk level, build a risk prediction model and conduct model training;

[0034] Obtaining the abnormal detection result of the target AC charging pile within a unit time, inputting it into the risk prediction model, predicting the evolution of the current abnormal operation status of the target AC charging pile in a future time period, and obtaining risk prediction information;

[0035] It is determined whether the target AC charging pile has an operation risk based on the risk prediction information. If there is an operation risk, a maintenance reminder is generated and pushed to the user end to remind the user to make a maintenance appointment.

[0036] In this solution, the scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target, which specifically includes:

[0037] Based on big data retrieval methods, different malicious behavior data of charging piles during charging are collected, and features of the collected malicious behavior data are extracted to obtain malicious behavior features;

[0038] The malicious behavior level is preset, and a fuzzy comprehensive evaluation method is introduced to evaluate the malicious behavior level of the acquired malicious behavior features. Based on the malicious behavior level evaluation results, the malicious behavior features are divided into different categories of data sets to obtain several malicious behavior feature category data sets;

[0039] A malicious behavior analysis model is constructed based on a deep neural network, a malicious behavior feature category data set is used to train the malicious behavior analysis model, and a back propagation method is used to optimize and adjust model parameters to obtain a malicious behavior analysis model that meets expectations;

[0040] Based on the scene monitoring device installed on the target AC charging pile, the scene monitoring video stream data during charging is obtained and input into the malicious behavior analysis model, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile to obtain malicious behavior analysis information;

[0041] It is determined whether malicious behavior occurs when the target AC charging pile is charging according to the malicious behavior analysis information. If malicious behavior occurs, the malicious behavior level is extracted according to the malicious behavior analysis information, and a control strategy is obtained to stop the malicious behavior.

[0042] In this solution, if there is malicious behavior, the malicious behavior level is extracted according to the malicious behavior analysis information, and a control strategy is obtained to stop the malicious behavior, which specifically includes:

[0043] If malicious behavior is detected in the current charging scenario, the malicious behavior level currently detected is extracted according to the malicious behavior analysis information, and the malicious behavior level is divided into three levels: low, medium, and high;

[0044] If the current malicious behavior level is low-level malicious behavior, a warning instruction is generated based on the preset warning strategy, the target AC charging pile is remotely warned and controlled, and an early warning is pushed to the user;

[0045] If the current malicious behavior level is medium-level malicious behavior, a warning instruction is generated based on the preset warning strategy, a power-off control instruction is generated according to the preset safety protection strategy, the target AC charging pile is remotely controlled, and an early warning is pushed to the user;

[0046] If the current malicious behavior level is high-level malicious behavior, a warning instruction is generated based on the preset warning strategy, and a power-off control instruction is generated according to the preset safety protection strategy to control the target AC charging pile to issue a warning voice and perform power-off protection, and the video stream data detected with malicious behavior is retained and pushed to the target user;

[0047] After the target AC charging pile performs warning processing according to the control instruction, continue to detect whether the malicious behavior in the current charging scene disappears. If the malicious behavior still exists, increase the malicious behavior level and obtain the corresponding level of control strategy to control the target AC charging pile.

[0048] A second aspect of the present invention provides an intelligent remote control system applied to an AC charging pile, the system comprising: a memory, a processor, the memory containing an intelligent remote control method program applied to an AC charging pile, the intelligent remote control method program applied to an AC charging pile being executed by the processor to implement the following steps:

[0049] Monitor the operating status of the target AC charging pile when it is working, obtain operating status monitoring information, pre-process the operating status monitoring information, and construct a real-time operating status feature sequence;

[0050] Building an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information;

[0051] Preset risk thresholds, compare the abnormal detection information with the preset risk thresholds, analyze whether there are operational risks, and conduct risk level assessment and control strategy formulation;

[0052] Obtain historical risk examples of AC charging piles, analyze the risk evolution characteristics of various risk events, build a risk prediction model and analyze whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provide maintenance warning prompts;

[0053] The scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target.

[0054] The present invention discloses an intelligent remote control method and system for AC charging piles, including: monitoring the operating status of the target AC charging pile when it is working, and constructing a real-time operating status feature sequence; constructing an abnormal detection model, inputting the real-time operating status feature sequence to detect abnormal operating data under the current operating status; presetting a risk threshold, analyzing whether there is an operating risk and conducting risk level assessment and management strategy formulation; analyzing the risk evolution characteristics of various risk events, constructing a risk prediction model and analyzing whether the abnormal operating status of the target AC charging pile will evolve into a risk event, and performing maintenance early warning prompts; obtaining scene monitoring video stream data during charging, performing malicious behavior detection and analysis on the charging scene of the target AC charging pile, and if there is malicious behavior, obtaining a control strategy to control the target AC charging pile to warn the malicious behavior target. Improve the operating safety of the charging pile and avoid charging safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0056] Figure 1 A flow chart of an intelligent remote control method applied to an AC charging pile provided by an embodiment of the present invention;

[0057] Figure 2 A flow chart of a malicious behavior control method applied to an AC charging pile provided in one embodiment of the present invention;

[0058] Figure 3 A block diagram of an intelligent remote control system applied to an AC charging pile provided by an embodiment of the present invention;

[0059] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0060] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0061] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0062] Figure 1 A flow chart of an intelligent remote control method applied to an AC charging pile provided by an embodiment of the present invention;

[0063] like Figure 1 As shown, the present invention provides a flow chart of an intelligent remote control method applied to an AC charging pile, including:

[0064] S102, monitoring the operating status of the target AC charging pile when it is working, obtaining operating status monitoring information, preprocessing the operating status monitoring information, and constructing a real-time operating status feature sequence;

[0065] S104, constructing an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information;

[0066] S106, preset a risk threshold, compare the abnormal detection information with the preset risk threshold, analyze whether there is an operational risk, and conduct risk level assessment and control strategy formulation;

[0067] S108, obtaining historical risk instances of AC charging piles, analyzing the risk evolution characteristics of various risk events, building a risk prediction model and analyzing whether the abnormal operation status of the target AC charging pile will evolve into a risk event, and providing maintenance warning prompts;

[0068] S110, acquiring scene monitoring video stream data during charging, detecting and analyzing malicious behaviors in the charging scene of the target AC charging pile, and if malicious behaviors exist, acquiring a control strategy to control the target AC charging pile to warn and stop the malicious behaviors.

[0069] It should be noted that the present invention provides an intelligent remote control method and system for AC charging piles, which obtains various types of operation status monitoring information related to its working state by real-time monitoring of the operation status of the target AC charging pile. After preprocessing, it is sorted in a time series to form a real-time operation status feature sequence. The generative adversarial network (GAN) is used to build an anomaly detection model. The generative adversarial network effectively learns the distribution characteristics of the normal operation status of the equipment through the game mechanism between the generator and the discriminator. In practical applications, the real-time operation status feature sequence is input into the anomaly detection model to analyze whether there is abnormal data under the current operation status, and finally obtain clear abnormal detection information to quickly locate the abnormal performance of the charging pile in operation. After the abnormal detection is completed, the detected abnormal information is compared with the preset risk threshold to determine whether there is an operation risk. If there is a risk, the risk level is evaluated according to the abnormal characteristics, and the corresponding control strategy is formulated in combination with the current operation status and risk level of the charging pile, and the remote control instruction is generated to ensure the safety and operation stability of the equipment, so as to realize the real-time identification and analysis of equipment abnormalities. In order to further improve the ability to prevent risk events, the historical risk instance data of AC charging piles is obtained, the evolution characteristics of different risk events are analyzed, and a risk prediction model is constructed accordingly. Using this model, the potential development trend of the current abnormal operation of the target AC charging pile is predicted to determine whether it may evolve into a more serious risk event. If the prediction result shows that there is a significant risk, an inspection and maintenance warning prompt is automatically issued so that relevant maintenance personnel or users can take timely measures to reduce potential losses. In addition, in order to fully ensure the safety of the charging process, the malicious behavior of the target AC charging pile charging scene is detected and analyzed in combination with the scene monitoring video stream data. Through video monitoring data, it is effective to identify whether there are malicious behaviors that damage charging facilities or endanger charging safety in the charging area. Once such behavior is detected, the malicious behavior target will be warned and stopped through visual or auditory signals based on the control strategy, thereby maintaining the safety and order of the charging pile and the charging environment.

[0070] Further, in a preferred embodiment of the present invention, the monitoring target AC charging pile's operating status when it is working, obtaining operating status monitoring information, preprocessing the operating status monitoring information, and constructing a real-time operating status feature sequence specifically includes:

[0071] Arrange a sensor array in the target AC charging pile, monitor the operating status of the target AC charging pile when it is working based on the sensor network, and obtain operating status monitoring information;

[0072] Calculate the data standard deviation according to the acquired operation status monitoring information, compare the calculated data standard deviation with the preset standard deviation range, and define the data corresponding to the data standard deviation that is not within the preset standard deviation range as an abnormal value;

[0073] The operation status monitoring data defined as outliers are deleted, and a linear interpolation algorithm is introduced to supplement the missing values ​​of the operation status monitoring data after outliers are deleted, so as to obtain preprocessing result information;

[0074] Performing data dimension reduction on the preprocessing result information based on the PCA algorithm, calculating the principal component score of each operating status monitoring data, and selecting the operating status monitoring data corresponding to the principal component score greater than a preset threshold as the principal component feature;

[0075] The principal component analysis results are obtained by using one-hot encoding to encode the principal component features, and the encoded principal component features are processed in time series to generate a time series feature sequence to obtain a real-time operating status feature sequence.

[0076] It should be noted that by arranging a sensor array in the target AC charging pile, a sensor network covering the operation of the target device is constructed, so that the various operating conditions of the AC charging pile during operation can be monitored in real time, and comprehensive operating status monitoring information can be obtained. Based on the monitoring data, the data standard deviation is calculated to reveal the outliers or abnormalities that may exist in the data. For the operating status data that are judged as outliers, in order to avoid adverse effects on the analysis results, these data are directly deleted. At the same time, in order to maintain the continuity and integrity of the data, the linear interpolation algorithm is introduced to supplement the missing values ​​caused by the deletion of outliers, and the missing values ​​are estimated by the interpolation method, so as to generate preprocessing result information with consistency and high credibility. After the data preprocessing is completed, the principal component analysis (PCA) algorithm is used to reduce the dimension of the monitoring data. The PCA method can identify and extract the main features that are most representative of the data changes. While reducing the data dimension, the data information that has the greatest impact on the operating status of the AC charging pile is retained. For the feature data obtained by principal component analysis, the unique hot encoding method is used to convert it into a structured data representation. The encoded data is then processed through time series to produce a time series that reflects changes in the equipment's operating status, ultimately generating a real-time operating status feature sequence.

[0077] Further, in a preferred embodiment of the present invention, the anomaly detection model is constructed based on the generative adversarial network, the real-time operating status feature sequence is input, the abnormal operating data under the current operating status is detected, and the anomaly detection information is obtained, which specifically includes:

[0078] Acquire historical normal operation status monitoring information of the target AC charging pile under different operation modes based on data retrieval, classify the historical normal operation status monitoring information according to the operation mode, and obtain first classification result information;

[0079] Performing principal component analysis based on the first classification result information using a PCA algorithm to obtain principal component characteristics of each operating mode category, and calculating the operating state transition probability based on the principal component characteristics of each operating mode category;

[0080] Generate a state transfer matrix according to the calculated operating state transition probability, construct a state transfer space in combination with the first classification result information, construct an anomaly detection model through a generative adversarial network, and perform model training through the state transfer space;

[0081] Obtain the real-time operating status feature sequence, input it into the trained anomaly detection model for analysis, match the state transfer matrix through the state transfer space, and the generator reconstructs the sequence according to the matched state transfer matrix to obtain the reconstructed sequence;

[0082] The reconstructed sequence is input into the discriminator, and combined with the matching state transfer matrix, it is determined whether the current reconstructed sequence is accepted. If the determination result is not accepted, the next reconstructed sequence of the generator is obtained for determination until it meets the state transfer rule of the matching state transfer matrix, and the final reconstructed sequence is output;

[0083] A reconstruction error between the final reconstruction sequence and the real-time operation status feature sequence is calculated, and abnormal operation status data is detected according to the calculated reconstruction error to obtain abnormality detection information.

[0084] It should be noted that the historical normal operating status monitoring information of the target AC charging pile under different operating modes is obtained through data retrieval, and the operating mode is used as the classification basis to classify and process these monitoring data to generate the first classification result information. By classifying and combing the historical data, the key operating features under each operating mode are extracted. The principal component analysis (PCA) algorithm is introduced to reduce the dimension of the first classification result information, and the principal component features of each operating mode category are extracted from it. These principal component features represent the most important state variables in each operating mode, and the transition probability of the operating state is further calculated to capture the state change law of the device when it is running in different modes. The state transition matrix is ​​generated by using the calculated operating state transition probability, which describes the dynamic change characteristics of the operating state of the target device. Combined with the first classification result information, the state transition space is constructed with the state transition matrix as the core, and the construction of the anomaly detection model is realized by generating adversarial networks (GAN). In practical applications, the real-time operating status feature sequence of the target AC charging pile is input into the trained anomaly detection model for analysis. The generator reconstructs the input sequence according to the matching state transition matrix to generate a candidate reconstructed sequence. Subsequently, the reconstruction sequence is input into the discriminator, and combined with the matching state transfer matrix, it is determined whether to accept the current reconstruction result. If it is determined not to be accepted, the generator generates the next reconstruction sequence and re-judges until the requirements of the state transfer law are met, thereby outputting the final reconstruction sequence. Finally, by calculating the reconstruction error between the final reconstruction sequence and the real-time operating status feature sequence, the abnormal operating status of the target AC charging pile is detected, and the abnormal detection information is obtained. Accurate monitoring and abnormal analysis of the equipment status are achieved.

[0085] Furthermore, in a preferred embodiment of the present invention, the preset risk threshold judges the abnormal detection information with the preset risk threshold, analyzes whether there is an operational risk, and performs risk level assessment and management strategy formulation, specifically including:

[0086] A preset risk threshold is set, and the abnormal detection information is compared with the preset risk threshold. If it is greater than the preset threshold, it indicates that there is a risk in the real-time operation state of the target AC charging pile, and risk control is performed;

[0087] Acquire abnormal detection information and perform feature extraction on the abnormal detection information, extract abnormal operation status features of the target AC charging pile in real time, and obtain abnormal operation status feature information;

[0088] Based on the abnormal operation condition characteristic information, a normal operation benchmark corresponding to the abnormal data is obtained, deviation calculation is performed to obtain the deviation, and a risk level is evaluated according to the calculated deviation value;

[0089] According to the risk level assessment result and the characteristic information of the abnormal operating condition, the corresponding risk management and control strategy is obtained through big data retrieval, and the risk management and control strategy with the highest frequency of use is selected to generate remote control instructions to control the target AC charging pile.

[0090] It should be noted that the abnormal detection information of the target AC charging pile is first obtained, and deep feature extraction is performed on it. The abnormal features in the current operation of the equipment are identified, including the abnormal change trend of key operating indicators, etc., to form a comprehensive characterization of the real-time abnormal operation status, and generate abnormal operation status feature information, providing accurate data basis for risk assessment and response. Subsequently, based on the abnormal operation status feature information, the normal operation benchmark value of the corresponding abnormal data is further obtained, and the deviation is calculated with the help of the benchmark value. By analyzing the deviation between the abnormal data and the normal benchmark value, the severity of the current abnormality is quantified. Based on the calculated deviation value, the risk level of the current operation status is assessed to provide a clear priority and response strategy for risk control. After the risk level assessment is completed, combined with the abnormal operation status feature information, historical control cases are analyzed through big data retrieval technology to quickly obtain risk control strategies similar to the current abnormal operation status. Among the various historical strategies, the control strategy with the highest frequency of adoption and the best performance in practice is selected as the reference for current risk control. Finally, the generated risk control strategy is converted into a remote control instruction and sent to the target AC charging pile through the communication network to achieve timely control of the abnormal state, thereby improving the stability and safety of equipment operation.

[0091] Further, in a preferred embodiment of the present invention, the acquisition of historical risk instances of AC charging piles, analysis of risk evolution characteristics of various risk events, construction of a risk prediction model and analysis of whether the abnormal operation of the target AC charging pile will evolve into a risk event, and maintenance warning prompts are specifically included:

[0092] Obtain historical risk instances of AC charging piles by means of big data, extract features of the historical risk instances to obtain feature information of the historical risk instances, and generate a time series feature sequence of each historical risk instance according to the feature information of the historical risk instances;

[0093] Calculate the Euclidean distance between the time series feature sequences of each historical risk instance, perform clustering based on the calculated Euclidean distance to obtain several clusters, and input the obtained clusters into the LSTM network for analysis;

[0094] The risk evolution characteristics are obtained by extracting features from various clusters through the LSTM network, and the risk evolution characteristics are reconstructed by feature encoding using a stacked autoencoder. The attention mechanism is introduced to weight the features to represent the importance of the risk evolution characteristics.

[0095] Obtain the reconstructed risk evolution characteristics, conduct risk level assessment on the reconstructed risk evolution characteristics, form a number of risk evolution characteristic data sets classified by risk level, build a risk prediction model and conduct model training;

[0096] Obtaining the abnormal detection result of the target AC charging pile within a unit time, inputting it into the risk prediction model, predicting the evolution of the current abnormal operation status of the target AC charging pile in a future time period, and obtaining risk prediction information;

[0097] It is determined whether the target AC charging pile has an operation risk based on the risk prediction information. If there is an operation risk, a maintenance reminder is generated and pushed to the user end to remind the user to make a maintenance appointment.

[0098] It should be noted that by extracting the features of historical risk instances, their feature information is generated, and based on these feature information, the time series feature sequences of each historical risk instance are constructed to form a time series data set that can reflect the dynamic evolution of risks. In the generated time series feature sequence, the Euclidean distance between each risk instance is calculated to analyze their similarities, and these risk instances are clustered based on the results of distance calculation and divided into several clusters. Each cluster represents a risk category with similar features and evolution patterns. On this basis, these clusters are input into the long short-term memory network (LSTM) for analysis, so as to extract the dynamic change characteristics of the risk clusters and deeply explore the risk evolution law of each type of risk instance. The stacked autoencoder is used to encode and reconstruct the risk evolution features extracted by LSTM, and the representation ability of the features is optimized by deep learning. At the same time, the attention mechanism is introduced to weight these features, evaluate and highlight the feature parts that play a key role in the risk evolution analysis, so as to more effectively understand and describe the risk evolution process. Based on the risk evolution characteristics after reconstruction and optimization, the risk level assessment operation is performed, and the assessment results are divided into multiple categories according to the risk level, and multiple risk evolution feature data sets are constructed based on these categories. The risk prediction model is constructed and trained through these data sets, so that the model can learn and predict the future change trends of different categories of risks, and improve the ability to assess the potential risks of the target AC charging pile. Subsequently, the abnormal detection results of the target AC charging pile in unit time are imported as input data into the trained risk prediction model for analysis. It is predicted whether the abnormal operation of the AC charging pile in the future will further develop into a risk event, and risk prediction information is generated. According to the risk prediction information, it is judged whether the target AC charging pile has potential operation risks. If it is confirmed that there is a risk, the maintenance reminder information will be generated immediately and sent to the user end through the push mechanism to remind the user to make a maintenance appointment for the equipment in time, which improves the safety and management efficiency of the charging pile operation.

[0099] Further, in a preferred embodiment of the present invention, the scene monitoring video stream data during charging is obtained, malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile, and if malicious behavior exists, a control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target, specifically including:

[0100] Based on big data retrieval methods, different malicious behavior data of charging piles during charging are collected, and features of the collected malicious behavior data are extracted to obtain malicious behavior features;

[0101] The malicious behavior level is preset, and a fuzzy comprehensive evaluation method is introduced to evaluate the malicious behavior level of the acquired malicious behavior features. Based on the malicious behavior level evaluation results, the malicious behavior features are divided into different categories of data sets to obtain several malicious behavior feature category data sets;

[0102] A malicious behavior analysis model is constructed based on a deep neural network, a malicious behavior feature category data set is used to train the malicious behavior analysis model, and a back propagation method is used to optimize and adjust model parameters to obtain a malicious behavior analysis model that meets expectations;

[0103] Based on the scene monitoring device installed on the target AC charging pile, the scene monitoring video stream data during charging is obtained and input into the malicious behavior analysis model, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile to obtain malicious behavior analysis information;

[0104] It is determined whether malicious behavior occurs when the target AC charging pile is charging according to the malicious behavior analysis information. If malicious behavior occurs, the malicious behavior level is extracted according to the malicious behavior analysis information, and a control strategy is obtained to stop the malicious behavior.

[0105] It should be noted that different types of malicious behavior data that may occur during the charging operation of the AC charging pile are collected by big data retrieval means, and the features of these data are extracted to obtain feature information that accurately describes the nature of the malicious behavior. The malicious behavior level is preset, and the acquired malicious behavior features are evaluated by the fuzzy comprehensive evaluation method. Fuzzy comprehensive evaluation can effectively deal with the ambiguity and uncertainty in the malicious behavior features and ensure the accuracy of the malicious behavior level evaluation. According to the evaluation results, the malicious behavior features are divided into different categories of data sets, thereby generating several malicious behavior feature category data sets to provide structured data support for classification modeling. Next, a dedicated malicious behavior analysis model is constructed based on a deep neural network, and the model is trained using the aforementioned malicious behavior feature category data sets. During the training process, the back propagation algorithm is used to optimize and adjust the parameters of the model to maximize the classification accuracy of the model and the analysis ability of malicious behavior. Through multiple iterative optimizations, a high-performance malicious behavior analysis model that meets expectations is finally obtained. Subsequently, real-time scene monitoring video stream data is obtained based on the scene monitoring device installed on the target AC charging pile, covering the behavior scenes involved during the charging operation. The video stream data is input into the trained malicious behavior analysis model to analyze the specific behaviors in the charging scene, accurately detect and identify whether there are malicious behaviors, and generate corresponding malicious behavior analysis information. If malicious behavior is confirmed, the level information of the malicious behavior is further extracted to clarify the degree of harm of the behavior. Combined with the level and nature of the specific malicious behavior, the corresponding control strategy is obtained, and control instructions are generated to effectively warn and stop malicious behaviors in a timely manner to ensure the safety and normal order of the charging pile operation environment.

[0106] Furthermore, in a preferred embodiment of the present invention, if there is malicious behavior, the malicious behavior level is extracted according to the malicious behavior analysis information, and a control strategy is obtained to stop the malicious behavior, which specifically includes:

[0107] If malicious behavior is detected in the current charging scenario, the malicious behavior level currently detected is extracted according to the malicious behavior analysis information, and the malicious behavior level is divided into three levels: low, medium, and high;

[0108] If the current malicious behavior level is low-level malicious behavior, a warning instruction is generated based on the preset warning strategy, the target AC charging pile is remotely warned and controlled, and an early warning is pushed to the user;

[0109] If the current malicious behavior level is medium-level malicious behavior, a warning instruction is generated based on the preset warning strategy, a power-off control instruction is generated according to the preset safety protection strategy, the target AC charging pile is remotely controlled, and an early warning is pushed to the user;

[0110] If the current malicious behavior level is high-level malicious behavior, a warning instruction is generated based on the preset warning strategy, and a power-off control instruction is generated according to the preset safety protection strategy to control the target AC charging pile to issue a warning voice and perform power-off protection, and the video stream data detected with malicious behavior is retained and pushed to the target user;

[0111] After the target AC charging pile performs warning processing according to the control instruction, continue to detect whether the malicious behavior in the current charging scene disappears. If the malicious behavior still exists, increase the malicious behavior level and obtain the corresponding level of control strategy to control the target AC charging pile.

[0112] It should be noted that when malicious behavior is detected in the current charging scenario, the malicious behavior level currently detected is extracted through the malicious behavior analysis information. Malicious behavior levels are divided into three categories: low level, medium level and high level, and corresponding control strategies are adopted according to different levels. For low-level malicious behavior, a warning instruction is generated according to the preset warning strategy, and the target AC charging pile is remotely controlled. This control is mainly based on issuing warnings, such as taking a simple buzzing sound for warning, and pushing relevant warning information to the user to remind the user to pay attention to potential problems. When medium-level malicious behavior is detected, a warning instruction is also generated based on the preset warning strategy. At this time, the warning effect can include voice warnings. At the same time, a power-off control instruction is further generated in combination with the safety protection strategy, and the target AC charging pile is powered off through remote control. Medium-level malicious behavior represents malicious behavior such as abnormal unplugging of charging guns. By giving a certain degree of warning on the scene and taking active measures to cut off the power supply, behaviors that may cause greater damage can be avoided. At the same time, the warning information will be pushed to the user synchronously to ensure that the user can know the problem in time. In the case of high-level malicious behavior, warning instructions generated by the preset warning strategy and power-off control instructions generated by the safety protection strategy are used to control the target AC charging pile to execute comprehensive protection measures. The charging pile will issue a warning voice to alert the surrounding personnel and immediately cut off the power supply to ensure the safety of the scene. In addition, the relevant video stream data of the detected malicious behavior will be saved and processed, and these data will be pushed to the user terminal to provide a complete chain of evidence and data support for subsequent disposal. After the target AC charging pile executes warnings or protection measures according to the control instructions, the malicious behavior status in the current charging scene is continuously monitored. If it is detected that the malicious behavior still exists, the malicious behavior level will be automatically increased, and the corresponding control strategy will be obtained according to the upgraded level, and a higher level of control operation will be performed on the target AC charging pile to ensure that the charging pile can operate under safe and controllable conditions, while eliminating safety hazards as much as possible.

[0113] Figure 2 A flow chart of a malicious behavior control method applied to an AC charging pile provided in one embodiment of the present invention;

[0114] like Figure 2 As shown, the present invention provides a flow chart of a malicious behavior control method applied to an AC charging pile, including:

[0115] S202, acquiring scene monitoring video stream data during charging based on a scene monitoring device installed at a target AC charging pile, and inputting the data into the malicious behavior analysis model to obtain malicious behavior analysis information;

[0116] S204, if there is malicious behavior, extracting the malicious behavior level according to the malicious behavior analysis information, and obtaining a control strategy to stop the malicious behavior;

[0117] S206, if the current malicious behavior level is a low-level malicious behavior, a warning instruction is generated based on a preset warning strategy, a remote warning control is performed on the target AC charging pile, and a warning is pushed to the user;

[0118] S208, if the current malicious behavior level is medium-level malicious behavior, a warning instruction is generated based on a preset warning strategy, a power-off control instruction is generated according to a preset safety protection strategy, the target AC charging pile is remotely controlled, and an early warning is pushed to the user;

[0119] S210, if the current malicious behavior level is a high-level malicious behavior, a warning instruction is generated based on the preset warning strategy, and a power-off control instruction is generated according to the preset safety protection strategy. The target AC charging pile is controlled to issue a warning voice and perform power-off protection, and the video stream data of the detected malicious behavior is retained and pushed to the target user.

[0120] Furthermore, in an intelligent remote control method applied to an AC charging pile provided by the present invention, the following steps are also included:

[0121] Obtaining historical energy consumption information of a target AC charging pile, performing feature extraction on the historical energy consumption information, obtaining energy consumption type features, consumption degree features, and consumption time features, and obtaining historical energy consumption feature information;

[0122] Building an analysis model based on SARIMA, inputting the historical energy consumption characteristic information into the analysis model to perform energy consumption trend analysis, and obtaining energy consumption trend analysis information;

[0123] Combining the historical energy consumption characteristic information and the energy consumption trend analysis information to construct an energy consumption profile of the target AC charging pile, and forming an energy consumption database based on the constructed energy consumption profile;

[0124] Performing real-time monitoring of energy consumption of a target AC charging pile, obtaining real-time energy consumption monitoring information, and extracting an energy consumption profile of the monitored target AC charging pile according to the energy consumption database;

[0125] Perform similarity calculation with the real-time energy consumption monitoring information to obtain a similarity value, and compare the calculated similarity value with a preset threshold to determine whether the energy consumption of the target AC charging pile is normal;

[0126] If the judgment result is abnormal, it means that the target AC charging pile has abnormal energy consumption. The energy consumption benchmark data of the target AC charging pile is extracted based on the energy consumption profile of the target AC charging pile.

[0127] The energy consumption baseline data of the target AC charging pile is calculated for deviation from the real-time energy consumption monitoring information, the degree of energy consumption abnormality is evaluated according to the calculated deviation value, and based on the degree of energy consumption abnormality, it is determined whether the target AC charging pile has energy leakage, and an early warning prompt is issued.

[0128] It should be noted that for AC charging piles, whether for home or public use, their charging energy consumption within a certain period of time is regular. When abnormal energy consumption occurs, it may be caused by abnormal personnel borrowing or AC charging pile leakage. Timely warnings and prompts for maintenance are required to avoid greater losses.

[0129] Figure 3 An intelligent remote control system 3 applied to an AC charging pile is provided in one embodiment of the present invention. The system includes: a memory 31 and a processor 32. The memory 31 contains an intelligent remote control method program applied to the AC charging pile. When the intelligent remote control method program applied to the AC charging pile is executed by the processor 32, the following steps are implemented:

[0130] Monitor the operating status of the target AC charging pile when it is working, obtain operating status monitoring information, pre-process the operating status monitoring information, and construct a real-time operating status feature sequence;

[0131] Building an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information;

[0132] Preset risk thresholds, compare the abnormal detection information with the preset risk thresholds, analyze whether there are operational risks, and conduct risk level assessment and control strategy formulation;

[0133] Obtain historical risk examples of AC charging piles, analyze the risk evolution characteristics of various risk events, build a risk prediction model and analyze whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provide maintenance warning prompts;

[0134] The scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target.

[0135] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0136] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0137] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0138] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, etc. Various media that can store program codes.

[0139] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0140] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent remote control method for AC charging piles, characterized in that: include: Monitor the operating status of the target AC charging pile when it is working, obtain operating status monitoring information, pre-process the operating status monitoring information, and construct a real-time operating status feature sequence; Building an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information; Preset risk thresholds, compare the abnormal detection information with the preset risk thresholds, analyze whether there are operational risks, and conduct risk level assessment and control strategy formulation; Obtain historical risk examples of AC charging piles, analyze the risk evolution characteristics of various risk events, build a risk prediction model and analyze whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provide maintenance warning prompts; The scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target.

2. The intelligent remote control method for AC charging pile according to claim 1, characterized in that: The monitoring of the operating status of the target AC charging pile during operation, obtaining operating status monitoring information, preprocessing the operating status monitoring information, and constructing a real-time operating status feature sequence specifically includes: Arrange a sensor array in the target AC charging pile, monitor the operating status of the target AC charging pile when it is working based on the sensor network, and obtain operating status monitoring information; Calculate the data standard deviation according to the acquired operation status monitoring information, compare the calculated data standard deviation with the preset standard deviation range, and define the data corresponding to the data standard deviation that is not within the preset standard deviation range as an abnormal value; The operation status monitoring data defined as outliers are deleted, and a linear interpolation algorithm is introduced to supplement the missing values ​​of the operation status monitoring data after outliers are deleted, so as to obtain preprocessing result information; Performing data dimension reduction on the preprocessing result information based on the PCA algorithm, calculating the principal component score of each operating status monitoring data, and selecting the operating status monitoring data corresponding to the principal component score greater than a preset threshold as the principal component feature; The principal component analysis results are obtained by using one-hot encoding to encode the principal component features, and the encoded principal component features are processed in time series to generate a time series feature sequence to obtain a real-time operating status feature sequence.

3. The intelligent remote control method for AC charging pile according to claim 1 is characterized in that: The abnormality detection model is constructed based on the generative adversarial network, the real-time operating status feature sequence is input, abnormal operating data under the current operating status is detected, and abnormality detection information is obtained, specifically including: Acquire historical normal operation status monitoring information of the target AC charging pile under different operation modes based on data retrieval, classify the historical normal operation status monitoring information according to the operation mode, and obtain first classification result information; Performing principal component analysis based on the first classification result information using a PCA algorithm to obtain principal component characteristics of each operating mode category, and calculating the operating state transition probability based on the principal component characteristics of each operating mode category; Generate a state transfer matrix according to the calculated operating state transition probability, construct a state transfer space in combination with the first classification result information, construct an anomaly detection model through a generative adversarial network, and perform model training through the state transfer space; Obtain the real-time operating status feature sequence, input it into the trained anomaly detection model for analysis, match the state transfer matrix through the state transfer space, and the generator reconstructs the sequence according to the matched state transfer matrix to obtain the reconstructed sequence; The reconstructed sequence is input into the discriminator, and combined with the matching state transfer matrix, it is determined whether the current reconstructed sequence is accepted. If the determination result is not accepted, the next reconstructed sequence of the generator is obtained for determination until it meets the state transfer rule of the matching state transfer matrix, and the final reconstructed sequence is output; A reconstruction error between the final reconstruction sequence and the real-time operation status feature sequence is calculated, and abnormal operation status data is detected according to the calculated reconstruction error to obtain abnormality detection information.

4. The intelligent remote control method for AC charging pile according to claim 1 is characterized in that: The preset risk threshold is used to judge the abnormal detection information against the preset risk threshold, analyze whether there is an operational risk, and conduct risk level assessment and control strategy formulation, specifically including: A preset risk threshold is set, and the abnormal detection information is compared with the preset risk threshold. If it is greater than the preset threshold, it indicates that there is a risk in the real-time operation state of the target AC charging pile, and risk control is performed; Acquire abnormal detection information and perform feature extraction on the abnormal detection information, extract abnormal operation status features of the target AC charging pile in real time, and obtain abnormal operation status feature information; Based on the abnormal operation condition characteristic information, a normal operation benchmark corresponding to the abnormal data is obtained, deviation calculation is performed to obtain the deviation, and a risk level is evaluated according to the calculated deviation value; According to the risk level assessment result and the characteristic information of the abnormal operating condition, the corresponding risk management and control strategy is obtained through big data retrieval, and the risk management and control strategy with the highest frequency of use is selected to generate remote control instructions to control the target AC charging pile.

5. The intelligent remote control method for AC charging pile according to claim 1, characterized in that: The acquisition of historical risk instances of AC charging piles, analysis of risk evolution characteristics of various risk events, construction of a risk prediction model and analysis of whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provision of maintenance warning prompts specifically include: Obtain historical risk instances of AC charging piles by means of big data, extract features of the historical risk instances to obtain feature information of the historical risk instances, and generate a time series feature sequence of each historical risk instance according to the feature information of the historical risk instances; Calculate the Euclidean distance between the time series feature sequences of each historical risk instance, perform clustering based on the calculated Euclidean distance to obtain several clusters, and input the obtained clusters into the LSTM network for analysis; The risk evolution characteristics are obtained by extracting features from various clusters through the LSTM network, and the risk evolution characteristics are reconstructed by feature encoding using a stacked autoencoder. The attention mechanism is introduced to weight the features to represent the importance of the risk evolution characteristics. Obtain the reconstructed risk evolution characteristics, conduct risk level assessment on the reconstructed risk evolution characteristics, form a number of risk evolution characteristic data sets classified by risk level, build a risk prediction model and conduct model training; Obtaining the abnormal detection result of the target AC charging pile within a unit time, inputting it into the risk prediction model, predicting the evolution of the current abnormal operation status of the target AC charging pile in a future time period, and obtaining risk prediction information; It is determined whether the target AC charging pile has an operation risk based on the risk prediction information. If there is an operation risk, a maintenance reminder is generated and pushed to the user end to remind the user to make a maintenance appointment.

6. The intelligent remote control method for AC charging pile according to claim 1, characterized in that: The scene monitoring video stream data during charging is obtained, malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile, and if malicious behavior exists, a control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target, specifically including: Based on big data retrieval methods, different malicious behavior data of charging piles during charging are collected, and features of the collected malicious behavior data are extracted to obtain malicious behavior features; The malicious behavior level is preset, and a fuzzy comprehensive evaluation method is introduced to evaluate the malicious behavior level of the acquired malicious behavior features. Based on the malicious behavior level evaluation results, the malicious behavior features are divided into different categories of data sets to obtain several malicious behavior feature category data sets; A malicious behavior analysis model is constructed based on a deep neural network, a malicious behavior feature category data set is used to train the malicious behavior analysis model, and a back propagation method is used to optimize and adjust model parameters to obtain a malicious behavior analysis model that meets expectations; Based on the scene monitoring device installed on the target AC charging pile, the scene monitoring video stream data during charging is obtained and input into the malicious behavior analysis model, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile to obtain malicious behavior analysis information; It is determined whether malicious behavior occurs when the target AC charging pile is charging according to the malicious behavior analysis information. If malicious behavior occurs, the malicious behavior level is extracted according to the malicious behavior analysis information, and a control strategy is obtained to stop the malicious behavior.

7. The intelligent remote control method for AC charging pile according to claim 6 is characterized in that: If there is malicious behavior, extract the malicious behavior level according to the malicious behavior analysis information, and obtain a control strategy to stop the malicious behavior, specifically including: If malicious behavior is detected in the current charging scenario, the malicious behavior level currently detected is extracted according to the malicious behavior analysis information, and the malicious behavior level is divided into three levels: low, medium, and high; If the current malicious behavior level is low-level malicious behavior, a warning instruction is generated based on the preset warning strategy, the target AC charging pile is remotely warned and controlled, and an early warning is pushed to the user; If the current malicious behavior level is medium-level malicious behavior, a warning instruction is generated based on the preset warning strategy, a power-off control instruction is generated according to the preset safety protection strategy, the target AC charging pile is remotely controlled, and an early warning is pushed to the user; If the current malicious behavior level is high-level malicious behavior, a warning instruction is generated based on the preset warning strategy, and a power-off control instruction is generated according to the preset safety protection strategy to control the target AC charging pile to issue a warning voice and perform power-off protection, and the video stream data detected with malicious behavior is retained and pushed to the target user; After the target AC charging pile performs warning processing according to the control instruction, continue to detect whether the malicious behavior in the current charging scene disappears. If the malicious behavior still exists, increase the malicious behavior level and obtain the corresponding level of control strategy to control the target AC charging pile.

8. An intelligent remote control system for AC charging piles, characterized in that: The system includes: a memory and a processor, wherein the memory contains an intelligent remote control method program applied to an AC charging pile, and when the intelligent remote control method program applied to an AC charging pile is executed by the processor, the following steps are implemented: Monitor the operating status of the target AC charging pile when it is working, obtain operating status monitoring information, pre-process the operating status monitoring information, and construct a real-time operating status feature sequence; Building an anomaly detection model based on a generative adversarial network, inputting the real-time operating status feature sequence, detecting abnormal operating data under the current operating status, and obtaining anomaly detection information; Preset risk thresholds, compare the abnormal detection information with the preset risk thresholds, analyze whether there are operational risks, and conduct risk level assessment and control strategy formulation; Obtain historical risk examples of AC charging piles, analyze the risk evolution characteristics of various risk events, build a risk prediction model and analyze whether the abnormal operation of the target AC charging pile will evolve into a risk event, and provide maintenance warning prompts; The scene monitoring video stream data during charging is obtained, and malicious behavior detection and analysis is performed on the charging scene of the target AC charging pile. If malicious behavior exists, the control strategy is obtained to control the target AC charging pile to warn and stop the malicious behavior target.

Images