Intelligent charging pile safety management system and method

Abstract

The present application relates to the field of charging pile safety management, and provides an intelligent charging pile safety management system and method, which uses data analysis technology based on deep learning to determine whether the charging pile is in a safe state by performing time series analysis on the insulation resistance value, temperature value, input three-phase voltage value, and output DC voltage value of the charging pile at multiple time points. In this way, the mutual relationship between different charging pile monitoring parameters is fully considered, and the safety state of the charging pile can be judged more accurately, which is conducive to reducing misjudgment and missed judgment.

Description

Technical Field

[0001] The present application relates to the field of charging pile safety management, and more specifically, to an intelligent charging pile safety management system and method. Background Art

[0002] As an important charging equipment for electric vehicles, the safety of charging piles is directly related to the safety of users' lives and property. If there are safety hazards in charging piles, it is very likely to cause safety accidents such as fires. Therefore, judging the safety status of charging piles is a necessary measure to prevent safety accidents. The traditional charging pile safety management system simply monitors the relevant parameters of each charging pile independently and judges whether the charging pile is safe based on a simple threshold mechanism. This method ignores the relationship between different monitoring parameters and is prone to misjudgment and missed judgment.

[0003] Therefore, a smart charging pile safety management solution is needed. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present application provides a smart charging pile safety management system and method.

[0005] A smart charging pile safety management method, comprising:

[0006] Obtaining insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​of the charging pile at multiple predetermined time points within a predetermined time period;

[0007] Performing a comprehensive time series analysis on the insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix;

[0008] The safety status classification result of the charging pile is obtained by fusing the information in the time series feature matrix according to the monitoring parameters of the charging pile.

[0009] An intelligent charging pile safety management system, comprising:

[0010] The charging pile safety monitoring parameter data acquisition module is used to obtain the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value of the charging pile at multiple predetermined time points within a predetermined time period;

[0011] A charging pile safety monitoring parameter data analysis module, used to perform a comprehensive time series analysis on the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix;

[0012] The charging pile safety monitoring result generation module is used to obtain the safety status classification result of the charging pile according to the charging pile monitoring parameters and the information in the time series feature matrix.

[0013] This application has significant technical effects due to the adoption of the above technical solutions:

[0014] The intelligent charging pile safety management system and method provided by the present application adopts data analysis technology based on deep learning, and judges whether the charging pile is in a safe state by performing time series analysis on the insulation resistance value, temperature value, input three-phase voltage value, and output DC voltage value of the charging pile at multiple time points. In this way, the mutual relationship between the monitoring parameters of different charging piles is fully considered, and the safety state of the charging pile can be judged more accurately, which is conducive to reducing the situation of misjudgment and missed judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 Flow chart of a smart charging pile safety management method according to an embodiment of the present application.

[0017] Figure 2 This is a flowchart of step S2 in the smart charging pile safety management method according to an embodiment of the present application.

[0018] Figure 3 This is a flowchart of step S21 in the smart charging pile safety management method according to an embodiment of the present application.

[0019] Figure 4 This is a flowchart of step S3 in the smart charging pile safety management method according to an embodiment of the present application.

[0020] Figure 5 This is a system block diagram of a smart charging pile safety management system according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] In today's era of booming electric vehicles, charging piles are an indispensable charging infrastructure for electric vehicles. The importance of safety is self-evident and is directly and closely related to the safety of users' lives and property. In daily life, electric vehicle charging piles are widely distributed in various places such as residential areas, commercial parking lots, and highway service areas. These places are densely populated with people and vehicles. Once there is a safety problem with the charging pile, the consequences will be disastrous.

[0023] For example, in some old residential areas, due to limited power capacity, some charging piles may not fully consider the overall power load of the community when they are installed. If there are safety hazards in the charging piles, such as internal circuit short circuits, heat dissipation system failures, etc., it is very likely to cause serious safety accidents such as fires. Once a fire occurs, it will not only burn the charging piles and surrounding electric vehicles, but the spread of the fire may also endanger residential buildings, causing huge casualties and property losses.

[0024] In view of this, accurate safety status judgment of charging piles has become a key and necessary measure to prevent safety accidents. At present, traditional charging pile safety management systems generally adopt a relatively basic monitoring method, which only independently monitors the voltage, current, temperature and other related parameters of each charging pile. Then, based on a simple threshold mechanism, it is judged whether the charging pile is in a safe state. For example, when the temperature of the charging pile is monitored to exceed a preset fixed threshold, it is determined that the charging pile may have a safety risk.

[0025] However, this approach has obvious drawbacks. In actual operation, there are actually complex relationships between different monitoring parameters. Taking temperature and current as an example, when the current suddenly increases, the resistance inside the charging pile will generate more heat due to the thermal effect of the current, which in turn causes the temperature to rise. However, traditional independent monitoring and simple threshold judgment methods often ignore the inherent connection between these parameters, which is prone to misjudgment and missed judgment. Therefore, a smart charging pile safety management solution is expected.

[0026] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for the safety management of smart charging piles.

[0027] Figure 1 Flow chart of the smart charging pile safety management method according to the embodiment of the present application. Figure 1As shown, according to the intelligent charging pile safety management method of the embodiment of the present application, the method includes: S1, obtaining the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value of the charging pile at multiple predetermined time points within a predetermined time period; S2, performing a comprehensive timing analysis on the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value at the multiple predetermined time points to obtain a charging pile monitoring parameter fusion timing feature matrix; S3, obtaining the safety status classification result of the charging pile according to the information in the charging pile monitoring parameter fusion timing feature matrix.

[0028] In step S1, the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value of the charging pile at multiple predetermined time points within a predetermined time period are obtained. It should be understood that insulation resistance is a key indicator for measuring the electrical insulation performance of the charging pile. During the operation of the charging pile, good insulation performance can prevent current from leaking to non-conductive parts, such as the outer shell. If the insulation resistance is too low, leakage may occur, making the outer shell of the charging pile charged, and an electric shock accident may occur when the user touches it. Moreover, leakage may also cause a short circuit, which in turn causes serious safety accidents such as fire. The charging pile generates heat when working, and its internal electronic components and charging modules have a certain operating temperature range. Excessive temperature may cause component performance degradation, accelerated aging or even damage. For example, the power semiconductor device in the charging module may experience thermal breakdown at high temperature, resulting in a short circuit. The three-phase voltage is the input power parameter of the charging pile. A stable input voltage is essential for the normal operation of the charging pile. Too high or too low input voltage may damage the circuits and components inside the charging pile. For example, when the input three-phase voltage is too high, it may exceed the withstand voltage of the transformer, rectifier and other components inside the charging pile, causing the components to be broken down and damaged, causing faults such as short circuits; while too low voltage may reduce the charging efficiency or even prevent normal charging. Long-term operation in this state may also affect the service life and safety of the charging pile. The output DC voltage is the voltage directly supplied to the electric vehicle for charging. Its stability is directly related to the safety of electric vehicle charging and the service life of the battery. Considering that the state of the charging pile is not static, but changes with time and usage. By obtaining the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value at multiple time points, the model can understand the changing trend of these indicators over time, so as to more comprehensively and accurately evaluate the safety status of the charging pile. Discover potential risks in time to ensure that the charging pile can operate in the optimal state.

[0029] In step S2, a comprehensive time series analysis is performed on the insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix. Specifically, Figure 2FIG. 1 is a flow chart of step S2 in the smart charging pile safety management method according to an embodiment of the present application. Figure 2 As shown, the step S2 includes: S21, extracting the timing characteristics of the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the multiple predetermined time points to obtain a timing characteristic matrix of charging pile monitoring parameters; S22, emphasizing the timing characteristics of the charging pile monitoring parameter timing characteristic matrix to obtain a charging pile monitoring parameter timing characteristic emphasis matrix; S23, fusing the charging pile monitoring parameter timing characteristic matrix and the charging pile monitoring parameter timing characteristic emphasis matrix to obtain the charging pile monitoring parameter fused timing characteristic matrix.

[0030] In step S21, the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the plurality of predetermined time points are subjected to time series feature extraction to obtain a charging pile monitoring parameter time series feature matrix. Specifically, Figure 3 FIG. 1 is a flow chart of step S21 in the smart charging pile safety management method according to an embodiment of the present application. Figure 3 As shown, the step S21 includes: S211, arranging the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the multiple predetermined time points according to the time dimension and the sample dimension to obtain a full-time input matrix of the charging pile monitoring parameters; S212, passing the full-time input matrix of the charging pile monitoring parameters through the charging pile monitoring parameter timing sensor to obtain the charging pile monitoring parameter timing characteristic matrix.

[0031] In step S211, the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the multiple predetermined time points are arranged according to the time dimension and the sample dimension to obtain the full time-series input matrix of the monitoring parameters of the charging pile. It should be understood that different monitoring indicators come from different sensors. In order to integrate the scattered data together to form a complete data set, it is helpful to evaluate the status of the charging pile from a global perspective, rather than looking at each indicator in isolation. In this application, the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the multiple predetermined time points need to be arranged according to the time dimension and the sample dimension. Through this arrangement, the time correlation and internal laws between the data can be better mined. In detail, in the time dimension, the change of each indicator over time can be observed, for example, whether the insulation resistance value gradually decreases over time, or whether the temperature rises abnormally in a certain time period. In the sample dimension, the situation of different indicators at the same time point can be compared at the same time, such as observing whether the input three-phase voltage also fluctuates while the temperature rises, so as to find the possible mutual influence relationship between different indicators. Each row in the full-time input matrix of the charging pile monitoring parameters represents a set of all monitoring parameters at a specific time point, and each column represents a sequence of values ​​of a specific monitoring parameter at all predetermined time points.

[0032] In step S212, the charging pile monitoring parameter full time series input matrix is ​​passed through the charging pile monitoring parameter time series sensor to obtain the charging pile monitoring parameter time series feature matrix. It should be understood that the charging pile monitoring parameter full time series input matrix contains the changing trend of different charging pile monitoring parameters over time and the relationship between different monitoring parameters. In order to capture this information from this original matrix, it is necessary to use the charging pile monitoring parameter time series sensor to encode the charging pile monitoring parameter full time series input matrix. Here, the charging pile monitoring parameter time series sensor described in this application is a convolutional neural network model in which adjacent layers use mutually transposed convolution kernels. Those of ordinary skill in the art should know that ordinary convolutional neural networks mainly obtain features of different scales through convolution kernels of different sizes and pooling operations, but when processing time series data, this method may lose some detailed information on the time series. The structure of adjacent layers using mutually transposed convolution kernels can better capture multi-scale time series features. Specifically, during the forward propagation process, a convolution layer can extract the detailed features of the local time series, while the adjacent transposed convolution layer can restore these detailed features to the scale of the original time series, while fusing more macroscopic contextual information to obtain a feature matrix containing details and contexts of different scales, which helps to fully understand the time series changes of the charging pile monitoring parameters. Moreover, in ordinary convolutional neural networks, feature information is mainly transmitted unidirectionally from the bottom layer to the top layer. As the number of layers increases, some detailed features of the bottom layer may gradually be lost. In the model where adjacent layers use mutually transposed convolution kernels, the transposed convolution layer can back-propagate the abstract features of the high layer to the bottom layer and fuse them with the detailed features of the bottom layer. This enables the model to simultaneously use high-level abstract features and bottom-level detailed features to better capture the information in the charging pile monitoring parameter matrix and improve the ability to recognize complex time series patterns.

[0033] In step S22, the timing feature matrix of the charging pile monitoring parameters is emphasized to obtain the timing feature emphasis matrix of the charging pile monitoring parameters. Specifically, in an embodiment of the present application, the step S22 includes: passing the timing feature matrix of the charging pile monitoring parameters through the key timing feature emphasizer of the charging pile monitoring parameters to obtain the timing feature emphasis matrix of the charging pile monitoring parameters. It should be understood that the information contained in the timing feature matrix of the charging pile monitoring parameters is relatively complex, and the importance of different features is different for judging the safety status of the charging pile. In order to highlight the timing features that play a key role in judging the safety status of the charging pile and weaken the secondary features, it is necessary to emphasize the timing features of the timing feature matrix of the charging pile monitoring parameters. That is, in actual operation, the safety hazards of the charging pile are often accompanied by certain abnormal timing features. These abnormal features may not be so obvious among many features and are easily masked by other normal features. Performing timing feature emphasis processing can focus on these abnormal features that may indicate safety problems, making it easier for the model to pay attention to these abnormal situations in the subsequent classification process. Specifically, the charging pile monitoring parameter timing feature matrix is ​​input into the charging pile monitoring parameter key timing feature emphasizer for emphasis processing. Here, the charging pile monitoring parameter key timing feature emphasizer is a convolutional neural network model using a spatial attention mechanism. The spatial attention mechanism can automatically learn and focus on the key areas in the feature matrix. In the charging pile monitoring parameter timing feature matrix, the importance of the features corresponding to different time points and parameters for the safety state judgment is different. The mechanism can learn to discover those feature areas that are closely related to potential safety hazards. For example, it may find that a set of features near a specific time point (such as a sudden change in the output DC voltage and a decrease in insulation resistance) is very critical for judging the safety state, thereby automatically emphasizing these areas. By adaptively adjusting the degree and range of emphasis according to the specific circumstances of the input data, it is possible to dynamically focus on the features that are most likely to reflect safety issues according to the different working conditions and states during the actual operation of the charging pile, which is conducive to more accurate identification of potential safety risks of the charging pile.

[0034] In step S23, the charging pile monitoring parameter timing feature matrix and the charging pile monitoring parameter timing feature emphasis matrix are fused to obtain the charging pile monitoring parameter fusion timing feature matrix. It should be understood that the charging pile monitoring parameter timing feature matrix contains the timing features extracted from the original data, which can reflect the various changes in the charging pile operating parameters in the time series, such as trends, cycles, etc. The charging pile monitoring parameter timing feature emphasis matrix is ​​a matrix that highlights the key timing features, and it focuses on those features that are most important for judging the safety status. The information contained in the two matrices is complementary. In order to combine comprehensive timing features with key emphasis features, so that the final feature matrix contains both complete timing information and highlights the key parts related to safety, it is necessary to fuse the charging pile monitoring parameter timing feature matrix and the charging pile monitoring parameter timing feature emphasis matrix in this application. By fusing the matrix, more detailed information can be provided for the safety assessment of the charging pile.

[0035] In step S3, the safety status classification result of the charging pile is obtained according to the information in the time series feature matrix of the charging pile monitoring parameters. Specifically, Figure 4 FIG. 1 is a flow chart of step S3 in the smart charging pile safety management method according to an embodiment of the present application. Figure 4 As shown, the step S3 includes: S31, expanding the charging pile monitoring parameter fusion timing feature matrix to obtain the charging pile monitoring parameter fusion timing feature vector; S32, performing target domain edge matching optimization based on energy capture on the charging pile monitoring parameter fusion timing feature vector to obtain an optimized charging pile monitoring parameter fusion timing feature vector; S33, passing the optimized charging pile monitoring parameter fusion timing feature vector through a safety state classifier to obtain the safety state classification result, and the safety state classification result is used to indicate whether the charging pile is in a safe state.

[0036] In step S31, the charging pile monitoring parameter fusion time series feature matrix is ​​expanded to obtain the charging pile monitoring parameter fusion time series feature vector. It should be understood that the charging pile monitoring parameter fusion time series feature matrix has a high dimension and is not convenient for subsequent calculations. In order to simplify the complexity of subsequent data calculations, it is necessary to expand the charging pile monitoring parameter fusion time series feature matrix. Through the expansion operation, the two-dimensional feature matrix is ​​expanded into a one-dimensional feature vector, which can achieve data dimensionality reduction while retaining the key information of the original feature matrix, that is, improve the calculation efficiency without affecting the safety status of the charging pile.

[0037] In step S32, the charging pile monitoring parameter fusion time series feature vector is optimized by edge matching of the target domain based on energy capture to obtain an optimized charging pile monitoring parameter fusion time series feature vector. In particular, considering that the environmental factors, models, aging degree and other conditions of the charging piles are different in actual applications, the feature distribution obtained by feature extraction of the charging pile monitoring parameters will have different significant differences in different target domains. If the traditional method ignores the differences in feature distribution of different target domains, the extracted features may not accurately reflect the actual situation of the charging pile, resulting in deviations in the judgment of the charging pile status, which will bring safety risks. Based on this, in the technical solution of the present application, it is necessary to perform target domain edge matching optimization on the charging pile monitoring parameter fusion time series feature vector based on energy capture to obtain an optimized charging pile monitoring parameter fusion time series feature vector.

[0038] Specifically, in an embodiment of the present application, the step S32 includes: mapping the charging pile monitoring parameter fusion time series feature vector to the intrinsic decomposition space to obtain the charging pile monitoring parameter fusion time series intrinsic decomposition feature vector; extracting the maximum eigenvalue and the minimum eigenvalue of the charging pile monitoring parameter fusion time series intrinsic decomposition feature vector; calculating the difference between the maximum eigenvalue and the minimum eigenvalue as the charging pile monitoring target domain edge anchor description operator; extracting the mean and standard deviation of the charging pile monitoring parameter fusion time series intrinsic decomposition feature vector, and dividing the mean by the standard deviation to obtain the charging pile monitoring optimization direction description operator; based on the charging pile monitoring target domain edge anchor description operator and the charging pile monitoring optimization direction description operator, matching and optimizing the charging pile monitoring parameter fusion time series feature vector to obtain the optimized charging pile monitoring parameter fusion time series feature vector.

[0039] In an embodiment of the present application, specifically, the charging pile monitoring parameter fusion time series feature vector is optimized based on energy capture target domain edge matching to obtain an optimized charging pile monitoring parameter fusion time series feature vector, including: processing the charging pile monitoring parameter fusion time series feature vector according to the following formula to obtain the optimized charging pile monitoring parameter fusion time series feature vector; wherein the formula is:

[0040]

[0041] α=v max -v min

[0042]

[0043] Wherein, V1 represents the charging pile monitoring parameter fusion time series feature vector, PCA(V1) represents mapping V1 to the intrinsic decomposition space, U1 is the sequence of the charging pile monitoring parameter fusion time series intrinsic decomposition vector, Λ1 is the charging pile monitoring parameter fusion time series diagonal matrix, U1 T is the transpose of U1, v 11 、v 12 、v 1m is the first, second and mth charging pile monitoring parameter fusion time series intrinsic decomposition vector of the sequence of charging pile monitoring parameter fusion time series intrinsic decomposition vector, λ 11 , 1m are the eigenvalues ​​of the first and mth positions of the diagonal matrix of the charging pile monitoring parameter fusion time series, V represents the eigendecomposition eigenvector of the charging pile monitoring parameter fusion time series, v max and v min They represent the maximum eigenvalue and minimum eigenvalue of the characteristic vector of the fusion time series intrinsic decomposition of the charging pile monitoring parameter, α represents the edge anchor description operator of the charging pile monitoring target domain, μ and σ represent the mean and standard deviation of the characteristic vector of the fusion time series intrinsic decomposition of the charging pile monitoring parameter, i represents the description operator of the optimization direction of the charging pile monitoring, ⊙ and They represent point-by-point addition, point-by-point multiplication, and point-by-point subtraction, respectively. exp represents the natural exponential function with the natural constant e as the base. V1 ⊙-1 represents the calculation of the inverse of each eigenvalue of the charging pile monitoring parameter fusion time series feature vector, and V′ represents the optimized charging pile monitoring parameter fusion time series feature vector.

[0044] In view of the above technical problems, in the technical solution of the present application, the target domain edge matching optimization based on energy capture is performed on the charging pile monitoring parameter fusion time series feature vector. The process first maps the charging pile monitoring parameter fusion time series feature vector to the intrinsic decomposition space. It should be understood that through intrinsic decomposition (such as eigenvalue decomposition, singular value decomposition, etc.), the high-dimensional original feature vector is projected into a low-dimensional space composed of orthogonal bases. In this low-dimensional space, the direction is defined by the eigenvector, and the eigenvalue characterizes the variance or intensity distribution of the data in different directions. This process not only achieves dimensionality reduction and removes redundant noise in the original features, but also maps the original vectors of complex distribution to decomposition dimensions with more intuitive physical meaning, making the structured information of the features clearer. For example, in the charging pile monitoring data, there may be some interference information that is irrelevant to the actual safety status. By projecting it to the intrinsic decomposition space, this redundant information can be removed, making the key features more prominent. That is, the subsequent feature data used to judge the safety status of the charging pile is more concise and effective, which can avoid misjudgment caused by noise interference, and make the safety status judgment based on these features more accurate and reliable.

[0045] After the mapping is completed, it is necessary to further extract the maximum eigenvalue and minimum eigenvalue of the eigenvector of the fusion time series intrinsic decomposition of the charging pile monitoring parameter. It should be understood that the maximum eigenvalue reflects the data characteristics of the eigenvector in the most significant direction, represents the proportion of data information in the main direction, and is the core explanatory factor of the characteristic distribution; the minimum eigenvalue is usually related to noise or data errors, and reflects the data characteristics in the least significant direction. In other words, the maximum eigenvalue can help determine the parameter direction that plays a dominant role in judging the safety status of the charging pile. If the direction corresponding to the maximum eigenvalue changes abnormally, it may mean that there is a safety problem with the charging pile. The minimum eigenvalue can assist in identifying possible noise or errors. If it increases abnormally, it may indicate that there is a problem with the measurement system or abnormal interference, affecting the accurate judgment of the safety status.

[0046] In order to construct further feature optimization indicators, it is necessary to calculate the difference between the maximum eigenvalue and the minimum eigenvalue, and define it as the edge anchoring description operator of the target domain of charging pile monitoring. It should be understood that from a geometric point of view, the difference between the maximum and minimum values ​​clarifies the distribution range and span of the charging pile monitoring feature in the target domain, which corresponds to the difference between the major axis and the minor axis in the intrinsic decomposition dimension, and can be regarded as an anchoring indicator of the feature distribution, reflecting the boundary characteristics that may appear in the target domain. By determining the boundary of the feature distribution, it is possible to more accurately judge whether the safety status of the charging pile exceeds the normal range. When the difference of the monitored eigenvalue deviates from the normal range, it may indicate that the charging pile is in an unsafe state, such as overload, short circuit, etc., thus providing an important basis for safety judgment. Moreover, the difference can not only describe the difference in feature data distribution, but also provide a measurement method for capturing the information singularity points that establish boundary performance in the feature space, supporting the subsequent division and adjustment of the target domain features, making the optimized feature vector more in line with the domain edge attributes, and enhancing its adaptability within the domain. For example, under different working modes of the charging pile, the operator can define the range of variation of the characteristic value and help determine the boundary between normal and abnormal states.

[0047] At the same time, it is also necessary to extract the mean and standard deviation from the characteristic vector of the fusion time series intrinsic decomposition of the charging pile monitoring parameters, and calculate their ratio to obtain the charging pile monitoring optimization direction description operator. It should be understood that the mean characterizes the concentration trend of the characteristic value in a specific direction, and the standard deviation describes its discrete characteristics within the overall range. The optimization direction description operator obtained by calculating the ratio of the two provides a normalization mechanism, which enables the feature optimization process to better cope with the imbalance of different data scales and ensure the robustness of the model. In other words, the charging pile monitoring optimization direction description operator can ensure that the characteristics of different monitoring parameters are analyzed under a unified standard to avoid misjudgment caused by differences in data scales. Moreover, in the high-dimensional feature space, the ratio of the mean and the standard deviation further provides a stable basis for directional optimization, ensuring that the optimization process will not be excessively disturbed by extreme values ​​or outliers. In other words, it can also make the judgment process more robust and not affected by individual abnormal data points, so as to more accurately identify the safety status of the charging pile and improve the reliability of safety judgment.

[0048] After the construction of the above key operators is completed, the final matching optimization operation of the charging pile monitoring parameter fusion time series feature vector uses the charging pile monitoring target domain edge anchor description operator and the charging pile monitoring optimization direction description operator to improve the resolution of the feature vector and obtain the optimized representation. It should be understood that by performing matching optimization, the edge attributes and directional attributes of the charging pile monitoring parameter fusion time series feature can be balanced at the same time to avoid overfitting or deviation. Specifically, the charging pile monitoring target domain edge anchor description operator ensures that the charging pile monitoring parameter fusion time series features can better reflect the global and local distribution patterns of the target domain, while the charging pile monitoring optimization direction description operator provides a direction adjustment reference, so that the optimization is promoted in a more effective gradient direction, improving the resolution of the feature vector and obtaining the optimized representation. The optimized charging pile monitoring parameter fusion time series feature vector can more accurately characterize the safety status of the charging pile and improve the system's ability to identify safety hazards of the charging pile.

[0049] In step S33, the optimized charging pile monitoring parameter fusion time series feature vector is passed through a safety state classifier to obtain the safety state classification result, and the safety state classification result is used to indicate whether the charging pile is in a safe state. It should be understood that although the optimized charging pile monitoring parameter fusion time series feature vector contains a large amount of information related to the safety state of the charging pile, this information exists in the form of a feature vector, which is relatively abstract and difficult to directly and intuitively judge the safety state of the charging pile. The role of the safety state classifier is to convert these complex features into simple and clear safety state judgments, that is, to clearly indicate whether the charging pile is in a safe state. The safety state classifier is usually built based on a specific algorithm and has powerful pattern recognition and classification capabilities. It can learn the complex relationship between different features and safety states in the fusion time series feature vector, and establish effective classification rules by training a large amount of charging pile data with known safety states. The safety state classification result clearly tells whether the charging pile is in a safe state. For the operation and maintenance personnel of the charging pile, the safety state classification result is an important basis for making maintenance decisions. If the classification results indicate that the charging pile is in an unsafe state, the operation and maintenance personnel can conduct detailed inspection and repair of the charging pile in a timely manner, locate and resolve potential safety hazards, and ensure the normal operation of the charging pile.

[0050] In summary, the intelligent charging pile safety management method based on the embodiment of the present application is explained, which uses data analysis technology based on deep learning to determine whether the charging pile is in a safe state by performing time series analysis on the insulation resistance value, temperature value, input three-phase voltage value, and output DC voltage value of the charging pile at multiple time points. In this way, the mutual relationship between different charging pile monitoring parameters is fully considered, and the safety state of the charging pile can be judged more accurately, which is conducive to reducing misjudgment and missed judgment.

[0051] Figure 5 : is a system block diagram of the intelligent charging pile safety management system according to an embodiment of the present application. Figure 5 As shown, according to the intelligent charging pile safety management system 100 of the embodiment of the present application, it includes: a charging pile safety monitoring parameter data acquisition module 110, which is used to obtain the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value of the charging pile at multiple predetermined time points within a predetermined time period; a charging pile safety monitoring parameter data analysis module 120, which is used to perform a comprehensive timing analysis on the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value at the multiple predetermined time points to obtain a charging pile monitoring parameter fusion timing feature matrix; a charging pile safety monitoring result generation module 130, which is used to obtain the safety status classification result of the charging pile according to the information in the charging pile monitoring parameter fusion timing feature matrix.

[0052] Here, those skilled in the art can understand that the specific functions and operations of each unit and module in the above-mentioned intelligent charging pile safety management system 100 have been referred to above. Figures 1 to 4 The description of the smart charging pile safety management method has been introduced in detail, and therefore, its repeated description will be omitted.

[0053] In summary, the intelligent charging pile safety management system 100 based on the embodiment of the present application is explained, which uses data analysis technology based on deep learning to determine whether the charging pile is in a safe state by performing time series analysis on the insulation resistance value, temperature value, input three-phase voltage value, and output DC voltage value of the charging pile at multiple time points. In this way, the mutual relationship between different charging pile monitoring parameters is fully considered, and the safety state of the charging pile can be judged more accurately, which is conducive to reducing misjudgment and missed judgment.

[0054] As described above, the smart charging pile safety management system 100 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server for smart charging pile safety management. In one example, the smart charging pile safety management system 100 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the smart charging pile safety management system 100 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the smart charging pile safety management system 100 can also be one of the many hardware modules of the wireless terminal.

[0055] Alternatively, in another example, the smart charging pile safety management system 100 and the wireless terminal may also be separate devices, and the smart charging pile safety management system 100 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

Claims

1. A smart charging pile safety management method, characterized in that: include: Obtaining insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​of the charging pile at multiple predetermined time points within a predetermined time period; Performing a comprehensive time series analysis on the insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix; The safety status classification result of the charging pile is obtained by fusing the information in the time series feature matrix according to the monitoring parameters of the charging pile.

2. The smart charging pile safety management method according to claim 1, characterized in that: Performing a comprehensive time series analysis on the insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix, including: Performing time series feature extraction on the insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter time series feature matrix; Performing timing feature emphasis on the charging pile monitoring parameter timing feature matrix to obtain a charging pile monitoring parameter timing feature emphasis matrix; The charging pile monitoring parameter time series feature matrix and the charging pile monitoring parameter time series feature emphasis matrix are fused to obtain the charging pile monitoring parameter fusion time series feature matrix.

3. The smart charging pile safety management method according to claim 2, characterized in that: The insulation resistance values, temperature values, input three-phase voltage values, and output DC voltage values ​​at the plurality of predetermined time points are subjected to time series feature extraction to obtain a charging pile monitoring parameter time series feature matrix, including: Arrange the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the plurality of predetermined time points according to the time dimension and the sample dimension to obtain a full time-series input matrix of the charging pile monitoring parameters; The charging pile monitoring parameter full time sequence input matrix is ​​passed through the charging pile monitoring parameter time sequence sensor to obtain the charging pile monitoring parameter time sequence characteristic matrix.

4. The smart charging pile safety management method according to claim 3, characterized in that: Performing timing feature emphasis on the charging pile monitoring parameter timing feature matrix to obtain the charging pile monitoring parameter timing feature emphasis matrix, comprising: passing the charging pile monitoring parameter timing feature matrix through a charging pile monitoring parameter key timing feature emphasizer to obtain the charging pile monitoring parameter timing feature emphasis matrix.

5. The smart charging pile safety management method according to claim 4, characterized in that: The charging pile monitoring parameter timing sensor is a convolutional neural network model in which adjacent layers use mutually transposed convolution kernels, and the charging pile monitoring parameter key timing feature emphasizer is a convolutional neural network model using a spatial attention mechanism.

6. The smart charging pile safety management method according to claim 5, characterized in that: According to the information in the time series feature matrix of the charging pile monitoring parameters, a safety status classification result of the charging pile is obtained, including: Expanding the charging pile monitoring parameter fusion time series feature matrix to obtain a charging pile monitoring parameter fusion time series feature vector; Performing target domain edge matching optimization based on energy capture on the charging pile monitoring parameter fusion time series feature vector to obtain an optimized charging pile monitoring parameter fusion time series feature vector; The optimized charging pile monitoring parameter is fused with the time series feature vector and passed through a safety state classifier to obtain the safety state classification result, and the safety state classification result is used to indicate whether the charging pile is in a safe state.

7. The smart charging pile safety management method according to claim 6, characterized in that: The charging pile monitoring parameter fusion time series feature vector is optimized based on energy capture target domain edge matching to obtain an optimized charging pile monitoring parameter fusion time series feature vector, including: Mapping the charging pile monitoring parameter fusion time series feature vector to the intrinsic decomposition space to obtain the charging pile monitoring parameter fusion time series intrinsic decomposition feature vector; Extract the maximum eigenvalue and the minimum eigenvalue of the eigenvalue decomposition eigenvector of the fusion time series intrinsic decomposition of the charging pile monitoring parameter; Calculate the difference between the maximum eigenvalue and the minimum eigenvalue as an edge anchoring description operator of a charging pile monitoring target domain; Extracting the mean and standard deviation of the eigenvalue decomposition eigenvector of the charging pile monitoring parameter fusion time series, and dividing the mean by the standard deviation to obtain a charging pile monitoring optimization direction description operator; Based on the charging pile monitoring target domain edge anchoring description operator and the charging pile monitoring optimization direction description operator, the charging pile monitoring parameter fusion time series feature vector is matched and optimized to obtain the optimized charging pile monitoring parameter fusion time series feature vector.

8. An intelligent charging pile safety management system, characterized in that: include: The charging pile safety monitoring parameter data acquisition module is used to obtain the insulation resistance value, temperature value, input three-phase voltage value and output DC voltage value of the charging pile at multiple predetermined time points within a predetermined time period; A charging pile safety monitoring parameter data analysis module, used to perform a comprehensive time series analysis on the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter fusion time series feature matrix; The charging pile safety monitoring result generation module is used to obtain the safety status classification result of the charging pile according to the charging pile monitoring parameters and the information in the time series feature matrix.

9. The intelligent charging pile safety management system according to claim 8, characterized in that: The charging pile safety monitoring parameter data analysis module includes: A charging pile monitoring parameter time series feature generation unit, used for extracting time series features of the insulation resistance values, temperature values, input three-phase voltage values ​​and output DC voltage values ​​at the plurality of predetermined time points to obtain a charging pile monitoring parameter time series feature matrix; A charging pile monitoring parameter timing feature emphasis unit, used for performing timing feature emphasis on the charging pile monitoring parameter timing feature matrix to obtain a charging pile monitoring parameter timing feature emphasis matrix; The charging pile monitoring parameter feature fusion unit is used to fuse the charging pile monitoring parameter time series feature matrix and the charging pile monitoring parameter time series feature emphasis matrix to obtain the charging pile monitoring parameter fusion time series feature matrix.

10. The intelligent charging pile safety management system according to claim 9, characterized in that: The charging pile safety monitoring result generating module includes: A charging pile monitoring parameter fusion feature dimension reduction unit, used for expanding the charging pile monitoring parameter fusion time series feature matrix to obtain a charging pile monitoring parameter fusion time series feature vector; A charging pile monitoring parameter fusion feature optimization unit, used to perform target domain edge matching optimization based on energy capture on the charging pile monitoring parameter fusion time series feature vector to obtain an optimized charging pile monitoring parameter fusion time series feature vector; The charging pile monitoring parameter fusion feature analysis unit is used to pass the optimized charging pile monitoring parameter fusion time series feature vector through a safety state classifier to obtain the safety state classification result, and the safety state classification result is used to indicate whether the charging pile is in a safe state.

Images