Charging Safety Detection Method, Device, Computer Equipment, and Storage Medium

Abstract

The present application relates to a charging safety detection method, device, computer device, and storage medium. The method includes: obtaining charging data of a charging pile; inputting the charging data into a multi-task prediction model to obtain a safety detection result of the charging pile; wherein the multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one common expert network and multiple private expert networks; the common expert network is used to perform feature extraction and fusion processing on the charging data related to physical faults and virtual faults of the charging pile to obtain a common feature vector; the private expert networks are used to respectively perform feature extraction on the charging data of the charging pile in different scenarios to obtain private feature vectors; the target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the common feature vector and the private feature vectors to obtain a safety detection result. In this way, the relevance of different charging characteristics is considered, and the accuracy of the safety detection result is improved.

Description

Technical Field

[0001] The present application relates to the technical field of charging detection, and particularly to a charging safety detection method, device, computer device, and storage medium. Background Art

[0002] With the popularization of new energy vehicles, charging piles have been more and more widely used to meet the charging needs of new energy vehicles. The charging pile itself is an electrical device and is usually placed outdoors with a relatively high failure rate. When a failure occurs and the charging function cannot be completed, it not only affects the user experience of new energy vehicle users, but also endangers the life and property safety of users in severe cases.

[0003] Currently, the safety diagnosis of charging pile equipment usually describes various fault situations through a fault list to enable maintenance personnel to quickly and accurately locate and diagnose faults. However, as a system integration device, the charging pile integrates various components such as a charging module, a control module, and a power regulation module. The faults that may occur during the charging process are diverse, and the faults may occur on different components at different time periods, that is, in different scenarios. Relying on the fault list for maintenance affects the timeliness and accuracy of fault diagnosis. Moreover, different faults affect each other, and the correlation between different fault factors cannot be considered. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a charging safety detection method, device, computer device, and storage medium.

[0005] In a first aspect, the present application provides a charging safety detection method, and the method includes:

[0006] Obtain the charging data of the charging pile; the charging data includes charging characteristics of different feature types;

[0007] Input the charging data into a multi-task prediction model to obtain the safety detection result of the charging pile;

[0008] Wherein, the multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one common expert network and multiple private expert networks; the common expert network is used to perform feature extraction and fusion processing on the charging data related to the physical faults and virtual faults of the charging pile to obtain a common feature vector; the private expert network is used to respectively perform feature extraction on the charging data of the charging pile in a specific scenario and / or under a specific task to obtain a private feature vector; the target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the common feature vector and the private feature vector to obtain the safety detection result.

[0009] In one embodiment, the multi-task detection model further includes a gating network; each of the expert networks is connected to a plurality of gating networks; the gating networks are connected to the target prediction network one by one; the gating network is used to obtain a weighting coefficient according to the common feature vector and / or the private feature vector; the weighting coefficient corresponds to the weight of each feature vector.

[0010] In one embodiment, the method further includes:

[0011] Calculating based on the first input feature and the first output feature of the common expert network to obtain first correlation information; wherein, the first input feature includes charging data related to physical faults and charging data related to virtual faults of the charging pile; the first output feature is the common feature vector;

[0012] Maximizing the first correlation information to obtain a first loss function based on the common expert network;

[0013] Calculating based on the first output feature and the second output features of each private expert network to obtain second correlation information; wherein, the second output feature is the private feature vector;

[0014] Minimizing the second correlation information to obtain a second loss function based on the private expert network;

[0015] Determining a target loss function of the multi-task prediction model according to the first loss function and the second loss function.

[0016] In one embodiment, the method further includes:

[0017] In the case where the safety detection result is a charging pile fault, obtaining fault characteristic parameters corresponding to the safety detection result; the fault characteristic parameters include at least one of a fault type, a fault location, and a fault range;

[0018] Generating a reminder task according to the fault characteristic parameters and determining the priority corresponding to the reminder task;

[0019] Outputting the alarm information corresponding to the reminder task in sequence according to the priority of the reminder task.

[0020] In one embodiment, the obtaining the fault characteristic parameters corresponding to the safety detection result includes:

[0021] Obtaining the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value that occur in the same charging pile within a preset time period;

[0022] Generating a reminder task according to the fault characteristic parameters and determining the priority corresponding to the reminder task includes:

[0023] Determining the priority corresponding to the reminder task according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value.

[0024] In one embodiment, determining the priority corresponding to the reminder task according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value includes:

[0025] Determining the charging piles with the number of faults greater than or equal to a first threshold as the first candidate charging piles;

[0026] Screening out the charging piles with the number of charging failures greater than or equal to a second threshold from the first candidate charging piles as the second candidate charging piles; wherein, the priority corresponding to the reminder task of the second candidate charging piles is higher than the priority corresponding to the reminder task of the first candidate charging piles;

[0027] Sorting the second candidate charging piles according to at least one of the charging fault temperature and the charging fault current value, and determining the priority corresponding to the reminder task of each second candidate charging pile in sequence according to the sorting result.

[0028] In one embodiment, sequentially outputting the alarm information corresponding to the reminder task includes at least one of the following:

[0029] The first type: sequentially outputting the alarm information through the voice device of the charging pile by voice;

[0030] The second type: sending the alarm information to a third-party terminal; the alarm information is used for display by the third-party terminal;

[0031] The third type: sending the alarm information to the third-party terminal; the alarm information is used for output by the motor vibration based on the third-party terminal.

[0032] In a second aspect, the present application also provides a charging safety detection device, and the device includes:

[0033] An acquisition module, configured to acquire the charging data of the charging pile; the charging data includes charging characteristics of different feature types;

[0034] A detection module, configured to input the charging data into a multi-task prediction model to obtain a safety detection result of the charging pile;

[0035] Among them, the multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one common expert network and multiple private expert networks; the common expert network is used to perform feature extraction and fusion processing on the charging data related to the physical faults of the charging pile and the charging data related to the virtual faults, so as to obtain a common feature vector; the private expert network is used to perform feature extraction on the charging data of the charging pile under specific scenarios and / or specific tasks respectively, so as to obtain a private feature vector; the target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the common feature vector and the private feature vector, so as to obtain a safety detection result.

[0036] In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the method described in any embodiment of the embodiments of the present application is implemented.

[0037] In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the method described in any embodiment of the present application is implemented.

[0038] The above-mentioned charging safety detection method, device, computer device, storage medium and computer program product can adapt to the charging characteristics corresponding to different scenarios / tasks by obtaining the charging data related to the physical faults of the charging pile and the charging data related to the virtual faults and inputting them into the multi-task prediction model to determine the safety detection result of the charging pile (for example, the probability of the charging pile having a fault), without the need to perform separate algorithm modeling for each characteristic, improving the processing performance; and, the correlation between different types of charging characteristics can be considered to improve the accuracy of the safety detection result. Description of the Drawings

[0039] Figure 1 is an application environment diagram of the charging safety detection method shown according to an exemplary embodiment;

[0040] Figure 2 is a flowchart of the charging safety detection method shown according to an exemplary embodiment;

[0041] Figure 3 is a structural diagram of the multi-task detection model shown according to an exemplary embodiment;

[0042] Figure 4 is a block diagram of the structure of the charging safety detection device shown according to an exemplary embodiment;

[0043] Figure 5 is an internal structure diagram of the computer device shown according to an exemplary embodiment. Detailed Embodiments

[0044] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] The terms "first", "second", and "third" in the embodiments of the present application are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first", "second", and "third" may explicitly or implicitly include at least one of such features. In the description of the present application, "at least one" is used to indicate one or more; the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0046] Referring to "embodiments" herein means that specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0047] The charging safety detection method provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the computer device 102 communicates with the server 104 through a network; optionally, the network can be a wired network or a wireless network. The charging pile 106 can obtain the charging data that the computer device 102 needs to process in real time during the charging process of the target vehicle. Among them, the computer device 102 can be any mobile terminal or fixed terminal. The terminal can be a device that provides voice and / or data connectivity to the user. Exemplarily, the terminal can be an Internet of Things terminal, such as a sensor device, a mobile phone (or referred to as a "cellular" phone), and a computer with an Internet of Things terminal. For example, it can be a fixed, portable, pocket-sized, handheld, computer-built-in or in-vehicle device. The server 104 can be an independent server or a server cluster composed of multiple servers, and can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0048] In one embodiment, as Figure 2 shown, a charging safety detection method is provided. Taking the terminal in Figure 1 as an example, the method includes the following steps:

[0049] Step S202, obtaining the charging data of the charging pile; the charging data includes charging characteristics of different feature types.

[0050] In the embodiment of the present application, the charging data indicates a set of information used to reflect various states during the charging process of the charging pile.

[0051] Optionally, the charging data can include at least one of historical charging data, real-time charging data, and charging limit data.

[0052] Exemplarily, the charging limit data can include at least one of charging power limit data, charging mode limit data, charging temperature limit data, and single charging time limit data. For example, the charging pile can support multiple charging modes, and the charging modes can include fast charging mode, slow charging mode, constant current charging mode, or constant voltage charging mode, etc. The charging pile can match a suitable charging mode according to the model of the target vehicle for charging.

[0053] In some embodiments, the charging data may include charging characteristics of different feature types. Optionally, the charging data may include, but is not limited to, time-series charging data, space-series charging data, and user behavior charging data; optionally, the charging data may include charging data related to physical faults and charging data related to virtual faults. Among them, the charging data related to physical faults may include, but is not limited to, at least one of voltage detection data, meter communication data, operating data of internal components of the charging pile (such as control boards, power modules, etc.), operating data of external components of the charging pile (such as plugs, sockets, charging cables, etc.), and external environment data; the charging data related to virtual faults may include, but is not limited to, at least one of operating data of internal software of the charging pile, system upgrade or update data of the charging pile, and network communication / configuration data. According to the charging data related to physical faults, the computer device can determine whether the voltage at the detection point of the charging pile is normal, whether the internal components of the charging pile are damaged, whether the charging cable or connector is faulty, and whether there is a device fault caused by an external extreme environment, etc.; according to the charging data related to virtual faults, the computer device can determine whether there is a fault in the internal software of the charging pile, whether there is a fault in the system update or upgrade, and whether the communication between the charging pile and the cloud server is normal, etc.

[0054] Step S204, input the charging data into the multi-task prediction model to obtain the safety detection result of the charging pile;

[0055] Among them, the multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one common expert network and multiple private expert networks; the common expert network is used to perform feature extraction and fusion processing on the charging data related to physical faults and the charging data related to virtual faults of the charging pile to obtain a common feature vector; the private expert network is used to respectively perform feature extraction on the charging data of the charging pile in a specific scenario and / or a specific task to obtain a private feature vector; the target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the common feature vector and the private feature vector to obtain the safety detection result.

[0056] In the embodiments of the present application, the multi-task detection model can complete multiple related or unrelated tasks at the same time, sharing the task feature extraction layer to improve the calculation efficiency and resource utilization rate.

[0057] In some embodiments, the multi-task detection model may include at least one of, but not limited to, a Multi-gate Mixture-of-Experts (MMoE), a Customized Gate Control (CGC), or a Progressive Layered Extraction (PLE).

[0058] In the embodiments of the present application, the expert network may be a feature extraction network, which can extract some or all of the features of the charging data input into the expert network.

[0059] In some embodiments, the charging habits of different users vary greatly. For example, users have different preferences for charging times. Some users tend to charge at night, while some users tend to charge during the day, corresponding to different charging scenarios; or, due to the wide variety of new energy vehicles, with different battery capacities, chemical compositions, and health conditions, the corresponding charging scenarios are also different; or, the grid load is different at different times. For example, the grid load is relatively large during peak hours, and the output power of the charging pile may correspond to different charging scenarios at different times.

[0060] In some embodiments, a specific task may include, but not be limited to, one of a task of paying attention to voltage anomalies, a task of paying attention to current fluctuations, and a task of paying attention to charging temperature fluctuations.

[0061] In some embodiments, there may be multiple expert networks, which are used to learn higher-order cross features from input features for fitting subsequent tasks. Generally, there should be at least two expert networks, and different expert networks can learn different types of cross features. Among them, the expert network can include a common expert network and private expert networks. The common expert network can include one or more fully connected networks arranged in parallel to extract and fuse features from the charging data to obtain a common feature vector; each private expert network processes the charging data under a specific scenario and / or a specific task to obtain a corresponding private feature vector; for example, the first private expert network receives the charging data corresponding to the task of focusing on voltage anomalies and obtains the corresponding first private feature vector; the second private expert network receives the charging data corresponding to the task of focusing on current fluctuations and obtains the corresponding second private feature vector; the third private expert network receives the charging data corresponding to the task of focusing on charging temperature fluctuations and obtains the corresponding third private feature vector; the target prediction network fuses the common feature vector and the private feature vectors and performs multi-task prediction to obtain the safety detection result corresponding to the charging pile. For example, the PLE model can be used as the multi-task prediction model, which helps to utilize shared features in the early stage and focus on feature specificity in the later stage, reducing interference between tasks. Exemplarily, as Figure 3 shown Figure 3 is a schematic structural diagram of the multi-task prediction model. The multi-task prediction model can include multiple layers of expert networks. Each layer of the expert network can include a common expert network and private expert networks corresponding to two tasks / scenarios. The bottom-layer expert network processes the original input (i.e., input features) or the output of the previous layer to extract low-level features; the high-layer expert network further extracts higher-level and more abstract features based on the bottom-layer output. Expert networks of different layers can capture information at different levels; the input features are the charging data after data preprocessing, and the first output feature and the second output feature are the safety detection results.

[0062] In the above charging safety detection method, by obtaining the charging data related to physical faults and virtual faults of the charging pile and inputting them into the multi-task prediction model to determine the safety detection result of the charging pile (for example, the probability of the charging pile having a fault), it can adapt to the charging characteristics corresponding to different scenarios / tasks, without the need to perform separate algorithm modeling for each feature, improving the processing performance; and, it can consider the correlation between different types of charging features and improve the accuracy of the safety detection result.

[0063] In one embodiment, the multi-task detection model further includes a gating network; each of the expert networks is connected to a plurality of gating networks; the gating networks are connected to the target prediction network one by one; the gating network is configured to obtain a weighting coefficient according to the common feature vector and / or the private feature vector; the weighting coefficient corresponds to the weight of each feature vector.

[0064] In the embodiments of the present application, the gating network screens according to the importance and task relevance of the feature vectors extracted by the expert network, classifies or normalizes some or all of the feature vectors, and outputs the weighting coefficients of each feature vector.

[0065] In some embodiments, after the expert network extracts the common feature vector and the private feature vector, they can be transmitted to the corresponding gating network to determine the corresponding attention weights; the gating network can dynamically adjust the weight distribution between different scenarios / tasks, so as to better adapt to the differences between different scenarios / tasks.

[0066] In some embodiments, the gating network can classify or normalize some or all of the common feature vector and the private feature vector output by the expert network, and output the weighting coefficients corresponding to each feature vector; the target prediction network can perform weighted processing on the feature vector with the corresponding weighting coefficient, for example, multiplying the first private feature vector by the weighting coefficient corresponding to the first private feature vector to obtain the weighted first private feature vector; multiplying the second private feature vector by the weighting coefficient corresponding to the second private feature vector to obtain the weighted second private feature vector; and so on until the weighted feature vectors are obtained; and, the target prediction network fuses the weighted feature vectors to obtain a target feature vector, and further performs classification / regression processing to obtain a safety detection result.

[0067] In this embodiment, the gating network allows the model to focus on the most useful information at an early stage, reduces redundant calculations, and helps to reduce interference between tasks; thus, it is applicable to charging safety prediction with fine control at the input level.

[0068] In some embodiments, the method further includes:

[0069] Calculating according to the first input feature and the first output feature of the common expert network to obtain first relevant information; wherein, the first input feature includes charging data related to physical faults of the charging pile and charging data related to virtual faults; the first output feature is the common feature vector;

[0070] Maximizing the first relevant information to obtain a first loss function based on the common expert network;

[0071] Calculations are performed based on the first output feature and the second output features of each of the private expert networks to obtain second relevant information; wherein, the second output feature is the private feature vector.

[0072] Minimize the second relevant information to obtain a second loss function based on the private expert network.

[0073] Determine the target loss function of the multi-task prediction model according to the first loss function and the second loss function.

[0074] In some embodiments, a computer device can measure the information correlation between two arbitrary random variables according to the mutual information (MI) method, which can well measure the degree of mutual dependence between two random variables. MI maximization can maximize the association between two random variables, and MI minimization can minimize the association between two variables.

[0075] Exemplarily, a way to determine mutual information is: ; wherein, denotes the first variable; denotes the second variable; denotes the joint distribution of the first variable and the second variable; and denotes the marginal distributions of the first variable and the second variable.

[0076] Exemplarily, a way to estimate the lower bound between two variables is: ; wherein, denotes the variational lower bound; denotes the auxiliary distribution; denotes the joint distribution of the first variable and the second variable; and denotes the marginal distributions of the first variable and the second variable; denotes the expectation of the joint distribution; denotes the expectation of the marginal distribution; .

[0077] Exemplarily, when maximizing the mutual information between variables to obtain the first loss function, the first variable can be replaced with the first input feature, and the second variable can be replaced with the first output feature; a way to determine the first loss function is: .

[0078] Exemplarily, when minimizing the mutual information between variables to obtain the second loss function, the first variable can be replaced with the first output feature, and the second variable can be replaced with the second output feature; one way to determine the second loss function is as follows: 。

[0079] In one embodiment, the computer device may use the weighted sum value of the first loss function, the second loss function, and the third loss function of the target prediction network as the target loss function; wherein, the third loss function can be used to indicate the error between the safety training result obtained by the multi-task prediction model based on the charging training sample data (with the safety detection result labeled) during the training process and the actual safety detection result of each charging training sample data; for example, the third loss function can be a cross-entropy function.

[0080] In the embodiments of the present application, through the first relevant information and the second relevant information related to the common expert network in the target loss function, the training of the multi-task prediction model can be well guided, information redundancy can be reduced, and the model performance and the accuracy of the safety detection result can be improved.

[0081] In some embodiments, the method further includes:

[0082] In the case where the safety detection result is a charging pile failure, obtain the fault feature parameters corresponding to the safety detection result; the fault feature parameters include at least one of the fault type, fault location, and fault range;

[0083] Generate a reminder task according to the fault feature parameters and determine the priority corresponding to the reminder task;

[0084] Output the alarm information corresponding to the reminder task in sequence according to the priority of the reminder task.

[0085] In the embodiments of the present application, the fault type may include at least one of, but not limited to, communication failure, electrical failure, mechanical failure, control logic failure, sensor failure, and display interface failure.

[0086] In the embodiments of the present application, the fault location may include at least one of, but not limited to, the internal components, external interfaces, internal software, and remote services (systems not located locally such as cloud services) of the charging pile.

[0087] In the embodiments of the present application, the fault range may include at least one of, but not limited to, local fault, systematic fault, temporary fault, and permanent fault.

[0088] In some embodiments, the fault feature parameters may further include at least one of, but not limited to, the number of faults, the number of charging failures, the charging fault temperature, the charging fault current value, and the number of communication failures.

[0089] In some embodiments, the computer device may determine the priority of reminder tasks corresponding to each fault characteristic parameter according to at least one of the importance level of the fault characteristic parameter, the order of occurrence time, and the urgency level. Exemplarily, the fault characteristic parameter of the first charging pile indicates that the number of faults is 1 and the fault type is a display interface fault; the fault characteristic parameter of the second charging pile indicates that the number of anomalies is 1 and the fault type is an electrical fault; the urgency level of the electrical fault is higher than that of the display interface fault. The computer device may determine that the priority of the reminder task of the second charging pile is higher than that of the reminder task of the first charging pile; the computer device may preferentially output the alarm information corresponding to the reminder task of the second charging pile.

[0090] In the embodiments of the present application, by outputting the alarm information corresponding to the reminder task, the safety detection result of the charging pile can be intuitively determined, which is convenient for the operator to directly determine the location, urgency level, and fault type of the abnormal fault in the charging pile. Therefore, it is beneficial for the operator to make corresponding processing in a timely manner according to the alarm information, reduce the inconvenience of use brought to users by the charging pile fault, and improve the user experience; moreover, it can also reduce the safety risks caused by the charging pile fault and ensure charging safety.

[0091] In some embodiments, the sequentially outputting the alarm information corresponding to the reminder task includes at least one of the following:

[0092] The first one: sequentially outputting the alarm information based on the voice of the voice device of the charging pile;

[0093] The second one: sending the alarm information to a third-party terminal; the alarm information is used for display on the third-party terminal;

[0094] The third one: sending the alarm information to the third-party terminal; the alarm information is used for output based on the motor vibration of the third-party terminal.

[0095] In some embodiments, the alarm information may be output in the form of a table, a list, a graph, a voice, or a vibration.

[0096] In some embodiments, the alarm information may be output in any form on the display component of the charging pile; or, the alarm information may be sent to a third-party terminal / platform and output on the third-party terminal / platform.

[0097] In the embodiments of the present application, by outputting the alarm information, the operator can be reminded in a timely and intuitive manner that the charging pile has a fault or anomaly, reducing the safety risks caused by the charging pile fault and reducing the failure rate of the charging pile.

[0098] In some embodiments, the obtaining of the fault characteristic parameter corresponding to the safety detection result includes:

[0099] Obtain the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value that occur in the same charging pile within a preset time period;

[0100] The generating a reminder task according to the fault characteristic parameters and determining the priority corresponding to the reminder task includes:

[0101] Determine the priority corresponding to the reminder task according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value.

[0102] In the embodiment of the present application, the number of faults can indicate the number of faults that occur in a charging pile. For example, if the first charging pile has both a communication fault and a display fault at the same time, the number of faults can be determined to be 2.

[0103] In the embodiment of the present application, the number of charging failures can indicate the number of charging failures that occur in the charging pile within a predetermined time period.

[0104] In the embodiment of the present application, the charging fault temperature can indicate that the temperature during charging of the charging pile exceeds the reference temperature range. For example, the reference temperature range is -20°C (degrees Celsius) to 50°C, and the current charging temperature of the charging pile is 60°C, then the current charging temperature can be determined to be the charging fault temperature.

[0105] In the embodiment of the present application, the charging fault current value can indicate that the charging current during charging of the charging pile exceeds the normal current range. For example, for a household charging pile, the maximum charging power is 7KW, and the normal current range is 10A (amperes) to 20A; the current charging current of the charging pile is 5A, then the current charging current can be determined to be the charging fault current value.

[0106] In some embodiments, since there can be multiple faulty charging piles and there can be multiple fault characteristic parameters of the faulty charging piles, in order to better display the warning information with a higher emergency level or a higher importance level, and facilitate the operator to process the corresponding faulty charging piles in a timely manner according to the warning information, the priority corresponding to the reminder task can be determined according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value; output the warning information corresponding to the reminder task in order according to the priority.

[0107] Exemplarily, the computer device can sort each faulty charging pile in descending order of the number of charging failures, determine the priority of the reminder task corresponding to each faulty charging pile; and display the warning information corresponding to each reminder task in sequence according to the priority.

[0108] In some embodiments, the computer device may assign weights to the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value respectively according to the importance of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value; the computer device scores each faulty charging pile in turn according to the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value to obtain a first score, a second score, a third score, and a fourth score; according to the first score and the corresponding weighted weight, the second score and the corresponding weighted weight, the third score and the corresponding weighted weight, and the fourth score and the corresponding weighted weight, determine the target score corresponding to the reminder task of the faulty charging pile; the computer device determines the target score corresponding to the reminder task of each faulty charging pile in turn, and sorts them from largest to smallest or from smallest to largest according to the target score to determine the priority.

[0109] In the embodiments of the present application, by dividing the priority of the reminder tasks of each faulty charging pile, the computer device can preferentially display the reminder tasks and corresponding alarm information of the faulty charging piles with higher fault severity, urgency, and / or importance, so that the operator can perform relevant processing such as maintenance in time, reducing the safety risks caused by faults.

[0110] In some embodiments, determining the priority corresponding to the reminder task according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value includes:

[0111] Determine the charging piles with the number of faults greater than or equal to the first threshold as the first candidate charging piles;

[0112] Screen out the charging piles with the number of charging failures greater than or equal to the second threshold from the first candidate charging piles as the second candidate charging piles; wherein, the priority corresponding to the reminder task of the second candidate charging piles is higher than the priority corresponding to the reminder task of the first candidate charging piles;

[0113] Sort the second candidate charging piles according to at least one of the charging fault temperature and the charging fault current value, and determine the priority corresponding to the reminder task of each second candidate charging pile in turn according to the sorting result.

[0114] In some embodiments, the computer device may sort the faulty charging piles according to the number of faults in descending order, and determine the faulty charging piles with the number of faults greater than or equal to the first threshold as the first candidate charging piles; for example, if the number of faults of the first to third faulty charging piles are 2, 3, and 3 respectively, and the first threshold is 3, then the second and third faulty charging piles can be determined as the first candidate charging piles. The computer device may sort according to the number of charging failures within a preset time period in descending order, and screen out the charging piles with the number of charging failures greater than or equal to the second threshold from the first candidate charging piles as the second candidate charging piles; for example, if the number of charging failures of the second faulty charging pile is 3 times, the number of charging failures of the third faulty charging pile is 1 time, and the second threshold is 2, then the second faulty charging pile can be determined as the second candidate charging pile. The computer device may determine that the priority of the reminder task corresponding to the second faulty charging pile is higher than the priority of the reminder task corresponding to the third faulty charging pile, and the priority of the reminder task corresponding to the third faulty charging pile is higher than the priority of the reminder task corresponding to the first faulty charging pile.

[0115] In some embodiments, the computer device may sort the second candidate charging piles in descending order according to the difference value between the charging fault temperature and the reference temperature range; or, the computer device may sort the second candidate charging piles in descending order according to the difference value between the charging fault current value and the normal current range, and determine the second candidate charging piles with the difference value greater than or equal to the third threshold as the target charging piles, and then the priority of the reminder task corresponding to the target charging piles can be determined to be the highest.

[0116] In the embodiments of the present application, the priorities of the reminder tasks corresponding to the faulty charging piles are determined in various ways. Compared with only considering single factors, charging piles with higher severity, urgency, and / or importance of faults, that is, higher priorities of reminder tasks, can be screened out from multiple dimensions, which can meet the charging safety detection requirements in different scenarios.

[0117] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least some of the steps or stages in other steps or other steps.

[0118] Based on the same inventive concept, an embodiment of the present application further provides a charging safety detection device for implementing the charging safety detection method involved above. The implementation solutions provided by this device to solve problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the charging safety detection device provided below can refer to the limitations on the charging safety detection method in the above text, and will not be repeated here.

[0119] In one embodiment, as Figure 4 shown, a charging safety detection device is provided, including:

[0120] An acquisition module 10, configured to acquire charging data of a charging pile; the charging data includes charging characteristics of different feature types;

[0121] A detection module 20, configured to input the charging data into a multi-task prediction model to obtain a safety detection result of the charging pile;

[0122] Among them, the multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one common expert network and multiple private expert networks; the common expert network is configured to perform feature extraction and fusion processing on the charging data related to the physical faults of the charging pile and the charging data related to the virtual faults to obtain a common feature vector; the private expert network is configured to perform feature extraction on the charging data of the charging pile in different scenarios respectively to obtain a private feature vector; the target prediction network is configured to perform multi-task prediction on the target feature vector obtained by fusing the common feature vector and the private feature vector to obtain a safety detection result.

[0123] In some embodiments, the multi-task detection model further includes a gating network; each expert network is connected to multiple gating networks; the gating networks are connected to the target prediction network one by one; the gating network is configured to obtain a weighting coefficient according to the common feature vector and / or the private feature vector; the weighting coefficient corresponds to the weight of the feature vector of each feature type.

[0124] In some embodiments, the charging safety detection device further includes:

[0125] A first calculation module, configured to calculate according to the first input feature and the first output feature of the common expert network to obtain first relevant information; wherein, the first input feature includes the charging data related to the physical faults of the charging pile and the charging data related to the virtual faults; the first output feature is the common feature vector;

[0126] A processing module, configured to maximize the first relevant information to obtain a first loss function based on the common expert network;

[0127] A second computing module, configured to perform computations based on the first output feature and the second output features of each private expert network to obtain second relevant information; wherein, the second output feature is a private feature vector;

[0128] A processing module, configured to minimize the second relevant information to obtain a second loss function based on the private expert network;

[0129] The processing module is further configured to determine an objective loss function of the multi-task prediction model according to the first loss function and the second loss function.

[0130] In some embodiments, the charging safety detection device further includes:

[0131] A first processing module, configured to obtain fault characteristic parameters corresponding to the safety detection result in the case where the safety detection result is a charging pile fault; the fault characteristic parameters include at least one of a fault type, a fault location, and a fault range;

[0132] A determination module, configured to generate a reminder task according to the fault characteristic parameters and determine the priority corresponding to the reminder task;

[0133] An output module, configured to sequentially output warning information according to the priority of the reminder task.

[0134] In some embodiments, the first processing module is configured to obtain the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value that occur in the same charging pile within a preset time period;

[0135] The determination module is configured to determine the priority corresponding to the reminder task according to at least one of the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value.

[0136] In some embodiments, the determination module is configured to perform the following steps:

[0137] Determine a charging pile with the number of faults greater than or equal to a first threshold as a first candidate charging pile;

[0138] Screen out a charging pile with the number of charging failures greater than or equal to a second threshold from the first candidate charging piles as a second candidate charging pile; wherein, the priority corresponding to the reminder task of the second candidate charging pile is higher than the priority corresponding to the reminder task of the first candidate charging pile;

[0139] Sort the second candidate charging piles according to at least one of the charging fault temperature and the charging fault current value, and sequentially determine the priorities corresponding to the reminder tasks of each second candidate charging pile according to the sorting result.

[0140] In some embodiments, the output module is configured to perform at least one of the following steps:

[0141] The first type: The voice device based on the charging pile sequentially outputs warning messages by voice;

[0142] The second type: Send the warning message to a third-party terminal; the warning message is used for display on the third-party terminal;

[0143] The third type: Send the warning message to a third-party terminal; the warning message is used for motor vibration output based on the third-party terminal.

[0144] Each module in the above charging safety detection device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0145] In one embodiment, a computer device is provided. The computer device can be a charging pile, and its internal structure diagram can be as Figure 5 shown. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. The computer program, when executed by the processor, implements a charging safety detection method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0146] Those skilled in the art can understand that Figure 5 the structure shown in

[0147] is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0148] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0149] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0150] The above embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A charging safety detection method, characterized in that: The method comprises: Acquiring charging data of a charging pile; the charging data including charging features of different feature types; Inputting the charging data into a multi-task prediction model to obtain a safety detection result of the charging pile; The multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one public expert network and multiple private expert networks; the public expert network is used to extract features and fuse charging data related to physical faults and charging data related to virtual faults of the charging pile to obtain a public feature vector; the private expert network is used to extract features from charging data of the charging pile in specific scenarios and / or specific tasks to obtain a private feature vector; Calculating according to the first input feature and the first output feature of the public expert network to obtain first relevant information; wherein the first input feature includes charging data related to the physical fault of the charging pile and charging data related to the virtual fault; the first output feature is the public feature vector; Maximizing the first relevant information to obtain a first loss function based on the public expert network; Calculating according to the first output feature and the second output feature of each of the private expert networks to obtain second related information; wherein the second output feature is the private feature vector; Minimizing the second relevant information to obtain a second loss function based on the private expert network; Determining a target loss function of the multi-task prediction model according to the first loss function and the second loss function; The target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the public feature vector and the private feature vector to obtain the security detection result; In the case where the safety detection result is a charging pile fault, obtaining a fault characteristic parameter corresponding to the safety detection result; the fault characteristic parameter includes at least one of a fault type, a fault location, and a fault range; obtaining the fault characteristic parameter corresponding to the safety detection result includes: obtaining the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value occurring in the same charging pile within a preset time period; Generating a reminder task and determining a priority corresponding to the reminder task according to the fault characteristic parameter includes: Determine the charging pile whose number of faults is greater than or equal to a first threshold as a first candidate charging pile; Selecting a charging pile whose number of charging failures is greater than or equal to a second threshold from the first candidate charging piles as a second candidate charging pile; wherein the priority corresponding to the reminder task of the second candidate charging pile is higher than the priority corresponding to the reminder task of the first candidate charging pile; sorting the second candidate charging piles according to at least one of the charging fault temperature and the charging fault current value, and determining the priority corresponding to the reminder task of each of the second candidate charging piles in turn according to the sorting result; According to the priority of the reminder task, the alarm information corresponding to the reminder task is output in sequence.

2. The method according to claim 1, characterized in that The multi-task prediction model also includes a gating network; each of the expert networks is connected to multiple gating networks; the gating networks are connected to the target prediction network one-to-one; the gating network is used to obtain weighting coefficients based on the public feature vectors and / or the private feature vectors; the weighting coefficients correspond to the weights of each feature vector.

3. The method according to claim 1, characterized in that The sequentially outputting the alarm information corresponding to the reminder task includes at least one of the following: The first method is to output the warning information in sequence based on the voice of the voice device of the charging pile; The second method: sending the alarm information to a third-party terminal; the alarm information is used for display by the third-party terminal; The third type: sending the alarm information to the third-party terminal; the alarm information is used for motor vibration output based on the third-party terminal.

4. A charging safety detection device, characterized in that: The device comprises: An acquisition module is used to acquire charging data of a charging pile; the charging data includes charging features of different feature types; The detection module is used to input charging data into the multi-task prediction model to obtain the safety detection results of the charging pile; The multi-task prediction model includes an expert network and a target prediction network; the expert network includes at least one public expert network and multiple private expert networks; the public expert network is used to extract and fuse the charging data related to the physical fault of the charging pile and the charging data related to the virtual fault to obtain a public feature vector; the private expert network is used to extract the features of the charging data of the charging pile in a specific scenario and / or a specific task to obtain a private feature vector; Calculating according to the first input feature and the first output feature of the public expert network to obtain first relevant information; wherein the first input feature includes charging data related to the physical fault of the charging pile and charging data related to the virtual fault; the first output feature is the public feature vector; Maximizing the first relevant information to obtain a first loss function based on the public expert network; Calculating according to the first output feature and the second output feature of each of the private expert networks to obtain second related information; wherein the second output feature is the private feature vector; Minimizing the second relevant information to obtain a second loss function based on the private expert network; Determining a target loss function of the multi-task prediction model according to the first loss function and the second loss function; The target prediction network is used to perform multi-task prediction on the target feature vector obtained by fusing the public feature vector and the private feature vector to obtain the security detection result; In the case where the safety detection result is a charging pile fault, obtaining a fault characteristic parameter corresponding to the safety detection result; the fault characteristic parameter includes at least one of a fault type, a fault location, and a fault range; obtaining the fault characteristic parameter corresponding to the safety detection result includes: obtaining the number of faults, the number of charging failures, the charging fault temperature, and the charging fault current value occurring in the same charging pile within a preset time period; Generating a reminder task and determining a priority corresponding to the reminder task according to the fault characteristic parameter includes: Determine the charging pile whose number of faults is greater than or equal to a first threshold as a first candidate charging pile; Selecting a charging pile whose number of charging failures is greater than or equal to a second threshold from the first candidate charging piles as a second candidate charging pile; wherein the priority corresponding to the reminder task of the second candidate charging pile is higher than the priority corresponding to the reminder task of the first candidate charging pile; sorting the second candidate charging piles according to at least one of the charging fault temperature and the charging fault current value, and determining the priority corresponding to the reminder task of each of the second candidate charging piles in turn according to the sorting result; According to the priority of the reminder task, the alarm information corresponding to the reminder task is output in sequence.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images