Power equipment state monitoring and fault early warning method and system based on edge computing

Abstract

The present application relates to the technical field of power equipment monitoring, and particularly relates to a power equipment state monitoring and fault early warning method and system based on edge computing, wherein the method classifies corresponding nodes according to running data after obtaining a large amount of running data, clusters multiple running data in the classification, and obtains a cluster center. The cluster center of multiple nodes is input into a data analysis and early warning model to obtain early warning information. After clustering the data, the data volume is greatly reduced while retaining as many features of the running data as possible, thereby reducing the data transmission volume and computation amount. The clustering is based on nodes, the early warning result is positioned to a specific node, the power equipment monitoring data transmission volume is reduced, the feature loss is reduced as much as possible, the data analysis and early warning model computation amount is reduced, the operation speed is accelerated, and the early warning information is positioned to the equipment node, thereby providing necessary data support for power equipment operation and maintenance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment monitoring, in particular to a power equipment state monitoring and fault early warning method and system based on edge computing. BACKGROUND

[0002] Power equipment mainly refers to various voltage level switching devices, transformers, mutual inductors, arresters, reactors, transmission lines and the like. When power equipment fails, the maintenance of the power equipment will bring great trouble to production and life.

[0003] Taking GIS equipment as an example, GIS equipment is a high-voltage electrical equipment with high operation reliability, less maintenance workload and long maintenance period, and its failure rate is only 20% to 40% of that of conventional equipment. However, due to the leakage of SF6 gas, the infiltration of external moisture, the existence of conductive impurities, the aging of insulators and other factors, internal flashover failure of GIS may occur. Because GIS uses a fully sealed structure, it is difficult to locate and maintain GIS equipment failure, and the maintenance work is complicated. The average outage maintenance time after an accident is longer than that of conventional equipment, and the outage range is large, involving many non-fault elements.

[0004] For another example, the transformer is the hub equipment of the power system, which undertakes the heavy responsibility of voltage transformation, power distribution and transfer. The normal operation of the transformer is an important guarantee for the safe, reliable, high-quality and economic operation of the power system. On the other hand, the transformer also has certain complexity and diversity. If the transformer is not properly managed and a fault occurs, it will usually lead to a large-scale and long-time power outage accident.

[0005] As can be seen, effective and accurate monitoring and diagnosis of power equipment is an effective way to improve power supply reliability and intelligent level of power grid operation. The power system urgently needs more accurate and rapid online monitoring and diagnosis technology of power transmission and transformation equipment, which can predict potential risks, reduce the possibility of power outage and further improve the quality of power supply.

[0006] However, in the prior art, there is a lack of a power equipment monitoring method based on processing of multiple state data in a period of time. One reason is that it is more cumbersome to extract valuable information from a large amount of data. Another reason is that it is difficult to transmit a large amount of data, and the network and computing burden are large.

[0007] Therefore, it is necessary to develop and design a power equipment state monitoring and fault early warning method based on edge computing. SUMMARY

[0008] The embodiment of the present application provides a power equipment state monitoring and fault early warning method and system based on edge computing, and is used for solving the problem that there is no effective information mining from big data to early warn the power equipment operation state in the prior art.

[0009] In a first aspect, the embodiment of the present application provides a power equipment state monitoring and fault early warning method based on edge computing, comprising:

[0010] obtaining a plurality of operation data of the power equipment and an early warning model, wherein the operation data comprises a plurality of state data of the power equipment in a preset time period;

[0011] clustering the plurality of operation data in units of nodes to obtain operation data centers of a plurality of nodes, wherein the operation data center of the node comprises a node identifier of the node and a clustering center of the plurality of operation data of the node;

[0012] inputting the operation data centers of the plurality of nodes into the early warning model to obtain early warning information, wherein the early warning information comprises a node identifier;

[0013] sending the early warning information to the node according to the node identifier of the early warning information.

[0014] In a possible implementation manner, the clustering the plurality of operation data to obtain the operation data centers of the plurality of nodes comprises:

[0015] for each node, the following steps are performed:

[0016] obtaining the node identifier, the plurality of operation data and a preset number of clustering centers corresponding to the node;

[0017] calculating distances between the plurality of operation data and each clustering center;

[0018] taking the node identifier and the clustering center with the minimum distance from the plurality of operation data as the operation data center of the node.

[0019] In a possible implementation manner, the obtaining the preset number of clustering centers corresponding to the node comprises:

[0020] obtaining the preset number and a plurality of samples;

[0021] classifying the plurality of samples into the preset number of classes;

[0022] for each class, the following steps are performed:

[0023] calculating a center point of each sample in the class as the clustering center.

[0024] In a possible implementation, the method further includes:

[0025] obtaining a class corresponding to the cluster center closest to the plurality of running data;

[0026] adding the plurality of running data as samples to the class;

[0027] calculating a center point of each sample in the class as the cluster center of the class.

[0028] In a possible implementation, the obtaining the early warning model includes:

[0029] obtaining a plurality of prediction samples, a plurality of labels, and a neural network model, wherein the labels correspond to the prediction samples and are used to identify an expectation of the prediction samples;

[0030] training the neural network model according to the plurality of prediction samples and the plurality of labels until a prediction accuracy of the neural network model reaches a first preset value;

[0031] fixing parameters of the neural network model as the early warning model.

[0032] In a possible implementation, the training the neural network model according to the plurality of prediction samples and the plurality of labels until the prediction accuracy of the neural network model reaches a preset value includes:

[0033] dividing the plurality of prediction samples and the plurality of labels into a training set and a verification set;

[0034] a training step of training the neural network model according to the training set until a prediction accuracy of the neural network model reaches a second preset value;

[0035] a verification step of fixing the parameters of the neural network model, verifying the prediction accuracy of the neural network model according to the verification set, and if the prediction accuracy is lower than the first preset value, changing a structure of the neural network model by reducing a number of intermediate layers, and returning to the training step.

[0036] In a second aspect, an embodiment of the present application provides an edge computing-based power equipment state monitoring and fault early warning system, including a sensor terminal, a node monitoring device, and a server, wherein the server runs an early warning model.

[0037] The sensor terminal is configured to obtain state data of the power equipment.

[0038] The node monitoring device is configured to form operation data according to the state data of the power equipment transmitted by the sensor terminal connected thereto, and to cluster the operation data to obtain node state data.

[0039] The server is configured to input the operation data of the plurality of nodes to the early warning model to obtain early warning information, and to send the early warning information to the node monitoring device according to the node identifier of the early warning information, wherein the early warning information comprises a node identifier.

[0040] In a possible implementation manner, the sensor terminal comprises at least one of the following: an ultrasonic partial discharge sensor, a very high frequency partial discharge sensor, a grounding current high frequency sensor, an SF6 gas monitoring device, an acoustic soundprint sensor, a mechanical vibration sensor, an oil chromatographic monitoring device, a switch operating mechanism mechanical characteristic monitoring device, a relative dielectric loss and capacitance online monitoring device, a resistive current online monitoring device, and a sensor for collecting power equipment operation related environmental quantities.

[0041] In a third aspect, an embodiment of the present application provides a node monitoring device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any possible implementation manner of the first aspect.

[0042] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, wherein the computer program is executable by a processor to implement the steps of the method according to any possible implementation manner of the first aspect.

[0043] Compared with the prior art, the embodiment of the present application has the following beneficial effects:

[0044] The embodiment of the present application discloses a power equipment state monitoring and fault early warning method based on edge computing. After a large amount of operation data is obtained, the operation data is classified according to corresponding nodes, then a plurality of operation data in the classification is clustered to obtain a clustering center. Next, the clustering centers of the plurality of nodes are input into an early warning model to obtain early warning information. After the data is clustered, the data quantity is greatly reduced, and at the same time, as many characteristics of the operation data as possible are retained. Therefore, the data transmission quantity and the calculation quantity are reduced. The clustering is based on nodes, and the specific node can be located according to the early warning result. The embodiment of the present application reduces the data transmission quantity, reduces the loss of characteristics as much as possible, reduces the operation quantity of the model, speeds up the operation speed, and locates the early warning information to the node, thereby providing necessary data support for the operation and maintenance of the power equipment. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0046] Figure 1 is a flowchart of the power equipment state monitoring and fault early warning method based on edge computing provided by the embodiments of the present application;

[0047] Figure 2 is a flowchart of clustering multiple operation data provided by the embodiments of the present application;

[0048] Figure 3 is a schematic diagram of re-computing a cluster center provided by the embodiments of the present application;

[0049] Figure 4 is a schematic diagram of a neural network structure provided by the embodiments of the present application;

[0050] Figure 5 is a functional block diagram of the power equipment state monitoring and fault early warning system based on edge computing provided by the embodiments of the present application;

[0051] Figure 6 is a functional block diagram of the node monitoring device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0052] In the following description, specific details are set forth in order to provide a thorough understanding of the embodiments of the present application. However, persons skilled in the art will understand that the present application can be practiced without these specific details. In other instances, well-known systems, methods, and devices have not been described in detail so as not to obscure the present application.

[0053] In order to make the objects, technical solutions and advantages of the present application clearer, the following will be described in conjunction with the drawings through specific embodiments.

[0054] The embodiments of the present application will be described in detail below, and the present example is implemented on the premise of the technical solutions of the present application, and detailed embodiments and specific operation processes are given, but the protection scope of the present application is not limited to the following embodiments.

[0055] Figure 1 is a flowchart of the power equipment state monitoring and fault early warning method based on edge computing provided by the embodiments of the present application.

[0056] AsFigure 1 As shown, it shows a flow chart of the implementation of the power equipment status monitoring and fault warning method based on edge computing provided by an embodiment of the present invention, which is detailed as follows:

[0057] In step 101, a plurality of operating data of the power equipment and an early warning model are obtained, wherein the operating data includes a plurality of status data of the power equipment within a preset time period.

[0058] For example, power equipment operating data is data that characterizes the operating status trends of the equipment. It includes multiple data points acquired by multiple sensors over a specific time period. In one application scenario, the sensor data collected is not uploaded in real time. Instead, the sensor data is aggregated and uploaded after a certain time interval, for example, every 12 or 24 hours. This approach reduces the amount of data transmitted, prevents network congestion, and reduces the amount of data computation required, while still meeting the requirements for characterizing the operating status trends of power equipment.

[0059] In step 102, the plurality of operation data are clustered in units of nodes to obtain operation data centers of the plurality of nodes, where the operation data centers of the nodes include: node identifiers of the nodes and cluster centers of the plurality of operation data of the nodes.

[0060] In some embodiments, clustering the plurality of operating data to obtain operating data centers of the plurality of nodes includes:

[0061] For each node, perform the following steps:

[0062] Obtaining the node identifier, a plurality of operation data, and a preset number of cluster centers corresponding to the node;

[0063] Calculating distances between the plurality of running data and each cluster center;

[0064] The node identifier and the cluster center with the smallest distance from the multiple operation data are used as the operation data center of the node.

[0065] In some implementations, obtaining a preset number of cluster centers corresponding to the nodes includes:

[0066] Obtaining the preset number and multiple samples;

[0067] Classifying the plurality of samples into the preset number of classes;

[0068] For each of these classes, perform the following steps:

[0069] Calculate the center point of each sample in the class as the cluster center.

[0070] In some embodiments, the node identification and the cluster center with the minimum distance from the plurality of running data are taken as the running data center of the node, and then the method comprises:

[0071] Obtaining a class corresponding to the cluster center with the minimum distance from the plurality of running data;

[0072] Adding the plurality of running data as samples to the class;

[0073] Calculating the center point of each sample in the class as the cluster center of the class.

[0074] Exemplarily, as described above, the data of the sensor is aggregated and then uploaded for processing, but there are various aggregation methods, such as calculating the maximum value, the minimum value, the change trend, and the average value, and taking the calculated value as the data representing the running state and uploading to the data processing end.

[0075] As can be seen, although the data is processed by calculation to obtain the data representing the running state, in some application scenarios, the calculation amount is too large, and these data for representing the running state still have defects, for example, for switch-type devices, the above data is difficult to express.

[0076] The present application provides a method that can be implemented, which uses as few data as possible to express more details, so as to restore the running state of the power equipment.

[0077] Figure 2 A flowchart for clustering a plurality of running data is shown, first, the plurality of running data is classified in units of nodes. In one application scenario, a node is a node monitoring device connected with a plurality of sensors. Then, the plurality of running data of the node is clustered, and the cluster center of the running data is taken as the center of the running data.

[0078] Specifically, first, a preset number of cluster centers are obtained, and the preset number is usually determined according to different types of abnormalities, such as determining the number of cluster centers according to the position of the abnormality.

[0079] Then, the distance between each running data of the node and the cluster center is calculated, and the cluster center with the closest distance to each running data is taken as the center of each running data of the node.

[0080] There are various distance calculation methods, such as the most commonly used Euclidean distance, which is a simple and easy-to-use calculation method.

[0081] For obtaining the cluster center, the known historical running data and the abnormality corresponding to the historical running data are usually taken as samples.

[0082] Specifically, a plurality of historical running data and a plurality of corresponding abnormality of the historical running data are obtained.

[0083] Then, the historical running data is classified according to the abnormality.

[0084] Finally, for each class, a center of the class is calculated as a clustering center of the class.

[0085] There are various ways to calculate the center of the class, one of which is determined according to the following formula:

[0086]

[0087] In the formula, μ is the center of the class, and x is a sample in the class C.

[0088] Figure 3 A clustering center recalculation diagram is shown in FIG. 3, and those skilled in the art should know that the diagram is only an example provided for understanding, not limitation.

[0089] The Figure 3 There are three classes in FIG. 2, and each element in each class has two attributes: X coordinate and Y coordinate. The first class is a rectangular box, and the center of the class is shown by the dashed circle. The dashed rectangle shown in the figure is calculated. The dashed rectangle is closest to the clustering center of the first class, so the first class is the closest class to the dashed rectangle, and the position of the dashed circle is obtained as the clustering center of the dashed rectangle. After obtaining the clustering center, the dashed rectangle is added to the first class, and the center of the first class is recalculated, as shown by the solid circle in the first class in the figure. The recalculated clustering center can be used next time for clustering.

[0090] In step 103, the running data center of the plurality of nodes is input into the early warning model to obtain early warning information, and the early warning information includes a node identifier.

[0091] In some embodiments, the early warning model is obtained by:

[0092] A plurality of prediction samples, a plurality of labels, and a neural network model are obtained, wherein the labels correspond to the prediction samples and are used to identify the expectations of the prediction samples.

[0093] The neural network model is trained according to the plurality of prediction samples and the plurality of labels until the prediction accuracy of the neural network model reaches a first preset value.

[0094] The parameters of the neural network model are fixed as the early warning model.

[0095] In some embodiments, the training of the neural network model according to the plurality of prediction samples and the plurality of labels until the prediction accuracy of the neural network model reaches a preset value comprises:

[0096] dividing the plurality of prediction samples and the plurality of labels into a training set and a validation set;

[0097] a training step: training the neural network model according to the training set until the prediction accuracy of the neural network model reaches a second preset value;

[0098] a validation step: fixing the parameters of the neural network model, verifying the prediction accuracy of the neural network model according to the validation set, if it is lower than the first preset value, changing the structure of the neural network model by reducing the number of intermediate layers, and returning to the training step.

[0099] Exemplarily, after obtaining the running data center of the plurality of nodes, the running data center can be input into the early warning model to obtain early warning information.

[0100] One embodiment of the early warning model is a neural network model trained by samples.

[0101] Figure 4 The schematic diagram of the neural network model is shown in the middle.

[0102] The circles on the left side of the diagram are input layers, the circles on the right side are output layers, and the boxes in the middle are intermediate layers, also known as hidden layers. The number of hidden layers can be adjusted according to different needs.

[0103] Specifically, neural networks (NN) are complex network systems formed by a large number of simple processing units (called neurons) widely interconnected, which reflect many basic characteristics of brain function. Neural networks are highly complex nonlinear dynamic learning systems. Neural networks have large-scale parallelism, distributed storage and processing, self-organization, self-adaptation and self-learning ability, and are particularly suitable for processing information processing problems that need to consider many factors and conditions, and are imprecise and fuzzy. The development of neural networks is related to neuroscience, mathematical science, cognitive science, computer science, artificial intelligence, information science, control theory, robotics, microelectronics, psychology, optical computing, molecular biology, etc. It is a new emerging interdisciplinary subject.

[0104] The basis of neural networks is neurons.

[0105] Neurons are biological models based on nerve cells of the biological nervous system. When people study the biological nervous system to explore the mechanism of artificial intelligence, neurons are mathematized, thereby generating a mathematical model of neurons.

[0106] A neural network is a highly nonlinear dynamic system composed of a large number of identical neurons connected together. While the structure and function of each neuron are relatively simple, the dynamic behavior of the neural network is highly complex. Therefore, neural networks can represent a wide range of phenomena in the real world.

[0107] Neural network models are described based on the mathematical model of neurons. An artificial neural network (ANN) describes the first-order characteristics of the human brain system. Simply put, it is a mathematical model. A neural network model is represented by network topology, node characteristics, and learning rules.

[0108] The great appeal of neural networks to people lies in the following points:

[0109] 1. Parallel distributed processing.

[0110] 2. Highly robust and fault-tolerant.

[0111] 3. Distributed storage and learning capabilities.

[0112] 4. Can fully approximate complex nonlinear relationships.

[0113] The control of uncertain systems has long been a central theme in control theory research, but this problem has remained largely unresolved. Leveraging the learning capabilities of neural networks, they can automatically learn the characteristics of uncertain systems while controlling them, thereby automatically adapting to temporal variations in their characteristics in order to achieve optimal control. This is clearly an exciting prospect and approach.

[0114] Figure 4 The neural network model in this paper is a supervised learning neural network model. Supervised learning refers to a learning network that, given input data and the corresponding output, seeks to determine the relationship between the output and the input data. When the input and output data are sufficient, a relationship between them can be established. This process is called learning, also known as training.

[0115] A common use of neural network models is to classify existing data, such as objects in an image, or to predict future developments based on existing data.

[0116] The neural network model applied in the embodiment of the present application is a supervised learning neural network. Through a large amount of running data and training of corresponding abnormal running data, the neural network establishes a relationship between the running data and the abnormal corresponding running data. The relationship is fixed, that is, the parameters of the fixed neural network in the embodiment of the present application.

[0117] The neural network has a significant overfitting tendency. Overfitting is to form a relationship between data and abnormality according to the trend of data, without considering whether the relationship has noise, that is, deviation. The overfitting neural network has the disadvantage of high classification accuracy for training samples, but low accuracy in practical application. The usual way to solve this problem is to regularize the neural network.

[0118] Therefore, the embodiment of the present application divides the training samples into two categories, one is the training set and the other is the validation set. The trained neural network model is verified by the validation set. When the verification accuracy is low, the neural network is regularized. The embodiment of the present application adjusts the verification accuracy by changing the number of intermediate layers of the neural network.

[0119] In step 104, according to the node identifier of the early warning information, the early warning information is sent to the node.

[0120] Exemplarily, in some embodiments, for the output early warning information, if the early warning information contains an abnormality, the early warning information containing the abnormality is sent to the corresponding node according to the node identifier in the early warning information.

[0121] The embodiment of the power equipment state monitoring and fault early warning method based on edge computing of the present application classifies the running data according to the corresponding nodes after obtaining a large amount of running data, then clusters the multiple running data in the classification to obtain the cluster center. Next, the cluster centers of multiple nodes are input into the early warning model to obtain the early warning information. After clustering the data, the data amount is greatly reduced, while the characteristics of as much running data as possible are retained. Therefore, the data transmission amount and the calculation amount are reduced. The clustering is based on nodes, which can be located to specific nodes according to the early warning result. The embodiment of the present application reduces the data transmission amount, reduces the loss of characteristics as much as possible, reduces the operation amount of the model, speeds up the operation speed, and locates the early warning information to the node, which provides necessary data support for the operation and maintenance of power equipment.

[0122] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0123] The following is the device embodiment of the present application, for which reference can be made to the corresponding method embodiments described above for details not described in detail.

[0124] Figure 5 is the functional block diagram of the edge computing-based power equipment state monitoring and fault early warning system provided by the embodiment of the present application, referring to Figure 5 , the edge computing-based power equipment state monitoring and fault early warning system comprises:

[0125] a sensor terminal 301, a node monitoring device 302 and a server 303, wherein the server 303 runs a warning model;

[0126] The sensor terminal 301 is configured to acquire state data of power equipment.

[0127] The node monitoring device 302 is configured to form operation data according to the state data of the power equipment transmitted by the sensor terminal 301 connected thereto, and to cluster the operation data to obtain node state data.

[0128] The server 303 is configured to input the operation data of the plurality of nodes to the warning model to obtain warning information, and to send the warning information to the node monitoring device 302 according to the node identifier of the warning information, wherein the warning information comprises a node identifier.

[0129] In some embodiments, the sensor terminal 301 comprises at least one of the following: an ultrasonic partial discharge sensor, a very high frequency partial discharge sensor, a ground current high frequency sensor, an SF6 gas monitoring device, an acoustic voiceprint sensor, a mechanical vibration sensor, an oil chromatographic monitoring device, a switch operating mechanism mechanical characteristic monitoring device, a relative dielectric loss and capacitance online monitoring device, a resistive current online monitoring device, and a sensor for collecting power equipment operation related environmental quantities.

[0130] Figure 6 is the functional block diagram of the node monitoring device provided by the embodiment of the present application. As Figure 6 shown, the node monitoring device 302 of the embodiment comprises a processor 400, a memory 401, and a computer program 402 stored in the memory 401 and executable on the processor 400. The processor 400 executes the computer program 402 to implement the steps in the above various edge computing-based power equipment state monitoring and fault early warning methods and embodiments, such as Figure 1 steps 101 to 104 shown in the figure.

[0131] For example, the computer program 402 can be divided into one or more modules / units, which are stored in the memory 401 and executed by the processor 400 to complete the present application.

[0132] The node monitoring device 302 can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The node monitoring device 302 can include, but is not limited to, a processor 400 and a memory 401. Those skilled in the art can understand that the node monitoring device 302 can include more or less components, or combine some components, or different components, for example, the terminal can also include an input / output device, a network access device, a bus, etc. Figure 6 The node monitoring device 302 is only an example and does not constitute a limitation on the node monitoring device 302, which can include more or less components than the illustration, or combine some components, or different components, for example, the terminal can also include an input / output device, a network access device, a bus, etc.

[0133] The processor 400 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0134] The memory 401 can be an internal storage unit of the node monitoring device 302, such as a hard disk or a memory of the node monitoring device 302. The memory 401 can also be an external storage device of the node monitoring device 302, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the memory 401 can include both the internal storage unit and the external storage device of the node monitoring device 302. The memory 401 is used to store the computer program and other programs and data required by the terminal. The memory 401 can also be used to temporarily store data that has been output or will be output.

[0135] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be realized in the form of hardware or software function unit. In addition, the specific name of each functional unit and module is only for the convenience of mutual distinction, and does not limit the protection scope of the present application. The specific working process of the unit and module in the above system can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

[0136] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0137] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0138] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / terminal and method can be implemented by other ways. For example, the above-mentioned apparatus / terminal embodiments are only schematic, and the division of the modules or units is only a logical function division, and there can be another division way in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0139] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

[0140] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0141] The integrated module / unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiments can be implemented by a computer program instructing related hardware, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned embodiments of the power equipment state monitoring and fault early warning method based on edge computing can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms, etc. The computer-readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0142] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for power equipment status monitoring and fault warning based on edge computing, characterized in that: include: Acquiring multiple operating data of an electric power device and an early warning model, wherein the operating data includes multiple status data of the electric power device within a preset time period; clustering the multiple operating data on a node-by-node basis to obtain operating data centers of the multiple nodes, wherein the operating data centers of the nodes include: node identifiers of the nodes and cluster centers of the multiple operating data of the nodes; inputting the operating data centers of the multiple nodes into the early warning model to obtain early warning information, wherein the early warning information includes node identifiers; and sending the early warning information to the nodes according to the node identifiers of the early warning information; Clustering the plurality of operation data to obtain operation data centers of the plurality of nodes includes: For each node, the following steps are performed: obtaining the node identifier, multiple operating data, and a preset number of cluster centers corresponding to the node; wherein the preset number is the number of types of anomalies; calculating the distance between the multiple operating data and each cluster center; and using the node identifier and the cluster center with the smallest distance to the multiple operating data as the operating data center of the node.

2. The power equipment status monitoring and fault early warning method based on edge computing according to claim 1 is characterized in that: The obtaining of a preset number of cluster centers corresponding to the nodes includes: Obtaining the preset number and multiple samples; Classifying the plurality of samples into the preset number of classes; For each of these classes, perform the following steps: Calculate the center point of each sample in the class as the cluster center.

3. The power equipment status monitoring and fault early warning method based on edge computing according to claim 1 is characterized in that: The node identifier and the cluster center with the smallest distance from the plurality of operation data are used as the operation data center of the node, and then the method further includes: Obtaining a class corresponding to a cluster center having the smallest distance from the plurality of running data; adding the plurality of operating data as samples to the class; Calculate the center point of each sample in the class as the cluster center of the class.

4. The power equipment status monitoring and fault early warning method based on edge computing according to any one of claims 1 to 3, characterized in that: The obtaining of the early warning model includes: Acquire a plurality of prediction samples, a plurality of labels, and a neural network model, wherein the labels correspond to the prediction samples and are used to identify expectations of the prediction samples; Training the neural network model according to the multiple prediction samples and the multiple labels until the prediction accuracy of the neural network model reaches a first preset value; The parameters of the neural network model are fixed as an early warning model.

5. The power equipment status monitoring and fault early warning method based on edge computing according to claim 4 is characterized in that: The training of the neural network model according to the plurality of prediction samples and the plurality of labels until the prediction accuracy of the neural network model reaches a preset value includes: Dividing the plurality of prediction samples and the plurality of labels into a training set and a validation set; Training step: training the neural network model according to the training set until the prediction accuracy of the neural network model reaches a second preset value; Verification step: fix the parameters of the neural network model, verify the prediction accuracy of the neural network model based on the verification set, if it is lower than a first preset value, change the structure of the neural network model by reducing the number of intermediate layers, and return to the training step.

6. A power equipment status monitoring and fault warning system based on edge computing, characterized in that: include: A sensor terminal, a node monitoring device, and a server, wherein an early warning model runs on the server; The sensor terminal is used to obtain status data of the power equipment; The node monitoring device is used to generate operation data based on the status data of the power equipment transmitted by the sensor terminal connected thereto, and cluster the operation data to obtain operation data centers of multiple nodes; The operation data center of the node includes: a node identifier of the node and a cluster center of multiple operation data of the node; The server is configured to input the operation data centers of the plurality of nodes into the early warning model to obtain early warning information, and send the early warning information to the node monitoring device according to the node identifier of the early warning information, wherein the early warning information includes the node identifier; For each node, the node monitoring device is specifically used to obtain the node identification, multiple operating data and a preset number of cluster centers corresponding to the node; wherein the preset number is the number of types of abnormalities; calculate the distance between the multiple operating data and each cluster center; and use the node identification and the cluster center with the smallest distance to the multiple operating data as the operating data center of the node.

7. The power equipment status monitoring and fault warning system based on edge computing according to claim 6 is characterized in that: The sensor terminal includes at least one of the following: an ultrasonic partial discharge sensor, an ultra-high frequency partial discharge sensor, a ground current high-frequency sensor, an SF6 gas monitoring device, an acoustic soundprint sensor, a mechanical vibration sensor, an oil chromatography monitoring device, a switch operating mechanism mechanical characteristic monitoring device, a relative dielectric loss and capacitance online monitoring device, a resistive current online monitoring device, and a sensor for collecting environmental quantities related to the operation of power equipment.

8. A node monitoring device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

9. A computer-readable storage medium storing a computer program, 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