A distribution network inspection intelligent analysis platform and method

Through real-time lossless compression of edge-end models and task grading decision-making, the problems of insufficient computing resources of edge-side equipment and insufficient coordination between cloud servers and edge-side resources in the intelligent inspection system of distribution network are solved, and efficient cloud-side collaborative inspection is achieved , improves the efficiency of inspection data analysis and management.

CN117220413BActive Publication Date: 2024-11-15STATE GRID INTELLIGENCE TECHNOLOGY CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202311226607.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-09-21
Publication Date
2024-11-15
Estimated Expiration
2043-09-21

AI Technical Summary

Technical Problem

In the existing intelligent inspection system of the distribution network, the computing resource performance of edge-side equipment is low and cannot support high-precision models. The lightweight model method lacks unified conversion tools under different training architectures, resulting in a decrease in model accuracy; at the same time, cloud servers The resource with edge side cannot be evaluated in real time, resulting in insufficient synergy.

Method used

The real-time lossless compression method of edge-end models is adopted, and the corresponding transformation of model weight parameters, network structure parameters and quantization parameters are designed to realize real-time quantification in the model inference process, and a task-level decision-making and resource adaptive scheduling method are proposed to realize cloud-edge collaboration three-dimensional inspection.

Benefits of technology

Solve the problem of the conflict between the performance of high-precision models and the edge-side device computing resources, improve the real-time evaluation and coordination of cloud servers and edge-side resources, realize the three-dimensional inspection of cloud edge collaborative, and improve the configuration. Efficiency of power grid patrol data analysis management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117220413B_ABST
    Figure CN117220413B_ABST
Patent Text Reader

Abstract

The present invention discloses an intelligent analysis platform and method for distribution network inspection, comprising: an end-side analysis subsystem acquires multi-channel inspection data and performs inspection analysis, and sends unprocessed inspection tasks, inspection analysis results and multi-channel inspection data to an edge-side analysis subsystem; when performing inspection analysis, a lightweight processing is performed by adopting a model for corresponding transformation of weight parameters and network structure parameters and quantization parameters; an edge-side analysis subsystem receives inspection tasks and performs inspection analysis, and sends the inspection analysis results to an end-side analysis subsystem, and sends unprocessed inspection tasks, the inspection analysis results of the end-side analysis subsystem and the inspection analysis results of the edge-side analysis subsystem to a cloud-side decision-making subsystem; the cloud-side decision-making subsystem summarizes the inspection analysis results and receives inspection tasks to make a final inspection task decision, which solves the current problem that cloud servers and edge-side resources cannot be evaluated in real time and lack coordination, and realizes three-dimensional inspection with cloud-edge coordination.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent inspection of distribution networks, and in particular to an intelligent analysis platform and method for inspection of distribution networks. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The current intelligent inspection system for distribution networks applies artificial intelligence recognition models, remote sensing data measurement and other technologies to relevant inspection equipment, which effectively improves the efficiency of distribution network inspections. However, the following problems still exist:

[0004] (1) The computing resources of edge devices have low performance, while high-precision models require high-performance computing resources, which is a contradiction. In addition, for existing model lightweighting methods, models generated under different training architectures lack a unified conversion tool during conversion, and cannot be uniformly adapted on related edge devices. After the model conversion, the model structure is trimmed, resulting in a decrease in model accuracy.

[0005] (2) Due to the lack of intelligence on the edge side and factors such as bandwidth, the current service-side usage is mainly in the form of single-node decision-making. Cloud servers and edge-side resources cannot be evaluated in real time, and the interaction time between the service side and the edge side is long and the coordination is insufficient. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes an intelligent analysis platform and method for distribution network inspection, proposes a real-time lossless compression method for edge models, designs corresponding conversion methods for model weight parameters, network structure parameters and quantization parameters, realizes real-time quantization in the model reasoning process, and solves the contradiction between the need for high-performance computing resources for high-precision models and the low performance of computing resources of edge devices; proposes task hierarchical decision-making and resource adaptive scheduling methods, solves the current problem that cloud servers and edge resources cannot be evaluated in real time and lack of coordination, and realizes three-dimensional inspection with cloud-edge collaboration.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a distribution network inspection intelligent analysis platform, comprising:

[0009] The end-side analysis subsystem is used to obtain multi-channel inspection data according to the inspection task, and use the artificial intelligence model corresponding to the inspection task to perform inspection analysis, and send the unfinished inspection tasks, inspection analysis results and multi-channel inspection data to the edge-side analysis subsystem; among them, the artificial intelligence model uses the weight parameters and network structure to convert the corresponding quantization parameters to perform lightweight processing when performing inspection analysis;

[0010] The edge analysis subsystem is used to receive the inspection tasks sent by the end-side analysis subsystem, perform inspection analysis, and send the inspection analysis results to the end-side analysis subsystem so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and sends the unfinished inspection tasks, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge analysis subsystem to the cloud-side decision-making subsystem;

[0011] The cloud-side decision-making subsystem is used to summarize the inspection analysis results and receive the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision.

[0012] As an optional implementation method, multiple inspection data are decoded and merged into a single data stream. After the inspection analysis of the single data stream is completed, the inspection analysis result of the single data stream is encoded and split into multiple data streams for output.

[0013] As an optional implementation method, the lightweight processing process includes: splitting the weight parameters into minimization parameters, splitting the network structure into the most simplified structural units; establishing a quantization dictionary, minimizing the parameters and structural units to achieve weight parameter quantization and network structure simplification through the quantization dictionary.

[0014] As an optional implementation, the weight parameter is quantized by quantizing the floating-point numbers in the minimization parameter into an integer parameter or a binary parameter through a quantization dictionary.

[0015] As an optional implementation, the network structure is simplified by reorganizing the network and pruning the structure through a quantized dictionary to further simplify the branch structure.

[0016] As an optional implementation, the quantization dictionary includes: a dictionary of floating-point number retention digits, that is, setting the number of decimal places retained for floating-point weight values; a floating-point number integer conversion dictionary, that is, converting floating-point weight values ​​into integer values; a floating-point weight parameter binarization parameter dictionary, that is, converting floating-point weight values ​​into 0 and 1; a network structure simplification dictionary, that is, simplifying network branches with weights corresponding to 0.

[0017] As an optional implementation method, the inspection tasks that are not processed and completed by the end-side analysis subsystem include: inspection tasks that are not supported by the end-side analysis subsystem; inspection tasks that require secondary analysis by the end-side analysis subsystem; inspection tasks that require overall scheduling by the end-side analysis subsystem; and non-emergency inspection tasks that are not supported by the end-side analysis subsystem.

[0018] As an optional implementation method, the inspection tasks that are not completed by the edge analysis subsystem include: defect types that cannot be determined by both the end-side analysis subsystem and the edge analysis subsystem, task scheduling that cannot be decided, and inspection tasks temporarily sent to the end-side analysis subsystem and the edge analysis subsystem.

[0019] As an optional implementation method, when the cloud-side decision-making subsystem and the edge-side analysis subsystem perform inspection and analysis, a multi-node distributed deployment is performed. The same nodes are deployed simultaneously in different areas, and information is shared between different nodes. All nodes are centrally managed and scheduled by the master node. When the upload bandwidth, computing power or computing resources of a node reach the critical value, the master node will re-schedule part of the inspection tasks and allocate them to other nodes for calculation. At the same time, the master node selects the corresponding node nearby for data upload management based on the location of the end-side device.

[0020] In a second aspect, the present invention provides a distribution network inspection intelligent analysis method, comprising:

[0021] The end-side analysis subsystem obtains multi-channel inspection data and uses the artificial intelligence model corresponding to the inspection task to perform inspection analysis. The unfinished inspection tasks, inspection analysis results and multi-channel inspection data are sent to the edge-side analysis subsystem. The artificial intelligence model uses the weight parameters and network structure to perform lightweight processing in the inspection analysis.

[0022] The edge-side analysis subsystem receives the inspection tasks sent by the end-side analysis subsystem, performs inspection analysis, and sends the inspection analysis results to the end-side analysis subsystem, so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and sends the unfinished inspection tasks, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge-side analysis subsystem to the cloud-side decision-making subsystem;

[0023] The cloud-side decision-making subsystem summarizes the inspection analysis results and receives the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The present invention innovatively proposes a task hierarchical decision-making and resource adaptive scheduling method, which is different from the current service-side usage that mainly relies on single-node decision-making. It solves the current problems of the inability to conduct real-time evaluation and insufficient coordination between cloud servers and edge-side resources, and realizes three-dimensional inspection with cloud-edge collaboration.

[0026] The present invention innovatively proposes a real-time lossless compression method for edge models, designs corresponding conversion methods for model weight parameters, network structure parameters and quantization parameters, realizes real-time quantization in the model reasoning process, and solves the contradiction between the high-precision model's requirement for high-performance computing resources and the low performance of edge device computing resources.

[0027] The present invention innovatively designs a "cloud-edge-end" distribution network intelligent inspection and analysis platform that integrates inspection data acquisition, equipment remote control, operation status tracking, data analysis, task hierarchical decision-making and other functions. Through the "cloud-edge-end" collaborative work division, it can not only achieve rapid response to tasks, but also reduce data transmission bandwidth and cloud-side service pressure, which can effectively improve the efficiency of distribution network inspection data analysis and management.

[0028] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0030] Figure 1 A schematic diagram of a distribution network inspection intelligent analysis platform provided in Example 1 of the present invention;

[0031] Figure 2 A schematic diagram of encoding and decoding of multi-channel inspection data provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0033] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0034] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "include" and "have" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0036] Example 1

[0037] This embodiment provides a distribution network inspection intelligent analysis platform, such as Figure 1 As shown, including:

[0038] The end-side analysis subsystem is used to obtain multi-channel inspection data according to the inspection task, and use the artificial intelligence model corresponding to the inspection task to perform inspection analysis, and send the unfinished inspection tasks, inspection analysis results and multi-channel inspection data to the edge-side analysis subsystem; among them, the artificial intelligence model uses the weight parameters and network structure to convert the corresponding quantization parameters to perform lightweight processing when performing inspection analysis;

[0039] The edge analysis subsystem is used to receive the inspection tasks sent by the end-side analysis subsystem, perform inspection analysis, and send the inspection analysis results to the end-side analysis subsystem so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and sends the unfinished inspection tasks, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge analysis subsystem to the cloud-side decision-making subsystem;

[0040] The cloud-side decision-making subsystem is used to summarize the inspection analysis results and receive the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision.

[0041] In this embodiment, the end-side analysis subsystem, the edge-side analysis subsystem, and the cloud-side decision-making subsystem are mainly used for hierarchical judgment and decision-making of inspection tasks, which is different from the process of data collection management and model training performed by the cloud system in the existing intelligent perception inspection system. Compared with substation inspection, distribution network inspection is mostly mobile and the inspection environment changes greatly, so the real-time and reliability of inspection task decision-making are required to be higher. Therefore, this embodiment designs task hierarchical decision-making, which can achieve rapid response to tasks and reduce data transmission bandwidth and cloud-side service pressure through the "cloud-edge-end" collaborative work division, effectively improving the efficiency of distribution network inspection data analysis and management.

[0042] In this embodiment, the end-side analysis subsystem includes actuators such as drones, robots, handheld devices and vehicle-mounted terminals, which are connected to relevant actuators through external communication devices and control mechanisms to achieve feedback control of various actuators, thereby realizing real-time tracking, autonomous collection or autonomous operation of distribution network targets.

[0043] For example, when combined with an operating robot arm, the robot can operate on a stationary target. First, the target is detected and located to determine the target position. At the same time, the operating robot arm provides real-time feedback of the current position. After the operation target is determined, the distance difference between the operating robot arm position and the target position is calculated in real time. The moving path and moving distance of the operating robot arm are given to control the operating robot arm to operate on the target.

[0044] For example, when combined with mobile mechanisms such as gimbal camera acquisition equipment, real-time tracking and image capture can be performed on targets with constantly changing positions. First, the video stream is acquired and analyzed in real time, and the target position is detected at the same time. The gimbal acquisition equipment is controlled not to move abnormally, the target is always kept within the field of view, and the target can be captured autonomously.

[0045] In this embodiment, the end-side analysis subsystem is used to perform intelligent analysis on urgent inspection tasks with clear status, and send the inspection analysis results and the collected multi-channel inspection data to the edge-side analysis subsystem; at the same time, for the simultaneous input analysis of multiple data streams, and in order to maximize the use of edge-side device performance while ensuring the real-time analysis of multiple-channel inspection data, this embodiment designs a multi-data high-speed encoding and decoding method;

[0046] Specifically:

[0047] Acquire multiple inspection data such as pictures and videos, decode and merge the multiple inspection data into a single data stream, complete the inspection analysis of the single data stream, and then encode and split the inspection analysis results of the single data stream into multiple data streams of the original input, so as to realize the real-time detection and output of the data stream at the edge of the distribution network, such as Figure 2 shown.

[0048] In response to the problems existing in the lightweighting of existing models, this embodiment proposes a real-time lossless compression method for edge models, and designs a corresponding conversion method for weight parameters, network structure parameters and quantization parameters. During the use of the model, the method of quantization first and then deployment is not adopted. Instead, a one-to-one conversion method is directly adopted to dynamically quantize the model during operation, preserve the model structure and parameters to the maximum extent, resolve the contradiction between the computing resources of edge devices and the model accuracy, achieve high-precision operation of the model on edge devices, and retain the accuracy of the model output to the greatest extent.

[0049] Specifically:

[0050] The weight parameters are split into minimization parameters, and a minimization parameter dictionary is established to split the network structure into the simplest structural units, that is, all network structures can be composed of the simplest structural units;

[0051] Establish a quantization dictionary, that is, the minimized parameters after the weight parameters are split and the simplified structural units are used to achieve parameter quantization and structural simplification;

[0052] Among them, parameter quantization is: if the minimization parameter contains floating-point numbers, the floating-point numbers are directly quantized into integer parameters or binary parameters through the quantization dictionary to reduce the amount of model calculation;

[0053] The structure is simplified as follows: the network structure can achieve network reorganization and structural pruning through the quantized dictionary, further simplify the branch structure in the model, reduce the model size, and thus reduce the resource usage in model reasoning.

[0054] The quantization dictionary includes: a dictionary of floating-point number retention digits, that is, setting the number of decimal places retained by floating-point weight values; a dictionary of floating-point number integer conversion, that is, converting floating-point weight values ​​into integer values; a dictionary of floating-point weight parameter binarization parameters, that is, directly converting floating-point weight values ​​into 0 and 1; a network structure simplification dictionary, that is, simplifying network branches with weights corresponding to 0, simplifying networks with branches of the same structure, and simplifying networks with branches of the same parameters.

[0055] After the conversion of the quantized dictionary, the relevant weight parameters of the original model can be retained to the greatest extent possible, and the structural parameters that need to be simplified can be quantized and simplified, which ensures the accuracy of the model and reduces the resource usage of the model operation.

[0056] In this embodiment, the inspection tasks that are not processed by the end-side analysis subsystem include: inspection tasks that the end-side analysis subsystem does not support analysis, such as models with high accuracy requirements that cannot be supported by the computing power of the end-side device; inspection tasks that the end-side analysis subsystem needs to perform secondary analysis on, such as the end-side needs to perform secondary confirmation on a certain defect type; inspection tasks that the end-side analysis subsystem needs to perform overall scheduling on, such as tasks that the end-side needs to perform comprehensive assessment and analysis on multiple defects; non-urgent inspection tasks that the end-side analysis subsystem does not support, etc.

[0057] The edge-side analysis subsystem provides further inspection analysis for the inspection tasks that have not been processed by the end-side analysis subsystem. The edge-side analysis subsystem performs inspection analysis on the above tasks based on the existing model, and then sends the analysis results to the end side.

[0058] In this embodiment, the cloud-side decision-making subsystem is mainly responsible for summarizing and managing the uploaded data, work status, and adjustment of work content on the end side and the edge side, and making final decisions on high-level decision-making tasks that the end side and the edge side cannot make decisions on; the inspection tasks handled by the cloud-side decision-making subsystem include: defect types that cannot be judged by both the end side and the edge side; task scheduling that cannot be decided on by the end side and the edge side, such as comprehensive analysis of all data information in a certain area; analysis, collection, and scheduling tasks that need to be temporarily sent to the end side and the edge side, etc.

[0059] In this embodiment, through the collaborative work of "cloud-edge-end", the inspection tasks can be decomposed according to the urgency and analysis difficulty. For urgent tasks with simple analysis, the analysis can be completed directly on the end side;

[0060] Among them, emergency inspection tasks are inspection tasks that require real-time analysis and decision-making, such as hidden defects in the distribution network that will affect the normal operation of the line;

[0061] Difficulties include: model recognition accuracy, such as defects with an accuracy rate lower than 80%; the end-side device cannot meet the computing power required for the analysis and decision-making process, such as conducting an overall analysis of equipment defects in a certain area; and tasks that exceed the decision-making authority of the end-side device.

[0062] Non-urgent tasks that require high computing power to analyze can be uploaded by the end-side device to the edge-side intelligent analysis service station for processing and analysis. Important decision-making tasks, batch data, and task management are aggregated from the end-side to the cloud side, which performs unified decision-making and management.

[0063] In this embodiment, an adaptive data scheduling and calculation method for multi-level distributed computing is designed. Through real-time evaluation of upload bandwidth, computing amount and existing computing resources, the current calculation is autonomously optimized and scheduled, and the nearest upload and computing nodes are selected to achieve rapid response scheduling of the platform.

[0064] Specific implementation method: Multi-node distributed deployment of resources of the cloud-side intelligent decision-making platform and the edge-side intelligent analysis service station, that is, the same nodes are deployed simultaneously in different areas, and data, resources and other information are shared between different nodes. At the same time, all nodes are uniformly managed and scheduled by the master node. The master node can obtain the operating status and resource occupancy information of all nodes. When the upload bandwidth, computing amount or computing resources of a node reaches the critical value, the master node can re-schedule part of the tasks and assign them to other nodes for calculation. At the same time, the master node can select the corresponding node nearby for data upload management according to the location of the end-side device, further improving the speed and flexibility of data upload and task operation.

[0065] Example 2

[0066] This embodiment provides a distribution network inspection intelligent analysis method, including:

[0067] The end-side analysis subsystem obtains multi-channel inspection data and uses the artificial intelligence model corresponding to the inspection task to perform inspection analysis. The unfinished inspection tasks, inspection analysis results and multi-channel inspection data are sent to the edge-side analysis subsystem. The artificial intelligence model uses the weight parameters and network structure to perform lightweight processing in the inspection analysis.

[0068] The edge-side analysis subsystem receives the inspection tasks sent by the end-side analysis subsystem, performs inspection analysis, and sends the inspection analysis results to the end-side analysis subsystem, so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and sends the unfinished inspection tasks, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge-side analysis subsystem to the cloud-side decision-making subsystem;

[0069] The cloud-side decision-making subsystem summarizes the inspection analysis results and receives the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision.

[0070] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A distribution network inspection intelligent analysis platform, characterized in that: include: The end-side analysis subsystem is used to obtain multi-channel inspection data according to the inspection task, and use the artificial intelligence model corresponding to the inspection task to perform inspection analysis, and send the unfinished inspection tasks, inspection analysis results and multi-channel inspection data to the edge-side analysis subsystem; among them, the artificial intelligence model uses the weight parameters and network structure to convert the corresponding quantization parameters to perform lightweight processing when performing inspection analysis; The edge analysis subsystem is used to receive the unfinished inspection tasks sent by the end-side analysis subsystem, perform inspection analysis, and send the inspection analysis results to the end-side analysis subsystem so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and send the unfinished inspection tasks of the edge analysis subsystem, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge analysis subsystem to the cloud-side decision-making subsystem; The cloud-side decision-making subsystem is used to summarize the inspection analysis results and receive the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision; Inspection tasks that are not completed by the end-side analysis subsystem include: inspection tasks that are not supported by the end-side analysis subsystem; inspection tasks that require secondary analysis by the end-side analysis subsystem; inspection tasks that require overall scheduling by the end-side analysis subsystem; and non-urgent inspection tasks that are not supported by the end-side analysis subsystem. The inspection tasks that are not completed by the edge analysis subsystem include: defect types that cannot be determined by both the device-side analysis subsystem and the edge analysis subsystem, task scheduling that cannot be decided, and inspection tasks temporarily sent to the device-side analysis subsystem and the edge analysis subsystem.

2. A distribution network inspection intelligent analysis platform as claimed in claim 1, characterized in that: The multi-channel inspection data is decoded and merged into a single data stream. After the inspection analysis of the single data stream is completed, the inspection analysis result of the single data stream is encoded and split into multiple data streams for output.

3. A distribution network inspection intelligent analysis platform as claimed in claim 1, characterized in that: The lightweight processing process includes: splitting the weight parameters into minimization parameters, splitting the network structure into the simplest structural units; establishing a quantization dictionary, minimizing the parameters and structural units to achieve weight parameter quantization and network structure simplification through the quantization dictionary.

4. A distribution network inspection intelligent analysis platform as claimed in claim 3, characterized in that: The weight parameter is quantized by quantizing the floating-point numbers in the minimized parameters into integer parameters or binary parameters through a quantization dictionary.

5. A distribution network inspection intelligent analysis platform as claimed in claim 3, characterized in that: The network structure is simplified by reorganizing the network and pruning the structure through a quantized dictionary, and the branch structure is further simplified.

6. A distribution network inspection intelligent analysis platform as claimed in claim 3, characterized in that: The quantization dictionary includes: a dictionary of floating-point number retention digits, that is, setting the number of decimal places retained by the floating-point weight value; a floating-point number integer conversion dictionary, that is, converting the floating-point weight value into an integer value; a floating-point weight parameter binary parameter dictionary, that is, converting the floating-point weight value into 0 and 1; a network structure simplification dictionary, that is, simplifying the network branches whose weights correspond to 0.

7. A distribution network inspection intelligent analysis platform as claimed in claim 1, characterized in that: When the cloud-side decision-making subsystem and the edge-side analysis subsystem perform inspection and analysis, multi-node distributed deployment is carried out. The same nodes are deployed simultaneously in different areas, and information is shared between different nodes. All nodes are centrally managed and scheduled by the master node. When the upload bandwidth, computing power or computing resources of a node reach the critical value, the master node will re-schedule part of the inspection tasks and allocate them to other nodes for calculation. At the same time, the master node selects the corresponding node nearby for data upload management based on the location of the end-side device.

8. A distribution network inspection intelligent analysis method, characterized in that: include: The end-side analysis subsystem obtains multi-channel inspection data and uses the artificial intelligence model corresponding to the inspection task to perform inspection analysis. The unfinished inspection tasks, inspection analysis results and multi-channel inspection data are sent to the edge-side analysis subsystem. The artificial intelligence model uses the weight parameters and network structure to perform lightweight processing in the inspection analysis. The edge-side analysis subsystem receives the unfinished inspection tasks sent by the end-side analysis subsystem, performs inspection analysis, and sends the inspection analysis results to the end-side analysis subsystem, so that the end-side analysis subsystem controls the execution mechanism to perform corresponding actions, and sends the unfinished inspection tasks of the edge-side analysis subsystem, the inspection analysis results of the end-side analysis subsystem, and the inspection analysis results of the edge-side analysis subsystem to the cloud-side decision-making subsystem; The cloud-side decision-making subsystem summarizes the inspection analysis results and receives the inspection tasks sent by the edge-side analysis subsystem to make the final inspection task decision; Inspection tasks that are not completed by the end-side analysis subsystem include: inspection tasks that are not supported by the end-side analysis subsystem; inspection tasks that require secondary analysis by the end-side analysis subsystem; inspection tasks that require overall scheduling by the end-side analysis subsystem; and non-urgent inspection tasks that are not supported by the end-side analysis subsystem. The inspection tasks that are not completed by the edge analysis subsystem include: defect types that cannot be determined by both the device-side analysis subsystem and the edge analysis subsystem, task scheduling that cannot be decided, and inspection tasks temporarily sent to the device-side analysis subsystem and the edge analysis subsystem.

Citation Information

Patent Citations

  • Power utilization inspection system and method

    CN108206587A

  • Intelligent sensing inspection system and method for transformer substation

    CN114665608A

  • Transformer substation inspection method based on end-side cooperation

    CN114693141A

  • Remote intelligent substation patrol system based on cloud-side-end three-level framework

    CN114784958A

  • Network device inspection method, apparatus, and device, and storage medium

    WO2022142931A1