Automatic inspection system and method for power distribution network based on artificial intelligence

Abstract

The invention discloses a power distribution network automatic inspection system and method based on artificial intelligence, and relates to the technical field of artificial intelligence, the method comprises the steps: carrying out the inspection of a power distribution network region through an unmanned plane, determining the position coordinates of power distribution equipment, and obtaining the time sequence data of each equipment through region networking; constructing a time sequence data slice matrix corresponding to the equipment; determining a period window to analyze the change rate of continuous time sequence data in each equipment period; a regional equipment observation matrix is constructed for each piece of equipment, and comprehensive distribution network time sequence data fluctuation value analysis is carried out in combination with the periodic time sequence data change rate of each piece of equipment; performing hierarchical division on each node based on the coordinate data of each node in combination with the positioning information of the source node of the distribution network; performing feature representation region construction on each adjacent node, and performing abnormal region detection by combining hierarchical division data of each node and regional distribution network data fluctuation analysis; and performing visual data monitoring feedback on each node and corresponding hierarchical division information, and warning abnormal areas.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based distribution network automated inspection system and method. Background Art

[0002] Distribution network automated inspection is a technology that integrates sensors, drones, robots and other equipment, combined with advanced data processing and analysis algorithms, to achieve real-time monitoring, fault detection and preventive maintenance of power grid equipment to ensure the stable operation and reliability of the power grid. Currently, the level of intelligent inspection of distribution networks is highly popular. Remote monitoring of the status of power grid equipment is achieved through drones and sensor equipment, and anomalies can be discovered in real time. However, the current monitoring system of distribution networks still maintains and supervises the power grid energy supply system by judging the abnormal status of independent equipment. However, when faced with complex power grid energy supply situations, most of the potential risks cannot be fully monitored only through abnormal monitoring of equipment. It lacks correlation analysis between power grid equipment and lacks power grid risk monitoring in small-scale areas. Summary of the invention

[0003] The purpose of the present invention is to provide an artificial intelligence-based distribution network automated inspection system and method to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions: An artificial intelligence-based distribution network automated inspection method, the method comprising the following steps: S100, inspecting the distribution network management area by using a drone, determining the location coordinates of the distribution equipment in the distribution network area, and acquiring the time series data of each device through regional networking based on the coordinate data of each device; S200, construct a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyze the change rate of the continuous time series data within the cycle of each device by determining the period window; construct a regional device observation matrix by conducting data overall planning for each device in the region, and analyze the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device; S300, based on the coordinate data of each node combined with the location information of the distribution network source node, each node is hierarchically divided; by constructing a feature characterization area for each adjacent node, abnormal area detection is performed in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; S400: Visual data monitoring and feedback are performed on each node and the corresponding hierarchical division information, and warnings are issued for abnormal areas.

[0005] The S100 inspects the distribution network management area through a drone, determines the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device. The specific steps are as follows: S101. By constructing a drone inspection platform, the inspection tasks for the distribution network area are formulated, the inspection routes for each drone are planned, and the equipment in the distribution network area is monitored by the drone; S102. Based on the drone monitoring of the distribution network area equipment, each device is spatially positioned to obtain the spatial coordinate information of the corresponding device, and the drone-borne communication sensor equipment is used to collect time series data of each device; the time series data is the operating process data of the equipment at each time point; the operating process data of the equipment includes temperature, voltage, current, power, load and other data.

[0006] The S200 constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of the continuous time series data within the cycle of each device by determining the period window; constructs a regional device observation matrix by comprehensively planning the data of each device in the region, and analyzes the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device. The specific analysis steps are as follows: S201, based on the time series data collected by each device, construct a time series data set of the corresponding device according to the data collected by each type of device at the corresponding time point; construct a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; use the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and use the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; S202, by calling the corresponding device time series data slice matrix, determine the period window to obtain the time series data set of each device in the corresponding period; analyze the degree of change of the device time series data in each device period, and the calculation formula is: ; Among them, RMn is the periodic time series data change rate of device number n; n is the number of devices; S (n,z) (t+1) and S (n,z) (t+1) are the device timing data of the corresponding data type number z at adjacent time points in the cycle; z is the number of device timing data types; t and t+1 are adjacent time points in the cycle; T is the cycle window; Combined with the change rate of the periodic time series data of each device in the region, the regional equipment monitoring matrix of the corresponding equipment in the distribution network area is constructed by coordinating the numbering data of each device; based on the regional equipment monitoring matrix, the fluctuation value of the comprehensive time series data within the distribution network area cycle is analyzed, and the calculation formula is: ; Among them, RC is the comprehensive time series data fluctuation value within the distribution network area period; Fn is the type influencing parameter corresponding to the device number n.

[0007] The S300 divides each node into levels based on the coordinate data of each node combined with the location information of the distribution network source node; constructs a feature characterization area for each adjacent node, and performs abnormal area detection in combination with the level division data of each node and the regional distribution network data fluctuation analysis. The specific steps are as follows: S301. Locate the source nodes of the distribution network in the distribution network area, and construct a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; divide each distribution network node into node hierarchies with the source node as the starting reference point; take the distribution network node directly connected to the source node as the first-level node, and then coordinate the position coordinates of each first-level node to construct a first-level node layer; traverse the lower directly connected nodes with the first-level node, and the traversed nodes are second-level nodes, and then coordinate the position coordinates of each second-level node to construct a second-level node layer; with this node division method, the distribution network nodes are fully traversed hierarchically divided; wherein the traversal of the lower directly connected point refers to taking the current node level as a reference, the node to which the current node receives energy transmission as the upper node, and the current node transmitting to the next directly connected node as the lower node, and the lower node is lower than the current node level; S302, taking each distribution network node as the center node, constructing the associated node pairing coordinate group corresponding to the current center node by traversing the distribution network nodes connected to the current center node; comprehensively constructing the associated node pairing coordinate group of the current center node to form a corresponding coordinate group set, and calculating the spatial distance of the distribution network nodes associated with the current center node in each coordinate group, and the calculation formula is: ; Among them, S(i,e:i) is the spatial distance between the current central node number i and the associated node number e; (i,e) is the associated node pairing coordinate group constructed by the current central node number i and the node number e; i(p) and e(p) correspond to the coordinate components of the nodes numbered i and e respectively; p takes the values ​​of X, Y, and Z; among them, X, Y, and Z are the node spatial coordinate directions respectively; the associated node spatial distance analysis is performed on the corresponding coordinate group sets associated with each central node respectively, and the associated nodes connected to the central node in the associated node pairing coordinate groups corresponding to MIN1[S(i,e:i)] and MIN2[S(i,e:i)] are taken to construct the feature representation area corresponding to each central node; among them, MIN1[S(i,e:i)] and MIN2[S(i,e:i)] correspond to the first minimum value and the second minimum value of the spatial distance calculation results between the central node and each associated node in the central node associated node pairing coordinate group respectively; Based on the characteristic representation areas in the distribution network area and the corresponding regional construction nodes, regional anomaly analysis is performed; each characteristic representation area is numbered and marked, and the regional time series data fluctuation deviation degree analysis is performed on each characteristic representation area and the corresponding construction node. The calculation formula is: ; Where FOV[K(i,e1,e2)] is the degree of fluctuation deviation of the regional time series data of the feature representation area numbered K; K is the number of feature representation areas; i, e1, e2 in K(i,e1,e2) are the construction node numbers of the corresponding feature representation areas numbered K; RM(i∈Wj), RM(e1∈Wj), RM(e2∈Wj) correspond to the periodic comprehensive time series data fluctuation values ​​of the level where nodes i, e1, e2 are located respectively; Wj corresponds to the level number; j is the number of levels; the periodic comprehensive time series data fluctuation values ​​of each level are calculated as follows: ; Among them, RM(Wj) corresponds to the fluctuation value of the periodic comprehensive time series data at the level numbered Wj; Based on the analysis results of the regional time series data fluctuation deviation degree of each feature characterization area, a warning threshold FOV(th) is introduced; if FOV[K(i,e1,e2)]≥FOV(th), there is power distribution data anomaly in the corresponding numbered feature characterization area; if FOV[K(i,e1,e2)]≥FOV(th) does not exist, the feature characterization area corresponding to the maximum value of the regional time series data fluctuation deviation degree of each feature characterization area is taken as the first monitoring object, and real-time abnormality monitoring is performed.

[0008] The specific steps of performing visual data monitoring and feedback on each node and the corresponding hierarchical division information and issuing warnings on abnormal areas are as follows: S401, visually outputting the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; S402: If there is an abnormal area, a warning feedback is given to the abnormal area; if there is no abnormality, the first monitoring object is output and priority real-time monitoring is performed.

[0009] An artificial intelligence-based distribution network automated inspection system, the system comprising a regional data acquisition module, a time series data fluctuation analysis module, a regional anomaly detection module and a data feedback module; The regional data acquisition module inspects the distribution network management area through drones to determine the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device; the time series data fluctuation analysis module constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of continuous time series data within the cycle of each device by determining the period window; constructs a regional equipment observation matrix by coordinating data planning for each device in the area, and analyzes the comprehensive distribution network time series data fluctuation value of each device in the area in combination with the change rate of the periodic time series data of each device; the regional anomaly detection module divides each node into levels based on the coordinate data of each node combined with the positioning information of the distribution network source node; constructs a feature characterization area for each adjacent node, and detects abnormal areas in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; the data feedback module performs visual data monitoring and feedback on each node and the corresponding hierarchical division information, and warns of abnormal areas.

[0010] The regional data acquisition module includes a drone inspection control unit and a regional distribution network data acquisition unit; The drone inspection control unit constructs a drone inspection platform, formulates inspection tasks for the distribution network area, plans inspection routes for each drone, and monitors the equipment in the distribution network area through the drone; The regional distribution network data acquisition unit spatially locates each device based on the distribution network regional equipment monitored by the drone to obtain the spatial coordinate information of the corresponding device, and collects time series data of each device through the drone's onboard communication sensor equipment; the time series data is the operating process data of the equipment at each time point.

[0011] The time series data fluctuation analysis module includes a time series data slice matrix construction unit and a regional node time series data fluctuation analysis unit; The time series data slice matrix construction unit constructs a time series data set of the corresponding device based on the time series data collected by each device and the data collected by each type of device at the corresponding time point; constructs a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; uses the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and uses the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; The regional node timing data fluctuation analysis unit retrieves the corresponding device timing data slice matrix, determines the period window, and obtains the timing data set of each device within the corresponding period; analyzes the degree of change of the device timing data operation within each device period; combines the period timing data change rate of each device in the region, and builds a regional equipment monitoring matrix for the equipment in the corresponding distribution network region by coordinating the numbering data of each device; and analyzes the comprehensive timing data fluctuation value within the distribution network region period based on the regional equipment monitoring matrix.

[0012] The regional anomaly detection module includes a regional node hierarchical division unit and a regional anomaly analysis unit; The regional node hierarchical division unit locates the distribution network source node in the distribution network area, and constructs a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; taking the source node as the starting reference point, each distribution network node is divided into node hierarchies; taking the distribution network node directly connected to the source node as the first-level node, the position coordinates of each first-level node are comprehensively constructed into a first-level node layer; traversing the nodes directly connected to the first-level node, the traversed nodes are second-level nodes, and the position coordinates of each second-level node are comprehensively constructed into a second-level node layer; in this node division method, the distribution network nodes are fully traversed and hierarchically divided; The regional anomaly analysis unit takes each distribution network node as a central point node, constructs an associated node pairing coordinate group corresponding to the current central node by traversing the distribution network nodes connected to the current central node; comprehensively constructs a corresponding coordinate group set for the associated node pairing coordinate group of the current central node, and respectively calculates the spatial distance of the distribution network nodes associated with the current central node in each coordinate group; respectively performs an associated node spatial distance analysis on the corresponding coordinate group set associated with each central node, takes the associated nodes connected to the central node in the associated node pairing coordinate group corresponding to the first minimum value and the second minimum value of the spatial distance analysis, and constructs a feature representation area corresponding to each central node; based on each characteristic node in the distribution network area, the corresponding coordinate group set associated with each central node is analyzed, and the associated nodes in the associated node pairing coordinate group corresponding to the first minimum value and the second minimum value of the spatial distance analysis are taken, and the characteristic representation area corresponding to each central node is constructed; based on each characteristic node in the distribution network area, the corresponding coordinate group set associated with each central node is analyzed, and the characteristic representation area corresponding to each characteristic node in the distribution network area ... The characteristic representation area is combined with the corresponding regional construction node to perform regional anomaly analysis; each characteristic representation area is numbered and marked respectively, and the regional time series data fluctuation offset degree analysis is performed on each characteristic representation area and the corresponding construction node; based on the regional time series data fluctuation offset degree analysis results of each characteristic representation area, a warning threshold is introduced; if there is a regional time series data fluctuation offset degree greater than or equal to the warning threshold, the corresponding numbered characteristic representation area has power distribution data anomaly; if there is no regional time series data fluctuation offset degree greater than or equal to the warning threshold, the characteristic representation area corresponding to the maximum value of the regional time series data fluctuation offset degree of each characteristic representation area is taken as the first monitoring object, and real-time anomaly monitoring is performed.

[0013] The data feedback module includes a data output unit and an abnormal warning unit; The data output unit visualizes and outputs the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; The abnormal warning unit determines whether there is an abnormal area; if there is an abnormal area, a warning feedback is given to the abnormal area; if there is no abnormality, the first monitoring object is output and priority real-time monitoring is performed.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention realizes data monitoring and collection of power grid area equipment by combining long-range drones and sensors, and performs comprehensive change rate analysis on the time series data of distribution network equipment in combination with artificial intelligence algorithms, and analyzes the fluctuation of regional time series data according to the change rate of equipment time series data; by hierarchical division of regional power grid equipment and combining adjacent nodes to build feature characterization areas, by miniaturizing the distribution network area, and by performing correlation fluctuation analysis on the component node equipment of each characterization area, the abnormal situation of the characterization area is judged; the present invention can adapt to the potential risks of the distribution network in various functional scenarios to perform correlation analysis between power grid equipment, and effectively monitor power grid risks by building small-scale areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The present invention is a schematic diagram of the structure of an artificial intelligence-based distribution network automated inspection and supervision system. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example: Figure 1 As shown, the present invention provides a technical solution: An artificial intelligence-based distribution network automated inspection method, the method comprising the following steps: S100, inspecting the distribution network management area by using a drone, determining the location coordinates of the distribution equipment in the distribution network area, and acquiring the time series data of each device through regional networking based on the coordinate data of each device; S200, construct a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyze the change rate of the continuous time series data within the cycle of each device by determining the period window; construct a regional device observation matrix by conducting data overall planning for each device in the region, and analyze the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device; S300, based on the coordinate data of each node combined with the location information of the distribution network source node, each node is hierarchically divided; by constructing a feature characterization area for each adjacent node, abnormal area detection is performed in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; S400: Visual data monitoring and feedback are performed on each node and the corresponding hierarchical division information, and warnings are issued for abnormal areas.

[0018] The S100 inspects the distribution network management area through a drone, determines the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device. The specific steps are as follows: S101. By constructing a drone inspection platform, the inspection tasks for the distribution network area are formulated, the inspection routes for each drone are planned, and the equipment in the distribution network area is monitored by the drone; S102, spatially locate each device in the distribution network area according to the drone monitoring equipment to obtain the spatial coordinate information of the corresponding device, and collect time series data of the equipment through the drone's onboard communication sensor equipment; the time series data is the operating process data of the equipment at each time point.

[0019] The S200 constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of the continuous time series data within the cycle of each device by determining the period window; constructs a regional device observation matrix by comprehensively planning the data of each device in the region, and analyzes the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device. The specific analysis steps are as follows: S201, based on the time series data collected by each device, construct a time series data set of the corresponding device according to the data collected by each type of device at the corresponding time point; construct a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; use the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and use the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; S202, by calling the corresponding device time series data slice matrix, determine the period window to obtain the time series data set of each device in the corresponding period; analyze the degree of change of the device time series data in each device period, and the calculation formula is: ; Among them, RMn is the periodic time series data change rate of device number n; n is the number of devices; S (n,z) (t+1) and S (n,z) (t+1) are the device timing data of the corresponding data type number z at adjacent time points in the cycle; z is the number of device timing data types; t and t+1 are adjacent time points in the cycle; T is the cycle window; Combined with the change rate of the periodic time series data of each device in the region, the regional equipment monitoring matrix of the corresponding equipment in the distribution network area is constructed by coordinating the numbering data of each device; based on the regional equipment monitoring matrix, the fluctuation value of the comprehensive time series data within the distribution network area cycle is analyzed, and the calculation formula is: ; Among them, RC is the comprehensive time series data fluctuation value within the distribution network area period; Fn is the type influencing parameter corresponding to the device number n.

[0020] The S300 divides each node into levels based on the coordinate data of each node combined with the location information of the distribution network source node; constructs a feature characterization area for each adjacent node, and performs abnormal area detection in combination with the level division data of each node and the regional distribution network data fluctuation analysis. The specific steps are as follows: S301, locate the distribution network source node in the distribution network area, and construct a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; take the source node as the starting reference point, and divide each distribution network node into node hierarchies; take the distribution network node directly connected to the source node as the first-level node, and then coordinate the position coordinates of each first-level node to construct a first-level node layer; traverse the nodes directly connected to the first-level node, and the traversed nodes are second-level nodes, and then coordinate the position coordinates of each second-level node to construct a second-level node layer; in this node division method, the distribution network nodes are fully traversed and divided into levels; S302, taking each distribution network node as the center node, constructing the associated node pairing coordinate group corresponding to the current center node by traversing the distribution network nodes connected to the current center node; comprehensively constructing the associated node pairing coordinate group of the current center node to form a corresponding coordinate group set, and calculating the spatial distance of the distribution network nodes associated with the current center node in each coordinate group, and the calculation formula is: ; Among them, S(i,e:i) is the spatial distance between the current central node number i and the associated node number e; (i,e) is the associated node pairing coordinate group constructed by the current central node number i and the node number e; i(p) and e(p) correspond to the coordinate components of the nodes number i and e respectively; p takes the values ​​of X, Y, and Z; among them, X, Y, and Z are the spatial coordinate directions of the nodes respectively; the associated node spatial distance analysis is performed on the corresponding coordinate group sets associated with each central node respectively, and the associated nodes connected to the central node in the associated node pairing coordinate groups corresponding to MIN1[S(i,e:i)] and MIN2[S(i,e:i)] are taken, and the feature representation area corresponding to each central node is constructed; Based on the characteristic representation areas in the distribution network area and the corresponding regional construction nodes, regional anomaly analysis is performed; each characteristic representation area is numbered and marked, and the regional time series data fluctuation deviation degree analysis is performed on each characteristic representation area and the corresponding construction node. The calculation formula is: ; Where FOV[K(i,e1,e2)] is the degree of fluctuation deviation of the regional time series data of the feature representation area numbered K; K is the number of feature representation areas; i, e1, e2 in K(i,e1,e2) are the construction node numbers of the corresponding feature representation areas numbered K; RM(i∈Wj), RM(e1∈Wj), RM(e2∈Wj) correspond to the periodic comprehensive time series data fluctuation values ​​of the level where nodes i, e1, e2 are located respectively; Wj corresponds to the level number; j is the number of levels; the periodic comprehensive time series data fluctuation values ​​of each level are calculated as follows: ; Among them, RM(Wj) corresponds to the fluctuation value of the periodic comprehensive time series data at the level numbered Wj; Based on the analysis results of the regional time series data fluctuation deviation degree of each feature characterization area, a warning threshold FOV(th) is introduced; if FOV[K(i,e1,e2)]≥FOV(th), there is power distribution data anomaly in the corresponding numbered feature characterization area; if FOV[K(i,e1,e2)]≥FOV(th) does not exist, the feature characterization area corresponding to the maximum value of the regional time series data fluctuation deviation degree of each feature characterization area is taken as the first monitoring object, and real-time abnormality monitoring is performed.

[0021] The specific steps of performing visual data monitoring and feedback on each node and the corresponding hierarchical division information and issuing warnings on abnormal areas are as follows: S401, visually outputting the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; S402: If there is an abnormal area, a warning feedback is given to the abnormal area; if there is no abnormality, the first monitoring object is output and priority real-time monitoring is performed.

[0022] An artificial intelligence-based distribution network automated inspection system, the system comprising a regional data acquisition module, a time series data fluctuation analysis module, a regional anomaly detection module and a data feedback module; The regional data acquisition module inspects the distribution network management area through drones to determine the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device; the time series data fluctuation analysis module constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of continuous time series data within the cycle of each device by determining the period window; constructs a regional equipment observation matrix by coordinating data planning for each device in the area, and analyzes the comprehensive distribution network time series data fluctuation value of each device in the area in combination with the change rate of the periodic time series data of each device; the regional anomaly detection module divides each node into levels based on the coordinate data of each node combined with the positioning information of the distribution network source node; constructs a feature characterization area for each adjacent node, and detects abnormal areas in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; the data feedback module performs visual data monitoring and feedback on each node and the corresponding hierarchical division information, and warns of abnormal areas.

[0023] The regional data acquisition module includes a drone inspection control unit and a regional distribution network data acquisition unit; The drone inspection control unit constructs a drone inspection platform, formulates inspection tasks for the distribution network area, plans inspection routes for each drone, and monitors the equipment in the distribution network area through the drone; The regional distribution network data acquisition unit spatially locates each device based on the distribution network regional equipment monitored by the drone to obtain the spatial coordinate information of the corresponding device, and collects time series data of each device through the drone's onboard communication sensor equipment; the time series data is the operating process data of the equipment at each time point.

[0024] The time series data fluctuation analysis module includes a time series data slice matrix construction unit and a regional node time series data fluctuation analysis unit; The time series data slice matrix construction unit constructs a time series data set of the corresponding device based on the time series data collected by each device and the data collected by each type of device at the corresponding time point; constructs a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; uses the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and uses the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; The regional node timing data fluctuation analysis unit retrieves the corresponding device timing data slice matrix, determines the period window, and obtains the timing data set of each device within the corresponding period; analyzes the degree of change of the device timing data operation within each device period; combines the period timing data change rate of each device in the region, and builds a regional equipment monitoring matrix for the equipment in the corresponding distribution network region by coordinating the numbering data of each device; and analyzes the comprehensive timing data fluctuation value within the distribution network region period based on the regional equipment monitoring matrix.

[0025] The regional anomaly detection module includes a regional node hierarchical division unit and a regional anomaly analysis unit; The regional node hierarchical division unit locates the distribution network source node in the distribution network area, and constructs a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; taking the source node as the starting reference point, each distribution network node is divided into node hierarchies; taking the distribution network node directly connected to the source node as the first-level node, the position coordinates of each first-level node are comprehensively constructed into a first-level node layer; traversing the nodes directly connected to the first-level node, the traversed nodes are second-level nodes, and the position coordinates of each second-level node are comprehensively constructed into a second-level node layer; in this node division method, the distribution network nodes are fully traversed and hierarchically divided; The regional anomaly analysis unit takes each distribution network node as a central point node, constructs an associated node pairing coordinate group corresponding to the current central node by traversing the distribution network nodes connected to the current central node; comprehensively constructs a corresponding coordinate group set for the associated node pairing coordinate group of the current central node, and respectively calculates the spatial distance of the distribution network nodes associated with the current central node in each coordinate group; respectively performs an associated node spatial distance analysis on the corresponding coordinate group set associated with each central node, takes the associated nodes connected to the central node in the associated node pairing coordinate group corresponding to the first minimum value and the second minimum value of the spatial distance analysis, and constructs a feature representation area corresponding to each central node; based on each characteristic node in the distribution network area, the corresponding coordinate group set associated with each central node is analyzed, and the associated nodes in the associated node pairing coordinate group corresponding to the first minimum value and the second minimum value of the spatial distance analysis are taken, and the characteristic representation area corresponding to each central node is constructed; based on each characteristic node in the distribution network area, the corresponding coordinate group set associated with each central node is analyzed, and the characteristic representation area corresponding to each characteristic node in the distribution network area ... The characteristic representation area is combined with the corresponding regional construction node to perform regional anomaly analysis; each characteristic representation area is numbered and marked respectively, and the regional time series data fluctuation offset degree analysis is performed on each characteristic representation area and the corresponding construction node; based on the regional time series data fluctuation offset degree analysis results of each characteristic representation area, a warning threshold is introduced; if there is a regional time series data fluctuation offset degree greater than or equal to the warning threshold, the corresponding numbered characteristic representation area has power distribution data anomaly; if there is no regional time series data fluctuation offset degree greater than or equal to the warning threshold, the characteristic representation area corresponding to the maximum value of the regional time series data fluctuation offset degree of each characteristic representation area is taken as the first monitoring object, and real-time anomaly monitoring is performed.

[0026] The data feedback module includes a data output unit and an abnormal warning unit; The data output unit visualizes and outputs the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; The abnormal warning unit determines whether there is an abnormal area; if there is an abnormal area, it gives a warning feedback to the abnormal area; if there is no abnormality, it outputs the first monitoring object and performs priority real-time monitoring; In the example: A regional distribution network control center adopts the distribution network automation inspection system based on artificial intelligence of the present invention to supervise the regional distribution network equipment; it constructs a drone inspection platform, formulates inspection tasks for the distribution network area, plans inspection routes for each drone, and monitors the distribution network area equipment through the drone; spatially locates each device based on the drone monitoring of the distribution network area equipment to obtain the spatial coordinate information of the corresponding device, and collects time series data of each device through the drone's airborne communication sensor equipment; the time series data is the operation process data of the equipment at each time point; Based on the time series data collected by each device, the time series data set of the corresponding device is constructed according to the data collected by each type of device at the corresponding time point; the time series data slice matrix of each device is constructed in combination with the spatial coordinate information and device type data of the corresponding device; the spatial coordinates and type information of the device are used as the first-level device identity data of the matrix, and the time series data set of the device at the corresponding time point is used as the second-level device status data of the corresponding matrix; by calling the corresponding device time series data slice matrix, the period window is determined to obtain the time series data set of each device in the corresponding period; the degree of change of the device time series data operation in each device period is analyzed, and the calculation formula is: ; Combined with the change rate of the periodic time series data of each device in the region, the regional equipment monitoring matrix of the corresponding equipment in the distribution network area is constructed by coordinating the numbering data of each device; based on the regional equipment monitoring matrix, the fluctuation value of the comprehensive time series data within the distribution network area cycle is analyzed, and the calculation formula is: ; The distribution network source nodes in the distribution network area are located, and the distribution network spatial node distribution model is constructed by combining the coordinates of each node for spatial mapping; the source node is used as the starting reference point to divide each distribution network node into node hierarchies; the distribution network nodes directly connected to the source node are regarded as the first-level nodes, and the position coordinates of each first-level node are comprehensively constructed into the first-level node layer; the nodes directly connected to the first-level node are traversed, and the traversed nodes are regarded as the second-level nodes, and the position coordinates of each second-level node are comprehensively constructed into the second-level node layer; in this way of node division, the distribution network nodes are fully traversed and hierarchically divided; Taking each distribution network node as the center node, the distribution network nodes connected to the current center node are traversed to build the corresponding node pairing coordinate group of the current center node; the corresponding coordinate group set is constructed for the current center node's associated node pairing coordinate group, and the spatial distance of the distribution network nodes associated with the current center node in each coordinate group is calculated. The calculation formula is: ; Perform spatial distance analysis on the corresponding coordinate group sets associated with each central node, take the associated nodes connected to the central node in the associated node pairing coordinate groups corresponding to MIN1[S(i,e:i)] and MIN2[S(i,e:i)], and construct the feature representation area corresponding to each central node; Based on the characteristic representation areas in the distribution network area and the corresponding regional construction nodes, regional anomaly analysis is performed; each characteristic representation area is numbered and marked, and the regional time series data fluctuation deviation degree analysis is performed on each characteristic representation area and the corresponding construction node. The calculation formula is: ; The calculation formula for the fluctuation value of the periodic comprehensive time series data at each level is: ; Based on the analysis results of the regional time series data fluctuation deviation degree of each feature characterization area, a warning threshold FOV(th) is introduced; if FOV[K(i,e1,e2)]≥FOV(th), there is power distribution data anomaly in the corresponding numbered feature characterization area; if FOV[K(i,e1,e2)]≥FOV(th) does not exist, the feature characterization area corresponding to the maximum value of the regional time series data fluctuation deviation degree of each feature characterization area is taken as the first monitoring object, and real-time abnormal monitoring is performed; the time series data, hierarchical division and time series data change rate analysis data of each node equipment in the distribution network area are visualized and output; if there is an abnormal area, a warning feedback is given to the abnormal area; if there is no abnormality, the first monitoring object is output and priority real-time monitoring is performed.

[0027] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. An automatic inspection method for distribution network based on artificial intelligence, characterized in that: The method comprises the following steps: S100, inspecting the distribution network management area by using a drone, determining the location coordinates of the distribution equipment in the distribution network area, and acquiring the time series data of each device through regional networking based on the coordinate data of each device; S200, construct a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyze the change rate of the continuous time series data within the cycle of each device by determining the period window; construct a regional device observation matrix by conducting data overall planning for each device in the region, and analyze the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device; S300, based on the coordinate data of each node combined with the location information of the distribution network source node, each node is hierarchically divided; by constructing a feature characterization area for each adjacent node, abnormal area detection is performed in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; S400: Visual data monitoring and feedback are performed on each node and the corresponding hierarchical division information, and warnings are issued for abnormal areas.

2. According to the artificial intelligence-based distribution network automatic inspection method of claim 1, it is characterized by: The S100 inspects the distribution network management area through a drone, determines the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device. The specific steps are as follows: S101. By constructing a drone inspection platform, the inspection tasks for the distribution network area are formulated, the inspection routes for each drone are planned, and the equipment in the distribution network area is monitored by the drone; S102. Based on the drone monitoring of the distribution network area equipment, each device is spatially positioned to obtain the spatial coordinate information of the corresponding device, and the drone-borne communication sensor equipment is used to collect time series data of each device; the time series data is the operating process data of the equipment at each time point.

3. The method for automatic inspection of distribution network based on artificial intelligence according to claim 2 is characterized in that: The S200 constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of the continuous time series data within the cycle of each device by determining the period window; constructs a regional device observation matrix by comprehensively planning the data of each device in the region, and analyzes the fluctuation value of the comprehensive distribution network time series data of each device in the region in combination with the change rate of the periodic time series data of each device. The specific analysis steps are as follows: S201, based on the time series data collected by each device, construct a time series data set of the corresponding device according to the data collected by each type of device at the corresponding time point; construct a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; use the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and use the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; S202, by calling the corresponding device time series data slice matrix, determine the period window to obtain the time series data set of each device in the corresponding period; analyze the degree of change of the device time series data in each device period, and the calculation formula is: ； Among them, RMn is the periodic time series data change rate of device number n; n is the number of devices; S (n,z) (t+1) and S (n,z) (t+1) are the device timing data of the corresponding data type number z at adjacent time points in the cycle; z is the number of device timing data types; t and t+1 are adjacent time points in the cycle; T is the cycle window; Combined with the change rate of the periodic time series data of each device in the region, the regional equipment monitoring matrix of the corresponding equipment in the distribution network area is constructed by coordinating the numbering data of each device; based on the regional equipment monitoring matrix, the fluctuation value of the comprehensive time series data within the distribution network area cycle is analyzed, and the calculation formula is: ； Among them, RC is the comprehensive time series data fluctuation value within the distribution network area period; Fn is the type influencing parameter corresponding to the device number n.

4. The method for automatic inspection of distribution network based on artificial intelligence according to claim 3 is characterized in that: The S300 divides each node into levels based on the coordinate data of each node combined with the location information of the distribution network source node; constructs a feature characterization area for each adjacent node, and performs abnormal area detection in combination with the level division data of each node and the regional distribution network data fluctuation analysis. The specific steps are as follows: S301, locate the distribution network source nodes in the distribution network area, and construct a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; take the source node as the starting reference point, divide each distribution network node into node hierarchies; take the distribution network nodes directly connected to the source node as the first-level node, and then coordinate the position coordinates of each first-level node to construct a first-level node layer; traverse the nodes directly connected to the first-level node, and the traversed nodes are second-level nodes, and then coordinate the position coordinates of each second-level node to construct a second-level node layer; Using this node division method, the distribution network nodes are fully traversed and divided into levels; S302, taking each distribution network node as the center node, constructing the associated node pairing coordinate group corresponding to the current center node by traversing the distribution network nodes connected to the current center node; comprehensively constructing the associated node pairing coordinate group of the current center node to form a corresponding coordinate group set, and calculating the spatial distance of the distribution network nodes associated with the current center node in each coordinate group, and the calculation formula is: ； Among them, S(i,e:i) is the spatial distance between the current central node number i and the associated node number e; (i,e) is the associated node pairing coordinate group constructed by the current central node number i and the node number e; i(p) and e(p) correspond to the coordinate components of the nodes number i and e respectively; p takes the values ​​of X, Y, and Z; among them, X, Y, and Z are the spatial coordinate directions of the nodes respectively; the associated node spatial distance analysis is performed on the corresponding coordinate group sets associated with each central node respectively, and the associated nodes connected to the central node in the associated node pairing coordinate groups corresponding to MIN1[S(i,e:i)] and MIN2[S(i,e:i)] are taken, and the feature representation area corresponding to each central node is constructed; Based on the characteristic representation areas in the distribution network area and the corresponding regional construction nodes, regional anomaly analysis is performed; each characteristic representation area is numbered and marked, and the regional time series data fluctuation deviation degree analysis is performed on each characteristic representation area and the corresponding construction node. The calculation formula is: ； Where FOV[K(i,e1,e2)] is the degree of fluctuation deviation of the regional time series data of the feature representation area numbered K; K is the number of feature representation areas; i, e1, e2 in K(i,e1,e2) are the construction node numbers of the corresponding feature representation areas numbered K; RM(i∈Wj), RM(e1∈Wj), RM(e2∈Wj) correspond to the periodic comprehensive time series data fluctuation values ​​of the level where nodes i, e1, e2 are located respectively; Wj corresponds to the level number; j is the number of levels; the periodic comprehensive time series data fluctuation values ​​of each level are calculated as follows: ； Among them, RM(Wj) corresponds to the fluctuation value of the periodic comprehensive time series data at the level numbered Wj; Based on the analysis results of the fluctuation deviation degree of regional time series data of each feature characterization area, the warning threshold FOV(th) is introduced; if FOV[K(i,e1,e2)]≥FOV(th), there is power distribution data anomaly in the corresponding numbered feature characterization area; if FOV[K(i,e1,e2)]≥FOV(th) does not exist, the feature characterization area corresponding to the maximum value of the regional time series data fluctuation deviation degree of each feature characterization area is taken as the first monitoring object, and real-time abnormality monitoring is performed.

5. The method for automatic inspection of distribution network based on artificial intelligence according to claim 4 is characterized in that: The specific steps of performing visual data monitoring and feedback on each node and the corresponding hierarchical division information and issuing warnings on abnormal areas are as follows: S401, visually outputting the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; S402, if there is an abnormal area, providing warning feedback for the abnormal area; If there is no abnormality, the first monitoring object will be output and prioritized for real-time monitoring.

6. An artificial intelligence-based distribution network automated inspection system, characterized in that: The system includes a regional data acquisition module, a time series data fluctuation analysis module, a regional anomaly detection module and a data feedback module; The regional data acquisition module inspects the distribution network management area through drones to determine the location coordinates of the distribution equipment in the distribution network area, and obtains the time series data of each device through regional networking based on the coordinate data of each device; the time series data fluctuation analysis module constructs a time series data slice matrix of the corresponding device according to the time series data of each device; comprehensively analyzes the change rate of continuous time series data within the cycle of each device by determining the period window; constructs a regional equipment observation matrix by coordinating data planning for each device in the area, and analyzes the comprehensive distribution network time series data fluctuation value of each device in the area in combination with the change rate of the periodic time series data of each device; the regional anomaly detection module divides each node into levels based on the coordinate data of each node combined with the positioning information of the distribution network source node; constructs a feature characterization area for each adjacent node, and detects abnormal areas in combination with the hierarchical division data of each node and the regional distribution network data fluctuation analysis; the data feedback module performs visual data monitoring and feedback on each node and the corresponding hierarchical division information, and warns of abnormal areas.

7. The distribution network automation inspection system based on artificial intelligence according to claim 6 is characterized by: The regional data acquisition module includes a drone inspection control unit and a regional distribution network data acquisition unit; The drone inspection control unit constructs a drone inspection platform, formulates inspection tasks for the distribution network area, plans inspection routes for each drone, and monitors the equipment in the distribution network area through the drone; The regional distribution network data acquisition unit spatially locates each device based on the distribution network regional equipment monitored by the drone to obtain the spatial coordinate information of the corresponding device, and collects time series data of each device through the drone's onboard communication sensor equipment; the time series data is the operating process data of the equipment at each time point.

8. The distribution network automatic inspection system based on artificial intelligence according to claim 7 is characterized by: The time series data fluctuation analysis module includes a time series data slice matrix construction unit and a regional node time series data fluctuation analysis unit; The time series data slice matrix construction unit constructs a time series data set of the corresponding device based on the time series data collected by each device and the data collected by each type of device at the corresponding time point; constructs a time series data slice matrix corresponding to each device in combination with the spatial coordinate information and device type data of the corresponding device; uses the spatial coordinates and type information of the device as the first-level device identity data of the matrix, and uses the time series data set of the device at the corresponding time point as the second-level device status data of the corresponding matrix; The regional node timing data fluctuation analysis unit retrieves the corresponding device timing data slice matrix, determines the period window, and obtains the timing data set of each device within the corresponding period; analyzes the degree of change of the device timing data operation within each device period; combines the period timing data change rate of each device in the region, and builds a regional equipment monitoring matrix for the equipment in the corresponding distribution network region by coordinating the numbering data of each device; and analyzes the comprehensive timing data fluctuation value within the distribution network region period based on the regional equipment monitoring matrix.

9. The distribution network automation inspection system based on artificial intelligence according to claim 8 is characterized by: The regional anomaly detection module includes a regional node hierarchical division unit and a regional anomaly analysis unit; The regional node hierarchical division unit locates the distribution network source node in the distribution network area, and constructs a distribution network spatial node distribution model by combining the coordinates of each node for spatial mapping; taking the source node as the starting reference point, each distribution network node is divided into node hierarchies; taking the distribution network node directly connected to the source node as the first-level node, the position coordinates of each first-level node are comprehensively constructed into a first-level node layer; traversing the nodes directly connected to the first-level node, the traversed nodes are second-level nodes, and the position coordinates of each second-level node are comprehensively constructed into a second-level node layer; Using this node division method, the distribution network nodes are fully traversed and divided into levels; The regional anomaly analysis unit takes each distribution network node as the central point node, and constructs the associated node pairing coordinate group corresponding to the current central node by traversing the distribution network nodes connected to the current central node; comprehensively constructs a corresponding coordinate group set for the associated node pairing coordinate group of the current central node, and respectively calculates the spatial distance of the distribution network nodes associated with the current central node in each coordinate group; respectively performs an associated node spatial distance analysis on the corresponding coordinate group set associated with each central node, and takes the associated nodes connected to the central node in the associated node pairing coordinate group corresponding to the first minimum value and the second minimum value of the spatial distance analysis, and constructs the characteristic representation area corresponding to each central node; performs regional anomaly analysis based on each characteristic representation area in the distribution network area in combination with the corresponding regional construction node; respectively number and mark each characteristic representation area, and perform regional time series data fluctuation offset degree analysis on each characteristic representation area and the corresponding construction node; based on the regional time series data fluctuation offset degree analysis results of each characteristic representation area, introduces a warning threshold; If the fluctuation deviation of regional time series data is greater than or equal to the warning threshold, the corresponding number feature indicates that there is abnormal power distribution data in the region; If there is no regional time series data fluctuation deviation greater than or equal to the warning threshold, the characteristic representation area corresponding to the maximum value of the regional time series data fluctuation deviation of each characteristic representation area is taken as the first monitoring object for real-time abnormal monitoring.

10. The distribution network automatic inspection system based on artificial intelligence according to claim 9 is characterized in that: The data feedback module includes a data output unit and an abnormal warning unit; The data output unit performs visual output of the time series data, hierarchical division and time series data change rate analysis data of each node device in the distribution network area; The abnormal warning unit determines whether there is an abnormal area; if there is an abnormal area, it gives a warning feedback to the abnormal area; If there is no abnormality, the first monitoring object will be output and prioritized for real-time monitoring.

Images