CN-121810130-B - Construction method of regional comprehensive energy system multi-level intelligent agent

CN121810130BCN 121810130 BCN121810130 BCN 121810130BCN-121810130-B

Abstract

The invention belongs to the technical field of energy intelligence, and provides a method for constructing a multi-level intelligent body of an area comprehensive energy system, which comprises the steps of carrying out space-time alignment processing on real-time operation data to obtain synchronous data streams; the method comprises the steps of mapping parameters in a synchronous data stream into specific visual attribute primitives respectively to obtain multi-level state semantic images, carrying out visual semantic segmentation on the multi-level state semantic images to obtain visual semantic units, constructing spatial correlation patterns among the visual semantic units, carrying out visual difference comparison on the spatial correlation patterns and historical contemporaneous spatial correlation patterns to obtain abnormal visual semantic units, carrying out multi-frame tracking verification on the abnormal visual semantic units, integrating abnormal unit identifications and abnormal types into an abnormal event report, and carrying out hierarchical decision on the abnormal event report and the spatial correlation patterns to obtain a collaborative regulation and control instruction. The invention can improve the efficiency of collaborative management of the regional comprehensive energy system.

Inventors

DOU ZHENLAN
ZHANG CHUNYAN
CHEN YUHUI
ZHANG YING
GUO LEI
HAN DONG

Assignees

国网上海市电力公司
国网上海综合能源服务有限公司

Dates

Publication Date: 20260508
Application Date: 20260310

Claims (10)

1. The method for constructing the multi-level intelligent agent of the regional comprehensive energy system is characterized by comprising the following steps of: S1, performing space-time alignment processing on real-time operation data of an area comprehensive energy system to obtain a synchronous data stream of the area comprehensive energy system; s2, according to a preset visual coding rule, mapping parameters in the synchronous data stream into specific visual attribute primitives of the regional comprehensive energy system respectively, and synthesizing the specific visual attribute primitives according to a system level of the regional comprehensive energy system to obtain a multi-level state semantic image of the regional comprehensive energy system; S3, performing visual semantic segmentation on the multi-level state semantic image to obtain visual semantic units of the regional comprehensive energy system, and constructing a spatial association map between the visual semantic units; s4, performing visual difference comparison on the spatial correlation map and the historical contemporaneous spatial correlation map of the regional comprehensive energy system to obtain an abnormal visual semantic unit of the regional comprehensive energy system; S5, carrying out multi-frame tracking verification on the abnormal visual semantic unit to obtain an abnormal unit identifier and an abnormal type of the regional comprehensive energy system, and integrating the abnormal unit identifier and the abnormal type into an abnormal event report of the regional comprehensive energy system; And S6, carrying out hierarchical decision on the abnormal event report and the spatial correlation map to obtain a collaborative regulation and control instruction of the regional comprehensive energy system.
2. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S1, a synchronous data stream of the area integrated energy system is obtained, comprising: collecting original operation data of an electric subsystem, a thermodynamic subsystem and a gas subsystem in the regional comprehensive energy system to obtain real-time operation data of the regional comprehensive energy system; Extracting an acquisition time stamp and a geographic position identifier of real-time operation data; consistency verification is carried out on the acquisition time stamps, and the data with time deviation are adjusted to a unified time reference axis to obtain time sequence operation data of the regional comprehensive energy system; and matching the time sequence operation data into a system space grid of the regional comprehensive energy system based on the geographic position identification to obtain a synchronous data stream of the regional comprehensive energy system.
3. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S2, mapping parameters in the synchronous data stream to specific visual attribute primitives of the area integrated energy system respectively comprises: Carrying out parameter identification on the synchronous data stream to obtain a parameter type identifier and a parameter value of the synchronous data stream; Selecting a corresponding primitive basic shape from a preset visual coding rule according to the parameter type identifier; Determining the thickness degree of the contour line of the basic shape of the primitive and the density degree of the internal filling texture according to the parameter values; and drawing images of the synchronous data stream based on the thickness degree of the contour lines, the density degree of the internal filling textures and the basic shape of the primitives, and obtaining the specific visual attribute primitives of the regional comprehensive energy system.
4. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S2, a multi-level state semantic image of the area integrated energy system is obtained, comprising: distributing the specific visual attribute primitives to equipment-level canvas, network-level canvas and energy-flow-level canvas of the regional comprehensive energy system according to node-level labels corresponding to the specific visual attribute primitives; Performing de-overlapping treatment on specific visual attribute primitives in the equipment hierarchy canvas to obtain an equipment state sub-layer of the regional comprehensive energy system; taking the equipment state sub-layer as a background, carrying out connection drawing on specific visual attribute primitives in the network hierarchy canvas to obtain a network topology sub-layer of the regional comprehensive energy system; Drawing specific visual attribute primitives in the energy flow layer canvas on the network topology sub-layer in a semitransparent superposition mode, and adjusting transparency according to filling texture density of the specific visual attribute primitives in the energy flow layer canvas to obtain an energy flow distribution sub-layer of the regional comprehensive energy system; And overlapping the device state sub-layer, the network topology sub-layer and the energy flow distribution sub-layer according to the sequence from bottom to top to obtain a multi-level state semantic image of the regional comprehensive energy system.
5. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S3, visual semantic units of the area integrated energy system are obtained, and a spatial correlation map between the visual semantic units is constructed, comprising: carrying out connected domain analysis on the equipment state sub-layers in the multi-level state semantic image to obtain candidate equipment areas of the equipment state sub-layers, and carrying out feature extraction on filling textures in the candidate equipment areas to obtain equipment visual semantic units of the area comprehensive energy system; Segment detection is carried out on a network topology sub-layer in the multi-level state semantic image, and a connection visual semantic unit of the regional comprehensive energy system is obtained; Taking the transparency value and the texture density of the pixels as growth criteria, and carrying out region growth segmentation on the energy flow distribution sub-layers in the multi-level state semantic image to obtain an energy flow visual semantic unit of the region comprehensive energy system; And constructing a spatial association map between the visual semantic units by taking the device visual semantic units, the connection visual semantic units and the energy flow visual semantic units as nodes and taking the connection relationship between devices and the membership relationship between layers in the regional comprehensive energy system as edges.
6. The method for constructing a multi-level agent for an area integrated energy system according to claim 5, wherein obtaining the visual semantic unit of the equipment of the area integrated energy system comprises: performing binarization processing on the equipment state sub-layer, marking pixel points with contour lines as foreground pixels and marking other pixel points as background pixels to obtain an equipment contour binary image of the regional comprehensive energy system; performing eight-neighborhood connected domain marking on the equipment outline binary image, merging the mutually connected foreground pixel points into the same connected domain, and extracting the minimum circumscribed rectangle of the connected domain as a candidate equipment region of the equipment outline binary image; carrying out gray value distribution analysis on pixels in the candidate equipment area to obtain texture contrast parameters, texture correlation parameters and texture inverse difference moment parameters of the candidate equipment area; comparing the texture contrast parameter, the texture correlation parameter and the texture inverse difference moment parameter with the equipment texture feature reference set of the area comprehensive energy system item by item to obtain equipment texture matching degree of the area comprehensive energy system; based on the equipment texture matching degree, cluster analysis is carried out on the candidate equipment areas, and unique equipment identification is allocated for the candidate equipment areas, so that the equipment visual semantic units of the area comprehensive energy system are obtained.
7. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S4, an abnormal visual semantic unit of the area integrated energy system is obtained, comprising: carrying out hierarchical analysis on the historical contemporaneous spatial association map of the regional comprehensive energy system based on the spatial association map to obtain a historical visual semantic unit of the regional comprehensive energy system; Performing dimension-by-dimension difference quantization on the attribute feature vectors in the visual semantic units and the historical attribute feature vectors in the historical visual semantic units to obtain attribute difference degrees of the regional comprehensive energy system; Traversing first-order neighbor nodes of the visual semantic unit, and recording unit identifiers and connecting edge directions of the first-order neighbor nodes to obtain neighbor adjacent sequences of the visual semantic unit; extracting a corresponding historical neighbor sequence from the historical visual semantic unit according to the neighbor sequence; comparing and analyzing the neighbor adjacent sequences with the historical neighbor adjacent sequences to obtain the topological structure difference of the regional comprehensive energy system; nonlinear combination is carried out on the attribute difference degree and the topological structure difference degree, and a global abnormality index of the visual semantic unit is obtained; and comparing the global abnormality index with a preset abnormality judgment threshold to obtain an abnormal visual semantic unit of the regional comprehensive energy system.
8. The method for constructing a multi-level agent for a regional integrated energy system of claim 7, wherein the global abnormality index is calculated by the formula: ; Wherein E represents a global abnormality index, alpha represents a preset first weight coefficient, beta represents a preset second weight coefficient, A represents an attribute difference degree, T represents a topological structure difference degree, E represents a natural constant, Representing a natural logarithmic function.
9. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S5, obtaining an abnormal unit identifier and an abnormal type of the area integrated energy system, and integrating the abnormal unit identifier and the abnormal type into an abnormal event report of the area integrated energy system, comprises: establishing a tracking file for the abnormal visual semantic unit, wherein the tracking file contains current frame image coordinates and current visual attribute parameters; acquiring a subsequent frame multi-level state semantic image of the region comprehensive energy system, and positioning a subsequent visual semantic unit which has the same image coordinate range and similar visual attribute parameters as the abnormal visual semantic unit in the subsequent frame multi-level state semantic image; taking the subsequent visual semantic unit as a tracking continuation unit of the abnormal visual semantic unit in a subsequent frame, and adding the image coordinates and visual attribute parameters of the tracking continuation unit into a tracking file to obtain a complete tracking file of the regional comprehensive energy system; determining the abnormal type of the abnormal visual semantic unit according to the change trend of the visual attribute parameters in the complete tracking file, and endowing the abnormal visual semantic unit with a unique abnormal unit identifier; and according to the hierarchical attribution relation and the frame appearance sequence of the abnormal visual semantic units in the spatial association map, arranging the abnormal types and the abnormal unit identifications to obtain an abnormal event report of the regional comprehensive energy system.
10. The method for constructing a multi-level agent for an area integrated energy system according to claim 1, wherein in S6, a coordinated regulation command of the area integrated energy system is obtained, comprising: Analyzing the abnormal event report to obtain an abnormal unit identifier, an abnormal type and an abnormal occurrence time of the abnormal visual semantic unit, and positioning a layer level and a space coordinate position of the abnormal visual semantic unit in the space correlation map according to the abnormal unit identifier; dividing the abnormal visual semantic units into an equipment-level abnormal unit, a network-level abnormal unit and a stream-level abnormal unit according to the layer level; clustering and grouping the equipment hierarchy abnormal units, the network hierarchy abnormal units and the energy hierarchy abnormal units based on the abnormal types to obtain a hierarchy abnormal grouping set of the regional comprehensive energy system; Based on the hierarchical abnormal group set, performing abnormal propagation path exploration on the space association map to obtain an association equipment set of the hierarchical abnormal group set; determining preliminary regulation parameter suggestions of the regional comprehensive energy system according to the associated equipment set and the abnormal type; Combining and de-duplication processing is carried out on the preliminary regulation parameter suggestion, and a hierarchical regulation instruction set of the regional comprehensive energy system is obtained; And step-by-step packaging is carried out on the level regulation instruction set, and linkage association identifiers among the regulation instructions in the level regulation instruction set are established according to membership relations among cross-layer units in the spatial association map in the packaging process, so that the cooperative regulation instructions of the regional comprehensive energy system are obtained.

Description

Construction method of regional comprehensive energy system multi-level intelligent agent Technical Field The invention belongs to the technical field of energy intelligence, and particularly relates to a method for constructing a multi-level intelligent body of an area comprehensive energy system. Background The regional comprehensive energy system integrates multiple subsystems such as electric power, heating power, fuel gas and the like, is an important carrier for intelligent energy management and control, but in the operation data acquisition stage of each subsystem, the problem that the acquisition dimension, the time reference and the space identification are not uniform exists, the prior art lacks a high-efficiency space-time alignment means for multisource real-time operation data, and is difficult to integrate scattered data into uniform synchronous data streams, so that the problem of data fracture exists in system operation state perception, and the overall operation situation of the system cannot be accurately reflected. Meanwhile, the prior art adopts a traditional numerical method to represent the running state of the system, so that the space association and energy flow interaction relation among all levels and units are difficult to intuitively embody, and obvious technical barriers are brought to the running state analysis and anomaly identification of the system. In the process of abnormality detection, regulation and control and treatment, the prior art also has obvious short plates, the abnormality detection is carried out by judging a threshold value aiming at a single dimension parameter, comprehensive analysis is carried out without combining the association relation between a system topological structure and units, the condition of abnormality missed judgment or misjudgment is extremely easy to occur, and a multi-frame tracking verification mechanism for an abnormal unit is lacking, so that the abnormality type and the continuous characteristic cannot be accurately defined. In the stage of exception handling, the formulation of the regulation and control instruction lacks layering and collaborative design, and accurate grading regulation and control are difficult to realize according to abnormal layering attribution and propagation paths, so that the response efficiency and suitability of system regulation and control are insufficient, and the management and control requirements of multi-level collaborative operation of the regional comprehensive energy system cannot be met. Disclosure of Invention According to the defects of the prior art, the invention aims to provide a multi-level intelligent agent construction method for an area comprehensive energy system, which can improve the efficiency of collaborative management of the area comprehensive energy system. In order to achieve the above purpose, the invention adopts the following technical scheme: a method for constructing a multi-level intelligent agent of a regional comprehensive energy system comprises the following steps: S1, performing space-time alignment processing on real-time operation data of an area comprehensive energy system to obtain a synchronous data stream of the area comprehensive energy system; s2, according to a preset visual coding rule, mapping parameters in the synchronous data stream into specific visual attribute primitives of the regional comprehensive energy system respectively, and synthesizing the specific visual attribute primitives according to a system level of the regional comprehensive energy system to obtain a multi-level state semantic image of the regional comprehensive energy system; S3, performing visual semantic segmentation on the multi-level state semantic image to obtain visual semantic units of the regional comprehensive energy system, and constructing a spatial association map between the visual semantic units; s4, performing visual difference comparison on the spatial correlation map and the historical contemporaneous spatial correlation map of the regional comprehensive energy system to obtain an abnormal visual semantic unit of the regional comprehensive energy system; S5, carrying out multi-frame tracking verification on the abnormal visual semantic unit to obtain an abnormal unit identifier and an abnormal type of the regional comprehensive energy system, and integrating the abnormal unit identifier and the abnormal type into an abnormal event report of the regional comprehensive energy system; And S6, carrying out hierarchical decision on the abnormal event report and the spatial correlation map to obtain a collaborative regulation and control instruction of the regional comprehensive energy system. Preferably, in S1, obtaining a synchronous data stream of the regional integrated energy system includes: collecting original operation data of an electric subsystem, a thermodynamic subsystem and a gas subsystem in the regional comprehensive energy system to obtain real-time operati