Power grid risk intelligent optimization system and method using large language model

Abstract

The present invention discloses a power grid risk intelligent optimization system and method using a large language model, and relates to the technical field of power grid data analysis. The present invention uses a drone to inspect the transmission lines of a regional distribution network, locates each transmission device and transmits the collected data to a monitoring terminal; uses the collected data to map a spatial node model, extracts the operation data of each node, and constructs a corresponding node data cycle form; analyzes the periodic operation status of the node equipment according to the data cycle form, and classifies each node equipment in combination with historical data; comprehensively analyzes the status analysis results of each node equipment, traverses the regional transmission lines in full, and constructs a set of proposed transmission lines; evaluates the transmission risk rate of each proposed transmission line, and outputs the planned transmission line according to the evaluation result; feedbacks the status of each node equipment in combination with the node equipment status analysis result, and outputs the regional transmission line in combination with the periodic transmission line planning situation.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid data analysis, and in particular to a power grid risk intelligent optimization system and method using a large language model. Background Art

[0002] The intelligent optimization of power grid using large language model is a smart grid risk management system. It collects data on the acceptance route, builds mapping relationships to form distribution network route information, and implements supervision and control of distribution network lines and equipment.

[0003] With the rapid development of the power system, users' demand for power resources is growing rapidly, and the distribution network's demand for regulating power resources is also increasing. Due to the increasing regional development and power demand, the monitoring demand for distribution networks has become more refined and intelligent. However, the current distribution network monitoring and management still only uses historical data and experience to forecast and allocate regional power demand, and performs fault detection on transmission equipment. This model has a large lag and is not conducive to the dynamic regulation of regional real-time demand. It is impossible to make effective and timely adjustments to potential transmission risks, which will cause excessive damage or abnormality in distribution network transmission. Summary of the invention

[0004] The purpose of the present invention is to provide a power grid risk intelligent optimization system and method using a large language model to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A power grid risk intelligent optimization method using a large language model comprises the following steps:

[0007] S100, inspect the transmission lines of the regional distribution network through the drone inspection system, spatially locate each transmission line equipment and transmit the collected image data to the monitoring end;

[0008] S200, using the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracting synchronous operation data of each node through the edge cloud port, and constructing a corresponding node data cycle form;

[0009] S300, analyzing the periodic operation status of node equipment according to the data period table of each node, and classifying the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzing the status results of each node equipment, traversing the regional power transmission lines in simulation, and constructing a set of proposed power transmission lines; evaluating the transmission risk rate of each proposed power transmission line, and outputting the planned power transmission line according to the evaluation results;

[0010] S400, providing status feedback for each node device in combination with the node device status analysis result, and outputting the regional power transmission line in combination with the periodic power transmission line planning situation.

[0011] The specific steps of the S100 inspecting the transmission lines of the regional distribution network through the drone inspection system, spatially locating each transmission line equipment and transmitting the collected image data to the monitoring end are as follows:

[0012] S101. Establishing a drone inspection network to locate the spatial position of the transmission lines and transfer transmission equipment in the regional distribution network, and collect image data of each transmission equipment; record the connection paths of adjacent transfer transmission equipment;

[0013] S102, using the drone inspection network to collect the image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment, and transmit them to the monitoring end using the regional communication network.

[0014] The S200 uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts synchronous operation data of each node through the edge cloud port, and constructs the corresponding node data cycle form in the following specific steps:

[0015] S201. Based on the image data of the transmission lines and transfer transmission equipment of the regional distribution network collected by the drone network, the transmission lines and transfer transmission equipment in each image data are marked with features and data dimension reduction is performed through a feature extraction model; feature points of each transfer transmission equipment are marked, and the connection paths between the feature points are explicitly marked; a three-dimensional virtual model is used to map and build a distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment;

[0016] S202. Build an access channel to the edge cloud database port of each node device through the monitoring end to extract the historical operation data of each node device; use the equipment operation time axis as the data series axis, annotate the node device operation data at each time point to construct a data cycle form corresponding to each node device.

[0017] The S300 analyzes the periodic operation status of the node equipment according to the periodic table of each node data, and classifies the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzes the status analysis results of each node equipment, simulates and fully traverses the regional power transmission lines, and constructs a set of proposed power transmission lines; and evaluates the power transmission risk rate of each proposed power transmission line, and outputs the planned power transmission line according to the evaluation results. The specific steps are as follows:

[0018] S301, call up the data cycle form of each node device, set the observation cycle, take the current real-time time point as the cycle acquisition starting point, take the monitoring cycle as the step length, and perform reverse observation cycle data acquisition on the node data cycle form time period; take the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extract the observation cycle data of each node, and construct the corresponding observation operation data matrix, extract the data before the observation cycle on the time axis in the node data cycle form as historical data, and construct the historical operation data matrix; wherein the operation data of the node device includes equipment operation environment data, equipment operation electrical data and equipment operation mechanical data; wherein the current time point is used as the observation cycle data acquisition starting point on the node data cycle form to perform reverse step length acquisition of data, because since the current time point is the latest time point data in the data cycle form, to obtain the observation data of the observation cycle time length, it is necessary to obtain data in reverse time axis;

[0019] Based on the observed operation data matrix of each node device, the real-time operation wear state index of the node device at each time point in the cycle is analyzed, and the calculation formula is:

[0020] ;

[0021] Among them, Wh (i,t) is the real-time operation wear state index of the corresponding node device numbered i at time point t within the observation period; α, β, γ are the equipment operation environment data wear coefficient, electrical data wear coefficient, and mechanical data wear coefficient respectively; n, q, z are the number of equipment operation environment data types, the number of operation electrical data types, and the number of operation mechanical data types respectively; M n , D q , J z They correspond to the operating environment data of each type of equipment, the electrical data of each type of equipment, and the mechanical data of each type of equipment; n (s), D q (s), J z(s) respectively correspond to the prescribed standard data for the operating environment data of each type of equipment, the prescribed standard data for the operating electrical data of each type of equipment, and the prescribed standard data for the operating mechanical data of each type of equipment; the real-time operating wear state index of the equipment at each time point in the observation period is combined to construct a periodic equipment wear state index curve; by using each time point in the observation period and the corresponding equipment real-time operating wear state index as coordinates, coordinate points are mapped in the plane coordinate system, and the mapped coordinates are connected by a smooth curve to construct a periodic equipment wear state index curve; by locating the midpoint of the curve between adjacent coordinate points, the coordinates corresponding to the midpoint between each adjacent coordinate point are obtained; the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment are traversed to construct a corresponding characteristic midpoint coordinate set; based on the characteristic midpoint coordinate set, the balanced real-time operating wear state index of the equipment in the observation period is calculated, and the calculation formula is:

[0022] ;

[0023] Among them, PWh i is the balanced real-time operation wear state index of the corresponding node device numbered i during the observation period; u is the number of elements in the characteristic midpoint coordinate set; Wh (i,t,t+1) The midpoints of the adjacent coordinate points of the periodic equipment wear state index curve correspond to the real-time operation wear state index; here, by constructing the periodic equipment wear state index curve, and collecting the midpoints between the adjacent coordinate points on the curve and comprehensively analyzing the balanced real-time operation wear state index of the equipment, the real-time wear state of the equipment can be analyzed more centrally and the calculation can be more accurate;

[0024] S302, combined with the observation period balanced real-time operation consumption state index analysis data of each node device, and compared with the historical operation data matrix, the real-time state of each node device is classified into levels; through the above analysis steps, the equipment balanced operation consumption state index of each historical period in the historical operation data matrix of each node device is extracted, and the historical comprehensive balanced operation consumption state index of each node device is obtained by performing mean analysis on each historical period; when PWh i <PWh i (L), the current node device is marked as low-loss operation state; when PWh i (L)≤PWh i ≤PWh i (Bm), the current node device is marked as medium-loss operation state; when PWh i >PWh i (Bm), the current node device is marked as a high-loss abnormal operation state; the node device in the high-loss abnormal operation state is marked as an abnormal node; Among them, PWh i(L) is the historical comprehensive balanced operation loss status index of the corresponding node equipment; PWh i (Bm) is the maximum health balance operation wear state index of the corresponding node;

[0025] Based on the normal operation of the node equipment, the real-time operation risk level of each node equipment is analyzed in combination with the real-time status level of each node equipment. The calculation formula is:

[0026] ;

[0027] Among them, Wf i is the real-time operation risk level of the device with the corresponding node number i; k e The value is k1 or k2, corresponding to the operation risk coefficient of low-loss level or medium-loss level node equipment respectively;

[0028] S303, in the regional distribution network node mapping model network, determine the location of the real-time distribution network demand transmission starting point and end point; screen the nodes in normal operation in the node network, and traverse the transmission transfer nodes according to the demand starting and end point location positioning results, and simulate the transmission route generation according to the traversed nodes to construct a set of simulated transmission routes;

[0029] Based on the proposed transmission route set, each proposed transmission route is extracted and combined with the operation risk degree of each operating equipment in the corresponding route to conduct risk assessment on each proposed transmission route. The calculation formula is:

[0030] ;

[0031] Among them, WL v is the transmission risk assessment rate of the proposed transmission line corresponding to number v; v is the transmission time of the simulated transmission line corresponding to number v; t min is the minimum transmission time of the historical transmission line corresponding to the target start and end points; ave(Wf i ) is the average risk level of each node in the region; i vmax and i vmin They correspond to the maximum and minimum number of nodes contained in the proposed transmission line; S v The transmission distance of the proposed transmission line corresponding to number v; S min The minimum transmission distance of the historical transmission line corresponding to the target starting and ending points; wl is the transmission loss rate per unit transmission distance; according to the transmission risk assessment rate analysis results of each proposed transmission line, sort in descending order to obtain MIN(WL v ) corresponds to the proposed transmission route which is the best transmission route in the current cycle.

[0032] The specific steps of S400 in which the status of each node device is fed back in combination with the node device status analysis result and the regional power transmission line is output in combination with the periodic power transmission line planning are as follows:

[0033] S401, obtaining the periodic operation status level of each node device in the area, and issuing warning prompts to abnormal devices; and providing real-time risk level feedback to normal operating devices;

[0034] S402: Display the proposed transmission route set and output the optimal transmission route periodically.

[0035] A power grid risk intelligent optimization system using a large language model, the system comprising a distribution network data acquisition module, a data model conversion and processing module, a distribution network comprehensive risk assessment module and an output feedback module;

[0036] The distribution network data acquisition module inspects the regional distribution network transmission lines through the drone inspection system, spatially locates each transmission line equipment and transmits the collected image data to the monitoring end; the data model conversion processing module uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts the synchronous operation data of each node through the edge cloud port, and constructs a corresponding node data cycle form; the distribution network comprehensive risk assessment module analyzes the periodic operation status of the node equipment according to the data cycle form of each node, and divides the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzes the status analysis results of each node equipment, simulates and fully traverses the regional transmission lines, and constructs a set of simulated transmission lines; evaluates the transmission risk rate of each simulated transmission line, and outputs the planned transmission line according to the evaluation results; the output feedback module provides status feedback for each node equipment in combination with the node equipment status analysis results, and outputs the regional transmission line in combination with the periodic transmission line planning situation.

[0037] The distribution network data acquisition module includes a distribution network data acquisition unit and a data communication unit;

[0038] The distribution network data acquisition unit locates the transmission lines and transfer transmission equipment of the regional distribution network by establishing a drone inspection network, and collects image data of each transmission equipment; records the connection paths of adjacent transfer transmission equipment;

[0039] The data communication unit transmits the collected image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment to the monitoring end through the regional communication network through the drone inspection network.

[0040] The data model conversion processing module includes a data model conversion unit and a data cycle form construction unit;

[0041] The data model conversion unit performs feature marking and data dimension reduction processing on the transmission lines and each transfer transmission equipment in each image data based on the image data of the regional distribution network transmission lines and transfer transmission equipment collected by the drone network through the feature extraction model; marks the feature points of each transfer transmission equipment, and explicitly marks the connection paths between the feature points; uses the three-dimensional virtual model to map and build a distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment, and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment;

[0042] The data cycle form construction unit builds an access channel to the edge cloud database port of each node device through the monitoring end, and extracts the historical operation data of each node device; using the equipment operation timeline as the data series axis, the node device operation data at each time point is annotated to construct a data cycle form corresponding to each node device.

[0043] The distribution network comprehensive risk assessment module includes a node equipment loss analysis unit, a node equipment risk assessment unit and a power transmission line comprehensive analysis unit;

[0044] The node device wear analysis unit calls the data cycle form of each node device, sets the observation cycle, takes the current real-time time point as the cycle acquisition starting point, takes the monitoring cycle as the step length, and performs reverse observation cycle data acquisition on the node data cycle form time period; takes the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extracts the observation cycle data of each node, and constructs the corresponding observation operation data matrix, extracts the data on the time axis before the observation cycle from the node data cycle form as historical data, and constructs the historical operation data matrix; based on the observation operation data matrix of each node device, analyzes the real-time operation wear status index of the node device at each time point in the cycle; Combine the real-time operation wear state index of the equipment at each time point in the observation period to construct a periodic equipment wear state index curve; use each time point in the observation period and the corresponding equipment real-time operation wear state index as coordinates, map coordinate points in a plane coordinate system, and connect each mapped coordinate through a smooth curve to construct a periodic equipment wear state index curve; locate the midpoint of the curve between adjacent coordinate points to obtain the coordinates corresponding to the midpoint between each adjacent coordinate point; traverse the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment to construct a corresponding characteristic midpoint coordinate set; calculate the balanced real-time operation wear state index of the equipment in the observation period based on the characteristic midpoint coordinate set;

[0045] The node equipment risk assessment unit combines the observation period balanced real-time operation wear state index analysis data of each node equipment, and classifies the real-time state of each node equipment by comparing it with the historical operation data matrix; through the above analysis steps, the equipment balanced operation wear state index of each historical period in the historical operation data matrix of each node equipment is extracted, and the historical comprehensive balanced operation wear state index of each node equipment is obtained by performing mean analysis on each historical period; combined with the historical comprehensive balanced operation wear state index and the maximum healthy balanced operation wear state index of each node equipment, the state level of each equipment is divided, and abnormal equipment is marked as abnormal node; based on the node equipment in normal operation, the real-time operation risk degree of each node equipment is analyzed in combination with the real-time state level of each node equipment;

[0046] The power transmission route comprehensive analysis unit determines the positions of the starting and ending points of the real-time distribution network demand transmission in the regional distribution network node mapping model network; screens the nodes in normal operation in the node network, and traverses the transmission transfer nodes according to the positioning results of the starting and ending points of the demand, and generates simulated power transmission routes according to the traversed nodes to construct a set of simulated power transmission routes; based on the set of simulated power transmission routes, extracts each simulated power transmission route and conducts risk assessment on each simulated power transmission route in combination with the operating risk degree of each operating equipment in the corresponding route; and sorts each simulated power transmission route in descending order according to the transmission risk assessment rate analysis results of each simulated power transmission route, and obtains the simulated power transmission route corresponding to the minimum value as the best power transmission route in the current cycle.

[0047] The output feedback module includes a device state feedback unit and a power transmission line output unit;

[0048] The device status feedback unit obtains the periodic operation status level of each node device in the area, issues warning prompts to abnormal devices, and provides real-time risk level feedback to normal operating devices;

[0049] The transmission line output unit displays the proposed transmission line set and outputs the optimal transmission line periodically.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention utilizes multi-module units to realize data collection, conversion processing, risk monitoring and route optimization analysis of the entire power transmission process of the regional distribution network; constructs a virtual distribution network regional node model by model mapping the distribution network, and realizes equipment status monitoring and level division of the distribution network power transmission process based on the node model; implements risk assessment of each node device based on the level of each device and the corresponding real-time operation data; constructs a real-time power transmission route in combination with the risk assessment traversal of the node device, performs risk optimization analysis, and outputs the real-time best route; the present invention combines the real-time status of each device in the regional distribution network to perform effective dynamic division of the power transmission route, which is conducive to improving the phenomenon of excessive power transmission or abnormality caused by delayed detection of abnormal equipment or failure to combine route planning with real-time equipment status in traditional planned routes. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a structural diagram of the power grid risk intelligent optimization system using a large language model according to the present invention. DETAILED DESCRIPTION

[0053] 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.

[0054] Example: Figure 1 As shown, the present invention provides a technical solution:

[0055] A power grid risk intelligent optimization method using a large language model comprises the following steps:

[0056] S100, inspect the transmission lines of the regional distribution network through the drone inspection system, spatially locate each transmission line equipment and transmit the collected image data to the monitoring end;

[0057] S200, using the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracting synchronous operation data of each node through the edge cloud port, and constructing a corresponding node data cycle form;

[0058] S300, analyzing the periodic operation status of node equipment according to the data period table of each node, and classifying the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzing the status results of each node equipment, traversing the regional power transmission lines in simulation, and constructing a set of proposed power transmission lines; evaluating the transmission risk rate of each proposed power transmission line, and outputting the planned power transmission line according to the evaluation results;

[0059] S400, providing status feedback for each node device in combination with the node device status analysis result, and outputting the regional power transmission line in combination with the periodic power transmission line planning situation.

[0060] The specific steps of the S100 inspecting the transmission lines of the regional distribution network through the drone inspection system, spatially locating each transmission line equipment and transmitting the collected image data to the monitoring end are as follows:

[0061] S101. Establishing a drone inspection network to locate the spatial position of the transmission lines and transfer transmission equipment in the regional distribution network, and collect image data of each transmission equipment; record the connection paths of adjacent transfer transmission equipment;

[0062] S102, using the drone inspection network to collect the image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment, and transmit them to the monitoring end using the regional communication network.

[0063] The S200 uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts synchronous operation data of each node through the edge cloud port, and constructs the corresponding node data cycle form in the following specific steps:

[0064] S201. Based on the image data of the transmission lines and transfer transmission equipment of the regional distribution network collected by the drone network, the transmission lines and transfer transmission equipment in each image data are marked with features and data dimension reduction is performed through a feature extraction model; feature points of each transfer transmission equipment are marked, and the connection paths between the feature points are explicitly marked; a three-dimensional virtual model is used to map and build a distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment;

[0065] S202. Build an access channel to the edge cloud database port of each node device through the monitoring end to extract the historical operation data of each node device; use the equipment operation time axis as the data series axis, annotate the node device operation data at each time point to construct a data cycle form corresponding to each node device.

[0066] The S300 analyzes the periodic operation status of the node equipment according to the periodic table of each node data, and classifies the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzes the status analysis results of each node equipment, simulates and fully traverses the regional power transmission lines, and constructs a set of proposed power transmission lines; and evaluates the power transmission risk rate of each proposed power transmission line, and outputs the planned power transmission line according to the evaluation results. The specific steps are as follows:

[0067] S301, call the data cycle form of each node device, set the observation cycle, take the current real-time time point as the cycle acquisition starting point, take the monitoring cycle as the step length, and perform reverse observation cycle data acquisition on the node data cycle form time period; take the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extract the observation cycle data of each node, and construct the corresponding observation operation data matrix, extract the data on the time axis before the observation cycle in the node data cycle form as historical data, and construct the historical operation data matrix;

[0068] Based on the observed operation data matrix of each node device, the real-time operation wear state index of the node device at each time point in the cycle is analyzed, and the calculation formula is:

[0069] ;

[0070] Among them, Wh (i,t) is the real-time operation wear state index of the corresponding node device numbered i at time point t within the observation period; α, β, γ are the equipment operation environment data wear coefficient, electrical data wear coefficient, and mechanical data wear coefficient respectively; n, q, z are the number of equipment operation environment data types, the number of operation electrical data types, and the number of operation mechanical data types respectively; M n , D q , J z They correspond to the operating environment data of each type of equipment, the electrical data of each type of equipment, and the mechanical data of each type of equipment; n (s), D q (s), J z (s) respectively correspond to the prescribed standard data for the operating environment data of each type of equipment, the prescribed standard data for the operating electrical data of each type of equipment, and the prescribed standard data for the operating mechanical data of each type of equipment; the real-time operating wear state index of the equipment at each time point in the observation period is combined to construct a periodic equipment wear state index curve; by using each time point in the observation period and the corresponding equipment real-time operating wear state index as coordinates, coordinate points are mapped in the plane coordinate system, and the mapped coordinates are connected by a smooth curve to construct a periodic equipment wear state index curve; by locating the midpoint of the curve between adjacent coordinate points, the coordinates corresponding to the midpoint between each adjacent coordinate point are obtained; the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment are traversed to construct a corresponding characteristic midpoint coordinate set; based on the characteristic midpoint coordinate set, the balanced real-time operating wear state index of the equipment in the observation period is calculated, and the calculation formula is:

[0071] ;

[0072] Among them, PWh iis the balanced real-time operation wear state index of the corresponding node device numbered i during the observation period; u is the number of elements in the characteristic midpoint coordinate set; Wh (i,t,t+1) The midpoints of the adjacent coordinate points of the periodic equipment wear state index curve correspond to the real-time operation wear state index;

[0073] S302, combined with the observation period balanced real-time operation consumption state index analysis data of each node device, and compared with the historical operation data matrix, the real-time state of each node device is classified into levels; through the above analysis steps, the equipment balanced operation consumption state index of each historical period in the historical operation data matrix of each node device is extracted, and the historical comprehensive balanced operation consumption state index of each node device is obtained by performing mean analysis on each historical period; when PWh i <PWh i (L), the current node device is marked as low-loss operation state; when PWh i (L)≤PWh i ≤PWh i (Bm), the current node device is marked as medium-loss operation state; when PWh i >PWh i (Bm), the current node device is marked as a high-loss abnormal operation state; the node device in the high-loss abnormal operation state is marked as an abnormal node; Among them, PWh i (L) is the historical comprehensive balanced operation loss status index of the corresponding node equipment; PWh i (Bm) is the maximum health balance operation wear state index of the corresponding node;

[0074] Based on the normal operation of the node equipment, the real-time operation risk level of each node equipment is analyzed in combination with the real-time status level of each node equipment. The calculation formula is:

[0075] ;

[0076] Among them, Wf i is the real-time operation risk level of the device with the corresponding node number i; k e The value is k1 or k2, corresponding to the operation risk coefficient of low-loss level or medium-loss level node equipment respectively;

[0077] S303, in the regional distribution network node mapping model network, determine the location of the real-time distribution network demand transmission starting point and end point; screen the nodes in normal operation in the node network, and traverse the transmission transfer nodes according to the demand starting and end point location positioning results, and simulate the transmission route generation according to the traversed nodes to construct a set of simulated transmission routes;

[0078] Based on the proposed transmission route set, each proposed transmission route is extracted and combined with the operation risk degree of each operating equipment in the corresponding route to conduct risk assessment on each proposed transmission route. The calculation formula is:

[0079] ;

[0080] Among them, WL v is the transmission risk assessment rate of the proposed transmission line corresponding to number v; v is the transmission time of the simulated transmission line corresponding to number v; t min is the minimum transmission time of the historical transmission line corresponding to the target start and end points; ave(Wf i ) is the average risk level of each node in the region; i vmax and i vmin They correspond to the maximum and minimum number of nodes contained in the proposed transmission line; S v The transmission distance of the proposed transmission line corresponding to number v; S min The minimum transmission distance of the historical transmission line corresponding to the target starting and ending points; wl is the transmission loss rate per unit transmission distance; according to the transmission risk assessment rate analysis results of each proposed transmission line, sort in descending order to obtain MIN(WL v ) corresponds to the proposed transmission route which is the best transmission route in the current cycle.

[0081] The specific steps of S400 in which the status of each node device is fed back in combination with the node device status analysis result and the regional power transmission line is output in combination with the periodic power transmission line planning are as follows:

[0082] S401, obtaining the periodic operation status level of each node device in the area, and issuing warning prompts to abnormal devices; and providing real-time risk level feedback to normal operating devices;

[0083] S402: Display the proposed transmission route set and output the optimal transmission route periodically.

[0084] A power grid risk intelligent optimization system using a large language model, the system comprising a distribution network data acquisition module, a data model conversion and processing module, a distribution network comprehensive risk assessment module and an output feedback module;

[0085] The distribution network data acquisition module inspects the regional distribution network transmission lines through the drone inspection system, spatially locates each transmission line equipment and transmits the collected image data to the monitoring end; the data model conversion processing module uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts the synchronous operation data of each node through the edge cloud port, and constructs a corresponding node data cycle form; the distribution network comprehensive risk assessment module analyzes the periodic operation status of the node equipment according to the data cycle form of each node, and divides the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzes the status analysis results of each node equipment, simulates and fully traverses the regional transmission lines, and constructs a set of simulated transmission lines; evaluates the transmission risk rate of each simulated transmission line, and outputs the planned transmission line according to the evaluation results; the output feedback module provides status feedback for each node equipment in combination with the node equipment status analysis results, and outputs the regional transmission line in combination with the periodic transmission line planning situation.

[0086] The distribution network data acquisition module includes a distribution network data acquisition unit and a data communication unit;

[0087] The distribution network data acquisition unit locates the transmission lines and transfer transmission equipment of the regional distribution network by establishing a drone inspection network, and collects image data of each transmission equipment; records the connection paths of adjacent transfer transmission equipment;

[0088] The data communication unit transmits the collected image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment to the monitoring end through the regional communication network through the drone inspection network.

[0089] The data model conversion processing module includes a data model conversion unit and a data cycle form construction unit;

[0090] The data model conversion unit performs feature marking and data dimension reduction processing on the transmission lines and each transfer transmission equipment in each image data based on the image data of the regional distribution network transmission lines and transfer transmission equipment collected by the drone network through the feature extraction model; marks the feature points of each transfer transmission equipment, and explicitly marks the connection paths between the feature points; uses the three-dimensional virtual model to map and build a distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment, and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment;

[0091] The data cycle form construction unit builds an access channel to the edge cloud database port of each node device through the monitoring end, and extracts the historical operation data of each node device; using the equipment operation timeline as the data series axis, the node device operation data at each time point is annotated to construct a data cycle form corresponding to each node device.

[0092] The distribution network comprehensive risk assessment module includes a node equipment loss analysis unit, a node equipment risk assessment unit and a power transmission line comprehensive analysis unit;

[0093] The node device wear analysis unit calls the data cycle form of each node device, sets the observation cycle, takes the current real-time time point as the cycle acquisition starting point, takes the monitoring cycle as the step length, and performs reverse observation cycle data acquisition on the node data cycle form time period; takes the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extracts the observation cycle data of each node, and constructs the corresponding observation operation data matrix, extracts the data on the time axis before the observation cycle from the node data cycle form as historical data, and constructs the historical operation data matrix; based on the observation operation data matrix of each node device, analyzes the real-time operation wear status index of the node device at each time point in the cycle; Combine the real-time operation wear state index of the equipment at each time point in the observation period to construct a periodic equipment wear state index curve; use each time point in the observation period and the corresponding equipment real-time operation wear state index as coordinates, map coordinate points in a plane coordinate system, and connect each mapped coordinate through a smooth curve to construct a periodic equipment wear state index curve; locate the midpoint of the curve between adjacent coordinate points to obtain the coordinates corresponding to the midpoint between each adjacent coordinate point; traverse the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment to construct a corresponding characteristic midpoint coordinate set; calculate the balanced real-time operation wear state index of the equipment in the observation period based on the characteristic midpoint coordinate set;

[0094] The node equipment risk assessment unit combines the observation period balanced real-time operation wear state index analysis data of each node equipment, and classifies the real-time state of each node equipment by comparing it with the historical operation data matrix; through the above analysis steps, the equipment balanced operation wear state index of each historical period in the historical operation data matrix of each node equipment is extracted, and the historical comprehensive balanced operation wear state index of each node equipment is obtained by performing mean analysis on each historical period; combined with the historical comprehensive balanced operation wear state index and the maximum healthy balanced operation wear state index of each node equipment, the state level of each equipment is divided, and abnormal equipment is marked as abnormal node; based on the node equipment in normal operation, the real-time operation risk degree of each node equipment is analyzed in combination with the real-time state level of each node equipment;

[0095] The power transmission route comprehensive analysis unit determines the positions of the starting and ending points of the real-time distribution network demand transmission in the regional distribution network node mapping model network; screens the nodes in normal operation in the node network, and traverses the transmission transfer nodes according to the positioning results of the starting and ending points of the demand, and generates simulated power transmission routes according to the traversed nodes to construct a set of simulated power transmission routes; based on the set of simulated power transmission routes, extracts each simulated power transmission route and conducts risk assessment on each simulated power transmission route in combination with the operating risk degree of each operating equipment in the corresponding route; and sorts each simulated power transmission route in descending order according to the transmission risk assessment rate analysis results of each simulated power transmission route, and obtains the simulated power transmission route corresponding to the minimum value as the best power transmission route in the current cycle.

[0096] The output feedback module includes a device state feedback unit and a power transmission line output unit;

[0097] The device status feedback unit obtains the periodic operation status level of each node device in the area, issues warning prompts to abnormal devices, and provides real-time risk level feedback to normal operating devices;

[0098] The power transmission line output unit displays the proposed power transmission line set and outputs the optimal power transmission line periodically;

[0099] In the examples:

[0100] A distribution network system needs to monitor the regional power resources. It adopts the intelligent optimization system for power grid risk using a large language model of the present invention. It spatially locates the transmission lines and transfer transmission equipment of the regional distribution network by establishing an unmanned aerial vehicle inspection network, and collects image data of each transmission equipment. It records the connection paths of adjacent transfer transmission equipment. It transmits the collected image data of the transmission lines and transfer transmission equipment of the regional distribution network and the connection path data between the corresponding transfer transmission equipment to the monitoring end through the unmanned aerial vehicle inspection network.

[0101] Based on the image data of regional distribution network transmission lines and transfer transmission equipment collected by drone networking, the transmission lines and transfer transmission equipment in each image data are marked with features through a feature extraction model and data dimension reduction is performed; the feature points of each transfer transmission equipment are marked, and the connection paths between the feature points are explicitly marked; the data processed by the feature extraction model are mapped to a distribution network virtual model using a three-dimensional virtual model, by nodeizing each transfer transmission equipment and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment; an access channel is established through the monitoring end to the edge cloud database port of each node device, and the historical operation data of each node device is extracted; the equipment operation timeline is used as the data series axis, and the node equipment operation data at each time point is marked to construct a data cycle form corresponding to each node device;

[0102] Retrieve the data cycle form of each node device, set the observation cycle, take the current real-time time point as the cycle acquisition starting point, take the monitoring cycle as the step length, and perform reverse observation cycle data acquisition on the node data cycle form time period; take the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extract the observation cycle data of each node, and construct the corresponding observation operation data matrix, extract the data on the time axis before the observation cycle in the node data cycle form as historical data, and construct the historical operation data matrix; based on the observation operation data matrix of each node device, analyze the real-time operation consumption status index of the node device at each time point in the cycle, and its calculation formula is:

[0103] ;

[0104] Combine the real-time operation wear state index of the equipment at each time point in the observation period to construct a periodic equipment wear state index curve; use each time point in the observation period and the corresponding equipment real-time operation wear state index as coordinates, map the coordinate points in the plane coordinate system, and connect the mapped coordinates through a smooth curve to construct the periodic equipment wear state index curve; locate the midpoint of the curve between adjacent coordinate points to obtain the coordinates corresponding to the midpoint between each adjacent coordinate point; traverse the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment to construct the corresponding feature midpoint coordinate set; calculate the balanced real-time operation wear state index of the equipment in the observation period based on the feature midpoint coordinate set, and the calculation formula is:

[0105] ;

[0106] Combined with the observation period balanced real-time operation consumption state index analysis data of each node device, the real-time state of each node device is divided into levels according to the historical operation data matrix; through the above analysis steps, the equipment balanced operation consumption state index of each historical period in the historical operation data matrix of each node device is extracted, and the historical comprehensive balanced operation consumption state index of each node device is obtained by performing mean analysis on each historical period; when PWh i <PWh i (L), the current node device is marked as low-loss operation state; when PWh i (L)≤PWh i ≤PWh i (Bm), the current node device is marked as medium-loss operation state; when PWh i >PWh i(Bm), the current node device is marked as a high-loss abnormal operation state; the node device in the high-loss abnormal operation state is marked as an abnormal node; based on the normally operating node devices, the real-time operation risk level of each node device is analyzed in combination with the real-time status level of each node device. The calculation formula is:

[0107] ;

[0108] In the regional distribution network node mapping model network, the starting and ending points of the real-time distribution network demand transmission are located; the nodes in normal operation in the node network are screened, and according to the positioning results of the starting and ending points of the demand, the transmission transfer nodes are traversed, and the simulated transmission lines are generated according to the traversed nodes to construct a set of simulated transmission lines; based on the set of simulated transmission lines, each simulated transmission line is extracted and combined with the operating risk degree of each operating equipment in the corresponding route to conduct risk assessment on each simulated transmission line, and the calculation formula is:

[0109] ;

[0110] According to the transmission risk assessment rate analysis results of each proposed transmission line, the MIN(WL v ) corresponds to the proposed power transmission route being the best power transmission route in the current cycle;

[0111] Obtain the periodic operation status level of each node device in the area, and issue warning prompts for abnormal equipment; provide real-time risk level feedback for normally operating equipment; display the set of proposed transmission routes, and output the periodic optimal transmission route.

[0112] 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. A power grid risk intelligent optimization method using a large language model, characterized by: The method comprises the following steps: S100, inspect the transmission lines of the regional distribution network through the drone inspection system, spatially locate each transmission line equipment and transmit the collected image data to the monitoring end; S200, using the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracting synchronous operation data of each node through the edge cloud port, and constructing a corresponding node data cycle form; S300, analyzing the periodic operation status of node equipment according to the data period table of each node, and classifying the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzing the status results of each node equipment, traversing the regional power transmission lines in simulation, and constructing a set of proposed power transmission lines; evaluating the transmission risk rate of each proposed power transmission line, and outputting the planned power transmission line according to the evaluation results; The specific steps of S300 are as follows: S301, call the data cycle form of each node device, set the observation cycle, take the current real-time time point as the cycle acquisition starting point, take the monitoring cycle as the step length, and perform reverse observation cycle data acquisition on the node data cycle form time period; take the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extract the observation cycle data of each node, and construct the corresponding observation operation data matrix, extract the data on the time axis before the observation cycle in the node data cycle form as historical data, and construct the historical operation data matrix; Based on the observed operation data matrix of each node device, the real-time operation wear state index of the node device at each time point in the cycle is analyzed, and the calculation formula is: ； Among them, Wh (i,t) is the real-time operation wear state index of the corresponding node device numbered i at time point t within the observation period; α, β, γ are the equipment operation environment data wear coefficient, electrical data wear coefficient, and mechanical data wear coefficient respectively; n, q, z are the number of equipment operation environment data types, the number of operation electrical data types, and the number of operation mechanical data types respectively; M n , D q , J z They correspond to the operating environment data of each type of equipment, the electrical data of each type of equipment, and the mechanical data of each type of equipment; n (s), D q (s), J z (s) respectively correspond to the prescribed standard data for the operating environment data of each type of equipment, the prescribed standard data for the operating electrical data of each type of equipment, and the prescribed standard data for the operating mechanical data of each type of equipment; the real-time operating wear state index of the equipment at each time point in the observation period is combined to construct a periodic equipment wear state index curve; by using each time point in the observation period and the corresponding equipment real-time operating wear state index as coordinates, coordinate points are mapped in the plane coordinate system, and the mapped coordinates are connected by a smooth curve to construct a periodic equipment wear state index curve; by locating the midpoint of the curve between adjacent coordinate points, the coordinates corresponding to the midpoint between each adjacent coordinate point are obtained; the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the equipment are traversed to construct a corresponding characteristic midpoint coordinate set; based on the characteristic midpoint coordinate set, the balanced real-time operating wear state index of the equipment in the observation period is calculated, and the calculation formula is: ； Among them, PWh i is the balanced real-time operation wear state index of the corresponding node device numbered i during the observation period; u is the number of elements in the characteristic midpoint coordinate set; Wh (i,t,t+1) The midpoints of the adjacent coordinate points of the periodic equipment wear state index curve correspond to the real-time operation wear state index; S302, combined with the observation period balanced real-time operation consumption state index analysis data of each node device, and compared with the historical operation data matrix, the real-time state of each node device is classified into levels; through the above analysis steps, the equipment balanced operation consumption state index of each historical period in the historical operation data matrix of each node device is extracted, and the historical comprehensive balanced operation consumption state index of each node device is obtained by performing mean analysis on each historical period; when PWh i <PWh i (L), the current node device is marked as low-loss operation state; when PWh i (L)≤PWh i ≤PWh i (Bm), the current node device is marked as medium-loss operation state; when PWh i >PWh i (Bm), the current node device is marked as a high-loss abnormal operation state; the node device in the high-loss abnormal operation state is marked as an abnormal node; Among them, PWh i (L) is the historical comprehensive balanced operation loss status index of the corresponding node equipment; PWh i (Bm) is the maximum health balance operation wear state index of the corresponding node; Based on the normal operation of the node equipment, the real-time operation risk level of each node equipment is analyzed in combination with the real-time status level of each node equipment. The calculation formula is: ； Among them, Wf i is the real-time operation risk level of the device with the corresponding node number i; k e The value is k1 or k2, corresponding to the operation risk coefficient of low-loss level or medium-loss level node equipment respectively; S303, in the regional distribution network node mapping model network, determine the location of the real-time distribution network demand transmission starting point and end point; screen the nodes in normal operation in the node network, and traverse the transmission transfer nodes according to the demand starting and end point location positioning results, and simulate the transmission route generation according to the traversed nodes to construct a set of simulated transmission routes; Based on the proposed transmission route set, each proposed transmission route is extracted and combined with the operation risk degree of each operating equipment in the corresponding route to conduct risk assessment on each proposed transmission route. The calculation formula is: ； Among them, WL v is the transmission risk assessment rate of the proposed transmission line corresponding to number v; v is the transmission time of the simulated transmission line corresponding to number v; t min is the minimum transmission time of the historical transmission line corresponding to the target start and end points; ave(Wf i ) is the average risk level of each node in the region; i vmax and i vmin They correspond to the maximum and minimum number of nodes contained in the proposed transmission line; S v The transmission distance of the proposed transmission line corresponding to number v; S min The minimum transmission distance of the historical transmission line corresponding to the target starting and ending points; wl is the transmission loss rate per unit transmission distance; according to the transmission risk assessment rate analysis results of each proposed transmission line, sort in descending order to obtain MIN(WL v ) corresponds to the proposed power transmission route being the best power transmission route in the current cycle; S400, providing status feedback for each node device in combination with the node device status analysis result, and outputting the regional power transmission line in combination with the periodic power transmission line planning situation.

2. The method for intelligent optimization of power grid risk using a large language model according to claim 1, characterized in that: The specific steps of the S100 inspecting the transmission lines of the regional distribution network through the drone inspection system, spatially locating each transmission line equipment and transmitting the collected image data to the monitoring end are as follows: S101. Establishing a drone inspection network to locate the spatial position of the transmission lines and transfer transmission equipment in the regional distribution network, and collect image data of each transmission equipment; record the connection paths of adjacent transfer transmission equipment; S102, using the drone inspection network to collect the image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment, and transmit them to the monitoring end using the regional communication network.

3. The method for intelligent optimization of power grid risk using a large language model according to claim 2 is characterized in that: The S200 uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts synchronous operation data of each node through the edge cloud port, and constructs the corresponding node data cycle form in the following specific steps: S201, based on the image data of the power transmission lines and transfer power transmission equipment of the regional distribution network collected by the drone network, feature-mark the power transmission lines and each transfer power transmission equipment in each image data through a feature extraction model and perform data dimensionality reduction processing; Mark the characteristic points of each transfer transmission equipment, and make the connection paths between the characteristic points explicit and annotated; use the three-dimensional virtual model to map and build the distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment; S202, building an access channel to the edge cloud database port of each node device through the monitoring terminal, and extracting the historical operation data of each node device; Taking the equipment operation time axis as the data series axis, the node equipment operation data at each time point is annotated to construct a data cycle form corresponding to each node equipment.

4. The method for intelligent optimization of power grid risk using a large language model according to claim 3 is characterized in that: The specific steps of S400 in which the status of each node device is fed back in combination with the node device status analysis result and the regional power transmission line is output in combination with the periodic power transmission line planning are as follows: S401, obtaining the periodic operation status level of each node device in the area, and issuing warning prompts to abnormal devices; and providing real-time risk level feedback to normal operating devices; S402: Display the proposed transmission route set and output the optimal transmission route periodically.

5. A power grid risk intelligent optimization system using a large language model, using a power grid risk intelligent optimization method using a large language model as claimed in any one of claims 1 to 4, characterized in that: The system includes a distribution network data acquisition module, a data model conversion processing module, a distribution network comprehensive risk assessment module and an output feedback module; The distribution network data acquisition module inspects the regional distribution network transmission lines through the drone inspection system, and spatially locates each transmission line equipment and transmits the collected image data to the monitoring end; the data model conversion processing module uses the regional distribution network transmission line and equipment image data to construct a spatial node model mapping, extracts the synchronous operation data of each node through the edge cloud port, and constructs a corresponding node data cycle form; The distribution network comprehensive risk assessment module analyzes the periodic operation status of node equipment according to the data period form of each node, and classifies the status level of each node equipment in combination with the historical equipment operation data; comprehensively analyzes the status analysis results of each node equipment, simulates and fully traverses the regional power transmission lines, and constructs a set of proposed power transmission lines; evaluates the transmission risk rate of each proposed power transmission line, and outputs the planned power transmission line according to the evaluation results; The output feedback module provides status feedback to each node device in combination with the node device status analysis result, and outputs the regional power transmission line in combination with the periodic power transmission line planning situation.

6. The power grid risk intelligent optimization system using a large language model according to claim 5 is characterized by: The distribution network data acquisition module includes a distribution network data acquisition unit and a data communication unit; The distribution network data acquisition unit locates the transmission lines and transfer transmission equipment of the regional distribution network by establishing a drone inspection network, and collects image data of each transmission equipment; records the connection paths of adjacent transfer transmission equipment; The data communication unit transmits the collected image data of the regional distribution network transmission lines and transfer transmission equipment and the connection path data between the corresponding transfer transmission equipment to the monitoring end through the regional communication network through the drone inspection network.

7. The power grid risk intelligent optimization system using a large language model according to claim 6 is characterized by: The data model conversion processing module includes a data model conversion unit and a data cycle form construction unit; The data model conversion unit performs feature marking and data dimension reduction processing on the transmission lines and each transfer transmission equipment in each image data through a feature extraction model based on the image data of the transmission lines and transfer transmission equipment of the regional distribution network collected by the drone network; Mark the characteristic points of each transfer transmission equipment, and make the connection paths between the characteristic points explicit and annotated; use the three-dimensional virtual model to map and build the distribution network virtual model for the data processed by the feature extraction model, by nodeizing each transfer transmission equipment and connecting the corresponding nodes according to the connection paths of each adjacent transfer transmission equipment; The data cycle form construction unit builds an access channel to the edge cloud database port of each node device through the monitoring end, and extracts the historical operation data of each node device; using the equipment operation timeline as the data series axis, the node device operation data at each time point is annotated to construct a data cycle form corresponding to each node device.

8. The power grid risk intelligent optimization system using a large language model according to claim 7 is characterized by: The distribution network comprehensive risk assessment module includes a node equipment loss analysis unit, a node equipment risk assessment unit and a power transmission line comprehensive analysis unit; The node device wear analysis unit calls the data cycle form of each node device, sets the observation cycle, takes the current real-time time point as the cycle acquisition starting point, takes the monitoring cycle as the step length, and performs reverse observation cycle data acquisition on the node data cycle form time period; takes the currently acquired observation cycle data as the real-time status analysis object of the corresponding node device; extracts the observation cycle data of each node, and constructs the corresponding observation operation data matrix, extracts the data on the time axis before the observation cycle from the node data cycle form as historical data, and constructs the historical operation data matrix; Based on the observed operation data matrix of each node device, the real-time operation wear state index of the node device at each time point in the cycle is analyzed; the periodic equipment wear state index curve is constructed in combination with the real-time operation wear state index of the device at each time point in the observation cycle; by using each time point in the observation cycle and the corresponding equipment real-time operation wear state index as coordinates, coordinate points are mapped in the plane coordinate system, and the periodic equipment wear state index curve is constructed by connecting each mapped coordinate through a smooth curve; by locating the midpoint of the curve between adjacent coordinate points, the coordinates corresponding to the midpoint between each adjacent coordinate point are obtained; the midpoint coordinates between all adjacent coordinates on the periodic equipment wear state index curve of the device are traversed to construct the corresponding feature midpoint coordinate set; Based on the characteristic midpoint coordinate set, the balanced real-time operation wear state index of the equipment within the observation period is calculated; The node equipment risk assessment unit combines the observation period balanced real-time operation wear state index analysis data of each node equipment, and classifies the real-time state of each node equipment by comparing it with the historical operation data matrix; through the above analysis steps, the equipment balanced operation wear state index of each historical period in the historical operation data matrix of each node equipment is extracted, and the historical comprehensive balanced operation wear state index of each node equipment is obtained by performing mean analysis on each historical period; combined with the historical comprehensive balanced operation wear state index and the maximum healthy balanced operation wear state index of each node equipment, the state level of each equipment is classified, and abnormal equipment is marked as abnormal node; Based on the normally operating node devices, the real-time operation risk level of each node device is analyzed in combination with the real-time status level of each node device; The power transmission route comprehensive analysis unit determines the location of the real-time distribution network demand transmission starting point and end point in the regional distribution network node mapping model network; screens the nodes in normal operation in the node network, and traverses the transmission transfer nodes according to the demand starting and end point location positioning results, and simulates the power transmission route generation according to the traversed nodes to construct a set of simulated power transmission routes; based on the set of simulated power transmission routes, extracts each simulated power transmission route and conducts risk assessment on each simulated power transmission route in combination with the operating risk degree of each operating equipment in the corresponding route; According to the transmission risk assessment rate analysis results of each proposed transmission line, the proposed transmission line corresponding to the minimum value is sorted in descending order and is the best transmission line in the current cycle.

9. The power grid risk intelligent optimization system using a large language model according to claim 8, characterized in that: The output feedback module includes a device state feedback unit and a power transmission line output unit; The device status feedback unit obtains the periodic operation status level of each node device in the area and issues warning prompts for abnormal devices; Provide real-time risk feedback for equipment in normal operation; The transmission line output unit displays the proposed transmission line set and outputs the optimal transmission line periodically.

Images