Demand perception and resource coordination system for power grid operation and maintenance

Abstract

The present application relates to the technical field of power intelligent operation and maintenance, and more particularly to a demand sensing and resource collaborative configuration system for power grid operation and maintenance, comprising an operation and maintenance scheduling platform, a database construction module, a three-layer coupling correlation model, a demand prediction module, a configuration scheme matching module and a self-adaptive optimization module; the present application is to preliminarily construct a multi-dimensional data source library, extract equipment health degradation features and environment-power grid coupling features, construct a three-layer coupling space-time correlation graph, combine a demand prediction model, output high-risk demand locations, demand types and resource demands, generate operation and maintenance demand heat maps and automatically trigger early warning events, and through a double-portrayal fusion matching model, realize accurate matching and dynamic optimization of personnel, vehicles and spare parts, finally output an optimal resource configuration scheme and execute it, realize active prediction of power grid operation and maintenance demand, accurate configuration of operation and maintenance resources and full-process closed-loop management and control, and improve fault early warning capability and resource scheduling efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for power grids, and in particular to a demand perception and resource collaborative allocation system for power grid operation and maintenance. Background Technology

[0002] With the rapid development of the power system, the scale of the power grid continues to expand, the number of equipment continues to increase, and the complexity of power grid operation and maintenance management is increasing day by day. Traditional power grid operation and maintenance methods mainly rely on manual experience, regular inspections and threshold alarms. Under the increasingly complex operating environment and diverse external influencing factors, many technical defects have been gradually exposed.

[0003] Currently, common practices in power grid operation and maintenance include: regular maintenance and preventive testing, threshold alarm mechanisms, manual experience-based dispatching, and independent use of meteorological early warning systems. However, actual technical analysis reveals the following shortcomings: 1. The use of threshold alarms makes it impossible to effectively model the gradual degradation process of equipment health. Equipment often undergoes a long-term performance degradation process from normal operation to failure. Traditional technologies struggle to capture the temporal dependencies and multi-feature coupling patterns in this process, resulting in delayed fault prediction and an inability to achieve proactive early warning and intervention. At the same time, there is a lack of quantitative modeling capabilities for the correlation between meteorological factors and equipment failures, making it impossible to assess the risk of cascading failures caused by meteorological disasters, resulting in insufficient emergency response capabilities under extreme weather conditions. 2. There is a lack of effective means to characterize the spatiotemporal distribution of operation and maintenance needs, making it difficult to quantify the intensity of operation and maintenance needs in different regions and time windows. Operation and maintenance decisions often rely on experience-based judgments and cannot intuitively present the distribution of high-risk areas and demand intensity, which affects the scientific deployment of operation and maintenance resources. 3. The scheduling needs of operation and maintenance resources (personnel, vehicles, spare parts) and the matching between resources lack quantitative standards, which can easily lead to mismatch between capabilities and untimely updates of resource status information, making it impossible to grasp the real-time availability of resources.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] The present invention aims to provide a demand perception and resource collaborative configuration system for power grid operation and maintenance, so as to solve the technical defects mentioned in the background art.

[0006] The objective of this invention can be achieved through the following technical solution: a demand perception and resource collaborative configuration system for power grid operation and maintenance, including an operation and maintenance scheduling platform, a database construction module, a three-layer coupled association model, a demand prediction module, a configuration scheme matching module, and an adaptive optimization module; The database construction module is used to collect multi-source heterogeneous data from the target area, preprocess the collected multi-source heterogeneous data, including data cleaning and standardization, build a multi-dimensional data source library, and send it to the operation and maintenance scheduling platform for storage. The three-layer coupling association model is used to extract equipment health degradation features and environment-grid coupling features respectively. With grid equipment as nodes and electrical connection relationships as edges, a grid topology graph is constructed, and a three-layer coupling graph structure of equipment-environment-space is also constructed. The demand forecasting module is used to retrieve the demand forecasting model based on the three-layer coupled spatiotemporal correlation graph, output three-dimensional forecasting results, including the location of high-risk demand, demand type and resource demand, and generate an operation and maintenance demand heat map and output operation and maintenance demand early warning events. The configuration scheme matching module is used to build a capability profile of early warning requirements based on early warning events of operation and maintenance requirements, and at the same time collect operation and maintenance resource information across the entire domain and build a resource capability profile. Combined with a preset dual-profile fusion matching model, it outputs a preliminary resource configuration scheme. The adaptive optimization module performs adaptive optimization on the initial configuration scheme, taking into account the dynamic status of resources and changes in operating conditions, while resolving resource conflicts and outputting the optimal resource configuration scheme.

[0007] Preferably, the analysis process for the health degradation characteristics of the extraction device is as follows: Based on equipment status data from a multidimensional data source library, equipment health degradation features are extracted, including electrical, chemical, thermal, and mechanical features. These features are then used as input layers to a preset equipment maintenance requirement prediction model. The model outputs the probability and type of equipment failure within a preset future time window, and the failure type corresponding to the maximum failure probability is set as the final predicted failure type. Retrieve the preset fault type-operation resource type association table, match the final predicted fault type with the preset fault type-operation resource type association table, and output the matched operation resource type.

[0008] Preferably, the analysis process for extracting the environment-power grid coupling features is as follows: Collect historical meteorological event data and corresponding equipment failure records, classify them by meteorological type and by equipment type, and statistically analyze the failure probability of each type of equipment under various meteorological conditions to form a vulnerability matrix M{m×n}, where m is the number of meteorological types and n is the number of equipment types; Extract spatial features from meteorological radar images to identify the direction of movement and the range of impact of severe convective weather; Based on the power grid topology, a directed graph of fault propagation is established. The probability P(j,i) of device j being affected after device i fails is calculated. Combined with the area affected by meteorological disasters, the cascade fault risk value R = ΣP(i)×P(j,i) of devices in the area is calculated. When the cascade fault risk value exceeds a preset threshold, it is marked as a multiple fault risk point. Spatiotemporal correlation graph construction: Using power grid equipment as nodes and electrical connection relationships as edges, a power grid topology graph is constructed. The predicted failure probability of equipment, meteorological influence factors, and historical work order density are used as node attributes and embedded into the power grid topology graph. By introducing geographic information system coordinates and using spatial proximity as additional edge weights, a three-layer coupled graph structure of device-environment-space is constructed.

[0009] Preferably, the analysis process of the demand forecasting module is as follows: Based on the three-layer coupled diagram structure of equipment-environment-space, the demand forecasting model is retrieved and the three-dimensional forecasting results are output, including the high-risk demand locations, demand types and resource requirements identified by equipment IDs. The target area is divided into equal-scale grids, and the failure probability, equipment importance weight, and user sensitivity weight are comprehensively calculated and predicted for each grid to form a comprehensive demand intensity index. The demand intensity index is mapped to color depth: demand intensity index < E1, low intensity, corresponding to light blue; E1 ≤ demand intensity index < E2, medium intensity, corresponding to yellow; E2 ≤ demand intensity index < E3, high intensity, corresponding to orange; demand intensity index ≥ E3, very high intensity, corresponding to dark red. Among them, 0 < E1 < E2 < E3 < 100. A dynamically updated heat map of operation and maintenance demand is generated and displayed in layers according to the time dimension. When the comprehensive demand intensity index is extremely high, an operation and maintenance demand warning event will be automatically generated.

[0010] Preferably, the analysis process for the early warning demand capability profile and resource capability profile is as follows: Based on the early warning events of operation and maintenance needs, collect power grid operation and maintenance early warning need information, and construct an early warning need capability profile based on the collected power grid operation and maintenance early warning need information; The system collects basic information, capability information, and dynamic status information of various operation and maintenance resources in real time, builds a global operation and maintenance resource pool, and creates a unique resource capability profile for each type of resource. The resource capability profiles specifically include: personnel resource profiles, vehicle resource profiles, and spare parts resource profiles.

[0011] Preferably, based on the demand capability profile and resource capability profile, a preset dual-profile fusion matching model is invoked, and the principles of capability similarity priority, priority priority, and distance priority are adopted to achieve accurate matching of demand and resources. The specific implementation process is as follows: Input to the preset dual-profile fusion matching model: Simultaneously input the completed requirement capability profile and resource capability profile into the preset dual-profile fusion matching model; Model fusion computation: The model automatically completes two core calculations—priority determination and capability similarity calculation; Directly output preliminary solution: Based on the fusion calculation results, the model automatically completes the priority matching and hierarchical matching of high-priority requirements, and directly outputs a complete preliminary resource configuration solution including personnel, vehicles and spare parts.

[0012] Preferably, the analysis process for the optimal resource allocation scheme is as follows: Real-time monitoring of resource status to determine if any resource is abnormal. If a resource is abnormal, the system automatically retrieves the second most similar available resource from the resource pool, replaces the abnormal resource, updates the configuration scheme, and simultaneously determines if the operating conditions have changed. If the operating conditions have changed, the system triggers a resource configuration adjustment command. Further conflict resolution measures were implemented for the initial configuration scheme: when multiple priority requests compete for the same resource, the conflict was resolved by prioritizing the use of high-priority requests and replacing low-priority requests with alternative resources. When requests of the same priority compete for the same resource, the more urgent and similar requests were prioritized based on the time limit and capability similarity, and the remaining requests were re-matched with alternative resources. Cost optimization: Under the premise of meeting the demand capacity requirements, a weighted fusion is performed based on the distance of resources and scheduling costs. The matching optimization coefficient is calculated based on the normalized distance × preset distance ratio coefficient + scheduling cost × preset cost ratio coefficient. The resource corresponding to the minimum value of the matching optimization coefficient is set as the target resource, and the optimal resource allocation scheme is finally output.

[0013] The beneficial effects of this invention are as follows: (1) This invention provides comprehensive and accurate data support for equipment status assessment, fault prediction and operation and maintenance decision-making by constructing a unified multidimensional data source library. Based on the multidimensional data source library, a three-layer coupled spatiotemporal correlation diagram integrating electrical topology, meteorological influence and geospatial is constructed to provide high-quality graph data support for demand forecasting. Furthermore, gridded demand intensity index calculation and heat map generation are adopted to realize the quantitative assessment and spatiotemporal dynamic display of operation and maintenance demand, and assist in quickly locating key protection areas.

[0014] (2) Based on operation and maintenance requirements, this invention constructs a dual profile of requirements and resources, adopts a multi-dimensional matching mechanism of capability similarity, priority, and distance, and realizes accurate configuration of personnel, vehicles, and spare parts through the dual profile fusion model, thereby improving resource utilization and response speed. At the same time, it monitors resource status and working condition changes in real time, automatically replaces abnormal resources, adopts a priority-first strategy to resolve resource conflicts, and combines a cost optimization model to ensure the reliability and economy of the scheduling scheme before finally issuing and executing it. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings; Fig. 1 This is a flowchart of the system of the present invention; Fig. 2 This is a reference analysis diagram of a three-layer coupled correlation model; Fig. 3 This is a schematic diagram illustrating the analysis of the three-dimensional prediction results of this invention; Fig. 4 This is a schematic diagram of the heat map analysis of operation and maintenance requirements in this invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments; Example 1: Please refer to Figs. 1 to 4 As shown, this invention is a demand perception and resource collaborative configuration system for power grid operation and maintenance, including an operation and maintenance scheduling platform, a database construction module, a three-layer coupled association model, a demand prediction module, a configuration scheme matching module, and an adaptive optimization module. The operation and maintenance scheduling platform and the database construction module have bidirectional communication connections, and the operation and maintenance scheduling platform has unidirectional communication connections with both the three-layer coupled association model and the configuration scheme matching module. The three-layer coupled association model has a unidirectional communication connection with the demand prediction module, the demand prediction module has a unidirectional communication connection with the operation and maintenance scheduling platform, the configuration scheme matching module has a unidirectional communication connection with the adaptive optimization module, and the adaptive optimization module has a unidirectional communication connection with the operation and maintenance scheduling platform. The database construction module is used to collect multi-source heterogeneous data such as power grid operation data, equipment online monitoring data, environmental meteorological data, geographic GIS data, historical fault work orders, load forecast data, and video surveillance data in the target area. It preprocesses the collected multi-source heterogeneous data, including data cleaning and standardization, builds a multi-dimensional data source database, and sends it to the operation and maintenance scheduling platform for storage. The three-layer coupling and association model is used to extract equipment health degradation features and environment-grid coupling features respectively. Using grid equipment as nodes and electrical connections as edges, it constructs a grid topology graph and simultaneously integrates and constructs a three-layer coupling graph structure of equipment, environment, and space, specifically including: Equipment-level requirement feature extraction: Based on equipment status data in a multi-dimensional data source library, a Long Short-Term Memory (LSTM) network is used to extract equipment health degradation features (including at least one of electrical, chemical, thermal, mechanical, and environmental features). The equipment health degradation features are used as input layers to a preset equipment operation and maintenance requirement prediction model, and the output is the probability and type of equipment failure within a preset future time window. The failure type corresponding to the maximum failure probability is set as the final predicted failure type. Retrieve the preset fault type-maintenance resource type association table, match the final predicted fault type with the preset fault type-maintenance resource type association table, and output the matched maintenance resource type (such as test personnel, crane, spare parts model). Environment-Power Grid Coupling Feature Extraction: Historical meteorological event data and corresponding equipment failure records are collected and classified by meteorological type, which includes at least one of typhoon, thunderstorm, icing, high temperature, and pollution. Equipment types are also classified by equipment type, which includes at least one of transformer, circuit breaker, transmission line, cable, and disconnector. The failure probability of each type of equipment under various meteorological conditions is statistically analyzed to form a vulnerability matrix M{m×n}, where m (m>0) is the number of meteorological types and n (n>0) is the number of equipment types. Spatial features of weather radar images are extracted using convolutional neural networks (CNNs) to identify spatial feature vectors such as the direction of movement and the range of impact of severe convective weather. Based on the power grid topology, a directed fault propagation graph is established. The probability P(j,i) of device j (j>0) being affected after device i (i>0) fails is calculated. Combined with the meteorological disaster impact area, the cascade fault risk value R = ΣP(i)×P(j,i) of the devices in the area is calculated. When the cascade fault risk value exceeds the preset threshold, it is marked as a multiple fault risk point. Spatiotemporal correlation graph construction: Using power grid equipment as nodes and electrical connection relationships as edges, a power grid topology graph is constructed. The predicted failure probability of equipment, meteorological influence factors (such as movement direction and influence range), and historical work order density are used as node attributes and embedded into the power grid topology graph. By introducing Geographic Information System (GIS) coordinates and using spatial proximity as an additional edge weight, a three-layer coupled graph structure of device-environment-space is constructed. The spatial proximity is determined by the Euclidean distance between devices. When the distance is less than a preset threshold, a spatial association edge is established, and the normalized Euclidean distance is set as the weight.

[0018] Example 2: The demand forecasting module is used to retrieve the demand forecasting model based on a three-layer coupled spatiotemporal correlation graph, output three-dimensional prediction results, including the location of high-risk demands, demand types, and resource requirements, and generate an operation and maintenance demand heatmap and output operation and maintenance demand early warning events, specifically including: Based on the three-layer coupling diagram structure of equipment-environment-space, the demand forecasting model is retrieved and the three-dimensional forecasting results are output, including the high-risk demand locations identified by equipment ID, demand types (including emergency repair / planned maintenance / special inspection / emergency power supply), and resource requirements (such as professional personnel, vehicle type, and spare parts model). The target area is divided into equal-scale grids (e.g., 1km×1km), and the failure probability, equipment importance weight, and user sensitivity weight are comprehensively calculated and predicted for each grid to form a comprehensive demand intensity index. Among them, the equipment importance weight is manually assigned based on voltage level, load level, whether it is a core site, and whether it is a power line to be protected, and normalized to [0, 1]. The larger the value, the higher the equipment importance. User sensitivity weights are manually assigned based on the type of power supply user (residential / critical power supply / industrial / hospital / government), and normalized to [0, 1]. The larger the value, the higher the user sensitivity. First, calculate the demand intensity S for each device in the grid = predicted failure probability × device importance weight × user sensitivity weight. That is, the higher the failure probability, the more important the device, and the more sensitive the user, the stronger the maintenance demand for the device. Sum the demand intensity S of all devices in a single grid and normalize it to obtain the grid comprehensive demand intensity index. The grid comprehensive demand intensity index is [0, 100]. The demand intensity index is mapped to color depth: demand intensity index < E1, low intensity, corresponding to light blue; E1 ≤ demand intensity index < E2, medium intensity, corresponding to yellow; E2 ≤ demand intensity index < E3, high intensity, corresponding to orange; demand intensity index ≥ E3, very high intensity, corresponding to dark red. Where 0 < E1 < E2 < E3 < 100, a dynamically updated heatmap of operation and maintenance demand is generated and displayed in layers according to the time dimension (such as current, next 1 hour, next 4 hours). When the comprehensive demand intensity index is extremely high, an operation and maintenance demand early warning event is automatically generated, and the three-dimensional prediction results and the operation and maintenance demand early warning event are sent to the operation and maintenance scheduling platform for storage.

[0019] Example 3: The configuration scheme matching module is used to construct a capability profile of early warning requirements based on operation and maintenance requirement early warning events. Simultaneously, it collects operation and maintenance resource information across the entire domain and constructs a resource capability profile. Combining this with a preset dual-profile fusion matching model, it outputs a preliminary resource configuration scheme, specifically including: Based on maintenance demand warning events, collect power grid maintenance demand information. The warning demand information includes at least: unique equipment ID, precise equipment location (latitude and longitude, work area, traffic conditions), demand type (emergency repair / planned maintenance / special inspection / emergency power supply), fault / hazard description, work time limit, resource demand details, and demand capability requirements. Based on the collected power grid operation and maintenance early warning demand information, a capability profile of early warning demand is constructed. Different capability requirements are defined for different demand types, forming a standardized set of demand capabilities, specifically including: Basic information: Equipment ID, location, demand type, operation time limit, and scope of impact; Personnel competency requirements: professional type, skill level, practical experience, emergency response capability (specific to emergency repair / emergency power supply), and process proficiency (specific to planned maintenance). Vehicle capability requirements: type, functional configuration, response speed, range / load capacity, environmental adaptability (specific to special patrol vehicles); Spare parts capability requirements: model, specifications, compatibility, quality grade, ease of replacement; Real-time collection of basic information, capability information, and dynamic status information of various operation and maintenance resources; construction of a global operation and maintenance resource pool; and creation of a unique resource capability profile for each type of resource to achieve visualized management and control of resource capabilities. The resource capability profile specifically includes: Personnel resource profile: Personnel ID, professional type, skill level, practical experience, emergency response cases, current on-duty status, available time period, current location, and workload; Vehicle resource profile: Vehicle ID, type, function configuration, response speed (historical average dispatch response time), range / capacity, current status (available / occupied / faulty), current location, maintenance records; Spare parts resource profile: spare parts ID, model, specifications, compatible device ID, quality grade, ease of replacement, warehouse, inventory quantity, availability, and delivery cycle; At the same time, a dynamic update mechanism for the entire domain operation and maintenance resource pool is established to synchronize resource status changes (such as personnel shifts, vehicle fault repair, spare parts issuance / replenishment) and capability improvements (such as personnel skill upgrades) in real time, ensuring the accuracy of resource capability profiles. Based on demand capability profiles and resource capability profiles, a pre-defined dual-profile fusion matching model is invoked. Following the principles of capability similarity priority, priority priority, and distance priority, accurate matching of demands and resources is achieved. The specific implementation process is as follows: Input to the preset dual-profile fusion matching model: Simultaneously input the completed demand capability profile (including equipment ID, location, demand type, and resource capability requirements, etc.) and resource capability profile (personnel, vehicles, spare parts) into the preset dual-profile fusion matching model; Model fusion calculation: The model automatically completes two core calculations—priority determination (combining demand type, scope of impact, and task time limit, dividing into first-level, second-level, and third-level categories, and automatically sorting them) and capability similarity calculation; Priority determination: Based on the type of need, scope of impact, and time limit of operation, the early warning needs are divided into Level 1 (emergency repair, emergency power supply), Level 2 (planned maintenance, special inspection of major hidden dangers), and Level 3 (routine special inspection, general equipment maintenance), and the priority is sorted from high to low. Capability similarity calculation: Calculate the similarity between the capability profiles of personnel, vehicles, and spare parts and the required capability requirements. Only when the similarity reaches a preset threshold (such as 80%) can it be included in the matching range. Directly output preliminary plan: Based on the fusion calculation results, the model automatically completes the priority matching and hierarchical matching of high-priority requirements (personnel-vehicle-spare parts), and directly outputs a complete preliminary resource configuration plan including personnel, vehicles and spare parts, clarifying the allocation quantity, adaptation basis and arrival time of each resource; The adaptive optimization module performs adaptive optimization on the initial configuration scheme, taking into account the dynamic status of resources and changes in operating conditions. It also resolves resource conflicts and outputs the optimal resource configuration scheme, specifically including: Dynamic optimization: Real-time monitoring of resource status to determine if any resource is abnormal. If a resource is abnormal (e.g., vehicle malfunction, personnel emergency, spare parts damage), the system automatically retrieves the second most similar available resource from the resource pool, replaces the abnormal resource, updates the configuration scheme, and simultaneously checks if the operating conditions have changed. If the operating conditions have changed (e.g., extreme weather increases the difficulty of operations), a resource configuration adjustment command is triggered, and the system responds to the preset early warning operations corresponding to the resource configuration adjustment command, reminding operation and management personnel to add emergency personnel, replace suitable vehicles, etc. Conflict resolution: When multiple priority requests compete for the same resource, the conflict is resolved by "high priority requests are given priority and low priority requests are replaced by alternative resources"; when requests of the same priority compete for the same resource, the more urgent and similar requests are prioritized for matching, and the remaining requests are re-matched with alternative resources, taking into account the time limit and capability similarity of the requests. Cost optimization: Under the premise of meeting the demand capacity requirements, the resource distance and scheduling cost are weighted and fused. The matching optimization coefficient is calculated based on the normalized distance × preset distance ratio coefficient + scheduling cost × preset cost ratio coefficient. The resource corresponding to the minimum value of the matching optimization coefficient is set as the target resource, and the optimal resource allocation scheme is finally output. The optimized final resource allocation plan (directly output from the dual-profile fusion model and optimized and adjusted) is distributed to the corresponding resource terminals (personnel mobile phones, vehicle dispatch terminals, and warehouse management systems) through the operation and maintenance scheduling platform, clarifying the work tasks, arrival time, capacity requirements, and work standards of each resource. In summary, by collecting heterogeneous data from multiple sources and performing cleaning and standardization, a unified multidimensional data source library is constructed, providing comprehensive and accurate data support for equipment status assessment, fault prediction, and operation and maintenance decisions. Simultaneously, LSTM is used to extract multidimensional health degradation characteristics of equipment, accurately predicting fault probability and type, and realizing a shift from passive response to proactive early warning. The fault-resource association table automatically matches operation and maintenance resources, shortening response time. Furthermore, by constructing a meteorological-equipment vulnerability matrix, multiple fault risk points are identified in advance, enhancing early warning capabilities under extreme weather conditions. Finally, a three-layer coupled spatiotemporal association map integrating electrical topology, meteorological influence, and geospatial data is constructed, providing high-quality graph data support for demand forecasting. Additionally, gridded demand intensity index calculation and heat map generation enable quantitative assessment and dynamic spatiotemporal display of operation and maintenance needs, assisting in the rapid location of key protection areas. Based on operational needs, a dual profile of needs and resources is constructed. A multi-dimensional matching mechanism based on capability similarity, priority, and distance is adopted. Through the dual profile fusion model, the precise configuration of personnel, vehicles, and spare parts is achieved, improving resource utilization and response speed. At the same time, resource status and operating condition changes are monitored in real time, abnormal resources are automatically replaced, and a priority-first strategy is adopted to resolve resource conflicts. Combined with a cost optimization model, the reliability and economy of the scheduling plan are ensured before finally being issued for execution.

[0020] The threshold is set for comparative analysis of results to determine whether they are good or bad. The value of the threshold is determined by a combination of large-scale model analysis of sample data and human experience. It can also be adjusted appropriately based on seasonal or common-sense influencing factors.

[0021] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A demand perception and resource collaborative allocation system for power grid operation and maintenance, characterized in that, It includes an operation and maintenance scheduling platform, a database construction module, a three-layer coupled association model, a demand prediction module, a configuration scheme matching module, and an adaptive optimization module; The database construction module is used to collect multi-source heterogeneous data from the target area, preprocess the collected multi-source heterogeneous data, including data cleaning and standardization, build a multi-dimensional data source library, and send it to the operation and maintenance scheduling platform for storage. The three-layer coupling association model is used to extract equipment health degradation features and environment-grid coupling features respectively. With grid equipment as nodes and electrical connection relationships as edges, a grid topology graph is constructed, and a three-layer coupling graph structure of equipment-environment-space is also constructed. The demand forecasting module is used to retrieve the demand forecasting model based on the three-layer coupled spatiotemporal correlation graph, output three-dimensional forecasting results, including the location of high-risk demand, demand type and resource demand, and generate an operation and maintenance demand heat map and output operation and maintenance demand early warning events. The configuration scheme matching module is used to build a capability profile of early warning requirements based on early warning events of operation and maintenance requirements, and at the same time collect operation and maintenance resource information across the entire domain and build a resource capability profile. Combined with a preset dual-profile fusion matching model, it outputs a preliminary resource configuration scheme. The adaptive optimization module performs adaptive optimization on the initial configuration scheme, taking into account the dynamic status of resources and changes in operating conditions, while resolving resource conflicts and outputting the optimal resource configuration scheme.

2. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 1, characterized in that, The analysis process for the health degradation characteristics of the extraction equipment is as follows: Based on equipment status data from a multidimensional data source library, equipment health degradation features are extracted, including electrical, chemical, thermal, and mechanical features. These features are then used as input layers to a preset equipment maintenance requirement prediction model. The model outputs the probability and type of equipment failure within a preset future time window, and the failure type corresponding to the maximum failure probability is set as the final predicted failure type. Retrieve the preset fault type-operation resource type association table, match the final predicted fault type with the preset fault type-operation resource type association table, and output the matched operation resource type.

3. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 2, characterized in that, The analysis process for extracting the environment-power grid coupling features is as follows: Collect historical meteorological event data and corresponding equipment failure records, classify them by meteorological type and by equipment type, and statistically analyze the failure probability of each type of equipment under various meteorological conditions to form a vulnerability matrix M{m×n}, where m is the number of meteorological types and n is the number of equipment types; Extract spatial features from meteorological radar images to identify the direction of movement and the range of impact of severe convective weather; Based on the power grid topology, a directed graph of fault propagation is established. The probability P(j,i) of device j being affected after device i fails is calculated. Combined with the area affected by meteorological disasters, the cascade fault risk value R = ΣP(i)×P(j,i) of devices in the area is calculated. When the cascade fault risk value exceeds a preset threshold, it is marked as a multiple fault risk point. Spatiotemporal correlation graph construction: Using power grid equipment as nodes and electrical connection relationships as edges, a power grid topology graph is constructed. The predicted failure probability of equipment, meteorological influence factors, and historical work order density are used as node attributes and embedded into the power grid topology graph. By introducing geographic information system coordinates and using spatial proximity as additional edge weights, a three-layer coupled graph structure of device-environment-space is constructed.

4. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 1, characterized in that, The analysis process of the demand forecasting module is as follows: Based on the three-layer coupled diagram structure of equipment-environment-space, the demand forecasting model is retrieved and the three-dimensional forecasting results are output, including the high-risk demand locations, demand types and resource requirements identified by equipment IDs. The target area is divided into equal-scale grids, and the failure probability, equipment importance weight, and user sensitivity weight are comprehensively calculated and predicted for each grid to form a comprehensive demand intensity index. The demand intensity index is mapped to color depth: demand intensity index < E1, low intensity, corresponding to light blue; E1 ≤ demand intensity index < E2, medium intensity, corresponding to yellow; E2 ≤ demand intensity index < E3, high intensity, corresponding to orange; demand intensity index ≥ E3, very high intensity, corresponding to dark red. Among them, 0 < E1 < E2 < E3 < 100. A dynamically updated heat map of operation and maintenance demand is generated and displayed in layers according to the time dimension. When the comprehensive demand intensity index is extremely high, an operation and maintenance demand warning event will be automatically generated.

5. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 1, characterized in that, The analysis process for the early warning demand capability profile and resource capability profile is as follows: Based on the early warning events of operation and maintenance needs, collect power grid operation and maintenance early warning need information, and construct an early warning need capability profile based on the collected power grid operation and maintenance early warning need information; The system collects basic information, capability information, and dynamic status information of various operation and maintenance resources in real time, builds a global operation and maintenance resource pool, and creates a unique resource capability profile for each type of resource. The resource capability profiles specifically include: personnel resource profiles, vehicle resource profiles, and spare parts resource profiles.

6. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 5, characterized in that, Based on demand capability profiles and resource capability profiles, a pre-defined dual-profile fusion matching model is invoked. Following the principles of capability similarity priority, priority priority, and distance priority, accurate matching of demands and resources is achieved. The specific implementation process is as follows: Input to the preset dual-profile fusion matching model: Simultaneously input the completed requirement capability profile and resource capability profile into the preset dual-profile fusion matching model; Model fusion computation: The model automatically completes two core calculations—priority determination and capability similarity calculation; Directly output preliminary solution: Based on the fusion calculation results, the model automatically completes the priority matching and hierarchical matching of high-priority requirements, and directly outputs a complete preliminary resource configuration solution including personnel, vehicles and spare parts.

7. The demand perception and resource collaborative allocation system for power grid operation and maintenance according to claim 1, characterized in that, The analysis process for the optimal resource allocation scheme is as follows: Real-time monitoring of resource status to determine if any resource is abnormal. If a resource is abnormal, the system automatically retrieves the second most similar available resource from the resource pool, replaces the abnormal resource, updates the configuration scheme, and simultaneously determines if the operating conditions have changed. If the operating conditions have changed, the system triggers a resource configuration adjustment command. Further conflict resolution measures were implemented for the initial configuration scheme: when multiple priority requests compete for the same resource, the conflict was resolved by prioritizing the use of high-priority requests and replacing low-priority requests with alternative resources. When requests of the same priority compete for the same resource, the more urgent and similar requests were prioritized based on the time limit and capability similarity, and the remaining requests were re-matched with alternative resources. Cost optimization: Under the premise of meeting the demand capacity requirements, a weighted fusion is performed based on the distance of resources and scheduling costs. The matching optimization coefficient is calculated based on the normalized distance × preset distance ratio coefficient + scheduling cost × preset cost ratio coefficient. The resource corresponding to the minimum value of the matching optimization coefficient is set as the target resource, and the optimal resource allocation scheme is finally output.

Images