Power business data tracking management system and method based on closed-loop management

Abstract

The invention discloses a power business data tracking management system and method based on closed-loop management, and belongs to the technical field of business tracking management. The method comprises the following steps: acquiring meteorological bureau data and power business data in real time, and establishing a three-dimensional elevation model; calling a three-dimensional elevation model, and carrying out risk area identification; acquiring an equipment set in the risk area, calling a path planning algorithm, and transferring to a safety area higher than the elevation of the plant area; establishing a graph structure for a risk-free equipment area; when the riverway water level exceeds the levee crest elevation, layering according to the layer number of the shortest path to the source node, and calculating the permeation amount layer by layer; a differentiation threshold is set based on the power business data, and when the permeation amount of any node exceeds the differentiation threshold, an early warning signal of a corresponding level is triggered; on the basis of the early warning signal, according to the real-time water level increase, calculating processing remaining time; and using an artificial intelligence model to adjust the processing remaining time based on the production operation data in the power business data to obtain the latest processing time.

Description

Technical Field

[0001] This invention relates to the field of business tracking and management technology, specifically to a power business data tracking and management system and method based on closed-loop management. Background Technology

[0002] In recent years, global climate change and frequent extreme weather events have made power companies crucial in typhoon and flood prevention efforts. As the core carriers of power production and transmission, the safe and stable operation of power equipment directly impacts the security of electricity supply and the well-being of the people. Typhoon and flood prevention work requires comprehensive consideration of factors such as meteorological changes, equipment characteristics, and the geographical environment of the power plant area. Traditional flood control management models are no longer sufficient to meet the safety requirements of modern power systems.

[0003] Traditional flood control methods rely heavily on single-point water level monitoring for assessing waterlogging risk, failing to comprehensively simulate the distribution of water within the plant area. The assessment of river backflow risk lacks scientific calculation methods, and early warnings often lag behind actual occurrences. Regarding equipment protection and relocation, existing solutions lack comprehensive consideration of equipment characteristics, production processes, and backup resources. Equipment relocation decisions do not take into account factors such as equipment weight, operating status, and production task priorities, leading to the delayed relocation of critical equipment or the waste of resources due to the relocation of unnecessary equipment. Flood control response times are often determined using fixed patterns, without dynamic adjustments based on real-time flood conditions, equipment operating status, and production needs. In the event of sudden changes in flood conditions or production plans, existing solutions struggle to respond quickly, failing to ensure both equipment safety and the continuity of power production. Summary of the Invention

[0004] The purpose of this invention is to provide a power business data tracking management system and method based on closed-loop management to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Firstly, this application provides a power business data tracking and management method based on closed-loop management, including the following steps: Real-time acquisition of meteorological and power business data, integration with geographic information in typhoon and flood prevention inspection cards, and establishment of a three-dimensional elevation model; When the meteorological bureau data shows a flood season signal, the three-dimensional elevation model is called to determine whether there is water accumulation in the equipment area that exceeds the danger threshold, and risk area identification is performed; the set of equipment in the risk area is obtained, and for equipment that meets the weight requirements and has a mobile base, the path planning algorithm of the factory AGV handling robot is called to transfer it to a safe area that is higher than the factory elevation. For equipment areas without risk, the plant area is divided into regular grids, with each grid serving as a node in a graph structure. Each node establishes directed edges with several adjacent grid nodes. When the river level exceeds the levee crest elevation, the seepage source nodes are initialized, and the layers are divided according to the shortest path from the source node, with the seepage amount calculated layer by layer. Differentiated thresholds are set based on power business data, and when the seepage amount of any node exceeds the differentiated threshold, the corresponding level of early warning signal is triggered. Based on the early warning signal and the real-time water level rise, the remaining processing time is calculated; using an artificial intelligence model, based on the production and operation data in the power business data, adjustments are made to the impact of equipment operation status, interventions are made to the continuity of production processes, and the availability of backup systems are adapted to adjust the remaining processing time and obtain the latest disposal time.

[0006] In conjunction with the first aspect, in the first embodiment of the first aspect of this application, the step of acquiring meteorological bureau data and power business data in real time, connecting with the geographic information in the typhoon and flood prevention inspection card, and establishing a three-dimensional elevation model includes: Meteorological data is obtained from the meteorological bureau's open data platform, and power business data is obtained from the power plant's internal system. The geographic information in the typhoon and flood prevention inspection card is analyzed, and the equipment location is converted into three-dimensional coordinates through GPS positioning. Based on the three-dimensional coordinates, the factory area is divided into plots, and the topographic contour lines in the factory area CAD drawings are used to extract grid elevation values. For areas without drawings, real-time terrain data is obtained through UAV aerial surveying, and grid elevation values ​​are extracted. A three-dimensional elevation model is created according to the actual size of the equipment, and the corresponding grid coordinates are located based on the grid elevation values ​​to mark the equipment foundation elevation.

[0007] In conjunction with the first aspect, in the second embodiment of the first aspect of this application, the step of calling a three-dimensional elevation model when a flood season signal appears in the meteorological bureau data to determine whether there is water accumulation in the equipment area exceeding the danger threshold and to identify the risk area includes: When the meteorological bureau data shows flood season signals, the three-dimensional elevation model is invoked, the expected rainfall and rainfall duration are input, and the surface runoff velocity is calculated in combination with the slope of the factory area; the river water level rise and soil permeability are input to simulate the groundwater infiltration path; for each grid, the water depth is calculated based on the grid elevation value and the expected rainfall; when the river water level exceeds the top of the levee, the diffusion of groundwater into the factory area is simulated from the boundary grid of the factory area according to the soil permeability. Calculate the water depth S in the grid, determine the type of equipment within the grid, and obtain the corresponding danger threshold T. When S≥T, mark the grid as a risk grid; otherwise, mark it as a safe grid.

[0008] In conjunction with the first aspect, in the third embodiment of the first aspect of this application, the step of obtaining the set of devices within the risk area, and for devices that meet the weight requirements and have a movable base, invoking the path planning algorithm of the factory AGV handling robot to transfer them to a safe area higher than the factory elevation, includes: From the risk area composed of several risk grids in the 3D elevation model, all equipment entities are extracted through spatial query. The weight of the equipment is compared with the maximum load of the AGV handling robot. If the weight exceeds the maximum load, it is marked as immovable, and other equipment is added to the equipment set. The base type field in the equipment ledger is checked, and only equipment with movable bases or detachable fixed bases is added to the equipment set. Based on a 3D elevation model, areas meeting the following criteria are searched: elevation and current water depth meet flood control standards; spatial dimensions meet equipment stacking requirements; enclosed spaces are prioritized, followed by open, high-altitude areas, generating a list of safe zones. Ground obstacle data is extracted from the 3D elevation model to generate a 2D grid map, marking prohibited areas and priority passageways. Personnel activity areas are read to avoid densely populated inspection routes, and the real-time positions of other AGV transport robots are synchronized to avoid path conflicts. Transferable equipment is grouped according to its safe zone position in the safe zone list. The required number of AGV transport robots is calculated based on equipment weight and AGV transport robot load capacity. For each group of equipment, the shortest path from the risk area to the safe zone is sorted. When the water depth of a road segment increases beyond a set threshold, the segment is automatically skipped, and the detour path is recalculated. When an AGV transport robot detects an obstacle while traveling, it is fed back to the system via LiDAR, dynamically adjusting its trajectory.

[0009] In conjunction with the first aspect, in the fourth embodiment of the first aspect of this application, for the risk-free equipment area, dividing the factory area into a regular grid, with each grid serving as a node in a graph structure, and each node establishing directed edges with several adjacent grid nodes, includes: The boundary coordinates of the risk-free area are extracted from the 3D elevation model; each grid is numbered and bound with corresponding geographic information; a graph structure is established, in which each grid node contains elevation, soil type, and surface cover, and is associated with equipment information within the grid and infrastructure is labeled; each node establishes directed edges with adjacent grid nodes, specifically nodes on the top, bottom, left, right, and diagonal lines, with the direction defined as: from high-elevation nodes to low-elevation nodes, and when adjacent nodes have the same elevation, they are connected in the default direction from north to south and from west to east; the weight of each edge reflects the water flow resistance and is determined by a combination of elevation difference, soil permeability coefficient, and surface cover.

[0010] In conjunction with the first aspect, in the fifth embodiment of the first aspect of this application, the step of initializing the seepage source node when the river water level exceeds the levee crest elevation, and calculating the seepage amount layer by layer according to the number of shortest path layers from the source node, includes: Select the boundary grids of the plant area adjacent to the river and the grid nodes whose elevations are lower than the current river water level from the graph structure as seepage source nodes; set an initial seepage amount Q for each seepage source node, with the formula: Q=RWL-NE, where RWL is the river water level and NE is the node elevation; A breadth-first search algorithm is used to divide the graph structure into layers. Layer 0 is the set of infiltration source nodes, represented as boundary nodes that are directly in contact with the river channel. Unlayered nodes adjacent to nodes in layers n and (n-1) are sorted by the shortest path layer from the infiltration source nodes, where n is the maximum number of layers in the graph structure. The following nodes are excluded during layering: nodes with an elevation not less than the river channel water level and nodes with no connected path to the source nodes. A layer list is then generated. At layer 0, the initial infiltration amount Q is used directly to output the infiltration amount of each node at layer 0 and mark it as calculated. For each node V at layer n, the infiltration amount is calculated according to the following logic: find the starting node of all directed edges pointing to V; collect the node values ​​of all upstream nodes and perform weighted summation considering edge weights; calculate the attenuation of upstream infiltration amount when flowing through the edge based on the soil type and elevation difference of node V; calculate in the order of the layer list to ensure that the infiltration amount of each layer depends only on the result of the previous layer. After the calculation of each layer is completed, the infiltration amount distribution table of that layer is generated; when a node has a drainage pump and it is running normally, the drainage capacity is deducted based on the infiltration amount distribution table.

[0011] In conjunction with the first aspect, in the sixth embodiment of the first aspect of this application, the step of setting a differentiated threshold based on power business data, and triggering a corresponding level of early warning signal when the penetration of any node exceeds the differentiated threshold, includes: Based on power business data, equipment attributes and business continuity are extracted, and differentiated thresholds are set for critical and non-critical equipment. When a node contains equipment, the importance level and preset threshold of the equipment are retrieved. When a node does not contain equipment, the general protection standards of the region are associated. When the penetration reaches the corresponding threshold, the corresponding level of early warning signal is triggered. Different levels of early warning correspond to differentiated handling measures.

[0012] In conjunction with the first aspect, in the seventh embodiment of the first aspect of this application, the step of calculating the remaining processing time based on the early warning signal and the real-time water level rise includes: The system acquires real-time river water levels and water depth within the factory area, plots water level versus time curves, identifies current rise characteristics, and, in conjunction with meteorological data, predicts future water level trends. Based on these future trends, it extrapolates the estimated time when the water level will reach a dangerous elevation. Finally, it calculates the time interval between the current moment and the estimated time, using this interval as the remaining processing time.

[0013] In conjunction with the first aspect, in the eighth embodiment of the first aspect of this application, the use of an artificial intelligence model, based on production and operation data in power business data, to adjust the impact of equipment operating status, intervene in production process continuity, and adapt the availability of backup systems, and adjust the remaining processing time to obtain the latest handling time, includes: The process involves acquiring production and operation data from power business data, performing data preprocessing and feature engineering; the artificial intelligence model is composed of a combination of a time-series prediction model and a decision tree model; the time-series prediction model is used to analyze equipment parameter change trends and assess fault risks; based on historical flood control scenario data, optimal handling time cases under different equipment states and production tasks are labeled to construct a training dataset; the input features in the time-series prediction model are: equipment health, production task urgency, and backup system availability; model parameters are adjusted through cross-validation to ensure its prediction accuracy under different scenarios. Based on the urgency of the current power generation task and grid dispatch instructions, the decision tree model coordinates production tasks with the dispatch system when the task cannot be interrupted, inserting a handling window during task gaps to extend the remaining time. When the task can be paused, priority is given to flood control measures to shorten the remaining time. The impact of equipment shutdown on other production processes is identified, and the handling sequence is optimized to reduce production interruptions. Combining the health status and switching time of backup equipment, the basic remaining time is maintained when the backup system is available. When debugging is required, the preparation time for the backup system is increased. The artificial intelligence model allocates resources based on the availability of backup equipment. The equipment status, production tasks, and backup system adjustment results are weighted and integrated to generate the latest handling time.

[0014] Secondly, this application provides a power business data tracking and management system based on closed-loop management, including: The data acquisition and elevation model establishment module includes a data acquisition unit and a three-dimensional elevation model establishment unit. The data acquisition unit acquires meteorological bureau data and power business data in real time, while the three-dimensional elevation model establishment unit connects to the geographic information in the typhoon and flood prevention inspection card to establish a three-dimensional elevation model. The area transfer module includes a risk area identification unit and an area transfer unit. When a flood season signal appears in the meteorological bureau data, the risk area identification unit calls a three-dimensional elevation model to determine whether there is water accumulation in the equipment area exceeding the danger threshold, and performs risk area identification. The area transfer unit obtains the set of equipment in the risk area. For equipment that meets the weight requirements and has a movable base, it calls the path planning algorithm of the factory AGV handling robot to transfer it to a safe area higher than the factory elevation. The early warning signal triggering module includes a graph structure construction unit, a permeability calculation unit, and an early warning signal triggering unit. The graph structure construction unit divides the plant area into regular grids for risk-free equipment areas, with each grid serving as a node in the graph structure, and each node establishing directed edges with several adjacent grid nodes. The permeability calculation unit initializes permeability source nodes when the river level exceeds the levee crest elevation, and calculates permeability layer by layer according to the shortest path distance to the source node. The early warning signal triggering unit sets differentiated thresholds based on power business data; when the permeability of any node exceeds the differentiated threshold, it triggers an early warning signal of the corresponding level. The latest handling time calculation module includes a remaining processing time calculation unit and a latest handling time calculation unit. The remaining processing time calculation unit calculates the remaining processing time based on the early warning signal and the real-time water level rise. The latest handling time calculation unit uses an artificial intelligence model to adjust the remaining processing time based on the production and operation data in the power business data, adjust the equipment operation status, intervene in the continuity of the production process, and adapt the standby system availability to obtain the latest handling time.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. Based on a three-dimensional elevation model and graph structure algorithm, this invention can simulate water accumulation and seepage paths under different flood conditions, accurately identify risk areas, and achieve graded early warning through differentiated threshold settings; whether it is the risk of water accumulation in the equipment area or the risk of backflow in the river in the risk-free area, it can be captured in a timely and accurate manner, effectively improving the timeliness and accuracy of early warning.

[0016] 2. For equipment within the risk area, this invention combines equipment attributes and utilizes AGV robots for intelligent transfer, planning the optimal path to improve the efficiency and safety of equipment transfer. At the same time, based on an artificial intelligence model, it comprehensively considers the equipment operating status, production process continuity, and backup system availability, dynamically adjusting the disposal strategy to achieve coordinated optimization of equipment disposal and power production.

[0017] 3. This invention forms a complete closed-loop management system from data collection, risk identification, early warning response to equipment handling and time optimization, which comprehensively improves the intelligent management level and emergency response capability of power companies in typhoon and flood prevention. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the steps of the power business data tracking and management method based on closed-loop management according to the present invention; Figure 2 This is a system architecture diagram of the power business data tracking and management system based on closed-loop management according to the present invention. Detailed Implementation

[0019] 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.

[0020] Example: Figures 1-2 As shown, the present invention provides a technical solution. like Figure 1 As shown, this application provides a power business data tracking and management method based on closed-loop management, including the following steps: Step S100: Acquire meteorological bureau data and power business data in real time, connect with the geographic information in the typhoon and flood prevention inspection card, and establish a three-dimensional elevation model; Specifically, meteorological data is obtained from the meteorological bureau's open data platform, and power business data is obtained from the power plant's internal system. The geographic information in the typhoon and flood prevention inspection card is analyzed, and the equipment location is converted into three-dimensional coordinates through GPS positioning. Based on the three-dimensional coordinates, the factory area is divided into plots, and the topographic contour lines in the factory area CAD drawings are used to extract grid elevation values. For areas without drawings, real-time terrain data is obtained through UAV aerial surveying, and grid elevation values ​​are extracted. A three-dimensional elevation model is created according to the actual size of the equipment, and the corresponding grid coordinates are located based on the grid elevation values ​​to mark the equipment foundation elevation.

[0021] In one specific embodiment, taking a coastal thermal power plant as an example, the plant area covers an area of ​​approximately 1.5 square kilometers, is located near the estuary, and has a generally flat terrain but with local elevation differences of 0.5-2 meters.

[0022] Real-time river water level data was obtained via the provincial meteorological bureau's API. The current water level at the estuary monitoring point is 7.8 meters, with an estimated rainfall of 55 millimeters in the next 24 hours. A blue typhoon warning has been issued (the center is 120 kilometers from the plant area). Operating parameters for 32 key pieces of equipment, including the main transformer and switchgear, were retrieved from the power plant's SCADA system. The equipment log shows that the foundation elevation of the #2 main transformer is 7.2 meters, and the bottom elevation of the No. 3 high-voltage cable trench is 6.5 meters, both located on the east side of the plant area. The typhoon and flood prevention inspection card records the GPS coordinates of the #1 booster station as 30°25′12″N, 121°38′45″E. After converting these coordinates to the plant's three-dimensional coordinate system (X=3526.8m, Y=2814.5m, Z=7.5m), the elevation of its foundation top surface is 7.8 meters.

[0023] The plant area was divided into regular grids with a precision of 5 meters x 5 meters, totaling approximately 60,000 grid cells. For areas with CAD drawings: the main plant area's CAD contour lines show a ground elevation of 7.5-7.8 meters, with the center point of a certain grid extracted to be 7.65 meters. Topographic data for the coal storage yard on the west side of the plant was obtained through UAV aerial surveying (100-meter altitude, 0.1-meter resolution). The measured elevation of grid numbered R25-C38 was 6.3 meters; this area has no fixed equipment but temporary cable supports are present.

[0024] The actual dimensions of the #2 main transformer are 7 meters long, 5 meters wide, and 3.5 meters high, with a foundation elevation of 7.2 meters. In the 3D model, it is located at (X=3650.2m, Y=2780.5m), and its top surface elevation is 7.2 + 3.5 = 10.7 meters, covering 4 grid cells. The #3 cable trench is 40 meters long, 1.8 meters wide, and 1.2 meters deep, with a bottom elevation of 6.5 meters. It is laid along the road on the east side of the plant area, covering 8 grid cells, and is marked as a semi-transparent blue entity in the model for easy observation of water accumulation risk.

[0025] Fifteen landmark points, including the factory gate and cooling tower, were selected. Elevations were measured on-site and compared with model data. The maximum deviation was 5 centimeters. For example, the measured elevation of the cooling tower foundation was 8.12 meters, while the model showed 8.17 meters, meeting the requirements for flood control simulation. Inputting a predicted rainfall of 55 millimeters, the model simulation showed that the water depth in the coal storage yard on the west side of the factory (elevation 6.3 meters) could reach 28 centimeters, while the water depth in the #2 main transformer area (elevation 7.2 meters) was only 10 centimeters. The coal storage yard was automatically marked as a yellow warning area, linking it to the protection requirements of temporary cable supports within this area.

[0026] The 3D model automatically correlates with meteorological data. When the river level rises from 7.8 meters to 8.0 meters, the model updates the infiltration volume in the cable trench area on the east side of the plant in real time. It calculates that the water depth in the area will reach 35 centimeters (exceeding the dangerous threshold of 30 centimeters above the equipment foundation elevation), triggering a red alert and generating equipment relocation recommendations.

[0027] Step S200: When a flood season signal appears in the meteorological bureau data, call the three-dimensional elevation model to determine whether there is water accumulation in the equipment area that exceeds the danger threshold and identify the risk area; obtain the set of equipment in the risk area, and for equipment that meets the weight requirements and has a mobile base, call the path planning algorithm of the factory AGV handling robot to transfer it to a safe area that is higher than the factory elevation. Specifically, when the meteorological bureau's data shows a flood season signal, a three-dimensional elevation model is invoked, and the expected rainfall and duration are input. The surface runoff velocity is calculated by combining the slope of the factory area's terrain. The rise in river level and soil permeability are input to simulate the groundwater infiltration path. For each grid, the water depth is calculated based on the grid elevation value and the expected rainfall. When the river level exceeds the top of the levee, the diffusion of groundwater into the factory area is simulated starting from the boundary grid of the factory area, according to the soil permeability. Calculate the water depth S in the grid, determine the type of equipment within the grid, and obtain the corresponding danger threshold T. When S≥T, mark the grid as a risk grid; otherwise, mark it as a safe grid.

[0028] Furthermore, from the risk area composed of several risk grids in the 3D elevation model, all equipment entities are extracted through spatial query. The weight of the equipment is compared with the maximum load of the AGV handling robot. If the weight exceeds the limit, the equipment is marked as immovable, and other equipment is added to the equipment set. The base type field in the equipment ledger is checked, and only equipment with movable bases or detachable fixed bases is added to the equipment set. Based on a 3D elevation model, areas meeting the following criteria are searched: elevation and current water depth meet flood control standards; spatial dimensions meet equipment stacking requirements; enclosed spaces are prioritized, followed by open, high-altitude areas, generating a list of safe zones. Ground obstacle data is extracted from the 3D elevation model to generate a 2D grid map, marking prohibited areas and priority passageways. Personnel activity areas are read to avoid densely populated inspection routes, and the real-time positions of other AGV transport robots are synchronized to avoid path conflicts. Transferable equipment is grouped according to its safe zone position in the safe zone list. The required number of AGV transport robots is calculated based on equipment weight and AGV transport robot load capacity. For each group of equipment, the shortest path from the risk area to the safe zone is sorted. When the water depth of a road segment increases beyond a set threshold, the segment is automatically skipped, and the detour path is recalculated. When an AGV transport robot detects an obstacle while traveling, it is fed back to the system via LiDAR, dynamically adjusting its trajectory.

[0029] In one specific embodiment, taking a substation in a densely river-networked area as an example, the substation covers an area of ​​approximately 0.8 square kilometers, is adjacent to a river, has a levee crest elevation of 8.0 meters, and an average elevation of 7.5 meters for the main equipment area. The following are experimental data and procedures based on the scenario: In the 3D model, with an input rainfall of 80 mm and an average slope of 1.2% for the plant area, the calculated surface runoff velocity is approximately 0.8 m / h. Runoff convergence is significant in the low-lying area west of the main equipment area. After the river level exceeds the levee crest, infiltration begins from the eastern boundary grid, reaching a depth of 50 meters within the plant area in the first hour, with a grid water depth of 0.3 meters. The grid where the #3 main transformer is located in the main equipment area has an elevation of 7.5 meters, and the expected water depth S = 0.4 meters (0.3 meters of river infiltration + 0.1 meters of surface runoff).

[0030] The #3 main transformer is a critical piece of equipment with an IP65 waterproof rating and a hazard threshold of T = 0.3 meters (foundation elevation + 0.3 meters). Since S = 0.4 meters ≥ T, this grid is marked as a risk grid. There are a total of 5 pieces of equipment within the risk area. The #3 main transformer weighs 20 tons, and the AGV has a maximum load of 5 tons, so it is marked as immovable. The 10kV switchgear weighs 0.8 tons, with a detachable and fixed base, and is included in the portable equipment set. The backup power cabinet weighs 1.2 tons, has a movable base, and is included in the portable equipment set.

[0031] The basement of the administration building (elevation 9.0 meters, space 20×15 meters) is selected as the first choice for the safe zone; the roof platform of the main control room (elevation 8.5 meters, space 10×8 meters) is selected as the second choice for the safe zone.

[0032] The two-dimensional grid map shows that the shortest path from the risk area to the basement of the administration building is through Road No. 3, where the initial water depth is 0.2 meters. The prohibited areas include the dense cable trench area (2 meters wide) and the area within 5 meters around the main transformer heat sink.

[0033] The switch cabinet (0.8 tons) and the backup power cabinet (1.2 tons) are grouped together, and two AGVs (each with a load of 5 tons) are used. The estimated travel time from the risk area to Route 3 and then to the basement of the administration building is 12 minutes.

[0034] After 5 minutes of driving, the water depth on Road 3 increased to 0.35 meters (exceeding the AGV safety threshold of 0.3 meters), and the system automatically detoured to Road 4 (where the water depth was 0.1 meters), increasing the driving time by 5 minutes. During the AGV's journey, the lidar detected temporarily piled sandbags, and the system adjusted its trajectory in real time to avoid the obstacles, increasing the travel time by 2 minutes.

[0035] AGV-01 first moves to the switch cabinet location, scans the QR code to confirm the equipment ID, raises the pallet to secure it, and then travels to the safe zone, with a positioning error of ±3 cm. AGV-02 then simultaneously transfers the backup power cabinet. The entire process takes 19 minutes, 7 minutes longer than initially planned. After the transfer, the measured elevation of the safe zone where the equipment is located is 9.0 meters, the current river level is 8.2 meters, and the water depth is 0.1 meters, all far below the equipment foundation elevation, thus eliminating the risk.

[0036] Step S300: For the risk-free equipment area, the plant area is divided into regular grids, each grid is a node in the graph structure, and each node establishes directed edges with several adjacent grid nodes; when the river water level exceeds the top elevation of the embankment, the seepage source node is initialized, and the layers are divided according to the number of shortest path layers from the source node, and the seepage amount is calculated layer by layer; a differentiated threshold is set based on power business data, and when the seepage amount of any node exceeds the differentiated threshold, the corresponding level of early warning signal is triggered; Specifically, the boundary coordinates of risk-free areas are extracted from the 3D elevation model; each grid is numbered and bound to corresponding geographic information; a graph structure is established, with each grid node containing elevation, soil type, and surface cover, and associated with equipment information within the grid, and infrastructure is labeled; each node establishes directed edges with adjacent grid nodes, specifically nodes on the top, bottom, left, right, and diagonal lines, with the direction defined as: from high-elevation nodes to low-elevation nodes, and when adjacent nodes have the same elevation, they are connected in the default direction of north to south and west to east; the weight of each edge reflects water flow resistance and is determined by a combination of elevation difference, soil permeability coefficient, and surface cover.

[0037] Furthermore, from the graph structure, the boundary grids of the plant area adjacent to the river and the grid nodes with node elevations lower than the current river water level are selected as seepage source nodes; an initial seepage amount Q is set for each seepage source node, with the formula: Q=RWL-NE, where RWL is the river water level and NE is the node elevation. A breadth-first search algorithm is used to divide the graph structure into layers. Layer 0 is the set of infiltration source nodes, represented as boundary nodes that are directly in contact with the river channel. Unlayered nodes adjacent to nodes in layers n and (n-1) are sorted by the shortest path layer from the infiltration source nodes, where n is the maximum number of layers in the graph structure. The following nodes are excluded during layering: nodes with an elevation not less than the river channel water level and nodes with no connected path to the source nodes. A layer list is then generated. At layer 0, the initial infiltration amount Q is used directly to output the infiltration amount of each node at layer 0 and mark it as calculated. For each node V at layer n, the infiltration amount is calculated according to the following logic: find the starting node of all directed edges pointing to V; collect the node values ​​of all upstream nodes and perform weighted summation considering edge weights; calculate the attenuation of upstream infiltration amount when flowing through the edge based on the soil type and elevation difference of node V; calculate in the order of the layer list to ensure that the infiltration amount of each layer depends only on the result of the previous layer. After the calculation of each layer is completed, the infiltration amount distribution table of that layer is generated; when a node has a drainage pump and it is running normally, the drainage capacity is deducted based on the infiltration amount distribution table.

[0038] Furthermore, based on power business data, equipment attributes and business continuity are extracted, and differentiated thresholds are set for critical and non-critical equipment; when a node contains equipment, the importance level and preset threshold of the equipment are retrieved; when a node does not contain equipment, the general protection standards of the region are associated; when the penetration reaches the corresponding threshold, the corresponding level of early warning signal is triggered; and different levels of early warning correspond to differentiated handling measures.

[0039] In one specific embodiment, taking a substation in a riverside industrial park as an example, the risk-free equipment area of ​​this substation is approximately 0.5 square kilometers, the levee crest elevation is 7.5 meters, the current river level is 7.8 meters (0.3 meters above the levee crest), the soil type is sandy soil, and the surface cover is mainly hardened road surface. Specific experimental data are as follows: The risk-free zone is divided into 500 grids of 10m x 10m each, numbered according to the rule R (row) - C (column), such as R10-C15. Nodes R15-C20: Elevation 7.0 meters, soil type sandy soil, hardened surface, non-critical equipment (lighting distribution box) within the grid, foundation elevation 7.2 meters. Nodes R16-C20: Elevation 6.8 meters, soil type sandy soil, grassy surface, no equipment, classified as a normal area.

[0040] Establish directed edges: R15-C20 (7.0 m) → R16-C20 (6.8 m), with the direction from high to low; R15-C19 (7.0 m) and R15-C20 (7.0 m) have the same elevation, and the edge direction is defined as west → east (R15-C19 → R15-C20).

[0041] The eastern boundary grid of the plant area, from R20-C10 to R20-C30, has an elevation of ≤7.0 meters (river water level 7.8 meters), comprising 11 nodes, which form the source node set S. The source node R20-C20 has an elevation of 6.5 meters, and Q = 7.8 - 6.5 = 1.3 meters (RWL = 7.8 meters, NE = 6.5 meters).

[0042] Layer 0: Source node S (R20-C10 to R20-C30); Layer 1: Inner grid adjacent to the source node (e.g., R19-C10 to R19-C30, R20-C9 to R20-C11); Layer 2: R18-C10 to R18-C30, etc., with a maximum number of layers n=5 (penetrating to the central area of ​​the plant).

[0043] Permeability of layer 0: Use Q=1.3 meters directly (e.g., node R20-C20). Calculation of permeability of node R19-C20 in layer 1: The upstream node is R20-C20 (Q=1.3 meters), and the edge weights take into account the permeability coefficient of sand (high) and hardened road surface (low resistance), with an attenuation coefficient of 0.7; permeability = 1.3 × 0.7 = 0.91 meters.

[0044] There is a drainage pump near node R19-C25 on the first layer, with a discharge rate of 0.5 m³ / h. After a 1-hour interval, the infiltration rate is calculated to be 0.91 - 0.5 = 0.41 m³.

[0045] Threshold settings: Critical equipment area (none); Non-critical equipment area (such as lighting distribution box of R15-C20): Threshold T=0.3 meters (equipment foundation elevation 7.2 meters - 7.0 meters + 0.1 meters safety margin); Ordinary area (such as R16-C20): Threshold T=1.0 meter.

[0046] A yellow alert is triggered when the penetration rate at R19-C20 nodes is 0.41 meters or greater than the non-critical equipment threshold of 0.3 meters; a yellow alert is also triggered when the penetration rate at R20-C20 nodes is 1.3 meters or greater than the ordinary area threshold of 1.0 meter.

[0047] After the yellow alert was triggered, a task was sent to the operations and maintenance department: install a waterproof baffle (0.5 meters high) on the lighting distribution box in the R19-C20 area; start the backup drainage pump to increase the drainage capacity of the area to 1.0 m / hour. One hour later, the infiltration rate of the R19-C20 node was recalculated: 0.41 - 1.0 (drainage) + 0.2 (new infiltration) = -0.39 m (a negative value indicates a reduction in water accumulation), and the alert was lifted. Actual inspection revealed that the water depth in the area was 0.1 m, with a deviation of ≤10% from the model calculation.

[0048] Step S400: Based on the early warning signal and the real-time water level rise, calculate the remaining processing time; using an artificial intelligence model, based on the production and operation data in the power business data, adjust the impact of equipment operation status, intervene in the continuity of production process, and adapt the availability of backup systems to adjust the remaining processing time and obtain the latest disposal time.

[0049] Specifically, the system acquires real-time river water levels and water depth within the factory area, plots water level versus time curves, identifies current rise characteristics, and, in conjunction with meteorological data, predicts future water level trends. Based on these future trends, it extrapolates the estimated time when the water level will reach a dangerous elevation. Finally, it calculates the time interval between the current moment and the estimated time, using this interval as the remaining processing time.

[0050] Furthermore, production and operation data from power business data are acquired, and data preprocessing and feature engineering are performed. The artificial intelligence model consists of a combination of a time-series prediction model and a decision tree model. The time-series prediction model is used to analyze the changing trends of equipment parameters and determine the risk of failure. Based on historical flood control scenario data, optimal handling time cases under different equipment states and production tasks are labeled to construct a training dataset. In the time-series prediction model, the input features are: equipment health, production task urgency, and backup system availability. The model parameters are adjusted through cross-validation to ensure its prediction accuracy under different scenarios. Based on the urgency of the current power generation task and grid dispatch instructions, the decision tree model coordinates production tasks with the dispatch system when the task cannot be interrupted, inserting a handling window during task gaps to extend the remaining time. When the task can be paused, priority is given to flood control measures to shorten the remaining time. The impact of equipment shutdown on other production processes is identified, and the handling sequence is optimized to reduce production interruptions. Combining the health status and switching time of backup equipment, the basic remaining time is maintained when the backup system is available. When debugging is required, the preparation time for the backup system is increased. The artificial intelligence model allocates resources based on the availability of backup equipment. The equipment status, production tasks, and backup system adjustment results are weighted and integrated to generate the latest handling time.

[0051] In one specific embodiment, real-time river water levels and water depth within the plant area are acquired, and a water level versus time curve is plotted through continuous monitoring. Initially, the river water level rises by 0.1 meters per hour, and some low-lying areas within the plant begin to accumulate water, with the water depth increasing by 0.05 meters per hour. Based on typhoon path and rainfall data released by the meteorological bureau, it is predicted that within the next 12 hours, due to continuous heavy rainfall, the river water level will continue to rise, with the rate of increase expected to accelerate to 0.2 meters per hour; the water depth within the plant area is expected to increase by 0.1 meters per hour. The hazardous elevation of the power plant's main equipment area is set at 8.5 meters. Currently, the ground elevation of the main equipment area is 8.2 meters, the real-time river water level is 8.0 meters, and the water depth in this area within the plant is 0.1 meters. Based on projected future water level changes, it is estimated that in 6 hours, the water depth inside the plant area will reach 0.7 meters (0.1 + 0.1 × 6). Combined with the rising river level, the total water level will reach 8.7 meters (8.0 + 0.2 × 6 + 0.7), exceeding the danger level of 8.5 meters. Therefore, the remaining processing time is 6 hours from the current moment. The process involves acquiring power plant production and operation data, including real-time operating parameters of the turbine generator units (such as speed, vibration, and temperature), current power generation tasks, and backup power status, and performing data preprocessing and feature engineering.

[0052] The artificial intelligence model consists of a combination of a time-series prediction model and a decision tree model. Using the time-series prediction model to analyze the parameter variation trends of the hydro-generator units, it was found that the vibration value of Unit 1 hydro-generator unit showed a gradual upward trend. Combined with historical data and an equipment health assessment model, it was determined that the unit has a potential failure risk, and the equipment health score has decreased. Currently, the power plant undertakes important regional peak-shaving tasks, and the production tasks are highly urgent; the backup power system is under maintenance, and the availability of the backup system is low. These characteristics are used as inputs, and the parameters of the time-series prediction model are adjusted through cross-validation.

[0053] Using a decision tree model for analysis, the current power generation task is to ensure peak electricity demand in the city, which is an uninterrupted task. Based on the grid dispatch instructions and in coordination with the dispatch system, the decision tree model determines that there is a brief dispatch gap within the next two hours. Therefore, the original 6-hour processing time is extended to 8 hours (6+2) to allow for equipment handling during the task gap.

[0054] Simultaneously, it was identified that immediately shutting down the equipment would lead to a decrease in the power output of the entire power generation system, affecting the stable operation of other units. The decision tree model optimized the handling sequence, prioritizing the protection of auxiliary equipment that was less affected. Considering that the backup power system requires 2 hours of commissioning before it can be put into use, an additional 2 hours of backup system preparation time was added to the remaining processing time.

[0055] Finally, the AI ​​model performs a weighted fusion based on equipment status, production tasks, and adjustments to the backup system. Potential equipment failure risk accounts for 40% of the weight, uninterrupted production tasks account for 30%, and the need for backup system debugging accounts for 30%. After comprehensive calculation, the latest possible response time is determined to be the second hour after the start of the power generation task gap (i.e., the eighth hour from the current moment). This ensures that necessary equipment handling can be completed before the water level reaches a dangerous elevation, while maximizing the continuity of power production.

[0056] like Figure 2 As shown, this application provides a power business data tracking and management system based on closed-loop management, including: The data acquisition and elevation model establishment module includes a data acquisition unit and a three-dimensional elevation model establishment unit. The data acquisition unit acquires meteorological bureau data and power business data in real time, while the three-dimensional elevation model establishment unit connects to the geographic information in the typhoon and flood prevention inspection card to establish a three-dimensional elevation model. The area transfer module includes a risk area identification unit and an area transfer unit. When a flood season signal appears in the meteorological bureau data, the risk area identification unit calls a three-dimensional elevation model to determine whether there is water accumulation in the equipment area exceeding the danger threshold, and performs risk area identification. The area transfer unit obtains the set of equipment in the risk area. For equipment that meets the weight requirements and has a movable base, it calls the path planning algorithm of the factory AGV handling robot to transfer it to a safe area higher than the factory elevation. The early warning signal triggering module includes a graph structure construction unit, a permeability calculation unit, and an early warning signal triggering unit. The graph structure construction unit divides the plant area into regular grids for risk-free equipment areas, with each grid serving as a node in the graph structure, and each node establishing directed edges with several adjacent grid nodes. The permeability calculation unit initializes permeability source nodes when the river level exceeds the levee crest elevation, and calculates permeability layer by layer according to the shortest path distance to the source node. The early warning signal triggering unit sets differentiated thresholds based on power business data; when the permeability of any node exceeds the differentiated threshold, it triggers an early warning signal of the corresponding level. The latest handling time calculation module includes a remaining processing time calculation unit and a latest handling time calculation unit. The remaining processing time calculation unit calculates the remaining processing time based on the early warning signal and the real-time water level rise. The latest handling time calculation unit uses an artificial intelligence model to adjust the remaining processing time based on the production and operation data in the power business data, adjust the equipment operation status, intervene in the continuity of the production process, and adapt the standby system availability to obtain the latest handling time.

[0057] It will be apparent to those skilled in the art that the present 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 its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A power business data tracking and management method based on closed-loop management, characterized in that, Includes the following steps: Real-time acquisition of meteorological and power business data, integration with geographic information in typhoon and flood prevention inspection cards, and establishment of a three-dimensional elevation model; When the meteorological bureau data shows a flood season signal, the three-dimensional elevation model is called to determine whether there is water accumulation in the equipment area that exceeds the danger threshold, and risk area identification is performed; the set of equipment in the risk area is obtained, and for equipment that meets the weight requirements and has a mobile base, the path planning algorithm of the factory AGV handling robot is called to transfer it to a safe area that is higher than the factory elevation. For equipment areas without risk, the plant area is divided into regular grids, with each grid serving as a node in a graph structure. Each node establishes directed edges with several adjacent grid nodes. When the river level exceeds the levee crest elevation, the seepage source nodes are initialized, and the layers are divided according to the shortest path from the source node, with the seepage amount calculated layer by layer. Differentiated thresholds are set based on power business data, and when the seepage amount of any node exceeds the differentiated threshold, the corresponding level of early warning signal is triggered. Based on the early warning signal and the real-time water level rise, the remaining processing time is calculated. Using artificial intelligence models, based on production and operation data in power business data, adjustments are made to the impact of equipment operating status, interventions are made to the continuity of production processes, and the availability of backup systems is adapted to adjust the remaining processing time and obtain the latest handling time.

2. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, The process of acquiring real-time meteorological and power business data, integrating it with geographic information from typhoon and flood prevention inspection cards, and establishing a three-dimensional elevation model includes: Meteorological data is obtained from the meteorological bureau's open data platform, and power business data is obtained from the power plant's internal system. The geographic information in the typhoon and flood prevention inspection card is analyzed, and the equipment location is converted into three-dimensional coordinates through GPS positioning. Based on the three-dimensional coordinates, the factory area is divided into plots, and the topographic contour lines in the factory area CAD drawings are used to extract grid elevation values. For areas without drawings, real-time terrain data is obtained through UAV aerial surveying, and grid elevation values ​​are extracted. A three-dimensional elevation model is created according to the actual size of the equipment, and the corresponding grid coordinates are located based on the grid elevation values ​​to mark the equipment foundation elevation.

3. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, When a flood season signal appears in the meteorological bureau data, a three-dimensional elevation model is invoked to determine whether there is water accumulation in the equipment area exceeding the danger threshold, and risk area identification is performed, including: When the meteorological bureau data shows flood season signals, the three-dimensional elevation model is invoked, the expected rainfall and rainfall duration are input, and the surface runoff velocity is calculated in combination with the slope of the factory area; the river water level rise and soil permeability are input to simulate the groundwater infiltration path; for each grid, the water depth is calculated based on the grid elevation value and the expected rainfall; when the river water level exceeds the top of the levee, the diffusion of groundwater into the factory area is simulated from the boundary grid of the factory area according to the soil permeability. Calculate the water depth S in the grid, determine the type of equipment within the grid, and obtain the corresponding danger threshold T. When S≥T, mark the grid as a risk grid; otherwise, mark it as a safe grid.

4. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, The process of acquiring the set of devices within the risk area, and for devices that meet the weight requirements and have a movable base, involves invoking the path planning algorithm of the factory AGV handling robot to move them to a safe area above the factory elevation, including: From the risk area composed of several risk grids in the 3D elevation model, all equipment entities are extracted through spatial query. The weight of the equipment is compared with the maximum load of the AGV handling robot. If the weight exceeds the maximum load, it is marked as immovable, and other equipment is added to the equipment set. The base type field in the equipment ledger is checked, and only equipment with movable bases or detachable fixed bases is added to the equipment set. Based on a 3D elevation model, areas meeting the following criteria are searched: elevation and current water depth meet flood control standards; spatial dimensions meet equipment stacking requirements; enclosed spaces are prioritized, followed by open, high-altitude areas, generating a list of safe zones. Ground obstacle data is extracted from the 3D elevation model to generate a 2D grid map, marking prohibited areas and priority passageways. Personnel activity areas are read to avoid densely populated inspection routes, and the real-time positions of other AGV transport robots are synchronized to avoid path conflicts. Transferable equipment is grouped according to its safe zone position in the safe zone list. The required number of AGV transport robots is calculated based on equipment weight and AGV transport robot load capacity. For each group of equipment, the shortest path from the risk area to the safe zone is sorted. When the water depth of a road segment increases beyond a set threshold, the segment is automatically skipped, and the detour path is recalculated. When an AGV transport robot detects an obstacle while traveling, it is fed back to the system via LiDAR, dynamically adjusting its trajectory.

5. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, For the risk-free equipment area, the factory area is divided into a regular grid, with each grid cell serving as a node in a graph structure. Each node establishes directed edges with several adjacent grid cells, including: The boundary coordinates of the risk-free area are extracted from the 3D elevation model; each grid is numbered and bound with corresponding geographic information; a graph structure is established, in which each grid node contains elevation, soil type, and surface cover, and is associated with equipment information within the grid and infrastructure is labeled; each node establishes directed edges with adjacent grid nodes, specifically nodes on the top, bottom, left, right, and diagonal lines, with the direction defined as: from high-elevation nodes to low-elevation nodes, and when adjacent nodes have the same elevation, they are connected in the default direction from north to south and from west to east; the weight of each edge reflects the water flow resistance and is determined by a combination of elevation difference, soil permeability coefficient, and surface cover.

6. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, When the river level exceeds the levee crest elevation, the seepage source nodes are initialized, and the seepage is calculated layer by layer according to the number of layers with the shortest path from the source node, including: Select the boundary grids of the plant area adjacent to the river and the grid nodes whose elevations are lower than the current river water level from the graph structure as seepage source nodes; set an initial seepage amount Q for each seepage source node, with the formula: Q=RWL-NE, where RWL is the river water level and NE is the node elevation; A breadth-first search algorithm is used to divide the graph structure into layers. Layer 0 is the set of infiltration source nodes, represented as boundary nodes that are directly in contact with the river channel. Unlayered nodes adjacent to nodes in layers n and (n-1) are sorted by the shortest path layer from the infiltration source nodes, where n is the maximum number of layers in the graph structure. The following nodes are excluded during layering: nodes with an elevation not less than the river channel water level and nodes with no connected path to the source nodes. A layer list is then generated. At layer 0, the initial infiltration amount Q is used directly to output the infiltration amount of each node at layer 0 and mark it as calculated. For each node V at layer n, the infiltration amount is calculated according to the following logic: find the starting node of all directed edges pointing to V; collect the node values ​​of all upstream nodes and perform weighted summation considering edge weights; calculate the attenuation of upstream infiltration amount when flowing through the edge based on the soil type and elevation difference of node V; calculate in the order of the layer list to ensure that the infiltration amount of each layer depends only on the result of the previous layer. After the calculation of each layer is completed, the infiltration amount distribution table of that layer is generated; when a node has a drainage pump and it is running normally, the drainage capacity is deducted based on the infiltration amount distribution table.

7. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, The method involves setting differentiated thresholds based on power business data. When the penetration rate of any node exceeds the differentiated threshold, a corresponding level of early warning signal is triggered, including: Based on power business data, equipment attributes and business continuity are extracted, and differentiated thresholds are set for critical and non-critical equipment. When a node contains equipment, the importance level and preset threshold of the equipment are retrieved. When a node does not contain equipment, the general protection standards of the region are associated. When the penetration reaches the corresponding threshold, the corresponding level of early warning signal is triggered. Different levels of early warning correspond to differentiated handling measures.

8. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, The calculation of the remaining processing time based on the early warning signal and the real-time water level rise includes: The system acquires real-time river water levels and water depth within the factory area, plots water level versus time curves, identifies current rise characteristics, and, in conjunction with meteorological data, predicts future water level trends. Based on these future trends, it extrapolates the estimated time when the water level will reach a dangerous elevation. Finally, it calculates the time interval between the current moment and the estimated time, using this interval as the remaining processing time.

9. The power business data tracking and management method based on closed-loop management according to claim 1, characterized in that, The use of artificial intelligence models, based on production and operation data from power business data, involves adjusting equipment operating status, intervening in production process continuity, and adapting backup system availability to adjust remaining processing time and obtain the latest handling time, including: The process involves acquiring production and operation data from power business data, performing data preprocessing and feature engineering; the artificial intelligence model is composed of a combination of a time-series prediction model and a decision tree model; the time-series prediction model is used to analyze equipment parameter change trends and assess fault risks; based on historical flood control scenario data, optimal handling time cases under different equipment states and production tasks are labeled to construct a training dataset; the input features in the time-series prediction model are: equipment health, production task urgency, and backup system availability; model parameters are adjusted through cross-validation to ensure its prediction accuracy under different scenarios. Based on the urgency of the current power generation task and grid dispatch instructions, the decision tree model coordinates production tasks with the dispatch system when the task cannot be interrupted, inserting a handling window during task gaps to extend the remaining time. When the task can be paused, priority is given to flood control measures to shorten the remaining time. The impact of equipment shutdown on other production processes is identified, and the handling sequence is optimized to reduce production interruptions. Combining the health status and switching time of backup equipment, the basic remaining time is maintained when the backup system is available. When debugging is required, the preparation time for the backup system is increased. The artificial intelligence model allocates resources based on the availability of backup equipment. The equipment status, production tasks, and backup system adjustment results are weighted and integrated to generate the latest handling time.

10. A power business data tracking and management system based on closed-loop management, using the power business data tracking and management method based on closed-loop management as described in any one of claims 1-9, characterized in that, include: The data acquisition and elevation model establishment module includes a data acquisition unit and a three-dimensional elevation model establishment unit. The data acquisition unit acquires meteorological bureau data and power business data in real time, while the three-dimensional elevation model establishment unit connects to the geographic information in the typhoon and flood prevention inspection card to establish a three-dimensional elevation model. The area transfer module includes a risk area identification unit and an area transfer unit. When a flood season signal appears in the meteorological bureau data, the risk area identification unit calls a three-dimensional elevation model to determine whether there is water accumulation in the equipment area exceeding the danger threshold, and performs risk area identification. The area transfer unit obtains the set of equipment in the risk area. For equipment that meets the weight requirements and has a movable base, it calls the path planning algorithm of the factory AGV handling robot to transfer it to a safe area higher than the factory elevation. The early warning signal triggering module includes a graph structure construction unit, a permeability calculation unit, and an early warning signal triggering unit. The graph structure construction unit divides the plant area into regular grids for risk-free equipment areas, with each grid serving as a node in the graph structure, and each node establishing directed edges with several adjacent grid nodes. The permeability calculation unit initializes permeability source nodes when the river level exceeds the levee crest elevation, and calculates permeability layer by layer according to the shortest path distance to the source node. The early warning signal triggering unit sets differentiated thresholds based on power business data; when the permeability of any node exceeds the differentiated threshold, it triggers an early warning signal of the corresponding level. The latest handling time calculation module includes a remaining processing time calculation unit and a latest handling time calculation unit. The remaining processing time calculation unit calculates the remaining processing time based on the early warning signal and the real-time water level rise. The latest handling time calculation unit uses an artificial intelligence model to adjust the remaining processing time based on the production and operation data in the power business data, adjust the equipment operation status, intervene in the continuity of the production process, and adapt the standby system availability to obtain the latest handling time.

Images