A multi-source fusion pumped storage power station underground chamber safety management and control method and system

Abstract

The application discloses a multi-source fusion pumped storage power station underground cavern safety management and control method and system, and relates to the technical field of digital and intelligent safety management and control of water conservancy and hydropower engineering. The application creates a fusion model with "BIM-UWB-personnel" dynamic data correlation as the core, realizes real-time and accurate mapping of personnel positions and engineering semantics through a four-dimensional dynamic correlation matrix; based on a dynamic risk assessment and electronic fence generation method of BIM-track coupling, an optimized LEC algorithm is used to calculate individual dynamic risk values, and a three-dimensional electronic fence updated with construction progress is automatically generated; an intelligent emergency escape guidance method based on real-time personnel positions and BIM structure is defined, a time-varying weighted graph is constructed, and an optimal evacuation route is planned. The application realizes the fusion of UWB high-precision positioning and BIM+GIS dynamic model, solves the problems of disconnection of personnel positioning and risk control, static electronic fence and low emergency evacuation efficiency in underground caverns.

Description

Technical Field

[0001] This invention belongs to the field of digital and intelligent safety management and control technology for water conservancy and hydropower projects, specifically relating to a multi-source integrated method and system for safety management and control of underground caverns in pumped storage power stations. Background Technology

[0002] The underground cavern complex of a pumped-storage power station, including underground powerhouses, main transformer tunnels, water diversion tunnels, and tailrace tunnels, is the core area for construction safety management. This type of workspace is characterized by confined working areas, dense personnel, overlapping construction procedures, and complex geological conditions, resulting in significantly higher safety risks than surface engineering projects. Currently, there are several technical deficiencies in the field of underground cavern construction safety management.

[0003] The tunnel construction environment is inherently highly dangerous, with harsh and complex conditions, unstable geological conditions, and multiple processes involved in excavation, including blasting, shotcreting, and support. These processes often take place before the tunnel has been stabilized and reinforced. Internal support conditions are limited, making it difficult to quickly locate personnel in the event of an accident, potentially missing the optimal window for search and rescue. Furthermore, the construction site involves numerous trades, high personnel mobility, overlapping operations of heavy machinery and personnel, complex pipeline layouts, and collaboration among multiple construction units, leading to complex personnel affiliations and posing practical difficulties for on-site inspections and supervision.

[0004] In terms of personnel positioning technology, most current engineering projects still rely on RFID-based personnel management systems. These systems only install turnstiles at work entrances and manage attendance by identifying RFID cards carried by personnel. However, this approach has fundamental limitations in cavern personnel positioning. The system itself lacks real-time positioning capabilities, making it impossible to track personnel movements within the site after entering the turnstiles. Attendance tracking can only confirm card swipes, not verify actual arrival at the designated work area. Safety alerts cannot determine entry into hazardous areas based on real-time personnel locations. Management feedback struggles to accurately grasp the on-site work status, resulting in low overall management efficiency. More critically, traditional positioning data is isolated and cannot be integrated with dynamically changing building information models. The cavern outline and layered structure continuously change with excavation, making it impossible to accurately map personnel positions in the 3D model. This makes it difficult for managers to intuitively understand the spatial correspondence between personnel distribution and cavern structure.

[0005] In terms of electronic fence technology, traditional solutions mostly use preset fixed areas, which cannot be automatically adjusted according to the progress of layered excavation of underground caverns, and are also difficult to respond to dynamic risk source changes such as fault zones revealed by geological forecasts, easily creating blind spots in control. As the excavation face advances forward, the original fenced area loses its control effectiveness, while newly formed risk areas lack corresponding control measures to cover them.

[0006] Emergency evacuation also has technical shortcomings. The underground caverns have complex structures and numerous passages. In emergencies such as landslides, sudden water inrushes, and fires, managers cannot quickly determine the precise location of people and cannot provide optimal evacuation route guidance for trapped personnel, directly threatening their lives.

[0007] Among existing related technologies, patent CN114529113A discloses a positioning and monitoring method that collects attribute information of monitored objects for statistical analysis and display. However, it is not integrated with building information models (BIM), and therefore cannot intuitively display the spatial relationship between personnel and cavern structures in a 3D engineering scene. Furthermore, the electronic fence is a static preset form, risk assessment relies on human experience, and it does not provide intelligent escape guidance in emergency situations. Patent CN113630866A focuses on optimizing ultra-wideband basic positioning algorithms, completing positioning calculations locally on mobile devices to protect privacy. However, the positioning data is not associated with a BIM, making it impossible to achieve accurate mapping of personnel positions within the engineering structure, and it does not address a dynamic update mechanism for the electronic fence. Patent CN117750300A discloses a fifth-generation mobile communication-integrated ultra-wideband wireless positioning method that adaptively switches between time-of-flight and time-of-arrival algorithms based on on-site environmental parameters. However, it also does not involve the integrated application of BIM and geographic information systems, and it does not achieve dynamic updates of the electronic fence as construction progresses. None of the aforementioned existing technologies have achieved deep integration of ultra-wideband high-precision positioning with building information modeling and geographic information system dynamic models, and have failed to fundamentally solve the technical problems of disconnect between personnel positioning and risk control in underground caverns, static electronic fences, and low emergency evacuation efficiency. Summary of the Invention

[0008] To address the aforementioned technical problems in the existing technology, this invention provides a multi-source integrated method and system for safety management of underground caverns in pumped storage power stations. This system solves the problems of low efficiency, poor accuracy, and delayed response in the construction safety management of underground caverns in pumped storage power stations, such as the disconnect between personnel positioning and BIM models in underground caverns, the inability to dynamically update static electronic fences with construction, the reliance on manual experience for risk assessment and the lack of individualization, and the lack of intelligent guidance for emergency evacuation.

[0009] To achieve the above objectives, the technical solution of the present invention is as follows: Firstly, a multi-source integrated method for safety management of underground caverns in pumped-storage power stations includes: S1. Create a fusion model with BIM-UWB-personnel dynamic data association as the core, construct a BIM+GIS dynamic three-dimensional scene synchronized with the construction progress, and use a hierarchical semantic grid data structure to store the spatial semantic information of the underground cavern. The three-dimensional coordinates and timestamps of personnel are obtained in real time through the UWB positioning system; a four-dimensional dynamic association matrix M = {ID, P, S, A} is constructed to dynamically bind the real-time coordinates of personnel with BIM semantic information and personnel attributes; S2. Based on BIM-trajectory coupling, dynamic risk assessment and electronic fence generation are performed. Construction progress and geological forecast data in the BIM model are analyzed in real time to automatically identify dynamic risk sources, generate a 3D electronic fence described by a spatiotemporal four-tuple, and define access rules using an attribute-based access control model; a BIM-trajectory coupling risk assessment model is constructed. By integrating UWB trajectory data with BIM structural parameters, and applying the optimization formula... Calculate the optimal value for risk exposure frequency according to the formula. Calculate the optimal value for risk severity and generate a tiered early warning system; S3. Based on real-time personnel location and BIM structure, intelligent emergency escape guidance is provided. In emergency situations, a time-varying weighted directed graph is constructed. edge weight Determined by real-time statistics of personnel density based on UWB positioning data; An improved Dijkstra algorithm is used to plan the optimal evacuation route on a time-varying map, with each person's real-time location as the starting point and the safety exit as the ending point; the route is then pushed to the 3D scene, management personnel terminals, and personnel positioning tag terminals.

[0010] Furthermore, the hierarchical semantic grid data structure in step S1 divides the underground space into several voxel units, each unit being associated with the name of its respective cavern, construction zone, risk level, and structural parameters.

[0011] Furthermore, the BIM structure adaptive weight coefficients α, β, and χ in step S2 are dynamically calculated from the span and height parameters of the cavity in the BIM model.

[0012] Furthermore, the dynamic electronic fence in step S2 is automatically regenerated as the BIM model is updated.

[0013] Furthermore, in step S3, the time-varying weighted graph contains personnel density in its edge weights. Using UWB positioning data as input, the number of people per unit area is calculated in real time.

[0014] Furthermore, the update response time of the three-dimensional electronic fence is ≤3 seconds; the generation time of the optimal evacuation route is ≤2 seconds; and the end-to-end delay of the route being pushed to the personnel positioning tag terminal is ≤500ms.

[0015] Secondly, a multi-source integrated safety management and control system for underground caverns of pumped-storage power stations, used to implement the multi-source integrated safety management and control method for underground caverns of pumped-storage power stations described in any of the preceding claims, characterized in that it includes: The BIM+GIS integration platform module is used to build and manage hierarchical semantic grid dynamic scenes. The UWB positioning integration module is used to receive, cache, and synchronize personnel positioning data. The dynamic association matrix construction module is used to create and maintain a four-dimensional dynamic association matrix M; The dynamic fence generation module is used to automatically generate and update 3D electronic fences according to the BIM construction progress. The risk assessment engine module is used to execute the BT-RAM model and calculate the individual dynamic risk value. The emergency escape simulation module is used to construct a time-varying weighted graph and plan the optimal evacuation route; The early warning and response module is used to generate tiered early warning information and push it to the terminal.

[0016] Thirdly, a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the multi-source fusion method for safety management of underground caverns in pumped-storage power stations as described in any of the preceding claims.

[0017] The beneficial effects of this invention are as follows: (1) Deep integration and complete state knowledge: Through a four-dimensional dynamic correlation matrix, the holographic, real-time and dynamic integration of personnel location, status and engineering structure and environment was realized for the first time in the field of pumped storage underground cavern construction.

[0018] (2) Dynamic defense and precise early warning: Through the dynamic risk assessment model, the security control is upgraded from static preset to dynamic individual quantitative early warning. The fence update and risk assessment are fully automated, and the electronic fence response update can be within 3 seconds.

[0019] (3) Intelligent guidance and efficient escape: Through the dynamic path planning algorithm based on time-varying weighted graph, the optimal evacuation guidance for each person in an emergency is realized. The static BIM model is transformed into a dynamic escape route network with real-time response. The path planning time is ≤2 seconds and the terminal push delay is ≤500ms. Personalized optimal path planning and active guidance are provided for each trapped person, which greatly improves the efficiency and success rate of emergency evacuation in complex underground spaces and significantly shortens the emergency response and evacuation time. Attached Figure Description

[0020] Figure 1 A flowchart of the multi-source fusion-based safety management method for underground caverns of pumped-storage power stations provided by the present invention; Figure 2 A flowchart for dynamic risk assessment and electronic fence generation methods; Figure 3 A flowchart for intelligent emergency escape guidance based on real-time personnel location and BIM structure; Figure 4 This is a schematic diagram of a four-dimensional dynamic association model. Detailed Implementation

[0021] The technical solution of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0022] It should be noted that, unless otherwise specifically stated, the relative arrangement and numerical expressions of the components and steps described in these embodiments should not be construed as limiting the scope of the invention.

[0023] The following description of exemplary embodiments is merely illustrative and is not intended to limit the invention or its application or use in any way. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail herein, but where applicable, such techniques, methods, and apparatus should be considered part of this specification.

[0024] Terminology Explanation: BIM: Building Information Modeling refers to a model that uses digital technology to digitally represent the physical and functional characteristics of a building project. It includes multi-dimensional data such as geometric information, material properties, construction information, and operation and maintenance information of building components, and can be used for the design, construction, management, and operation and maintenance of a building throughout its entire life cycle. GIS: Geographic Information System refers to a technical system that, with the support of computer hardware and software, collects, stores, manages, analyzes, displays, and outputs geographic data with spatial attributes. It is capable of processing and analyzing the correspondence between geographic spatial location and attribute information. UWB (Ultra-Wideband) features large bandwidth, low power consumption, strong penetration, and high positioning accuracy (up to centimeter level), and can be used for indoor high-precision positioning, people / object tracking, wireless communication, and other scenarios.

[0025] Example 1 The method described in this embodiment includes three steps: fusion model construction, dynamic risk assessment and electronic fence generation, and intelligent emergency escape guidance. Each step is executed sequentially to form a closed loop; specifically, it includes: S1. Create a fusion model centered on Building Information Modeling (BIM), Ultra-Wideband Positioning (UWB) and dynamic data association of personnel information. Based on the BIM and Geographic Information System (GIS) integration platform, construct a dynamic 3D scene of the underground cavern complex that is updated synchronously with the construction progress. The scene includes the spatial topology of the cavern's layered structure, passage layout, exit locations, and risk areas.

[0026] The model employs a hierarchical semantic grid data structure, dividing the underground space into several voxel units with engineering semantics. Each unit is associated with structural parameter attributes such as the name of its associated cavern, construction zone, risk level, span, height, and slope. Using the deployed ultra-wideband positioning base station network and an improved Taylor series least squares positioning algorithm, the model calculates the 3D coordinates and timestamps of personnel carrying positioning tags in real time, forming the raw positioning data stream.

[0027] Construct a four-dimensional dynamic association matrix M = {ID, P, S, A}, where ID is the unique identifier of the personnel, P is the ultra-wideband real-time positioning coordinates, S is the semantic information quickly retrieved from the hierarchical semantic grid model through a spatial indexing algorithm, including the region, risk level, structural parameters, and distance to the nearest exit, and A is the personnel attribute vector, including identity, qualifications, health status, and physical fitness level.

[0028] The update frequency of this correlation matrix is ​​synchronized with the ultra-wideband positioning refresh rate, with a typical value of no less than 1 Hz. The system uses a bidirectional hash mapping table to achieve real-time conversion from coordinates to semantics, ensuring that the location of any person at any time can be accurately bound to the semantic environment of the building information model engineering in which they are located.

[0029] S2, see reference Figure 2 The system performs dynamic risk assessment and generates electronic fences based on the coupling of Building Information Modeling (BIM) and trajectory. It also analyzes the construction schedule and advanced geological forecast data associated with the BIM in real time.

[0030] Define dynamic identification rules for risk sources. When the construction status of a certain area in the building information model changes to excavation or blasting operations, the system automatically extracts the spatial outline of the area and generates a first-level dynamic fence. When the geological forecast data is updated to show a fault fracture zone, the system automatically expands the buffer distance outward and generates a second-level dynamic fence.

[0031] The mathematical model of the fence is defined as a spatiotemporal quadruple F = {region, time window, access rule, risk level}, where the region is a closed surface in the building information model space, the access rule adopts an attribute-based access control model, and the fence is automatically regenerated as the building information model is updated. After the dynamic fence is defined, the system initiates an individual risk assessment for each person entering or approaching the fence, using a risk assessment model coupled with the building information model and trajectory to structurally optimize the traditional job condition hazard assessment method.

[0032] The formula for calculating the optimal value of risk exposure frequency is: E'=E_base×(1+α×t / T+β×f+χ×Σ(1 / d_k)) t: The cumulative time a person spends in the risk area (obtained by continuous integration of the UWB trajectory); T: Preset exposure duration threshold (determined by historical accident statistics for this area in the BIM model); f: Frequency of entering the risk zone (obtained from UWB trajectory entry and exit counts); d_k: Spatial distance between personnel and the kth hazard source (such as blasting point, fault zone); α, β, χ: Adaptive weighting coefficients for BIM structure, whose values ​​are dynamically calculated from the engineering structural parameters in the BIM model. α = α0 × (span / span_ref) × (height / height_ref), meaning that the larger the span and the higher the height of the cavern, the greater the weight.

[0033] The formula for calculating the optimal value of risk severity is: C'=C_base×γ Where γ = γ0 × (1 + δ × (rock_class)) × (1 + ε × (water_inflow)), and γ is the amplification factor based on geological parameters such as surrounding rock grade and water inflow in BIM.

[0034] Accident probability L: The value used in the traditional LEC is adopted, but it can be dynamically fine-tuned according to the support status in BIM.

[0035] The final formula for calculating the individual dynamic risk value is: R'=L×E'×C' The system automatically classifies warning levels based on risk values ​​and generates warning information that includes personnel identification, location, risk level, and recommended measures.

[0036] Preferably, the technical path for updating the 3D electronic fence within 3 seconds is as follows: Each component in the BIM model has an R-tree spatial index pre-established; when the model is updated, a bounding box intersection query is performed... The system can quickly locate the affected area within a short timeframe; it can expand the boundary of the changed area outward by a specified radius and generate a new fence closed surface using a planar scanning algorithm, with a calculation time of ≤200ms; the admission rule check and geometric calculation are performed in parallel, with a multi-threaded evaluation time of ≤100ms; through the comprehensive application of the above technical paths, the end-to-end time is ensured to be stably controlled within 3 seconds.

[0037] S3, see reference Figure 3The system provides intelligent emergency escape guidance based on real-time personnel location and building information model (BIM) structure. Upon receiving an emergency command or distress alarm, the system immediately initiates the emergency response process, utilizing a four-dimensional dynamic correlation matrix to construct a time-varying weighted directed graph. .

[0038] The node set extracts the spatial coordinates of all passageway intersections, stairwells, and exit locations from the building information model; the edge set is the passable path between adjacent nodes, and its passability is determined by geometric parameters such as passageway width and slope, as well as real-time status such as door opening / closing and collapse. The time-varying weight function is defined as the passage cost of each edge, and is defined as follows:

[0039] in, Let be the physical length of the edge. The population density along this edge was obtained from ultra-wideband positioning data. This is the obstacle coefficient, which takes a larger value when obstacles are present, and zero otherwise. These are the weighting coefficients.

[0040] Multi-objective dynamic optimal path planning: Starting from the real-time location of each trapped person (obtained from the M matrix) and ending at all safe exits, a time-varying weighted directed graph is used. On running the improved Algorithm. This algorithm recalculates the current time during each relaxation operation. Edge weights below This reflects changes in crowd density over time (because crowd movement alters congestion). The algorithm outputs the optimal path sequence for each person. And the estimated travel time.

[0041] Multi-channel differentiated push: The system will push the planned optimal route through the following methods: In the 3D BIM+GIS scene, it is displayed as a superimposed highlight dynamic line; it is pushed to the mobile terminal (large screen or mobile phone) of the management personnel for global command; if the UWB positioning tag has display or beep function, the path guidance is simplified and coded and pushed to the tag terminal.

[0042] Transforming static BIM models into real-time responsive dynamic escape routes provides each trapped individual with personalized optimal path planning and proactive guidance, greatly improving the efficiency and success rate of emergency evacuation from complex underground spaces.

[0043] Preferably, the technical approach for generating the optimal evacuation route with a time ≤2 seconds and a terminal push delay ≤500ms is as follows: The static road network in the BIM model is pre-extracted and stored as an adjacency list, eliminating the need to re-parse the BIM file in emergencies, with a loading time ≤0.1 seconds; UWB positioning data is aggregated by channel grid, and a kernel density estimation algorithm is used to generate real-time personnel density distribution within 0.2 seconds; an independent thread is allocated for parallel computation of the path planning task for each trapped person; with a 1-second cycle, only local paths in density-changing areas are recalculated to avoid recalculating the entire map. After the optimal route is generated, it is pushed to the management personnel's large screen, mobile terminals, and personnel tag terminals via a dedicated network, with an end-to-end transmission delay ≤500ms.

[0044] Example 2 See Figure 1 The system described in this embodiment includes five layers: hardware infrastructure layer, data resource layer, technology platform layer, business application layer, and user decision layer. Each layer is connected sequentially from bottom to top to form a closed-loop control architecture.

[0045] The hardware infrastructure layer forms the foundation for the system's data acquisition, including ultra-wideband base stations, positioning tags, video surveillance equipment, and network and computing storage facilities. Ultra-wideband base stations are deployed in a grid pattern throughout the underground caverns; positioning tags are carried by construction personnel; video surveillance equipment is deployed in key passageways and work areas; and network and computing storage facilities provide data transmission and computational support.

[0046] The data resource layer aggregates various types of basic data, including Building Information Modeling (BIM) data, Geographic Information System (GIS) data, Ultra-Wideband (UWB) positioning data, and video surveillance data. BIM data contains 3D structural information of underground cavern complexes and construction progress data; GIS data provides geographic coordinate benchmarks for surface and underground spaces; UWB positioning data provides real-time personnel coordinate streams; and video surveillance data provides supplementary on-site visualization.

[0047] The technology platform layer provides the core computing engines, including a Building Information Modeling (BIM) and Geographic Information System (GIS) engine, a spatial analysis engine, a risk assessment engine, and a path planning engine. The BIM and GIS engine is responsible for the construction and updating of dynamic 3D scenes, the spatial analysis engine performs coordinate-to-semantic spatial index queries, the risk assessment engine runs the risk assessment model, and the path planning engine performs shortest path algorithm calculations.

[0048] The business application layer integrates three core functional modules, corresponding to three innovative methods. The first module is the dynamic data association model module, which uses the dynamic data association of Building Information Modeling (BIM), ultra-wideband positioning, and personnel information as its core to construct and maintain a four-dimensional dynamic association matrix. The second module is the dynamic risk assessment and electronic fence module, which, based on the coupling mechanism between BIM and trajectory, realizes dynamic electronic fence generation and individual risk assessment. The third module is the intelligent emergency escape guidance module, which, based on the real-time location of personnel and the structure of the BIM model, constructs a time-varying weighted graph and plans the optimal evacuation route.

[0049] The user decision-making process is tailored to different user groups, including the large screen in the conference center, mobile terminals for managers, and personnel tag terminals. The large screen in the conference center displays the overall 3D scene and summarized data, the mobile terminals for managers receive early warning information and conduct dispatch and command, and the personnel tag terminals provide distress triggering and simple escape guidance.

[0050] The specific composition of each module in the system is as follows: the Building Information Modeling (BIM) plus Geographic Information System (GIS) integration platform module is used to construct and manage hierarchical semantic grid dynamic scenes; the ultra-wideband positioning integration module is used to receive, cache, and synchronize personnel positioning data; the dynamic association matrix construction module is used to create and maintain a four-dimensional dynamic association matrix; the dynamic fence generation module is used to automatically generate and update three-dimensional electronic fences based on the construction progress of the BIM model; the risk assessment engine module is used to execute the BIM and trajectory coupled risk assessment model to calculate individual dynamic risk values; the emergency escape simulation module is used to construct a time-varying weighted graph and plan the optimal evacuation route; and the early warning and response module is used to generate graded early warning information and push it to the terminal.

[0051] Example 3 To make the objectives, technical solutions, and advantages of this invention clearer, the following detailed description, using the layered excavation construction of the underground powerhouse of the Wuhai pumped storage power station as a specific embodiment, is provided in conjunction with the accompanying drawings. In this embodiment, hardware facilities such as UWB positioning base stations, positioning tags, and video surveillance have been procured and deployed according to design requirements (specific parameters: one-dimensional positioning base station spacing in the tunnel ≤ 150m, two-dimensional positioning base station spacing in the powerhouse ≤ 30m, positioning accuracy 0.3~0.5m; tags have functions such as SOS button, low battery alarm, and offline cache ≥ 1000 entries). The following focuses on describing the three core innovative methods of this invention implemented based on this hardware foundation. First, based on UAV oblique photography (accuracy 0.05m) and reverse BIM technology, a BIM+GIS dynamic model of the three-layer structure (generator layer, busbar layer, and volute layer) of the underground powerhouse (dimensions 170.0m×25.0m×53.8m) was constructed with an accuracy of ±0.1m. A hierarchical semantic grid (HSM) data structure was adopted to label each voxel unit with attributes such as "floor, construction zone, risk level, span, and height". Then, UWB positioning tags were distributed to 500 construction workers, and the tag IDs were bound to personnel identity, job type, blasting qualification, and health status within the system. The UWB base stations were deployed in a two-dimensional grid (spacing ≤30m, triple coverage), with a positioning refresh rate of 2Hz.

[0052] Key Implementation: The system backend establishes a four-dimensional dynamic association matrix M. Taking worker "Zhang San" (ID: ZS001, blaster) as an example, when he enters the No. 2 unit section of the generator layer, the real-time UWB coordinates (x, y, z) are queried from the HSM model using a spatial indexing algorithm to obtain semantic information S={zone:“generator layer - No. 2 unit”, risk_level:“Level 1 blasting zone”, structure_params:{span:25m,height:8m},nearest_exit_distance:“85m to No. 1 access tunnel”}. Personnel attributes A={identity:“Zhang San”,qualification:“blasting certificate”,energy:“98%”}. This association matrix is ​​updated in real time at a frequency of 2Hz and visualized in the three-dimensional scene (e.g., Figure 4 (As shown), click on the person's icon to view their complete dynamic profile.

[0053] The system analyzes the BIM construction schedule in real time. When the schedule shows that blasting operations are about to take place in the "Generator Layer - Unit 4 Section", the system automatically extracts the spatial outline of the area and generates a Level 1 dynamic fence F={Region: Closed surface of Unit 4 Section, Time_window:14:00-14:30, Access rule:(qualification="blasting permit")∧(energy≥20%), risk_level:1}. The fence is displayed as a red semi-transparent area in the GIS scene.

[0054] The system detected that "Li Si," an ordinary worker without blasting qualifications, approached the fence and lingered for more than one minute. The BT-RAM model was triggered. Calculated from the UWB trajectory: t = 65 seconds, f = 2 times / hour; From the BIM structural parameters: span = 25m, height = 8m, calculate α = 0.2 × (25 / 20) × (8 / 6) = 0.333, take β = 0.15, χ = 0.1; The calculation of E' = E_base × (1 + 0.333 × 65 / 60 + 0.15 × 2 + 0.1 × Σ(1 / distance)) shows a significantly higher result than that of the normal region. Meanwhile, C' = C_base × γ, where γ is taken as 1.5 according to the surrounding rock grade IV in the BIM.

[0055] Finally, R'=150, triggering an orange level 2 warning. The system pushes the warning information to the safety officer's terminal and highlights Li Si's location in the 3D scene.

[0056] Assuming the system detects a fire in the "Generator Floor - Unit 3 Section", it will automatically trigger a Level 1 emergency response. The system will execute: Construct a time-varying weighted map: Extract all passages, staircases, and exits of the floor from the BIM model as nodes and edges; calculate the current personnel density ρ of each passage from UWB positioning data (e.g., main passage ρ=0.3 people / m², secondary passage ρ=0.05 people / m²); set the obstacle coefficient σ=10 on the edge where the fire point is located.

[0057] Calculate the edge weights: the weight of the main channel is W = L × (1 + 2 × 0.3 + 0.5 × 0) = 1.6L, and the weight of the secondary channel is 1.1L. Although the secondary channel is slightly longer, its weight is lower.

[0058] Path planning: Starting from the real-time location of the trapped person "Wang Wu", the improved Dijkstra algorithm is run to output the optimal route as "detour through the secondary passage to the No. 2 traffic tunnel", with an estimated time of 2 minutes and 30 seconds.

[0059] Push notification: The route is highlighted in green in the 3D scene and pushed to Wang Wu's tag terminal (with a beep and simple arrow guidance) and the management personnel's large screen.

[0060] The above specific embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station, characterized in that, include: S1. Create a fusion model with BIM-UWB-personnel dynamic data association as the core, construct a BIM+GIS dynamic three-dimensional scene synchronized with the construction progress, and use a hierarchical semantic grid data structure to store the spatial semantic information of the underground cavern. The three-dimensional coordinates and timestamps of personnel are obtained in real time through the UWB positioning system; a four-dimensional dynamic association matrix M = {ID, P, S, A} is constructed to dynamically bind the real-time coordinates of personnel with BIM semantic information and personnel attributes; S2. Based on BIM-trajectory coupling, dynamic risk assessment and electronic fence generation are carried out. The construction progress and geological forecast data in the BIM model are analyzed in real time, dynamic risk sources are automatically identified, a three-dimensional electronic fence described by a spatiotemporal quadruple is generated, and an attribute-based access control model is used to define access rules. A BIM-trajectory coupled risk assessment model BT-RAM is constructed, which integrates UWB trajectory data and BIM structural parameters. The individual dynamic risk value C'=C_base×γ is calculated according to the optimization formula E'=E_base×(1+α×t / T+β×f+χ×Σ(1 / d_k)) to generate a graded early warning. S3. Intelligent emergency escape guidance based on real-time personnel location and BIM structure. In emergency situations, a time-varying weighted directed graph G(V,E,W(t)) is constructed, where the edge weight W_e(t)=L_e×(1+μ×ρ_e(t)+ν×σ_e), and ρ_e(t) is determined by the personnel density statistically analyzed in real time from UWB positioning data. An improved Dijkstra algorithm is used to plan the optimal evacuation route on a time-varying map, with each person's real-time location as the starting point and the safety exit as the ending point; the route is then pushed to the 3D scene, management personnel terminals, and personnel positioning tag terminals.

2. The method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station according to claim 1, characterized in that, The hierarchical semantic grid data structure in step S1 divides the underground space into several voxel units, each of which is associated with the name of the cavern to which it belongs, the construction zone, the risk level, and the structural parameters.

3. The method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station according to claim 1, characterized in that, The adaptive weighting coefficients α, β, and χ of the BIM structure in step S2 are dynamically calculated from the span and height parameters of the cavity in the BIM model.

4. The method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station according to claim 1, characterized in that, The dynamic electronic fence in step S2 is automatically regenerated as the BIM model is updated.

5. The method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station according to claim 1, characterized in that, The time-varying weighted graph in step S3 has a population density in its edge weights. Using UWB positioning data as input, the number of people per unit area is calculated in real time.

6. The method for safety management and control of underground caverns in a multi-source integrated pumped-storage power station according to claim 1, characterized in that, The update response time of the three-dimensional electronic fence is ≤3 seconds; the generation time of the optimal evacuation route is ≤2 seconds; and the end-to-end delay of the route being pushed to the personnel positioning tag terminal is ≤500ms.

7. A multi-source integrated pumped-storage power station underground cavern safety management system, used to implement the multi-source integrated pumped-storage power station underground cavern safety management method according to any one of claims 1 to 6, characterized in that, include: The BIM+GIS integration platform module is used to build and manage hierarchical semantic grid dynamic scenes. The UWB positioning integration module is used to receive, cache, and synchronize personnel positioning data. The dynamic association matrix construction module is used to create and maintain a four-dimensional dynamic association matrix M; The dynamic fence generation module is used to automatically generate and update 3D electronic fences according to the BIM construction progress. The risk assessment engine module is used to execute the BT-RAM model and calculate the individual dynamic risk value. The emergency escape simulation module is used to construct a time-varying weighted graph and plan the optimal evacuation route; The early warning and response module is used to generate tiered early warning information and push it to the terminal.

Images