A substation three-dimensional patrollable space modeling method and system

Abstract

This invention discloses a method and system for three-dimensional inspectable spatial modeling of substations, belonging to the field of power system inspection technology. The method includes: constructing a three-dimensional scene model and spatially discretizing it to obtain a set of spatial units; constructing a multimodal dynamic knowledge graph, associating entity objects and their semantic attributes with corresponding spatial units, and generating dynamic spatial constraints that act on the corresponding spatial units; obtaining a set of effective observation spatial units based on reachability and effective observation determination, and dividing the observation area based on the spatial units to obtain the area division result; constructing a three-dimensional inspectable spatial situation map on the three-dimensional scene model based on the entity objects and dynamic spatial constraints associated with the area division result, and updating the three-dimensional inspectable spatial situation map based on changes in entity objects or relationships between entity objects. This invention solves the problems in the prior art where inspection path and observation point planning relies on manual experience, has safety blind spots, and cannot be dynamically adjusted according to operating conditions.

Description

Technical Field

[0001] This invention relates to a method and system for three-dimensional inspectable spatial modeling of substations, belonging to the field of power system inspection technology. Background Technology

[0002] In recent years, with the accelerated construction of new power systems, the power industry's demand for unmanned and minimally staffed operation and maintenance models has become increasingly urgent. In the field of substation inspection, current methods mainly include manual inspection, drone inspection, and robotic inspection. Manual inspection is labor-intensive, relying heavily on the experience of personnel to judge equipment status, and its quality and efficiency are significantly affected by subjective factors. Drone and robotic inspections, on the other hand, require pre-planned flight paths or routes to execute inspection tasks according to preset trajectories. Regarding 3D scene modeling, existing technologies utilize laser point clouds, oblique photography, and video reconstruction to construct 3D models of substations, and path planning can be performed based on these models.

[0003] However, existing technologies still have the following shortcomings. First, the planning of inspection paths and observation points heavily relies on human experience and lacks systematic quantitative analysis methods. The deployment of existing intelligent inspection systems mainly depends on human experience, resulting in incomplete coverage of shooting angles. There are many blind spots in areas such as high-altitude structures, equipment tops, and the backs of pipelines, making the site survey work of engineering designers extremely demanding. Robots and drones can only operate according to preset routes or paths and cannot automatically determine the optimal inspection path and observation points based on the actual spatial distribution of equipment and observation needs. Second, path planning is based on static maps and cannot be dynamically adjusted according to operating conditions and work status. The path planning of existing inspection robots is usually based on static maps, which makes it difficult to cope with unmodeled environmental information and real-time changes in equipment status and work range. When temporary work, changes in energized status, or sudden alarms occur in the substation, the preset inspection path cannot automatically adapt to new spatial constraints and lacks the ability to dynamically adjust the path based on real-time spatial information. Third, the 3D model lacks dynamic correlation with the real-time operating status. Most existing 3D models of substations are static geometric models, used only for visualization. Real-time sensor data, manual inspection records, equipment ledgers, etc. are scattered in different business systems. There is a lack of unified spatial positioning standards and data semantic specifications, making it difficult to transform the real-time status of equipment, including energized status, working status, and risk status, into dynamically updatable spatial constraints. As a result, the 3D model only has visualization functions but no computational capabilities, and cannot provide real-time and accurate spatial decision-making basis for the dynamic planning of inspection paths and the adaptive adjustment of observation points. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for three-dimensional inspectable spatial modeling of substations. By constructing a multimodal dynamic knowledge graph that integrates spatial geometric data, business semantic data, multimodal perception data, and real-time status data, and associating entity objects and their semantic attributes with discretized spatial units, spatial constraint rules that dynamically change with equipment and operating status are generated. Based on the joint determination of accessibility and effective observability, the spatial units are divided into regions to solve the problems in the prior art where inspection path and observation point planning rely on manual experience, have safety blind spots, and cannot be dynamically adjusted according to operating conditions.

[0005] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution.

[0006] In a first aspect, the present invention provides a method for modeling a three-dimensional traversable spatial domain of a substation, comprising:

[0007] Acquire spatial geometric data, business semantic data, multimodal sensing data, and real-time status data of the substation;

[0008] Based on the spatial geometric data, a three-dimensional scene model of the substation is constructed and spatially discretized to obtain a set of spatial units;

[0009] A multimodal dynamic knowledge graph is constructed based on the business semantic data, multimodal perception data, and real-time status data. The multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects.

[0010] The entity object and its corresponding semantic attributes are associated with the corresponding spatial unit, and dynamic spatial constraints are generated according to the current state of the entity object and the preset safety procedures and on-site operation rules, and applied to the corresponding spatial unit.

[0011] Based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of reachable spatial units. Based on the set of reachable spatial units, effective observation is determined based on preset field of view constraints and observation parameter conditions corresponding to the inspection target to obtain a set of effective observation spatial units. Then, the observation area is divided based on the spatial units to obtain the area division result.

[0012] Based on the entity objects associated with the region division results and dynamic spatial constraints, a three-dimensional patrolable spatial situation map is constructed on the three-dimensional scene model, and the three-dimensional patrolable spatial situation map is updated based on the changes in entity objects or the relationships between entity objects.

[0013] Furthermore, the spatial geometric data includes at least one of the following: laser point cloud data, oblique photogrammetry model data, and video reconstruction model data;

[0014] The business semantic data includes at least one of the following: equipment ledger data, equipment code data, equipment attribute data, operation mode data, safety procedure data, area division data, on-site operation rules, and inspection task data;

[0015] The multimodal sensing data includes at least two or more of the following: visible light image data, infrared image data, audio data, and text data.

[0016] The real-time status data includes at least one of the following: equipment disconnection / connection status, equipment operation status, energized status, operation status, operation scope, and risk status.

[0017] Furthermore, a three-dimensional scene model of the substation is constructed based on the spatial geometric data and then spatially discretized to obtain a set of spatial units, including:

[0018] After the spatial geometric data is preprocessed, it is unified to the substation coordinate system and converted into a preset three-dimensional model format. The spatial geometric data after model repair is obtained through model repair.

[0019] Based on the equipment type, area type, obstacle type and passage type in the substation, semantic mapping is performed on the spatial geometric data after model repair to obtain a 3D scene model with semantic information.

[0020] The initial spatial unit set is obtained by initially discretizing the spatial bounding range of the 3D scene model.

[0021] Based on the spatial intersection relationship between the entity object and the initial spatial unit, determine the spatial units to be subdivided that intersect with the physical spatial boundary of the entity object; the physical spatial boundary includes at least an outer envelope boundary, a spatial boundary, a region boundary, an obstacle boundary, and a passage boundary;

[0022] The spatial unit to be subdivided is subdivided based on a preset spatial unit scale to obtain spatial units corresponding to the physical spatial boundaries of the entity object.

[0023] A spatial index is established based on the spatial location of spatial units corresponding to the physical spatial boundary of the entity object, and an adjacency relationship is established based on the coplanarity, common edge, or preset adjacency distance between spatial units to obtain a set of spatial units.

[0024] Furthermore, the entity objects include at least equipment entities, area entities, obstacle entities, passage entities, risk entities, operation entities, and inspection target entities. The relationships between these entity objects include at least spatial affiliation mapping relationships, business association relationships, and state association relationships. A multimodal dynamic knowledge graph is constructed based on the business semantic data, multimodal perception data, and real-time state data, including:

[0025] Modal features are extracted from the multimodal sensing data, and feature alignment is performed based on the spatiotemporal correlation between different modal features. The aligned modal features are then mapped to a unified feature space for correlation and fusion to obtain the correlation and fusion result.

[0026] Based on the correlation fusion results, identify one or more of the following: equipment status, on-site obstacle, operation status and risk status, and associate the identification results with the corresponding equipment entity, obstacle entity, operation entity or risk entity;

[0027] Establish spatial attribution mapping relationships based on the spatial location, spatial enclosing range, region, and topological connection relationships of entity objects;

[0028] Establish business relationships based on equipment attributes, operating modes, safety procedures, on-site operation rules, and inspection task data, and associate the real-time status data with the corresponding entity objects to establish status relationships;

[0029] When the association fusion result or the real-time status data changes, the entity objects and the relationships between entity objects are updated to obtain a multimodal dynamic knowledge graph.

[0030] Furthermore, the entity object and its corresponding semantic attributes are associated with the corresponding spatial unit, and dynamic spatial constraints are generated based on the current state of the entity object and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial unit, including:

[0031] Based on the entity objects and relationships between them in the multimodal dynamic knowledge graph, as well as the safety procedures and on-site operation rules associated with the current entity object, determine the currently effective spatial constraint rules;

[0032] Based on the currently effective spatial constraint rules, determine the constraint source entity, constraint type, and spatial action parameters, and based on the spatial range of the constraint source entity, determine the spatial action range of the dynamic spatial constraint.

[0033] Based on the spatial range of the dynamic spatial constraints and the spatial intersection relationship between spatial units, the currently effective dynamic spatial constraints are associated with the corresponding spatial units to determine the current passage constraint state of each spatial unit.

[0034] When the constraint type and spatial action parameters of the constraint source entity change, the difference in the corresponding action range of the entity object that will cause the dynamic spatial constraint to change before and after the change is determined as the spatial constraint change region.

[0035] When the constraint type and spatial action parameters of the constraint source entity do not meet the currently effective spatial constraint rules, the corresponding dynamic spatial constraint is invalidated.

[0036] When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the currently effective spatial constraint rules, the spatial action parameters of the corresponding dynamic spatial constraint are updated according to the constraint type of the constraint source entity.

[0037] When the constraint type and spatial action parameters of the constraint source entity change from not satisfying to satisfying the currently effective spatial constraint rule, a corresponding dynamic spatial constraint is generated and applied to the corresponding spatial unit according to the constraint type and spatial action parameters of the constraint source entity.

[0038] Once the spatial constraint change region is determined, an initial set of affected spatial units is determined based on the spatial intersection relationship between the spatial constraint change region and the spatial units, and the passage constraint state of the corresponding spatial units is updated based on the initial set of affected spatial units.

[0039] Furthermore, the spatial range of the dynamic spatial constraint is expressed as:

[0040] ;

[0041] In the formula, Represents the physical device At any moment The spatial range of the dynamic spatial constraints formed Indicates the current time; This indicates the spatial location of the device entity within the 3D scene model; Represents three-dimensional Euclidean space; Represents the physical device The spatial outer envelope in a 3D scene model; Represents the spatial position of the device entity in the 3D scene model. Spatial outer envelope of the device entity in the 3D scene model The minimum distance; Indicates according to Relationships with other entity objects, and relationships with device entities. The current distance limits determined by the associated safety procedures and on-site operating rules, and When the constraint type and spatial action parameters of the constraint source entity no longer meet the current effective conditions, the corresponding dynamic spatial constraint becomes invalid.

[0042] Furthermore, the spatial constraint change region and the initial set of affected spatial units are respectively represented as:

[0043] ;

[0044] ;

[0045] In the formula, Indicates the current time The corresponding spatial constraint variation region; Indicates the current time The set of entity objects that cause changes in dynamic space constraints; and Representing the physical devices The corresponding scope of action under the dynamic space constraints before and after the state change; Represents the symmetric difference operation between two spatial sets; Represents the initial set of affected spatial units; Represents a set of spatial units; Indicates the first spatial units This represents the intersection operation of sets; This represents the empty set.

[0046] Furthermore, based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of reachable spatial units. Based on this set, and considering preset field-of-view constraints and the observation parameters corresponding to the patrol target, effective observation is determined to obtain a set of effective observation spatial units. Finally, observation areas are divided based on these spatial units to obtain the area division results, including:

[0047] Spatial units that do not meet the connectivity and preset static spatial constraints are eliminated to obtain a set of candidate passable spatial units;

[0048] Based on the candidate passable spatial unit set, an adjacency graph is constructed according to the adjacency relationship between the candidate passable spatial units. A connectivity search is performed starting from the spatial unit where the current inspection target entity is located. The candidate passable spatial units connected to the starting point are determined as the reachable spatial unit set, and the spatial units in the spatial unit set other than the reachable spatial unit set are determined as the unreachable spatial unit set.

[0049] Based on the target type of the current inspection target entity and its associated observation parameter conditions, the corresponding target observation constraints are determined; the target observation constraints include one or more of the following constraints that must be satisfied when effectively observing the inspection target: observation distance constraint, observation angle constraint, visibility constraint, and occlusion constraint.

[0050] Candidate observation spatial units that can be used as observation positions of the inspection equipment are determined from the set of spatial units. Based on the three-dimensional scene model, it is determined whether there is occlusion between the candidate observation spatial units and the inspection target entity. Target observation constraint determination is performed on the candidate observation spatial units without occlusion.

[0051] Candidate observation spatial units that satisfy the target observation constraints are determined as the set of effective observation spatial units;

[0052] Spatial units that belong to both the reachable spatial unit set and the effective observation spatial unit set are divided into reachable and effective observation regions; spatial units that belong to the reachable spatial unit set but not to the effective observation spatial unit set are divided into reachable but ineffective observation regions; spatial units that belong to both the inaccessible spatial unit set and the effective observation spatial unit set are divided into inaccessible but effective observation regions; spatial units that belong to both the inaccessible spatial unit set and the ineffective observation spatial unit set are divided into inaccessible and ineffective observation regions, thus obtaining the region division results.

[0053] Furthermore, based on the entity objects associated with the region division results and dynamic spatial constraints, a three-dimensional patrolable spatial situation map is constructed on the three-dimensional scene model, and the three-dimensional patrolable spatial situation map is updated based on changes in entity objects or relationships between entity objects, including:

[0054] When changes in the device's open / closed state, device's operating state, energized state, work state, work range, the spatial location of the inspection target entity, the observation parameter conditions, or the spatial location of the obstacle entity in the multimodal dynamic knowledge graph cause changes in dynamic spatial constraints, the corresponding dynamic spatial constraints are updated according to the changed entity objects and the relationships between entity objects, and the range of dynamic spatial constraints before and after the update is compared.

[0055] For entity objects whose scope of dynamic spatial constraints changes before and after the update, the region of difference in scope of dynamic spatial constraints before and after the change is defined as the spatial constraint change region.

[0056] Based on the spatial intersection relationship between the spatial constraint change region and the spatial unit, an initial set of affected spatial units is determined, and the connectivity and reachability of each spatial unit in the initial set of affected spatial units are re-determined.

[0057] When the reachability state of a spatial unit in the initial set of affected spatial units changes and affects the connectivity of adjacent spatial units, the reachability of the affected spatial units continues to be determined until the reachability state of the affected spatial units no longer changes.

[0058] When the spatial location, observation parameters, or spatial location and enclosing range of the target entity change, effective observation is re-determined within the candidate observation spatial unit associated with the target entity based on the changed field of view constraints and target observation constraints.

[0059] Based on the updated accessibility determination results and effective observation determination results, the reachable and effective observation areas, the reachable but ineffective observation areas, the inaccessible but effective observation areas, and the inaccessible but ineffective observation areas are redefined, and the corresponding areas in the three-dimensional patrolable spatial situation map are updated.

[0060] Secondly, the present invention provides a three-dimensional roving domain space modeling system for substations, used to implement the three-dimensional roving domain space modeling method for substations described in the first aspect, including:

[0061] The data acquisition module is used to acquire spatial geometric data, business semantic data, multimodal sensing data, and real-time status data of the substation.

[0062] The spatial model construction module is used to construct a three-dimensional scene model within the substation based on the spatial geometric data and to spatially discretize it to obtain a set of spatial units.

[0063] The graph construction module is used to construct a multimodal dynamic knowledge graph based on the business semantic data, multimodal perception data and real-time status data. The multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects.

[0064] The constraint construction module is used to associate the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generate dynamic spatial constraints based on the current state of the entity object and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial unit.

[0065] The region division module is used to determine the reachability of spatial units based on the connectivity of spatial units, preset static spatial constraints and the dynamic spatial constraints, to obtain a set of reachable spatial units. Based on the set of reachable spatial units, it determines the effective observation based on preset field of view constraints and observation parameter conditions corresponding to the inspection target, to obtain a set of effective observation spatial units. Based on the spatial units, it divides the observation area to obtain the region division result.

[0066] The dynamic update module is used to construct a three-dimensional patrolable spatial situation map on the three-dimensional scene model based on the entity objects associated with the region division results and dynamic spatial constraints, and to update the three-dimensional patrolable spatial situation map based on the changes in entity objects or the relationships between entity objects.

[0067] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0068] 1. This invention constructs a multimodal dynamic knowledge graph and associates entity objects with spatial units. Based on the current state of the entity objects and preset safety procedures and on-site operation rules, dynamic spatial constraints are generated. Spatial units are divided based on reachability and effective observation. A dynamically updatable three-dimensional patrolable spatial situation map is constructed. This transforms the planning of inspection paths and observation points, which relies on manual experience, into a systematic quantitative analysis process based on real-time spatial data. This effectively overcomes the visual blind spots that exist in manual point placement. At the same time, through the real-time updating of dynamic spatial constraints and the dynamic linkage of the situation map, the inspection path can be automatically adjusted according to changes in operating conditions and work status, significantly improving the scientificity, adaptability, and safety of substation inspection path planning.

[0069] 2. This invention discretizes the substation space into a set of spatial units, and determines accessibility based on the connectivity of the spatial units, preset static spatial constraints, and dynamic spatial constraints to obtain a set of accessible spatial units. Simultaneously, it determines effective observation based on the field of view constraints and the observation parameters corresponding to the inspection target to obtain a set of effective observation spatial units. Furthermore, the spatial units are divided into accessible and effectively observable areas, accessible but not effectively observable areas, inaccessible but effectively observable areas, and inaccessible but not effectively observable areas. This provides a systematic quantitative analysis method for inspection path planning and observation point layout, overcoming the problems of existing technologies that rely on manual experience for site surveys and cannot quantitatively assess blind spots.

[0070] 3. This invention constructs a multimodal dynamic knowledge graph, associating entity objects and their semantic attributes with corresponding spatial units. It generates dynamic spatial constraints based on the current state of the entity objects and preset safety procedures and on-site operation rules, applying them to the corresponding spatial units. When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the currently effective spatial constraint rules, the spatial range of the corresponding dynamic spatial constraint is updated according to the constraint type and spatial action parameters of the constraint source entity. When the constraint type and spatial action parameters of the constraint source entity no longer satisfy the currently effective spatial constraint rules, the corresponding dynamic spatial constraint becomes invalid. When the constraint type and spatial action parameters of the constraint source entity change from not satisfying to satisfying the effective spatial constraint rules, the corresponding dynamic spatial constraint is generated and applied to the corresponding spatial unit. This achieves automatic updates of dynamic spatial constraints in real-time with changes in equipment operating status, energized status, operational status, operational range, and risk status, solving the problem in existing technologies where static maps cannot adapt to dynamic changes in operating conditions.

[0071] 4. This invention constructs a three-dimensional patrolable spatial situation map on a three-dimensional scene model. When changes in the operating status, energized status, operational status, operational range, spatial location of the patrol target entity, observation parameter conditions, or spatial location of the obstacle entity in the multimodal dynamic knowledge graph cause changes in dynamic spatial constraints, the corresponding dynamic spatial constraints are updated based on the changed entity objects and the relationships between them, and the spatial constraint change area is determined. Based on the spatial constraint change area, an initial set of affected spatial units is determined, and the accessibility of the initial affected spatial units is re-determined. When the accessibility status changes and affects the connectivity of adjacent spatial units, the affected range is expanded along the adjacency relationship, and accessibility determination continues until the accessibility status of the newly added spatial units no longer changes. Simultaneously, when the spatial location of the patrol target entity, observation parameter conditions, or the spatial location or spatial enclosure of the obstacle entity changes, effective observation determination is re-performed within the candidate observation spatial units. Finally, based on the updated accessibility determination results and effective observation determination results, the region is re-divided, and the corresponding region in the situation map is updated. This achieves real-time linkage and updating of the three-dimensional patrolable spatial situation map and dynamic spatial constraints, providing real-time and accurate spatial decision-making basis for intelligent substation inspection. Attached Figure Description

[0072] Figure 1 This is a flowchart illustrating a method for modeling a three-dimensional patrolable space of a substation, as provided in an embodiment of the present invention. Detailed Implementation

[0073] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0074] Example 1

[0075] like Figure 1 As shown in the figure, this embodiment introduces a method for modeling a three-dimensional patrolable domain space of a substation, including:

[0076] Step 1: Acquire the substation's spatial geometric data, business semantic data, multimodal sensing data, and real-time status data.

[0077] This embodiment acquires spatial geometric data, business semantic data, multimodal perception data, and real-time status data of the substation, enabling 3D patrolable spatial modeling to simultaneously integrate the substation's geometric morphology information, semantic information such as equipment ledgers and procedures, perception information such as images and audio, and real-time dynamic information such as equipment operation and work. This avoids the problem of spatial models being out of touch with reality due to incomplete information from a single data source.

[0078] Step 2: Construct a three-dimensional scene model of the substation based on the spatial geometric data and perform spatial discretization to obtain a set of spatial units.

[0079] This embodiment constructs a three-dimensional scene model within the substation based on spatial geometric data, and then spatially discretizes the three-dimensional scene model to obtain a set of spatial units. This transforms the continuous three-dimensional space into discretized spatial units, enabling the three-dimensional space to change from a "visual display" to a "computable unit," thus solving the problem that continuous space cannot be quantitatively analyzed.

[0080] Step 3: Construct a multimodal dynamic knowledge graph based on the business semantic data, multimodal perception data, and real-time status data.

[0081] In this embodiment, the multimodal dynamic knowledge graph includes entity objects and the relationships between them. This embodiment constructs a multimodal dynamic knowledge graph based on multimodal perception data, business semantic data, and real-time status data. It extracts modal features from the multimodal perception data and performs correlation and fusion. Based on the correlation and fusion results, it identifies equipment status, on-site obstacles, operational status, and risk status and associates them with corresponding entity objects. It establishes spatial affiliation mapping relationships based on the spatial location, spatial enclosure, region, and topological connections of entity objects. It establishes business correlation relationships based on equipment attributes, operating modes, safety procedures, on-site operation rules, and inspection task data. Finally, it associates real-time status data with corresponding entity objects to establish status correlation relationships. This achieves unified correlation of multi-source heterogeneous data at the spatial semantic level, providing a semantic mapping foundation for the correlation between spatial units and entity objects.

[0082] Step 4: Associate the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generate dynamic spatial constraints based on the current state of the entity object and the preset safety procedures and on-site operation rules, and apply them to the corresponding spatial unit.

[0083] This embodiment associates entity objects and their semantic attributes with corresponding spatial units, and generates dynamic spatial constraints based on the current state of the entity objects and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial units. The currently effective spatial constraint rules are determined based on the current state of the entity objects and the associated safety procedures and on-site operation rules. The spatial range of the dynamic spatial constraints is determined based on the constraint source entity. The dynamic spatial constraints are associated with the corresponding spatial units based on the spatial intersection relationship between the spatial range of the dynamic spatial constraints and the spatial units to determine the passage constraint status of each spatial unit. When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the effective spatial constraint rules, the spatial range is updated. When the constraints are no longer satisfied, the effective spatial constraint rules are invalidated. When the constraints change from not satisfying to satisfying, new effective spatial constraint rules are generated. This realizes that spatial constraints are automatically updated according to the dynamic changes of equipment operating status, energized status, operation status, and risk status.

[0084] Step 5: Based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of accessible spatial units. Based on the set of accessible spatial units, effective observation is determined based on preset field-of-view constraints and observation parameter conditions corresponding to the inspection target to obtain a set of effective observation spatial units. Then, the observation area is divided based on the spatial units to obtain the area division result.

[0085] This embodiment determines the reachability of spatial units based on connectivity, preset static spatial constraints, and dynamic spatial constraints. It then determines the effective observation of spatial units based on preset field-of-view constraints and observation parameters corresponding to the inspection target. Finally, it divides the spatial units into reachable and effectively observable areas, reachable but not effectively observable areas, inaccessible but effectively observable areas, and inaccessible but not effectively observable areas, based on the reachability and effective observation determinations. By combining reachability determination with effective observation determination, it achieves a joint analysis of the two dimensions of "accessibility" and "effective observability" in the substation space, providing a systematic quantitative basis for inspection path planning and observation point deployment.

[0086] Step Six: Based on the entity objects and dynamic spatial constraints associated with the region division results, construct a three-dimensional patrolable spatial situation map on the three-dimensional scene model, and update the three-dimensional patrolable spatial situation map based on changes in entity objects or relationships between entity objects.

[0087] This embodiment constructs a three-dimensional patrolable spatial situation map on a three-dimensional scene model based on the entity objects and dynamic spatial constraints associated with the region division results. When changes in entity objects or the relationships between entity objects cause changes in dynamic spatial constraints, the corresponding dynamic spatial constraints are updated according to the changed entity objects and the relationships between entity objects, and the spatial constraint change area is determined. The initial set of affected spatial units is determined according to the spatial constraint change area, and accessibility is determined. When the spatial position of the patrol target entity, the observation parameter conditions, or the spatial position of the obstacle changes, the effective observation determination is re-performed, and the four types of regions are re-divided according to the updated determination results, and the corresponding regions in the situation map are updated. This realizes the real-time linkage update of the three-dimensional patrolable spatial situation map and dynamic spatial constraints, providing real-time and accurate spatial situation information for intelligent substation inspection.

[0088] Example 2

[0089] Based on the same inventive concept as Embodiment 1, this embodiment describes the implementation steps of a three-dimensional patrolable spatial modeling method for substations, including:

[0090] Step 1: Acquire the substation's spatial geometric data, business semantic data, multimodal sensing data, and real-time status data.

[0091] In this embodiment, the spatial geometric data includes at least one of laser point cloud data, oblique photogrammetry model data, and video reconstruction model data; the business semantic data includes at least one of equipment ledger data, equipment coding data, equipment attribute data, operation mode data, safety procedure data, area division data, on-site operation rules, and patrol task data; the multimodal perception data includes at least any two or more of visible light image data, infrared image data, audio data, and text data; and the real-time status data includes at least one of equipment disconnection / connection status, equipment operation status, energized status, operation status, operation range, and risk status.

[0092] Step 2: Construct a three-dimensional scene model of the substation based on the spatial geometric data and perform spatial discretization to obtain a set of spatial units.

[0093] This embodiment constructs a three-dimensional scene model within the substation based on one or more of the laser point cloud data, oblique photogrammetry model data, or video reconstruction model data in the spatial geometric data.

[0094] Step 2.1: After obtaining the spatial geometric data and preprocessing it, convert it to a preset three-dimensional model format in the substation coordinate system, and obtain the spatial geometric data after model repair through model repair.

[0095] Step 2.2: Based on the equipment type, area type, obstacle type, and passage type within the substation, perform semantic mapping on the spatial geometric data after model repair to obtain a 3D scene model with semantic information.

[0096] Step 2.3: Perform initial discretization based on the spatial bounding range of the 3D scene model to obtain an initial set of spatial units.

[0097] Step 2.4: Based on the spatial intersection relationship between the entity object and the initial spatial unit, determine the spatial units to be subdivided that intersect with the physical spatial boundary of the entity object; the physical spatial boundary includes at least the outer envelope boundary, spatial boundary, region boundary, obstacle boundary and channel boundary.

[0098] Step 2.5: Subdivide the spatial unit to be subdivided based on the preset spatial unit scale to obtain spatial units corresponding to the physical spatial boundaries of the entity object.

[0099] Step 2.6: Establish a spatial index based on the spatial position of the spatial unit corresponding to the physical spatial boundary of the entity object, and establish adjacency relationship between spatial units according to coplanarity, common edge or preset adjacency distance to obtain a set of spatial units.

[0100] Step 3: Construct a multimodal dynamic knowledge graph based on the business semantic data, multimodal perception data, and real-time status data.

[0101] In this embodiment, the multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects. The entity objects include at least equipment entities, region entities, obstacle entities, channel entities, risk entities, operation entities, and inspection target entities. The relationships between entity objects include at least spatial affiliation mapping relationships, business association relationships, and state association relationships.

[0102] Step 3.1: Extract modal features from the multimodal sensing data respectively, align features according to the spatiotemporal correlation between different modal features, and map the aligned modal features to a unified feature space for correlation fusion to obtain the correlation fusion result.

[0103] Step 3.2: Identify one or more of the following based on the association fusion results: equipment status, on-site obstacle, operation status, and risk status, and associate the identification results with the corresponding equipment entity, obstacle entity, operation entity, or risk entity.

[0104] Step 3.3: Establish spatial attribution mapping relationships based on the spatial location, spatial enclosing range, region, and topological connection relationships of the entity object.

[0105] Step 3.4: Establish business relationships based on equipment attributes, operating modes, safety procedures, on-site operation rules, and inspection task data, and associate the real-time status data with the corresponding entity objects to establish status relationships.

[0106] Step 3.5: When the association fusion result or the real-time status data changes, update the entity objects and the relationships between entity objects to obtain a multimodal dynamic knowledge graph.

[0107] Step 4: Associate the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generate dynamic spatial constraints based on the current state of the entity object and the preset safety procedures and on-site operation rules, and apply them to the corresponding spatial unit.

[0108] Step 4.1: Based on the entity objects and relationships between entity objects in the multimodal dynamic knowledge graph, as well as the safety procedures and on-site operation rules associated with the current entity object, determine the currently effective spatial constraint rules.

[0109] Step 4.2: Based on the currently effective spatial constraint rules, determine the constraint source entity, constraint type, and spatial action parameters, and based on the spatial range of the constraint source entity, determine the spatial action range of the dynamic spatial constraint.

[0110] In this embodiment, the spatial range of the dynamic spatial constraint is represented as:

[0111] ;

[0112] In the formula, Represents the physical device At any moment The spatial range of the dynamic spatial constraints formed Indicates the current time; This indicates the spatial location of the device entity within the 3D scene model; Represents three-dimensional Euclidean space; Represents the physical device The spatial outer envelope in a 3D scene model; Represents the spatial position of the device entity in the 3D scene model. Spatial outer envelope of the device entity in the 3D scene model The minimum distance; Indicates according to Relationships with other entity objects, and relationships with device entities. The current distance limits determined by the associated safety procedures and on-site operating rules, and When the constraint type and spatial action parameters of the constraint source entity no longer meet the current effective conditions, the corresponding dynamic spatial constraint becomes invalid.

[0113] Step 4.3: Based on the spatial range of the dynamic spatial constraints and the spatial intersection relationship between spatial units, associate the currently effective dynamic spatial constraints with the corresponding spatial units to determine the current passage constraint status of each spatial unit.

[0114] Step 4.4: When the constraint type and spatial action parameters of the constraint source entity change, the difference in the corresponding action range of the entity object that causes the dynamic spatial constraint to change before and after the change is determined as the spatial constraint change region.

[0115] Step 4.5: When the constraint type and spatial action parameters of the constraint source entity do not meet the currently effective spatial constraint rules, the corresponding dynamic spatial constraint is invalidated.

[0116] Step 4.6: When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the currently effective spatial constraint rules, update the spatial action parameters of the corresponding dynamic spatial constraint according to the constraint type of the constraint source entity.

[0117] Step 4.7: When the constraint type and spatial action parameters of the constraint source entity change from not satisfying to satisfying the currently effective spatial constraint rule, generate the corresponding dynamic spatial constraint according to the constraint type and spatial action parameters of the constraint source entity and apply it to the corresponding spatial unit.

[0118] In this embodiment, once the spatial constraint change region is determined, an initial set of affected spatial units is determined based on the spatial intersection relationship between the spatial constraint change region and the spatial units, and the passage constraint state of the corresponding spatial units is updated based on the initial set of affected spatial units.

[0119] In this embodiment, the spatial constraint change region and the initial set of affected spatial units are respectively represented as:

[0120] ;

[0121] ;

[0122] In the formula, Indicates the current time The corresponding spatial constraint variation region; Indicates the current time The set of entity objects that cause changes in dynamic space constraints; and Representing the physical devices The corresponding scope of action under the dynamic space constraints before and after the state change; Represents the symmetric difference operation between two spatial sets; Represents the initial set of affected spatial units; Represents a set of spatial units; Indicates the first spatial units This represents the intersection operation of sets; This represents the empty set.

[0123] Step 5: Based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of accessible spatial units. Based on the set of accessible spatial units, effective observation is determined based on preset field-of-view constraints and observation parameter conditions corresponding to the inspection target to obtain a set of effective observation spatial units. Then, the observation area is divided based on the spatial units to obtain the area division result.

[0124] Step 5.1: Eliminate spatial units that do not meet the connectivity and preset static spatial constraints of spatial units to obtain a set of candidate passable spatial units.

[0125] Step 5.2: Based on the candidate passable spatial unit set, construct an adjacency graph according to the adjacency relationship between the candidate passable spatial units, and perform connectivity search starting from the spatial unit where the current inspection target entity is located. The candidate passable spatial units connected to the starting point are determined as the reachable spatial unit set, and the spatial units in the spatial unit set other than the reachable spatial unit set are determined as the unreachable spatial unit set.

[0126] Step 5.3: Determine the corresponding target observation constraints based on the target type of the current inspection target entity and its associated observation parameter conditions.

[0127] In this embodiment, the target observation constraints include one or more of the following constraints that must be met when effectively observing the patrol target: observation distance constraint, observation angle constraint, visibility constraint, and occlusion constraint.

[0128] Step 5.4: Determine candidate observation spatial units from the set of spatial units that can be used as observation positions for the inspection equipment, determine whether there is occlusion between the candidate observation spatial units and the inspection target entity based on the three-dimensional scene model, and perform target observation constraint determination on the candidate observation spatial units without occlusion.

[0129] Step 5.5: Determine the candidate observation spatial units that satisfy the target observation constraints as the set of effective observation spatial units.

[0130] Step 5.6: Divide the spatial units that belong to both the reachable spatial unit set and the effective observation spatial unit set into reachable and effective observation regions; divide the spatial units that belong to the reachable spatial unit set but not the effective observation spatial unit set into reachable but ineffective observation regions; divide the spatial units that belong to both the inaccessible spatial unit set and the effective observation spatial unit set into inaccessible but effective observation regions; divide the spatial units that belong to both the inaccessible spatial unit set and the ineffective observation spatial unit set into inaccessible and ineffective observation regions, thereby obtaining the region division results.

[0131] Step 6: Based on the entity objects and dynamic spatial constraints associated with the region division results, construct a three-dimensional patrolable spatial situation map on the three-dimensional scene model, and update the three-dimensional patrolable spatial situation map based on changes in entity objects or relationships between entity objects.

[0132] In this embodiment, when the relationship between entity objects changes, the spatial affiliation mapping relationship, business association relationship or state association relationship is updated, and steps 6.1 to 6.6 are executed according to the updated relationship.

[0133] Step 6.1: When changes in the device's separation / combination state, device operation state, energized state, operation state, operation range, spatial location of the inspection target entity, observation parameter conditions, or spatial location of the obstacle entity in the multimodal dynamic knowledge graph cause changes in dynamic spatial constraints, update the corresponding dynamic spatial constraints according to the changed entity objects and the relationships between entity objects, and compare the range of dynamic spatial constraints before and after the update.

[0134] Step 6.2: For entity objects whose scope of dynamic spatial constraints changes before and after the update, the region of difference in scope of dynamic spatial constraints before and after the change is defined as the spatial constraint change region.

[0135] Step 6.3: Based on the spatial intersection relationship between the spatial constraint change region and the spatial unit, determine the initial set of affected spatial units, and re-determine the connectivity and reachability of each spatial unit in the initial set of affected spatial units.

[0136] Step 6.4: When the reachability state of a spatial unit in the initial set of affected spatial units changes and affects the connectivity of adjacent spatial units, continue to determine the reachability of the affected spatial units until the reachability state of the affected spatial units no longer changes.

[0137] Step 6.5: When the spatial position, observation parameters, or spatial position and enclosing range of the target entity change, a new effective observation determination is made within the candidate observation spatial unit associated with the target entity, based on the changed field of view constraints and target observation constraints.

[0138] Step 6.6: Based on the updated reachability determination results and effective observation determination results, redetermine the reachable and effective observation area, the reachable but ineffective observation area, the inaccessible but effective observation area, and the inaccessible but ineffective observation area, and update the corresponding areas in the three-dimensional patrolable spatial situation map.

[0139] Example 3

[0140] Based on the same inventive concept as other embodiments, this embodiment introduces a three-dimensional traversable spatial modeling system for substations, used to implement the three-dimensional traversable spatial modeling method for substations described in Embodiment 1 or 2, including:

[0141] The data acquisition module is used to acquire spatial geometric data, business semantic data, multimodal sensing data, and real-time status data of the substation.

[0142] The spatial model construction module is used to construct a three-dimensional scene model within the substation based on the spatial geometric data and to spatially discretize it to obtain a set of spatial units.

[0143] The graph construction module is used to construct a multimodal dynamic knowledge graph based on the business semantic data, multimodal perception data and real-time status data. The multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects.

[0144] The constraint construction module is used to associate the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generate dynamic spatial constraints based on the current state of the entity object and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial unit.

[0145] The region division module is used to determine the reachability of spatial units based on the connectivity of spatial units, preset static spatial constraints and the dynamic spatial constraints, to obtain a set of reachable spatial units. Based on the set of reachable spatial units, it determines the effective observation based on preset field of view constraints and observation parameter conditions corresponding to the inspection target, to obtain a set of effective observation spatial units. Based on the spatial units, it divides the observation area to obtain the region division result.

[0146] The dynamic update module is used to construct a three-dimensional patrolable spatial situation map on the three-dimensional scene model based on the entity objects associated with the region division results and dynamic spatial constraints, and to update the three-dimensional patrolable spatial situation map based on the changes in entity objects or the relationships between entity objects.

[0147] The specific functions of each module described above are explained in the relevant content of Embodiment 1 or 2, and will not be repeated here.

[0148] In summary, this invention constructs a multimodal dynamic knowledge graph and associates entity objects with spatial units. It generates dynamic spatial constraints based on the current state of the entity objects and preset safety procedures and on-site operation rules. Spatial units are divided based on reachability and effective observation criteria, and a dynamically updatable three-dimensional patrolable spatial situation map is constructed. This transforms the manual experience-dependent planning of inspection paths and observation points into a systematic quantitative analysis process based on real-time spatial data, effectively overcoming the visual blind spots inherent in manual point placement. Furthermore, through real-time updates of dynamic spatial constraints and dynamic linkage with the situation map, the inspection path can be automatically adjusted according to changes in operating conditions and work status, significantly improving the scientific rigor, adaptability, and safety of substation inspection path planning.

[0149] This invention discretizes the substation space into a set of spatial units, and determines accessibility based on the connectivity of the spatial units, preset static spatial constraints, and dynamic spatial constraints to obtain a set of accessible spatial units. Simultaneously, it determines effective observation based on the field of view constraints and the observation parameters corresponding to the inspection target to obtain a set of effective observation spatial units. Furthermore, the spatial units are divided into accessible and effectively observable areas, accessible but not effectively observable areas, inaccessible but effectively observable areas, and inaccessible but not effectively observable areas. This provides a systematic quantitative analysis method for inspection path planning and observation point layout, overcoming the problems of existing technologies that rely on manual experience for site surveys and cannot quantitatively assess blind spots.

[0150] This invention constructs a multimodal dynamic knowledge graph, associating entity objects and their semantic attributes with corresponding spatial units. It then generates dynamic spatial constraints based on the current state of the entity objects and preset safety procedures and on-site operation rules, applying these constraints to the corresponding spatial units. When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the currently effective spatial constraint rules, the spatial range of the corresponding dynamic spatial constraint is updated according to the constraint type and spatial action parameters of the constraint source entity. When the constraint type and spatial action parameters of the constraint source entity no longer satisfy the currently effective spatial constraint rules, the corresponding dynamic spatial constraint is invalidated. When the constraint type and spatial action parameters of the constraint source entity change from not satisfying to satisfying the effective spatial constraint rules, the corresponding dynamic spatial constraint is generated and applied to the corresponding spatial unit. This achieves automatic updates of dynamic spatial constraints in real-time with changes in equipment operating status, energized status, operational status, operational range, and risk status, solving the problem in existing technologies where static maps cannot adapt to dynamic changes in operating conditions.

[0151] This invention constructs a 3D patrolable spatial situation map on a 3D scene model. When changes in the operating status, energized status, operational status, operational range, spatial location of the patrol target entity, observation parameter conditions, or spatial location of the obstacle entity in the multimodal dynamic knowledge graph cause changes in dynamic spatial constraints, the corresponding dynamic spatial constraints are updated based on the changed entity objects and the relationships between them, and the spatial constraint change area is determined. Based on the spatial constraint change area, an initial set of affected spatial units is determined, and the accessibility of the initial affected spatial units is re-determined. When the accessibility status changes and affects the connectivity of adjacent spatial units, the affected range is expanded along the adjacency relationship, and accessibility determination continues until the accessibility status of the newly added spatial units no longer changes. Simultaneously, when the spatial location of the patrol target entity, observation parameter conditions, or the spatial location or spatial enclosure of the obstacle entity changes, effective observation determination is re-performed within the candidate observation spatial units. Finally, based on the updated accessibility determination results and effective observation determination results, the region is re-divided, and the corresponding region in the situation map is updated. This achieves real-time linkage and updating of the 3D patrolable spatial situation map and dynamic spatial constraints, providing real-time and accurate spatial decision-making basis for intelligent substation inspection.

[0152] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0153] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0154] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0155] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0156] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for modeling a three-dimensional traversable spatial domain of a substation, characterized in that, include: Acquire spatial geometric data, business semantic data, multimodal sensing data, and real-time status data of the substation; Based on the spatial geometric data, a three-dimensional scene model of the substation is constructed and spatially discretized to obtain a set of spatial units; A multimodal dynamic knowledge graph is constructed based on the business semantic data, multimodal perception data, and real-time status data. The multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects. The entity object and its corresponding semantic attributes are associated with the corresponding spatial unit, and dynamic spatial constraints are generated according to the current state of the entity object and the preset safety procedures and on-site operation rules, and applied to the corresponding spatial unit. Based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of reachable spatial units. Based on the set of reachable spatial units, effective observation is determined based on preset field of view constraints and observation parameter conditions corresponding to the inspection target to obtain a set of effective observation spatial units. Then, the observation area is divided based on the spatial units to obtain the area division result. Based on the entity objects associated with the region division results and dynamic spatial constraints, a three-dimensional patrolable spatial situation map is constructed on the three-dimensional scene model, and the three-dimensional patrolable spatial situation map is updated based on the changes in entity objects or the relationships between entity objects.

2. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 1, characterized in that, The spatial geometric data includes at least one of the following: laser point cloud data, oblique photogrammetry model data, and video reconstruction model data; The business semantic data includes at least one of the following: equipment ledger data, equipment code data, equipment attribute data, operation mode data, safety procedure data, area division data, on-site operation rules, and inspection task data; The multimodal sensing data includes at least two or more of the following: visible light image data, infrared image data, audio data, and text data. The real-time status data includes at least one of the following: equipment disconnection / connection status, equipment operation status, energized status, operation status, operation scope, and risk status.

3. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 2, characterized in that, Based on the spatial geometric data, a three-dimensional scene model of the substation is constructed and spatially discretized to obtain a set of spatial units, including: After the spatial geometric data is preprocessed, it is unified to the substation coordinate system and converted into a preset three-dimensional model format. The spatial geometric data after model repair is obtained through model repair. Based on the equipment type, area type, obstacle type and passage type in the substation, semantic mapping is performed on the spatial geometric data after model repair to obtain a 3D scene model with semantic information. The initial spatial unit set is obtained by initially discretizing the spatial bounding range of the 3D scene model. Based on the spatial intersection relationship between the entity object and the initial spatial unit, determine the spatial units to be subdivided that intersect with the physical spatial boundary of the entity object; the physical spatial boundary includes at least an outer envelope boundary, a spatial boundary, a region boundary, an obstacle boundary, and a passage boundary; The spatial unit to be subdivided is subdivided based on a preset spatial unit scale to obtain spatial units corresponding to the physical spatial boundaries of the entity object. A spatial index is established based on the spatial location of spatial units corresponding to the physical spatial boundary of the entity object, and an adjacency relationship is established based on the coplanarity, common edge, or preset adjacency distance between spatial units to obtain a set of spatial units.

4. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 3, characterized in that, The entity objects include at least equipment entities, area entities, obstacle entities, passage entities, risk entities, operation entities, and inspection target entities. The relationships between these entity objects include at least spatial affiliation mapping relationships, business association relationships, and status association relationships. A multimodal dynamic knowledge graph is constructed based on the business semantic data, multimodal perception data, and real-time status data, including: Modal features are extracted from the multimodal sensing data, and feature alignment is performed based on the spatiotemporal correlation between different modal features. The aligned modal features are then mapped to a unified feature space for correlation and fusion to obtain the correlation and fusion result. Based on the correlation fusion results, identify one or more of the following: equipment status, on-site obstacle, operation status and risk status, and associate the identification results with the corresponding equipment entity, obstacle entity, operation entity or risk entity; Establish spatial attribution mapping relationships based on the spatial location, spatial enclosing range, region, and topological connection relationships of entity objects; Establish business relationships based on equipment attributes, operating modes, safety procedures, on-site operation rules, and inspection task data, and associate the real-time status data with the corresponding entity objects to establish status relationships; When the association fusion result or the real-time status data changes, the entity objects and the relationships between entity objects are updated to obtain a multimodal dynamic knowledge graph.

5. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 4, characterized in that, Associating the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generating dynamic spatial constraints based on the current state of the entity object and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial unit, including: Based on the entity objects and relationships between them in the multimodal dynamic knowledge graph, as well as the safety procedures and on-site operation rules associated with the current entity object, determine the currently effective spatial constraint rules; Based on the currently effective spatial constraint rules, determine the constraint source entity, constraint type, and spatial action parameters, and based on the spatial range of the constraint source entity, determine the spatial action range of the dynamic spatial constraint. Based on the spatial range of the dynamic spatial constraints and the spatial intersection relationship between spatial units, the currently effective dynamic spatial constraints are associated with the corresponding spatial units to determine the current passage constraint state of each spatial unit. When the constraint type and spatial action parameters of the constraint source entity change, the difference in the corresponding action range of the entity object that will cause the dynamic spatial constraint to change before and after the change is determined as the spatial constraint change region. When the constraint type and spatial action parameters of the constraint source entity do not meet the currently effective spatial constraint rules, the corresponding dynamic spatial constraint is invalidated. When the constraint type and spatial action parameters of the constraint source entity change but still satisfy the currently effective spatial constraint rules, the spatial action parameters of the corresponding dynamic spatial constraint are updated according to the constraint type of the constraint source entity. When the constraint type and spatial action parameters of the constraint source entity change from not satisfying to satisfying the currently effective spatial constraint rule, a corresponding dynamic spatial constraint is generated and applied to the corresponding spatial unit according to the constraint type and spatial action parameters of the constraint source entity. Once the spatial constraint change region is determined, an initial set of affected spatial units is determined based on the spatial intersection relationship between the spatial constraint change region and the spatial units, and the passage constraint state of the corresponding spatial units is updated based on the initial set of affected spatial units.

6. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 5, characterized in that, The spatial range of the dynamic spatial constraint is expressed as follows: ； In the formula, Represents the physical device At any moment The spatial scope of the dynamic spatial constraints formed; Indicates the current time; This indicates the spatial location of the device entity within the 3D scene model; Represents three-dimensional Euclidean space; Represents the physical device The spatial outer envelope in a 3D scene model; Represents the spatial position of the device entity in the 3D scene model. Spatial outer envelope of the device entity in the 3D scene model The minimum distance; Indicates based on the device entity Relationships with other entity objects, and relationships with device entities. The current distance limits determined by the associated safety procedures and on-site operating rules, and ; When the constraint type and spatial action parameters of the constraint source entity no longer meet the current effective conditions, the corresponding dynamic spatial constraint becomes invalid.

7. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 6, characterized in that, The spatial constraint change region and the initial set of affected spatial units are respectively represented as follows: ； ； In the formula, Indicates the current time The corresponding spatial constraint variation region; Indicates the current time The set of entity objects that cause changes in dynamic space constraints; and Representing the physical devices The corresponding scope of action under the dynamic space constraints before and after the state change; Represents the symmetric difference operation between two spatial sets; Represents the initial set of affected spatial units; Represents a set of spatial units; Indicates the first One spatial unit; This represents the intersection operation of sets; This represents the empty set.

8. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 7, characterized in that, Based on the connectivity of spatial units, preset static spatial constraints, and the dynamic spatial constraints, accessibility is determined to obtain a set of reachable spatial units. Then, based on this set, and considering preset line-of-sight constraints and the observation parameters corresponding to the patrol target, effective observation is determined to obtain a set of effective observation spatial units. Finally, observation areas are divided based on these spatial units to obtain the area division results, including: Spatial units that do not meet the connectivity and preset static spatial constraints are eliminated to obtain a set of candidate passable spatial units; Based on the candidate passable spatial unit set, an adjacency graph is constructed according to the adjacency relationship between the candidate passable spatial units. A connectivity search is performed starting from the spatial unit where the current inspection target entity is located. The candidate passable spatial units connected to the starting point are determined as the reachable spatial unit set, and the spatial units in the spatial unit set other than the reachable spatial unit set are determined as the unreachable spatial unit set. Based on the target type of the current inspection target entity and its associated observation parameter conditions, the corresponding target observation constraints are determined; the target observation constraints include one or more of the following constraints that must be satisfied when effectively observing the inspection target: observation distance constraint, observation angle constraint, visibility constraint, and occlusion constraint. Candidate observation spatial units that can be used as observation positions of the inspection equipment are determined from the set of spatial units. Based on the three-dimensional scene model, it is determined whether there is occlusion between the candidate observation spatial units and the inspection target entity. Target observation constraint determination is performed on the candidate observation spatial units without occlusion. Candidate observation spatial units that satisfy the target observation constraints are determined as the set of effective observation spatial units; Spatial units that belong to both the reachable spatial unit set and the effective observation spatial unit set are divided into reachable and effective observation regions; spatial units that belong to the reachable spatial unit set but not to the effective observation spatial unit set are divided into reachable but ineffective observation regions; spatial units that belong to both the inaccessible spatial unit set and the effective observation spatial unit set are divided into inaccessible but effective observation regions; spatial units that belong to both the inaccessible spatial unit set and the ineffective observation spatial unit set are divided into inaccessible and ineffective observation regions, thus obtaining the region division results.

9. The method for modeling a three-dimensional patrolable domain space of a substation according to claim 8, characterized in that, Based on the entity objects and dynamic spatial constraints associated with the region division results, a three-dimensional patrolable spatial situation map is constructed on the three-dimensional scene model. The three-dimensional patrolable spatial situation map is then updated based on changes in entity objects or relationships between entity objects, including: When changes in the device's open / closed state, device's operating state, energized state, work state, work range, the spatial location of the inspection target entity, the observation parameter conditions, or the spatial location of the obstacle entity in the multimodal dynamic knowledge graph cause changes in dynamic spatial constraints, the corresponding dynamic spatial constraints are updated according to the changed entity objects and the relationships between entity objects, and the range of dynamic spatial constraints before and after the update is compared. For entity objects whose scope of dynamic spatial constraints changes before and after the update, the region of difference in scope of dynamic spatial constraints before and after the change is defined as the spatial constraint change region. Based on the spatial intersection relationship between the spatial constraint change region and the spatial unit, an initial set of affected spatial units is determined, and the connectivity and reachability of each spatial unit in the initial set of affected spatial units are re-determined. When the reachability state of a spatial unit in the initial set of affected spatial units changes and affects the connectivity of adjacent spatial units, the reachability of the affected spatial units continues to be determined until the reachability state of the affected spatial units no longer changes. When the spatial location, observation parameters, or spatial location and enclosing range of the target entity change, effective observation is re-determined within the candidate observation spatial unit associated with the target entity based on the changed field of view constraints and target observation constraints. Based on the updated accessibility determination results and effective observation determination results, the reachable and effective observation areas, the reachable but ineffective observation areas, the inaccessible but effective observation areas, and the inaccessible but ineffective observation areas are redefined, and the corresponding areas in the three-dimensional patrolable spatial situation map are updated.

10. A three-dimensional, inspectable spatial modeling system for substations, characterized in that, The method for implementing the three-dimensional patrolable spatial modeling method for substations according to any one of claims 1 to 9 includes: The data acquisition module is used to acquire spatial geometric data, business semantic data, multimodal sensing data, and real-time status data of the substation. The spatial model construction module is used to construct a three-dimensional scene model within the substation based on the spatial geometric data and to spatially discretize it to obtain a set of spatial units. The graph construction module is used to construct a multimodal dynamic knowledge graph based on the business semantic data, multimodal perception data and real-time status data. The multimodal dynamic knowledge graph includes entity objects and the relationships between entity objects. The constraint construction module is used to associate the entity object and its corresponding semantic attributes with the corresponding spatial unit, and generate dynamic spatial constraints based on the current state of the entity object and preset safety procedures and on-site operation rules, which are then applied to the corresponding spatial unit. The region division module is used to determine the reachability of spatial units based on the connectivity of spatial units, preset static spatial constraints and the dynamic spatial constraints, to obtain a set of reachable spatial units. Based on the set of reachable spatial units, it determines the effective observation based on preset field of view constraints and observation parameter conditions corresponding to the inspection target, to obtain a set of effective observation spatial units. Based on the spatial units, it divides the observation area to obtain the region division result. The dynamic update module is used to construct a three-dimensional patrolable spatial situation map on the three-dimensional scene model based on the entity objects associated with the region division results and dynamic spatial constraints, and to update the three-dimensional patrolable spatial situation map based on the changes in entity objects or the relationships between entity objects.

Images