Narrow-space-oriented intelligent planning and conflict detection method for maintenance path of gas turbine unit

Abstract

The invention discloses a gas unit maintenance path intelligent planning and conflict detection method for a narrow space, and belongs to the technical field of intelligent maintenance. According to the method, on the basis of a three-dimensional digital model of the gas unit and maintenance entity information, a topological self-adaptive routing algorithm of the gas unit is adopted, the component density, the space narrow degree and the maintenance entity motion characteristics of the gas unit are comprehensively considered, a safe and efficient maintenance path is planned, and optimization selection is carried out through multi-dimensional cost evaluation. And accurate conflict early warning is realized through multi-level collision detection. According to the method, a multi-dimensional priority evaluation model is introduced, factors such as component criticality, fault emergency degree, operation risk and component aging are integrated, the maintenance task priority is dynamically adjusted, and path planning is optimized accordingly. According to the method, the safety and efficiency of gas unit overhaul are remarkably improved, the collision risk is effectively avoided, resource allocation is optimized, the downtime is shortened, and intelligentization and benefit maximization of gas unit overhaul are achieved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of intelligent maintenance, and particularly relates to a method for intelligent planning of maintenance paths and conflict detection of gas turbine units in narrow spaces. BACKGROUND

[0002] As the core equipment of power production and industrial driving, the safe and stable operation of gas turbine units is of great importance. In order to ensure the long-term and efficient operation of gas turbine units, regular and emergency maintenance work is essential. However, the internal structure of gas turbine units is highly complex, with densely arranged components, and often accompanied by dangerous or difficult-to-reach areas such as high temperature, high pressure, and narrow space, which poses a serious challenge to traditional maintenance.

[0003] Currently, the maintenance work of gas turbine units still largely relies on manual operation. Manual maintenance not only faces high personal safety risks, low work efficiency, and unstable maintenance quality due to human factors, but also in extremely narrow or high-risk areas, maintenance personnel are difficult to enter or stay for a long time, which easily causes missed inspection or failure to effectively perform maintenance tasks. In order to improve the efficiency and safety of maintenance, the industry has begun to explore the use of automated or semi-automated equipment, such as maintenance robots, to assist in maintenance. However, existing technologies for robot navigation and path planning are mostly suitable for relatively open or structured environments. In the special environment of gas turbine units, which has a highly complex three-dimensional structure, dense obstacles, and extremely narrow space, traditional path planning algorithms, such as those based solely on Euclidean distance or simple obstacle avoidance principles, often fail to generate efficient, safe, and practical maintenance paths. These algorithms usually fail to fully consider the actual size and kinematic constraints of the maintenance entity, as well as the fragility of the components inside the gas turbine unit and the specific safety distance requirements. In addition, existing conflict detection methods may be too rough to accurately identify potential collisions between multiple entities and components in narrow spaces, or can only provide passive conflict warnings rather than actively avoiding risks during the path planning stage. Furthermore, there is a lack of a dynamic priority allocation mechanism that can consider multiple factors such as component criticality, fault urgency, operation risk, resource consumption, and component aging and degradation state for maintenance of gas turbine units, resulting in the inability to intelligently optimize maintenance paths when multiple tasks are parallel or emergency faults occur, thereby affecting overall maintenance efficiency and downtime and increasing operating costs. Therefore, there is an urgent need for an advanced method that can adapt to the characteristics of narrow and complex spaces in gas turbine units, achieve intelligent path planning and fine conflict detection, and dynamically adjust the priority of maintenance tasks. SUMMARY

[0004] In view of the above problems existing in the prior art, the present application provides a method for intelligent planning of maintenance paths and conflict detection of gas turbine units in narrow spaces, comprising the following steps: Step S1, a three-dimensional digital model of the gas turbine unit is constructed to obtain structural data of the gas turbine unit in a narrow space, including geometric and physical models of components of the gas turbine unit such as a gas turbine, a compressor, a combustion chamber, a waste heat boiler, a steam turbine, a generator, auxiliary pipelines, valves, support structures, and a maintenance platform, and the geometric and physical models are integrated to construct a three-dimensional digital model of the gas turbine unit.

[0005] Step S2, a maintenance task and a maintenance entity are defined, at least one of the maintenance entity includes a maintenance personnel, a maintenance tool, and a maintenance robot is determined according to a type of the maintenance task to be performed, and a three-dimensional envelope geometric model and kinematic constraint parameters of each of the maintenance entity are obtained.

[0006] Step S3, an intelligent planning of a maintenance path is performed, an initial maintenance path from a starting maintenance point to a target maintenance point is planned based on a topological adaptive pathfinding algorithm of the gas turbine unit according to the three-dimensional digital model of the gas turbine unit, the type of the maintenance task, the three-dimensional envelope geometric model of the maintenance entity, and the kinematic constraint parameters of the maintenance entity.

[0007] Step S4, a path cost optimization and selection is performed, a multi-dimensional cost evaluation is performed on the initial maintenance path, at least one of cost evaluation factors includes a path length cost, a maintenance safety cost, and an energy consumption cost, and an optimal recommended maintenance path is selected according to a preset path optimization objective function.

[0008] Step S5, a collision detection and early warning of the maintenance path is performed, a real-time or quasi-real-time collision detection is performed on the recommended maintenance path, whether a spatial conflict occurs between the maintenance entity and components of the gas turbine unit, or other static obstacles, or other dynamic maintenance entities during movement along the recommended maintenance path is determined, and a conflict early warning information is sent according to a detection result.

[0009] Further, the step S1 specifically includes the following steps: Step S201, spatial coordinate data, geometric size data, and material attribute data of components of the gas turbine unit in a narrow space are obtained through a three-dimensional laser scanner, a high-precision point cloud data acquisition device, or design drawings and structural drawings of the gas turbine unit.

[0010] Step S202, a three-dimensional entity modeling is performed on the components of the gas turbine unit according to the spatial coordinate data, the geometric size data, and the material attribute data to obtain geometric and physical models of the components of the gas turbine unit.

[0011] Step S203, the geometric and physical models of the components of the gas turbine unit are integrated in the same three-dimensional coordinate system to construct a three-dimensional digital model of the gas turbine unit including all key components of the gas turbine unit and gaps therebetween, and the three-dimensional digital model of the gas turbine unit defines passable areas, forbidden areas, and maintenance work surfaces.

[0012] Further, the step S3 specifically comprises the following steps: Step S301, gas turbine unit maintenance space map construction, according to the three-dimensional digital model of the gas turbine unit, the key maintenance nodes and space reachable points in the passable area are extracted, and the passable connection relationship between the key maintenance nodes and the space reachable points is established to form a gas turbine unit maintenance space map.

[0013] Step S302, path search process initialization and iteration selection, on the gas turbine unit maintenance space map, the path search process is initialized, the starting maintenance point is taken as the first node to be explored, and the node with the minimum current comprehensive optimal value is selected as the current exploration node from the set of nodes to be explored.

[0014] Step S303, adjacent reachable node identification and preliminary screening, all adjacent reachable nodes of the current exploration node are identified; for each adjacent reachable node, it is detected in advance whether the maintenance entity will have potential collision with the gas turbine unit components or the preset static obstacles when passing through the path segment connecting the current exploration node and the adjacent reachable node; and the path segment with potential collision is excluded.

[0015] Step S304, path segment adaptive optimal value calculation, for the adjacent reachable nodes that have not been excluded, the comprehensive optimal value of the path segment connecting the current exploration node and the adjacent reachable node is calculated, which comprehensively considers the length of the path segment, the gas turbine unit component density when the maintenance entity passes through, the space narrowness, and the energy consumption evaluation factor of the maintenance entity passing through the path segment, and combines the spatial distance from the adjacent reachable node to the target maintenance point and the estimated gas turbine unit component safety evaluation factor along the way.

[0016] Step S305, set of nodes to be explored updating and path generation, according to the comprehensive optimal value, the state of the adjacent reachable node in the set of nodes to be explored is updated, and the iteration is continued until the target maintenance point is explored, and finally the initial maintenance path is formed.

[0017] Further, in the step S304, the path segment comprehensive optimal value calculated by the gas turbine unit topology adaptive pathfinding algorithm is calculated as follows: , wherein represents the actual path cost accumulated from the starting maintenance point to the current path node; represents the estimated optimal cost from the current path node to the target maintenance point; the calculation formula of the is as follows: , wherein represents the number of path segments passed through from the starting maintenance point to the current path node; representing the length of the segment path; representing the gas turbine unit component safety assessment factor of the segment path; representing the obstacle proximity penalty factor of the segment path; representing the energy consumption assessment factor of the segment path; representing the weight coefficient of the gas turbine unit component safety assessment factor; representing the weight coefficient of the obstacle proximity penalty factor; representing the weight coefficient of the energy consumption assessment factor.

[0018] Further, the step S4 comprises: Step S401, multi-objective optimization evaluation, taking the path length cost, maintenance safety cost and energy consumption cost as the input of the multi-objective optimization function, and comprehensively evaluating the initial maintenance path through weight distribution.

[0019] Step S402, recommended maintenance path selection, selecting the recommended maintenance path meeting the preset safety threshold and optimization target from the multi-objective optimization evaluation result, the recommended maintenance path minimizing the maintenance time or resource consumption under the premise of ensuring maintenance safety.

[0020] Further, the step S5 specifically comprises: Step S501, coarse-grained collision detection based on envelope, judging whether there is an intersection area between the three-dimensional envelope geometric model of the maintenance entity and the three-dimensional envelope of the gas turbine unit component, the three-dimensional envelope of the static obstacle and the three-dimensional envelope of the dynamic maintenance entity at each key path node or time step of the recommended maintenance path.

[0021] Step S502, accurate collision detection based on refined grid, if the coarse-grained collision detection based on envelope judges that there is an intersection area, then the refined grid model collision detection of the maintenance entity and the gas turbine unit component or the obstacle is performed in the intersection area to determine whether there is actual geometric interference.

[0022] Step S503, conflict warning and feedback, if the refined grid model collision detection judges that there is actual geometric interference, then the maintenance path conflict warning information is sent out, the maintenance path conflict warning information including the conflict occurrence position, the conflict entity, the conflict severity, and suggesting to perform maintenance path re-planning.

[0023] Further, the conflict warning information further comprises conflict estimation time and potential damage risk assessment.

[0024] Further, the method further comprises: maintenance task priority calculation and path optimization adjustment, before intelligent planning of the maintenance path, priority is allocated to a plurality of to-be-executed gas unit maintenance tasks, and the weight of the cost evaluation factor or the path optimization objective function is adjusted according to the maintenance task priority, so as to affect the generation of the recommended maintenance path.

[0025] Further, the maintenance task priority calculation method calculates the maintenance task priority value through the following multi-dimensional priority evaluation model: The multi-dimensional priority evaluation model comprehensively considers at least three influencing factors and dynamically weights and non-linearly comprehensively calculates the following factors: operation emergency and component criticality comprehensive factor, which reflects the importance of the to-be-maintained gas unit component in the entire gas unit operation, and is non-linearly enhanced in combination with the emergency level of the fault.

[0026] Operation risk and environmental proximity comprehensive factor, which evaluates the safety or environmental risk possibly brought in the maintenance task execution process, and adjusts the risk weight according to the proximity of the operation area to the high-value or fragile components in the gas unit.

[0027] Resource timeliness and downtime constraint factor, which comprehensively considers the resource input degree such as manpower and special maintenance tools required by the maintenance task, and is inversely proportional to the available downtime window.

[0028] Component health status and aging degradation factor, which is based on the operation of the component since the last maintenance time, and is non-linearly accumulated with time when the preset safe operation time threshold is exceeded, the maintenance task priority value is calculated by giving different weights to the above influencing factors and using adaptive weighted summation or multi-criteria decision analysis method, to ensure that tasks with high risk, high emergency or facing strict time limit obtain higher priority.

[0029] Further, according to the maintenance task priority adjustment path optimization objective function and / or the weight of the cost evaluation factor, if the maintenance task priority value is higher, the weight of the maintenance safety cost and the path length cost in the path optimization objective function is increased in the path cost optimization and selection process, so as to preferentially select a shorter and safer maintenance path, thereby shortening the maintenance time and meeting the emergency requirement of high-priority tasks.

[0030] Compared with the prior art, the present application has the following advantages: The present application significantly improves the safety and efficiency of gas turbine unit maintenance operations in narrow and complex spaces. Based on a refined three-dimensional model of the gas turbine unit and a topology adaptive pathfinding algorithm, the present application deeply integrates maintenance entity size, kinematic characteristics, and safety requirements of internal components of the gas turbine unit to plan a safe, efficient, and low-energy recommended maintenance path. This path planning goes beyond existing simple obstacle avoidance, optimizes adaptability to complex topological structures of the gas turbine unit, and combines multi-level collision detection to provide real-time early warning of potential conflicts, significantly reducing operational risks and equipment damage, and effectively addressing the safety and feasibility deficiencies of existing technologies in complex and narrow space operations.

[0031] The present application realizes intelligent management and resource optimization of gas turbine unit maintenance tasks. The innovative multi-dimensional priority evaluation model comprehensively considers factors such as component criticality, fault emergency level, operation risk, resource timeliness, and component aging degradation, performs dynamic weighting and nonlinear comprehensive calculation, and assigns adaptive priority. This mechanism effectively addresses the complexity and urgency of gas power plant maintenance. Based on this priority, the present application dynamically adjusts the path planning target weight, prioritizes generating shorter and safer maintenance paths for high-priority tasks, maximizes downtime reduction and operational loss reduction, ensures efficient execution of critical tasks, and significantly improves maintenance intelligence level and economic benefits. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 An exemplary step flowchart for the method of the present application.

[0033] Figure 2 An exemplary step flowchart for constructing a three-dimensional model of the present application.

[0034] Figure 3 An exemplary step flowchart for maintenance path planning of the present application.

[0035] Figure 4 An exemplary step flowchart for path cost optimization of the present application.

[0036] Figure 5 An exemplary step flowchart for maintenance path conflict detection of the present application. DETAILED DESCRIPTION

[0037] The present application will be further described below in conjunction with specific embodiments.

[0038] This application provides an intelligent planning and conflict detection system for gas turbine maintenance paths in confined spaces. This system aims to significantly improve the safety, efficiency, and accuracy of maintenance operations by intelligently planning, optimizing, and detecting conflicts in real time during gas turbine maintenance. It is particularly suitable for complex environments with limited space. This system effectively addresses the inefficiency, collision-proneness, and safety challenges of traditional manually planned paths during maintenance operations within compact and complex gas turbines.

[0039] like Figure 1 As shown, the method for intelligent planning and conflict detection of gas unit maintenance paths in narrow spaces is characterized by comprising the following steps: Step S1: construct a three-dimensional digital model of the gas turbine unit. The structural data of the gas turbine unit in a narrow space is obtained, including the geometric and physical models of the gas turbine unit components such as the gas turbine, compressor, combustion chamber, waste heat boiler, steam turbine, generator, auxiliary pipelines, valves, support structure and maintenance platform, and the geometric and physical models are integrated into a three-dimensional digital model of the gas turbine unit.

[0040] In one embodiment, this step forms the foundation of the entire method, providing a precise digital blueprint for maintenance path planning and collision detection. By constructing a detailed 3D model of the gas turbine, the system accurately identifies the geometry, physical properties, and spatial relationships of all components.

[0041] like Figure 2 FIG. 1 is a flowchart of an exemplary step of constructing a three-dimensional model in step S1 of this embodiment, including the following steps: Step S201 , obtaining spatial coordinate data, geometric dimension data, and material attribute data of gas turbine components in a narrow space through a 3D laser scanner, high-precision point cloud data acquisition equipment, or gas turbine design drawings and structural drawings.

[0042] In one embodiment, this step is intended to obtain the raw data required to build a high-precision three-dimensional model from multiple sources of data. Three-dimensional laser scanners and high-precision point cloud data acquisition equipment are capable of acquiring accurate three-dimensional point cloud data of actual physical objects. These data contain high-density spatial coordinate information. For example, ground-based laser scanners, handheld three-dimensional scanners, or drone-mounted lidars can be used for data acquisition. The design drawings and structural drawings of the gas unit provide the theoretical geometric dimensions, material properties, and assembly relationships of the components. Material property data, such as density, elastic modulus, etc., may be crucial for subsequent physical simulation or energy consumption evaluation. By integrating data from these different sources, the comprehensiveness and accuracy of the model can be ensured.

[0043] Step S202, according to the spatial coordinate data, geometric size data and material attribute data, the three-dimensional entity modeling of the gas turbine unit components is carried out to obtain the geometric physical model of the gas turbine unit components. In an embodiment, this step is to convert the original data obtained in S201 into a computer- processable three-dimensional entity model. This usually involves gridding, surface reconstruction of point cloud data, or lifting two-dimensional information in design drawings to three-dimensional entities. The goal of modeling is to create an accurate geometric physical model for each gas turbine unit component, which can accurately reflect its actual shape, size and physical properties in three-dimensional space.

[0044] Step S203, the geometric physical model of the gas turbine unit components is integrated in the same three-dimensional coordinate system to build a three-dimensional digitized model of the gas turbine unit containing all key components of the gas turbine unit and the gaps therebetween, and the three-dimensional digitized model of the gas turbine unit defines passable areas, forbidden areas and maintenance work surfaces.

[0045] In an embodiment, this step is to assemble and integrate the individual gas turbine unit components independently modeled in S202 according to their real assembly relationship in a unified three-dimensional coordinate system. This forms a complete three-dimensional digitized model of the gas turbine unit that can be used for subsequent path planning and conflict detection. In this integrated model, the system further defines passable areas, forbidden areas and maintenance work surfaces. The definition of these areas is crucial for intelligent pathfinding and collision avoidance, and they serve as constraints and decision-making basis for path planning algorithms.

[0046] Step S2, maintenance task and maintenance entity definition, according to the type of maintenance task to be performed, the maintenance entity required for the maintenance work is determined, the maintenance entity includes at least one of maintenance personnel, maintenance tools and maintenance robots, and the three-dimensional envelope geometric model and kinematic constraint parameters of each maintenance entity are obtained.

[0047] In an embodiment, it involves identifying the personnel, tools or robots required to perform a specific maintenance task, and establishing their representation and motion capability in the virtual environment. Specifically, the system intelligently determines or specifies by the user the maintenance entity required to complete the task according to the specific type of maintenance task. These entities can include but are not limited to: maintenance personnel, various manual or electric maintenance tools, and automated maintenance robots. For each maintenance entity, the system will obtain or construct its three-dimensional envelope geometric model and set its kinematic constraint parameters.

[0048] Step S3, intelligent planning of maintenance path, according to the three-dimensional digitized model of the gas turbine unit, the type of maintenance task, the three-dimensional envelope geometric model of the maintenance entity and the kinematic constraint parameters of the maintenance entity, based on the gas turbine unit topology adaptive pathfinding algorithm, the initial maintenance path from the starting maintenance point to the target maintenance point is planned.

[0049] As Figure 3 An exemplary step flow chart of the path planning in step S3 is shown, including the following steps: Step S301, gas turbine unit maintenance space map construction, according to the three-dimensional digital model of the gas turbine unit, the key maintenance nodes and space reachable points in the passable area are extracted, and the passable connection relationship between the key maintenance nodes and space reachable points is established, forming the gas turbine unit maintenance space map.

[0050] In one embodiment, the maintenance space map is an abstract representation of the complex topology structure inside the gas turbine unit, which converts the passable area in three-dimensional space into discrete nodes and connection edges. The system will automatically identify or manually annotate all key maintenance nodes and space reachable points that can be passed by the maintenance entity according to the three-dimensional digital model of the gas turbine unit constructed in S1. These nodes and reachable points can be predefined, or automatically generated through space discretization, skeleton extraction or visual graph construction algorithm. Then, the system will evaluate whether there is an unobstructed connection path between these nodes and reachable points, and establish the passable connection relationship between them, for example, through AABB or OBB collision detection. These connection relationships form a network graph composed of nodes and edges, effectively simplifying the path search problem in three-dimensional space, so that it can be efficiently solved under the framework of graph theory.

[0051] Step S302, path search process initialization and iteration selection, on the gas turbine unit maintenance space map, initialize the path search process, take the starting maintenance point as the first node to be explored, and iteratively select the node with the smallest current comprehensive optimal value from the set of nodes to be explored as the current exploration node.

[0052] Step S303, adjacent reachable node identification and preliminary screening, identify all adjacent reachable nodes of the current exploration node; for each adjacent reachable node, pre-detect whether the maintenance entity will have potential collision with the gas turbine unit components or the preset static obstacles when passing through the path segment connecting the current exploration node and the adjacent reachable node; and exclude the path segment with potential collision.

[0053] In one embodiment, after the system selects a current exploration node, this step identifies all adjacent reachable nodes directly connected to the node, and the connection information is defined in the spatial graph in S301. The system performs preliminary pre-collision detection for the path segment connecting the current exploration node and each adjacent reachable node. This preliminary detection usually uses coarse-grained bounding box collision detection to quickly determine whether the maintenance entity will potentially collide with the gas turbine unit components or the preset static obstacles when passing through the path segment. Any path segment detected as potentially colliding is immediately excluded to avoid wasting computational resources on these infeasible paths and ensure that the subsequent search path segments are initially collision-free, thereby significantly improving search efficiency.

[0054] Step S304, path segment adaptive preference value calculation: for the adjacent reachable nodes that are not excluded, the comprehensive preference value of the path segment connecting the current exploration node and the adjacent reachable node is calculated, which comprehensively considers the length of the path segment, the gas turbine unit component density when passing through, the space narrowness, and the energy consumption evaluation factor of the maintenance entity passing through the path segment, and combines the spatial distance from the adjacent reachable node to the target maintenance point and the estimated gas turbine unit component safety evaluation factor along the way.

[0055] In step S304, the calculation method of the path segment comprehensive preference value calculated by the gas turbine unit topology adaptive pathfinding algorithm is as follows: wherein, represents the actual path cost accumulated from the starting maintenance point to the current path node; represents the estimated optimization cost from the current path node to the target maintenance point; The calculation formula of is as follows: wherein represents the number of path segments passed through from the starting maintenance point to the current path node; represents the length of the th path segment; represents the gas turbine unit component safety evaluation factor of the th path segment; represents the obstacle proximity penalty factor of the th path segment; represents the energy consumption evaluation factor of the th path segment; represents the weight coefficient of the gas turbine unit component safety evaluation factor; represents the weight coefficient of the obstacle proximity penalty factor; represents the weight coefficient of the energy consumption evaluation factor.

[0056] Step S305, the update of the set of nodes to be explored and the generation of the path, according to the comprehensive preference value, updating the state of the adjacent reachable nodes in the set of nodes to be explored, and continuously iterating until the target maintenance point is explored, and finally forming the initial maintenance path.

[0057] In one embodiment, this step is the core of the iteration of the path search algorithm. According to the comprehensive preference value P calculated in S304, the system updates the state of the adjacent reachable nodes in the set of nodes to be explored. The search process continues to iterate until any of the following conditions is met: the target maintenance point is explored, or the set of nodes to be explored is empty. Once the target maintenance point is explored, the system traces the path back from the target point to the starting point, thereby generating the final initial maintenance path. This process ensures that the pathfinding algorithm can find an initial path that is conflict-free in the preliminary collision detection and meets the multi-objective constraints.

[0058] Step S4, path cost optimization and selection, performing multi-dimensional cost evaluation on the initial maintenance path, the cost evaluation factors including at least one of path length cost, maintenance safety cost, and energy consumption cost, and selecting the optimal recommended maintenance path according to the preset path optimization objective function.

[0059] As Figure 4 The following is an exemplary step flowchart of path cost optimization in step S4 of the embodiment, including: Step S401, multi-objective optimization evaluation, taking the path length cost, the maintenance safety cost, and the energy consumption cost as inputs of a multi-objective optimization function, and performing comprehensive evaluation on the initial maintenance path through weight distribution.

[0060] In one embodiment, this step takes the initial maintenance path generated in S3 as input and performs comprehensive multi-objective evaluation on it. The path length cost usually refers to the total length of the physical path and is a basic indicator of efficiency. The maintenance safety cost may include the degree of passing through dangerous areas, the minimum distance from obstacles, operation complexity, and potential risks to machine components, aiming to ensure the safety of personnel and equipment. The energy consumption cost evaluates the energy required by the maintenance entity during movement along the path, aiming to reduce operating costs. These cost factors are considered as inputs of a multi-objective optimization function. The system performs comprehensive evaluation on these costs through preset weight distribution, generates a comprehensive path evaluation score, and provides a quantitative basis for path selection.

[0061] Step S402, recommended maintenance path selection, selecting the recommended maintenance path that meets the preset safety threshold and optimization objective from the multi-objective optimization evaluation results, the recommended maintenance path minimizing the maintenance time or resource consumption under the premise of ensuring maintenance safety.

[0062] In one embodiment, this step is to select the recommended maintenance path that best meets the preset safety threshold and optimization target from the multi-objective optimization evaluation results of S401. The system will iterate through all evaluated paths or optimize a single path until it meets all safety constraints and achieves the goal of minimizing maintenance time or resource consumption in terms of overall cost. This process may involve post-processing such as path smoothing, redundant joint optimization, etc. to ensure that the finally selected recommended maintenance path is not only safe but also highly feasible, efficient and economical in actual operation.

[0063] Step S5, maintenance path conflict detection and warning, real-time or quasi-real-time collision detection of the recommended maintenance path, to determine whether the maintenance entity will collide with the gas turbine components, or other static obstacles, or other dynamic maintenance entities during movement along the recommended maintenance path, and issue a conflict warning message based on the detection results.

[0064] As Figure 5 The following is an exemplary step flowchart of maintenance path conflict detection in step S5 of the present embodiment, which includes: Step S501, coarse-grained collision detection based on envelope, at each key path node or time step of the recommended maintenance path, determine whether there is an intersection region between the three-dimensional envelope geometric model of the maintenance entity and the three-dimensional envelope of the gas turbine components, the three-dimensional envelope of the static obstacles, and the three-dimensional envelope of the dynamic maintenance entities.

[0065] Step S502, precise collision detection based on refined mesh, if the coarse-grained collision detection based on envelope determines that there is an intersection region, then perform refined mesh model collision detection of the maintenance entity and the gas turbine components or obstacles in the intersection region to determine whether there is actual geometric interference.

[0066] In one embodiment, if the coarse-grained collision detection of S501 determines that there is an intersection region, it indicates that there may be a real collision in that region. At this time, the system will focus on this specific intersection region and start higher-precision collision detection. This involves using the refined mesh model of the maintenance entity, gas turbine components or obstacles. The precise collision detection algorithm will directly calculate on these refined mesh data to determine whether there is actual geometric interference. This two-stage collision detection strategy can balance efficiency and accuracy, avoiding high-cost precise detection at all time steps while ensuring the discovery of real collisions.

[0067] Step S503, conflict warning and feedback, if the refined mesh model collision detection determines that there is actual geometric interference, issue a maintenance path conflict warning message, which includes the conflict location, conflict entity, conflict severity, and suggests re-planning the maintenance path.

[0068] In one embodiment, if the refined mesh model collision detection of S502 confirms the existence of actual geometric interference, the system will immediately issue a maintenance path conflict warning information. The warning information is detailed, intuitive and instructive. Exemplarily, it can include: the precise three-dimensional position of the conflict: for example, the coordinate point or the involved components. The conflict entity: clearly indicates which maintenance entity has a conflict with which obstacle or component. Conflict severity: quantifies the severity level of the conflict according to the collision depth, the vulnerability of the involved components, the motion speed, etc. Conflict prediction time: predicts when the actual collision will occur if the current trend continues, providing a time window for the operator to intervene in advance. More importantly, the system will intelligently propose avoidance suggestions based on the type and severity of the conflict, such as suggesting that the operator pause the operation, adjust the motion speed, retreat, fine-tune the maintenance entity pose, or directly trigger the maintenance path re-planning process. These suggestions aim to guide the operator to take timely measures to avoid or mitigate actual equipment damage or personnel injury, thereby ensuring job safety.

[0069] In this embodiment, the conflict warning information also includes the conflict prediction time and the potential damage risk assessment. In one embodiment, in addition to real-time conflict detection, the system also has the ability to predict conflicts and assess risks. The conflict prediction time refers to predicting when a collision will occur with an obstacle based on the current maintenance entity motion trend and path. This provides a time window for the operator to intervene in advance, allowing them to respond more calmly. The potential damage risk assessment quantitatively assesses the degree of equipment damage that may be caused according to the type of components involved in the conflict, the collision speed, and the estimated collision depth, so that the operator can understand the severity of the risk and take appropriate measures, such as outputting "moderate damage risk, may affect sensor function".

[0070] The method of this embodiment also includes maintenance task priority calculation and path optimization adjustment. Before intelligent planning of the maintenance path, the priority of multiple gas turbine unit maintenance tasks to be executed is assigned, and the weight of the path optimization objective function or cost evaluation factor is adjusted according to the maintenance task priority, thereby affecting the generation of the recommended maintenance path.

[0071] The maintenance task priority calculation method calculates the maintenance task priority value through the following multi-dimensional priority evaluation model: The multi-dimensional priority evaluation model considers at least three influencing factors and performs dynamic weighting and nonlinear comprehensive calculation on them, The operation urgency and component criticality comprehensive factor reflects the importance of the gas turbine unit component to be maintained in the entire gas turbine unit operation, and combines the nonlinear enhancement of the fault emergency level.

[0072] A job risk and environment proximity factor, which evaluates the safety or environmental risks that may arise during the execution of the maintenance task and adjusts the risk weight according to the proximity of the job area to high-value or fragile components within the gas unit.

[0073] A resource time effectiveness and downtime constraint factor, which synthesizes the degree of resource input required for the maintenance task, such as manpower and special maintenance tools, and is inversely proportional to the available downtime window.

[0074] A component health status and aging degradation factor, which is based on the operation of the component since the last maintenance, and accumulates non-linearly with time when the pre-set safe operation time threshold is exceeded. The maintenance task priority value is calculated by assigning different weights to the above-mentioned factors and using adaptive weighted summation or multi-criteria decision analysis methods to ensure that tasks with high risk, high urgency, or strict time constraints receive higher priority.

[0075] The weights of the path optimization objective function and / or cost evaluation factors are adjusted according to the maintenance task priority as follows: If the maintenance task priority value is higher, the weights of the maintenance safety cost and path length cost in the path optimization objective function are increased during the path cost optimization and selection process, to preferentially select shorter and safer maintenance paths, thereby shortening the maintenance time and meeting the urgency requirements of high-priority tasks.

[0076] In one embodiment, the maintenance task priority and path planning process are highly linked. If the maintenance task priority value is higher, the system dynamically increases the weights of the maintenance safety cost and the path length cost in the path optimization objective function during the path cost optimization and selection process in S4. This means that for high-priority tasks, the system will prefer paths that are shorter and safer, even if it may mean making slight compromises on other costs. Through this adaptive weight adjustment mechanism, the system can ensure that the optimal maintenance path that best meets the urgency requirements of high-priority tasks is generated, thereby improving overall maintenance efficiency and safety.

[0077] Those skilled in the art will understand that the above embodiments are only exemplary, and various modifications and equivalent replacements can be made without departing from the spirit and scope of the present application. For example, specific feature point algorithms, optimizer selection, distortion model details, etc. can be adjusted according to actual needs. Therefore, the scope of protection of the present application should be defined by the appended claims.

Claims

1. An intelligent planning and conflict detection method for gas unit maintenance paths in narrow spaces, characterized by: The following steps are involved: Step S1: constructing a 3D digital model of the gas turbine unit. Structural data of the gas turbine unit within a narrow space is obtained, including geometric and physical models of gas turbine components such as the gas turbine, compressor, combustion chamber, waste heat boiler, steam turbine, generator, auxiliary pipelines, valves, support structures, and maintenance platforms. These are then integrated into a 3D digital model of the gas turbine unit. Step S2: Define the maintenance task and maintenance entity. According to the type of maintenance task to be performed, determine the maintenance entity required for the maintenance operation. The maintenance entity includes at least one of a maintenance person, a maintenance tool, and a maintenance robot. Obtain the three-dimensional envelope geometric model and kinematic constraint parameters of each maintenance entity. Step S3, intelligent maintenance path planning, planning an initial maintenance path from the starting maintenance point to the target maintenance point based on the three-dimensional digital model of the gas unit, the maintenance task type, the three-dimensional envelope geometric model of the maintenance entity, and the kinematic constraint parameters of the maintenance entity, and based on the gas unit topology adaptive path finding algorithm; Step S4, path cost optimization and selection, performs a multi-dimensional cost evaluation on the initial maintenance path, where the cost evaluation factors include at least one of path length cost, maintenance safety cost, and energy consumption cost, and selects the optimal recommended maintenance path based on a preset path optimization objective function; Step S5, maintenance path conflict detection and warning, performs real-time or quasi-real-time collision detection on the recommended maintenance path, determines whether the maintenance entity has spatial conflicts with gas unit components, other static obstacles, or other dynamic maintenance entities while moving along the recommended maintenance path, and issues conflict warning information based on the detection results.

2. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The step S1 includes the following steps: Step S201: Acquire spatial coordinate data, geometric dimension data, and material attribute data of gas turbine components within a narrow space using a 3D laser scanner, high-precision point cloud data acquisition equipment, or gas turbine design drawings and structural drawings; Step S202, performing three-dimensional solid modeling on the gas turbine component according to the spatial coordinate data, geometric dimension data, and material attribute data to obtain a geometric physical model of the gas turbine component; In step S203, the geometric physical models of the gas turbine components are integrated in the same three-dimensional coordinate system to construct a three-dimensional digital model of the gas turbine that includes all key components of the gas turbine and their gaps. The three-dimensional digital model of the gas turbine defines passable areas, prohibited areas, and maintenance work surfaces.

3. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The step S3 includes the following steps: Step S301: constructing a gas turbine maintenance space map. Based on the three-dimensional digital model of the gas turbine, key maintenance nodes and spatially accessible points within the traversable area are extracted, and traversable connection relationships between the key maintenance nodes and spatially accessible points are established to form a gas turbine maintenance space map. Step S302: Initializing the path search process and iterative selection. On the gas turbine maintenance spatial map, the path search process is initialized, the starting maintenance point is used as the first node to be explored, and the node with the smallest current comprehensive optimal value is iteratively selected from the set of nodes to be explored as the current exploration node. Step S303: Identify and preliminarily screen adjacent reachable nodes to identify all adjacent reachable nodes of the currently explored node; for each adjacent reachable node, pre-check whether the maintenance entity has a potential collision with gas turbine components or preset static obstacles when passing through the path segment connecting the currently explored node and the adjacent reachable node; and exclude path segments that have a potential collision. Step S304: Calculate the optimal value of path segment adaptability. For the adjacent reachable nodes that are not excluded, calculate the comprehensive optimal value of the path segment connecting the current exploration node and the adjacent reachable node. The comprehensive optimal value comprehensively considers the length of the path segment, the density of gas turbine components when the maintenance entity passes through, the narrowness of the space, and the energy consumption assessment factor of the maintenance entity passing through the path segment, in addition to the spatial distance from the adjacent reachable node to the target maintenance point and the estimated safety assessment factor of the gas turbine components along the path. Step S305, updating the set of nodes to be explored and generating a path, updating the status of adjacent reachable nodes in the set of nodes to be explored according to the comprehensive optimal value, and continuously iterating until the target maintenance point is explored, and finally forming an initial maintenance path.

4. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 3 is characterized by: In step S304, the path segment comprehensive optimization value calculated by the gas unit topology adaptive routing algorithm The calculation method is: ,in, represents the actual path cost accumulated from the starting maintenance point to the current path node; represents the estimated optimization cost from the current path node to the target maintenance point; The calculation formula is: ,in Indicates the number of path segments from the starting inspection point to the current path node; Indicates the The length of the segment path; Indicates the Safety assessment factors of gas turbine components for segment paths; Indicates the The obstacle proximity penalty factor for the segment path; Indicates the Energy consumption evaluation factor of segment path; Represents the weight coefficient of the safety assessment factor of the gas unit components; Represents the weight coefficient of the obstacle approach penalty factor; Represents the weight coefficient of the energy consumption assessment factor.

5. The method for intelligent planning and conflict detection of maintenance paths for gas turbine units in narrow spaces according to claim 1 is characterized by: The step S4 includes: Step S401, multi-objective optimization evaluation, taking the path length cost, maintenance safety cost, and energy consumption cost as inputs of a multi-objective optimization function, and performing a comprehensive evaluation of the initial maintenance path through weight allocation; Step S402 , recommended maintenance path selection, selects a recommended maintenance path that meets the preset safety threshold and optimization target from the multi-objective optimization evaluation results, and the recommended maintenance path minimizes maintenance time or resource consumption while ensuring maintenance safety.

6. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The step S5 includes: Step S501: Based on coarse-grained collision detection of envelopes, at each critical path node or time step of the recommended maintenance path, it is determined whether there is an intersection area between the geometric model of the three-dimensional envelope of the maintenance entity and the three-dimensional envelope of the gas unit components, the three-dimensional envelope of the static obstacle, and the three-dimensional envelope of the dynamic maintenance entity; Step S502: performing precise collision detection based on the refined mesh. If the coarse-grained collision detection based on the envelope volume determines that an intersection exists, performing collision detection on the refined mesh model of the maintenance entity and the gas unit components or obstacles in the intersection to determine whether there is actual geometric interference. Step S503, conflict warning and feedback: if the refined grid model collision detection determines that there is actual geometric interference, a maintenance path conflict warning message is issued. The maintenance path conflict warning message includes the location of the conflict, the conflicting entity, and the severity of the conflict, and it is recommended to re-plan the maintenance path.

7. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The conflict warning information also includes an estimated conflict time and a potential damage risk assessment.

8. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: It also includes maintenance task priority calculation and path optimization adjustment. Before performing intelligent maintenance path planning, priorities are assigned to multiple gas unit maintenance tasks to be executed, and the path optimization objective function or the weight of the cost evaluation factor is adjusted according to the maintenance task priority, thereby affecting the generation of the recommended maintenance path.

9. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The maintenance task priority calculation method calculates the maintenance task priority value through the following multi-dimensional priority evaluation model: The multi-dimensional priority evaluation model comprehensively considers at least three of the following influencing factors and performs dynamic weighting and nonlinear comprehensive calculations on them: A comprehensive factor of operational urgency and component criticality, which reflects the importance of the component to be repaired in the operation of the entire gas turbine unit and is nonlinearly enhanced based on the urgency level of the fault; A comprehensive factor of operational risk and environmental proximity, which assesses the potential safety or environmental risks associated with the maintenance task and adjusts the risk weight based on the proximity of the work area to high-value or vulnerable components within the gas turbine unit; Resource timeliness and downtime constraint factor, which integrates the resource input required for the maintenance task, such as manpower and specialized maintenance tools, and is inversely proportional to the available downtime window; The component health status and aging degradation factor is based on the component's operation since the last maintenance. When the preset safe operation time threshold is exceeded, the factor shows a nonlinear cumulative growth over time. The maintenance task priority value is calculated by assigning different weights to the above-mentioned influencing factors and adopting adaptive weighted summation or multi-criteria decision analysis methods to ensure that high-risk, high-urgency or time-constrained tasks are given higher priority.

10. The method for intelligent planning and conflict detection of gas turbine maintenance paths in narrow spaces according to claim 1 is characterized by: The weights of the path optimization objective function and / or the cost evaluation factors are adjusted according to the priority of the maintenance task. If the priority value of the maintenance task is higher, the weights of the maintenance safety cost and the path length cost in the path optimization objective function are increased during the path cost optimization and selection process to give priority to shorter and safer maintenance paths, thereby shortening the maintenance time and meeting the urgency requirements of high-priority tasks.

Images