Dynamic path planning and real-time monitoring method for hydropower station GCB disassembly and assembly

Abstract

The invention discloses a dynamic path planning and real-time monitoring method for hydropower station GCB disassembly and assembly, and belongs to the field of hydropower station equipment overhaul, and the method comprises the steps: firstly carrying out initial condition setting and parameter initialization, and then generating a preliminary GCB disassembly and assembly path by adopting a fast exploration random tree RRT algorithm; and then track optimization is carried out on the generated disassembly and assembly path by using a constraint gradient optimization motion planning CHOMP algorithm, so that the disassembly and assembly efficiency and precision are improved. In the disassembly and assembly process, relevant parameters are dynamically adjusted through a real-time monitoring technology, and it is ensured that disassembly and assembly work is smoothly conducted according to plans; according to the method, the path exploration capability of the RRT algorithm and the trajectory optimization capability of the CHOMP algorithm are combined, dynamic planning and real-time monitoring of the GCB dismounting path are achieved, the safety and efficiency of dismounting operation are effectively improved, the operation risk and human errors are reduced, and the method is of great significance to maintenance of hydropower station equipment.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of hydropower station equipment maintenance, and particularly relates to a dynamic path planning and real-time monitoring method for GCB disassembly and assembly of a hydropower station. BACKGROUND

[0002] As a core component of the power equipment in a hydropower station, the GCB bears the key protection and control function. In the long-term operation process, the GCB may have faults such as insulation performance degradation due to environmental factors, and therefore needs to be maintained. However, the maintenance process usually needs to rely on manual hoisting operation, and the operator needs to have high environmental perception and equipment coordination ability to avoid potential collision risks, equipment damage and personal safety accidents.

[0003] In the prior art GCB disassembly and assembly, the operation risk and error possibility exist due to the reliance on manual experience and intuitive judgment. Meanwhile, due to the complexity and particularity of the GCB, the surrounding environment and equipment need to be paid attention to in the disassembly and assembly process to avoid accidents such as collision; therefore, a dynamic path planning and real-time monitoring method for GCB disassembly and assembly of a hydropower station is needed to solve the above problems. SUMMARY

[0004] The technical problem to be solved by the application is to provide a dynamic path planning and real-time monitoring method for GCB disassembly and assembly of a hydropower station, aiming to solve the problems of high operation risk and large error in the prior art GCB disassembly and assembly process relying on manual operation, and to have the characteristics of quickly and accurately generating and optimizing the disassembly and assembly path of the GCB, ensuring that the entire disassembly and assembly operation is strictly performed according to the predetermined plan, which is safe and efficient.

[0005] To solve the above technical problems, the technical solution adopted by the application is: A dynamic path planning and real-time monitoring method for GCB disassembly and assembly of a hydropower station, comprising the following steps: S1, initial condition setting and parameter initialization; S2, generating a preliminary GCB disassembly and assembly path by using an RRT algorithm; S3, optimizing the GCB disassembly and assembly trajectory by using a CHOMP algorithm; S4, real-time monitoring the GCB disassembly and assembly process and dynamically adjusting parameters; S5, outputting the adjusted algorithm parameters and trajectory optimization results.

[0006] Preferably, in step S1, the initial condition setting and parameter initialization comprises the following steps: S101: defining the disassembly and assembly starting point of the GCB workpiece and the target area . S102: configuring wall and hoisting equipment obstacle information in the work space ; S103: setting random sampling range of RRT algorithm , extension step , and maximum iteration number ; S104: setting smoothness cost weight , obstacle avoidance cost weight , iteration number , and gradient descent step parameters of CHOMP algorithm.

[0007] Preferably, the disassembly starting point of the GCB workpiece .

[0008] Preferably, the wall and hoisting equipment obstacle information .

[0009] Preferably, the random sampling range of the RRT algorithm is: .

[0010] Preferably, in step S2, generating the preliminary GCB disassembly path by using the RRT algorithm includes the following steps: S201: initializing a node set , generating a sampling point by randomly sampling in the work space using the RRT algorithm; S202: finding the nearest node to the sampling point by nearest neighbor search in , and trying to extend the step in the direction of the sampling point to generate a new node ; S203: checking whether the new node collides with any obstacle in . If collision occurs, the new node is abandoned and random sampling is performed again; if no collision occurs, it is added to the node set ; S204: repeating steps S201-203 until a preliminary path connecting the starting point and the target area is found or the preset iteration number is reached, and outputting the preliminary path .

[0011] Preferably, in step S3, optimizing the GCB disassembly trajectory by using the CHOMP algorithm includes the following steps: S301: Parameterize trajectory: parameterize the trajectory as a continuous function ; S302: Define cost function: include smoothness cost function and obstacle avoidance cost function , calculate total cost function ; S303: Calculate gradient: take derivative of with respect to , get gradient ; S304: Update trajectory using gradient descent method; S305: Collision detection and adjustment: after each update of the trajectory, check whether it collides with . If collision, adjust parameters and return to S303; S306: Repeat S303-S305 until the maximum number of iterations is reached or the cost function value no longer decreases significantly.

[0012] Preferably, in the smoothness cost function: ; ; ; wherein is a continuous function.

[0013] Preferably, updating the trajectory using the gradient descent method includes: ; wherein denotes the trajectory at the kth iteration.

[0014] Preferably, in step S4, the GCB disassembly process is monitored in real time, and the sampling range of the RRT algorithm and the cost function weight of the CHOMP algorithm and other parameters are dynamically adjusted, including the following steps: S401: Obtain position information, attitude information and velocity information of the GCB workpiece using sensors; S402: Calculate the deviation between the actual trajectory and the optimized trajectory ; S403: Dynamically adjust the sampling range of the RRT algorithm and the cost function weight of the CHOMP algorithm and other parameters according to the real-time monitoring data and error correction results.

[0015] The beneficial effects of the present application are as follows: ​This method overcomes the limitations of traditional algorithms by constructing a cascade reservoir optimal scheduling model based on graph neural networks and reinforcement learning. Specifically, it employs a graph neural network architecture, utilizing message passing neural networks (MPNN) to extract local relationships between reservoirs and capturing global relationships through the self-attention mechanism of the Transformer encoder, thus comprehensively modeling the complex spatiotemporal dependencies between cascade reservoirs. Furthermore, by combining a PPO reinforcement learning model, it can autonomously learn scheduling strategies in a dynamic environment, gradually approaching the global optimum. Compared with traditional methods, this method has significant advantages in computational efficiency, global optimum, and adaptability to complex scheduling scenarios, providing an innovative solution to the cascade reservoir optimal scheduling problem. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the flowchart of the present invention; Figure 2 This is a schematic diagram showing the effect of comparing the present invention with the prior art in an embodiment of the present invention. Detailed Implementation

[0017] Example 1: like Figure 1 As shown, a dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations includes the following steps: S1, Initial condition setting and parameter initialization; S2, using the RRT algorithm to generate the initial GCB assembly / disassembly path; S3, using the CHOMP algorithm to optimize the GCB disassembly and assembly trajectory; S4 monitors the GCB disassembly and assembly process in real time and dynamically adjusts parameters; S5 outputs the adjusted algorithm parameters and trajectory optimization results.

[0018] Example 2: Step S1, initial condition setting and parameter initialization, includes the following steps: S101: Define the start point for disassembling and assembling GCB workpieces. and target area ; S102: Configure obstacle information within the workspace, including walls and hoisting equipment. ; S103: Set the random sampling range for the RRT algorithm Expanding step size and maximum number of iterations ; S104: Set the smoothness cost weights for the CHOMP algorithm Obstacle avoidance cost weight Number of iterations and gradient descent step size parameters.

[0019] Preferably, the disassembly starting point of the GCB workpiece ; target region .

[0020] Preferably, the wall and hoisting equipment obstacle information .

[0021] Preferably, the random sampling range of the RRT algorithm is: .

[0022] Preferably, in step S2, generating the preliminary GCB disassembly path using the RRT algorithm includes the following steps: S201: initialize a node set , generate a sampling point using the RRT algorithm to randomly sample in the workspace ; S202: find the nearest node to the sampling point by nearest neighbor search in , and try to expand the step size in the direction of the sampling point to generate a new node ; S203: check whether the new node collides with any obstacle in . If collision, discard the new node and re-sample randomly; if no collision, add it to the node set ; S204: repeat steps S201-203 until a preliminary path connecting the starting point and the target region is found or the preset number of iterations is reached , output the preliminary path .

[0023] Preferably, in step S3, optimizing the GCB disassembly trajectory using the CHOMP algorithm includes the following steps: S301: parameterize the trajectory: parameterize as a continuous function ; S302: define the cost function: include the smoothness cost function and the obstacle avoidance cost function , calculate the total cost function ; S303: calculate the gradient: take the derivative of with respect to to get the gradient ; S304: updating the trajectory using gradient descent method; S305: collision detection and adjustment: after updating the trajectory each time, check whether it collides with the obstacles. If collision, adjust the parameters and return to S303; S306: repeat S303-S305 until the maximum number of iterations or the cost function value no longer decreases significantly is reached.

[0024] Preferably, in the smoothness cost function: ; ; ; wherein, is a continuous function.

[0025] Preferably, updating the trajectory using gradient descent method comprises: ; wherein, represents the trajectory at the kth iteration.

[0026] Preferably, in step S4, real-time monitoring the GCB disassembly process and dynamically adjusting the sampling range of the RRT algorithm and the cost function weight of the CHOMP algorithm and other parameters include the following steps: S401: obtaining the position information, attitude information and velocity information of the GCB workpiece by using sensors; S402: calculating the deviation between the actual trajectory and the optimized trajectory ; S403: dynamically adjusting the sampling range of the RRT algorithm and the cost function weight of the CHOMP algorithm and other parameters according to the real-time monitoring data and the error correction results.

[0027] Embodiment three: In the GCB disassembly operation of a certain hydropower station, the method is used for path planning, and the specific process is as follows: I. The key parameters are set as follows: p_start = [0, 0, 0];% starting point coordinates (cm); R_goal = [100,400; 50,350; 20,220]; % target area; obstacles = [500,100,50,200,150,200;% Obstacle 1 [x,y,z, length, width, height]; 1200,300,100,150,100,150]; % obstacle2; R_sample = [-10,2000; -10,500; -10,300]; % sampling range; α_RRT = 50; % RRT step (cm); λ_smooth = 0.7; % smoothing weight; λ_obstacle = 0.3; % obstacle avoidance weight.

[0028] II. The performance of the proposed method is compared with the traditional method as shown in Table 1 below: Table 1: Comparison of the proposed method with the traditional manual method;

[0029] From Table 1 above and Figure 2 It can be seen that the present application significantly improves the efficiency of GCB disassembly in hydropower stations by fusing RRT rapid exploration and CHOMP trajectory optimization. Compared with the traditional manual method: The planning time is compressed from 120 minutes to 15 minutes, with an efficiency improvement of 87.5%, thanks to the RRT algorithm for quickly generating a feasible path; the path length is shortened by 28.5%, thanks to the continuous trajectory optimization of CHOMP; the collision risk is reduced to zero, thanks to the real-time monitoring and dynamic adjustment of parameters; the trajectory smoothness is improved by 61.8%, ensuring stable operation of the equipment; and the positioning accuracy is ±3 cm, with an improvement of 80%.

[0030] The core innovation of the proposed method is that the RRT-CHOMP two-stage optimization considers both efficiency and quality; the dynamic parameter adjustment mechanism responds to environmental changes in real time; and the three-dimensional obstacle avoidance model accurately handles complex spatial constraints. Experiments show that the success rate of the proposed method is 99.5% in 50 disassembly tasks, which is 14.5% higher than that of the manual method, significantly reducing the safety risk and providing an intelligent solution for equipment maintenance in hydropower stations.

Claims

1. A dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations, characterized in that, Includes the following steps: S1, Initial condition setting and parameter initialization; S2, using the RRT algorithm to generate the initial GCB assembly / disassembly path; S3, using the CHOMP algorithm to optimize the GCB disassembly and assembly trajectory; S4 monitors the GCB disassembly and assembly process in real time and dynamically adjusts parameters; S5 outputs the adjusted algorithm parameters and trajectory optimization results.

2. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 1, characterized in that, Step S1, initial condition setting and parameter initialization, includes the following steps: S101: Define the start point for disassembling and assembling GCB workpieces. and target area ; S102: Configure obstacle information within the workspace, including walls and hoisting equipment. ; S103: Set the random sampling range for the RRT algorithm Expanding step size and maximum number of iterations ; S104: Set the smoothness cost weights for the CHOMP algorithm Obstacle avoidance cost weight Number of iterations and gradient descent step size parameter.

3. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 2, characterized in that, GCB workpiece assembly and disassembly starting point Target area .

4. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 2, characterized in that, Obstacle information for walls and hoisting equipment .

5. The dynamic path planning and real-time monitoring method for the disassembly and assembly of GCBs in hydropower stations according to claim 2, characterized in that, The random sampling range of the RRT algorithm is: 。 6. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 1, characterized in that, In step S2, generating the preliminary GCB assembly / disassembly path using the RRT algorithm includes the following steps: S201: Initialize the node set Using the RRT algorithm in the workspace Internal random sampling to generate sampling points ; S202: Using nearest neighbor search in Find the distance to the sampling point nearest node And try to move towards the sampling point. Expanding step size Generate new nodes ; S203: Check new node Whether or not The node is randomly selected to collide with any obstacle in the node set. If a collision occurs, the node is discarded and a new random selection is performed; otherwise, it is added to the node set. middle; S204: Repeat steps S201-203 until a preliminary path connecting the starting point and the target region is found or the preset number of iterations is reached. Output the initial path .

7. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 1, characterized in that, Step S3, optimizing the GCB assembly / disassembly trajectory using the CHOMP algorithm, includes the following steps: S301: Parametric trajectory: Parameterization as a continuous function ; S302: Define the cost function: including the smoothness cost function. and obstacle avoidance cost function Calculate the total cost function ; S303: Calculate gradient: for about Take the derivative to obtain the gradient. ; S304: Update the trajectory using gradient descent; S305: Collision Detection and Adjustment: After each trajectory update, check... Whether or not Collision; if a collision occurs, adjust the parameters and return to S303; S306: Repeat S303-S305 until the maximum number of iterations is reached. Or the cost function value no longer decreases significantly.

8. The dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 7, characterized in that, In the smoothness cost function: ； ； ； In the formula, It is a continuous function.

9. A dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 7, characterized in that, Updating the trajectory using gradient descent includes: ； in This represents the trajectory at the k-th iteration.

10. A dynamic path planning and real-time monitoring method for the assembly and disassembly of GCBs in hydropower stations according to claim 1, characterized in that, In step S4, the real-time monitoring of the GCB assembly / disassembly process and the dynamic adjustment of parameters such as the sampling range of the RRT algorithm and the cost function weights of the CHOMP algorithm include the following steps: S401: Use sensors to acquire position, attitude, and velocity information of the GCB workpiece; S402: Calculate the actual trajectory and the optimized trajectory Deviation between; S403: Based on real-time monitoring data and error correction results, dynamically adjust the application range of the RRT algorithm and the cost function weights of the CHOMP algorithm, among other parameters.

Images