CN114021745A

CN114021745A - Virtual overhaul flow optimization method for hydropower station equipment

Info

Publication number: CN114021745A
Application number: CN202111235753.7A
Authority: CN
Inventors: 付文龙; 顾嘉豪; 李佰霖; 袁朝晖
Original assignee: China Three Gorges University CTGU
Current assignee: China Three Gorges University CTGU
Priority date: 2021-10-22
Filing date: 2021-10-22
Publication date: 2022-02-08
Anticipated expiration: 2041-10-22
Also published as: CN114021745B

Abstract

A virtual overhaul flow optimization method for hydropower station equipment comprises the following steps: step S1: establishing a three-dimensional model of the equipment; step S2: obtaining a constraint matrix model of the equipment; step S3: establishing a disassembly evaluation index; step S4: establishing an objective function and calculating a workload matrix; step S5: solving by using an improved discrete whale algorithm to obtain an optimal disassembly sequence; step S6: and automatically playing the disassembly animation by adopting the Unity3D recognition sequence. The whale algorithm provided by the invention combines optimization heuristic variation and round trip optimization operators, can well balance global search capacity and local search capacity, can efficiently solve NP problem, and Unity3D can complete three-dimensional visualization of a disassembly process by means of scripts with rich functions, thereby being beneficial to improving maintenance efficiency.

Description

Virtual overhaul flow optimization method for hydropower station equipment

Technical Field

The invention relates to the technical field of equipment maintenance and virtual reality, in particular to a virtual maintenance flow optimization method for hydropower station equipment.

Background

The hydropower station has good peak regulation and frequency modulation capabilities in a power system network and plays an important role in the safe and stable operation of a power grid. As a hydraulic turbine unit which is one of important hydropower station equipment, a water inlet ball valve arranged at a water inlet of the hydraulic turbine unit can ensure that the hydraulic turbine unit can still stably run for a period of time when a system fails. Because the ball valve that intakes is opening and the closed in-process for a long time, the sealing washer can be inevitable and lose sealed effect because of wearing and tearing, then need change ball valve seal this moment. However, the structure and the working principle of the water inlet ball valve serving as a main water inlet system are complex, and the difficulty in overhauling and maintaining is very high, so that the establishment of an excellent ball valve part disassembling sequence plan to improve the overhauling efficiency and shorten the overhauling time becomes very critical.

In previous researches, the problem of the disassembly sequence planning of hydroelectric equipment is a discrete space optimization problem, and the expression of equipment components is often applied to the theory of graph theory, wherein the theory comprises a directed graph model, an undirected graph model, an and-or graph model and a Petri network model. However, as the number of parts of the equipment increases, the graph theory models are easy to become large and complex, the constraint relation among the parts of the equipment is difficult to express, an improved directed graph model containing the combination nodes is proposed, and the image model is converted into a constraint matrix mathematical model for subsequent processing by a computer. After the device is represented digitally, an objective function needs to be established and an algorithm is adopted to solve an optimal disassembly sequence, and four attributes of a disassembly direction, a disassembly method, disassembly requirements and part materials are always considered. However, for the maintenance of large equipment in a hydropower station, the time for the personnel to move back and forth is longer than that for small equipment, so the distance between parts is a non-negligible factor.

The problem of the disassembled sequence planning belongs to a complex completely uncertain polynomial problem, and the problems like the problem can use an exhaustion method to check results one by one, and finally, an answer can be obtained. However, an exhaustive method in which the computation time increases exponentially with the complexity of the problem becomes inapplicable. Therefore, the population-based intelligent algorithm is increasingly applied to the planning of disassembly sequences of waste products and equipment maintenance by virtue of the advantages of simple concept and easy implementation.

So far, the final result of the disassembly sequence problem is a disassembly work instruction book, the visual expression of which is limited and the disassembly process cannot be displayed, so that the visual expression can be enhanced by making a visual disassembly software by combining unity3D software and an intelligent algorithm, and the function of maintenance training becomes particularly important.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides a virtual overhaul process optimization method for hydropower station equipment, which is characterized in that an improved model is established to evaluate the overhaul process of the hydropower station equipment more reasonably, then an improved discrete whale algorithm is used for optimizing the solving process, and finally, Unity3D is adopted to make visual software to automatically play and disassemble animation. The whale algorithm provided by the invention combines optimization heuristic variation and round trip optimization operators, can well balance global search capacity and local search capacity, can efficiently solve NP problem, and Unity3D can complete three-dimensional visualization of a disassembly process by means of scripts with rich functions, thereby being beneficial to improving maintenance efficiency.

The technical scheme adopted by the invention is as follows:

a virtual overhaul flow optimization method for hydropower station equipment comprises the following steps:

step S1: establishing a three-dimensional model of the equipment;

step S2: obtaining a constraint matrix model of the equipment;

step S3: establishing a disassembly evaluation index;

step S4: establishing an objective function and calculating a workload matrix;

step S5: solving by using an improved discrete whale algorithm to obtain an optimal disassembly sequence;

step S6: and automatically playing the disassembly animation by adopting the Unity3D recognition sequence.

In the step S1, three-dimensional modeling is carried out on hydropower station equipment by utilizing Solidworks according to the mechanical drawing CAD design drawing information to obtain a complete 1:1 three-dimensional model.

In step S2, the constraint matrix model of the device is represented by a matrix C ═ C_ij]_N×NIs shown, in which: i, j are nodes of disassembled parts, N is total number of nodes, c_ijThe values of (a) can be expressed as:

in the step S3, the disassembly evaluation index includes a disassembly tool change cost, a disassembly direction change cost, and a disassembly position change cost;

when different parts are disassembled, the disassembling tool can be replaced therewith, and the matrix T is used_ij＝[t_ij]_N×NRepresents the tool change cost, t, between disassembled part node i and disassembled part node j_ijCan be defined as:

the part disassembly direction can be defined by { + x, -x, + y, -y, + z, -z } and matrix D_ij＝[d_ij]_N×NThe cost of the change of the disassembly direction of the ball valve part is represented as follows:

the distance of the disassembly position is measured by three-dimensional software, and the change cost p of the disassembly position_ijThe difference generated by the position change of a person from the node i of the disassembled part to the node j of the disassembled part is represented by the following formula:

wherein, the central coordinate of the part node i is (x)_i,y_i,z_i) The center coordinate of the part node j is (x)_j,y_j,z_j). If the center coordinate of the part node i is the origin of coordinates (0,0,0) and the center coordinate of the part node j is (1,1,1), the difference generated by the position change of the maintenance personnel from the disassembled part node i to the disassembled part node j is 1.732.

In step S4, a one-dimensional disassembly sequence set ds is given to represent a part disassembly sequence required for completing the equipment maintenance project, and the objective function may be defined as:

where k is the part position in the disassembly sequence, ds (k) denotes the number of the kth part, T_{ds(k),ds(k+1)}Representing the cost of tool change between the kth part and the (k +1) th part;

the tool change cost, the position change cost and the direction change cost are added to form a target function, and the finally obtained disassembly work cost is smaller and better.

Step S4, establishing an objective function and calculating a workload matrix (T)_ij+P_ij+D_ij) It is a two-dimensional matrix of i rows and j columns, i.e. the sum of tool change cost, position change cost and direction change cost, representing the differential workload of the mutual disassembly and transfer of all nodes of the equipment (assuming that the equipment contains n nodes), and can use OP_ijTo indicate.

The step S5 includes the steps of:

s5.1: determining a disassembly node, and quickly generating an initial population by using a hierarchical combination method;

s5.2: calculating the workload of the individual;

s5.3: establishing a priority protection constraint crossing mechanism;

s5.4: selecting individuals from parents by using a discrete whale algorithm and generating more excellent individuals through crossing;

s5.5: adding heuristic mutation and round trip optimization operators;

s5.6: judging whether the set iteration times reach the maximum iteration number or not, and if not, returning to the step S5.2; otherwise, outputting the disassembly sequence with the minimum workload.

In step S5.4, the discrete whale algorithm for selecting a male parent from the population is:

wherein, X (t +1) refers to the next generation of the current iteration, namely the descendant individual, and f (-) refers to the method of selecting nodes from the parent by adopting the criterion of priority protection of the cross PPX and generating the disassembly sequence of the descendant; x^*(t) refers to the individual with the minimum workload, i.e. the minimum disassembly cost, which is found from the population in the current iteration process; x_rand(t) refers to the individuals randomly selected from the population in the current iterative process; g (t) refers to a set of the first B individuals in the current iteration process, wherein the population is arranged from small to large according to the disassembly cost; a is coefficient vector

It takes on the value of

It is decided that, in the course of an iteration,

the linear decrease from 2 to 0 is obtained,

is a random variable from 0 to 1, and p is also a random variable from 0 to 1.

The discrete whale algorithm needs to determine three parent individuals, namely, the disassembled sequences with the best population fitness value and the minimum disassembly cost in the current iterationBody X^*Randomly selected individuals X_randAnd arranging the first B cell arrays G of the population from small to large according to the disassembly cost.

The length of the cell array G is determined by B, which is defined as follows:

wherein: n is a radical of_pDenotes the size of the population, t_maxIs the maximum value of iteration [, ]]Is the rounding of the whole fraction, and B is a variable that decreases to 1 as the number of iterations increases.

In the step S6, after the discrete whale algorithm is improved in the step S5 and iterated for a fixed number of times, an optimal disassembly sequence is obtained, which means that each iteration can find an individual X with the minimum workload, i.e., the minimum disassembly cost, from the population in the current iteration process^*And (t), when iteration reaches a certain number of times, the workload of searching the individual with the minimum disassembly cost can not be reduced. Then, editing the script by using Unity3D, leading each model imported into Unity3D to define the disassembly direction and the disassembly distance, taking the optimal disassembly sequence as input, moving the three-dimensional model by using transform () method in Unity3D script for the first part in the sequence, detecting the moving distance by using vector3 distance () method, if the moving distance reaches the defined value, hiding the first part by using gap object.setactive (), then controlling the second part, thereby moving all parts in the optimal disassembly sequence by using the same method.

The invention discloses a virtual overhaul process optimization method for hydropower station equipment, which has the following technical effects:

1) the constraint matrix and the directed graph model which are used for quantitatively representing the equipment can completely and effectively express the constraint relation among the parts of the hydropower station equipment after the combination nodes are added.

2) The position change cost is added into the evaluation index, and two different working conditions of a large tool and a small tool are added into the tool change cost, so that the finally established objective function is closer to the actual overhaul operating environment, and the distance between each part led into the Unity3D can be accurately measured.

3) The improved discrete whale algorithm combines optimization heuristic variation and a round-trip optimization operator, can well balance global search capacity and local search capacity, can efficiently solve NP problem, and the Unity3D can complete three-dimensional visualization of a disassembly process by means of a script with rich functions.

Drawings

Fig. 1 is a detailed flowchart of the present embodiment.

Fig. 2 is an exploded view of the downstream pipeline of the water inlet ball valve of the present embodiment.

Fig. 3 is a downstream pipeline directed graph defined in the present embodiment.

FIG. 4 is a downstream pipeline constraint matrix diagram according to this embodiment.

Fig. 5 is a directed graph of the present embodiment.

Fig. 6 is a diagram for generating a feasible solution of the present embodiment.

Fig. 7 is a priority protection constraint cross-plot of the present embodiment.

Fig. 8 is a schematic diagram of the disassembled visualization of Unity3D in this embodiment.

FIG. 9(1) is a first graph (item 1) of the algorithm optimization of this embodiment;

fig. 9(2) is a second graph (item 2) of the algorithm optimization of this embodiment;

fig. 9(3) is a third graph (item 3) of the algorithm optimization of this embodiment.

Detailed Description

As shown in fig. 1, a virtual overhaul process optimization method for hydropower station equipment includes the following steps:

s1: establishing a three-dimensional model of equipment by utilizing Solidworks;

and carrying out three-dimensional modeling on hydropower station equipment by utilizing Solidworks according to drawing information to obtain a complete 1:1 three-dimensional model.

Fig. 2 is an exploded view of a downstream pipeline of a water inlet ball valve, the number of original parts of the downstream pipeline of the ball valve is 315, in order to simplify the processing flow, the common method is to group 3 stud nuts and use the central positions of the nuts as coordinates, the expression of the downstream pipeline is simpler through replacement, and the number of the finally disassembled nodes is 73.

Fig. 3 is a directed diagram of a downstream pipeline, which represents the constraint relationship between components, and the arrow represents the disassembly direction, i.e. the component before the arrow must be disassembled before the component after the arrow. The node types are as follows:

(1) free node: nodes that can be disassembled directly, such as 2, 25, 51;

(2) leaf node: refers to nodes that do not pose constraints on other nodes, such as 59, 60;

(3) and (3) common nodes: the nodes of other nodes are restrained by other nodes;

(4) combining nodes: like

nodes

1, 12, 24, the composite node does not actually exist, it exists in its entirety on behalf of its children nodes.

S2: obtaining a constraint matrix model of the equipment;

the constraint relation between two disassembly nodes is converted into a mathematical model and can be represented by a constraint matrix C ═ C_ij]_N×NWherein i, j is the node of the disassembled part, N is the total number of the nodes, c_ijThe values of (a) can be expressed as:

the effect of the combination nodes is to make the model representation more compact, so that the constraint matrix of the upstream pipeline can be calculated, as shown in fig. 4.

S3: establishing a disassembly evaluation index;

and establishing three evaluation indexes of cost change of a disassembling tool, cost change of a disassembling direction and cost change of a disassembling position.

A certain part on the ball valve corresponds to a disassembling tool, when different parts are disassembled, the tool can be replaced along with the disassembling tool, and the matrix T is used for the purpose_ij＝[t_ij]_N×NRepresents the tool change cost, t, between disassembled part node i and disassembled part node j_ijCan be defined as:

in a three-dimensional rectangular coordinate system, the disassembly direction of the part can be defined by { + x, -x, + y, -y, + z, -z } and can be defined by a matrix D_ij＝[d_ij]_N×NThe cost of the change of the disassembly direction of the ball valve part is represented as follows:

a ball valve model successfully imported into Unity3D automatically adds a transform component, wherein a ternary variable position is formed in the component and represents the position of a model part in three-dimensional space, for example, the position of a part i is (x)_i,y_i,z_i) The position of part j is (x)_j,y_j,z_j) Then the available matrix P_ij＝[p_ij]_N×NExpressed by the following formula:

s4: establishing an objective function and calculating a workload matrix;

firstly, a one-dimensional disassembly sequence group ds is given, which represents the disassembly sequence of the parts required by completing the equipment maintenance project, and the objective function can be defined as:

S5: solving by using an improved discrete whale algorithm to obtain an optimal disassembly sequence:

s5.2: calculating the workload of the individual;

s5.3: establishing a priority protection constraint crossing mechanism;

s5.4: selecting individuals from parents using a whale algorithm and generating more excellent individuals through crossover;

s5.5: adding heuristic mutation and round trip optimization operators;

s5.6: judging whether the set iteration times reach the maximum iteration number or not, and if not, returning to the step S5.2; otherwise, outputting a disassembly sequence with the minimum workload;

in step S5.1, the directed graph model is converted into a constraint matrix as shown in fig. 6, according to the example in fig. 5.

S5.1.1: assuming that the node to be disassembled is 5, column 5 retrieves 1 with 2,3, 4 nodes, meaning that 2,3, 4 nodes must be disassembled prior to 5, and then 5 is placed in the last level C;

s5.1.2: next,

search

2,3, 4 columns, find that they all point to 1, and then

place

2,3, 4 in layer B;

s5.1.3: searching the 1 st column, finding no restriction, and stopping searching by placing 1 in the A layer;

s5.1.4: in order to obtain different initial population individual sequences, nodes in the three layers are randomly scattered and arranged, and finally A, B, C layers are sequentially placed in a disassembled sequence group ds, so that the initial population of the disassembled sequence individuals meeting the constraint relationship can be quickly obtained by repeating the operation.

In step S5.2, the sequence in ds is substituted into the following equation:

the workload of each individual in the population can be determined.

In step S5.3, as shown in fig. 7, two parents are extracted from the generated initial population, namely, false 1 ═ 1,2,4,3,5, and false 2 ═ 1,4,3,2, 5. All the nodes meet the constraint relation among the nodes, a random variable p between 0 and 1 is given, and the rule is that if p is larger than 0.5, the nodes are selected from a male parent 2; if p is less than 0.5, a node is selected from the parent 1.

S5.3.1: when p is 0.8147>0.5 and p is 0.623>0.5, selecting the

initial nodes

1 and 4 from the parent 2 and storing the nodes in the offspring;

s5.3.2: when p is 0.1270<0.5, selecting from the male parent 1, deleting the existing 1 and 4 nodes, leaving only 2,3 and 5 nodes in the male parent 1, and selecting the node 2 in sequence and storing in the offspring;

s5.3.3: when p is 0.9134>0.5, deleting 1,4,2 in the male parent 2 and leaving 3,5, and storing 3 in the offspring; when p is 0.0971<0.5, 5 in the male parent 1 is selected and stored in the offspring, and finally the offspring child satisfying the constraint relationship is obtained [1,4,2,3,5 ].

In step S5.4, the whale algorithm for selecting a male parent from the population is:

wherein, X (t +1) refers to the next generation of the current iteration, i.e. the descendant individuals, and f (-) refers to the method of selecting nodes from the parent by adopting the criterion of priority protection of the cross PPX and generating the disassembly sequence of the descendant.

Discrete whale algorithm, three parent individuals need to be determined, namely a disassembly sequence individual X with the best population fitness value and the minimum disassembly cost in current iteration^*Randomly selected individuals X_randAnd arranging the first B cell arrays G of the population from small to large according to the disassembly cost.

The length of the cell array G is determined by B, which is represented by the following equation:

wherein N is_pIs the population size, t_maxIs the maximum value of iteration [, ]]Is the rounding of the whole decimal. B is a variable that decreases to 1 as the number of iterations increases.

In step s55, heuristic mutation is aimed at enhancing the algorithm local search capability, and the steps are as follows:

s5.5.1: setting the variation number N_mHalf of the population number;

s5.5.2: randomly selecting a disassembled sequence seq from a disassembled sequence set with small relative workload as a variant individual;

s5.5.3: randomly selecting an Mp node from the selected variant individual seq as a variant node, then searching a position range in which the node can be reinserted, and in order to reduce the number of times of judgment, firstly, checking a seq (Mp) th column in a constraint matrix, recording a node number of a column corresponding to data equal to 1 as a Start node set Start, checking a seq (Mp) th row of the constraint matrix, and recording a node number of a row corresponding to data equal to 1 as an End node set End; then checking a disassembly sequence seq, wherein the insertable position is an insertable position after the node in the Start set and before the node in the End set;

s5.5.4: calculating the workload values corresponding to all the pluggable positions, and reserving the individual H with the minimum workload_bestFor updating the population;

s5.5.5: repeat S5.5.2 to S5.5.4 for a total of N_mNext, the variation between two generations of a population is completed.

S6: automatically playing and disassembling the animation by adopting a Unity3D recognition sequence;

after the discrete whale algorithm is improved and iterated for a fixed number of times, an optimal or near optimal dismantling sequence can be obtained, as shown in fig. 8, the ID of the part node is input in the upper right input field of the visual dismantling software, for example, the ID of the working seal ring of the ball valve is 37; then, using the edited script to identify the ID, generating an initial population by using a hierarchical combination method, then using a discrete whale algorithm to iterate for 250 times, and keeping the minimum workload of each iteration; finally, the optimization curve of the algorithm can be seen by drawing by using an Xcharts plug-in unit 3D shop and taking the number of iterations as an abscissa and the workload as an ordinate. And then sequentially controlling the three-dimensional models to move according to the optimal disassembly sequence, and recording the model moved each time and displaying the model in a right status bar.

Example (b):

taking a water inlet ball valve of a water inlet system in hydropower station equipment as an example, in long-term operation, the problems of seal clearance expansion and seal part aging and abrasion exist in the upstream overhaul seal and the downstream working seal of the water inlet ball valve, so when the ball valve cannot play a sealing role due to aging parts and expanded clearances, the upstream overhaul seal ring and the downstream working seal ring of the ball valve need to be replaced. To complete the overhaul, three projects need to be disassembled.

Item 1 is the disassembly of the bypass line. The bypass pipeline mainly comprises a bypass pipe, a hydraulic valve, a manual valve and bolt and nut fasteners thereof, and is connected with the upstream pipeline and the downstream pipeline, so that the bypass pipeline must be disassembled firstly when the upstream pipeline and the downstream pipeline are disassembled.

Item 2 is the replacement of the maintenance seal ring. The surface of the sealing ring can be damaged to a certain extent when the maintenance sealing ring in the upstream pipeline is eroded by high-pressure and high-speed water for a long time in operation. If the surface of the sealing ring is seriously damaged and the damaged area is overlarge, and the sealing water leakage of the ball valve is increased, the sealing ring needs to be replaced.

Item 3 is to replace the working seal ring. The working seal rings in the downstream pipeline are also damaged in the operation process like the maintenance seal rings and also need to be replaced.

And then, testing the performance of each algorithm by using the items, wherein in a test experiment, each algorithm is executed 100 times, the population size Np and the iteration number T are recorded in an experiment result, and the minimum value MIN, the maximum value MAX, the average value AVG, the standard deviation STD, the average running time T and the proportion ROM of the minimum-cost disassembled sequence in the disassembled sequence result are obtained through planning. The operation environment is as follows: test software-matlab 2016; encoding language-c language; processor-intel core i 5-6300 hq 4, main frequency-2.3 GHz, memory-16 GB. The experiments were as follows:

DWOA was tested for project 1, project 2, and project 3, respectively, and the effects of HM, FBOO on DWOA convergence speed, optimization capability, and algorithm stability were tested.

Table 1 test experiment results of item 1, item 2 and item 3

The test results are shown in table 1. In item 1, the minimum value of the disassembly sequence is taken by the FBOO or the DWOA of the HM in parallel, which shows that the HM and the FBOO can enhance the optimization capability of the algorithm, and items two and three show that after the HM and the FBOO are increased, the STD and the AVG in the result are smaller, the ROM is larger, and especially the effect is more obvious after the HM is increased, thus the convergence speed and the optimization capability of the discrete whale algorithm are obviously improved by the HM. The optimal image data of the four algorithms for achieving the minimum dismantling cost is shown in fig. 9(1), fig. 9(2) and fig. 9(3), and it can be seen that the DWOA algorithm after adding HM and FBOO can find the optimal value more quickly.

Claims

1. a virtual maintenance process optimization method for hydropower station equipment, is characterized in that comprising the following steps:

Step S1: establish a three-dimensional model of the device;

Step S2: obtain the constraint matrix model of the device;

Step S3: establishing dismantling evaluation indicators;

Step S4: establish the objective function and calculate the workload matrix;

Step S5: use the improved discrete whale algorithm to solve to obtain the optimal dismantling sequence;

Step S6: using the Unity3D recognition sequence to automatically play the dismantling animation.

2. A virtual maintenance process optimization method for hydropower station equipment according to claim 1, characterized in that: in the step S2, the constraint matrix model of the equipment is represented by a matrix C=[c _ij ] _N×N , wherein: i, j are the dismantling component nodes, N is the total number of nodes, and the value of c _ij can be expressed as:

3. a kind of virtual maintenance process optimization method for hydropower station equipment according to claim 1, is characterized in that: in described step S3, dismantling evaluation index comprises dismantling tool change cost, dismantling direction change cost and dismantling position change the price;

When dismantling different parts, the dismantling tool will be replaced accordingly. Here, the matrix T _ij =[t _ij ] _N×N represents the tool change cost between the dismantling part node i and the dismantling part node j, and t _ij can be defined as:

The disassembly direction of the part can be defined by {+x,-x,+y,-y,+z,-z}, and the matrix D _ij =[d _ij ] _N×N represents the cost of changing the disassembly direction of the part, as follows :

The dismantling position change cost p _ij represents the difference between the position of the dismantling part node i to the dismantling part node j, and its formula is as follows:

Among them, the center coordinates of part node i are ( _xi , y _i , z _i ), and the center coordinates of part node j are (x _j , y _j , z _j ).

4. a kind of virtual maintenance process optimization method oriented to hydropower station equipment according to claim 1, it is characterized in that: in described step S4, first provide one-dimensional dismantling sequence array ds, in order to represent the needs of completing equipment maintenance project The part disassembly sequence of , the objective function can be defined as:

Among them, k is the part position in the dismantling sequence, ds(k) represents the number of the kth part, T _{ds(k), ds(k+1)} represents the distance between the kth part and the k+1th part the cost of tool change;

Adding the tool change cost, position change cost and direction change cost is the objective function, and the final dismantling work cost is as small as possible.

5. a kind of virtual maintenance process optimization method for hydropower station equipment according to claim 1, is characterized in that:

The step S5 includes the following steps:

S5.1: Determine the dismantling node, and use the hierarchical combination method to quickly generate the initial population;

S5.2: Calculate the workload of the individual;

S5.3: Establish a priority protection and constraint intersection mechanism;

S5.4: Use the discrete whale algorithm to select individuals from the male parent and generate better individuals through crossover;

S5.5: Add heuristic mutation and round-trip optimization operators;

S5.6: Determine whether the set number of iterations reaches the maximum number of iterations, if not, return to step S5.2; otherwise, output the dismantling sequence with the smallest workload.

6. a kind of virtual maintenance process optimization method facing hydropower station equipment according to claim 5, is characterized in that:

In the step S5.4, the discrete whale algorithm for selecting the male parent from the population is:

Among them, X(t+1) refers to the next generation of the current iteration, that is, the individual child, and f( ) refers to the disassembly sequence of selecting nodes from the parent and generating the child using the priority protection cross-PPX criterion;

The discrete whale algorithm needs to determine three parent individuals, which are the dismantling sequence individual X ^* with the best population fitness value and the smallest dismantling cost in the current iteration, and the randomly selected individual X _rand . The first B cell array G arranged from small to large;

The length of cell array G is determined by B, which is defined as follows:

Among them: N _p represents the population size, t _max is the maximum value of the iteration, [] is the rounding with all decimals rounded off, and B is a variable that is continuously reduced to 1 with the increase of the number of iterations.

7. a kind of virtual maintenance process optimization method for hydropower station equipment according to claim 1, is characterized in that: in described step S6, after improving discrete whale algorithm through fixed number of iterations, obtains an optimal dismantling sequence , using the Unity3D editing script, taking the optimal dismantling sequence as input, and controlling the movement of the model in turn.

8. The method for obtaining the optimal dismantling sequence based on the improved discrete whale algorithm is characterized by comprising the following steps:

A1: Determine the dismantling node, and use the hierarchical combination method to quickly generate the initial population;

A2: Calculate the workload of the individual;

A3: Establish a priority protection and restraint crossover mechanism;

A4: Use the whale algorithm to select individuals from the male parent and generate better individuals through crossover;

A5: Add heuristic mutation and round-trip optimization operators;

A6: Determine whether the set number of iterations reaches the maximum number of iterations, if not, return to step A2; otherwise, output the dismantling sequence with the smallest workload.

9. a kind of method based on improved discrete whale algorithm solution according to claim 8 and obtain optimal dismantling sequence, it is characterized in that: in step A4, from population, the whale algorithm of selecting male parent is:

The length of cell array G is determined by B, which is represented by:

Among them, N _p is the population size, t _max is the maximum value of the iteration, [] is the rounding of all decimals; B is a variable that decreases to 1 with the increase of the number of iterations.

10. a kind of method based on improved discrete whale algorithm solution according to claim 9 and obtain optimal dismantling sequence, it is characterized in that: in step A5, heuristic variation is intended to enhance algorithm local search ability, and step is as follows:

S5.5.1: Set the number of mutation N _m as half of the population;

S5.5.2: Randomly select a disassembled sequence seq from the disassembled sequence set with relatively small workload as a mutant individual;

S5.5.3: Randomly select the Mpth node from the selected mutant individual seq as the mutant node, and then find the position range where the node can be reinserted. In order to reduce the number of judgments, first, check the seq(Mp)th column in the constraint matrix, The node number of the column corresponding to the record data equal to 1 is used as the starting node set Start, the seq(Mp) row of the constraint matrix is checked, and the node number of the row corresponding to the record data equal to 1 is used as the end node set End; then check the dismantling sequence seq, Then the insertable position is the insertable position after the node in the Start collection and before the node in the End collection;

S5.5.4: Calculate the value of the workload corresponding to all insertable positions, and keep the individual H _best with the smallest workload for updating the population;

S5.5.5: Repeat S5.5.2 to S5.5.4 for a total of N _m times to complete the variation between two generations of a population.