CN116167740A

CN116167740A - Digital twin model construction method for intelligent maintenance of pumped storage power station

Info

Publication number: CN116167740A
Application number: CN202211606679.XA
Authority: CN
Inventors: 刘雷; 张彬桥; 杨洋; 雷均
Original assignee: China Three Gorges University CTGU
Current assignee: China Three Gorges University CTGU
Priority date: 2022-12-13
Filing date: 2022-12-13
Publication date: 2023-05-26

Abstract

A digital twin model construction method for intelligent maintenance of a pumped storage power station is characterized by comprising the following steps of: the method comprises the following steps that firstly, a three-dimensional model is built by adopting a pumped storage power station physical entity digital twin body module; step two, forming an overhaul knowledge graph by adopting a digital twin intelligent pumped storage power station overhaul planning module; step three, adopting a digital twin intelligent pumped storage power station maintenance implementation module to carry out automatic maintenance; and step four, carrying out quantitative evaluation on the model overhaul effect by adopting a digital twin intelligent pumped storage power station overhaul evaluation module. The invention provides a digital twin model construction method for intelligent overhaul of a pumped storage power station, which combines a three-dimensional model of main equipment and hydraulic scene of the pumped storage power station with visualization and dynamics of an overhaul task, and further realizes digitization and standardization of the overhaul task through a knowledge graph technology.

Description

Digital twin model construction method for intelligent maintenance of pumped storage power station

Technical Field

The invention relates to the technical field of hydropower station overhaul, in particular to a digital twin model construction method for intelligent overhaul of a pumped storage power station.

Background

Along with the continuous development of the economy in China, the energy supply and demand patterns are greatly changed, and a clean, low-carbon, safe and efficient energy system is constructed to become a new trend of the energy development in China. The pumped storage power station has the functions of peak regulation, frequency modulation, phase modulation, accident standby, black start and the like, and plays an important role in the energy internet

At present, the three-dimensional simulation technology and the digitizing technology are widely applied in the hydropower industry, a plurality of expert students study the technologies of equipment three-dimensional modeling, overhaul three-dimensional simulation, operation visual simulation and the like, a plurality of results are obtained in the simulation and training fields of main and auxiliary systems such as water turbines, generators, transformers, speed regulators, ball valves, hydraulic buildings and the like, overhaul process simulation mainly comprising two-dimensional or three-dimensional model animation in the overhaul three-dimensional digitizing aspect of a pumped storage power station is still achieved, the open interactive overhaul preview requirement in a virtual environment is not achieved, the tasks are more in the overhaul process of the pumped storage power station, the construction period is compact and the tasks are heavy, in addition, the complex space structure of an underground cavern brings higher challenges to the applicability and reliability of the whole overhaul process of the traditional three-dimensional simulation model, on the other hand, the traditional three-dimensional simulation is quite insufficient in aspects such as overhaul time, space, resource and the like, and the aspects such as overhaul strategy are lack in the aspects such as overhaul strategy optimization. In view of the above, aiming at the problems of poor overhauling interactivity, insufficient strategy optimization capability and the like at present, the technical difficulties of overhauling digital operation specifications, overhauling decision selection and the like of the pumped storage power station are solved based on the key technologies such as a three-dimensional digital technology, an evaluation algorithm technology and the like.

Disclosure of Invention

The technical problem to be solved by the invention is to provide a digital twin model construction method for intelligent maintenance of a pumped storage power station, which combines a three-dimensional model of main equipment and hydraulic scene of the pumped storage power station with visualization and dynamics of maintenance tasks, and further realizes digitization and standardization of the maintenance tasks through a knowledge graph technology.

In order to solve the technical problems, the invention adopts the following technical scheme: a digital twin model construction method for intelligent maintenance of a pumped storage power station, which comprises the following steps,

step one, a three-dimensional model is established by adopting a pumped storage power station physical entity digital twin body module;

step two, forming an overhaul knowledge graph by adopting a digital twin intelligent pumped storage power station overhaul planning module;

step three, adopting a digital twin intelligent pumped storage power station maintenance implementation module to carry out automatic maintenance;

and step four, carrying out quantitative evaluation on the model overhaul effect by adopting a digital twin intelligent pumped storage power station overhaul evaluation module.

Preferably, the first step includes building a model of a water pump turbine, a generator motor, a speed regulator, a generator outlet switch, a booster station and a switching station, and building a three-dimensional model of a generator layer, a water turbine, an outlet layer, a waterwheel chamber, a room partition and an overhaul tool.

Preferably, the second step is to form an overhaul knowledge graph according to the overhaul history of the pumped storage power station; then pushing the required overhaul tools and overhaul personnel for the corresponding overhaul scheme; the reliability and economy of the solutions are analyzed before pushing, and the maintenance solutions are optimized through analyzing the risk and cost of the equipment in the running process.

Preferably, the third step is to push the maintenance task to the three-dimensional engine and display the maintenance task in the form of a task table; and then, based on the maintenance task, roaming maintenance is carried out, roles in the three-dimensional engine are called, the movement and execution tasks of the three-dimensional engine are controlled, and the deduction of major repair, minor repair and electrified maintenance of the pumped storage power station, the deduction of driving operation or hoisting operation is carried out.

Preferably, after the maintenance task is finished, evaluating the maintenance effect; the method comprises the steps of carrying out comprehensive evaluation on the condition monitoring data before and after equipment overhaul according to the vibration, pressure pulsation and temperature signals of the equipment. The monitoring data used in this module may be data derived from the actual pumped-storage power station before and after service, primarily for the purpose of verifying the rationality of the service scheme.

The invention provides a digital twin model construction method for intelligent maintenance of a pumped storage power station, which aims at the aspects of insufficient digital capacity expression, insufficient maintenance interactivity, insufficient maintenance strategy optimization and the like in the aspects of maintenance time, space, resources and the like of the traditional pumped storage power station. A virtual power station based on an actual pumped storage power station is created by adopting a digital twin technology, immersive overhaul operation can be realized in three-dimensional engine software of a computer, an overhaul scheme (an overhaul personnel and a required overhaul tool) can be directly displayed on a man-machine interaction interface by adopting a knowledge graph technology, visualization and dynamics of an overhaul task are realized, the overhaul task is further digitized and normalized by adopting the knowledge graph technology, the limitation of overhaul time, space and resources can not be met in a virtual entity, and the overhaul interactivity is improved.

And pushing the optimal overhaul strategy by using an evaluation algorithm, and providing a ready basis for the execution of the following roaming overhaul task. And finally, evaluating the overhaul effect, providing a scheme for optimizing a follow-up overhaul strategy, and providing technical support for the key technical problem of intelligent overhaul of the pumped storage power station.

Drawings

The invention is further illustrated by the following examples in conjunction with the accompanying drawings:

FIG. 1 is a schematic diagram of the structure of the present invention;

fig. 2 is a schematic diagram of an overhaul knowledge graph structure of the invention.

Detailed Description

As shown in fig. 1, a digital twin model construction method for intelligent maintenance of a pumped storage power station comprises the following steps,

The pumped storage power station physical entity digital twin body module comprises: this part can be divided into a model of a power plant house and a model of equipment and tools.

Space data registration of a main power plant of the power station is obtained through a three-dimensional laser scanning technology, and a model of the power plant is built;

and (3) performing preliminary modeling on equipment parts and maintenance tools by using a 3D MAX or Solidworks modeling tool, correcting the appearance by comparing with a drawing, and assembling the part models into a unit integral model according to the physical structure of unit equipment. And mapping and light and shadow baking are carried out on the equipment model, so that the reality degree of the model is improved.

The main models comprise static models such as medium hydraulic architecture, main equipment and auxiliary equipment which do not participate in animation production, and dynamic models which have game objects such as equipment which participate in dynamic demonstration, dynamic personas, particle special effects and the like. And during scene management, importing the Unity3D engine in the fbx format, merging the equipment model and the factory building model, and completing the complete and refined model of the pumped storage power station.

Digital twin wisdom pumped storage power station overhauls plan module: firstly, according to a unit overhaul case in the pumped storage power station industry, an overhaul flow and a knowledge base query idea are combined, and then an overhaul knowledge map is constructed by utilizing a knowledge map technology. The pumped storage unit maintenance case and the core data source of the maintenance flow are relatively fixed, so that one of three methods in the knowledge graph is adopted to construct the unit maintenance knowledge graph in a bottom-up mode, namely, the entity, the attribute and the relation are extracted from the data source and added into the data layer, the elements are generalized, the elements are abstracted into concept rules, and finally, a mode layer is formed. The data layer acquires the relevant knowledge and case of the pumped storage unit overhaul from the data source, analyzes and preprocesses the relevant knowledge and case, extracts entities, relations and attributes, and completes knowledge storage and fusion. The mode layer acquires a knowledge set of the data layer, integrates the corresponding relation between the overhaul case and the resource, generalizes the knowledge set into an overhaul rule, and establishes a unit overhaul strategy through the overhaul rule to realize intelligent generation of an overhaul plan. The overhaul knowledge graph structure is shown in figure 2.

After the maintenance strategy is generated, various schemes can be adopted, and the transformer is taken as an example and can be classified into major maintenance, minor maintenance and electrified maintenance. The technical nature and economy of the transformer maintenance strategy are comprehensively considered, maintenance cost and risks are combined, and an improved analytic hierarchy process is adopted in the combined method.

Further, starting from the whole operation period of the transformer, a cost analysis model based on the cost of the whole life cycle is established. The method for establishing the model adopts an engineering method: the equipment system is divided into a plurality of components, each cost is calculated one by one according to each parameter, and finally the total cost is obtained by accumulation. The LCC of the transformer represents the cost over the life cycle and its calculation formula can be expressed by formula 1:

LCC＝C ₁ +C ₂ +C ₃ +C ₄ (1)

wherein: c1 is the initial input cost, C2 is the operation maintenance cost, C3 is the fault cost, and C4 is the retirement cost.

The initial input cost is the cost paid before the transformer is put into operation, and the cost only occurs in the initial stage, including purchasing cost, installation cost, debugging cost, other cost and the like. Can be expressed by equation 2:

C ₁ ＝C _p +C _i +C _o (2)

the purchase cost, the installation and debugging cost and other costs are sequentially arranged.

The operation maintenance cost refers to the sum of costs (including operation costs and maintenance costs) of the costs required for the transformer to ensure its safe and stable operation throughout its lifetime. Wherein the operating costs include labor and unavoidable loss costs due to operation. The loss cost can be expressed by equation 3:

C _F ＝a×P _g ×(8760T _s ) (3)

a is electricity selling price, pg is fixed power loss, and Ts is annual outage time. The operator cost can be expressed by equation 4:

where CF is the fixed power loss cost, cv is the variable power loss portion, and Cn is the operator cost. L is the operational lifetime of the transformer. The maintenance cost is the cost of maintaining the power transformer during the operational cycle, in this case, three different maintenance strategies, namely major maintenance, minor maintenance, and live maintenance.

The fault cost mainly comprises power failure loss cost and fault repair cost:

wherein C31 represents a power outage loss cost; c32 represents the cost of fault repair; b represents the electricity selling profit of the unit electric quantity; SN represents the rated capacity of the power transformer; beta represents an average load factor;

representing the average power factor; λ represents the average number of failures of the device; CM represents the cost of fault repair; t represents the device average repair time.

The decommissioning cost only occurs at the end, and means the cost of disassembly and recovery of the transformer, including the cost associated with disposal and the residual value at the time of decommissioning. Denoted as C4:

c1i is the installation direct engineering fee, C1e is the purchase fee, and L can be generally taken for 20 years.

On the basis of modeling based on full life cycle cost, because failure rate or maintenance cost are different for different maintenance modes (major maintenance, minor maintenance, live maintenance), a better scheme needs to be determined according to actual cost calculation and is combined with risk assessment which will be described later.

The equipment risk assessment is an indispensable step of maintenance strategy formulation, and the influence degree of potential risks can be better shown by data through the risk assessment, wherein a plurality of factors such as assets, environment and safety are involved, and the equipment attribute is considered from the technical point of view, and the attribute as an asset is considered from the economic point of view. The risk assessment takes a risk value as an index, combines the possibly lost asset and fault rate, is dimensionless, and has a calculation formula as follows:

R(t)＝LE(t)×P(t) (7)

wherein R is a transformer risk value at the moment t, LE is an asset which is possibly lost, and P is a transformer fault probability. The possible lost assets LE are divided into different levels according to different transformer models, and the product of the assets and the loss degree is considered to obtain the potential loss, wherein the formula is as follows:

LE＝A×F (8)

a is the asset and F is the asset loss level. Asset a will be classified into different levels depending on the type of power transformer. The parameters of the relevant transformer model can be queried to obtain. The asset loss degree F is related to cost, environment and safety, where each loss degree is determined jointly by the element loss and the element loss probability. The specific formula is as follows:

wherein j comprises 1 to 4, which respectively represent cost, environment, personal safety and power grid safety; k comprises 1 to 6, representing the element loss level; fj is the degree of element loss; IOFjk and POFjk are element loss values and element loss probabilities, and specific values can be obtained by inquiring the dividing standard of element loss. The asset loss level is a weighted sum of the loss levels of each element as follows:

F＝w ₁ ×F ₁ +w ₂ ×F ₂ +w ₃ ×(F ₃ +F ₄ ) (10)

wherein F1, F2, F3 and F4 respectively represent the cost of the asset, the environment, the personal safety and the loss degree of the power grid safety; w1, w2, w3 represent element loss weight coefficients, and are respectively 0.4, 0.2, and 0.4.

After the risk value is obtained, a standard is needed to determine the risk level and the acceptability, in the risk judging standard, the application of the ALARP principle is wide, and two important dividing lines, namely an intolerable line and a negligible line, exist in the grading, so that the intolerable line, the AIARP area and the acceptable area are further divided.

Firstly, determining the risk level according to the ALARP criterion and the actual condition of the current power transformer, and further judging which risk level the current risk value of the transformer is in, thereby providing basis for maintenance decision. Wherein the negligible risk level line is the product of the lowest probability of failure to be emphasized level of failure and the lowest possible loss asset level of failure to be emphasized level of failure:

the lowest possible loss asset level should be emphasized:

LE _min ＝A _min ×F (12)

thus, the negligible risk level for a transformer is:

R _min ＝LE _min ×P _min (13)

unacceptable risk levels are:

R _max ＝LE _max ×P _min ,LE _max ＝A _max ×F (14)

the maintenance strategy can then be selected according to the calculated R value, and when the risk value is greater than Rmax, great care must be taken and maintenance is immediately scheduled so as not to jeopardize the safe operation of the whole power system. When the risk value is between the Rmin and Rmax values, the current risk state of the transformer is at a medium level in the ALARP receiving area, and the transformers at medium and high risks need to be timely overhauled in order to ensure safe and stable operation of the power grid. When the risk value is smaller than Rmin, the current state of the transformer is in an acceptable area and is at low risk, and the transformer can be considered to safely and stably run without maintenance work.

And (3) using an improved analytic hierarchy process (APH), integrating risk and cost factors under one frame for comprehensive analysis, and achieving the aim of overhauling and optimizing. When constructing an APH, there are several steps: constructing a hierarchical structure, constructing a judgment matrix and sequencing hierarchical sheets.

Modeling according to APH related theory, dividing the APH related theory into three layers: a target layer A, a criterion layer C and a scheme layer P, wherein the criterion layer comprises a reduced risk value and a full life cycle cost; the scheme layer comprises three schemes of electrified overhaul, minor overhaul and major overhaul; the target layer is a reasonable overhaul decision.

Constructing a judgment matrix and ordering a hierarchical list: the form of the judgment matrix P is shown in equation 15, and it can be used to compare the influence of n factors of a certain layer on the factors of the previous layer.

The judgment matrix is a matrix with equal rows and columns and has symmetry, so that when filling, the main diagonal elements of the matrix are the same elements, namely the importance is the same, and the ratio is 1. For the transformer, the judging matrix of the target layer can be obtained by two criteria of the risk value and the whole life cycle cost, and the maximum characteristic root lambda of the target layer can be obtained by calculation _Max Then by using

Eigenvalue methods, i.e. by the eigenvalue pw=λ _max The corresponding feature vector W is obtained by xw, and W is normalized. Known as hierarchical single ordering; meaning that the same level factor is relative to the upper level factorAnd (5) sorting importance. The priority of the service strategy may be arranged based on the importance ranking of the judgment matrix P of the three scheme layers (overhaul, minor overhaul, live overhaul).

Digital twin wisdom pumped storage power station overhauls implementation module: based on the previous step, the overhaul strategy is prioritized, and the overhaul task can be issued and displayed on the Unity3D interface. The method comprises the steps of dynamically displaying overhaul scenes, realizing roaming overhaul of virtual personas in a three-dimensional engine based on the overhaul scenes, controlling actions of the virtual personas by using a keyboard and a mouse, displaying pictures in front of a user through a computer display when the scenes irradiated by a main camera in Unity3D are running, and controlling the movement of the main camera by writing program codes to realize free movement of the virtual personas in a virtual environment.

Taking the maintenance of the transformer as an example, the free movement of the character can be controlled on the basis of the previous step, and the operations of oil tank sampling, oil tank uncovering, oil discharging, dismantling other parts and the like can be included after the character reaches the vicinity of the target transformer. Mainly uses the trigger function of the collider of the Unity3D and an animation system. Placing a box-type collision device on the oil tank cover part, subsequently realizing the identification of the mouse cursor to the part, utilizing the Animation system of the system to produce Animation in the Unity3D, and generating a state controller file to add an Animation key frame for the Animation after the Animation tab is established, namely, each time point, wherein the time points correspond to the attribute information such as the position, the direction and the like of the part. And setting the part position information, angle information and the like of each key frame, and debugging, running and dismantling the animation. Disassembly of other parts such as bushings, coolers, etc. may also be operated as well. After the disassembly work is completed, the subsequent electric torch, the magnetic rod, the endoscope and the like are checked to check the internal condition of the transformer.

In addition to control of personas, automated overhaul of the task at hand is included. And a basic interface of the Unity3D self-contained GUI module layering is adopted, and navigation components required by overhauling the top, left and bottom modes of the screen in different scenes are clicked by a mouse to enter a specific overhauling task. For example, the overhaul hoisting process of the pumped storage unit comprises the steps of selecting a hoisting tool, penetrating a hoisting wire rope into a hoisting ring of the top cover, and carrying out test hoisting on the top cover within the range of 100 meters. These processes are based on previous step animation, which requires that the animation be debugged and saved in advance, and the sequencing of each step is controlled by using the c# script. And (5) presenting vivid maintenance process to the user.

Digital twin wisdom pumped storage power station overhauls evaluation module: taking repair of a water pump turbine A of a pumped storage unit as an example, firstly, an evaluation index system is established, and in order to accurately represent the real state of the water pump turbine A before and after repair of the water pump turbine A, the average value sj of quartile intervals of all index state monitoring data [ Q1, Q3] is obtained based on the state monitoring data of the bottom indexes of the water pump turbine A before and after repair of the water pump turbine A, wherein [ Q1, Q3] are respectively the quartiles of all index state monitoring data.

S _j ＝∑Q _i n _i /∑n _i

(16)

Wherein Qi is the number of the [ Q1, Q3] intervals; ni is the number of Qi; i is the number of Qi values in the interval.

The degradation degree is often used for representing the damage degree of the component, and the degradation degree of each bottom layer index is calculated according to the average value of the quartile range of each bottom layer index before and after the water pump turbine A is repaired and the limit value of each index. The degradation degree calculation formula of each bottom layer index is as follows:

in the method, in the process of the invention,

a j-th index degradation degree; gamma ray _max 、γ ₀ All are index limit values. The component maintenance effect corresponding to the indexes of vibration swing degree, temperature and pulsating pressure degradation degree can be obtained according to the formula.

Adopting a degradation degree self-adaptive entropy weighting method, an Analytic Hierarchy Process (AHP) andthe weight comprehensive calculation model obtains the comprehensive weight of each layer of index. For the AHP method, already mentioned in the previous method of selection of the maintenance strategy of the transformer, the main steps are also: 1) Constructing index judgment matrixes V of all layers; 2) Calculating the index weight of each layer to obtain the characteristic vector of each index weight

3) And carrying out consistency test on the index weight feature vectors of each layer. Specific formulas are not described in detail; and (3) constructing index judgment matrixes of each layer in sequence according to the comprehensive evaluation index system of the repair effect of the water pump turbine A, calculating weights of the indexes, carrying out consistency test, and calculating the comprehensive degradation degree of the water pump turbine A before and after repair according to the degradation degree of the indexes of each layer. Taking a pressure pulsation signal as an example, a judgment matrix for obtaining the pressure pulsation signal by adopting a nine-fold scale method is V1, and subscripts 1, 2, 3, 4 and 5 in the V1 respectively represent main shaft sealing pressure, guide vane water inlet pressure, guide vane water outlet pressure, draft tube pressure and volute pressure. Carrying out normalization processing on V1 to obtain a normalization matrix V1, further adding V1 in the same row to obtain W1, and carrying out normalization processing to obtain a weight feature vector +.>

Then, the maximum eigenvalue of the judgment matrix V1 is found to be lambda _max And further, a consistency test result is obtained and whether the consistency requirement is met is judged.

The degradation degree adaptive entropy weighting method adaptively weights each index through index degradation degree variability. The method mainly comprises the following steps: 1) Constructing a bottom layer index degradation degree decision matrix X according to n evaluation indexes of two typical working conditions of water pump turbine power generation and water pumping:

2) And (5) normalizing the decision matrix. For index x _ij Normalization processing is carried out, and the calculation formula is as follows:

thereby obtaining a standard normalization matrix Y; 3) Calculating an index entropy value hj, wherein the calculation formula is as follows:

wherein k is Boltzmann coefficient, and mainly comprises two working conditions of power generation and water pumping. So k=1/lnm, m=2;

3) And calculating the index weight uj. The formula is as follows:

thereby obtaining an index weight matrix U. And establishing a bottom index degradation degree decision matrix according to the degradation degree of each index of the power generation and pumping working conditions before and after the water pump turbine A is repaired, calculating the self-adaptive weight of the degradation degree of the bottom index before and after the water pump turbine A is repaired through (19) to (21), further calculating the degradation degree corresponding to each layer of index in the evaluation index system, and constructing a new degradation decision matrix of the next layer of index until the comprehensive degradation degree of the water pump turbine A before and after the repair is calculated.

And (5) weight synthesis. In order to obtain more objective and accurate weights, the objective weights of all indexes obtained by a degradation degree self-adaptive entropy weight method and the subjective weights of all indexes obtained by an AHP method are integrated to obtain more objective and reliable integrated weights T, and the calculation formula is as follows:

wherein, a is the assignment of AHP and degradation degree self-adaptive entropy weight method through expert experience. And integrating the index weights of all layers before and after the repair of the water pump turbine A according to weights obtained by the analytic hierarchy process and the degradation degree self-adaptive entropy weight process to obtain the comprehensive weight of all layers of indexes. Taking the pressure pulsation signal of the power generation working condition before maintenance as an example for analysis, and calculating the pressure pulsation signal AHP method corresponding weight

The degradation degree self-adaptive entropy weight method corresponds to weight U1, comprehensive weight T1 and degradation degree of power generation working condition before repairing pressure pulsation signal A>

And is made up of->

The overall degradation degree is obtained. And the vibration signal, the temperature signal and the comprehensive weight of each layer of index under different working conditions before and after the repair of the index A can be obtained by the same method, and finally, the comprehensive degradation degree of the power generation and water pumping working conditions before and after the repair of the water pump turbine A by the degradation degree self-adaptive entropy weight analytic hierarchy process is obtained.

And (5) comprehensively evaluating the repair effect of the water pump turbine A. And carrying out comprehensive evaluation on the A repair effect of the water pump turbine on the basis of comprehensive degradation evaluation results of the water pump turbine A before and after repair power generation, a water pumping working condition AHP method, a degradation self-adaptive entropy weight method and a degradation self-adaptive entropy weight analytic hierarchy process.

The above embodiments are only preferred embodiments of the present invention, and should not be construed as limiting the present invention, and the scope of the present invention should be defined by the claims, including the equivalents of the technical features in the claims. I.e., equivalent replacement modifications within the scope of this invention are also within the scope of the invention.

Claims

1. A digital twin model construction method for intelligent maintenance of a pumped storage power station is characterized by comprising the following steps of: comprises the steps of,

2. The digital twin model construction method for intelligent overhaul of a pumped storage power station according to claim 1, which is characterized in that: the first step comprises the steps of building a water pump turbine, a generator motor, a speed regulator, a generator outlet switch, a booster station and a switching station model, and building a three-dimensional model of a generator layer, a water turbine, an outlet layer, a waterwheel chamber, a room partition and maintenance tools.

3. The digital twin model construction method for intelligent overhaul of a pumped storage power station according to claim 1, which is characterized in that: step two, forming an overhaul knowledge graph according to the overhaul history of the pumped storage power station; then pushing the required overhaul tools and overhaul personnel for the corresponding overhaul scheme; the reliability and economy of the solutions are analyzed before pushing, and the maintenance solutions are optimized through analyzing the risk and cost of the equipment in the running process.

4. The digital twin model construction method for intelligent overhaul of a pumped storage power station according to claim 1, which is characterized in that: pushing the maintenance task into a three-dimensional engine and displaying the maintenance task in a task form mode; and then, based on the maintenance task, roaming maintenance is carried out, roles in the three-dimensional engine are called, the movement and execution tasks of the three-dimensional engine are controlled, and the deduction of major repair, minor repair and electrified maintenance of the pumped storage power station, the deduction of driving operation or hoisting operation is carried out.

5. The digital twin model construction method for intelligent overhaul of a pumped storage power station according to claim 1, which is characterized in that: step four, after the maintenance task is finished, evaluating the maintenance effect; the method comprises the steps of carrying out comprehensive evaluation on the condition monitoring data before and after equipment overhaul according to the vibration, pressure pulsation and temperature signals of the equipment.