Hydropower station safety comprehensive management system and method based on digital twin scene

Abstract

The present application relates to a hydropower station safety comprehensive management system based on a digital twin scene, designs a hydropower station data resource coding system, establishes a hydropower engineering digital twin base, integrates all relevant geographic spatial data, basic data, monitoring data and business management data of the hydropower station, and constructs a hydropower station digital twin scene; predicts the displacement and deformation of the hydropower station building structure and the change trend of the state index of important facilities and equipment of the hydropower station, and predicts and judges the safety risks therein; forecasts or predicts the hydropower station reservoir inflow, upstream and downstream river section water level, flow change, hydropower station power generation output and power generation capacity, provides corresponding plans for flood warning or weather disaster warning or water level overrun warning, simulates the key links of the plans in combination with a simulation engine; and simulates and calculates different power generation and reservoir dispatching strategies to obtain an optimal dispatching scheme. The present application realizes the intelligentization of hydropower station engineering safety, dispatching safety and power generation safety management.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of intelligent hydropower stations, and particularly relates to a hydropower station safety comprehensive management system based on a digital twin scene. BACKGROUND

[0002] Digital twin, also known as digital mapping or digital mirror, is a simulation process integrating multi-discipline, multi-physical quantity, multi-scale and multi-probability in a virtual space by making full use of physical models, sensor updates and operation history data. The digital twin of a hydropower station is an intelligent solution using digital technology, which simulates the facilities, equipment and hydrological data of the hydropower station to realize monitoring, optimization and prediction of the hydropower station, and is conducive to improving the construction, operation and safety management level of the hydropower station. However, the existing hydropower station safety management systems or methods using digital twin technology have low intelligence and wisdom, such as the invention application “A reservoir safety life cycle management system based on digital twin” (CN116362689A) which discloses a reservoir safety full life cycle management system including data acquisition, data support, data processing, application support, application management and visualization modules, which collects, analyzes and processes various data of the reservoir to realize monitoring and management of all factors of the reservoir. The system disclosed in the invention application focuses on data collection, analysis, processing and monitoring and management of reservoir operation, mainly provides query display services, and the application in engineering safety prediction, early warning, pre-play, pre-plan and other aspects is not sufficient, and the simulation, simulation and deduction of safety risks and operating conditions cannot be realized in combination with the digital twin scene, which is difficult to assist the operation and maintenance unit in safety comprehensive management and dispatching decision-making. SUMMARY

[0003] The application aims at the above problems, and provides a hydropower station safety comprehensive management system based on a digital twin scene, designs a hydropower station data resource coding system, establishes a hydropower engineering digital twin base, integrates all related geographic space data, basic data, monitoring data and business management data of the hydropower station, constructs a hydropower station digital twin scene, and provides a three-dimensional visual carrier for hydropower station safety management; the AI model is used to predict the displacement and deformation of the hydropower station building structure and the change trend of the state index of important facilities and equipment of the hydropower station, and to predict and judge the safety risks; the hydrological forecasting model, the hydrodynamic model and the scheduling calculation model are used to forecast or predict the inflow of the hydropower station, the water level and flow change of the upstream and downstream river sections, and the power generation output and capacity of the hydropower station, and corresponding plans are provided for flood warning, weather disaster warning or water level overrun warning, and the key links of the plans are simulated by combining with a simulation engine; different power generation and reservoir scheduling strategies are simulated and calculated to obtain an optimal scheduling scheme; the digital twin scene of the hydropower station is combined to simulate the dynamic effects of flood evolution, gate discharge, unit power generation, water level rise and fall and submergence influence; and the digital twin scene based hydropower station safety comprehensive management system is used to realize intelligent management of hydropower station engineering safety, scheduling safety and power generation safety management.

[0004] In order to achieve the above purpose, the technical scheme provided by the application is:

[0005] The hydropower station safety comprehensive management system based on a digital twin scene comprises an engineering digital twin base and an engineering safety subsystem, a scheduling safety subsystem and a power generation safety subsystem.

[0006] The engineering safety subsystem adopts an AI model for the deformation monitoring points of the hydropower station building to realize deformation prediction of key areas of the hydropower station building; the finite element analysis model is used to calculate the displacement and stress distribution of the hydropower station building structure under design conditions, typical operation conditions, extreme conditions, self-defined conditions or different load combinations, and the displacement field and stress field of the hydropower station building structure are displayed in the form of a rendering cloud map; based on the digital twin scene, abnormal measuring points and risk areas with safety risks are located and labeled.

[0007] The scheduling safety subsystem adopts a hydrological forecasting model to predict the inflow process of the hydropower station, calculates the discharge flow, reservoir water level and unit output under different scheduling schemes, compares the calculation results of different scheduling strategies, selects an optimal scheduling scheme, and simulates the dynamic effects of gate discharge, unit power generation and flood detention by using a digital twin engine.

[0008] The power generation safety subsystem uses an AI model to predict the change trend of the state index of unit equipment; when the unit equipment has safety risks, the abnormal measuring points and risk equipment are located in the digital twin scene, relevant plans are recommended, and the plan disposal process is simulated and practiced.

[0009] The engineering digital twin base integrates all relevant geographic spatial data, basic data, monitoring data, business management data, and external shared data of the hydropower station; through spatiotemporal multi-scale data mapping, a digital twin scene of the hydropower station is constructed to provide a three-dimensional visualization carrier for various business applications.

[0010] Further, the engineering digital twin base comprises a data acquisition module, a data management module, a data model module, and a data sharing module.

[0011] The data acquisition module integrates all relevant geographic spatial data, basic data, monitoring data, business management data, and external shared data of the hydropower station according to a hydropower station data resource coding system.

[0012] The data management module realizes resource directory management, data quality management, and data security management based on a data standard system.

[0013] The data model module constructs a digital twin scene of the hydropower station that is unified in basic data, integrated in monitoring data, and integrated in two and three dimensions, to provide a three-dimensional visualization carrier for various business applications.

[0014] The data sharing module provides an information channel between the management department of the hydropower station and the upper and lower management units and external relevant units, to realize cross-network, cross-industry, and cross-level data sharing.

[0015] Preferably, the hydropower station data resource coding system adopts a three-level coding AB1B2C1C2D1D2.E1E2F1F2G1G2.H1H2H3, the first-level coding AB1B2C1C2D1D2 is the type coding of the data resource, wherein A represents the major category of the data resource, the major category of the data resource includes basic data, monitoring data, business management data, cross-industry shared data, and geographic spatial data; B1B2C1C2D1D2 represents the minor category of the data resource, wherein B1B2 represents the type of monitoring data, the type of monitoring data includes engineering safety monitoring, environmental quantity monitoring, and equipment operation monitoring; C1C2 represents the subdivided type of monitoring data, the subdivided type of monitoring data includes deformation, seepage, hydrology, meteorology, water machine equipment operation, electrical equipment operation, and gold joint equipment operation type; D1D2 represents the sub-type of the subdivided type of monitoring data; the second-level coding E1E2F1F2G1G2 is the hierarchical coding of the data resource, which is set according to the business hierarchy and the subordinate hierarchy; the third-level coding H1H2H3 is the sequential coding of the data resource, which is coded according to the equipment sequence and the measuring point sequence.

[0016] Further, the engineering safety subsystem comprises an engineering safety prediction module, an engineering safety early warning module, an engineering safety pre-play module, and an engineering safety pre-plan module.

[0017] The engineering safety prediction module constructs a gradient boosting regression tree model for deformation prediction of key areas of the hydropower station building for deformation monitoring indicators of the hydropower station building; the prediction results obtained by the gradient boosting regression tree model are rendered into the digital twin scene to form a heat map that can reflect the future deformation distribution of the key section of the building, and the rendering results are published to the WEB end for display.

[0018] The engineering safety warning module publishes warning information when the measured value or the predicted value of the monitoring indicator exceeds the threshold value, displays the warning time, associated measuring points, reasons and data changes, distinguishes the warning levels by different colors, and locates the abnormal measuring points and risk areas in the three-dimensional digital twin model of the engineering structure.

[0019] The engineering safety rehearsal module simulates and calculates the overall displacement and stress and strain of the structure under the preset working conditions and self-defined working conditions using the finite element analysis model of the hydropower station building structure; the user selects the design working condition, the typical operation working condition, the extreme working condition, the self-defined working condition or the self-defined live load, the temperature load and the wind load combination, calculates the displacement and stress distribution of the hydropower station building structure, and displays the displacement field and stress field of the hydropower station building structure in the form of rendered cloud map; the finite element results are rendered on the server side by the visualization engine to generate the corresponding heat map, and the rendering results are published to the WEB end for display.

[0020] The engineering safety plan module searches the built-in plan library when the engineering safety monitoring, prediction and rehearsal results trigger an alarm, finds the plan information suitable for the alarm scenario, and displays the processing method and steps of the plan; combined with the simulation engine, the key links of the plan are simulated, including material preparation, personnel evacuation, equipment inspection and operation dynamic process.

[0021] Further, the method of rendering the deformation prediction results of the gradient boosting regression tree model to the digital twin scene in the engineering safety prediction module includes:

[0022] 1) Estimate the distribution density of the profile measuring points by the coordinate position of the measuring points;

[0023] 2) Set a color gradient to map the displacement prediction data to different color bands and process the color value of each measuring point according to the numerical range;

[0024] 3) Render the displacement prediction results on the server side by the visualization engine to form a heat map that can reflect the future deformation distribution of the key section of the building;

[0025] 4) Publish the rendering results to the WEB end for display, integrate the heat map with the building twin scene, and drag the time axis to view the future deformation trend of the building at different times.

[0026] Further, in the engineering safety pre-play module, the method of fusing stress and strain data obtained by the finite element analysis model with the digital twin scene includes:

[0027] I. Data preparation;

[0028] The calculation data of the finite element analysis model includes node information, element information, and physical properties. The node information includes node number and x, y, z coordinates. The element information includes element number and node number of the constituent element. Each node or element contains corresponding calculated physical properties;

[0029] Before rendering the finite element analysis results, convert the file of the finite element analysis model calculation results into a JSON data file. The JSON data file contains node coordinates, element index, and result values;

[0030] II. Data standardization processing;

[0031] In order to map the finite element calculation results to different color intervals, standardize the displacement and stress data to normalize them to (0, 1);

[0032] III. Read JSON data;

[0033] Read the vertex coordinates, triangle index, and calculation result data in the JSON file through the visualization engine;

[0034] IV. Create triangle patches;

[0035] Reconstruct the node and element information of the finite element model to split the triangles and quadrilaterals contained in the mesh into triangle patches;

[0036] V. Define color mapping;

[0037] Map displacement and stress data to color values, set color gradients, specify color bands for physical properties using RGB codes, and process color values for each node according to numerical range;

[0038] VI. Draw triangle patches and colorize;

[0039] Load the JSON data and triangle patches of the finite element object in the visualization engine, use the previously defined color mapping, render displacement and stress calculation data on the server side, and generate the corresponding heat map;

[0040] VII. Achievement display;

[0041] Publish the rendering results to the WEB side for display, view the displacement and stress distribution cloud map of the XYZ direction of the surface of the hydropower station building structure, and further render the static cloud map into a deformation animation for display through the interpolation algorithm.

[0042] Further, the dispatching safety subsystem comprises a dispatching safety prediction module, a dispatching safety early warning module, a dispatching safety rehearsal module and a dispatching safety plan module.

[0043] The dispatching safety prediction module takes rainfall prediction as input, predicts the inflow of the hydropower station through a hydrological prediction model, takes the predicted inflow as input, predicts the changes of the cross-section water level and flow of the upstream and downstream river channels through a hydrodynamic model, and predicts the power generation output and power generation capacity of the hydropower station through a dispatching calculation model with the input of the gate unit regulation mode.

[0044] The dispatching safety early warning module compares the real-time water level with the characteristic water level, which includes the flood control water level, the normal storage water level, the design flood water level and the check flood water level, and triggers an alarm when the characteristic water level limit is exceeded; when receiving the early warning information of natural disasters such as floods, heavy rains, gales and high temperatures, the early warning information is displayed in a flashing and highlighted manner in the digital twin scene, and the early warning information includes the early warning level, the predicted time and the early warning value, and an early warning message is sent to the relevant person in charge.

[0045] The dispatching safety rehearsal module inputs the inflow process, and the user selects real-time, historical, predicted and preset inflow for calculation, or uploads a template to customize the inflow for calculation; according to the dispatching rules or dispatching instructions, the user selects the control mode of water level, output and discharge flow, inputs the starting water level and control value; through the dispatching calculation model, the changes of the discharge flow, the reservoir water level, the reservoir capacity and the output are calculated; through the hydrodynamic model, the changes of the water level and flow of the cross-section of the upstream and downstream river channels of the hydropower station are calculated; the calculation results of the dispatching calculation model are integrated into the digital twin scene to simulate the dynamic effects of flood evolution, gate discharge, unit power generation, water level rise and fall and inundation influence; the simulation calculation results of different dispatching strategies are compared to select the optimal dispatching scheme and give the best operation mode of the gate and unit.

[0046] The dispatching safety plan module searches the built-in plan library to find the plan information suitable for the scene when the hydropower station encounters a super-standard flood or a related accident, displays the disposal methods and steps of different levels and types of dangerous situations, and simulates the key links of the plan in combination with the simulation engine, and the simulation of the plan includes the dynamic processes of dispatching measures, material preparation, flood control and rescue, personnel evacuation, equipment inspection and plugging and repair.

[0047] Further, the power generation safety subsystem comprises a power generation safety prediction module, a power generation safety early warning module and a power generation safety plan module.

[0048] The power generation safety prediction module predicts the trend of the state quantity index of the important facility equipment of the hydropower station through gradient boosting regression tree or AI model, taking the equipment operation condition as input.

[0049] The power generation safety warning module issues a warning information when the state quantity index of the important equipment of the hydropower station exceeds the threshold value, reminds the user to pay attention in the form of flashing and highlighting, displays the warning time, associated measuring points, data changes, and disposal situation, distinguishes the warning levels by different colors, and locates the abnormal measuring points in the three-dimensional digital twin model of the facility equipment.

[0050] The power generation safety plan module searches the built-in plan library to find the plan information suitable for the alarm scene when the hydropower station encounters an alarm event, displays the plan disposal method and steps, and simulates the key links of the plan by combining the simulation engine. The simulation of the plan includes the dynamic process of fault maintenance, personnel evacuation, and equipment control.

[0051] Further, in the hydropower station safety comprehensive management system, remote sensing images, oblique photography, and BIM modeling data are used to create a digital twin scene, in which the geographic information, civil engineering, electrical equipment and facilities, water turbines, and metal structures of the hydropower station and its affected areas are presented in the form of digital twin models.

[0052] The engineering safety control method of the hydropower station safety comprehensive management system based on the digital twin scene includes the following steps:

[0053] Step 1: Predict the deformation of the safety monitoring point, perform a pre-rehearsal of the displacement and stress-strain of the hydropower station building structure, analyze the abnormalities, and predict the safety risk;

[0054] Step 1.1: According to the deformation monitoring index of the hydropower station building, predict the deformation of the safety monitoring point of the hydropower station, predict the future deformation trend, predict the potential risk, and identify the abnormal measuring point;

[0055] Step 1.2: Select the working condition or load combination mode, perform finite element analysis on the hydropower station building structure, perform a pre-rehearsal of the displacement and stress-strain of the building structure, and analyze the safety risk of the hydropower station building structure according to the displacement and stress-strain of the building structure obtained by the pre-rehearsal;

[0056] Step 2: If the measured value or predicted value of the monitoring index of the safety monitoring point exceeds the threshold value or the method of step 1 identifies an abnormality or safety risk, issue a warning information through the safety comprehensive management system, display the warning time, associated measuring points, cause analysis, and data changes, distinguish the warning levels by different colors, and locate the abnormal measuring points and risk areas in the digital twin model;

[0057] Step 3: search the preplan library to find the preplan information suitable for the early warning alarm scene of step 2, show the processing method and steps of the preplan, and give the recommended preplan for the early warning alarm scene of step 2 to the user;

[0058] Step 4: according to the preplan obtained in step 3, take corresponding safety measures to handle the exception of step 1, control the safety risk, and ensure the safe operation of the hydropower station building structure.

[0059] The dispatching safety control method of the hydropower station safety comprehensive management system based on the digital twin scene comprises the following steps:

[0060] Step 1: predict and forecast the inflow of the hydropower station, the power generation output, and the water level and flow of the upstream and downstream river channels;

[0061] Step 2: issue a warning for real-time water level overrun or predicted water level overrun or standard flood or weather disaster forecast;

[0062] Step 2.1: if the real-time water level or the predicted water level obtained in step 1 exceeds the water level limit value, trigger an alarm and issue a water level limit value warning;

[0063] Step 2.2: if it is predicted in step 1 that the inflow of the hydropower station will exceed the standard flood, issue a flood warning;

[0064] Step 2.3: issue a weather disaster warning for weather disaster forecast;

[0065] Step 3: search the preplan library to find the preplan suitable for the early warning scene of step 2;

[0066] Step 4: preplay the preplan obtained in step 3;

[0067] Step 4.1: select the predicted inflow or the preset inflow; set the starting water level and the control value;

[0068] Step 4.2: calculate the change process of the outflow, the reservoir water level, the reservoir capacity, and the output through the dispatching calculation model; calculate the change process of the water level and flow of the upstream and downstream river channel sections of the hydropower station through the hydrodynamic model;

[0069] Step 4.3: compare the simulation calculation results of different dispatching strategies, and select the optimal dispatching scheme in combination with the preplan;

[0070] Step 5: control the safe and stable operation of the hydropower station according to the optimal dispatching scheme obtained in step 4.

[0071] Compared with the prior art, the beneficial effects of the present application include:

[0072] 1) The application provides a standardized hydropower station digital twin base construction method and a hierarchical data resource coding system, realizes the integration of all related geographic spatial data, basic data, monitoring data and business management data of the hydropower station, and constructs a hydropower station digital twin scene; a hydropower station safety comprehensive management system based on the digital twin scene is established, realizing intelligent management of all-round safety prediction, early warning, preplan, preplay of hydropower station engineering safety, power generation safety, dispatching safety, etc., facilitating hydropower station operation and management personnel to master the hydropower station engineering safety situation, change trend and safety risk, improving the accuracy of early warning and the scientific nature of decision-making, providing technical support in reducing safety hazards, controlling safety risks and assisting management decision-making, which can enhance the comprehensive management ability of the hydropower station, improve the advanced nature, timeliness and intelligent degree of safety management, and provide guidance and reference for efficient, safe and reliable operation of the hydropower station.

[0073] 2) The application provides a hydropower station data resource coding system comprising data resource type coding, monitoring data type coding and data resource level coding, which integrates and integrates multi-dimensional and multi-type data through the digital twin model and digital twin scene of the hydropower station entity, realizes the sharing and collaborative management of various data, and improves the integration and utilization efficiency of hydropower station data resources.

[0074] 3) The application establishes a hierarchical promotion regression tree model for the deformation monitoring points of the hydropower station building, realizes the deformation prediction of the key area of the hydropower station building, constructs a finite element analysis model for the structure of the hydropower station building, calculates the displacement and stress distribution of the building structure under different working conditions and different load combinations, and displays them in the form of rendering cloud chart; when a safety risk occurs, the positioning of abnormal measuring points and risk areas in the hydropower station digital twin scene is used to recommend relevant plans, realize the simulation of the plan disposal process, and greatly improve the intelligent degree and wisdom level of the hydropower station engineering safety.

[0075] 4) The application predicts the inflow process of the hydropower station through a hydrological forecasting model; takes the predicted inflow as input, calculates the outflow, reservoir water level and unit output process under different dispatching schemes through a dispatching calculation model; compares the calculation results of different dispatching strategies, and selects the optimal dispatching scheme; simulates the dynamic effects of gate discharge, unit power generation and flood detention through a digital twin engine; after the occurrence of super-standard flood or related accidents, the early warning information is displayed in the form of flashing and highlighting in the hydropower station digital twin scene, and relevant plans are recommended to realize the simulation of the plan disposal process, greatly improving the intelligent degree and wisdom level of the hydropower station dispatching safety.

[0076] 5) The application predicts the change trend of temperature, vibration and other equipment state indicators of the unit of the hydropower station through intelligent algorithms such as gradient boosting decision tree and long short-term neural network; when there is a risk in the operation of the equipment, the abnormal measuring point and the risk equipment are located in the digital twin scene, and the relevant plan is recommended, thereby realizing the simulation of the plan disposal process, and greatly improving the intelligent degree and intelligent level of the power generation safety of the hydropower station.

[0077] 6) The engineering safety control method provided by the application can discover potential safety risks and abnormal measuring points in advance through the prediction of the deformation of the safety monitoring point and the pre-rehearsal of the displacement and stress-strain of the building structure of the hydropower station, so as to achieve early prevention and early treatment, and effectively avoid or reduce the occurrence of safety accidents; when the measured value or the predicted value of the monitoring index exceeds the threshold value, or the abnormality or safety risk is identified, the early warning information can be issued, and the early warning levels can be distinguished by different colors, and the abnormal measuring point and the risk area can be located in the digital twin model; this visual management mode makes the early warning information more intuitive and easy to understand, which helps to quickly respond to and handle safety problems; the applicable plan information including the treatment method and steps can be quickly found and displayed according to the early warning alarm scene; the user is provided with a targeted solution, which helps to quickly take measures to control safety risks and ensure the safe operation of the building structure of the hydropower station in an emergency.

[0078] 7) The dispatching safety control method provided by the application can master the key parameters and future trend of the operation of the hydropower station in advance through comprehensive prediction and forecast of the reservoir inflow, power generation output, and water level and flow of the upstream and downstream river channels, which helps to discover potential safety risks in time; specific early warnings are issued for different situations such as real-time water level overrun, predicted water level overrun, super-standard flood and meteorological disasters; this precise early warning mechanism can ensure that relevant personnel respond quickly and take necessary dispatching measures to effectively control safety risks and ensure the safe and stable operation of the hydropower station. The system has a built-in plan library, which can quickly find and provide applicable plans according to the early warning scene, and provides the user with a scientific decision basis and reference, which helps to quickly develop and implement effective dispatching schemes in an emergency. Through the dispatching calculation model and the hydrodynamic model, the change process of the outflow, reservoir water level, reservoir capacity, output, and water level and flow of the upstream and downstream river channel sections under different dispatching strategies can be simulated, thereby providing the user with intuitive simulation results, which helps to compare the effects of different dispatching strategies and select the optimal dispatching scheme. According to the optimal dispatching scheme, the safe and stable operation of the hydropower station can be controlled, which helps to ensure that the power generation output, water level and flow of the hydropower station fluctuate within a safe range, avoid the occurrence of safety accidents, and ensure the long-term stable operation of the hydropower station. BRIEF DESCRIPTION OF DRAWINGS

[0079] The application will be further described below in combination with the drawings and examples.

[0080] Figure 1 This is a flowchart illustrating the safety control method for hydropower station projects according to an embodiment of the present invention.

[0081] Figure 2 This is a flowchart illustrating the hydropower station dispatching safety control method according to an embodiment of the present invention.

[0082] Figure 3 This is a flowchart illustrating the hydropower generation safety control method according to an embodiment of the present invention.

[0083] Figure 4 This is a flowchart illustrating the method for constructing a digital twin model of a hydropower station in an embodiment of the present invention.

[0084] Figure 5 This is a simulation interface diagram of the deformation of the ship lift and lock structure of a hydropower station according to an embodiment of the present invention.

[0085] Figure 6 This is a diagram of the reservoir scheduling calculation interface according to an embodiment of the present invention.

[0086] Figure 7 This is a diagram of the hydropower station unit power generation prediction interface according to an embodiment of the present invention. Detailed Implementation

[0087] The hydropower station safety integrated management system based on digital twin scenarios includes an engineering digital twin base as well as an engineering safety subsystem, a dispatch safety subsystem, and a power generation safety subsystem.

[0088] The engineering safety subsystem uses a gradient boosting regression tree model to predict deformation in key areas of hydropower station structures, targeting deformation monitoring points. It also uses a finite element analysis model to calculate the displacement and stress distribution of the hydropower station structure under design conditions, typical operating conditions, extreme conditions, custom conditions, or different load combinations, and displays the displacement and stress fields of the hydropower station structure using rendered cloud maps. Furthermore, it locates and marks abnormal measuring points and risk areas with safety risks based on a digital twin scenario.

[0089] The engineering safety subsystem includes an engineering safety prediction module, an engineering safety early warning module, an engineering safety simulation module, and an engineering safety contingency plan module.

[0090] The engineering safety prediction module constructs a gradient boosting regression tree model for predicting deformation in key areas of hydropower station structures, based on deformation monitoring indicators. The prediction results obtained from the gradient boosting regression tree model are rendered into a digital twin scene to form a heat map that reflects the future deformation distribution of key sections of the structure, and the rendering results are published to the web for display.

[0091] The engineering safety early warning module issues early warning information, shows early warning time, associated measuring points, reasons and data changes, and locates abnormal measuring points and risk areas in the three-dimensional digital twin model of the engineering structure, in the case that the measured value or the predicted value of the monitoring index exceeds the threshold value.

[0092] The engineering safety pre-play module simulates and calculates the overall displacement and stress and strain of the structure under preset working conditions and self-defined working conditions using the finite element analysis model of the hydropower station building structure; the user selects a design working condition, a typical operation working condition, an extreme working condition, a self-defined working condition or a self-defined live load, temperature load and wind load combination, calculates the displacement and stress distribution of the hydropower station building structure, and shows the displacement field and stress field of the hydropower station building structure in the form of rendered cloud maps; the finite element results are rendered on the server side through a visualization engine to generate corresponding heat maps, and the rendering results are published to the WEB side for display.

[0093] The engineering safety pre-plan module searches the built-in pre-plan library when the engineering safety monitoring, prediction and pre-play results trigger an alarm, finds pre-plan information suitable for the alarm scenario, and shows the handling method and steps of the pre-plan; in combination with a simulation engine, the key links of the pre-plan are simulated, including material preparation, personnel evacuation, equipment inspection and operation dynamic process.

[0094] The dispatching safety subsystem predicts the inflow process of the hydropower station using a hydrological prediction model, calculates the discharge, reservoir water level and unit output under different dispatching schemes, compares the calculation results of different dispatching strategies, selects the optimal dispatching scheme, and simulates the dynamic effects of gate discharge, unit power generation and flood retention through a digital twin engine.

[0095] The dispatching safety prediction module predicts the inflow of the hydropower station through a hydrological prediction model with rainfall prediction as input; predicts the water level and flow changes of the upstream and downstream river sections through a hydrodynamic model with predicted inflow as input; predicts the power generation output and capacity of the hydropower station through a dispatching calculation model with gate unit control mode as input.

[0096] The dispatching safety early warning module compares real-time water levels with characteristic water levels including flood control water level, normal storage water level, design flood water level and checking flood water level, and triggers an alarm when the characteristic water level limit value is exceeded; when receiving natural disaster early warning information such as flood, heavy rain, gale and high temperature, the early warning information is displayed in a flashing and highlighted manner in the digital twin scene, and early warning information including early warning level, expected time and early warning value is sent to relevant responsible persons in the form of a short message.

[0097] The dispatching safety rehearsal module inputs the reservoir inflow process, and a user selects real-time, history, forecast, preset reservoir inflow for calculation, or performs calculation in a self-defined reservoir inflow manner by uploading a template; according to a dispatching regulation or a dispatching instruction, the user selects a control mode of a water level, an output, and a reservoir outflow, and inputs a starting regulation water level and a control value; through a dispatching calculation model, reservoir outflow, reservoir water level, reservoir capacity, and output change processes are calculated; through a hydrodynamic model, water level and flow change processes of upstream and downstream river channel sections of the hydropower station are calculated; the dispatching calculation model calculation result is integrated into a digital twin scene to simulate dynamic effects of flood evolution, gate discharge, unit power generation, water level rise and fall, and inundation influence; simulation calculation results of different dispatching strategies are compared to select an optimal dispatching scheme, and a best operation mode of a gate and a unit is given.

[0098] The dispatching safety plan module, when the hydropower station encounters a super-standard flood or a related accident, searches a built-in plan library to find plan information applicable to the scene, and shows disposal methods and steps of different levels and types of dangerous situations; in combination with a simulation engine, key links of the plan are simulated, and the plan simulation includes dynamic processes of dispatching measures, material preparation, flood control and rescue, personnel evacuation, equipment inspection, and plugging and repair.

[0099] The power generation safety subsystem uses an LSTM network to predict a change trend of a unit equipment state index; when the unit equipment has a safety risk, an abnormal measuring point and a risk equipment are located in the digital twin scene, a related plan is recommended, and a plan disposal process is simulated and practiced.

[0100] The power generation safety subsystem includes a power generation safety prediction module, a power generation safety early warning module, and a power generation safety plan module.

[0101] The power generation safety prediction module, for a state quantity index of important facility equipment of the hydropower station, uses a gradient boosting regression tree or an AI model to predict a change trend of the equipment state index by taking equipment operation conditions as input.

[0102] The power generation safety early warning module, when the state quantity index of the important equipment of the hydropower station exceeds a threshold value, issues a warning information to remind a user to pay attention in a flashing and highlighting manner, shows a warning time, an associated measuring point, data change, and disposal conditions, and distinguishes warning levels by different colors, and locates the abnormal measuring point in a three-dimensional digital twin model of the facility equipment.

[0103] The power generation safety plan module, when the hydropower station encounters an alarm event, searches a built-in plan library to find plan information applicable to the alarm scene, and shows a plan disposal method and steps; in combination with a simulation engine, key links of the plan are simulated, and the plan simulation includes dynamic processes of fault maintenance, personnel evacuation, and equipment control.

[0104] The engineering digital twin base integrates all relevant geographic spatial data, basic data, monitoring data, business management data and external shared data of the hydropower station; through time-space multi-scale data mapping, a digital twin scene of the hydropower station is constructed, and a three-dimensional visualization carrier is provided for various business applications.

[0105] The engineering digital twin base comprises a data acquisition module, a data management module, a data model module and a data sharing module.

[0106] The data acquisition module integrates all relevant geographic spatial data, basic data, monitoring data, business management data and external shared data of the hydropower station according to a hydropower station data resource coding system.

[0107] In the embodiment, the hydropower station data resource coding system is divided into three levels, denoted as AB1B2C1C2D1D2.E1E2F1F2G1G2.H1H2H3. The first seven bits are the type code of the data resource, the middle six bits are the hierarchical code of the data resource, and the last three bits are the sequence code of the data resource, which are distinguished by “.”.

[0108] The code A refers to the large category code of the data resource, including five categories of basic data, monitoring data, business management data, cross-industry shared data and geographic spatial data, and the code is set as 1-5.

[0109] The code B1B2C1C2D1D2 refers to the small category code of the data resource, which is subdivided for each type of resource. Taking the monitoring data as an example, the code B1B2 divides the monitoring data into three categories of engineering safety monitoring, environmental quantity monitoring and equipment operation monitoring; the code C1C2 subdivides the monitoring types under the above classification, such as deformation, seepage, hydrology, meteorology, water machine equipment operation, electrical equipment operation and gold joint equipment operation; and the code D1D2 subdivides the monitoring indicators under the above classification, such as displacement, settlement, uplift pressure, foundation reaction, water level, flow, rainfall, vibration, working temperature, power and opening degree. Each level of code is set as 01-09.

[0110] The code E1E2F1F2G1G2 refers to the hierarchical code of the data resource, which can be set according to the business hierarchy and subordinate hierarchy. Taking the monitoring data as an example, the code E1E2 refers to the engineering or equipment object to which the measuring point belongs, such as dam body, dam crest, dam foundation, gate head, water turbine unit, gate and transformer; the code F1F2 refers to the decomposition structure of the engineering or equipment, such as dam section #1, dam section #2 and dam section #3; and the code G1G2 refers to the next level of decomposition of F1F2, such as unit #1 which can be divided into water guide bearing, rack, top cover, stator, spiral case and runner. Each level of code is set as 01-09.

[0111] Encoding H1H2H3 refers to the sequential code of the data resource, which is encoded according to the device sequence and the measurement point sequence, and is set as 001-009. Taking monitoring data as an example, the encoding system of the data resource of the hydropower station is analyzed, as shown in Table 1.

[0112] Table 1

[0113]

[0114] The data management module realizes resource directory management, data quality management and data security management based on the data standard system.

[0115] The data resource directory is established based on the database construction and the resource encoding system, and provides directory and index services for user objects and various business systems. By using data virtualization and metadata technology, unified maintenance of data is realized, and functions such as data catalog management, data catalog publishing, data catalog service and data catalog maintenance are provided.

[0116] Data quality management: Analyze various data quality problems in the whole life cycle of data collection, storage and application, including key data identification, measurement scheme rule design, data quality measurement and data quality measurement report publishing, to provide effective support for improving the data quality of each system.

[0117] Data standard system: Based on the demands of unified data caliber, indicating data position, analyzing data relationship and managing model changes, the basic common standard, data classification standard, encoding system, key technology standard, tool / platform standard, evaluation standard and security standard are formed to support efficient management of hydropower station data assets.

[0118] Data security management: Based on the data resource directory, data classification and identification are carried out, data classification management regulations are formulated, and data storage, access, backup and disposal are managed.

[0119] The data model module constructs a unified basic data, integrated monitoring data and two / three-dimensional integrated hydropower station digital twin scene through space-time multi-scale data mapping, and provides a three-dimensional visualization carrier for various business applications.

[0120] In the embodiment, remote sensing images, oblique photography and BIM modeling data are used to create a digital twin scene, in which the geographic information, civil engineering, electrical equipment and facilities, water turbine and metal structure of the hydropower station and its influence area are presented in the form of digital twin models.

[0121] As shown in Figure 4 the construction process of the hydropower station digital twin scene includes:

[0122] Model processing: Fuse the BIM model of the hydropower station engineering structure and auxiliary mechanical and electrical equipment, the three-dimensional model, the oblique photography model of the surrounding influence area, and other related remote sensing, DOM / DEM model, and convert various models into a unified coordinate system such as CGCS2000 coordinate system, 1985 national elevation datum, and a unified three-dimensional general standard data format.

[0123] Model editing and beautification: Use vector drawing to crop conflicting terrain images, achieve seamless splicing and fusion of digital twins, and refine and map the textures of buildings, equipment, roads, terrain, vegetation, water systems, etc. that do not conform to reality, and beautify the model presentation effect.

[0124] Model data fusion: Establish the association between business, monitoring, and other data and digital twins to form a hydropower station digital twin scene, allowing external physical entities to react in real-time on the model. The basic data of engineering structures and equipment, as well as monitoring data from measuring points, are linked to the three-dimensional model, the state of the gate and unit is converted into the opening and closing action of the gate and unit, the scheduling execution effect is converted into simulation scenarios such as flood discharge, power generation, and water level rise, and alarm and warning situations are converted into highlighted flashing points in the scene.

[0125] Data sharing module, provides information channels between hydropower station management departments and upper and lower management units, external related units, realizes cross-network, cross-industry, cross-level data sharing.

[0126] Follow the data sharing standards and specifications, form data exchange links, realize the reporting, issuing and synchronization of related data information between hydropower station management departments and upper and lower management units, external related units, realize cross-network, cross-industry, cross-level data sharing. Through clear data asset subject responsibility, authorization time granting and updating mechanism, etc., realize unified access, exchange state monitoring, data query, data subscription, visual report data sharing services, meet real-time data services, batch data sharing needs.

[0127] In the engineering safety prediction module, the input variables of the gradient boosting regression tree model for deformation prediction in key areas of hydropower station buildings include upstream water level, downstream water level, upstream and downstream water level difference, rainfall and environmental temperature, denoted as x ; the output variable is the deformation monitoring index, denoted as y .

[0128] The calculation process of the gradient boosting regression tree model is as follows:

[0129] a) The training set is denoted as , where represents the input variable of the i-th sample data, represents the output variable of the i-th sample data. i ​

[0130] The expression of the gradient boosting regression tree model is:

[0131] ;

[0132] wherein denotes the number of regression trees, denotes the parameters of the th regression tree, denotes the th regression tree classifier,

[0133] The expression of the th regression tree is:

[0134] ;

[0135] The gradient boosting regression tree model adopts a square loss function:

[0136] ;

[0137] b) Initialize the first regression tree :

[0138] ;

[0139] In the formula, argmin denotes the parameter that takes the minimum value, and c is the class of the root node;

[0140] c) For data samples i , i =1,2,3... N , calculate the negative gradient of the loss function as the estimate of the residual,

[0141] ;

[0142] Calculate the residual value , , and take it as the new true value of the sample, and as the next set of training data, fit a new regression tree, and obtain the leaf node area of the th regression tree, wherein is the serial number of the leaf node;

[0143] d) For =1,2,3... J , J is the number of leaf nodes, calculate the value of the leaf node area to minimize the loss function, and calculate the optimal fitting value:

[0144] ;

[0145] where represents the fitted value of the jth leaf node of the tth regression tree;

[0146] e) Update the trapezoidal regression tree model:

[0147]

[0148] where is an indicator function, and determines whether it belongs to the leaf node region

[0149] f) Get the final regression tree, that is, add the leaf node values of each tree:

[0150]

[0151] g) After the regression tree model is built, the training set is used to train the regression tree model to obtain the optimal trapezoidal regression tree model;

[0152] h) The optimal trapezoidal regression tree model is tested and evaluated using the validation set. The mean absolute error MAE, mean square error MSE, root mean square error RMSE, and determination coefficient R 2 of the regression tree model are calculated. The closer MAE, MSE, and RMSE are to 0, and the closer R 2 is to 1, the higher the fitting degree of the model to the data. According to the validation result, the final regression tree model is determined.

[0153] By analyzing the safety monitoring of the hydropower station building, the key cross sections and longitudinal sections that affect the safety of the building and have dense measuring points are selected. For each displacement measuring point on the longitudinal section and the cross section, a corresponding gradient regression tree model is constructed. The gradient regression tree model takes the predicted upstream and downstream water levels, rainfall, and temperature as inputs to calculate the displacement change process of each measuring point on the section within the prediction period.

[0154] After the calculation is completed, the prediction results need to be rendered to the three-dimensional scene. The specific steps are as follows: estimate the distribution density of the cross section measuring points through the measuring point coordinate positions; set a color gradient to map the displacement prediction data to different color bands and process the color value of each measuring point according to the numerical range; render the displacement prediction results on the server side through the visualization engine to form a heat map that can reflect the future deformation distribution of the key sections of the building.

[0155] Publish the rendering results to the WEB side for display, and fuse the heat map with the building twin scene. Drag the time axis to view the future deformation trend of the building at different times.

[0156] ​​​In the engineering safety pre-play module, the method of fusing stress and strain data obtained from the finite element analysis model with the digital twin scene includes:

[0157] a) Data preparation;

[0158] The finite element model calculation data is composed of node information, element information, and physical properties. The node information is composed of node number and x, y, z coordinates, and the element information is composed of element number and node number of the element. Each node or element will have corresponding calculated physical properties, such as displacement, stress, etc.

[0159] Before rendering the finite element results, the finite element model result file needs to be converted into a JSON data file. The JSON data file contains node coordinates, element indices, and result values.

[0160] Node coordinates is a two-dimensional array, where each element is a sub-array containing x, y, z coordinates.

[0161] Element index is a two-dimensional array, where each element is the index of three nodes or four nodes, used to form triangles or quadrilaterals.

[0162] Result value is an array of numerical values corresponding to triangles or quadrilaterals, including displacement values and stress values in x, y, z directions.

[0163] b) Data standardization processing;

[0164] In order to map the finite element calculation results to different color intervals, the displacement, stress and other result data need to be standardized to normalize them to (0, 1). The specific steps are:

[0165] Calculate the minimum and maximum values of the result values, subtract the minimum value from each result value, and divide by the difference between the maximum and minimum values to get the standardized result value.

[0166] c) Read JSON data;

[0167] Read the vertex coordinates, triangle indices, and calculation result data in the JSON file through the visualization engine to perform subsequent numerical calculation and processing.

[0168] d) Create triangle patches;

[0169] The solid elements of the finite element model cannot be directly rendered to the WEB, and need to be converted into renderable 3D graphics by the visualization engine. The node and element information of the finite element model is reconstructed, and the triangles and quadrilaterals contained in the mesh are split into triangle patches. For example, a hexahedral element can be split into 12 triangle patches, and a hexahedral node can be split into 36 triangle patch nodes.

[0170] e) defining a color mapping;

[0171] Map the displacement, stress, etc. result data to color values, set a color gradient, specify a color band for the physical property with RGB codes, and process the color values for each node according to the numerical range.

[0172] f) draw the triangular facets and color them;

[0173] Load the processed JSON data and triangular facets of the finite element object in the visualization engine, use the color mapping defined earlier, render the displacement, stress calculation data on the server side, and generate the corresponding heat map.

[0174] g) show the results;

[0175] Publish the rendered results to the WEB for display, you can view the displacement and stress distribution cloud map of the XYZ direction of the surface of the hydropower station building structure, and the static cloud map can be further rendered into a deformation animation for display.

[0176] At the same time, the system supports the model sectioning function, and the section formed after sectioning is displayed as a cloud map using the calculation node data inside the structure, which can view and analyze the mechanical properties of the entire model.

[0177] As shown in Figure 1 The engineering safety control method of the hydropower station safety comprehensive management system based on the digital twin scene includes the following steps:

[0178] Step 1: predict the deformation of the safety monitoring point, pre-play the displacement and stress strain of the hydropower station building structure, analyze the abnormality, and predict the safety risk;

[0179] Step 1.1: According to the deformation monitoring index of the hydropower station building, predict the deformation of the safety monitoring point of the hydropower station, predict the future deformation trend, predict the potential risk, and identify the abnormal measuring point;

[0180] Step 1.2: Select the working condition or load combination method, perform finite element analysis on the building structure, pre-play the displacement and stress strain of the building structure, and analyze the safety risk of the building structure according to the displacement and stress strain of the building structure obtained by pre-playing;

[0181] Step 2: If the measured value or predicted value of the monitoring index of the safety monitoring point exceeds the threshold value or the method of step 1 identifies an abnormality or safety risk, issue a warning information through the safety comprehensive management system of the application, display the warning time, associated measuring point, cause analysis and data change, distinguish the warning level by different colors, and locate the abnormal measuring point and risk area in the digital twin model;

[0182] Step 3: Search the pre-plan library for pre-plan information suitable for the early warning alarm scenario of Step 2, display the handling method and steps of the pre-plan, and give the user a recommended pre-plan for the early warning alarm scenario of Step 2;

[0183] Step 4: According to the pre-plan obtained in Step 3, take appropriate safety measures to handle the abnormality of Step 1, control safety risks, and ensure the safe operation of the hydropower station building structure.

[0184] In the embodiment, the deformation of the building structure of the ship lift of the hydropower station under the self-defined working condition is predicted, as shown in Figure 5 .

[0185] As shown in Figure 2 , the dispatching safety control method of the hydropower station safety comprehensive management system based on the digital twin scene includes the following steps:

[0186] Step 1: Predict and forecast the inflow of the hydropower station, the power generation output, and the water level and flow of the upstream and downstream river channels;

[0187] Step 2: Issue a warning for real-time water level exceeding the limit or predicted water level exceeding the limit or super-standard flood or meteorological disaster forecast;

[0188] Step 2.1: If the real-time water level or the predicted water level obtained in Step 1 exceeds the water level limit, trigger an alarm and issue a water level limit warning;

[0189] Step 2.2: If it is predicted in Step 1 that the inflow of the hydropower station will exceed the standard flood, issue a flood warning;

[0190] Step 2.3: Issue a meteorological disaster warning for meteorological disaster forecast;

[0191] Step 3: Search the pre-plan library to find a pre-plan suitable for the warning scenario of Step 2;

[0192] Step 4: Pre-act the pre-plan obtained in Step 3;

[0193] Step 4.1: Select the predicted inflow or the preset inflow; set the starting water level and control value;

[0194] Step 4.2: Calculate the change process of the outflow, reservoir water level, reservoir capacity, and output through the dispatching calculation model; calculate the water level and flow change process of the upstream and downstream river channels of the hydropower station through the hydrodynamic model;

[0195] Step 4.3: Compare the simulation calculation results of different dispatching strategies, and select the optimal dispatching scheme in combination with the pre-plan;

[0196] Step 5: Control the safe and stable operation of the hydropower station according to the optimal dispatching scheme obtained in Step 4.

[0197] In an embodiment, the reservoir dispatching pre-play of the hydropower station under flood warning is as shown in Figure 6 .

[0198] The power generation safety control method of the embodiment is as shown in Figure 3 , wherein the prediction results of the thrust pad temperature, the upper guide pad temperature, and the water guide pad temperature of the hydropower station are as shown in Figure 7 .

Claims

1. A safety comprehensive management system for a hydropower station based on a digital twin scene, characterized in that, The engineering digital twin base and an engineering safety subsystem, a dispatch safety subsystem and a power generation safety subsystem are included. The engineering safety subsystem uses an AI model to realize deformation prediction of key areas of the hydropower station building for deformation monitoring points of the hydropower station building. The finite element analysis model is used to calculate the displacement and stress distribution of the hydropower station building structure under design conditions, typical operating conditions, extreme conditions, self-defined conditions or different load combinations, and the displacement field and stress field of the hydropower station building structure are displayed in the form of rendered cloud maps. Based on the digital twin scene, abnormal measurement points and risk areas with safety risks are located and labeled. The engineering safety subsystem includes an engineering safety prediction module, an engineering safety early warning module, an engineering safety pre-play module and an engineering safety pre-plan module. The engineering safety prediction module constructs a gradient boosting regression tree model for deformation prediction of key areas of the hydropower station building for deformation monitoring indicators of the hydropower station building. The prediction results obtained by the gradient boosting regression tree model are rendered into the digital twin scene to form a heat map that can reflect the future deformation distribution of the key sections of the building, and the rendering results are published to the WEB end for display. The engineering safety early warning module issues warning information when the measured value or predicted value of the monitoring indicator exceeds the threshold, displays the warning time, associated measurement points, reasons and data changes, and distinguishes the warning levels by different colors, and locates abnormal measurement points and risk areas in the three-dimensional digital twin model of the engineering structure. The engineering safety pre-play module uses the finite element analysis model of the hydropower station building structure to simulate and calculate the overall displacement and stress-strain of the structure under the preset conditions and self-defined conditions. The user selects design conditions, typical operating conditions, extreme conditions, self-defined conditions or custom live loads, temperature loads and wind load combinations to calculate the displacement and stress distribution of the hydropower station building structure, and the displacement field and stress field of the hydropower station building structure are displayed in the form of rendered cloud maps. The finite element results are rendered by the visualization engine on the server side to generate corresponding heat maps, and the rendering results are published to the WEB end for display. When the engineering safety monitoring, prediction and pre-play results trigger an alarm, the engineering safety pre-plan module searches the built-in pre-plan library for pre-plan information suitable for the alarm scenario, displays the processing methods and steps of the pre-plan, and simulates the key links of the pre-plan, including material preparation, personnel evacuation, equipment inspection and operation dynamic process, using the simulation engine. The dispatch safety subsystem predicts the inflow process of the hydropower station, calculates the discharge flow, reservoir water level and unit output under different dispatching schemes, compares the calculation results of different dispatching strategies, selects the optimal dispatching scheme, and simulates the dynamic effects of gate discharge, unit power generation and flood detention through the digital twin engine. The power generation safety subsystem uses an AI model to predict the trend of unit equipment state indicators. When the unit equipment has safety risks, the abnormal measurement points and risk equipment are located in the digital twin scene, relevant pre-plans are recommended, and the pre-plan disposal process is simulated and practiced. The engineering digital twin base integrates all relevant geographic spatial data, basic data, monitoring data, business management data and external shared data of the hydropower station; through spatiotemporal multi-scale data mapping, a digital twin scene of the hydropower station is constructed to provide a three-dimensional visualization carrier for various business applications; The engineering digital twin base comprises a data acquisition module, a data management module, a data model module and a data sharing module; The data acquisition module integrates all relevant geographic spatial data, basic data, monitoring data, business management data and external shared data of the hydropower station according to a hydropower station data resource coding system; The data management module realizes resource directory management, data quality management and data security management based on a data standard system; The data model module constructs a unified basic data, monitoring data collection and two-three-dimensional integrated digital twin scene of the hydropower station through spatiotemporal multi-scale data mapping to provide a three-dimensional visualization carrier for various business applications; The data sharing module provides an information channel between the hydropower station management department and the upper and lower management units and external relevant units to realize cross-network, cross-industry and cross-level data sharing; The hydropower station data resource coding system adopts a three-level coding AB1B2C1C2D1D2.E1E2F1F2G1G2.H1H2H3, the first level coding AB1B2C1C2D1D2 is the type coding of the data resource, wherein A represents the major category of the data resource, the major category of the data resource includes basic data, monitoring data, business management data, cross-industry shared data and geographic spatial data; B1B2C1C2D1D2 represents the minor category of the data resource, wherein B1B2 represents the monitoring data type, the monitoring data type includes three types of engineering safety monitoring, environmental quantity monitoring and equipment operation monitoring; C1C2 represents the subdivided type of the monitoring data, the subdivided type of the monitoring data includes deformation, seepage, hydrology, meteorology, water machine equipment operation, electrical equipment operation and metal structure operation type; D1D2 represents the sub-type of the subdivided type of the monitoring data; The second level coding E1E2F1F2G1G2 is the hierarchical coding of the data resource, which is set according to the business level and the subordinate level; The third level coding H1H2H3 is the sequential coding of the data resource, which is coded according to the equipment sequence and the measuring point sequence; In the hydropower station safety comprehensive management system, remote sensing images, oblique photography and BIM modeling data are used to create a digital twin scene, in which the geographic information, civil engineering, electrical equipment and facilities, water turbine and metal structure of the hydropower station and its influence area are presented in the form of digital twin models; The construction process of the digital twin scene includes model data preparation, model editing and beautification and model data fusion.

2. The digital-twin-scene-based safety integrated management system for a hydropower station according to claim 1, characterized in that, The input variables of the gradient boosting regression tree model for deformation prediction of key areas of the hydropower station building in the engineering safety prediction module include upstream water level, downstream water level, upstream and downstream water level difference, rainfall and environmental temperature, denoted as x ; The output variable is a deformation monitoring index, denoted as y ; The calculation process of the gradient boosting regression tree model is as follows: a) The training set is denoted as ,in The input variable representing the i-th sample data is... Indicates the first i Output variables for each sample data; The expression of the gradient boosting regression tree model is as follows: ； wherein denotes the number of regression trees, denotes the parameters of the th regression tree, denotes the th regression tree classifier, The expression of one regression tree is: ； The gradient boosting regression tree model adopts a square loss function: ； b) initializing the first regression tree : ； In the formula, argmin represents the parameter that obtains the minimum value, and c is the class of the root node; c) for data samples i , i =1,2,3... N , compute the negative gradient of the loss function as an estimate of the residual, ； Compute residual values , , and as the new true values of the samples, as the next set of training data, fit a new regression tree to obtain the leaf node regions of the th regression tree , where is the leaf node number. d) for = 1,2,3... J , J is the number of leaf nodes, the values of the leaf nodes are computed to minimize the loss function, and the optimal fit value is computed: ； In the formula Ftj represents the fitted value of the jth leaf node of the tth regression tree. e) updating the trapezoidal regression tree model: ； In the formula is an indicator function, which for an input variable determines whether it belongs to the region of a leaf node ; f) obtaining the final regression tree, ； g) After the regression tree model is built, the regression tree model is trained using the training set to obtain an optimal trapezoidal regression tree model; h) The optimal trapezoidal regression tree model is tested and evaluated using the validation set.

3. The digital-twin-scene-based safety integrated management system for a hydropower station according to claim 1, characterized in that, In the engineering safety rehearsal module, the method for fusing stress and strain data obtained from the finite element analysis model with the digital twin scene includes: I. Data preparation; The calculation data of the finite element analysis model includes node information, element information and physical properties, the node information includes node number and x, y, z coordinates, the element information includes element number and node number of the composed element, and each node or element contains corresponding calculated physical properties; Before rendering the finite element analysis results, convert the file of the finite element analysis model calculation results into a JSON data file, and the JSON data file contains node coordinates, element index and result value; II. Data standardization processing; Standardize the displacement and stress data to normalize them to (0, 1); III. Read JSON data; Read the vertex coordinates, triangle index and calculation result data in the JSON file through the visualization engine; IV. Create a triangular patch; Reconstruct the node and element information of the finite element model, and split the triangles and quadrilaterals contained in the grid into triangular patches; V. Define color mapping; Map the displacement and stress data to color values, set the color gradient, specify the color band of the physical property with RGB code, and process the color value of each node according to the numerical range; VI. Draw triangular patches and color them; Load the JSON data and triangular patches of the finite element object in the visualization engine, use the previously defined color mapping, render the displacement and stress calculation data on the server side, and generate the corresponding heat map; VII. Achievement display; Publish the rendering results to the WEB side for display, view the displacement and stress distribution cloud map of the XYZ direction of the surface of the hydropower station building structure, and further render the static cloud map into a deformation animation for display through the interpolation algorithm.

4. The digital-twin-scene-based safety integrated management system for a hydropower station according to claim 1, characterized in that, The dispatching safety subsystem includes a dispatching safety prediction module, a dispatching safety early warning module, a dispatching safety rehearsal module and a dispatching safety plan module; The dispatching safety prediction module takes rainfall prediction as input, predicts the inflow of the hydropower station through a hydrological prediction model, takes the predicted inflow as input, predicts the change of the water level and flow of the upstream and downstream river sections through a hydrodynamic model, and takes the gate unit control mode as input, predicts the power generation output and capacity of the hydropower station through a dispatching calculation model; The dispatching safety early warning module compares the real-time water level with the characteristic water level, which includes the flood control water level, the normal storage water level, the design flood water level and the check flood water level, and triggers an alarm when the characteristic water level limit is exceeded; when receiving natural disaster warning information, display the warning information in a flashing and highlighted manner in the digital twin scene, the warning information includes warning level, predicted time and warning value, and send warning messages to relevant responsible persons; The dispatching safety pre-play module inputs the reservoir inflow process, and a user selects real-time, history, forecast, preset reservoir inflow for calculation, or uploads a template to perform self-defined reservoir inflow calculation; according to a dispatching regulation or a dispatching instruction, the user selects a control mode of water level, output, and reservoir outflow, inputs a starting water level and a control value; a dispatching calculation model is used to calculate a change process of a dispatching element; a hydrodynamic model is used to calculate a change process of a water level and a flow of a river section upstream and downstream of the hydropower station; the calculation result of the dispatching calculation model is integrated into a digital twin scene to simulate dynamic effects of flood evolution, gate discharge, unit power generation, water level rise and fall, and inundation influence; simulation calculation results of different dispatching strategies are compared to select an optimal dispatching scheme, and a best operation mode of a gate and a unit is given; The dispatching safety pre-plan module searches a built-in pre-plan library to find pre-plan information suitable for the scene when the hydropower station encounters a super-standard flood or a related accident, and shows disposal methods and steps of different levels and types of dangerous situations; a simulation engine is combined to simulate key links of the pre-plan, and simulation of the pre-plan includes dynamic processes of dispatching measures, material preparation, flood control and rescue, personnel evacuation, equipment inspection, and plugging and repair.

5. The digital-twin-scene-based safety integrated management system for a hydropower station according to claim 1, characterized in that, The power generation safety subsystem includes a power generation safety prediction module, a power generation safety early warning module, and a power generation safety pre-plan module; The power generation safety prediction module predicts a change trend of a state quantity index of important facility equipment of the hydropower station by using a gradient boosting regression tree or an AI model with equipment operation conditions as input; The power generation safety early warning module issues a warning information when a state quantity index of important equipment of the hydropower station exceeds a threshold value, reminds a user to pay attention in a flashing and highlighted manner, shows a warning time, an associated measuring point, data change, and disposal situation, and distinguishes warning levels by different colors and locates an abnormal measuring point in a three-dimensional digital twin model of the facility equipment; The power generation safety pre-plan module searches a built-in pre-plan library to find pre-plan information suitable for the alarm scene when the hydropower station encounters an alarm event, and shows disposal methods and steps of the pre-plan; a simulation engine is combined to simulate key links of the pre-plan, and simulation of the pre-plan includes dynamic processes of fault maintenance, personnel evacuation, and equipment control.

6. The digital-twin-scene-based safety integrated management system for a hydropower station according to claim 1, characterized in that, The construction process of the digital twin scene includes: (1) Model data preparation: fuse BIM models and three-dimensional model data of engineering structures and auxiliary mechanical and electrical equipment of the hydropower station, tilt photography model data of a surrounding influence area, and remote sensing, DOM / DEM model data, and convert various model data into a unified coordinate system and a unified three-dimensional general standard data format; (2) Model editing and beautification: use vector drawing to crop conflicting terrain images to realize seamless splicing and fusion of the digital twin model; and refine and map the textures of buildings, equipment, roads, terrain, vegetation, and water systems that do not conform to reality to beautify the presentation effect of the digital twin model; (3) Model data fusion: Establish the association between business, monitoring data and corresponding digital twin model, form the digital twin scene of the hydropower station, and let the outside world's effect on the physical entity be reflected in real time on the digital twin model; Hang the basic data of engineering structures, equipment and facilities and the monitoring data of measuring points to the digital twin model, convert the state of gates and units into the opening and closing actions of gates and units, convert the dispatch execution effect into the simulation scene of flood discharge, power generation and water level rise, and convert the alarm and early warning situation into a highlighted flashing point in the scene.

7. The engineering safety control method of the hydropower station safety comprehensive management system based on digital twin scene according to any one of claims 1-6, characterized in that, The method comprises the following steps: Step 1: Predict the deformation of the safety monitoring point, pre-play the displacement and stress-strain of the hydropower station building structure, analyze the abnormality, and predict the safety risk; Step 1.1: According to the deformation monitoring index of the hydropower station building, predict the deformation of the safety monitoring point of the hydropower station, predict the future deformation trend, predict the potential risk, and identify the abnormal measuring point; Step 1.2: Select the working condition or load combination mode, perform finite element analysis on the hydropower station building structure, pre-play the displacement and stress-strain of the building structure, analyze the safety risk of the hydropower station building structure according to the displacement and stress-strain of the building structure obtained by pre-playing; Step 2: If the measured value or predicted value of the monitoring index of the safety monitoring point exceeds the threshold value or the method of step 1 identifies and judges that it is abnormal and has a safety risk, the safety comprehensive management system issues an early warning information, displays the early warning time, associated measuring point, cause analysis and data change, and distinguishes the early warning level by color, and locates the abnormal measuring point and risk area in the digital twin model; Step 3: Search the preplan library to find the preplan information suitable for the early warning alarm scene of step 2, display the processing method and steps of the preplan, and give the recommended preplan for the early warning alarm scene of step 2 to the user; Step 4: According to the preplan obtained in step 3, take corresponding safety measures to handle the abnormality of step 1, control the safety risk, and ensure the safe operation of the hydropower station building structure.

Images