Comprehensive monitoring method and system for hydropower projects based on digital twins

Abstract

The present invention provides a method and system for comprehensive monitoring of hydropower projects based on digital twins, which relate to the technical field of hydropower project monitoring. A digital simulation model is constructed by monitoring parameters and a background map of the hydropower project. The monitoring parameters in the digital simulation model are optimized based on the background map to obtain a feedback particle group of the hydropower project, and an engineering knowledge map of the hydropower project is obtained. The gradient residual vector in the process of monitoring and modeling the hydropower project is determined based on the engineering knowledge map. The edge error of the feedback particle group is optimized based on the gradient residual vector to obtain an error compensation parameter in the process of monitoring and modeling the hydropower project. The digital simulation model is subjected to distortion compensation by the error compensation parameter to obtain a twin monitoring model of the hydropower project. The twin monitoring model is visualized to monitor the hydropower project, thereby solving the problem of low credibility of the existing digital simulation model corresponding to the operating status of the hydropower project. The present invention is suitable for monitoring hydropower projects.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower engineering monitoring, and in particular to a hydropower engineering comprehensive monitoring method and system based on digital twins. Background Art

[0002] Hydropower projects are infrastructure that rely on the energy of water flow to generate electricity. They usually include dams, reservoirs, turbines, generators, transmission equipment and other important components. Comprehensive monitoring of data in the monitoring area of ​​hydropower projects can help engineering personnel to timely understand the operating status of equipment, identify potential failures or risks, and provide data support for optimizing operating strategies.

[0003] Digital twin technology is a technology that reflects physical entities by establishing virtual digital simulation models. This technology is applied to the modeling and monitoring of the operating status of hydropower projects. By taking the digital simulation model of the hydropower project as the core component and collecting real-time monitoring data parameters on site, it can predict future operating trends. However, due to the lack of accuracy of the digital simulation model, error propagation, uncertainty of system parameters and other issues, the digital simulation model may have large deviations in actual applications, thus affecting the accuracy and credibility of project monitoring. Therefore, how to pre-compensate the distortion in the modeling process of the digital simulation model when monitoring hydropower projects in order to improve the credibility of comprehensive monitoring of hydropower projects has become a difficult problem faced by the industry. Summary of the invention

[0004] Technical problem solved by the present invention: The present invention provides a comprehensive monitoring method and system for hydropower projects based on digital twins to solve the problem of low credibility of digital simulation models corresponding to the operating status of existing hydropower projects.

[0005] The technical solution adopted by the present invention to solve the above technical problems is: a comprehensive monitoring method for hydropower projects based on digital twins, comprising the following steps:

[0006] S1. Determine the background map of the hydropower project and collect monitoring parameters of the hydropower project;

[0007] S2. constructing a digital simulation model based on the monitoring parameters and the background map of the hydropower project;

[0008] S3, optimizing the monitoring parameters in the digital simulation model based on the background image to obtain a feedback particle swarm of the hydropower project;

[0009] S4, obtaining an engineering knowledge graph of the hydropower project, and determining a gradient residual vector in the process of monitoring and modeling the hydropower project based on the engineering knowledge graph;

[0010] S5, performing edge error optimization on the feedback particle swarm based on the gradient residual vector to obtain error compensation parameters in the process of hydropower project monitoring modeling;

[0011] S6. Compensate the digital simulation model for distortion using error compensation parameters to obtain a twin monitoring model of the hydropower project, visualize the twin monitoring model, and monitor the hydropower project.

[0012] Furthermore, in S1, the monitoring parameters are collected by sensors, and the sensors include water level sensors, flow sensors, temperature sensors, humidity sensors, vibration sensors and pressure sensors. The monitoring parameters include water flow parameters, water level parameters, equipment operation parameters and environmental parameters. The equipment operation parameters include vibration parameters and pressure parameters. The environmental parameters include temperature parameters and humidity parameters.

[0013] Furthermore, in S1, the background image of the hydropower project is a digital orthophoto of an aerial photo of the hydropower project.

[0014] Furthermore, in S2, a digital simulation model is constructed according to the monitoring parameters and the background map of the hydropower project, including the following steps:

[0015] S21. Construct a digital simulation model space based on the background map of the hydropower project;

[0016] S22, classifying and aggregating the monitoring parameters to obtain regional information of different monitoring areas of the hydropower project and monitoring information of different monitoring categories of the hydropower project;

[0017] S23. Digitally reconstruct the regional information of different monitoring areas of the hydropower project and the monitoring information of different monitoring categories of the hydropower project to obtain a digital simulation model.

[0018] Furthermore, in S3, the feedback particle swarm of the hydropower project is output based on the background image, including the following steps:

[0019] S31, dividing the hydropower project into different feedback areas according to the background map;

[0020] S32, locally optimizing the monitoring parameters in each feedback zone to obtain feedback particles in each feedback zone;

[0021] S33. Determine the feedback particle group through all the feedback particles.

[0022] Furthermore, in S4, the engineering knowledge graph includes monitoring parameters, which are stored and managed through a graph database.

[0023] Further, in S4, determining the gradient residual vector in the process of hydropower engineering monitoring modeling based on the engineering knowledge graph includes the following steps:

[0024] S41. Determine the observed value of each monitoring parameter in the process of hydropower project monitoring modeling through a linear regression analysis model;

[0025] S42, determining the residual value of each monitoring parameter in the hydropower project monitoring modeling process according to the engineering knowledge graph and the observed value, and determining the gradient relationship of all residual values;

[0026] S43. Determine the gradient residual vector in the hydropower project monitoring modeling process based on the gradient relationship.

[0027] Further, in S5, the edge error optimization of the feedback particle swarm is performed based on the gradient residual vector, comprising the following steps:

[0028] S51, determining an update rate of each feedback particle in the feedback particle group based on a bidirectional gated recurrent neural network;

[0029] S52, drawing an update rate scatter plot through all update rates, and determining all edge penalty items in the feedback particle swarm based on the update rate scatter plot;

[0030] S53, optimizing all edge penalty items according to the gradient residual vector to obtain error compensation parameters.

[0031] Further, in S6, the digital simulation model is subjected to distortion compensation by using the error compensation parameter, comprising the following steps:

[0032] S61, using the monitoring parameters in the digital simulation model as feedback parameters;

[0033] S62, using the feedback parameter as the forward input of the BiGRU neural network, using the error compensation parameter as the backward input of the BiGRU neural network, and adaptively compensating the feedback parameter through the gating mechanism of the BiGRU neural network to obtain a distortion compensation value corresponding to the feedback parameter;

[0034] S63. Replace and compensate all distortion compensation values ​​with original feedback parameters in the digital simulation model to obtain a twin monitoring model of the hydropower project.

[0035] The present invention also provides a comprehensive monitoring system for hydropower projects based on digital twins, which realizes the comprehensive monitoring method for hydropower projects based on digital twins as described above. The system includes a data acquisition module, a digital simulation model establishment module, a model optimization module, a gradient residual vector acquisition module, an error compensation module and a visualization module; the data acquisition module is used to determine the background map of the hydropower project and collect the monitoring parameters of the hydropower project; the digital simulation model establishment module is used to construct a digital simulation model based on the monitoring parameters and the background map of the hydropower project; the model optimization module is used to optimize the monitoring parameters in the digital simulation model and output the feedback particle group of the hydropower project based on the background map; the gradient residual vector acquisition module is used to obtain the engineering knowledge graph of the hydropower project and determine the gradient residual vector in the process of monitoring modeling of the hydropower project based on the engineering knowledge graph; the error compensation module is used to optimize the edge error of the feedback particle group based on the gradient residual vector to obtain the error compensation parameter in the process of monitoring modeling of the hydropower project; the visualization module is used to perform distortion compensation on the digital simulation model through the error compensation parameter to obtain the twin monitoring model of the hydropower project, and visualize the twin monitoring model to monitor the hydropower project.

[0036] Beneficial effects of the invention: The present invention provides a comprehensive monitoring method and system for hydropower projects based on digital twins, which constructs a digital simulation model through monitoring parameters and a background map of the hydropower project, optimizes the monitoring parameters in the digital simulation model based on the background map to obtain a feedback particle group of the hydropower project, obtains an engineering knowledge graph of the hydropower project, determines the gradient residual vector in the process of hydropower project monitoring modeling based on the engineering knowledge graph, optimizes the edge error of the feedback particle group based on the gradient residual vector, obtains the error compensation parameter in the process of hydropower project monitoring modeling, compensates for the distortion of the digital simulation model through the error compensation parameter, obtains a twin monitoring model of the hydropower project, visualizes the twin monitoring model, and monitors the hydropower project, thereby solving the problem of low credibility of the existing digital simulation model corresponding to the operating status of the hydropower project. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of a comprehensive monitoring method for hydropower projects based on digital twins provided by the present invention;

[0038] Figure 2 It is a schematic diagram of a process of constructing a digital simulation model in a comprehensive monitoring method for hydropower projects based on digital twins provided by the present invention;

[0039] Figure 3It is a schematic diagram of a flow chart of optimizing the monitoring parameters in the digital simulation model based on a background image to obtain a feedback particle swarm of a hydropower project in a comprehensive monitoring method of a hydropower project based on a digital twin provided by the present invention;

[0040] Figure 4 It is a schematic diagram of a flow chart of determining a gradient residual vector in a hydropower project monitoring modeling process according to the engineering knowledge graph in a hydropower project comprehensive monitoring method based on digital twins provided by the present invention;

[0041] Figure 5 It is a schematic diagram of a process of optimizing the edge error of the feedback particle swarm based on the gradient residual vector in a comprehensive monitoring method for hydropower projects based on digital twins provided by the present invention;

[0042] Figure 6 It is a flow chart of performing distortion compensation on the digital simulation model through error compensation parameters in a comprehensive monitoring method for hydropower projects based on digital twins provided by the present invention. DETAILED DESCRIPTION

[0043] In order to solve the problem of low credibility of digital simulation models corresponding to the operating status of existing hydropower projects, the present invention provides a comprehensive monitoring method and system for hydropower projects based on digital twins. The core of the method and system is as follows: a digital simulation model is constructed through monitoring parameters and a background map of the hydropower project, the monitoring parameters in the digital simulation model are optimized based on the background map to obtain a feedback particle swarm of the hydropower project, an engineering knowledge graph of the hydropower project is obtained, a gradient residual vector in the process of monitoring and modeling of the hydropower project is determined according to the engineering knowledge graph, edge error optimization of the feedback particle swarm is performed based on the gradient residual vector, error compensation parameters in the process of monitoring and modeling of the hydropower project are obtained, distortion compensation is performed on the digital simulation model through the error compensation parameters, a twin monitoring model of the hydropower project is obtained, the twin monitoring model is visualized, and the hydropower project is monitored, thereby improving the accuracy of the twin monitoring model of the hydropower project.

[0044] like Figure 1 As shown, the present invention provides a comprehensive monitoring method for hydropower projects based on digital twins, comprising the following steps:

[0045] S1. Determine the background map of the hydropower project and collect monitoring parameters of the hydropower project.

[0046] Specifically, the background map of the hydropower project is a digital orthophoto of the hydropower project aerial photo, which provides a spatial reference for subsequent models. The digital orthophoto can be used to locate the monitoring area of ​​the hydropower project, such as dams, pipelines, generators, etc. The monitoring area of ​​the hydropower project can be photographed by drone, and the aerial image can be geometrically corrected and coordinate calibrated through existing modeling software to obtain a digital orthophoto.

[0047] The monitoring data is collected by sensors, and the sensors include water level sensors, flow sensors, temperature sensors, humidity sensors, vibration sensors and pressure sensors. The monitoring data include water flow parameters, water level parameters, equipment operation parameters and environmental parameters. The equipment operation parameters include vibration parameters and pressure parameters, and the environmental parameters include temperature parameters and humidity parameters. According to the characteristics of different monitoring points, arrays composed of corresponding sensors are installed, and the deployment is based on the specific needs of the hydropower project monitoring plan. For example, a sensor array composed of water level sensors and vibration sensors is installed in the dam area, a sensor array composed of flow sensors and pressure sensors is installed in the pipeline area, an array composed of temperature sensors and humidity sensors is installed in the generator set area, and so on.

[0048] S2. Constructing a digital simulation model based on the monitoring parameters and the background map of the hydropower project.

[0049] Specifically, a digital simulation model is constructed based on the monitoring parameters and the background map of the hydropower project, such as Figure 2 As shown, the following steps are included:

[0050] S21. Construct a digital simulation model space based on the background map of the hydropower project.

[0051] Specifically, the monitoring area or monitoring point can be accurately located from the background image.

[0052] S22. Classify and aggregate the monitoring parameters to obtain regional information of different monitoring areas of the hydropower project and monitoring information of different monitoring categories of the hydropower project.

[0053] Specifically, the monitoring parameters are classified and divided based on different monitoring areas of the hydropower project through a decision tree algorithm to obtain regional information of different monitoring areas of the hydropower project. The different monitoring areas of the hydropower project include: dams, pipelines, generator sets, substations and other areas.

[0054] The monitoring parameters are normalized and then input into a support vector machine model, and the support vector machine model is used to output monitoring information of different monitoring categories of the hydropower project, wherein the different monitoring categories of the hydropower project include: water flow, water level, equipment operation parameters and environmental parameters.

[0055] In this way, complex data patterns can be identified through the decision tree algorithm and support vector machine model, which is conducive to more accurate classification and aggregation of monitoring areas and monitoring categories.

[0056] S23. Digitally reconstruct the regional information of different monitoring areas of the hydropower project and the monitoring information of different monitoring categories of the hydropower project to obtain a digital simulation model.

[0057] The regional information of different monitoring areas of the hydropower project and the monitoring information of different monitoring categories of the hydropower project are digitally reconstructed through the spline interpolation algorithm to obtain a digital simulation model. The digital simulation model contains a basis function matrix and node distribution parameters, and can perform surface fitting on discrete monitoring data parameters. In specific implementation, first, a spline interpolation model is initialized based on the spline interpolation algorithm; secondly, the regional information of different monitoring areas of the hydropower project and the monitoring information of different monitoring categories of the hydropower project are respectively used as input information of the spline interpolation model; then, the regional rule data corresponding to the regional information of different monitoring areas of the hydropower project and the category rule data corresponding to the monitoring information of different monitoring categories of the hydropower project are output through the spline interpolation model. Finally, the regional rule data and category rule data under different rule constraints are output through the spline interpolation model, and the output regional rule data and category rule data are imported into the existing digital twin engine, such as Unity3D, to construct a digital simulation model of the hydropower project in a data-driven manner. The digital simulation model contains static parameters and dynamic parameters, wherein the static parameters include engineering structure size, equipment model, etc., and the dynamic parameters include monitoring parameters. Among them, the regional rule data can reflect the parameter distribution under the boundary constraints of each monitoring area, such as the water flow velocity gradient distribution in the generator set area, and the category rule data can describe the quantitative relationship between different monitoring categories in space, such as the spatial coupling characteristics of water level parameters and temperature parameters; the digital simulation model is a model that converts the characteristics, dynamic changes and interactions of physical objects, systems or environments into digital form. The digital simulation model can be virtually simulated during the monitoring process of hydropower projects. By adjusting the parameters in the digital simulation model, it is beneficial to monitor the operating status of the hydropower project and optimize the operating efficiency of the hydropower project. However, due to the existence of errors, the credibility of the digital simulation model is low, so it still needs to be optimized.

[0058] S3. Optimizing the monitoring parameters in the digital simulation model based on the background image to obtain a feedback particle swarm of the hydropower project.

[0059] Specifically, Figure 3 As shown, the following steps are included:

[0060] S31. Divide the hydropower project into different feedback areas according to the background map.

[0061] Specifically, the existing edge segmentation algorithm is used to divide the background image into different feedback areas according to the monitoring area, such as: reservoir area, power generation area and equipment area.

[0062] S32, locally optimizing the monitoring parameters in each feedback zone to obtain feedback particles in each feedback zone;

[0063] Specifically, a multimodal optimization algorithm is used to coordinately optimize the monitoring parameters in the digital simulation model, and a particle swarm algorithm is used to locally optimize the monitoring parameters in each feedback area to obtain the local optimal solution of each feedback area, and the local optimal solution is used as the feedback particle. The feedback particle swarm realizes the dynamic coordinated adjustment of water flow parameters, water level parameters, equipment operation parameters and environmental parameters through a multimodal optimization algorithm.

[0064] S33. Determine the feedback particle group through all the feedback particles.

[0065] Specifically, the composition set of all feedback particles is taken as the feedback particle group of the hydropower project.

[0066] S4. Obtain an engineering knowledge graph of the hydropower project, and determine a gradient residual vector in the process of monitoring and modeling the hydropower project based on the engineering knowledge graph.

[0067] Specifically, the engineering knowledge graph of the hydropower project can be obtained from the graph database of the hydropower project. The engineering knowledge graph refers to a structured knowledge base in the hydropower project. The engineering knowledge graph contains knowledge in multiple dimensions such as engineering data, models, design standards, equipment parameters, monitoring parameters, etc. related to the hydropower project. That is, the engineering knowledge graph contains monitoring parameters, which are stored and managed through the graph database, so that the relationship between different types of data can be stored through the structure of nodes in the graph database.

[0068] The gradient residual vector in the hydropower project monitoring modeling process is determined based on the engineering knowledge graph, such as Figure 4 As shown, the following steps are included:

[0069] S41. Determine the observed value of each monitoring parameter in the process of hydropower project monitoring modeling through linear regression analysis model.

[0070] Specifically, each monitoring parameter refers to each of a water flow parameter, a water level parameter, a vibration parameter, a pressure parameter, a temperature parameter, and a humidity parameter. Each collected parameter is substituted into a linear regression analysis model for linear regression analysis, and then the linear regression analysis model outputs the observation value corresponding to each monitoring parameter.

[0071] S42. Determine the residual value of each monitoring parameter in the hydropower project monitoring modeling process according to the engineering knowledge graph and the observation value, and determine the gradient relationship of all residual values.

[0072] Specifically, the estimated value corresponding to each monitoring parameter in the process of hydropower project monitoring modeling is output through the autoregression analysis model, and the difference between the estimated value corresponding to each monitoring parameter and the observed value corresponding to the parameter is used as the residual value of the parameter, thereby obtaining the residual value of each monitoring parameter in the process of hydropower project monitoring modeling. The autoregression analysis model is an existing statistical model for processing time series data. It uses a fractional normalization formula to normalize all residuals, and calculates the gradient relationship of the standardized residuals through the finite difference method.

[0073] S43. Determine the gradient residual vector in the hydropower project monitoring modeling process based on the gradient relationship.

[0074] Specifically, the gradient relationship includes the dimension of all residual values, which can be used as the direction of the gradient residual vector. The gradient relationship also includes the change amplitude of all residual values, which can be used as the size of the gradient residual vector, thereby obtaining the gradient residual vector in the process of hydropower project monitoring modeling. Determining the gradient residual vector is conducive to reflecting the change trend of the monitoring parameter error in all monitoring parameter dimensions, and is convenient for parameter calibration for the monitoring parameter error to ensure the matching of parameter values. The gradient residual vector can provide an error compensation basis for the subsequent optimization and update of the digital simulation model, thereby optimizing the model and improving accuracy.

[0075] S5. Optimizing the edge error of the feedback particle swarm based on the gradient residual vector to obtain error compensation parameters in the process of hydropower project monitoring modeling.

[0076] Specifically, the edge error of the feedback particle swarm is optimized based on the gradient residual vector, such as Figure 5 As shown, the following steps are included:

[0077] S51. Determine an update rate of each feedback particle in the feedback particle group based on a bidirectional gated recurrent neural network.

[0078] Specifically, the bidirectional gated recurrent neural network, namely the BiGRU neural network, is composed of forward and backward GRU units. The bidirectional structure of the BiGRU neural network can simultaneously capture past information and predict future information. In the BiGRU neural network, each GRU unit has a gating mechanism to control the flow of information, wherein the gating mechanism includes an update gate and a reset gate. The update gate determines whether the input at the current moment updates the current state, and the reset gate determines how to combine the past state with the current input. Therefore, the feedback particle group can be used as the forward input of the BiGRU neural network, and the update rate of each feedback particle in the feedback particle group can be adaptively identified through the gating mechanism of the BiGRU neural network.

[0079] S52: Draw an update rate scatter plot using all update rates, and determine all edge penalty items in the feedback particle swarm based on the update rate scatter plot.

[0080] Specifically, an update rate scatter plot is drawn based on all update rates through a visualization tool, and the update rate scatter plot is used to describe the changes in update rates of different feedback particles. For example, the update rate scatter plot can be drawn using the Matplotlib library, which is conducive to intuitively showing the changes in update rates of different feedback particles.

[0081] In the drawn update rate scatter plot, the slope of the lines connecting all adjacent update rates is calculated by the slope calculation formula, and the lines whose slopes are greater than the average of the slopes of all lines are taken as edge lines, and then the feedback particles corresponding to the update rates at both ends of all edge lines are taken as edge penalty items, where the edge penalty item refers to the monitoring parameters that change greatly or slightly in the process of hydropower project monitoring modeling. Optimizing the edge penalty item is conducive to improving the accuracy of hydropower project monitoring modeling.

[0082] S53, optimizing all edge penalty items according to the gradient residual vector to obtain error compensation parameters.

[0083] Specifically, the gradient boosting algorithm is used to optimize the errors of all edge penalty items according to the gradient residual vector. The specific process includes: taking the direction of the gradient residual vector as the optimization direction of the gradient boosting algorithm, taking the size of the gradient residual vector as the optimization step size of the gradient boosting algorithm, and optimizing all edge penalty items by setting the optimization direction and optimization step size of the gradient boosting algorithm, and then taking all the optimized edge penalty items as error compensation parameters to obtain error compensation parameters. By combining the dynamic adjustment of the gradient boosting algorithm, the modeling error caused by excessive updating of monitoring parameter changes can be avoided.

[0084] S6. Compensate the digital simulation model for distortion using error compensation parameters to obtain a twin monitoring model of the hydropower project, visualize the twin monitoring model, and monitor the hydropower project.

[0085] Specifically, the digital simulation model is subjected to distortion compensation through error compensation parameters, such as Figure 6 As shown, the following steps are included:

[0086] S61. Using the monitoring parameters in the digital simulation model as feedback parameters.

[0087] S62, using the feedback parameter as the forward input of the BiGRU neural network, using the error compensation parameter as the backward input of the BiGRU neural network, and adaptively compensating the feedback parameter through the gating mechanism of the BiGRU neural network to obtain a distortion compensation value corresponding to the feedback parameter.

[0088] S63. Replace and compensate all distortion compensation values ​​with original feedback parameters in the digital simulation model to obtain a twin monitoring model of the hydropower project.

[0089] The distortion compensation value output by the BiGRU neural network can describe the error of the original feedback parameter, which is conducive to eliminating the monitoring parameter error caused by environmental changes, equipment failure and other factors, and improving the credibility of hydropower project monitoring.

[0090] The twin monitoring model is visualized through a visualization platform or a visualization tool. In specific implementation, Three.js in the prior art can be used to visualize the twin monitoring model.

[0091] It can be seen that in the present application, the digital simulation model is distorted and compensated by error compensation parameters to obtain a twin monitoring model of the hydropower project; first, by classifying, aggregating and digitally reconstructing the collected monitoring parameters, a digital simulation model that can dynamically reflect the operating status of the hydropower project is provided, and the monitoring parameters in the digital simulation model are optimized based on the background image to obtain the feedback particle group of the hydropower project, which is convenient for matching the actual working conditions; secondly, the gradient residual vector is determined by referring to the engineering knowledge graph, and the deviation between the monitoring parameters and the actual parameters is analyzed by the gradient residual vector, and the feedback particle group is optimized based on the deviation obtained by the analysis to obtain the error compensation parameter, which can avoid the modeling error caused by excessive updating of parameter changes; finally, the digital simulation model is distorted and compensated by the error compensation parameter to obtain the twin monitoring model of the hydropower project, which can reduce the distortion of the monitoring data caused by environmental changes or equipment failures to ensure the accuracy of the twin monitoring model; in summary, the present application scheme can compensate for the distortion in the twin modeling process during the monitoring of the hydropower project to improve the credibility of the comprehensive monitoring of the hydropower project.

[0092] The present invention also provides a comprehensive monitoring system for hydropower projects based on digital twins, which realizes the above-mentioned comprehensive monitoring method for hydropower projects based on digital twins. The system includes a data acquisition module, a digital simulation model establishment module, a model optimization module, a gradient residual vector acquisition module, an error compensation module and a visualization module; the data acquisition module is used to determine the background map of the hydropower project and collect the monitoring parameters of the hydropower project; the digital simulation model establishment module is used to construct a digital simulation model based on the monitoring parameters and the background map of the hydropower project; the model optimization module is used to optimize the monitoring parameters in the digital simulation model and output the feedback particle group of the hydropower project based on the background map; the gradient residual vector acquisition module is used to obtain the engineering knowledge graph of the hydropower project, and determine the gradient residual vector in the process of monitoring modeling of the hydropower project based on the engineering knowledge graph; the error compensation module is used to optimize the edge error of the feedback particle group based on the gradient residual vector to obtain the error compensation parameter in the process of monitoring modeling of the hydropower project; the visualization module is used to perform distortion compensation on the digital simulation model through the error compensation parameter to obtain the twin monitoring model of the hydropower project, and visualize the twin monitoring model to monitor the hydropower project.

Claims

1. A comprehensive monitoring method for hydropower projects based on digital twins, characterized in that: The following steps are involved: S1. Determine the background map of the hydropower project and collect monitoring parameters of the hydropower project, wherein the background map of the hydropower project is a digital orthophoto of an aerial photo of the hydropower project; S2. constructing a digital simulation model based on the monitoring parameters and the background map of the hydropower project; S3, optimizing the monitoring parameters in the digital simulation model based on the background image to obtain a feedback particle swarm of the hydropower project; S4, obtaining an engineering knowledge graph of a hydropower project, and determining a gradient residual vector in a hydropower project monitoring modeling process based on the engineering knowledge graph, including the following steps: S41. Determine the observed value of each monitoring parameter in the process of hydropower project monitoring modeling through a linear regression analysis model; S42, determining the residual value of each monitoring parameter in the process of hydropower project monitoring modeling according to the engineering knowledge graph and the observed value, and determining the gradient relationship of all residual values; outputting the estimated value corresponding to each monitoring parameter in the process of hydropower project monitoring modeling through the autoregressive analysis model, and then taking the difference between the estimated value corresponding to each monitoring parameter and the observed value corresponding to the parameter as the residual value of the parameter, thereby obtaining the residual value of each monitoring parameter in the process of hydropower project monitoring modeling; S43, determining the gradient residual vector in the process of hydropower project monitoring modeling according to the gradient relationship; S5, performing edge error optimization on the feedback particle swarm based on the gradient residual vector to obtain error compensation parameters in the process of hydropower project monitoring modeling, including the following steps: S51, determining an update rate of each feedback particle in the feedback particle group based on a bidirectional gated recurrent neural network; S52, drawing an update rate scatter plot through all update rates, and determining all edge penalty items in the feedback particle swarm based on the update rate scatter plot; the edge penalty item refers to the monitoring parameter that changes greatly or slightly during the hydropower project monitoring modeling process; S53, optimizing all edge penalty items according to the gradient residual vector to obtain error compensation parameters; S6, compensating the distortion of the digital simulation model by using error compensation parameters to obtain a twin monitoring model of the hydropower project, visualizing the twin monitoring model, and monitoring the hydropower project; compensating the distortion of the digital simulation model by using error compensation parameters to obtain a twin monitoring model of the hydropower project, including the following steps: S61, using the monitoring parameters in the digital simulation model as feedback parameters; S62, using the feedback parameter as the forward input of the BiGRU neural network, using the error compensation parameter as the backward input of the BiGRU neural network, and adaptively compensating the feedback parameter through the gating mechanism of the BiGRU neural network to obtain a distortion compensation value corresponding to the feedback parameter; S63. Replace and compensate all distortion compensation values ​​with original feedback parameters in the digital simulation model to obtain a twin monitoring model of the hydropower project.

2. The method for comprehensive monitoring of hydropower projects based on digital twins according to claim 1 is characterized in that: In S1, the monitoring parameters are collected by sensors, and the sensors include water level sensors, flow sensors, temperature sensors, humidity sensors, vibration sensors and pressure sensors. The monitoring parameters include water flow parameters, water level parameters, equipment operation parameters and environmental parameters. The equipment operation parameters include vibration parameters and pressure parameters. The environmental parameters include temperature parameters and humidity parameters.

3. The method for comprehensive monitoring of hydropower projects based on digital twins according to claim 1 is characterized in that: In S2, a digital simulation model is constructed according to the monitoring parameters and the background map of the hydropower project, including the following steps: S21. Construct a digital simulation model space based on the background map of the hydropower project; S22, classifying and aggregating the monitoring parameters to obtain regional information of different monitoring areas of the hydropower project and monitoring information of different monitoring categories of the hydropower project; S23. Digitally reconstruct the regional information of different monitoring areas of the hydropower project and the monitoring information of different monitoring categories of the hydropower project to obtain a digital simulation model.

4. The method for comprehensive monitoring of hydropower projects based on digital twins according to claim 1 is characterized in that: In S3, the monitoring parameters in the digital simulation model are optimized based on the background image to obtain a feedback particle swarm of the hydropower project, including the following steps: S31, dividing the hydropower project into different feedback areas according to the background map; S32, locally optimizing the monitoring parameters in each feedback zone to obtain feedback particles in each feedback zone; S33. Determine the feedback particle group through all the feedback particles.

5. The method for comprehensive monitoring of hydropower projects based on digital twins according to claim 1 is characterized in that: In S4, the engineering knowledge graph includes monitoring parameters and is stored and managed through a graph database.

6. The integrated monitoring system for hydropower projects based on digital twins is characterized by: To implement the comprehensive monitoring method for hydropower projects based on digital twins as described in claim 1, the system includes a data acquisition module, a digital simulation model building module, a model optimization module, a gradient residual vector acquisition module, an error compensation module and a visualization module; the data acquisition module is used to determine the background map of the hydropower project and collect the monitoring parameters of the hydropower project; the digital simulation model building module is used to construct a digital simulation model based on the monitoring parameters and the background map of the hydropower project; the model optimization module is used to optimize the monitoring parameters in the digital simulation model and output the feedback particle swarm of the hydropower project based on the background map; The gradient residual vector acquisition module is used to acquire the engineering knowledge graph of the hydropower project, and determine the gradient residual vector in the hydropower project monitoring modeling process based on the engineering knowledge graph; the error compensation module is used to optimize the edge error of the feedback particle swarm based on the gradient residual vector to obtain the error compensation parameters in the hydropower project monitoring modeling process; the visualization module is used to compensate for the distortion of the digital simulation model through the error compensation parameters to obtain the twin monitoring model of the hydropower project, and visualize the twin monitoring model to monitor the hydropower project.

Images