Digital twin hydraulic engineering operation and maintenance monitoring system and method

Abstract

The invention relates to the technical field of computers, and discloses a digital twin hydraulic engineering operation and maintenance monitoring system which comprises a data acquisition module, a twin modeling module, a principal stress extraction module, a stress evolution prediction module, a safety domain judgment module, a regulation and control decision module, a control execution module and a state feedback module. The invention also discloses a digital twin hydraulic engineering operation and maintenance monitoring method, which comprises the following steps: data acquisition: real-time sensing data of a hydraulic engineering structure area is acquired through the data acquisition module, and the sensing data comprises strain, water pressure, temperature and displacement information and is used for representing the current state of the structure; and carrying out twin modeling, and receiving the real-time sensing data transmitted by the data acquisition module. According to the method, intelligent sensing and dynamic regulation and control of the structure state are realized by constructing a closed-loop system of principal stress prediction, regulation and control decision, control execution and state feedback.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and specifically to a digital twin water conservancy project operation and maintenance monitoring system and method. Background Technique

[0002] During the long-term operation of large-scale water conservancy projects, structural safety issues have always been the core factors affecting the project's stability and service life; due to the uncertainty of the hydrological environment, the nonlinearity of load effects, and the annual degradation of material properties, the structural stress state is often in a dynamic change. If these changes cannot be sensed and response-controlled in a timely manner, it is extremely easy to cause local damage or even structural failure.

[0003] At present, some structural monitoring systems can realize the basic perception of the operation status of water conservancy facilities. The typical method is to deploy sensors such as stress, displacement, and water level, and perform over-limit alarms through threshold setting. Such systems have achieved good results in improving the visualization ability of structural risks. Some technologies also support simple sluice control or pump station linkage, and have a certain engineering response ability. There are also some Internet of Things-based data acquisition platforms that can realize remote monitoring and early warning push, significantly improving the convenience and controllability of operation management.

[0004] However, the existing technologies still face many key limitations, especially in the generation and feedback update of control strategies; on the one hand, the existing methods mostly rely on static threshold triggering judgments, cannot capture the evolution trend of the principal stress, and are prone to judgment delays under sudden load actions. On the other hand, the current control methods are generally relatively rigid, often only facing single device operations, lacking the perception of the overall structural stress field, and there are blind spots in the control results. At the same time, most systems do not build a closed-loop feedback mechanism, and the actual structural response after regulation cannot be transmitted back for model update, resulting in the digital model gradually deviating from the real state, and the prediction accuracy decaying over time. Summary of the Invention

[0005] Aiming at the deficiencies of the existing technologies, the present invention provides a digital twin water conservancy project operation and maintenance monitoring system and method, which solves the problems of lagging structural state evaluation, single control strategy, lack of feedback closed-loop, and inability to dynamically update the digital model in the existing technologies.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A digital twin water conservancy project operation and maintenance monitoring system, including:

[0007] A data acquisition module, used to collect real-time sensing data of the water conservancy project structure area, and the real-time sensing data includes structural strain, water pressure, temperature, and displacement information;

[0008] The twin modeling module receives the real-time sensing data transmitted by the data acquisition module, constructs a stress tensor model of the key structural area based on the pre-built three-dimensional finite element model, and then generates the three-dimensional stress field information inside the structure;

[0009] The principal stress extraction module receives the stress tensor model, extracts the principal stress trajectories at the key structural positions, and generates the principal stress spectrum data for structural safety assessment;

[0010] The stress evolution prediction module predicts the change trend of the principal stress at each key structural position within the future time interval based on the principal stress spectrum data, and sends the principal stress prediction result to the safety domain judgment module;

[0011] The safety domain judgment module receives the principal stress prediction result, compares and analyzes whether the future principal stress exceeds the safety limit according to the preset structural safety stress threshold. If it is judged that there is a risk of principal stress exceeding the limit, it sends a control trigger instruction to the regulation decision-making module;

[0012] The regulation decision-making module, after receiving the control trigger instruction, generates a control strategy for optimizing the structural stress state according to the principal stress change trend and the structural operation parameters;

[0013] The control execution module controls the physical execution devices in the water conservancy project to perform regulation operations according to the received control strategy. The execution devices include the sluice opening adjustment device, the pump station control device, and the structural unloading device;

[0014] The state feedback module collects the structural state data after regulation and feeds the data back to the twin modeling module to update the digital twin system model, realizing the closed-loop control of principal stress prediction and structural regulation.

[0015] Preferably, the data acquisition module includes:

[0016] The strain acquisition unit is used to obtain the strain change information on the surface or inside of the structure;

[0017] The water pressure monitoring unit is used to collect the acting force of the water body around the structure on the surface of the structure;

[0018] The temperature acquisition unit is used to collect the temperature data of the environment and inside the structural materials;

[0019] The displacement measurement unit is used to collect the absolute or relative displacement changes of the key parts of the structure.

[0020] Preferably, the twin modeling module includes:

[0021] The finite element modeling unit is used to construct the static or dynamic mechanical model of the structure;

[0022] A tensor calculation unit for dynamically calculating the stress tensor distribution according to the sensing data;

[0023] A model correction unit for comparing the real-time monitoring data with the simulation model and correcting the calculation result of the stress field.

[0024] Preferably, the principal stress extraction module includes:

[0025] A tensor eigenvalue analysis unit for extracting the principal stress values of the stress tensor;

[0026] A stress trajectory construction unit for constructing an evolution trajectory based on the principal stress values at multiple time points;

[0027] A spectrogram generation unit for generating a principal stress spectrogram or a principal stress distribution map reflecting the structural state.

[0028] Preferably, the stress evolution prediction module includes:

[0029] A time series modeling unit for modeling the historical sequence of principal stresses;

[0030] A trend prediction unit for predicting the future change of the principal stress based on the current state and the historical evolution trend;

[0031] A prediction result output unit for transmitting the prediction result to the safety domain judgment module.

[0032] Preferably, the safety domain judgment module includes:

[0033] A threshold setting unit for setting the safety stress limit of the structural material;

[0034] A principal stress comparison unit for comparing the predicted principal stress with the set threshold;

[0035] A regulation determination unit for judging whether there is an overlimit risk according to the comparison result and generating a control signal.

[0036] Preferably, the regulation decision module includes:

[0037] A control variable definition unit for determining the external parameters available for regulation;

[0038] An objective function construction unit for constructing a regulation optimization objective based on the structural safety and the operation cost;

[0039] A strategy solving unit for solving the optimal control strategy by using an optimization algorithm.

[0040] Preferably, the control execution module includes:

[0041] An instruction parsing unit for converting the control strategy into executable instructions;

[0042] An execution device control unit for sending control instructions to physical devices of a water conservancy project;

[0043] A regulation effect detection unit for feedback confirmation of the device state after execution.

[0044] Preferably, the state feedback module includes:

[0045] A state perception unit for real-time perception of the stress and deformation states of the structure after control;

[0046] A data synchronization unit for transmitting the feedback state back to the twin model update module;

[0047] A cyclic update unit for starting a new prediction and control cycle.

[0048] The present invention also provides a digital twin water conservancy project operation and maintenance monitoring method, including the following steps:

[0049] Data acquisition, obtaining real-time sensing data of the water conservancy project structure area through a data acquisition module, where the sensing data includes strain, water pressure, temperature, and displacement information, used to characterize the current state of the structure;

[0050] Twin modeling, receiving the real-time sensing data transmitted by the data acquisition module, and constructing a stress tensor model of the water conservancy project structure area in the twin modeling module based on a pre-constructed three-dimensional finite element model, generating corresponding three-dimensional stress field information;

[0051] Principal stress extraction, extracting principal stress trajectory data from the stress tensor model generated by the twin modeling module, where the principal stress trajectory reflects the stress changes at key positions of the structure, forming a principal stress spectrum for subsequent analysis;

[0052] Stress evolution prediction, using a stress evolution prediction module to predict the trend of the principal stress trajectory, and predicting the stress change trend of the structure within a future period based on historical data and the current principal stress state;

[0053] Safety assessment and judgment, based on the predicted principal stress trajectory, through a safety domain judgment module, comparing and analyzing with the structural safety stress threshold to determine whether there is a risk of principal stress exceeding the limit;

[0054] Control strategy construction, if it is determined that the principal stress exceeds the limit, the regulation decision-making module generates an optimal control strategy according to the over-limit conditions and the operating parameters of the structure, and the control strategy aims to adjust the stress state of the structure;

[0055] Execution of regulation and feedback, executing the control strategy through a control execution module, adjusting the physical devices in the water conservancy project, and collecting feedback state data after regulation through a state feedback module;

[0056] Model update and loop, transmit the feedback status data back to the twin modeling module, update the digital twin model, and achieve continuous monitoring and optimization adjustment of the structure.

[0057] The present invention provides a digital twin operation and maintenance monitoring system and method for water conservancy projects. It has the following beneficial effects:

[0058] 1. The present invention adopts a prediction and judgment model based on the evolution trend of principal stress, and constructs a corresponding dynamic identification mechanism for the safety domain, achieving the technical effect of real-time identification and early warning of the structural operation risk. Compared with the method relying on a single stress threshold discrimination in the prior art, it solves the problems of lagging risk judgment and large error under dynamic working conditions.

[0059] 2. The present invention introduces a regulation decision-making module. By perceiving the current operating parameters and stress changes of the structure, it actively generates an adaptive regulation strategy to effectively intervene in the structural state under complex loads. Traditional solutions usually only trigger a simple response through a preset threshold and cannot adjust the strategy according to the stress development trend. This invention breaks through the limitations of this static response method.

[0060] 3. The present invention drives various types of physical regulation devices through the control execution module, including the sluice opening, pump station working conditions, and structural unloading system, realizing multi-dimensional adjustment of the water conservancy structure state. Most of the prior art takes a single execution unit as the control object, and the adjustment method lacks flexibility and pertinence, while this solution solves the problems of single control response and poor adaptability.

[0061] 4. The present invention constructs a closed-loop mechanism for structural state feedback and twin model update, and real-time feedbacks the regulation results to the digital twin system for the next round of prediction and regulation decision update, significantly improving the adaptive ability of the model. Compared with the traditional digital modeling system with low update frequency and relying on manual correction, the present invention effectively solves technical defects such as model aging and error accumulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the system construction of the present invention;

[0063] Figure 2 It is a framework diagram of the data acquisition module of the present invention;

[0064] Figure 3 It is a framework diagram of the twin modeling module of the present invention;

[0065] Figure 4 It is a framework diagram of the principal stress extraction module of the present invention;

[0066] Figure 5 It is a framework diagram of the stress evolution prediction module of the present invention;

[0067] Figure 6Frame diagram of the security domain judgment module of the present invention;

[0068] Figure 7 Frame diagram of the regulation and decision-making module of the present invention;

[0069] Figure 8 Frame diagram of the control execution module of the present invention;

[0070] Figure 9 Frame diagram of the status feedback module of the present invention;

[0071] Figure 10 Schematic diagram of the method flow of the present invention. Detailed implementation manners

[0072] Next, in combination with the accompanying drawings of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0073] Please refer to the attached Figure 1 - attached Figure 9 , the embodiments of the present invention provide a digital twin water conservancy project operation and maintenance monitoring system and method, including:

[0074] A data acquisition module, configured to acquire real-time sensing data of the water conservancy project structure area, where the real-time sensing data includes structural strain, water pressure, temperature, and displacement information;

[0075] The data acquisition module is one of the front-end basic modules of the entire system. Its main function is to continuously, stably, and frequently acquire key state parameters of the water conservancy project structure area, and provide accurate and real-time original data support for subsequent twin modeling, stress field construction, and evolution prediction modules.

[0076] Generally, the water conservancy project structure is usually in a complex environmental condition, such as high water pressure impact, day-night temperature difference change, and long-term load action. Therefore, higher requirements are put forward for the accuracy and response speed of real-time data acquisition. As an option, the data acquisition module needs to have the ability to parallelly acquire multi-source heterogeneous sensing data, and at the same time should be able to adjust the sampling frequency and acquisition area according to the dynamic characteristics of the structural force change.

[0077] Specifically, in this embodiment, the data acquisition module is used to collect real-time sensing data of the water conservancy project structure area, and the real-time sensing data includes but is not limited to structural strain, water pressure, temperature, and displacement information. The data acquisition coverage includes the structural surface layer, internal load-bearing components, and the structural surrounding environment area, and appropriate sensor types and deployment strategies are adopted according to different sensing targets.

[0078] In some embodiments, the acquisition of structural strain data adopts the method of arranging a strain gauge array. The strain gauges can include resistance strain gauges, fiber Bragg grating sensors, etc., and are deployed at the concrete gravity dam body, steel structure connection nodes, and important support parts to capture local strain changes under the action of water load, temperature gradient, or operating vibration.

[0079] In this embodiment, the strain can be represented by the strain tensor ε ij , where i, j = 1, 2, 3, representing the direction components in three-dimensional space. The composition of the strain tensor includes normal strains ε 11 , ε 22 , ε 33 , and shear strains ε 12 , ε 13 , ε 23 . In the calculation, the following expression can be used for conversion:

[0080]

[0081] where ε ij is the strain tensor component; u i represents the displacement component of the structure in the i-th direction; x j is the coordinate in the j-th direction; u j represents the displacement component of the structure in the j-th direction; x i represents the coordinate in the i-th direction.

[0082] For the acquisition of water pressure, multi-point distributed pressure sensors are usually adopted, which can be arranged on the water-facing surface of the hydraulic structure, the flood discharge channel, and the water seal area inside the structure. Generally, the measured water pressure data p can be expressed as the relationship with the water depth h and the liquid density ρ:

[0083] p = ρgh;

[0084] where p is the water pressure data; ρ represents the density of water; g represents the acceleration due to gravity; h is the vertical distance from the measurement point to the free liquid surface. Such pressure data helps to construct the load distribution model under actual working conditions and serves as an important boundary condition for subsequent stress tensor calculations.

[0085] The acquisition of temperature data is mainly achieved through thermocouples, resistance temperature sensors (RTDs), and distributed fiber optic temperature sensors. In a possible implementation, the temperature sensors are buried inside the concrete structure, near key welding parts, and on the project surface to form a spatially continuous temperature field acquisition system. The temperature field data is not only used to correct the strain error caused by material thermal expansion but also serves as a reference for evaluating the aging and fatigue of structural materials.

[0086] The acquisition of displacement information relies on high-precision displacement sensors, such as LVDTs, laser displacement meters, or GNSS structural monitoring units. In actual layout, the displacement sensors can be installed at key nodes such as the dam crest, the top of the gate pier, and the slope foundation to monitor the changes in their absolute and relative displacements. The measured displacement data u i can also be used to inversely calculate the strain tensor distribution of the structure and compare it with the values collected by strain gauges for verification, thereby improving the overall measurement accuracy.

[0087] As an implementation, the data acquisition module also includes an edge computing unit, which is used to perform preliminary filtering, noise reduction, and time synchronization processing on the original sensing data to ensure high availability and timeliness of the data before it enters the twin modeling module. The edge computing unit is built with a clock synchronization algorithm and combines with the GPS time synchronization mechanism to construct a unified time reference for cross-sensor data.

[0088] In some embodiments, the acquisition module has the ability of dynamic adaptive frequency modulation. That is, when the structural vibration frequency or the strain growth rate exceeds the set threshold, the system automatically increases the sampling frequency to achieve high-density recording of key working conditions. This strategy is beneficial to improving the system's response ability to emergencies and capturing its weak signals in the early stage of structural damage.

[0089] In this embodiment, all sensor nodes are connected to the central control node through wired (such as RS485, fiber optic Ethernet) or wireless (such as LoRa, NB-IoT) methods. The control node is responsible for unified data scheduling and uploading, building a multi-level redundant data path, and improving the stability and fault tolerance of the acquisition system.

[0090] Looking further, the acquisition module can further introduce machine learning methods to extract features and perform pattern recognition on the long-term collected original data, so as to identify abnormal strain patterns, sudden changes in water pressure, or displacement drift trends, provide auxiliary data labels for the subsequent stress prediction module, and strengthen the system's full-life cycle monitoring ability.

[0091] The twin modeling module receives the real-time sensing data transmitted by the data acquisition module and constructs a stress tensor model of the key structural area based on the pre-built three-dimensional finite element model, and then generates the three-dimensional stress field information inside the structure;

[0092] After completing the acquisition of real-time sensing data such as strain, water pressure, temperature, and displacement of the hydraulic engineering structure, the system needs to conduct in-depth modeling and analysis of these raw data to reveal the true stress state inside the structure. For this purpose, the twin modeling module serves as the core computing unit that links the preceding and the following in the system. Its function is to fuse the measured data with the simulation model and generate a three-dimensional stress field of the key parts of the structure through tensor construction means. The module and the data acquisition module are connected through a stable data interface for data flow docking to ensure timeliness and spatial consistency, providing a necessary stress tensor basis for subsequent principal stress extraction, evolution prediction, and safety judgment.

[0093] In this embodiment, the twin modeling module receives real-time sensing data from the data acquisition module and conducts stress tensor modeling on the key areas of the hydraulic engineering structure based on the three-dimensional finite element model pre-constructed by the system. This three-dimensional finite element model is usually constructed using structural drawings, material parameters, and typical working conditions during the engineering design stage and is calibrated and adjusted on-site to improve its correspondence accuracy with the actual engineering state.

[0094] Generally, the three-dimensional finite element model is divided into multiple unit bodies, and each unit body node corresponds to an actual measurement point in the physical structure. The strain and displacement data collected by the sensors are applied as input boundary conditions to the model nodes, and the stress responses at each node are solved through the finite element calculation process.

[0095] Specifically, the construction of the stress tensor σ ij is based on the material constitutive relationship and the equilibrium condition. Under the condition of isotropic linear elastic materials, the generalized Hooke's law is used to describe the stress-strain relationship:

[0096] σ ij =λ·δ ij ·ε kk +2μ·ε ij ;

[0097] where, σ ij represents the stress tensor component; ε ij is the strain tensor component; δ ij is the Kronecker function; λ and μ are the Lame constants respectively; ε kk represents the trace of the strain tensor, that is, the volumetric strain.

[0098] The Lame constants can be converted through the Young's modulus E and Poisson's ratio ν of the material:

[0099]

[0100] where, λ and μ are the Lame constants respectively; E is the elastic modulus; ν is the Poisson's ratio, indicating the degree of deformation in the perpendicular direction when the material is pulled or compressed in one direction.

[0101] In a possible implementation, the finite element solver uses the step-by-step iteration method or the matrix decomposition method (such as the conjugate gradient method, LU decomposition method) to calculate the entire structural model and obtain the stress tensor distribution at each node. During the solution process, temperature is introduced as a thermal load term into the material volume expansion model, and the overall stress result is adjusted by coupling the thermal strain term.

[0102] In some embodiments, the twin modeling module is not limited to static analysis but also supports structural dynamic response analysis. By applying a time-varying load function and combining with the transient analysis method, the stress field evolution process of the structure under dynamic disturbances such as earthquakes and water hammers can be obtained. During the dynamic stress analysis process, the system introduces the damping matrix C, the mass matrix M, and the stiffness matrix K to establish the motion differential equation:

[0103]

[0104] where u(t), respectively represent the changes of displacement, velocity, and acceleration with time; F(t) is the time-dependent load applied to the structure; E is the elastic modulus; M is the mass matrix.

[0105] As an option, to improve the real-time performance and data fusion accuracy of the model, the twin modeling module uses numerical interpolation methods to expand the discrete sensing point data, such as three-dimensional spline interpolation, Kriging interpolation, or radial basis function methods, to map the measured data to the entire finite element grid and improve the modeling density and resolution.

[0106] In a specific implementation, the system is also configured with a model correction module to correct the parameters of the original finite element model based on historical monitoring data and on-site maintenance results. Through the least squares fitting or Bayesian update algorithm, the material parameters, boundary conditions, and load distribution are iteratively optimized and gradually converge to a "dynamic twin model" that highly matches the measured state.

[0107] In addition, in some embodiments, the twin modeling module can integrate GPU parallel acceleration technology to use the graphics processing unit to execute matrix operations and field distribution calculation tasks, significantly shortening the modeling time and meeting the high-frequency data processing requirements.

[0108] The principal stress extraction module receives the stress tensor model, extracts the principal stress trajectories at key structural positions, and generates the principal stress spectrum data for structural safety assessment;

[0109] After the construction of the structural stress tensor is completed based on the aforementioned twin modeling module, the system needs to further decouple it mathematically and extract engineering information to identify the true stress path inside the structure. The principal stress extraction module plays a key role in this process. It is mainly used to receive the stress tensor data generated by the three-dimensional finite element model, extract the principal stress distribution at each key position inside the structure through the principal value decomposition method, and generate the principal stress spectrum data for structural safety analysis and subsequent trend judgment. This module interfaces with the twin modeling module in the format of tensor data and serves as one of the basic input sources for the prediction and evaluation module, with an indispensable intermediate calculation function.

[0110] In this embodiment, the principal stress extraction module receives the stress tensor model of the key areas of the structure output by the twin modeling module, and based on the tensor eigenvalue analysis method, extracts the principal stress components and corresponding principal directions of the target area. The principal stress refers to the diagonalization result of the stress tensor under the condition of no shear, reflecting the maximum and minimum normal stress values borne by the structural unit in the local coordinate system.

[0111] Generally, the principal stress extraction module performs principal stress analysis according to the following steps: First, at each finite element node or the specified key analysis position, extract the three-dimensional stress tensor σ, which is in the form of a symmetric second-order tensor. By solving the eigenvalue problem corresponding to the stress tensor:

[0112] det(σ - λI) = 0;

[0113] where σ is the stress tensor, representing the stress state of each point inside the object; λ is the eigenvalue, which is a scalar of the matrix; I is the identity matrix, with all elements on the diagonal being 1; det is the determinant operator.

[0114] Three real eigenvalues σ1, σ2, σ3 are obtained, representing the maximum principal stress, the intermediate principal stress, and the minimum principal stress respectively. Here, λ is the eigenvalue of the stress tensor, and I is the identity tensor.

[0115] As an option, for each principal stress value, further calculate its direction vector (i.e., eigenvector) to construct the spatial distribution of the stress transfer path in the structure. This direction vector is usually used to subsequently construct the principal stress trajectory map and correlate it with the actual crack evolution or failure mode.

[0116] Specifically, when the principal stress extraction module constructs the principal stress trajectory, it uses linear interpolation or spline interpolation algorithms to extend the discrete principal stress direction field to a continuous path in space. During the visualization analysis process, the streamline construction method can be selected to combine the principal stress direction with the structure shape to display the stress concentration area and stress trend distribution.

[0117] In some embodiments, to enhance the stability and usability of the principal stress spectrum, the principal stress extraction module introduces a multi-time window sliding calculation mechanism to perform average and standard deviation analysis on continuous time-series data and construct statistical spectrum information of the principal stress. The principal stress spectrum uses time as the horizontal axis and the principal stress amplitude as the vertical axis to form a three-dimensional dataset of the trend of the principal stress evolving over time.

[0118] As a possible implementation, the principal stress extraction module organizes the principal stress values at each position into a principal stress spectrum data structure in the following form:

[0119]

[0120] Wherein, represents the stress state at a specific spatial position and time point; (x, y, z) represents the structural spatial coordinate position; t represents the time point; and σ1, σ2, σ3 are the principal stress values at the corresponding moment. Such spectral data can be used not only for single-point analysis but also for operations such as regional aggregation analysis, trend prediction, and dangerous area identification.

[0121] In actual engineering, the principal stress direction in some structural regions rotates over time, and there is an obvious stress redistribution phenomenon. To identify such behavior, the principal stress extraction module supports the introduction of the principal direction rotation rate index θ′(t), which is calculated by comparing the direction changes of the principal stress eigenvector at adjacent time points and can be used to identify disturbed or damaged regions.

[0122] In a specific implementation, the system can further construct principal stress isosurfaces based on the principal stress tensor characteristics. The isosurfaces are nested and drawn in the three-dimensional structural model through a voxelization method to represent the structural volume regions where the principal stress exceeds the threshold, assisting in judging potential failure positions.

[0123] In addition, to improve the data stream processing efficiency, the principal stress extraction module supports a parallel computing strategy to concurrently perform eigenvalue decomposition and trajectory generation on tensor data of multiple nodes under a high-performance computing architecture, thereby meeting the real-time requirements under a large-scale structural model.

[0124] The stress evolution prediction module, based on the principal stress spectrum data, predicts the change trend of the principal stress at each key structural position within a future time interval and sends the principal stress prediction result to the safety domain judgment module;

[0125] After the construction of the principal stress spectrum is completed, to achieve a forward-looking control of the state of hydraulic engineering structures, it is necessary to further predict the evolution trend of the principal stress over time. The stress evolution prediction module undertakes this task. Its core lies in identifying the time-series characteristics of stress changes in key structural areas based on the existing principal stress spectrum data, and then estimating the trend of the dynamic evolution state of the principal stress in a future period of time, and transmitting the prediction results to the safety domain judgment module in real time to provide a necessary basis for subsequent risk identification and control decisions. This module relies on the high-resolution and continuous principal stress spectrum information generated by the principal stress extraction module, constituting the prediction link of structural health assessment.

[0126] In this embodiment, the stress evolution prediction module receives the principal stress spectrum data output by the principal stress extraction module, extracts the principal stress change trajectories in the time dimension for each key monitoring position in the structure, and predicts the evolution trend of the principal stress values in a future given time interval based on the time-series modeling method.

[0127] Generally, the principal stress is affected by multiple factors during the structural operation period, including water pressure fluctuations, temperature changes, operating load disturbances, and material creep, etc., with obvious non-linear and periodic characteristics. Therefore, this module adopts a combined modeling strategy, integrating statistical regression and machine learning methods for multi-scale prediction.

[0128] Specifically, in one implementation, the stress evolution prediction module uses an autoregressive moving average model (ARMA) or its extended form (such as the ARIMA model) to fit and model the principal stress time series. Assuming the time series is {σ(t1), σ(t2),..., σ(t n )}, then its prediction model can be expressed as:

[0129]

[0130] where ∈(t-j) is the white noise term at the jth past time step; σ(t) represents the output variable or prediction variable at time t; α i , β j are the regression coefficients respectively; ∈(t) is the white noise disturbance term; p and q are the orders of the autoregressive term and the moving average term respectively; i is the time step index of the autoregressive term; j is the time step index of the moving average term.

[0131] As an option, for the principal stress trajectory of a structure with obvious non-linear trends or mutations, a long short-term memory network (LSTM) can be used to model it. In this implementation, the principal stress spectrum enters the recurrent neural network as the input sequence, and through the gating mechanism inside the network unit, the coupling characteristics of long-term dependencies and short-term changes are captured, so as to generate the principal stress prediction sequence in the next few time steps.

[0132] In some embodiments, to further improve the prediction accuracy, this module introduces a multi-variable collaborative modeling method, taking multi-dimensional sensing data such as principal stress, temperature, and water pressure as input variables simultaneously to construct a multi-input single-output (MISO) prediction model. For example, the following non-linear regression model can be constructed:

[0133] σ(t + τ) = f(σ(t), T(t), p(t), d(t));

[0134] where σ(t + τ) represents the stress tensor at the prediction moment of τ time after the current time t; τ is the prediction time lag; T(t), p(t), and d(t) represent the temperature, water pressure, and displacement data at the current moment respectively, and the function f can be implemented by a neural network or support vector regression (SVR). This method can effectively consider the interaction effects between multiple physical fields and improve the physical consistency of stress prediction.

[0135] In a possible implementation, the stress evolution prediction module supports flexible setting of the prediction interval. For example, it supports prediction step sizes with different granularities such as the next 1 hour, 6 hours, or 24 hours. The prediction results are represented in the form of a multi-channel time series vector and are consistent with the original data structure of the principal stress spectrum, facilitating direct invocation by downstream modules.

[0136] As an implementation mechanism, all prediction results are transmitted to the safety domain judgment module in real time through a formatted interface. This interface follows a unified data structure standard and includes fields such as the structure position number, prediction moment, predicted value, and confidence interval, which are used to support the subsequent stress limit comparison and out-of-bounds identification in the safety domain judgment module.

[0137] In addition, to avoid interference from historical outliers in the prediction process, this module integrates an outlier detection mechanism, using methods such as Z-score analysis, outlier detection, or density evaluation to identify and eliminate non-physical mutation points, thereby improving the robustness of the model.

[0138] In a specific implementation, the prediction module also supports an online learning function. By feeding the prediction error back into the model weight optimization process, adaptive parameter adjustment is achieved. This process uses a sliding window method to update the training data set in real time, enabling the model to gradually approach the true operating characteristics of the structure over time.

[0139] The safety domain judgment module receives the principal stress prediction results, and based on the preset structural safety stress threshold, compares and analyzes whether the future principal stress exceeds the safety limit. If it is determined that there is a risk of principal stress exceeding the limit, a control trigger instruction is sent to the control decision-making module;

[0140] After the stress evolution prediction module completes the trend deduction of the future principal stress state, the system needs to perform a safety judgment on the predicted data to achieve timely response and pre-control of potential structural instability or stress overrun risks. The safety domain judgment module serves as the judgment core in the structural state assessment process. Based on the set safety stress domain limit values, it conducts point-by-point comparison and analysis of the predicted principal stresses to determine whether there is a risk of principal stress overrun at key structural positions in the future time period. Through data linkage with the regulation and decision-making module, this module realizes an intelligent control mechanism from state monitoring to active intervention and is a key logical node in building a full-cycle digital twin safety closed-loop.

[0141] In this embodiment, the safety domain judgment module receives the future principal stress prediction values of each structural position output by the stress evolution prediction module, including the maximum principal stress σ max (t + τ), the minimum principal stress σ min (t + τ) and other information, and performs a principal stress overrun judgment operation based on the preset upper and lower threshold values of the principal stress safety in the structural design benchmark. Execute the principal stress overrun judgment operation.

[0142] Generally, the principal stress overrun determination follows the following logic: If there exists any time t + τ that satisfies the following conditions:

[0143] Or

[0144] Where, σ max (t + τ) represents the maximum principal stress at the prediction time of τ time after the current time t; σ min (t + τ) represents the minimum principal stress at the prediction time of τ time after the current time t; Is the maximum stress threshold; Is the minimum stress threshold.

[0145] Specifically, the judgment module integrates a multi-channel safety assessment logic unit, which can perform overrun analysis on multiple monitoring points simultaneously. This module supports logical aggregation judgment based on area division, that is, by setting area-level safety indicators, it realizes the overall risk situation identification of key structural areas (such as crown, floor, side piers, etc.).

[0146] As an option, the safety domain judgment module supports a dynamic threshold adjustment mechanism. This mechanism automatically adjusts the upper and lower threshold values according to the current operating state, aging degree or working condition level of the structure to meet the tolerance change requirements caused by the degradation of material properties during long-term operation.

[0147] In some embodiments, the judgment module introduces a risk level classification strategy to quantify the degree of principal stress overrun, and according to the difference interval between the predicted overrun value and the threshold value, it is divided into mild, moderate and severe risk levels, which are defined as follows:

[0148]

[0149] wherein, Δσ represents the overrun amplitude index; σ max (t + τ) represents the maximum principal stress at the prediction time of τ after the current time t; is the maximum stress threshold. If Δσ falls into different risk level intervals, different levels of regulation trigger instructions are generated accordingly, and the structural state of high-risk areas is preferentially regulated.

[0150] In a possible implementation manner, the judgment module supports sending a standardized regulation instruction to the regulation decision-making module, and the instruction data structure includes the following fields: structural unit ID, prediction time, overrun principal stress value, corresponding threshold, overrun amplitude, recommended response level, etc. This structure ensures that the regulation module receives complete context information, facilitating subsequent operations such as load unloading, structural adjustment, or alarm broadcasting.

[0151] In addition, the security domain judgment module supports comparison and analysis with historical security domain data to determine whether the structure shows a continuous overrun trend, further strengthening the system's ability to identify abnormal persistence. This function can be achieved by introducing a sliding window cumulative risk index as follows:

[0152]

[0153] wherein, t is the current time; is the cumulative overrun index; i is the time step index of the autoregressive term; γ i is the weight coefficient; 1 {σi>σth} is the indicator function for determining whether there is an overrun; w is the sliding window length. Whether the structure is in a long-term progressive risk state is judged through such cumulative indicators.

[0154] After receiving the regulation trigger instruction, the regulation decision-making module generates a control strategy for optimizing the structural stress state according to the principal stress change trend and structural operation parameters;

[0155] As a key feedback control link, the regulation decision-making module is responsible for taking corresponding structural regulation measures according to the real-time safety assessment results. Once the security domain judgment module issues a regulation trigger instruction, indicating that there is a risk of overrun of the principal stress, the regulation decision-making module will generate a corresponding control strategy based on this instruction, the principal stress change trend, and other structural operation parameters to optimize the structural stress state and prevent further structural instability or damage. The regulation decision-making module is closely connected with the previous stress evolution prediction module and security domain judgment module, and its role is to reduce or eliminate structural risks through effective control means and ensure the safety and stability of the structure.

[0156] In this embodiment, after receiving the regulation trigger instruction from the security domain judgment module, the regulation decision-making module first parses the instruction content, and combines the principal stress prediction trend of the current structure, operating parameters (such as temperature, humidity, load, etc.), and historical monitoring data to analyze the current structure state and the potential risks of future stress changes. Based on these input information, the regulation decision-making module will formulate specific control strategies.

[0157] Generally, the goal of the regulation decision-making module is to reduce the stress concentration area of the structure by optimizing the load distribution of the structure, applying external control forces, etc., so as to keep the structure within the safe domain in the future period. The regulation method can be flexibly selected according to the specific structure type and risk type. For the area where the principal stress prediction exceeds the limit, the regulation decision-making module first identifies the principal stress concentration positions in these areas, and adopts appropriate control means, such as unloading, adding supports or adjusting external loads, etc., to redistribute the internal stress of the structure.

[0158] As an option, the regulation decision-making module adopts the model predictive control (MPC) method. MPC can calculate the optimal control strategy in real time based on the prediction information of the system state. In this mode, the regulation decision-making module will use the following optimization problem to solve the control strategy according to the principal stress change trend in the future period:

[0159]

[0160] where, is the minimization operator; y(k) is the system state within the prediction time step; y*(k) is the target state; u(k) is the displacement change; λ is the eigenvalue, which is a scalar of the matrix; N is the length of the prediction horizon. By solving this optimization problem, the regulation decision-making module can generate a series of control inputs u(k) to optimize the stress distribution in the future period and ensure that the structure does not exceed the safe stress threshold.

[0161] Specifically, the regulation decision-making module sets corresponding regulation goals based on the principal stress prediction value and the structure operation parameters. For the structural area with a higher risk of exceeding the limit, the regulation decision-making module will give priority to taking adjustment measures, such as adjusting the water pressure through the pumping system, adjusting the load distribution, or changing the position of the structure stress point, etc. to intervene. By implementing these adjustments, the phenomenon of excessive stress concentration can be effectively alleviated, the service life of the structure can be extended, and the probability of structural instability can be reduced.

[0162] In one possible implementation, the control decision module supports a combination of multiple strategies. Through the collaborative work of multiple control methods, the module can automatically select the most suitable control method under different working conditions. For example, under conditions with uniform load distribution, fine-tuning through external support is given priority; under high-risk conditions, joint adjustment can be carried out through diversified means such as unloading, load reduction, and temperature control.

[0163] In some embodiments, the control decision module also integrates a prediction error correction mechanism. When there is a large deviation between the predicted principal stress value and the actual measured value, the module will automatically start the error correction process and adjust the model parameters or control strategy to maintain the high efficiency and stability of the system.

[0164] The control execution module controls the physical execution devices in the water conservancy project to perform regulation operations according to the received control strategy. The execution devices include the sluice opening adjustment device, the pump station control device and the structure unloading device.

[0165] As the core execution unit, the control execution module is responsible for driving various physical execution devices in the water conservancy project and implementing control operations according to the control strategy output by the control decision module. By coordinating with execution devices such as sluice gates, pump stations and structural unloading, the control execution module can accurately adjust the stress state of the structure to ensure that the project structure always remains within a safe range during operation. The efficient collaboration of this module with the control decision module and other monitoring modules constitutes a complete real-time control closed loop, which greatly enhances the safety and stability of water conservancy projects.

[0166] In this embodiment, the control execution module accurately controls the sluice gate opening adjustment device, the pump station control device and the structure unloading device according to the instructions of the control decision module. For each execution device, the control execution module generates a control signal through a specific control strategy and transmits it to the corresponding physical device to ensure the stability of the structure and the water conservancy system.

[0167] Generally, the sluice gate opening adjustment device adjusts the water level by controlling the opening and closing state of the gate, thereby affecting the water pressure on the structure. After receiving the instructions from the control decision module, the control execution module timely adjusts the sluice gate opening according to the predicted trend of the principal stress change to control the water pressure within the safety threshold and prevent the structure from being overloaded or unstable due to excessive water level. Specifically, the control execution module calculates the expected sluice gate opening value according to the control strategy, and adjusts the gate opening in real time through the actuator to achieve precise control of the water pressure.

[0168] As an option, the control execution module can further optimize the opening adjustment process by establishing a relationship model between the sluice opening and the structural principal stress. For example, the relationship between the sluice opening θ(t) and the structural principal stress σ(t) can be expressed by the following model:

[0169] σ(t) = f(θ(t), p(t), T(t));

[0170] Where, θ(t) is the control parameter; σ(t) represents the output variable or prediction variable at time t; p(t) is the pressure influence variable; T(t) represents the temperature influence; f is a non - linear relationship function. Through this model, the control execution module can consider the influence of other external factors on the structural stress while adjusting the sluice opening, so as to achieve more precise control.

[0171] In some embodiments, the control execution module can also be linked with the pump station control device to adjust the working state of the water pump, so as to effectively regulate the water flow and water pressure. When the system determines that the principal stress in a certain key structural area approaches the threshold value, the control execution module will send a signal to adjust the pump station flow and reduce the impact of water pressure on the structure. The working mechanism of the pump station control device is usually achieved by adjusting the pump speed, start - stop control and the opening of the flow regulating valve. In this case, the control execution module will convert the flow regulation instruction of the regulation decision - making module into specific pump station control instructions to ensure that the working state of the water pump can be synchronized with the structural stress requirements.

[0172] Specifically, the control execution module converts the regulation requirements of the regulation decision - making module into the following control instructions:

[0173] Q target = f(σ(t), p target , T(t));

[0174] Where, Q target represents the target flow rate after adjustment of the pump station; p target is the preset target water pressure; T(t) is the ambient temperature at the time point; σ(t) represents the output variable or prediction variable at time t; f is a non - linear relationship function.

[0175] As a possible implementation, when a serious principal stress exceeds the limit, the control execution module can also activate the structural unloading device. This device usually includes an adjustable support device, which changes the stress state of the structure by increasing or decreasing the support force. The control execution module precisely adjusts the support force of the structural unloading device according to the instructions of the regulation decision - making module, so as to effectively reduce stress concentration and relieve the structural load.

[0176] In some embodiments, the control execution module automatically adjusts the unloading support strength by real - time monitoring the change of structural stress and combining with the feedback mechanism of the unloading device. This process can be expressed by the following formula:

[0177] F unload (t) = g(σ(t), p(t), T(t));

[0178] Among them, F unload (t) represents the supporting force of the structure unloading device, g is the unloading adjustment function calculated according to factors such as structural stress, external water pressure, and temperature; σ(t) represents the output variable or predicted variable at time t; p(t) is the pressure influence variable; T(t) represents the temperature influence.

[0179] In the above way, the control execution module can achieve precise regulation operations, not only meeting the real-time changing structural safety requirements, but also effectively coping with dynamic working condition changes and enhancing the long-term stability of water conservancy projects.

[0180] The state feedback module collects the structural state data after regulation and feeds the data back to the twin modeling module to update the digital twin system model and realize the closed-loop control of principal stress prediction and structural regulation.

[0181] The state feedback module undertakes the key information feedback function. Its role is to monitor the structural operation state in real time after the implementation of the regulation strategy and feed the collected structural response data back to the twin modeling module in a timely manner, so as to realize the dynamic update of the digital twin system. A tight data path is formed between this module and the control execution module to ensure that the regulation effect can be quickly known after the control operation is executed, and then the closed-loop of principal stress prediction and regulation based on actual feedback is realized, improving the overall timeliness and accuracy of the system.

[0182] In this embodiment, after the control execution module completes the regulation operation, the state feedback module immediately starts the state acquisition process and collects multi-dimensional data including principal stress values, structural displacements, vibration responses, water pressure changes, and temperature and humidity environments.

[0183] Generally, the state feedback module realizes the comprehensive perception of the structural state by means of multi-type sensor fusion. Specifically, it includes fiber Bragg grating sensors (FBG), distributed fiber sensors, resistance strain gauges, MEMS accelerometers, laser displacement gauges, etc. The data collected by various sensors are unified in time stamps and spatially registered through a spatio-temporal synchronization mechanism to ensure the structural consistency and timeliness and accuracy of the feedback data.

[0184] As an option, after data collection, the state feedback module first performs denoising processing on various sensor signals and uses a multi-scale fusion algorithm to perform normalization conversion on different types of data. In this processing process, the following fusion model is used to reconstruct the collected data:

[0185]

[0186] Among them, X fused represents the fused state feedback data, X s is the strain-related data, X ais the acceleration response data, X d is the displacement observation data, is the fusion function, which can usually be implemented based on principal component analysis (PCA), autoencoder network or graph neural network.

[0187] Specifically, the state feedback module uploads the processed structural state data to the twin modeling module and triggers the dynamic update mechanism of the twin system. The twin modeling module corrects the previous model parameters according to the feedback data, including key variables such as material parameters, boundary conditions, load models and damping coefficients. In this way, the model prediction error can be significantly reduced and the accuracy of subsequent principal stress prediction can be improved.

[0188] In a possible implementation, the state feedback module also integrates an error inversion mechanism. The system calculates the residual between the predicted state and the feedback state and iteratively corrects the model parameters using the least squares or Bayesian update method. This process is implemented through the following model:

[0189] θ k+1 = θ k + K k ·(y obs - y pred );

[0190] where θ k is the parameter vector of the k-th iteration, K k is the update gain matrix, y obs is the actual observation value, y pred is the model prediction value. Such a feedback update method can effectively improve the time-dependent adaptive ability of the digital twin model.

[0191] In some embodiments, the state feedback module also uses the feedback data to update the training samples of the self-learning model to enhance the generalization ability of the principal stress prediction model under different working conditions. After the data is updated, the model can fully consider the latest structural state in the next round of prediction to form a complete closed-loop control path.

[0192] Specifically, the update data provided by the state feedback module is not limited to the instantaneous state after the current regulation, but also includes the dynamic evolution sequence during the regulation process. This sequence data is constructed through a sliding window mechanism and transmitted as a time series input to the twin modeling module to support the learning of the time series characteristics of the model.

[0193] A digital twin water conservancy project operation and maintenance monitoring method described below can be correspondingly referred to with a digital twin water conservancy project operation and maintenance monitoring system described above.

[0194] Please refer to the appendix Figure 10 , the present invention also provides a sensorless DC brushless motor control method, including the following steps:

[0195] S1. Data acquisition: Obtain real-time sensing data of the water conservancy project structure area through the data acquisition module. The sensing data includes strain, water pressure, temperature, and displacement information, which is used to characterize the current state of the structure.

[0196] S2. Twin modeling: Receive the real-time sensing data transmitted by the data acquisition module, and build a stress tensor model of the water conservancy project structure area in the twin modeling module based on the pre-constructed three-dimensional finite element model, generating corresponding three-dimensional stress field information.

[0197] S3. Principal stress extraction: Extract principal stress trajectory data from the stress tensor model generated by the twin modeling module. The principal stress trajectory reflects the stress changes at key positions of the structure, forming a principal stress spectrum for subsequent analysis.

[0198] S4. Stress evolution prediction: Use the stress evolution prediction module to predict the trend of the principal stress trajectory. Based on historical data and the current principal stress state, predict the stress change trend of the structure in the future for a period of time.

[0199] S5. Safety assessment and judgment: Based on the predicted principal stress trajectory, compare and analyze with the structural safety stress threshold through the safety domain judgment module to determine whether there is a risk of principal stress exceeding the limit.

[0200] S6. Control strategy construction: If it is determined that the principal stress exceeds the limit, the regulation decision module generates an optimal control strategy according to the over-limit conditions and the operating parameters of the structure. The control strategy aims to adjust the stress state of the structure.

[0201] S7. Execution of regulation and feedback: Execute the control strategy through the control execution module to adjust the physical devices in the water conservancy project, and collect the feedback status data after regulation through the status feedback module.

[0202] S8. Model update and loop: Transmit the feedback status data back to the twin modeling module to update the digital twin model, realizing continuous monitoring and optimal adjustment of the structure.

[0203] The method of this embodiment can be used to implement the above system embodiment, and its principle and technical effects are similar, so they will not be elaborated here.

[0204] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A digital twin operation and maintenance monitoring system for water conservancy projects, characterized in that, It includes: A data acquisition module for acquiring real-time sensing data of the water conservancy project structure area, where the real-time sensing data includes structural strain, water pressure, temperature, and displacement information; A digital twin modeling module that receives the real-time sensing data transmitted by the data acquisition module and constructs a stress tensor model of the key structural area based on a pre-built three-dimensional finite element model, and then generates three-dimensional stress field information inside the structure; A principal stress extraction module that receives the stress tensor model, extracts the principal stress trajectories at key structural positions, and generates principal stress spectrum data for structural safety assessment; A stress evolution prediction module that, based on the principal stress spectrum data, predicts the change trend of the principal stress at each key structural position within a future time interval and sends the principal stress prediction result to the safety domain judgment module; A safety domain judgment module that receives the principal stress prediction result, compares and analyzes whether the future principal stress exceeds the safety limit according to the preset structural safety stress threshold, and if it is judged that there is a risk of principal stress exceeding the limit, sends a control trigger instruction to the regulation decision-making module; A regulation decision-making module that, after receiving the control trigger instruction, generates a control strategy for optimizing the structural stress state according to the principal stress change trend and the structural operation parameters; A control execution module that, according to the received control strategy, controls the physical execution devices in the water conservancy project to perform regulation operations, and the execution devices include sluice opening adjustment devices, pumping station control devices, and structural unloading devices; A state feedback module that acquires the regulated structural state data and feeds the data back to the digital twin modeling module to update the digital twin system model and achieve closed-loop control of principal stress prediction and structural regulation.

2. The digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The data acquisition module includes: A strain acquisition unit for obtaining strain change information on the surface or inside of the structure; A water pressure monitoring unit for acquiring the acting force of the water body around the structure on the structure surface; A temperature acquisition unit for acquiring the temperature data of the environment and inside the structural material; A displacement measurement unit for acquiring the absolute or relative displacement changes at key parts of the structure.

3. A digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The digital twin modeling module includes: A finite element modeling unit for constructing a static or dynamic mechanical model of the structure; A tensor calculation unit for dynamically calculating the stress tensor distribution according to the sensing data; A model correction unit for comparing the real-time monitoring data with the simulation model and correcting the stress field calculation result.

4. A digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The principal stress extraction module includes: A tensor eigenvalue analysis unit for extracting the principal stress values of the stress tensor; A stress trajectory construction unit for constructing an evolution trajectory according to the principal stress values at multiple time points; A spectrum diagram generation unit for generating a principal stress spectrum diagram or principal stress distribution diagram reflecting the structural state.

5. A digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The stress evolution prediction module includes: A time series modeling unit for modeling the principal stress historical sequence; A trend prediction unit for predicting the future principal stress change based on the current state and historical evolution trend; A prediction result output unit for transmitting the prediction result to the safety domain judgment module.

6. The digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The safety domain judgment module includes: A threshold setting unit for setting the safety stress limit of the structural material; A principal stress comparison unit for comparing the predicted principal stress with the set threshold; A regulation determination unit, which is used to determine whether there is an overlimit risk according to the comparison result and generate a control signal.

7. The digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The regulation decision-making module includes: A control variable definition unit, which is used to determine the external parameters available for regulation; An objective function construction unit, which is used to construct a regulation optimization objective based on structural safety and operation cost; A strategy solution unit, which is used to solve the optimal control strategy by using an optimization algorithm.

8. A digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The control execution module includes: An instruction parsing unit, which is used to convert the control strategy into executable instructions; An execution device control unit, which is used to send control instructions to the physical devices of the water conservancy project; A regulation effect detection unit, which is used to confirm the feedback of the device state after execution.

9. A digital twin water conservancy project operation and maintenance monitoring system according to claim 1, characterized in that, The state feedback module includes: A state perception unit, which is used to perceive the stress and deformation state of the structure in real time after control; A data synchronization unit, which is used to transmit the feedback state back to the digital twin model update module; A loop update unit, which is used to start a new prediction and control cycle.

10. A method for operation and maintenance monitoring of a digital twin water conservancy project, according to the digital twin water conservancy project operation and maintenance monitoring system described in any one of claims 1-9, characterized in that, It includes the following steps: Data acquisition, obtaining real-time sensing data of the water conservancy project structure area through the data acquisition module, where the sensing data includes strain, water pressure, temperature and displacement information, which is used to characterize the current state of the structure; Digital twin modeling, receiving the real-time sensing data transmitted by the data acquisition module, and constructing a stress tensor model of the water conservancy project structure area in the digital twin modeling module based on a pre-constructed three-dimensional finite element model, and generating corresponding three-dimensional stress field information; Principal stress extraction, extracting principal stress trajectory data from the stress tensor model generated by the digital twin modeling module, where the principal stress trajectory reflects the stress change at the key positions of the structure, and forming a principal stress spectrum for subsequent analysis; Stress evolution prediction, using the stress evolution prediction module to predict the trend of the principal stress trajectory, and predicting the stress change trend of the structure in a future period of time based on historical data and the current principal stress state; Safety assessment and judgment, based on the predicted principal stress trajectory, comparing and analyzing with the structural safety stress threshold through the safety domain judgment module to determine whether there is a risk of principal stress overlimit; Control strategy construction, if it is judged that the principal stress is overlimit, the regulation decision-making module generates an optimal control strategy according to the overlimit condition and the operation parameters of the structure, and the control strategy is aimed at adjusting the stress state of the structure; Execute regulation and feedback, execute the control strategy through the control execution module to adjust the physical devices in the water conservancy project, and collect the feedback state data after regulation through the state feedback module; Model update and loop, transmit the feedback state data back to the digital twin modeling module, update the digital twin model, and realize the continuous monitoring and optimization adjustment of the structure.

Images