Wind power mixed tower damage prediction method and system based on digital twinning

Abstract

The invention provides a wind power mixed tower damage prediction method and system based on digital twinning. The method comprises the steps of obtaining multi-source data of a wind power mixed tower; wherein the multi-source data comprises structure response data, environment load data and historical operation and maintenance data; based on a building information model and finite element model fusion technology, constructing a digital twinborn model of the wind power mixing tower; inputting the multi-source data into a digital twinborn model to realize real-time mapping of the wind power mixed tower and the digital twinborn model, and obtaining real-time structure state data of the wind power mixed tower through the digital twinborn model; inputting the multi-source data and the real-time structure state data into a pre-trained LSTM-GRU hybrid neural network model, and predicting a damage evolution trend of the wind power hybrid tower in combination with an attention mechanism; and on the basis of the damage evolution trend, a maintenance scheme of the wind power mixed tower is generated by using a multi-target particle swarm algorithm, so that balanced optimization of maintenance cost, power generation loss and risk level is realized, and the operation and maintenance efficiency and safety of the wind power mixed tower are integrally improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment damage prediction, and in particular to a wind power hybrid tower damage prediction method and system based on digital twins. Background Art

[0002] With the rapid development of the wind power industry, the structural safety and reliability of wind turbine hybrid towers, as key components of wind power systems, are paramount. Exposed to complex natural environments for extended periods, wind turbine hybrid towers face numerous damage risks, such as cracks, corrosion, and deformation. These damages not only impact the service life of wind turbine hybrid towers but can also lead to serious safety incidents. Therefore, timely and accurate damage prediction and maintenance are crucial to ensuring the stable operation of wind power systems.

[0003] However, traditional methods for detecting and maintaining wind turbine hybrid tower damage have numerous limitations. For one thing, traditional detection methods rely primarily on manual inspections and scheduled maintenance, which are inefficient and difficult to detect early damage. Furthermore, existing damage prediction methods often rely on a single data source, such as structural response data, which fails to fully reflect the actual operating status of wind turbine hybrid towers, resulting in insufficient prediction accuracy. Summary of the Invention

[0004] The present invention provides a wind power hybrid tower damage prediction method and system based on digital twin, which is used to solve the technical problem that the damage degree of wind power hybrid tower cannot be accurately predicted in the prior art.

[0005] In one aspect, the present invention provides a wind turbine hybrid tower damage prediction method based on digital twins, comprising: Acquire multi-source data of the wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data and historical operation and maintenance data; Based on the fusion technology of building information model and finite element model, a digital twin model of the wind turbine hybrid tower is constructed; Inputting the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtaining real-time structural status data of the wind turbine hybrid tower through the digital twin model; Inputting the multi-source data and the real-time structural status data into a pre-trained LSTM-GRU hybrid neural network model, and combining the attention mechanism to predict the damage evolution trend of the wind turbine hybrid tower; Based on the damage evolution trend, a maintenance plan for the wind turbine hybrid tower is generated using a multi-objective particle swarm algorithm.

[0006] According to a wind power hybrid tower damage prediction method based on digital twin provided by the present invention, a digital twin model of the wind power hybrid tower is constructed based on the fusion technology of building information model and finite element model, including: Use building information modeling to build a three-dimensional geometric model of the wind turbine hybrid tower; Use finite element model to simulate the mechanical behavior of wind turbine hybrid tower; The building information model and the finite element model are integrated to obtain a digital twin model of the wind turbine hybrid tower.

[0007] According to a digital twin-based wind power hybrid tower damage prediction method provided by the present invention, the multi-source data is input into the digital twin model to achieve real-time mapping between the wind power hybrid tower and the digital twin model, including: Inputting the multi-source data into the digital twin model, and dynamically adjusting the model parameters through a parameter update mechanism in the digital twin model, so that the digital twin model reflects the operating status of the wind turbine hybrid tower in real time; Predicting multi-source prediction data at the next moment based on the input multi-source data through the digital twin model; Comparing the multi-source prediction data at the next moment with the multi-source data to obtain a comparison result; The digital twin model is corrected based on the comparison result.

[0008] According to a wind turbine hybrid tower damage prediction method based on digital twins provided by the present invention, the LSTM-GRU hybrid neural network model is trained in the following manner: Collecting sample data of wind turbine hybrid towers, and dividing the sample data into training sample data and verification sample data; The LSTM-GRU hybrid neural network model is trained using the training sample data, and the input training sample data is weighted using an attention mechanism to automatically focus on sensitive parameters; Input the verification sample data into the trained LSTM-GRU hybrid neural network model to obtain a prediction result; Determine the error index between the predicted results and the actual data; When the error indicator meets the preset rule, the training of the LSTM-GRU hybrid neural network model is stopped.

[0009] According to a digital twin-based wind turbine hybrid tower damage prediction method provided by the present invention, based on the damage evolution trend, a multi-objective particle swarm algorithm is used to generate a maintenance plan for the wind turbine hybrid tower, including: determining a damage index of the damage according to the damage evolution trend; In combination with the operating status and maintenance objectives of the wind turbine hybrid tower, multiple optimization objectives are set; wherein the optimization objectives include minimizing maintenance costs, minimizing power generation losses, and minimizing risk levels; Inputting the damage index and the optimization objective into a multi-objective particle swarm algorithm to find an optimal solution that satisfies all optimization objectives, thereby obtaining an optimal solution set; Selecting a target solution from the optimal solution set based on actual maintenance requirements and resource constraints; The target solution is converted into a maintenance plan.

[0010] According to a wind turbine hybrid tower damage prediction method based on digital twin provided by the present invention, the damage index is determined according to the damage evolution trend, including: Extracting damage location, damage type, and damage degree as damage characteristics from the damage evolution trend, wherein the damage type includes but is not limited to cracks, corrosion, and deformation, and the damage degree includes but is not limited to crack length, corrosion depth, and deformation; For each injury type, the rate of change of its injury severity over time was calculated to quantify the rate of injury development; The degree of injury is divided into three levels according to its impact on function: mild, moderate and severe; The development speed of the injury is divided into three levels according to its speed: slow, medium and fast; The remaining useful life is divided into three levels according to its length: long, medium and short; The quantitative results of damage degree, development speed and remaining service life are used as damage indicators.

[0011] A wind turbine hybrid tower damage prediction method based on digital twins provided by the present invention also includes: Based on damage evolution trends and maintenance plans, a full life cycle health profile of the wind turbine hybrid tower is constructed; wherein the full life cycle health profile includes but is not limited to initial design parameters, manufacturing process data, installation and commissioning records, real-time monitoring data during operation, damage prediction records, maintenance decision records, and post-retirement assessment data; Utilize blockchain technology to store the entire life cycle health records in a distributed manner and record them in an unalterable manner; When the damage evolution trend reaches the preset damage threshold, the maintenance task is automatically triggered through the smart contract, and a maintenance work order is generated and assigned to the corresponding maintenance team.

[0012] According to the present invention, a wind turbine hybrid tower damage prediction method based on digital twins is provided. Based on damage evolution trends and maintenance plans, a full life cycle health record of the wind turbine hybrid tower is constructed, which also includes: When new monitoring data or injury prediction results are generated, the data update process is automatically triggered and fed back to the health record in real time; Develop an interactive health record visualization platform and use augmented reality technology to display the health status of wind turbine hybrid towers on the health record visualization platform.

[0013] A wind turbine hybrid tower damage prediction method based on digital twins provided by the present invention also includes: Based on the digital twin model of the wind turbine hybrid tower and the damage prediction results, a virtual reality decision-making support system is constructed; The virtual reality decision-making system is used to visualize the real-time structural status, damage evolution trend, and maintenance plan of the wind turbine hybrid tower to the operation and maintenance personnel. The implementation process of different maintenance plans is simulated in the virtual reality decision-making auxiliary system.

[0014] On the other hand, the present invention also provides a wind power hybrid tower damage prediction system based on digital twin, comprising: A data acquisition module is used to obtain multi-source data of the wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data and historical operation and maintenance data; A model construction module, used to construct a digital twin model of the wind turbine hybrid tower based on the fusion technology of building information model and finite element model; A model mapping module is used to input the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtain real-time structural status data of the wind turbine hybrid tower through the digital twin model; A damage prediction module, configured to input the multi-source data and the real-time structural status data into a pre-trained LSTM-GRU hybrid neural network model, and predict the damage evolution trend of the wind turbine hybrid tower in combination with an attention mechanism; A maintenance generation module is used to generate a maintenance plan for the wind turbine hybrid tower based on the damage evolution trend using a multi-objective particle swarm algorithm.

[0015] The digital twin-based wind turbine hybrid tower damage prediction method and system provided by this invention achieves real-time status monitoring and damage evolution trend prediction of wind turbine hybrid towers by acquiring multi-source data and constructing a digital twin model. Damage prediction is performed using an LSTM-GRU hybrid neural network combined with an attention mechanism, improving prediction accuracy. Based on the damage evolution trend, an optimized maintenance plan is generated using a multi-objective particle swarm algorithm, achieving a balanced optimization of maintenance costs, power generation losses, and risk levels, thereby improving the overall operation and maintenance efficiency and safety of wind turbine hybrid towers. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 This is a flow chart of a wind power hybrid tower damage prediction method based on digital twins provided by an embodiment of the present invention; Figure 2 Schematic diagram of the structure of a wind power hybrid tower damage prediction system based on digital twins provided by an embodiment of the present invention; Figure 3 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0019] Figure 1 It is a flow chart of the wind power hybrid tower damage prediction method based on digital twin provided by an embodiment of the present invention. A wind power hybrid tower refers to a hybrid structure tower for supporting wind turbines. It combines the advantages of traditional steel towers and concrete towers, and is usually composed of steel tower sections and concrete tower sections. The steel tower section has good wind resistance and construction convenience, while the concrete tower section has high rigidity and compressive strength. Steel is generally used for the upper structure of the tower. Concrete is generally used for the lower structure of the tower. The steel tower section and the concrete tower section are usually connected by high-strength bolts, prestressed steel bars or welding to ensure the integrity and stability of the structure.

[0020] See also Figure 1 ,The wind turbine hybrid tower damage prediction method based on digital twin ,may include the following steps.

[0021] Step 101: Acquire multi-source data of a wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data, and historical operation and maintenance data.

[0022] In this step, structural response data refers to the structural response data generated by various loads during the operation of the wind turbine hybrid tower. This data reflects the actual operating state of the hybrid tower, including but not limited to strain, displacement, acceleration, and vibration frequency. This data is typically collected in real time using sensors installed on the hybrid tower (such as strain gauges, accelerometers, and displacement sensors). By analyzing this structural response data, we can understand the stress distribution, deformation, and dynamic characteristics of the hybrid tower under different operating conditions, thereby determining whether there is structural damage or potential safety hazards.

[0023] Environmental load data refers to the loads generated by natural environmental factors acting on hybrid wind turbine towers. This data reflects the complexity and uncertainty of the tower's environment and includes data such as wind speed, direction, temperature, humidity, rainfall, snow load, and seismic activity. This data can be obtained through equipment such as meteorological stations and environmental monitoring sensors. Environmental loads are a key factor contributing to damage to hybrid wind turbine towers. By analyzing environmental load data, we can understand the external loads acting on the hybrid tower, assess their impact on the structure, and provide a basis for damage prediction.

[0024] Historical O&M data refers to the maintenance, inspection, and troubleshooting data accumulated during the operation of a wind turbine hybrid tower. This data reflects the hybrid tower's operational history and maintenance, including equipment repair records, fault diagnosis reports, component replacement records, and regular inspection results. This data is typically stored in the wind farm's O&M management system. Historical O&M data can provide information on the damage history and maintenance experience of the hybrid tower, helping to analyze damage trends and patterns, and providing a reference for developing appropriate maintenance plans.

[0025] Step 102: Based on the fusion technology of building information model and finite element model, a digital twin model of the wind turbine hybrid tower is constructed.

[0026] Step 102 may include: Use building information modeling to build a three-dimensional geometric model of the wind turbine hybrid tower; Specifically, Building Information Modeling (BIM) is an integrated digital representation method used to create and manage information about buildings or infrastructure projects. It encompasses not only geometric information but also the properties, relationships, and behaviors of building components. In wind turbine hybrid tower applications, BIM is used to establish a three-dimensional geometric model of the tower. The goal of this step is to create an accurate virtual geometric model that reflects the tower's actual physical structure. Professional BIM software (such as Revit and ArchiCAD) can be used to create a three-dimensional geometric model of the tower based on the wind turbine hybrid tower's design drawings and parameters. This model includes geometric information such as the tower's shape, dimensions, segment connections, and foundation structure.

[0027] Use finite element model to simulate the mechanical behavior of wind turbine hybrid tower; Specifically, the finite element model (FEM) is a numerical analysis method used to simulate the mechanical behavior of complex structures under various loads. It divides the structure into a finite number of small units and solves the governing equations on these units to obtain the structural mechanical responses, such as stress, strain, and displacement. In wind turbine hybrid tower applications, FEM is used to simulate the mechanical behavior of hybrid towers under loads such as wind, gravity, and seismic forces. Finite element analysis software (such as ANSYS and ABAQUS) can be used to establish a finite element model based on the geometric model and material properties of the wind turbine hybrid tower. Boundary conditions and load conditions, such as wind speed, wind direction, gravity, and seismic motion, can be defined. Numerical calculations can then be performed to obtain the mechanical responses of the hybrid tower, such as stress distribution and deformation, under different operating conditions.

[0028] The building information model is integrated with the finite element model to obtain a digital twin model of the wind turbine hybrid tower; Specifically, model fusion involves combining the geometric and attribute information of the BIM model with the mechanical analysis results of the FEM model to create a comprehensive digital twin model. The digital twin model contains not only geometric information but also the mechanical behavior and performance information of the structure. The goal of this step is to create a virtual model that can reflect the operational status and mechanical behavior of the wind turbine hybrid tower in real time for subsequent real-time monitoring, damage prediction, and maintenance decision-making. Existing data interfaces and software tools can be developed or used to enable data exchange and fusion between the BIM model and the FEM model, mapping the mechanical analysis results of the FEM model (such as stress, strain, displacement, etc.) to the geometric structure of the BIM model.

[0029] Step 103: Input multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtain real-time structural status data of the wind turbine hybrid tower through the digital twin model.

[0030] Step 103 may include: Input multi-source data into the digital twin model, and dynamically adjust the model parameters through the parameter update mechanism in the digital twin model, so that the digital twin model can reflect the operating status of the wind turbine hybrid tower in real time; Specifically, one of the core functions of a digital twin model is to reflect the operating status of a wind turbine hybrid tower in real time. To achieve this, the model needs to dynamically adjust its internal parameters based on real-time multi-source data input. The parameter update mechanism refers to the algorithm or rules within the model that adjust the model's parameters based on new data to ensure that the model's predictions are consistent with the actual operating status. For example, multi-source data (structural response data, environmental load data, historical operation and maintenance data) is input into the digital twin model. The model's internal algorithms (such as Kalman filters and machine learning algorithms) use this data to dynamically adjust model parameters, such as material properties, boundary conditions, and load distribution.

[0031] Use the digital twin model to predict the multi-source prediction data at the next moment based on the input multi-source data; Specifically, the model uses its internal prediction algorithms (such as time series analysis and neural networks) to process the current input data and generate multi-source prediction data for the next moment. Multi-source prediction data can include structural responses (such as strain, displacement, vibration frequency, etc.) and environmental loads (such as wind speed and temperature).

[0032] Comparing the multi-source prediction data at the next moment with the multi-source data to obtain a comparison result; Specifically, the data predicted by the model is compared with the actual measured data, and the error (such as mean square error, absolute error, etc.) is calculated.

[0033] Correct the digital twin model based on the comparison results.

[0034] Specifically, if there is a large deviation between the predicted data and the actual data, use an optimization algorithm (such as gradient descent, genetic algorithm, etc.) to adjust the model parameters, or update the model's prediction algorithm to better adapt to the actual operating conditions.

[0035] Real-time structural status data is calculated by the digital twin model based on multi-source input data. It reflects the current structural health of the wind turbine hybrid tower. For example, it can include information such as current stress distribution, deformation level, and vibration status. This data is updated in real time, helping operators to understand the operating status of the wind turbine hybrid tower and make appropriate maintenance decisions.

[0036] Step 104: Input the multi-source data and real-time structural status data into the pre-trained LSTM-GRU hybrid neural network model, and combine it with the attention mechanism to predict the damage evolution trend of the wind turbine hybrid tower.

[0037] Specifically, the LSTM-GRU hybrid neural network model can be trained in the following ways: Collect sample data of wind turbine hybrid towers and divide the sample data into training sample data and verification sample data; Specifically, the sample data may include structural response data of the wind turbine hybrid tower, environmental load data, historical operation and maintenance data, etc. These data reflect the operating status and damage of the hybrid tower under different working conditions.

[0038] The LSTM-GRU hybrid neural network model is trained using training sample data, and the attention mechanism is used to weight the input training sample data, automatically focusing on sensitive parameters; Specifically, LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) are two commonly used recurrent neural network (RNN) architectures that can process time series data and capture temporal dependencies within the data. The LSTM-GRU hybrid neural network model combines LSTM and GRU, leveraging their respective strengths to improve model performance. The attention mechanism is a technique that enables a model to automatically focus on important portions of input data. By assigning different weights to different input data, the model's sensitivity to key features is increased. Sensitive parameters are input parameters that have a significant impact on wind turbine tower damage prediction. For example, strain data, wind speed, and maintenance records are sensitive parameters because they significantly influence crack damage prediction. By identifying and weighting these sensitive parameters, the model can more accurately predict crack damage in wind turbine towers.

[0039] Input the verification sample data into the trained LSTM-GRU hybrid neural network model to obtain the prediction results; Specifically, the validation set data does not participate in model training and can therefore be used to objectively evaluate the generalization ability of the model.

[0040] Determine the error index between the predicted results and the actual data; Specifically, error metrics are used to quantify the difference between the model's predictions and the actual data. Common error metrics include mean squared error (MSE) and mean absolute error (MAE).

[0041] When the error index meets the preset rules, the training of the LSTM-GRU hybrid neural network model is stopped.

[0042] Step 105: Based on the damage evolution trend, a maintenance plan for the wind turbine hybrid tower is generated using a multi-objective particle swarm algorithm.

[0043] Step 105 may include: Determine the damage index based on the damage evolution trend; Based on the operating status and maintenance objectives of the wind turbine hybrid tower, multiple optimization goals are set; among them, the optimization goals include minimizing maintenance costs, minimizing power generation losses, and minimizing risk levels; The damage index and optimization objectives are input into the multi-objective particle swarm algorithm to find the optimal solution that meets all optimization objectives and obtain the optimal solution set; Select the target solution from the optimal solution set based on actual maintenance requirements and resource constraints; Convert the target solution into a maintenance plan.

[0044] In this example, by acquiring multi-source data and building a digital twin model, real-time condition monitoring and damage evolution trend prediction of wind turbine hybrid towers were achieved. Damage prediction was performed using an LSTM-GRU hybrid neural network combined with an attention mechanism, improving prediction accuracy. Based on the damage evolution trend, an optimized maintenance plan was generated using a multi-objective particle swarm algorithm. This balanced maintenance cost, power generation loss, and risk level, improving the overall operational efficiency and safety of wind turbine hybrid towers.

[0045] In one embodiment of this specification, determining a damage indicator based on the damage evolution trend includes: Extracting damage location, damage type, and damage degree as damage characteristics from the damage evolution trend, where damage types include but are not limited to cracks, corrosion, and deformation, and damage degrees include but are not limited to crack length, corrosion depth, and deformation; For each injury type, the rate of change of its injury degree over time is calculated to quantify the development speed of the injury; for example, the change of injury degree at adjacent time points is calculated by the difference method; The degree of damage is classified into three levels: minor, moderate, and severe, based on its impact on functionality. For example, minor damage may not affect normal operation, while severe damage may require immediate repair, and moderate damage requires close observation. Classify the rate of injury progression as slow, moderate, or rapid; for example, slow progression may mean the injury will not significantly worsen over a long period of time, while rapid progression requires close attention; The remaining useful life is divided into three levels according to its length: long, medium and short; The quantified results of the damage degree, development rate, and remaining service life are used as damage indicators. For example, the quantified results of the damage degree, development rate, and remaining service life are integrated into a damage indicator. Specifically, these factors can be weighted and summed to obtain a comprehensive damage indicator.

[0046] In this embodiment, those skilled in the art will understand the meaning of the above-mentioned levels, and will not be elaborated here. The damage evolution trend refers to the law and direction of the development and change of the wind turbine hybrid tower structure damage over time. It reflects the damage from the initial state to the current state, as well as the possible development in the future. By analyzing the damage evolution trend, the severity of the damage, the development speed, and the impact on the safety and service life of the wind turbine hybrid tower structure can be predicted. The damage evolution trend may include damage location, damage type, damage degree, damage development speed, remaining service life, etc.

[0047] This embodiment provides a more detailed and comprehensive indicator system for damage assessment by quantifying damage characteristics (such as damage location, type, extent, development speed, and remaining service life). This helps to more accurately judge the severity and urgency of the damage and provide stronger support for maintenance decisions.

[0048] In one embodiment of this specification, the wind turbine hybrid tower damage prediction method based on digital twins further includes: Based on damage evolution trends and maintenance plans, a full life cycle health record for wind turbine hybrid towers is constructed. This life cycle health record includes but is not limited to initial design parameters, manufacturing process data, installation and commissioning records, real-time monitoring data during operation, damage prediction records, maintenance decision records, and post-retirement assessment data. Use blockchain technology to store and tamper-proof health records throughout the life cycle; When the damage evolution trend reaches the preset damage threshold, the maintenance task is automatically triggered through the smart contract, and a maintenance work order is generated and assigned to the corresponding maintenance team.

[0049] In this embodiment, a smart contract is a self-executing contract clause deployed on the blockchain in the form of computer code. It can automatically trigger and execute the operations defined in the contract when the preset conditions are met. A maintenance work order is a specific description of a maintenance task, which may include task description, damage location, priority, estimated completion time, etc. This embodiment constructs a full life cycle health record for the wind turbine hybrid tower and uses blockchain technology for distributed storage and tamper-proof recording, ensuring the authenticity and integrity of the data. Automatically triggering maintenance tasks through smart contracts improves the automation level and response speed of operation and maintenance, and reduces the operation and maintenance risks caused by human factors.

[0050] In one embodiment of this specification, a full life cycle health profile of a wind turbine hybrid tower is constructed based on damage evolution trends and maintenance plans, further comprising: When new monitoring data or injury prediction results are generated, the data update process is automatically triggered and fed back to the health record in real time; Develop an interactive health record visualization platform and use augmented reality technology to display the health status of wind turbine hybrid towers on the health record visualization platform.

[0051] In this embodiment, real-time updating and visual display of health records are achieved. Augmented reality technology is used to enhance the operation and maintenance personnel's intuitive perception of the health status of the wind turbine hybrid tower, improve the scientificity and timeliness of operation and maintenance decisions, and further enhance operation and maintenance efficiency and equipment reliability.

[0052] In one embodiment of this specification, the wind turbine hybrid tower damage prediction method based on digital twins further includes: Build a virtual reality decision-making support system based on the digital twin model of the wind turbine hybrid tower and damage prediction results; The virtual reality decision-making system can visualize the real-time structural status, damage evolution trend, and maintenance plan of the wind turbine hybrid tower to the operation and maintenance personnel; The implementation process of different maintenance plans is simulated in the virtual reality decision-making system.

[0053] In this embodiment, the real-time status, damage prediction and maintenance plan of the wind turbine hybrid tower are displayed to the operation and maintenance personnel in a visual form through a virtual reality decision-making support system, and the implementation process of different maintenance plans is simulated, providing the operation and maintenance personnel with more intuitive decision support, reducing the risk of maintenance decision-making, and improving the scientificity and safety of operation and maintenance.

[0054] In some other embodiments of this specification, the damage index and optimization objectives are input into the multi-objective particle swarm algorithm to find the optimal solution that meets all optimization objectives, and the optimal solution set is obtained, including: Initialize the particle swarm: Create an initial particle swarm. Each particle represents a possible maintenance plan, including specific maintenance measures, maintenance schedule, maintenance resource allocation, and other information. These initial plans are randomly generated and used as the starting point of the algorithm.

[0055] Fitness function evaluation: Each particle (maintenance solution) is evaluated to determine whether its performance meets the optimization goal. The evaluation is based on the fitness function, which takes into account multiple factors: Repair cost: Evaluate the cost of the maintenance plan, including the cost of materials, labor, and equipment; Power generation loss: Assess the potential reduction in power generation that may result from the implementation of the maintenance plan; Risk level: Assess the safety risks brought by the maintenance plan, such as the risk of accidents during the maintenance process; Damage indicators: Evaluate the effectiveness of the maintenance plan on the current damage, such as the reduction of damage severity and the slowdown of development speed; Through the comprehensive evaluation of these factors, each particle is assigned a fitness value, which indicates the quality of the solution.

[0056] Update particle speed and position: Based on each particle's fitness, the particle's "speed" and "position" are adjusted. "Speed" and "position" here are metaphors, representing the direction and state of a particle's movement in the solution space. If a particle has a good fitness, it's close to the optimization target and will continue moving in that direction. If its fitness is poor, it will adjust its direction and move toward a more optimal direction. This process is achieved through internal rules within the algorithm, with the goal of gradually moving the particle closer to the optimization target.

[0057] Iterative Optimization: This process of evaluation and updating is repeated over and over again. Each iteration reevaluates the fitness of each particle and updates its state based on the new fitness value. This process continues until a pre-defined condition is met, such as reaching a certain number of iterations or until the particle's fitness value stabilizes and no longer shows significant improvement.

[0058] Obtaining the optimal solution set: After multiple iterations, the algorithm generates an optimal solution set consisting of multiple maintenance plans that strike a good balance between multiple optimization objectives, such as maintenance cost, power generation loss, risk level, and damage index.

[0059] In some other embodiments of this specification, multi-source data and real-time structural status data are input into a pre-trained LSTM-GRU hybrid neural network model, and combined with an attention mechanism to predict the damage evolution trend of a wind turbine hybrid tower, including: The topological relationship diagram of the wind turbine hybrid tower is constructed using a graph convolutional network (GCN) for structural response data (such as strain and displacement), and the spatial correlation features of damage-sensitive areas are extracted; Input environmental load data (such as wind speed and temperature) and historical operation and maintenance data into the bidirectional GRU network to capture the temporal dependency of loads; The spatial features and temporal features are fused through the cross-attention mechanism to generate a spatiotemporal joint feature vector; The joint feature vector is input into the fully connected layer, and the prediction result of the damage evolution trend is output.

[0060] In this embodiment, by combining spatial features and temporal features and using the attention mechanism to focus on key information, this method can more accurately predict the damage evolution trend of wind turbine hybrid towers.

[0061] In some other embodiments of this specification, after acquiring multi-source data, performing multi-source data adaptive enhancement and alignment specifically includes: Data quality grading: Structural response data is categorized into high-confidence (e.g., standard deviation ≤ 0.05) and low-confidence (e.g., standard deviation > 0.05) based on sensor accuracy and historical failure rates. Low-confidence data is augmented using a generative adversarial network (GAN) to generate synthetic data with a distribution consistent with the high-confidence data.

[0062] Spatiotemporal alignment optimization: For data with different sampling frequencies (e.g., strain data at 1kHz, wind speed data at 10Hz), the Dynamic Time Warping (DTW) algorithm is used to align timestamps and interpolate to generate time series data with a uniform frequency. For unevenly distributed sensor data (e.g., sparse measurement points at the top of a tower), radial basis function (RBF) interpolation is used to reconstruct the spatially continuous field.

[0063] Self-repair of abnormal data: Build an anomaly detection model based on Isolation Forest to automatically identify sensor drift or communication packet loss data; replace abnormal data with simulation results of the digital twin model and mark it as a virtual data source.

[0064] In this embodiment, through data quality grading, GAN data enhancement, spatiotemporal alignment optimization, and abnormal data self-repair, the quality, consistency, and spatiotemporal alignment accuracy of multi-source data are significantly improved, and the interference of abnormal data is reduced, thereby providing higher quality and more reliable data support for wind turbine hybrid tower damage prediction, effectively improving the accuracy of damage prediction and the overall performance of the system.

[0065] Based on the same general inventive concept, the present invention also protects a wind power hybrid tower damage prediction system based on digital twin, such as Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of a wind power hybrid tower damage prediction system based on digital twins provided by an embodiment of the present invention. The wind power hybrid tower damage prediction system based on digital twins provided by the present invention is described below. The wind power hybrid tower damage prediction system based on digital twins described below and the wind power hybrid tower damage prediction method based on digital twins described above can be used in conjunction with each other.

[0066] The wind power hybrid tower damage prediction system based on digital twin includes a data acquisition module 201, a model construction module 202, a model mapping module 203, a damage prediction module 204 and a maintenance generation module 205.

[0067] The data acquisition module 201 is used to obtain multi-source data of the wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data and historical operation and maintenance data; The model construction module 202 is used to construct a digital twin model of the wind turbine hybrid tower based on the fusion technology of building information model and finite element model; The model mapping module 203 is used to input the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtain real-time structural status data of the wind turbine hybrid tower through the digital twin model; The damage prediction module 204 is used to input the multi-source data and the real-time structural status data into a pre-trained LSTM-GRU hybrid neural network model, and combine the attention mechanism to predict the damage evolution trend of the wind turbine hybrid tower; The maintenance generation module 205 is used to generate a maintenance plan for the wind turbine hybrid tower based on the damage evolution trend using a multi-objective particle swarm algorithm.

[0068] Figure 3 It is a structural diagram of an electronic device provided by an embodiment of the present invention.

[0069] like Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the wind power hybrid tower damage prediction method based on digital twins.

[0070] Furthermore, the logic instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0071] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the wind power hybrid tower damage prediction method based on digital twins provided by the above methods.

[0072] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the wind power hybrid tower damage prediction method based on digital twin provided by the above methods.

[0073] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0074] Through the description of the above embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A wind turbine hybrid tower damage prediction method based on digital twin, characterized in that: include: Acquire multi-source data of the wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data and historical operation and maintenance data; Based on the fusion technology of building information model and finite element model, a digital twin model of the wind turbine hybrid tower is constructed; Inputting the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtaining real-time structural status data of the wind turbine hybrid tower through the digital twin model; Inputting the multi-source data and the real-time structural status data into a pre-trained LSTM-GRU hybrid neural network model, and combining the attention mechanism to predict the damage evolution trend of the wind turbine hybrid tower; Based on the damage evolution trend, a maintenance plan for the wind turbine hybrid tower is generated using a multi-objective particle swarm algorithm.

2. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: Based on the fusion technology of building information model and finite element model, a digital twin model of the wind turbine hybrid tower is constructed, including: Use building information modeling to build a three-dimensional geometric model of the wind turbine hybrid tower; Use finite element model to simulate the mechanical behavior of wind turbine hybrid tower; The building information model and the finite element model are integrated to obtain a digital twin model of the wind turbine hybrid tower.

3. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: Inputting the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model includes: Inputting the multi-source data into the digital twin model, and dynamically adjusting the model parameters through a parameter update mechanism in the digital twin model, so that the digital twin model reflects the operating status of the wind turbine hybrid tower in real time; Predicting multi-source prediction data at the next moment based on the input multi-source data through the digital twin model; Comparing the multi-source prediction data at the next moment with the multi-source data to obtain a comparison result; The digital twin model is corrected based on the comparison result.

4. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: The LSTM-GRU hybrid neural network model is trained in the following way: Collecting sample data of wind turbine hybrid towers, and dividing the sample data into training sample data and verification sample data; The LSTM-GRU hybrid neural network model is trained using the training sample data, and the input training sample data is weighted using an attention mechanism to automatically focus on sensitive parameters; Input the verification sample data into the trained LSTM-GRU hybrid neural network model to obtain a prediction result; Determine the error index between the predicted results and the actual data; When the error indicator meets the preset rule, the training of the LSTM-GRU hybrid neural network model is stopped.

5. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: Based on the damage evolution trend, a maintenance plan for the wind turbine hybrid tower is generated using a multi-objective particle swarm algorithm, including: determining a damage index of the damage according to the damage evolution trend; In combination with the operating status and maintenance objectives of the wind turbine hybrid tower, multiple optimization objectives are set; wherein the optimization objectives include minimizing maintenance costs, minimizing power generation losses, and minimizing risk levels; Inputting the damage index and the optimization objective into a multi-objective particle swarm algorithm to find an optimal solution that satisfies all optimization objectives, thereby obtaining an optimal solution set; Selecting a target solution from the optimal solution set based on actual maintenance requirements and resource constraints; The target solution is converted into a maintenance plan.

6. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 5 is characterized in that: Determine the damage index of the damage according to the damage evolution trend, including: Extracting damage location, damage type, and damage degree as damage characteristics from the damage evolution trend, wherein the damage type includes but is not limited to cracks, corrosion, and deformation, and the damage degree includes but is not limited to crack length, corrosion depth, and deformation; For each injury type, the rate of change of its injury severity over time was calculated to quantify the rate of injury development; The degree of injury is divided into three levels according to its impact on function: mild, moderate and severe; The development speed of the injury is divided into three levels according to its speed: slow, medium and fast; The remaining useful life is divided into three levels according to its length: long, medium and short; The quantitative results of damage degree, development speed and remaining service life are used as damage indicators.

7. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: Also includes: Based on damage evolution trends and maintenance plans, a full life cycle health profile of the wind turbine hybrid tower is constructed; wherein the full life cycle health profile includes but is not limited to initial design parameters, manufacturing process data, installation and commissioning records, real-time monitoring data during operation, damage prediction records, maintenance decision records, and post-retirement assessment data; Utilize blockchain technology to store the entire life cycle health records in a distributed manner and record them in an unalterable manner; When the damage evolution trend reaches the preset damage threshold, the maintenance task is automatically triggered through the smart contract, and a maintenance work order is generated and assigned to the corresponding maintenance team.

8. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 7 is characterized in that: Based on damage evolution trends and maintenance plans, a full life cycle health profile for wind turbine hybrid towers is constructed, including: When new monitoring data or injury prediction results are generated, the data update process is automatically triggered and fed back to the health record in real time; Develop an interactive health record visualization platform and use augmented reality technology to display the health status of wind turbine hybrid towers on the health record visualization platform.

9. The wind turbine hybrid tower damage prediction method based on digital twin according to claim 1 is characterized in that: Also includes: Based on the digital twin model of the wind turbine hybrid tower and the damage prediction results, a virtual reality decision-making support system is constructed; The virtual reality decision-making system is used to visualize the real-time structural status, damage evolution trend, and maintenance plan of the wind turbine hybrid tower to the operation and maintenance personnel. The implementation process of different maintenance plans is simulated in the virtual reality decision-making auxiliary system.

10. A wind power hybrid tower damage prediction system based on digital twin, characterized by: include: A data acquisition module is used to obtain multi-source data of the wind turbine hybrid tower; wherein the multi-source data includes structural response data, environmental load data and historical operation and maintenance data; A model construction module, used to construct a digital twin model of the wind turbine hybrid tower based on the fusion technology of building information model and finite element model; A model mapping module is used to input the multi-source data into the digital twin model to achieve real-time mapping between the wind turbine hybrid tower and the digital twin model, and obtain real-time structural status data of the wind turbine hybrid tower through the digital twin model; A damage prediction module, configured to input the multi-source data and the real-time structural status data into a pre-trained LSTM-GRU hybrid neural network model, and predict the damage evolution trend of the wind turbine hybrid tower in combination with an attention mechanism; A maintenance generation module is used to generate a maintenance plan for the wind turbine hybrid tower based on the damage evolution trend using a multi-objective particle swarm algorithm.

Images