Wind power plant high-precision three-dimensional terrain modeling method and device

Abstract

The invention belongs to the technical field of wind power plant terrain modeling, and discloses a wind power plant high-precision three-dimensional terrain modeling method and device, and the method comprises the steps: S1, building a wind power plant multi-scale three-dimensional terrain model through dynamic adaptive modeling; s2, microtopographic feature extraction is carried out by using a graph neural network; s3, adopting a multi-scale fusion modeling method to carry out seamless transition on macroscopic landform and microtopographic features; and S4, performing uncertainty quantification through Bayesian inference, and performing optimization by adopting an active learning strategy to obtain a high-precision three-dimensional terrain model of the wind power plant. According to the high-precision three-dimensional terrain modeling method for the wind power plant, through terrain fusion, the precision of the terrain model can be remarkably improved, and micro-terrain features influencing the wind power plant can be accurately captured.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind farm terrain modeling, and in particular relates to a high-precision three-dimensional terrain modeling method and device for a wind farm. Background Art

[0002] The site selection and design of wind farms are crucial to the efficiency of wind power generation, and terrain modeling is a key link. High-precision three-dimensional terrain models can provide important support for the planning, construction and operation of wind farms. Traditional terrain modeling methods mainly rely on terrain surveys and digital elevation models (DEMs). These data are used to build terrain models of wind farms, providing basic information for wind turbine site selection and wind farm design.

[0003] However, with the continuous development of wind power technology and the continuous expansion of wind farms, traditional terrain modeling methods have gradually revealed their limitations. The current technology mainly has the following problems: first, the data collection method is single, and it is difficult to fully capture the detailed features of complex terrain; second, the modeling accuracy is insufficient, especially under complex terrain conditions, it is difficult to accurately reflect the impact of micro-topography characteristics on wind farms; third, the existing modeling methods lack dynamic update capabilities and are difficult to reflect terrain changes in a timely manner; finally, in the modeling process of large-scale wind farms, the calculation efficiency is low and it is difficult to meet the needs of rapid decision-making.

[0004] In view of this, there is an urgent need for a wind farm 3D terrain modeling method that can provide high precision, dynamic update, and multi-scale fusion. This method should be able to comprehensively collect terrain data, accurately capture micro-topography features, achieve real-time updates of the model, and have efficient computing and visualization capabilities. This will help improve the accuracy of wind farm site selection, optimize wind turbine layout, and significantly improve the overall power generation efficiency of wind farms.

[0005] The Chinese patent publication number is CN117993178A, and the name is a patent application for a method for fine three-dimensional terrain mapping and analysis of a wind farm. The patent application uses a drone equipped with a three-dimensional laser scanner to obtain oblique photography images and laser point cloud data to form a city-level real-scene three-dimensional model; uses satellite remote sensing technology to obtain satellite remote sensing data to form a terrain-level real-scene three-dimensional model; combines the city-level real-scene three-dimensional model with the terrain-level real-scene three-dimensional model to generate an integrated three-dimensional digital terrain map; conducts a fine analysis of wind energy resources based on the three-dimensional digital terrain map; uses the three-dimensional digital terrain map to optimize the layout of wind turbines; and monitors and manages the operation status of wind farms based on the three-dimensional digital terrain map. Although the patent application can perform three-dimensional modeling of wind farms, it cannot achieve fine modeling including micro-topographic features. Summary of the invention

[0006] In order to overcome the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a high-precision three-dimensional terrain modeling method and device for a wind farm, which can achieve fine modeling of the wind farm terrain by extracting and integrating micro-topography features with macro-topography.

[0007] To achieve the above object, the technical solution adopted by the present invention is: In a first aspect, the present invention provides a method for high-precision three-dimensional terrain modeling of a wind farm, comprising the following steps: S1: Constructing a multi-scale 3D terrain model of a wind farm through dynamic adaptive modeling; S2: Micro-topography feature extraction using graph neural network; S3: Use multi-scale fusion modeling methods to seamlessly transition macro-topography and micro-topography features; S4: Uncertainty quantification is performed through Bayesian inference and active learning strategy optimization is adopted to obtain a high-precision three-dimensional terrain model of the wind farm.

[0008] Optionally, before step S1, the wind farm terrain data is acquired through a data acquisition system of the wind farm.

[0009] Optionally, before step S1, the wind farm terrain data is preprocessed and fused, wherein the preprocessing includes geometric correction, atmospheric correction, noise removal and data smoothing of the data; the data fusion adopts a deep learning network architecture based on U-Net++ to perform multi-source data fusion, the network introduces spatial and channel attention modules, and the loss function combines L1 loss, structural similarity loss and adversarial loss.

[0010] Optionally, in step S1, the dynamic adaptive modeling adopts an octree structure to implement a wind field-aware adaptive grid system, and dynamically adjusts the grid density according to the wind speed gradient, turbulence intensity and terrain complexity; a real-time update mechanism is implemented based on edge computing devices, and an incremental model update algorithm is adopted, including particle filtering for state estimation, online stochastic gradient descent for online learning, and knowledge distillation technology for model compression.

[0011] Optionally, after step S4, the obtained high-precision three-dimensional terrain model of the wind farm is visualized through an augmented reality system; the augmented reality system uses a Microsoft HoloLens 2 device in combination with an ORB-SLAM3 algorithm to achieve real-time positioning and mapping.

[0012] Optionally, after step S4, the obtained high-precision three-dimensional terrain model of the wind farm is integrated into a digital twin wind farm management system for use in wind farm site optimization and operation management.

[0013] In a second aspect, the present invention provides a high-precision three-dimensional terrain modeling system for a wind farm, comprising: Model building module, used to build a multi-scale three-dimensional terrain model of the wind farm through dynamic adaptive modeling; Micro-topography extraction module, used to extract micro-topography features using graph neural networks; The fusion module is used to seamlessly transition macro-topography and micro-topography features using a multi-scale fusion modeling approach; The optimization module is used to quantify uncertainty through Bayesian inference and optimize using active learning strategies to obtain a high-precision three-dimensional terrain model of the wind farm.

[0014] In a third aspect, the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the high-precision three-dimensional terrain modeling method for a wind farm when executing the computer program.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for high-precision three-dimensional terrain modeling of a wind farm is implemented.

[0016] In a fifth aspect, the present invention provides a computer program product comprising a computer-readable medium, wherein the computer-readable medium contains computer-readable program code, and the program code executes the method for high-precision three-dimensional terrain modeling of a wind farm.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The high-precision three-dimensional terrain modeling method for wind farms of the present invention can significantly improve the accuracy of terrain models through terrain fusion, and can accurately capture the micro-topographic features that affect the wind farm. The present invention enables the model to reflect terrain changes in a timely manner through dynamic adaptive modeling, thereby improving the adaptability of wind farm planning and operation. The present invention achieves a seamless transition from macro-topography to micro-topographic features through a multi-scale fusion modeling method, providing a more comprehensive basis for wind turbine site selection. The present invention improves the reliability and decision-making support capabilities of the model through micro-topographic feature extraction and uncertainty quantification.

[0018] Furthermore, by integrating high-precision terrain models into the digital twin management system, the present invention provides strong support for the full life cycle management of wind farms, helps to optimize wind turbine layout, improve power generation efficiency, and reduce operation and maintenance costs, thereby significantly improving the overall economic and environmental benefits of wind farms.

[0019] Furthermore, the present invention significantly improves the accuracy and comprehensiveness of the terrain model through multi-source data collection and deep learning fusion technology, and can accurately capture the micro-topography features that affect the wind field.

[0020] Furthermore, the present invention provides an intuitive and convenient means for model verification and interaction through an augmented reality system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings described herein are for illustration purposes only and are not intended to limit the scope of the present disclosure in any way.

[0022] In the attached picture: Figure 1 The figure is a flow chart of a method for high-precision three-dimensional terrain modeling of a wind farm according to the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0025] It should be noted that in the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The present application may be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names. In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined. The present invention is described in detail below with reference to the accompanying drawings.

[0026] like Figure 1 As shown, a high-precision three-dimensional terrain modeling method for a wind farm of the present invention comprises the following steps: S1: Constructing a multi-scale 3D terrain model of a wind farm through dynamic adaptive modeling; S2: Micro-topography feature extraction using graph neural network; S3: Use multi-scale fusion modeling methods to seamlessly transition macro-topography and micro-topography features; S4: Uncertainty quantification is performed through Bayesian inference and active learning strategy optimization is adopted to obtain a high-precision three-dimensional terrain model of the wind farm.

[0027] The present invention uses dynamic adaptive modeling to enable the model to reflect terrain changes in a timely manner, thereby improving the adaptability of wind farm planning and operation. By extracting and integrating micro-topography features with macro-topography, fine modeling of wind farm terrain is achieved. The multi-scale fusion modeling method achieves a seamless transition from macro-topography to micro-topography features, providing a more comprehensive basis for wind turbine site selection.

[0028] Example 1 Combine the following Figure 1 The specific implementation modes of the present invention are further described.

[0029] See also Figure 1 The high-precision three-dimensional terrain modeling method for a wind farm of the present invention comprises the following main steps: Multi-source data acquisition: A multi-source data acquisition system is used to obtain wind farm terrain data.

[0030] Data preprocessing and fusion: Preprocess and fuse the acquired data.

[0031] Dynamic adaptive modeling: Build three-dimensional terrain models based on dynamic adaptive modeling technology.

[0032] Micro-topography feature extraction: Micro-topography feature extraction using graph neural network.

[0033] Multi-scale fusion modeling: Multi-scale fusion modeling method is used to achieve seamless transition from macro-topography to micro-topography features.

[0034] Uncertainty quantification and model optimization: Uncertainty quantification is performed through Bayesian inference, and active learning strategies are used to optimize the model.

[0035] High Performance Computing and Visualization: Use high performance computing technology and augmented reality technology for model calculation and visualization.

[0036] Application integration: The obtained 3D terrain model is integrated into the digital twin wind farm management system for wind farm site optimization and operation management.

[0037] Among them, multi-source data acquisition uses a multi-source data acquisition system to obtain wind farm terrain data, including the following three main data sources: Satellite remote sensing: Use high-resolution optical satellite data and SAR satellite data with a resolution better than 0.5m. Optical satellites mainly obtain visible light and near-infrared band data, while SAR satellites provide all-weather and all-day surface information.

[0038] Drone aerial photography: Drone aerial photography systems include fixed-wing drones and multi-rotor drones: Fixed-wing UAV: ​​used for rapid large-scale mapping, with a flight altitude of 200-300m.

[0039] Multi-rotor drone: used for local detailed mapping, with a flight altitude of 50-100m.

[0040] Both drones are equipped with high-resolution cameras and multispectral sensors.

[0041] Data preprocessing includes the following steps: Geometric Correction: Orthorectification using ground control points and digital elevation models (DEM).

[0042] Atmospheric correction: FLAASH algorithm is used for atmospheric correction.

[0043] Noise removal: Use a combination of wavelet transform and Gaussian filtering to remove data noise.

[0044] Data smoothing: Adaptive median filter is used for data smoothing.

[0045] The deep learning data fusion algorithm uses a deep learning network architecture based on U-Net++ to fuse multi-source data. The network introduces spatial and channel attention modules to improve the ability to extract features of different scales. The loss function combines L1 loss, structural similarity (SSIM) loss, and adversarial loss, where the total loss function L_total is defined as follows:

[0046] Among them, α, β, and γ are The weight coefficients of loss, SSIM loss and adversarial loss are used to determine the optimal values ​​through cross-validation.

[0047] Dynamic adaptive modeling includes a wind farm-aware adaptive grid system, which uses an octree structure to implement the wind farm-aware adaptive grid system. The grid refinement strategy is based on the following criteria: Wind speed gradient: Refine the grid when the wind speed difference between adjacent grids exceeds 2m / s.

[0048] Turbulence Intensity: Regions with turbulence intensity above 15% are automatically refined.

[0049] Terrain Complexity: Dynamically adjusts mesh density based on the Terrain Roughness Index (TRI).

[0050] The calculation formula of terrain roughness index (TRI) is as follows:

[0051] in, is the elevation of the central grid cell, is the elevation of the eight adjacent grid cells.

[0052] The real-time update mechanism is implemented based on edge computing devices (NVIDIA Jetson AGX Xavier modules). An incremental model update algorithm is used, including: particle filtering for state estimation, online stochastic gradient descent (Online SGD) for online learning, and knowledge distillation technology for model compression.

[0053] Micro-terrain feature extraction uses the Graph Attention Network (GAT) to extract micro-terrain features. First, a terrain map is constructed based on Delaunay triangulation. The node features include elevation, slope, aspect and roughness, and the edge features include elevation difference, distance and direction.

[0054] Multi-scale fusion modeling includes multi-scale representation based on wavelet transform.

[0055] The multi-scale representation based on wavelet transform uses biorthogonal wavelets for multi-scale decomposition, with 4-6 decomposition layers. The signal f(x) after wavelet transform can be expressed as:

[0056] in, is the scale function, is the wavelet function, and are the scale coefficient and wavelet coefficient respectively, and J is the coarsest decomposition scale.

[0057] The SIBERIA geomorphic evolution model is used to simulate topographic changes in the next 5-30 years. The Monte Carlo method is used to perform parameter uncertainty analysis and generate multiple possible future topographic scenarios.

[0058] Uncertainty quantification and model optimization use Bayesian inference uncertainty quantification and active learning strategies.

[0059] Bayesian inference uncertainty quantification uses No-U-Turn Sampler (NUTS) for sampling and weak information priors. Posterior analysis is performed by calculating credible intervals and prediction intervals to quantify the uncertainty of the model.

[0060] Active learning strategy uses maximum entropy sampling strategy for active learning. Sampling stops when uncertainty decreases to a preset threshold or reaches the budget limit. It is integrated with the drone path planning algorithm to automatically collect supplementary data.

[0061] The high-performance computing system adopts a hybrid cloud architecture, which is a three-layer hybrid cloud architecture, including edge layer, fog layer and cloud layer. The edge layer uses NVIDIA Jetson devices for real-time processing. The fog layer deploys NVIDIA DGX Station for intermediate processing. The cloud layer uses GPU clusters (AWS EC2 P4 instances) for large-scale computing.

[0062] The high-performance computing system uses distributed data processing based on Apache Spark, uses Hadoop Distributed File System (HDFS) to store data, and uses Kubernetes for container orchestration.

[0063] The augmented reality system uses the Microsoft HoloLens 2 device and the ORB-SLAM3 algorithm to achieve real-time positioning and mapping. Interaction is achieved through gesture recognition and voice commands.

[0064] The digital twin wind farm management system integrates high-precision terrain models, meteorological data, wind turbine operation data and power grid data. The digital twin wind farm management system includes real-time monitoring, predictive maintenance, power prediction and optimization scheduling modules. It uses a 3D visualization interface based on the Unity engine.

[0065] The method of integrating genetic algorithm and deep reinforcement learning is used for micro-site optimization. The optimization objectives include maximizing annual power generation, minimizing wake impact, and reducing construction and maintenance costs. At the same time, multiple constraints such as terrain, wind conditions, environmental impact, and grid access are considered.

[0066] Example 2 In this embodiment, the high-resolution optical satellite data uses WorldView-3 satellite data; and the SAR satellite data uses TerraSAR-X satellite data.

[0067] Ground laser scanning uses vehicle-mounted LiDAR or portable LiDAR: Vehicle-mounted LiDAR: including Velodyne HDL-64E, with a scanning frequency of up to 20Hz, used to quickly obtain large-scale terrain data. Portable LiDAR, using LeicaBLK360, is used for detailed scanning of complex terrain and can obtain point cloud data with centimeter-level accuracy.

[0068] Example 3 Based on the high-precision three-dimensional terrain modeling method of a wind farm in Example 1, a high-precision three-dimensional terrain modeling system for a wind farm is disclosed, comprising: A model building module, used to construct a multi-scale three-dimensional terrain model of a wind farm through dynamic adaptive modeling; Micro-topography extraction module, used to extract micro-topography features using graph neural networks; The fusion module is used to seamlessly transition macro-topography and micro-topography features using a multi-scale fusion modeling approach; The optimization module is used to quantify uncertainty through Bayesian inference and optimize using active learning strategies to obtain a high-precision three-dimensional terrain model of the wind farm.

[0069] Example 4 The purpose of this embodiment is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the high-precision three-dimensional terrain modeling method for a wind farm when executing the computer program.

[0070] Example 5 The purpose of this embodiment is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for high-precision three-dimensional terrain modeling of a wind farm is implemented.

[0071] Example 6 The purpose of this embodiment is to provide a computer program product including a computer-readable medium, on which a computer-readable program code is contained, and the program code executes the method for high-precision three-dimensional terrain modeling of a wind farm.

[0072] The steps involved in the devices of the above embodiments 3, 4, 5 and 6 correspond to those of the method embodiment 1. For the specific implementation methods, please refer to the relevant description part of embodiment 1.

[0073] It should be understood by those skilled in the art that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes. These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0074] Unless otherwise specified, the working modes or control modes involved in the above embodiments are all conventional working modes or control modes in the art.

[0075] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application. Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in the field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A high-precision three-dimensional terrain modeling method for a wind farm, characterized in that: The following steps are involved: S1: Constructing a multi-scale 3D terrain model of a wind farm through dynamic adaptive modeling; S2: Micro-topography feature extraction using graph neural network; S3: Use multi-scale fusion modeling methods to seamlessly transition macro-topography and micro-topography features; S4: Uncertainty quantification is performed through Bayesian inference and active learning strategy optimization is adopted to obtain a high-precision three-dimensional terrain model of the wind farm.

2. A high-precision three-dimensional terrain modeling method for a wind farm according to claim 1, characterized in that: Before step S1, the wind farm terrain data is acquired through the data acquisition system of the wind farm.

3. A high-precision three-dimensional terrain modeling method for a wind farm according to claim 1, characterized in that: Before step S1, the wind farm terrain data is preprocessed and data fused, wherein the preprocessing includes geometric correction, atmospheric correction, noise removal and data smoothing; Data fusion uses a deep learning network architecture based on U-Net++ for multi-source data fusion. The network introduces spatial and channel attention modules, and the loss function combines L1 loss, structural similarity loss and adversarial loss.

4. A high-precision three-dimensional terrain modeling method for a wind farm according to claim 1, characterized in that: In step S1, the dynamic adaptive modeling uses an octree structure to implement a wind field-aware adaptive grid system, and dynamically adjusts the grid density according to wind speed gradient, turbulence intensity, and terrain complexity; A real-time update mechanism is implemented based on edge computing devices, and an incremental model update algorithm is adopted, including particle filtering for state estimation, online stochastic gradient descent for online learning, and knowledge distillation technology for model compression.

5. A high-precision three-dimensional terrain modeling method for a wind farm according to claim 1, characterized in that: After step S4, the obtained high-precision three-dimensional terrain model of the wind farm is visualized through an augmented reality system; the augmented reality system uses a Microsoft HoloLens 2 device and combines the ORB-SLAM3 algorithm to achieve real-time positioning and mapping.

6. A high-precision three-dimensional terrain modeling method for a wind farm according to claim 1, characterized in that: After step S4, the obtained high-precision three-dimensional terrain model of the wind farm is integrated into the digital twin wind farm management system for use in wind farm site optimization and operation management.

7. A high-precision three-dimensional terrain modeling system for a wind farm, characterized in that: include: A model building module, used to construct a multi-scale three-dimensional terrain model of a wind farm through dynamic adaptive modeling; Micro-topography extraction module, used to extract micro-topography features using graph neural networks; The fusion module is used to seamlessly transition macro-topography and micro-topography features using a multi-scale fusion modeling approach; The optimization module is used to quantify uncertainty through Bayesian inference and optimize using active learning strategies to obtain a high-precision three-dimensional terrain model of the wind farm.

8. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the high-precision three-dimensional terrain modeling method for a wind farm as claimed in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for high-precision three-dimensional terrain modeling of a wind farm according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer readable medium, characterized in that The computer-readable medium contains computer-readable program code, and the program code executes the method for high-precision three-dimensional terrain modeling of a wind farm as claimed in any one of claims 1 to 6.

Images