Digital twin modeling and mapping method for offshore wind plant

Abstract

The invention discloses a digital twin modeling and mapping method for an offshore wind plant, and relates to the technical field of new energy, and the method comprises the steps: S1, constructing a digital twin model of the offshore wind plant, S2, mapping the real-time operation data of a fan into a virtual environment based on the digital twin model, and S3, carrying out the mapping of the real-time operation data of the fan into the virtual environment based on a mapping result. S4, on the basis of an analysis result, generating a regulation and control instruction to optimize the power generation efficiency of the fan; according to the digital twin modeling and mapping method for the offshore wind plant, the power generation efficiency of the fan is regulated and controlled according to the real-time operation data of the offshore wind plant so as to solve the problem of energy conversion loss.

Description

Technical Field

[0001] The present invention relates to the field of new energy technologies, and in particular to a digital twin modeling and mapping method for an offshore wind farm. Background Art

[0002] With the growing demand for renewable energy, offshore wind farms, as a key component of clean energy, have been widely adopted worldwide. Offshore wind power systems typically consist of multiple wind turbines located in specific areas of the ocean. These systems convert wind energy into electricity through a grid-connected system and transmit it to the power grid. In practice, the operating state of wind turbines is influenced by both marine environmental conditions (such as wind speed, direction, humidity, and temperature) and the performance of the equipment itself, resulting in fluctuations in energy conversion efficiency.

[0003] To improve the operational stability and power generation efficiency of wind power systems, existing technologies generally use sensors to collect operational data from wind turbines and use monitoring systems to perform status sensing and fault warnings on wind turbines. While these methods can improve system reliability to a certain extent, most are limited to static analysis or scheduled maintenance strategies at the local device level, lacking the ability to globally and dynamically model and optimize the overall operational status of the entire offshore wind farm. Furthermore, in actual operation, the long-term operation of wind turbines is susceptible to various factors such as mechanical wear, climate disturbances, and control system lags, resulting in reduced power generation efficiency and unstable power output. Particularly in marine environments with frequent wind speed fluctuations, traditional control strategies struggle to respond quickly to real-time operational data and are unable to achieve real-time regulation of wind turbine performance. This ultimately leads to reduced wind energy utilization and increased operation and maintenance costs, severely restricting the economic benefits and operational efficiency of wind farms. Summary of the Invention

[0004] The purpose of the present invention is to provide a digital twin modeling and mapping method for offshore wind farms, which can regulate the power generation efficiency of wind turbines according to the real-time operation data of offshore wind farms to solve the problem of energy conversion loss.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for modeling and mapping digital twins of offshore wind farms, the method comprising:

[0006] S1. Build a digital twin model of an offshore wind farm, including obtaining the operating level of each wind turbine in the wind farm, obtaining the topological distance between wind turbines, setting the modeling coupling factor, calculating the coupling strength of the wind turbines, and constructing the coupling network structure of the wind farm based on the coupling strength of each pair of wind turbines to form the logical framework of the twin model;

[0007] S2. Based on the digital twin model, the real-time operation data of the wind turbine is mapped into the virtual environment. This includes collecting all the current key state parameters of the wind turbine, calculating the current characteristic value of each wind turbine, normalizing and calculating the probability of the wind turbine state in the twin map, and mapping the graphical state of the wind turbine based on the probability of each wind turbine state in the twin map.

[0008] S3. Based on the mapping results, analyze the wind turbine power generation efficiency and identify energy conversion losses, including continuously recording the power generation efficiency of each wind turbine, calculating the efficiency change between adjacent time periods, and calculating the efficiency change rate. If the absolute value of the efficiency change rate is greater than a preset threshold, mark the wind turbine with an efficiency change rate greater than the preset threshold.

[0009] S4. Based on the analysis results, generate control instructions to optimize the wind turbine power generation efficiency, including constructing three candidate control schemes, performing multi-dimensional scoring on each scheme, setting the weight of each dimension, calculating the comprehensive score of the control scheme, and selecting the control strategy with the highest score as the current control instruction.

[0010] Preferably, the specific formula for calculating the coupling strength of the fan in S1 is:

[0011]

[0012] Among them, F ij Indicates the coupling strength of the fan, M i Indicates the operating level of fan i, M j Indicates the operating level of fan j, D ij represents the topological distance between wind turbine i and wind turbine j, and G represents the modeling coupling factor.

[0013] Preferably, the specific formula for normalizing the probability of the wind turbine state in the twin map in S2 is:

[0014]

[0015] Among them, P(w i ) represents the probability of the wind turbine state in the twin map, f i Represents the current characteristic value of the wind turbine, and n represents the number of wind turbine states currently participating in the normalization calculation.

[0016] Preferably, the specific formula for calculating the efficiency change rate in S3 is:

[0017]

[0018] Where R represents the efficiency change rate, ΔE represents the efficiency change between adjacent time periods, and Δt represents the time period.

[0019] Preferably, the specific formula for calculating the comprehensive score of the control scheme in S4 is:

[0020] S=w1x1+w2x2+w3x3;

[0021] Among them, S represents the comprehensive score of the control scheme, x1, x2, x3 represent the energy-saving score, safety score, and responsiveness score respectively, and w1, w2, w3 represent the energy-saving weight, safety weight, and responsiveness weight respectively.

[0022] Preferably, S1 also includes establishing initial feature templates for wind turbines of different models in the wind farm based on historical operating data; comparing the differences between real-time collected data and the feature templates, and adjusting the parameter matching degree of the twin model; according to the modeling accuracy requirements, enabling multi-level structure modeling for high-load wind turbines and adopting a simplified modeling method for low-load wind turbines; completing the automatic generation of the model structure and visually displaying it on the modeling platform.

[0023] Preferably, S2 also includes setting the spatial layout of wind turbines in the virtual environment to maintain a one-to-one correspondence with the real wind farm; binding each wind turbine to its corresponding operating status data source; setting the numerical range and color mapping relationship of the status parameter; when a parameter exceeds the set threshold, triggering the corresponding highlight, flashing or alarm graphic effect in the virtual model.

[0024] Preferably, S3 also includes performing sliding window processing on the wind turbine power generation efficiency according to daily, weekly and monthly cycles; calculating the average efficiency and peak efficiency in each cycle, and calculating the difference between the two; if the difference is greater than the preset deviation threshold, recording it as a potential efficiency loss event; associating the event with the corresponding wind turbine and marking it as a key tracking target in the twin system.

[0025] Preferably, the S4 also includes retrieving corresponding records of historical control instructions and efficiency responses based on the identified efficiency loss wind turbine; using a scoring model to calculate the comprehensive performance of multiple optional control schemes in different dimensions; judging the adaptability of each scheme under the current working conditions and sorting them; selecting the optimal scheme to form a control instruction, and sending it to the target wind turbine control unit through the control interface.

[0026] Preferably, the scoring model includes a rule-based scoring mechanism and an experience-based weight adjustment mechanism, wherein the rule-based scoring mechanism performs a preliminary score on the control scheme according to a preset threshold, and the experience-based weight adjustment mechanism automatically corrects the weight value of each scoring dimension according to historical execution results.

[0027] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0028] This offshore wind farm digital twin modeling and mapping method constructs a digital twin model of the offshore wind farm. Based on the digital twin model, it maps real-time wind turbine operating data into a virtual environment. Based on the mapping results, it analyzes wind turbine power generation efficiency and identifies energy conversion losses. Based on the analysis results, it generates control instructions to optimize wind turbine power generation efficiency. This method breaks through the static operation and maintenance model of traditional wind power systems, which focuses on local monitoring and periodic maintenance. It achieves dynamic modeling and real-time mapping based on the global state of the wind farm for the first time, enabling comprehensive perception of the operating status and energy efficiency trends of multiple wind turbines in complex marine environments. By constructing a coupled network diagram and digital twin model, it not only reveals the operational relationships between wind turbines but also provides structural support for the coordinated optimization of the entire system. This effectively improves system-level operational visibility and control capabilities, enabling timely identification of potential efficiency losses or abnormally operating wind turbines and rapid generation of adaptive control instructions. This enhances the ability to adjust wind turbine performance in real time, helping to continuously improve energy conversion efficiency and reduce equipment failure rates and operation and maintenance costs. Through the above means, the method of the present invention effectively solves the problem that traditional offshore wind power systems are difficult to achieve global modeling, real-time regulation and intelligent optimization, improves wind energy utilization and power generation stability, enhances the overall economic benefits and operational reliability of wind farms, and has strong engineering application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0030] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0031] like Figure 1 As shown, the present invention provides a technical solution: a digital twin modeling and mapping method for an offshore wind farm, the method comprising:

[0032] S1. Build a digital twin model of an offshore wind farm, including obtaining the operating level of each wind turbine in the wind farm, obtaining the topological distance between wind turbines, setting the modeling coupling factor, calculating the coupling strength of the wind turbines, and constructing the coupling network structure of the wind farm based on the coupling strength of each pair of wind turbines to form the logical framework of the twin model;

[0033] S2. Based on the digital twin model, the real-time operation data of the wind turbine is mapped into the virtual environment. This includes collecting all the current key state parameters of the wind turbine, calculating the current characteristic value of each wind turbine, normalizing and calculating the probability of the wind turbine state in the twin map, and mapping the graphical state of the wind turbine based on the probability of each wind turbine state in the twin map.

[0034] S3. Based on the mapping results, analyze the wind turbine power generation efficiency and identify energy conversion losses, including continuously recording the power generation efficiency of each wind turbine, calculating the efficiency change between adjacent time periods, and calculating the efficiency change rate. If the absolute value of the efficiency change rate is greater than a preset threshold, mark the wind turbine with an efficiency change rate greater than the preset threshold.

[0035] S4. Based on the analysis results, generate control instructions to optimize the wind turbine power generation efficiency, including constructing three candidate control schemes, performing multi-dimensional scoring on each scheme, setting the weight of each dimension, calculating the comprehensive score of the control scheme, and selecting the control strategy with the highest score as the current control instruction.

[0036] This implementation method uses digital twin technology to build a virtual mapping system for offshore wind farms. Its core lies in accurately projecting the operating status of the real wind farm into the virtual environment by modeling the coupling relationship and real-time status between wind turbines. Specifically:

[0037] First, in the modeling stage, the system obtains the operating level of each wind turbine in the wind farm, such as the operating load level, health status level or wind speed adaptation level, and obtains the topological distance between the wind turbines based on the actual geographical layout or communication structure of the wind turbines. On this basis, by setting the modeling coupling factors (such as spatial distance weight, wind energy mutual influence factor, network topology coefficient), the coupling strength between the wind turbines is calculated using mathematical formulas (such as Gaussian function, inverse function or custom coupling model). The coupling strength reflects the degree of interaction between wind turbines, and the larger the value, the stronger the coupling relationship. The system constructs an undirected weighted graph structure based on the coupling strength between all wind turbine pairs to form a wind farm coupling network. This network graph serves as the core framework of the digital twin model to support subsequent dynamic simulation and control.

[0038] During the mapping phase, the system regularly collects key operating status parameters of each wind turbine, such as output power, blade yaw angle, grid current, voltage, speed, etc., and extracts features from these multi-dimensional data to construct a multi-dimensional vector reflecting the operating status of the wind turbine. Through normalization processing and embedded mapping technology, the wind turbine state is converted into a measurable feature space distribution. In the mapping algorithm, the deviation between the historical operating trajectory of the wind turbine and the current state is taken into account, and the probability density function is combined to calculate the distribution probability of its state in the twin space, forming a graphic pattern with physical meaning (such as color depth, node size, edge width, etc.) to intuitively present the real-time overall picture of the virtual wind farm.

[0039] During the efficiency analysis phase, the system continuously records the power generation efficiency of each wind turbine (such as output power per unit wind speed), calculates the efficiency difference between adjacent time periods (such as 5 minutes, 10 minutes or 1 hour), and then calculates the efficiency change rate. By comparing with the set threshold (such as ±5%), wind turbines with abnormal efficiency changes can be quickly identified. These wind turbines may have problems such as blade ice accumulation, wind direction mismatch, or abnormal electronic control system. The marked wind turbines will be highlighted in the twin system to prompt maintenance or dispatch personnel to intervene in time.

[0040] During the control phase, the system generates multiple control schemes for each abnormal or inefficient wind turbine, such as adjusting the yaw angle, changing the speed setting, switching the control algorithm, etc. Through a multi-dimensional scoring system (which may include energy efficiency, structural response, load changes, risk level, operating costs, etc.), each scheme is assigned a weight and a comprehensive score is calculated. Ultimately, the control scheme with the highest score is selected as the control instruction and sent to the target wind turbine through the wind farm scheduling system or remote communication protocol to achieve targeted optimization control. This process can form a closed-loop feedback, that is, after the control is completed, the efficiency changes are continuously monitored to verify the control effect and continuously optimize.

[0041] This implementation method, through the construction of a digital twin model based on a coupled network, enables intuitive, real-time presentation of wind farm status information, enhancing the system's understanding and controllability of wind power operating status. During the mapping process, a normalized probability distribution is used to map wind turbine states, improving the model's sensitivity and accuracy to state changes. By setting a rate-of-change threshold and automatically identifying wind turbines with abnormal efficiency, early warning of potential wind turbine failures or abnormal energy consumption is achieved. A multi-dimensional scoring system is introduced into the control process to comprehensively evaluate the effectiveness of control strategies from multiple dimensions, improving the targeted and scientific nature of the control. Ultimately, this achieves dynamic optimization of wind farm power generation efficiency, enhancing overall energy efficiency and operational stability.

[0042] The specific formula for calculating the coupling strength of the fan in S1 is:

[0043]

[0044] Among them, F ij Indicates the coupling strength of the fan, M i Indicates the operating level of fan i, M j Indicates the operating level of fan j, D ij represents the topological distance between wind turbine i and wind turbine j, and G represents the modeling coupling factor.

[0045] This implementation method introduces a mathematical modeling mechanism to construct a digital twin model for offshore wind farms, thereby achieving accurate mapping and dynamic optimization and regulation of the wind farm's operating status. In the process of building the twin model, the coupling strength between wind turbines is a key parameter of the model, which is defined as follows:

[0046]

[0047] Among them, F ij Indicates the coupling strength of the fan, M i Indicates the operating level of fan i, M j Indicates the operating level of fan j, D ij represents the topological distance between wind turbine i and wind turbine j, and G represents the modeling coupling factor.

[0048] The design of the above formula can reasonably reflect the degree of mutual influence between fans: that is, the higher the operating level and the closer the distance between the two fans, the greater the coupling strength, which means that the mutual influence in control, disturbance or load fluctuation is more obvious. ij The calculation results are used to construct an undirected weighted graph, which is then used as the logical framework to establish the digital twin topology of the wind farm. Based on this graph structure, the system collects wind turbine operating data in real time, performs mapping processing, and visualizes its status. During the mapping process, the status of each wind turbine is normalized into a state vector by extracting key features (such as speed, power factor, yaw angle, etc.), and the graph nodes are adjusted based on the coupling strength graph weight. Ultimately, the operating status of each wind turbine and its coupling relationship are dynamically presented in a virtual interface. Changes in the mapping graph will reflect local disturbances, regional aggregation characteristics, or energy efficiency fluctuation trends in real time.

[0049] This implementation method improves the scientificity and operability of digital twin modeling of wind farms by introducing a clear coupling strength calculation formula. Using the wind turbine operating level and topological distance as the main parameters, the coupling strength formula realizes the quantitative expression of the interaction between wind turbines, making the twin model more accurate in structural construction, and can effectively reflect the actual correlation between wind turbines in the wind farm in terms of power coordination, disturbance conduction and energy conversion efficiency. In addition, the introduction of the inverse square form improves the response capability to spatial distribution sensitivity, allowing the system to more accurately identify closely coupled regional groups, providing precise structural support for subsequent state mapping and efficiency diagnosis. Compared with traditional empirical methods or fuzzy modeling methods, this method has the advantages of clear models, controllable parameters, and strong versatility, which improves the performance of the overall twin system in terms of modeling accuracy and dynamic response, thereby contributing to more efficient and intelligent wind farm operation management and control strategy optimization.

[0050] The specific formula for normalizing the probability of the wind turbine state in the twin map in S2 is:

[0051]

[0052] Among them, P(w i ) represents the probability of the wind turbine state in the twin map, f iRepresents the current characteristic value of the wind turbine, and n represents the number of wind turbine states currently participating in the normalization calculation.

[0053] This implementation method normalizes the eigenvalues ​​of wind turbine states and determines their probabilities within the twin map, driving the dynamic representation of wind turbine graphical patterns within the wind farm virtual environment. During implementation, the system first collects key state parameters for all wind turbines currently participating in the mapping and calculates the current eigenvalues ​​for each wind turbine based on predefined rules or algorithms. These eigenvalues ​​comprehensively reflect the wind turbine's operating status at a specific moment, such as power output, vibration amplitude, and speed fluctuations.

[0054] The system then normalizes the current eigenvalues ​​of all turbines and calculates the relative contribution of each turbine's state to the entire mapping system. Specifically, the eigenvalue of the target turbine is compared with the sum of the eigenvalues ​​of all turbines participating in the normalization calculation to derive the relative probability of that turbine's state. This normalized probability reflects the degree and intensity of the wind turbine's influence on the overall operating characteristics of the twin system under its current operating state.

[0055] These normalized probabilities guide the evolution of wind turbine graphical patterns within the digital twin system, including but not limited to visualization attributes such as node size, color depth, transparency, and activity. By mapping the probabilistic information corresponding to wind turbine states onto graphical parameters, the system enables dynamic response and behavioral visualization of the wind farm virtual model in the graphical dimension. This process not only enhances the system's ability to perceive and distinguish wind turbine operating states but also provides a reliable and unified quantitative foundation for subsequent anomaly identification, cluster analysis, and control strategy recommendations.

[0056] Through the above method, the system achieves unified measurement and real-time response of wind turbine status information without introducing additional complex calculation models, ensuring the interpretability, stability and scalability of the mapping results, and providing a solid foundation for establishing a high-precision, high-dynamic response capability digital twin mapping system for offshore wind farms.

[0057] This implementation method achieves the quantitative expression of wind turbine status in digital twin mapping by introducing a clear normalized probability calculation formula, thereby improving the interpretability and accuracy of the mapping process. The mapping probability of each wind turbine is calculated by the ratio of the current characteristic value of the wind turbine to the sum of the characteristic values ​​of all wind turbines, which can effectively avoid the distortion caused by the difference in a single indicator, thereby ensuring the consistency and fairness of the relative position and performance of each wind turbine status in the virtual mapping. This probability model has the characteristics of normalization, comparability and scalability, which helps to achieve unified quantification and dynamic visualization of status differences in a multi-wind turbine system. It is particularly suitable for constructing a visual mapping of node size, color or activity in a graphical interface. Compared with the traditional method of using static thresholds or fuzzy scoring, this formula can dynamically respond to real-time changes in wind turbine status, making the twin system more sensitive and adaptable, thereby providing a highly reliable input basis for subsequent state identification, cluster analysis and control strategy formulation, and significantly enhancing the level of intelligent data-driven wind farm operation.

[0058] The specific formula for calculating the efficiency change rate in S3 is:

[0059]

[0060] Where R represents the efficiency change rate, ΔE represents the efficiency change between adjacent time periods, and Δt represents the time period.

[0061] This embodiment calculates the efficiency change rate in step S3 to determine the time-varying trend of the wind turbine's power generation efficiency, thereby automatically identifying energy conversion anomalies. During implementation, the system first continuously records the target wind turbine's power generation efficiency data for different time periods. This power generation efficiency can be calculated from the relationship between the output power per unit time and the theoretical wind energy utilization value. Subsequently, the system extracts the difference in efficiency data between two adjacent time periods to represent the efficiency change. At the same time, the system normalizes the efficiency change based on the time interval between the two time periods to obtain the efficiency change rate.

[0062] The efficiency change rate is a key parameter for measuring the stability of a wind turbine's operating status and the consistency of its energy conversion. By comparing efficiency changes against time intervals, the system can identify wind turbines experiencing significant efficiency fluctuations within a short period of time, thereby determining whether there are operational anomalies, mechanical failures, control deviations, or other potential issues. The inclusion of the efficiency change rate not only makes the system more responsive to wind turbine status fluctuations but also provides a key basis for subsequent control.

[0063] The logical rigor of this processing method avoids the one-sidedness caused by the judgment of a single efficiency value, enhances the time dimension expression capability of data analysis, enables the changing process of power generation efficiency to be dynamically quantified and continuously tracked, and ensures the digital twin system's ability to respond to the operating data in a timely manner and capture trends.

[0064] This implementation method can accurately reflect the dynamic change trend of wind turbine power generation performance in different time periods through the calculation method of efficiency change rate. Compared with the method of judging only by a single efficiency value, the efficiency change rate introduces the time dimension, which helps to identify short-term fluctuations or gradually degraded operational problems and improves the system's detection sensitivity to energy efficiency anomalies. At the same time, this method realizes the quantitative expression and trend modeling of efficiency change behavior in the digital twin system, providing a clear and quantitative judgment basis for the automatic generation of subsequent control strategies, significantly enhancing the intelligent operation capability and adaptive optimization level of the wind farm, and improving the operational stability and reliability of the energy conversion system.

[0065] The specific formula for calculating the comprehensive score of the control scheme in S4 is:

[0066] S=w1x1+w2x2+w3x3;

[0067] Among them, S represents the comprehensive score of the control scheme, x1, x2, x3 represent the energy-saving score, safety score, and responsiveness score respectively, and w1, w2, w3 represent the energy-saving weight, safety weight, and responsiveness weight respectively.

[0068] This embodiment calculates the comprehensive score of the control scheme in step S4 to select the optimal control strategy from multiple candidate control schemes to optimize the operating performance and power generation efficiency of the wind turbine. During the specific implementation process, the system first evaluates the scores of each candidate control scheme in the three dimensions of energy saving, safety, and responsiveness. The energy-saving score is used to reflect the performance of the scheme in reducing energy consumption and improving power generation efficiency; the safety score is used to measure its ability to ensure the structural safety, electrical protection, and system stability of the wind turbine; and the responsiveness score is used to evaluate the flexibility and timeliness of the scheme in the face of operating fluctuations, wind speed changes, or fault response.

[0069] After obtaining the scores for each of the above dimensions, the system weights and summarizes the scores for each dimension based on the preset energy-saving weight, safety weight, and responsiveness weight, and calculates the comprehensive score of the corresponding control scheme. This comprehensive score is a quantitative result of the overall performance of the control scheme under multiple evaluation indicators. The weight value can be set according to the current operating strategy or priority target of the wind farm. For example, the safety weight can be increased during the high-load operation phase, and the energy-saving weight can be increased during the energy efficiency optimization cycle. The system compares the comprehensive scores of all candidate control schemes, selects the control strategy with the highest score as the final control instruction, and automatically sends it to the corresponding wind turbine to achieve intelligent control at the system level.

[0070] Through the above processing flow, the system has established a standardized and quantitative control strategy evaluation and decision-making mechanism to ensure that the final selected control solution has comprehensive advantages in various key performance indicators and meets the optimal requirements of the current operating scenario.

[0071] This implementation method realizes the unified quantitative evaluation of multi-dimensional control indicators by introducing a calculation method for the comprehensive score of the control scheme, providing a scientific and systematic evaluation basis for the selection of control strategies. Compared with the traditional control method that relies on manual experience or single indicator priority sorting, this method takes into account the three indicators of energy saving, safety and responsiveness, and realizes adaptive adjustment to different operating goals by flexibly setting weights, with higher intelligence and decision-making accuracy. Through the comprehensive scoring method, the system can dynamically evaluate and accurately screen the optimal control scheme, significantly improving the effectiveness and pertinence of wind turbine regulation. At the same time, this mechanism also enhances the self-learning and self-optimization capabilities of the digital twin system, making the operation of wind farms more efficient, safe and stable, and improving the overall utilization of wind energy and system operation reliability.

[0072] S1 also includes establishing initial feature templates for different types of wind turbines in the wind farm based on historical operating data; comparing the differences between real-time collected data and the feature templates, and adjusting the parameter matching degree of the twin model; enabling multi-level structure modeling for high-load wind turbines and using simplified modeling for low-load wind turbines according to modeling accuracy requirements; completing the automatic generation of the model structure and visually displaying it on the modeling platform.

[0073] This implementation further integrates step S1 by introducing historical operating data to model wind turbine characteristics, achieving dynamic adaptation and structural optimization of the twin model. Specifically, the system first analyzes and processes the historical operating data of different wind turbine models deployed in the wind farm, extracting key indicators reflecting their operating patterns, power response characteristics, vibration characteristics, or fault behaviors, and establishing a representative initial feature template. This template serves as a benchmark for subsequent dynamic model adjustments, used to determine the degree of deviation between the wind turbine's current state and its long-term performance baseline.

[0074] During the real-time modeling process, the system collects the wind turbine's current operating parameters and compares them with the initial characteristic template of the corresponding model. If the difference exceeds a set threshold, the system automatically triggers parameter reconstruction or weight adjustment of the twin model to ensure that the model accurately matches the actual operating characteristics of the current wind turbine, achieving dynamic optimization of parameter matching.

[0075] Furthermore, to balance modeling accuracy and computational efficiency, the system implements a hierarchical modeling strategy based on the current modeling accuracy requirements and the turbine's operating load level. For turbines with higher loads and stricter operational status monitoring requirements, the system uses a multi-level structural modeling approach, combining multiple sub-models to refine modeling across dimensions such as turbine structure, control logic, and environmental impact. For turbines with lower operating loads, a modeling strategy with streamlined parameters and simplified structure is employed to reduce model dimensions and computational complexity.

[0076] Ultimately, the system will automatically generate the corresponding structural configuration for the completed twin model and visualize it in a graphical interface on the modeling platform, including the connection relationship between wind turbines, the model hierarchical structure and dynamic status information, making it convenient for users to view, adjust and analyze the overall operating status of the wind farm and the twin system architecture in real time.

[0077] This implementation method improves the adaptability of the digital twin model to the characteristics of the wind turbine itself and the accuracy of the initialization modeling by establishing an initial feature template based on historical operating data. The introduction of a dynamic comparison mechanism between real-time data and templates enables the model parameters to be adaptively adjusted, thereby improving the accuracy and stability of the twin system in complex operating environments. A differentiated modeling strategy based on load status is adopted to ensure the modeling accuracy of high-load wind turbines while reducing the computing resource usage of low-load wind turbines, thereby optimizing the overall computing efficiency and resource allocation of the system. The automatic generation and visual display function of the model structure further improves the operability and transparency of the system, allowing operation and maintenance personnel to intuitively grasp the operating structure and model evolution process of the wind farm. Overall, this method enhances the intelligence, flexibility and practicality of the twin modeling platform, and effectively supports the refined management and efficient operation of large-scale offshore wind farms.

[0078] S2 also includes setting the spatial layout of wind turbines in the virtual environment to maintain a one-to-one correspondence with the real wind farm; binding each wind turbine to its corresponding operating status data source; setting the numerical range and color mapping relationship of the status parameters; when a parameter exceeds the set threshold, triggering the corresponding highlight, flashing or alarm graphic effect in the virtual model.

[0079] This implementation further improves the mapping mechanism of wind turbine status in the virtual environment based on step S2, enhancing the model's performance in terms of spatial consistency, data binding, and anomaly visualization. Specifically, the system first sets the spatial layout of the wind turbines in the virtual environment, ensuring a one-to-one correspondence with the geographic coordinates of each wind turbine in the actual wind farm. This spatial synchronization allows the layout of wind turbines in the twin system to fully correspond to the physical scene, ensuring that subsequent status displays are closely linked to their actual locations.

[0080] On this basis, the system binds each wind turbine in the virtual environment to its corresponding operating status data source. This data source includes key operating parameters collected in real time, such as power output, wind speed, temperature, current, and voltage. Each wind turbine's status data is transmitted to the twin system in real time, ensuring that the virtual model remains synchronized with the physical object.

[0081] To enhance the intuitive presentation of status information, the system presets numerical ranges for status parameters and assigns corresponding color mappings to each range. For example, a gradient from green to red can represent power load levels, while a gradient from blue to orange can represent temperature conditions. This color change allows users to quickly identify whether the wind turbine's current operating status is safe or abnormal.

[0082] When the system detects that a turbine's status parameter exceeds a preset threshold, it automatically triggers graphical responses within the virtual model, including visual cues such as highlighting, flashing borders, and warning symbols. These graphical effects are non-invasively overlaid on the model view, allowing users to instantly identify potential anomalies without disrupting the overall layout. This allows for rapid location of the problematic turbine and guides subsequent O&M responses.

[0083] This implementation method achieves precise synchronization between the physical layout and the twin mapping by setting a one-to-one correspondence between the wind turbines and the actual wind farm in the virtual environment, laying the foundation for the spatial visualization of status data. The binding of the operating status data source enables the real-time operating status of each wind turbine to be accurately and independently displayed in the virtual platform, avoiding the problem of data confusion or mapping deviation. Through the mapping relationship between the status parameter interval and the color, the system has a clear and intuitive status recognition capability, which helps users quickly obtain key information in the global view. The graphical prompt mechanism for abnormal status further enhances the interactivity and early warning capabilities of the system, so that abnormal operation of the wind farm can be identified and responded to in the first time. Overall, this method improves the spatial consistency, dynamic interaction capability and status response efficiency of the digital twin system, and provides an efficient and visual support platform for real-time monitoring and intelligent decision-making of wind farm operations.

[0084] S3 also includes sliding window processing of wind turbine power generation efficiency according to daily, weekly and monthly cycles; calculating the average efficiency and peak efficiency in each cycle, and calculating the difference between the two; if the difference is greater than the preset deviation threshold, it is recorded as a potential efficiency loss event; the event is associated with the corresponding wind turbine and marked as a key tracking target in the twin system.

[0085] This implementation, building on step S3, introduces a time-period-based sliding window processing mechanism for deeper trend analysis and anomaly identification of wind turbine power generation efficiency. The system first applies sliding window processing to wind turbine power generation efficiency data at three different time scales: daily, weekly, and monthly. The sliding window collects wind turbine efficiency data over different time periods, using continuous time intervals as units, ensuring data continuity and coverage to accommodate trend extraction requirements at different analysis granularities.

[0086] During each sliding cycle, the system calculates the average and peak values ​​of the wind turbine's power generation efficiency to reflect the overall energy efficiency level and maximum performance during that cycle. By comparing the average efficiency with the peak efficiency, the system further calculates the difference between the two, which reflects the degree of performance fluctuation or potential loss range of the wind turbine during that cycle. If the system determines that the difference is greater than the preset deviation threshold, it indicates that the wind turbine has a large efficiency deviation during that cycle, which may be due to structural fatigue, environmental interference, or control strategy failure.

[0087] Once the difference is confirmed to be outside the specified range, the system records the cycle as a potential efficiency loss event and associates it with the corresponding wind turbine. To enhance subsequent operational oversight, the system tags the corresponding wind turbine within the twin system interface, such as by adding an identifier, changing the icon status, or focusing the view to prioritize it for tracking. This tagging not only supports manual O&M intervention but also triggers subsequent data analysis or strategy reassessment.

[0088] This implementation method improves the continuity and depth of wind turbine efficiency evaluation in the time dimension by introducing a sliding window analysis mechanism, and can identify potential operational anomalies that are difficult to detect through instantaneous data alone. By evaluating the difference between average efficiency and peak efficiency at multiple cycle scales, the system can keenly capture fluctuations and anomalies in the long-term operating trend of wind turbines, providing a basis for judging structural losses or control deviations. The automatic recording of difference-exceeding events and the wind turbine binding mechanism realize data-driven target tracking, enabling the twin system to dynamically identify and mark key objects of concern, and optimize subsequent maintenance resource allocation and control strategy formulation. Overall, this method enhances the intelligent capabilities of the twin system in operation monitoring, early warning identification, and operation and maintenance decision-making, and improves the efficiency and stability of long-term operations of wind farms.

[0089] S4 also includes retrieving the corresponding records of historical control instructions and efficiency responses based on the identified efficiency-loss fans; using the scoring model to calculate the comprehensive performance of multiple optional control schemes in different dimensions; judging the adaptability of each scheme under the current working conditions and sorting them; selecting the optimal scheme to form a control instruction, and sending it to the target fan control unit through the control interface.

[0090] This implementation further enhances the control strategy's adaptability assessment and historical feedback analysis capabilities, building on step S4 to establish a control solution recommendation mechanism centered on efficiency response. Specifically, the system automatically retrieves historical control command records and corresponding efficiency response data for wind turbines identified as experiencing efficiency loss. These records encompass control strategies executed at different times, under different environmental or load conditions, and the resulting changes in power generation efficiency, serving as an empirical basis for evaluating the feasibility of the current control strategy.

[0091] Based on historical data, the system uses a pre-set scoring model to comprehensively evaluate multiple candidate control solutions. This scoring model covers multiple dimensions, such as energy efficiency improvement, load balancing, control response speed, and strategy execution risk. Based on the scores for each dimension and the set weights, the system calculates the overall performance of each control solution and, based on this, determines its adaptability to the current operating conditions.

[0092] After the adaptability assessment is complete, the system ranks all control solutions based on their comprehensive scores and selects the control strategy with the highest score and best match for the current wind turbine operating state as the optimal solution. Ultimately, this optimal control solution is converted into specific control instructions through the control interface and issued to the target wind turbine's control unit, enabling dynamic adjustment and optimized control of the wind turbine's operating state.

[0093] This implementation method realizes data-driven and experience-related control strategy selection by retrieving historical control behaviors and efficiency response records, significantly improving the matching and reliability of the control scheme. By introducing a multi-dimensional scoring model and an adaptive judgment mechanism, the system can comprehensively measure the effectiveness and stability of the control scheme under the current working conditions, avoiding the problem of traditional schemes relying on static rules or single-target optimization. The sorting and optimal selection process of the control scheme gives the control instructions a clear priority basis, improving the scientific nature and effectiveness of the control decision. Ultimately, the control interface is used to achieve accurate issuance and rapid execution of instructions, enhancing the closed-loop control capability of the wind turbine system, effectively suppressing efficiency loss, and improving the operating efficiency, response speed and autonomous adjustment level of the wind farm.

[0094] The scoring model includes a rule-based scoring mechanism and an experience-based weight adjustment mechanism. The rule-based scoring mechanism gives a preliminary score to the control scheme based on a preset threshold, and the experience-based weight adjustment mechanism automatically corrects the weight value of each scoring dimension based on historical execution results.

[0095] This implementation further refines the scoring model's composition. By combining a rule-based scoring mechanism with an empirical weight adjustment mechanism, it enhances the rationality and dynamic adaptability of the control scheme's comprehensive scoring. Specifically, the rule-based scoring mechanism, as the first stage of the scoring model, provides preliminary scoring of candidate control schemes across various scoring dimensions based on the system's preset control performance thresholds, optimization target ranges, or safe operation standards. This mechanism establishes clear standard limits to quantitatively evaluate the performance of each control strategy across dimensions such as energy efficiency, safety, and responsiveness, forming a foundational scoring system.

[0096] Based on the rule-based scoring mechanism, the system introduces an experience-based weight adjustment mechanism to dynamically optimize the weight distribution of each scoring dimension. This mechanism performs a retrospective analysis based on the historical control scheme execution results stored in the system, compares the deviation between the historical scores and the actual efficiency response, and identifies the actual contribution of each scoring dimension to the results in different operating scenarios. When a scoring dimension is found to have a stronger impact on efficiency improvement, the system automatically increases the weight of this dimension; otherwise, its weight is reduced. This mechanism enables the scoring model to continuously self-correct and optimize based on actual operational feedback, thereby more accurately evaluating the adaptability and expected effects of the control scheme in different scenarios.

[0097] The two mechanisms work together to ensure that the scoring model has a structured rule basis and can dynamically adapt to changes in strategy feedback in complex operating scenarios, thereby improving the accuracy, robustness and evolutionary capabilities of the scoring model.

[0098] This implementation method effectively solves the problems of strong subjectivity and poor adaptability in control scheme evaluation by constructing a dual scoring model consisting of a rule scoring mechanism and an experience weight adjustment mechanism. The rule scoring mechanism provides a controllable and reusable basic scoring logic to ensure the stability and traceability of the scoring process; the experience weight adjustment mechanism introduces a feedback learning mechanism to dynamically adjust the importance of each scoring dimension based on historical execution results, thereby achieving continuous optimization and personalized adaptation of the scoring model. The combination of the two mechanisms makes the control scheme selection closer to the actual operating characteristics of the wind farm, improves the scientific nature of the control instructions and the accuracy of the execution effect prediction, and ultimately helps the digital twin system form a highly intelligent control system with self-learning and self-regulation capabilities, thereby improving the operating efficiency and strategy response quality of offshore wind farms.

[0099] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A digital twin modeling and mapping method for an offshore wind farm, characterized in that: The method comprises: S1. Build a digital twin model of an offshore wind farm, including obtaining the operating level of each wind turbine in the wind farm, obtaining the topological distance between wind turbines, setting the modeling coupling factor, calculating the coupling strength of the wind turbines, and constructing the coupling network structure of the wind farm based on the coupling strength of each pair of wind turbines to form the logical framework of the twin model; S2. Based on the digital twin model, the real-time operation data of the wind turbine is mapped into the virtual environment. This includes collecting all the current key state parameters of the wind turbine, calculating the current characteristic value of each wind turbine, normalizing and calculating the probability of the wind turbine state in the twin map, and mapping the graphical state of the wind turbine based on the probability of each wind turbine state in the twin map. S3. Based on the mapping results, analyze the wind turbine power generation efficiency and identify energy conversion losses, including continuously recording the power generation efficiency of each wind turbine, calculating the efficiency change between adjacent time periods, and calculating the efficiency change rate. If the absolute value of the efficiency change rate is greater than a preset threshold, mark the wind turbine with an efficiency change rate greater than the preset threshold. S4. Based on the analysis results, generate control instructions to optimize the wind turbine power generation efficiency, including constructing three candidate control schemes, performing multi-dimensional scoring on each scheme, setting the weight of each dimension, calculating the comprehensive score of the control scheme, and selecting the control strategy with the highest score as the current control instruction.

2. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The specific formula for calculating the coupling strength of the fan in S1 is: Among them, F ij Indicates the coupling strength of the fan, M i Indicates the operating level of fan i, M j Indicates the operating level of fan j, D ij represents the topological distance between wind turbine i and wind turbine j, and G represents the modeling coupling factor.

3. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The specific formula for normalizing the probability of the wind turbine state in the twin map in S2 is: Among them, P(w i ) represents the probability of the wind turbine state in the twin map, f i Represents the current characteristic value of the wind turbine, and n represents the number of wind turbine states currently participating in the normalization calculation.

4. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The specific formula for calculating the efficiency change rate in S3 is: Where R represents the efficiency change rate, ΔE represents the efficiency change between adjacent time periods, and Δt represents the time period.

5. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The specific formula for calculating the comprehensive score of the control scheme in S4 is: S=w1x1+w2x2+w3x3; Among them, S represents the comprehensive score of the control scheme, x1, x2, x3 represent the energy-saving score, safety score, and responsiveness score respectively, and w1, w2, w3 represent the energy-saving weight, safety weight, and responsiveness weight respectively.

6. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The S1 also includes establishing an initial feature template for different types of wind turbines in the wind farm based on historical operating data; comparing the difference between the real-time collected data and the feature template, and adjusting the parameter matching degree of the twin model; According to the modeling accuracy requirements, multi-level structure modeling is enabled for high-load fans, and a simplified modeling method is adopted for low-load fans; the model structure is automatically generated and visualized on the modeling platform.

7. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The S2 also includes setting the spatial layout of wind turbines in the virtual environment to maintain a one-to-one correspondence with the real wind farm; binding each wind turbine to its corresponding operating status data source; setting the numerical range and color mapping relationship of the status parameter; when a parameter exceeds the set threshold, triggering the corresponding highlight, flashing or alarm graphic effect in the virtual model.

8. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The S3 also includes performing sliding window processing on the wind turbine power generation efficiency according to daily, weekly and monthly cycles; calculating the average efficiency and peak efficiency in each cycle, and calculating the difference between the two; if the difference is greater than the preset deviation threshold, recording it as a potential efficiency loss event; associating the event with the corresponding wind turbine and marking it as a key tracking target in the twin system.

9. The offshore wind farm digital twin modeling and mapping method according to claim 1, characterized in that: The S4 also includes searching for corresponding records of historical control instructions and efficiency responses based on the identified efficiency loss fans; calculating the comprehensive performance of multiple optional control schemes in different dimensions using a scoring model; judging the adaptability of each scheme under the current working conditions and ranking them; The optimal solution is selected to form a control instruction, which is then sent to the target wind turbine control unit through the control interface.

10. The offshore wind farm digital twin modeling and mapping method according to claim 9, characterized in that: The scoring model includes a rule-based scoring mechanism and an experience-based weight adjustment mechanism, wherein the rule-based scoring mechanism preliminarily scores the control scheme according to a preset threshold, and the experience-based weight adjustment mechanism automatically corrects the weight value of each scoring dimension based on historical execution results.

Images