Offshore high-power wind power plant rapid prediction system and method based on wind resource refined evaluation and field network integration

Abstract

The invention discloses an offshore high-power wind power plant rapid prediction system and method based on wind resource refined evaluation and field network integration. The system comprises a wind resource fine evaluation module used for comprehensively considering wind speed, wind direction, turbulence intensity, atmospheric stability and marine environment parameters to obtain spatial and temporal distribution of wind energy resources; the high-power wind turbine comprehensive modeling module is used for establishing a wind power plant unit power output model in combination with the power curve and a Jensen wake flow model; the neural network prediction module is used for realizing rapid prediction of future power; and the field network integrated rapid solution and grid connection module combines the predicted power with the power grid constraint condition, and outputs the adjusted power curve and power grid operation characteristics by adopting a rapid solution and optimization method. According to the method, the impact of wind power fluctuation on the power grid can be effectively reduced, the wind power grid-connected friend performance and the power grid stability are improved, rapid solving and dispatching optimization of the generating capacity of the high-power wind power plant in the deep and far sea are achieved, and the method has high engineering application value.

Description

Technical Field

[0001] This invention belongs to the field of new energy and power system technology, specifically relating to a rapid prediction system and method for high-power offshore wind farms based on refined wind resource assessment and grid integration. Background Technology

[0002] Wind energy, as an important clean and renewable energy source, is developing rapidly worldwide, especially offshore wind energy, which boasts vast reserves and even greater development potential. With the continuous increase in the capacity of individual wind turbine units and the large-scale construction of deep-sea wind farms, offshore wind power has become an important component of the power structure transformation. However, wind power output itself exhibits significant randomness, volatility, intermittency, and anti-peak-shaving characteristics. Its power output changes rapidly with environmental conditions such as wind speed, wind direction, atmospheric stability, and sea state, making it difficult to maintain stability.

[0003] When wind power is connected to the grid, the uncertainty and volatility of wind power can easily cause problems such as voltage fluctuations, frequency deviations, and power flow imbalances in the lines, thereby impacting the safety and stability of the power grid system. Especially in the scenario of high-power wind turbines and large-scale wind farms, the output of a single unit is concentrated, and the overall power fluctuation is large. Without effective prediction and coordination methods, direct grid connection not only increases the peak-shaving pressure on the power grid, but may also reduce the grid's ability to absorb renewable energy.

[0004] Existing methods either ignore the wake effect and turbine response of wind farms, have insufficient prediction accuracy, low computational efficiency, or lack coupling with the grid level, and cannot meet the needs of rapid prediction and grid-connected operation of high-power wind farms in deep-sea areas. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a rapid prediction system and method for high-power offshore wind power based on integrated grid-field management. This system involves: incorporating marine environmental factors (atmosphere, sea level, topography, etc.) for refined wind resource assessment; employing the Jensen wake model for comprehensive modeling of high-power wind turbines; utilizing deep neural networks to predict power generation; and combining this with a rapid grid-connection solution method to directly apply the prediction results to grid operation reference. This invention significantly improves the accuracy and real-time performance of wind power prediction and effectively mitigates the impact of wind power fluctuations on grid stability through the integrated grid-field management approach, thereby enhancing the grid-friendliness of wind power and the level of renewable energy absorption, providing crucial technical support for large-scale offshore wind power development.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: On the one hand, this invention proposes a rapid prediction system for high-power offshore wind farms based on refined wind resource assessment and integrated wind farm-grid system, comprising the following modules: The refined wind resource assessment module is used to comprehensively consider marine atmospheric conditions, sea level conditions, and topographic factors, and combine wind speed, wind direction, turbulence intensity, and atmospheric stability parameters to conduct multi-dimensional refined modeling and assessment of marine wind resources. High-power wind turbine integrated modeling module: used to build a high-power wind turbine model, and combine the Jensen wake model to simulate the wake effect of the wind turbine, so as to realize the accurate calculation of the power generation of the wind farm; The neural network prediction module is used to predict the output power of wind farms based on wind resource data and wind turbine modeling results, using a deep neural network to achieve nonlinear feature extraction and high-precision fitting. The integrated wind farm and grid rapid solution and grid connection module is used to take the above predicted power as input, establish a unified simulation model of wind farm and grid, and realize the stability analysis and coordinated operation between wind farm and grid through grid connection dynamic response simulation and rapid solution algorithm.

[0007] Furthermore, the wind resource refined assessment module further includes a data acquisition unit, a data processing unit, a key parameter calculation unit, and a wind resource characteristic database unit. The high-power wind turbine integrated modeling module includes a single wind turbine modeling unit, a unit for determining the location of upstream wind turbines, a unit for calculating the wake of a single upstream wind turbine, a unit for calculating the wake superposition effect, and a unit for calculating the power generation. The neural network prediction module includes an input data unit, a dataset partitioning unit, and an output power prediction curve unit. The integrated grid-grid rapid solution and grid connection module includes an input prediction power unit, a stability analysis unit, a stability judgment unit, and an output grid connection dynamic response result unit.

[0008] On the other hand, this invention also proposes a rapid prediction method for high-power offshore wind farms based on refined wind resource assessment and integrated wind farm and grid systems. This method is based on the system described above and includes the following process steps: S1: Taking into account marine atmospheric conditions, sea level and topographic factors, and combining wind speed, wind direction, turbulence intensity and atmospheric stability parameters, we conduct multi-dimensional and refined modeling and evaluation of marine wind resources. S2: Establish a high-power wind turbine model and combine it with the Jensen wake model to simulate the wake effect of the wind turbine in order to achieve accurate calculation of the power generation of the wind farm. S3: Based on wind resource data and wind turbine modeling results, a deep neural network is used to predict the output power of the wind farm, achieving nonlinear feature extraction and high-precision fitting; S4: Using the above predicted power as input, establish a unified simulation model of wind farm and power grid. Through grid-connected dynamic response simulation and fast solution algorithm, realize the stability analysis and coordinated operation between wind farm and power grid.

[0009] Furthermore, step S1 includes the following processes: data acquisition, data processing, key parameter calculation, and wind resource characteristic database storage; specifically, The data collected in the data acquisition process includes sea level height, wave characteristics, ocean current speed, wind speed, wind direction, air temperature, air pressure, sea surface temperature, topographic and sea state parameters, and relative humidity; The data processing technology performs time alignment, spatial interpolation and noise filtering on the acquired data, completes missing value repair and anomaly removal, and unifies the spatial scale of different data sources through coordinate transformation and grid mapping. The key parameters calculated in the process include turbulence intensity, atmospheric stability, air density, and wind power density. In the wind resource feature database storage process, the preprocessed data and key parameter results obtained above are stored in the wind resource feature database for subsequent module calls.

[0010] Furthermore, step S2 includes the following processes: single wind turbine modeling, determining the location of upstream wind turbines, calculating the wake of a single upstream wind turbine, calculating the wake superposition effect, and calculating the power generation.

[0011] Furthermore, in step S2, In the single-wind turbine modeling process, the model is built by determining the turbine parameters, which include power curve, thrust coefficient curve, rotor diameter, hub height, blade length, rated power, cut-in / cut-out wind speed, rated wind speed, yaw angle, pitch angle, and induction factor. In the process of determining the location of upstream wind turbines, based on the prevailing wind direction at the current moment (from the wind resource module) and the latitude and longitude coordinates of all wind turbines in the wind farm, geometric relationship judgments are made to find all wind turbines located in the wind direction of the target wind turbine T. In calculating the wake process of a single upstream wind turbine, the wake velocity deficit is calculated using the Jensen wake formula. The formula is as follows: Where R is the rotor radius; x is the distance from downstream to the rotor surface; k is the wake diffusion coefficient, taken as 0.05 at sea; V0 is the incoming wind speed; and C T ΔVut1, ΔVut2, ΔVut3... are the thrust coefficients; the wind speed deficits [ΔVut1, ΔVut2, ΔVut3...] caused by all upstream wind turbines to the target wind turbine T. In calculating the wake superposition effect, the energy superposition formula is used: (ΔVtotal / V0)² = Σ (ΔVut / V0) 2The total velocity loss ΔVtotal is calculated, and the wake superposition effect Veffective is calculated using Veffective = V0 - ΔVtotal, where V0 is the incoming wind speed; ΔVut is the wind speed loss caused by a single upstream turbine u at the target turbine t; and ΔVtotal is the total wind speed loss at the target turbine location. In the process of calculating power generation, the output power P of a single wind turbine is... W =1 / 2ρAV³Cp where ρ is the air density, A is the swept area, V is the effective wind speed after wake correction (Veffective), and Cp is the power coefficient; unit output power. , where P W This refers to the output power of a single fan.

[0012] Furthermore, in step S3, the data from the wind resource refined assessment module and the unit integrated modeling module are integrated into a database and fed into the neural network for training, with the wind farm output power as the output value; the data in the database is divided into a 70% training set and a 30% test set; the training set is used to train the neural network model, while the test set is used to verify the model's prediction effect on the wind farm output power.

[0013] Furthermore, step S4 includes the following processes: inputting predicted power, stability analysis, stability judgment, and outputting grid-connected dynamic response results; specifically: The power prediction process inputs the power prediction results from the neural network module. In the stability analysis process, the voltage and frequency stability of the system are rapidly evaluated based on power flow equations and dynamic response models. In the stability judgment process, the system voltage and frequency stability are quickly evaluated, and the stability analysis results are judged. If the judgment result meets the requirements, the system enters the next unit. If the result does not meet the requirements, the evaluation signal is fed back to the high-power wind turbine integrated modeling module. The process outputs the voltage response, frequency fluctuation, and total grid-connected power dynamic curves of key nodes in the power grid, providing a basis for decision-making in wind power dispatch and control.

[0014] Furthermore, after receiving feedback, the high-power wind turbine integrated modeling module adjusts the operating status of the upstream wind turbine based on the stability analysis results, including yaw angle adjustment, pitch angle correction, and induction factor optimization. The adjusted results are then fed back into the neural network prediction module and the integrated field-grid rapid solution and grid connection module for a new round of prediction and evaluation, forming a closed-loop iterative optimization.

[0015] The present invention also proposes a computer-readable storage medium, characterized in that the computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the rapid prediction method for high-power offshore wind farms based on refined wind resource assessment and integrated wind farm and grid as described above.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) Significantly improve prediction accuracy: By combining deep learning with integrated modeling of wind turbines, the spatiotemporal evolution of wind power can be more accurately depicted.

[0017] (2) Enhance grid-friendliness: Through integrated grid design, grid dispatch can be flexibly adjusted based on more accurate forecast results, thereby reducing the adverse effects of wind power fluctuations on the grid.

[0018] (3) Support for large-scale application: The proposed fast solution and collaborative optimization method can be adapted to large-scale wind farm grid connection scenarios and has good computational efficiency and scalability.

[0019] (4) Improve system stability: Through the organic combination of power prediction and grid connection optimization, the impact of wind farm output on grid frequency and voltage stability is significantly reduced.

[0020] (5) Achieving closed-loop adaptive optimization: By introducing a feedback adjustment mechanism into the integrated grid module, the system can automatically adjust the wind turbine operating status based on the grid stability analysis results. This closed-loop feedback mechanism realizes the adaptive response and real-time optimization of the wind farm to grid disturbances, significantly improving the overall robustness and operational stability of the system. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall framework of the system of the present invention.

[0022] Figure 2 This is a schematic diagram of the structure of the wind resource refined assessment module in the system of the present invention.

[0023] Figure 3 This is a schematic diagram of the principle of the high-power wind turbine integrated modeling module in the system of the present invention.

[0024] Figure 4 This is a schematic diagram of the neural network prediction module in the system of the present invention.

[0025] Figure 5 This is a schematic diagram of the principle of the integrated field-grid rapid solution and grid connection module in the system of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0027] Please see the appendix Figure 1 This invention provides an integrated, refined assessment and grid connection system for offshore wind farms, mainly comprising four functional modules: a refined wind resource assessment module; a comprehensive modeling module for high-power wind turbines (i.e., a comprehensive modeling module for turbine units); a neural network prediction module; and an integrated, rapid solution and grid connection module. Specifically: Module 1: Refined Wind Resource Assessment Module: This module comprises four parts: a data acquisition unit, a data processing unit, a key parameter calculation unit, and a wind resource feature database unit. It aims to achieve refined feature extraction and spatial distribution modeling of offshore wind resources, providing foundational data for subsequent power prediction and wind farm coupling modeling.

[0028] Module Two: High-Power Wind Turbine Integrated Modeling Module: This module comprehensively considers the parameter characteristics and wake effect of wind turbines to construct a combined aerodynamic-power model of high-power wind turbines. It mainly includes the following five parts: a single wind turbine modeling unit, a unit for determining the location of upstream wind turbines, a unit for calculating the wake of a single upstream wind turbine, a unit for calculating the wake superposition effect, and a unit for calculating power generation. By introducing the aerodynamic characteristics and wake effect of wind turbine units, the overall power output characteristics of the wind farm are comprehensively modeled, improving the realism and accuracy of power simulation.

[0029] Module 3: Neural Network Prediction Module: This module is used to achieve short-term rapid prediction of wind power, and includes an input data unit, a dataset partitioning unit, and an output power prediction curve unit.

[0030] Module 4: Integrated Wind Farm and Grid Connection Module: This module is used to realize the dynamic coupling and rapid response analysis of wind farm output power and grid operation characteristics. It mainly includes four parts: input predicted power, stability analysis unit, stability judgment unit, and output grid connection dynamic response results.

[0031] The four modules are connected sequentially according to the data flow, realizing integrated calculation of the entire process from wind energy resource assessment to wind farm power prediction and wind farm grid connection.

[0032] Please see the appendix Figure 2The refined wind resource assessment module involves four steps: data acquisition, data processing, key parameter calculation, and wind resource characteristic database storage. Its aim is to achieve refined feature extraction and spatial distribution modeling of offshore wind resources. Specifically: Step S201: Obtain marine wind field data by combining multi-source observations and reanalysis data, including numerical weather prediction (NWP) data, reanalysis data (ERA5), buoy observation data, and lidar wind measurement data. Data types include: sea level height, wave characteristics, ocean current velocity, wind speed, wind direction, air temperature, air pressure, sea surface temperature, topographic and sea state parameters, and relative humidity.

[0033] Step S202: Preprocess, clean, and fuse the raw data, checking for missing and outlier values. For missing or outlier data, use interpolation or removal methods. Standardize and unify the units for variables such as wind speed, wind direction, and temperature. Align data from different data sources (wind measurement towers, NWPs, buoys, LiDAR) along the time dimension. In the spatial dimension, use Kriging interpolation to map sparse site data to grid points or wind farm areas of interest.

[0034] Step S203: After obtaining the preprocessed wind resource data and environmental characteristics, the module further calculates the key parameters required for wind power applications, which are then used as input for subsequent modules. These include: turbulence intensity, atmospheric stability, air density, and wind power density.

[0035] Step S204: Store the preprocessed data and key parameter results obtained above into the wind resource feature database for subsequent module calls.

[0036] Please refer to Figure 3. The process of the high-power wind turbine integrated modeling module includes the following five steps: single wind turbine modeling, determining the location of upstream wind turbines, calculating the wake of a single upstream wind turbine, calculating the wake superposition effect, and calculating the power generation; specifically: Step S301: Determine the basic parameters of the wind turbine, including power curve, thrust coefficient curve, rotor diameter, hub height, blade length, rated power, cut-in / cut-out wind speed, rated wind speed, yaw angle, pitch angle, and induction factor.

[0037] Step S301: After receiving feedback from Module 4, the wind turbine operating status will be adjusted based on the stability analysis results, including: yaw angle adjustment, pitch angle correction, and induction factor optimization.

[0038] Step S301: The adjusted results are re-entered into Module 3 and Module 4 for a new round of prediction and evaluation, forming a closed-loop iterative optimization.

[0039] Step S302: Based on the prevailing wind direction at the current moment (from the wind resource module) and the latitude and longitude coordinates of all wind turbines in the wind farm, perform geometric relationship judgment to find all wind turbines located in the wind direction above the target wind turbine T.

[0040] Step S303: Based on the thrust coefficient of the upstream wind turbine, the rotor diameter, and the distance along the wind direction between the upstream wind turbine and the target wind turbine, the Jensen wake model is used to calculate the wake effect of a single upstream wind turbine.

[0041] Step S303: In the wake process calculation for a single upstream wind turbine, the wake velocity deficit is calculated using the Jensen wake formula. The formula is as follows: Where R is the rotor radius; x is the distance from downstream to the rotor surface; k is the wake diffusion coefficient, taken as 0.05 at sea; V0 is the incoming wind speed; and C T ΔVut1, ΔVut2, ΔVut3... are the thrust coefficients; the wind speed deficits [ΔVut1, ΔVut2, ΔVut3...] caused by all upstream wind turbines to the target wind turbine T.

[0042] Step S304: Using the energy superposition formula: (ΔVtotal / V0)² = Σ (ΔVut / V0) 2 The total velocity loss ΔVtotal is calculated, and the wake superposition effect Veffective is calculated using Veffective = V0 - ΔVtotal, where V0 is the incoming wind speed; ΔVut is the wind speed loss caused by a single upstream turbine u at the target turbine t; and ΔVtotal is the total wind speed loss at the target turbine location.

[0043] Step S305: Calculate the fan output power: P W =1 / 2ρAV³Cp where ρ is the air density, A is the swept area, V is the effective wind speed after wake correction (Veffective), and Cp is the power coefficient; unit output power. , where P W This refers to the output power of a single fan.

[0044] Please see the appendix Figure 4 The process of a neural network prediction module includes the following three steps: inputting data, dividing the dataset, and outputting a power prediction curve. Specifically: Step S401: Integrate the data from the wind resource refinement assessment module and the unit integrated modeling module into a single database, feed it into the neural network for training, and use the wind farm output power as the output value. Step S402: Divide the database data into a 70% training set and a 30% test set. The training set is used to train the neural network model, while the test set is used to verify the model's predictive performance on wind farm output power.

[0045] Step S403: The output result is the predicted total power generation of the entire power plant.

[0046] Please see the appendix Figure 5 The integrated power grid rapid solution and grid connection module related processes mainly include four steps: input predicted power, stability analysis unit, stability judgment unit, and output grid connection dynamic response results. Specifically: Step S501: Receive the short-term wind power prediction results output by the neural network prediction module as the input power sequence of the wind farm.

[0047] Step S502: Based on the power flow equation and dynamic response model, quickly evaluate the system voltage and frequency stability.

[0048] Step S503: Judge the stability analysis results. When the obtained grid-connected node voltage deviation is within 0.95-1.05 pu (±5%) and the frequency is within 49.8-50.2 Hz (±0.2 Hz), and the system's weakest mode damping ratio ζ_min ≥ 0.03 and RoCoF ≤ 0.5 Hz / s, it is judged to meet the requirements, and the system proceeds to the next unit. If the results do not meet the requirements, the evaluation signal is fed back to the unit integrated modeling module.

[0049] Step S504: Summarize the optimized system output power, voltage and frequency dynamic results to form the grid-connected dynamic response curve of the wind farm, and output the grid-connected dynamic response results.

[0050] In summary, the integrated power generation prediction and rapid solution system and method for offshore high-power wind turbines proposed in this invention achieves full-link modeling from offshore wind energy resource feature identification to power prediction and grid response analysis through the organic integration of modules such as refined wind resource assessment, comprehensive modeling of high-power wind turbines, neural network power generation prediction, and integrated grid connection solution.

[0051] The system of this invention can accurately predict the output power of wind farms while taking into account the complex marine environment and the wake effect of wind turbines. It can also perform rapid calculations by combining the dynamic response model of the power grid, thereby effectively mitigating the impact of wind power output fluctuations on the stability of the power grid and improving grid friendliness and system operation reliability.

[0052] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A fast prediction system for offshore large-power wind farm based on wind resource refined assessment and field network integration, characterized in that, Comprise the following modules: Wind resource fine evaluation module: for comprehensive consideration of offshore atmospheric conditions, sea level state and topographic factors, combined with wind speed, wind direction, turbulence intensity and atmospheric stability parameters, multi-dimensional fine modeling and evaluation of offshore wind resources; High-power wind turbine comprehensive modeling module: for establishing a high-power wind turbine model, simulating the wind turbine wake effect by combining the Jensen wake model to achieve accurate calculation of wind farm power generation; Neural network prediction module: for predicting the output power of the wind farm based on wind resource data and wind turbine modeling results using deep neural networks to achieve nonlinear feature extraction and high-precision fitting; Field network integrated fast solution and grid connection module: for taking the above predicted power as input, establishing a unified simulation model of the wind farm and the power grid, and through grid dynamic response simulation and fast solution algorithm, realizing the stability analysis and coordinated operation between the wind farm and the power grid.

2. The offshore high-power wind farm fast prediction system based on wind resource fine evaluation and field network integration according to claim 1, characterized in that: The wind resource fine evaluation module further comprises a data acquisition unit, a data processing unit, a key parameter calculation unit, and a wind resource feature database unit; The high-power wind turbine comprehensive modeling module includes a single wind turbine modeling unit, a determination of upstream wind turbine position unit, a calculation of single upstream wind turbine wake unit, a calculation of wake superposition effect unit, and a calculation of power generation unit; The neural network prediction module includes an input data unit, a data set division unit, and an output power prediction curve unit; The field network integrated fast solution and grid connection module includes an input predicted power unit, a stability analysis unit, a stability judgment unit, and an output grid dynamic response result unit.

3. A method for offshore large-power wind farm fast prediction based on wind resource refined evaluation and field network integration, based on the system of claim 1 or 2, characterized in that, Comprise the following process steps: S1: Comprehensive consideration of offshore atmospheric conditions, sea level state and topographic factors, combined with wind speed, wind direction, turbulence intensity and atmospheric stability parameters, multi-dimensional fine modeling and evaluation of offshore wind resources; S2: Establish a high-power wind turbine model, simulate the wind turbine wake effect by combining the Jensen wake model to achieve accurate calculation of wind farm power generation; S3: Based on wind resource data and wind turbine modeling results, use deep neural networks to predict the output power of the wind farm, realize nonlinear feature extraction and high-precision fitting; S4: Taking the above predicted power as input, establishing a unified simulation model of the wind farm and the power grid, and through grid dynamic response simulation and fast solution algorithm, realizing the stability analysis and coordinated operation between the wind farm and the power grid.

4. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and field network integration according to claim 3, characterized in that, In step S1, the following processes are included: data acquisition, data processing, key parameter calculation, and wind resource feature database storage; Specifically, The collected data in the data acquisition process includes sea level height, wave characteristics, current velocity, wind speed, wind direction, air temperature, air pressure, sea surface temperature, topographic and sea state parameters, and relative humidity; The data processing process aligns the acquired data in time, interpolates in space, and filters out noise, completes missing value repair and abnormal rejection, and unifies the spatial scale of different data sources through coordinate transformation and gridding mapping; The key parameters calculated in the key parameter calculation process include turbulence intensity, atmospheric stability, air density, and wind power density. In the wind resource feature database storage process, the preprocessed data and the key parameter results obtained above are stored in a wind resource feature database for subsequent module calling.

5. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and integrated with field network according to claim 3, characterized in that, The step S2 includes the following processes: single wind turbine modeling, determination of upstream wind turbine position, calculation of single upstream wind turbine wake, calculation of wake superposition effect, and calculation of power generation.

6. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and integrated with field network according to claim 5, characterized in that, In the step S2, In the single wind turbine modeling process, modeling is performed by determining wind turbine parameters, including power curve, thrust coefficient curve, wind wheel diameter, hub height, blade length, rated power, cut-in / cut-out wind speed, rated wind speed, yaw angle, pitch angle, and induction factor. In the determination of upstream wind turbine position process, according to the dominant wind direction at the current time and the latitude and longitude coordinates of all wind turbines in the wind farm, geometric relationship is judged to find all wind turbines located in the target wind turbine T. In the process of calculating the single upstream wind turbine wake, the wake velocity deficit is calculated by Jensen wake formula, and the calculation formula is where R is the radius of the wind wheel; x is the distance from the downstream to the wind wheel surface; k is the wake diffusion coefficient, which is 0.05 at sea; V0 is the incoming wind speed, C T is the thrust coefficient; and all the wind speed deficit values [ΔVut1, ΔVut2, ΔVut3...] caused by the upstream wind turbine to the target wind turbine T. In the process of calculating the wake superposition effect, the total speed loss ΔVtotal is calculated by the energy superposition formula: (ΔVtotal / V0)² = Σ (ΔVut / V0) 2 , and the wake superposition effect Veffective is calculated by Veffective = V0 - ΔVtotal, where V0 is the incoming wind speed. ΔVut: wind speed loss value caused by a single upstream wind turbine u at the target wind turbine t; ΔVtotal: total wind speed loss value at the target wind turbine position; In the process of calculating the power generation, the single fan output power P W =1 / 2ρAV³Cp wherein ρ is the air density, A is the swept area, V is the effective wind speed Veffective after the wake correction, and Cp is the power coefficient; the unit output power , wherein P W is the single fan output power.

7. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and integrated with field network according to claim 3, characterized in that, In the step S3, the data of the wind resource fine evaluation module and the unit comprehensive modeling module are integrated into a database, which is brought into a neural network for training, with the wind farm output power as the output value; the data of the database is divided into a 70% training set and a 30% test set; the training set is used to train the neural network model, and the test set is used to verify the prediction effect of the model on the wind farm output power.

8. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and integrated with field network according to claim 3, characterized in that, In the step S4, the following processes are included: input predicted power, stability analysis, stability judgment, and output grid-connected dynamic response results; specifically: In the input predicted power process, the power prediction result from the neural network module is input; In the stability analysis process, the system voltage and frequency stability are quickly evaluated based on the power flow equation and the dynamic response model; In the stability judgment process, the system voltage and frequency stability are quickly evaluated, and the stability analysis result is judged; if the judgment result is that the requirements are met, the system enters the next unit, and if the result does not meet the requirements, the evaluation signal is fed back to the high-power wind turbine comprehensive modeling module; In the output grid-connected dynamic response result process, the voltage response of the key nodes of the power grid, the frequency fluctuation, and the total grid-connected power dynamic curve are output, providing decision basis for wind power dispatching and control.

9. The offshore large-power wind farm fast prediction method based on wind resource refined evaluation and integrated with field network according to claim 8, characterized in that, After receiving the feedback, the high-power wind turbine comprehensive modeling module adjusts the operating state of the upstream wind turbine according to the stability analysis result, including yaw angle adjustment, pitch angle correction, and induction factor optimization; the adjusted result is re-input into the neural network prediction module and the field network integrated fast solution and grid-connected module for a new round of prediction and evaluation, forming a closed-loop iterative optimization.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium includes a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the offshore high-power wind farm fast prediction method based on wind resource fine evaluation and field network integration as claimed in any one of claims 3-9 when the computer program runs.

Images