Downstream Turbine Research Articles

Research on the dynamic characteristics of wind turbine gearboxes under the spatiotemporal inhomogeneous in the wake

In response to the problem of increasing fatigue load and frequent faults caused by the wake effect, the dynamic characteristics and fatigue life analysis of wind turbine gearbox under the Spatial-temporal non-uniformity of wake region were carried out in this paper. A rigid-flexible coupling model of gears transmission system for a 2 MW wind turbine was established. Based on the three-dimensional Station-temporal inhomogeneous wake model proposed in our previous work, the wind speed distribution of six typical locations in the wake region was obtained. Results show that, the spatiotemporal inhomogeneous effects of the wake expansion have great impact on the dynamic characteristics of the gearboxes for the downwind turbines located in different positions. Vibration energy representing the magnitude of impact energy during the operation of the gearbox is significantly higher under wake conditions than under upstream conditions; meanwhile, the gearbox vibration signal contains more impact component in the wake. The velocity deficit and increased turbulence intensity results in less intense but more complex vibration non-smoothness in the turbine transmission system. As the downwind distance downwind increases, the vibration frequency distribution of the transmission system at each moment is larger, but the vibration energy is reduced. Under the 3D spatiotemporal inhomogeneous effects of the wake expansion, the downstream turbine with staggered arrangement is less affected by wake and has a smoother vibration than the tandem arrangement. At the same downstream distance, the gear life is better in a tandem arrangement than in a staggered arrangement (a maximum value happens at the downstream distance of 6D). Research in this paper can provide references for the study of the dynamic characteristics of large wind turbine transmission systems under the wake effects as well as for the micro-siting of wind farms.

Prediction of multiple-wake velocity and wind power using a cosine-shaped wake model

It is well accepted the wakes induced by upstream turbines have an adverse impact on the power production of downstream turbines and this ubiquitous phenomenon severely affects wind farm performance. The aim of the study is to apply a new analytical wake model to prediction of multiple-wake velocity and power output inside wind farms and evaluate its performance. Unlike the classical top-hat model (Jensen model) widely embedded in industry-standard software, the newly proposed analytical wake model, which is derived based on the conservation of both mass and momentum, assumes a trigonometric distribution for the velocity deficit in the wakes of a stand-alone wind turbine and adopts a variable wake growth rate relating ambient turbulence as well as rotor-generated turbulence. Two superposition models including the SED (superposition of energy deficits) and the SVD (superposition of velocity deficits) are employed to simulate the wake-interaction flows and the performance of the new model with the superposition methods is evaluated in two cases: (1) wake interactions of two wind turbines; (2) power prediction of the Horns Rev offshore wind farm. The open-source wind-farm simulation tool FLORIS is employed to make a further comparison of different wake models. It is found that the new and the Gaussian models with the SED are in reasonably good agreement with large-eddy simulation (LES) and measurements compared with other wake models, whereas lower velocity as well as power output is predicted when using the SVD counterpart. This research proposes the SED model as an effective selection strategy for evaluating power output in wind farms when using the cosine and Gaussian models.

The impact of platform motion phase differences on the power and load performance of tandem floating offshore wind turbines

The interactions between tandem floating offshore wind turbines remain an essential research topic with a view to optimizing future large-scale offshore wind farms. The focus of past floating offshore wind turbine aerodynamic research was centred on platform motion impact on rotor aerodynamics for standalone or in-phase tandem turbines. The question of how the differences in the surge motion phase of tandem turbines influence the performance of the downstream turbine remains open. Recent investigations reveal that these motions might not invariably be in-phase. This study seeks to determine the repercussions of platform motion phase differences on power, thrust, and wake recovery. Notably, it postulates that phase differences could detrimentally affect fatigue. Using a validated Navier–Stokes actuator line model rooted in the OpenFOAM® toolbox, the study aims to address this hypothesis. It was found that the out-of-plane root bending moment amplitude increase by 100% when the rotors are out of phase. A 72% increase in amplitude was also observed for CT. Both of these observations have also been linked to an increase in angle of attack amplitude over the blade span. This work stimulates future research in floating offshore turbine control strategies and provides guidance to minimize turbine failures and downtime.

A Reinforcement Learning framework for Wake Steering of Wind Turbines

Ideally, optimum power for a single turbine is obtained when the wind-turbine is aligned with the wind direction. However in multi-turbine wind-farm set-up, wake effects lead to decreased power production from downstream turbine [1, 2, 3, 4, 5]. Hence, a control strategy based on wake steering involves misalignment of upstream turbines with the wind direction causing their wakes to deflect away from downstream wind turbines needs to be investigated. A great deal of work has been put into dynamically controlling the orientation of the individual wind turbines to maximize the power output of the farm [6, 7, 8, 9]. In the wake-steering based control, the misaligned wind turbines produce less power, while the performance of downstream turbines gets enhanced which increases overall net power gain for the wind power plant. Traditionally, the benefits of wake steering have been demonstrated assuming fixed wind directions (e.g., using high-fidelity modeling). Amongst the most recent techniques, particularly promising is the use of Reinforcement learning (RL), which is a branch of machine learning where models are trained to make decisions based on observations of their environment. It is a flexible framework for devising strategies for solving optimal control problems in a broad range of applications across the sciences. Early attempts at using Reinforcement learning for wake steering have been carried out [7, 8, 9], and show promising results. In practice, however, wake-steering controllers must operate in dynamic wind environments in which the wind conditions are estimated from imperfect measurements. Hence, a reinforcement learning framework is developed in this work for dynamic wind conditions. The results show that the framework is promising, and we compare the deep reinforcement learning approach against a considerably more expensive traditional optimization approach which serves as a good baseline. Future work could include looking at more realistic wake models, steering in the presence of noisy observations, and incorporating weather predictions.

An optimum design and economic feasibility analysis of wind farms in Kuwait using different wind generation technologies

Kuwait plans to produce 15 % of its electricity from renewable resources by 2030. This paper aims at designing a 300-MW wind farm in six different sites in Kuwait. The study uses the measured wind data at Kuwait International Airport to predict the wind profile (speed and power density) at the selected sites. To validate the work, two different types of wind generation technologies are used in the design. The first type is Full-Converter Wind Turbine Generators (FCWTGs), and the second type is Doubly Fed Induction Generators (DFIGs). The designed wind farms are simulated using the Wind Atlas Analysis and Application Program (WAsP®) to estimate the Annual Energy Production (AEP) for each type of wind farms in each selected site. The location of every wind turbine in each wind farm (both types) for every selected site is accurately determined while considering needed spacing between rows and columns of turbines. The wake effect for downstream wind turbines is considered and estimated in the design. A comparison between the different wind farms in the six sites using the DIFGs and the FCWTGs generators is carried out. The economic feasibility of the designed wind farms is studied using the estimated Annual Energy Production (AEP) in each one of the designed wind farms.

Performance analysis of vertical-axis wind turbine clusters: Effect of inter-turbine spacing and turbine rotation

Wind energy has emerged as a viable alternative to fossil fuels, with vertical-axis wind turbines (VAWTs) gaining popularity due to their efficiency and adaptability. Combining the actuator line method (ALM) with large eddy simulation (LES) enables accurate performance evaluations, facilitating the design and optimization of wind turbines. The present study invokes ALM-based methodology to perform calculations for the VAWTs. The results of the LES simulations of the VAWTs have been extensively validated against the available experimental and numerical data. The study further explores a VAWT cluster of three turbines by investigating the influence of turbine spacing (in both inline and staggered configuration) on cluster performance. This study shows that the configuration with a streamwise separation (Xsep) of 0.34 D and a transverse separation (Ysep) of 2.5 D exhibits superior performance to other combinations owing to increased kinetic energy in the wake for the downstream turbines. Furthermore, we have presented the effect of varying the rotation direction (in combinations of Clockwise and Counterclockwise rotation) for the individual turbines in the 3-turbine cluster for the two configurations: inline (Xsep = 0 D, Ysep = 2.5 D) and staggered (Xsep = 0.34 D, Ysep = 2.5 D). Staggered counter-rotating turbine cases show reduced performance compared to co-rotating cases, specifically, the clockwise co-rotating (C-C-C) configuration. In the inline configuration, counter-rotating setups outperform co-rotating ones. Counter-rotation analysis reveals that reducing streamwise separation allows turbines to align in line without sacrificing performance, thereby increasing the power density of the turbine cluster.

Wake physics of two-dimensional flapping-hydrofoil turbines

We present a numerical study on the wake of two-dimensional flapping-hydrofoil turbines using Reynolds averaged Navier–Stokes method with shear stress transport k–ω model. The adaptive mesh refinement was applied for vortex simulations. The pitching amplitude ranges from 50° to 90°, and the reduced frequency ranges from 0.10 to 0.20. By varying the reduced frequency and pitching amplitude, we visualized three different types of wakes, and they are the von Karman wake, the mixed wake, and the chaotic wake. We found that there is a critical value of the reduced frequency to determine whether the wake will eventually develop into a standard von Karman wake. When the vortices leave the hydrofoil, they first form a classical staggered arrangement and then develop into a stable double-row configuration. The regular motions of vortices along specific trajectories are explained by analyzing the resultant velocity using velocity polygon, taking into account the effects of vortex interactions and environmental factors. The main component of vortex induced velocities at specific locations is always opposite to the freestream velocity, which is the cause of velocity attenuation in the wake. With the increase in the reduced frequency and pitching amplitude, the velocity attenuation is getting worse. The maximum velocity attenuation usually occurs farther downstream from the hydrofoil for large reduced frequencies and large pitching amplitudes. The wake of a flapping-hydrofoil turbine is divided into four feature zones by studying the time-varying characteristics of velocities, which deepens the understanding of the wake and can provide important references in selecting the optimal location for downstream turbines.

A Combined Delayed Detached Eddy Simulation and Linearized Navier–Stokes Equation Study on the Generation and Reduction of Aerodynamic Noises Inside Steam Turbine Control Valve With Acoustic Liner

Abstract This study was aimed at numerically investigating the source, generation mechanism, and strategy for reducing aerodynamic noises inside a steam turbine control valve. A delayed detached eddy simulation was performed to extract the three-dimensional unsteady turbulent flow structures formed within the serpentine flow passage of the turbine valve. Acoustic analogies, spatial Fourier transform, and spectral proper orthogonal decomposition on the delayed detached eddy simulation-simulated flow data were complementarily combined to clarify the generation mechanism of tonal and broadband aerodynamic noises. The results showed that broadband noises were produced by wall-attached jet flow and turbulent mixing flow between the annular wall jets and central reverse flow. High-intensity tonal noises were generated by the excitation of multi-order natural acoustic modes of the bell-shaped valve spindle. The intensive acoustic pressure pulsations concentrated inside the bell jar and propagated along the diffuser to the downstream turbine chamber. A novel ring acoustic liner was designed using the acoustic impedance model to reduce the valve noises without sacrificing the flow performance. The noise reduction effectiveness was evaluated by solving the linearized Navier–Stokes equations in the frequency domain.

Assessment of yaw-control effects on wind turbine-wake interaction: A coupled unsteady vortex lattice method and curled wake model analysis

Downstream wind turbines in a wind farm are affected by the wake of the upstream turbine. Therefore, wake steering has been introduced to alleviate the wake effect by yawing the shaft of the upstream wind turbine. In this scenario, the upstream and downstream wind turbines encounter non-uniform inflow in wake steering. In this study, the unsteady vortex lattice method (UVLM) was coupled with the curled wake model to consider the unsteady aerodynamics resulting from the wake steering. The proposed method was validated with the National Renewable Energy Laboratory (NREL) Phase VI experiment and Large Eddy Simulation (LES) data simulated with two-turbine case. Three NREL 5 MW wind turbines were simulated in wake steering conditions, and the power performance of the individual turbines and their root flapwise bending moments were investigated. The flow field investigation indicated that the counter-rotating vortices significantly influence the effective angle of attack and aerodynamic loads. In addition, a secondary effect was observed on the effective angle of attack of the rotor and wake deflection. The simulation results demonstrated the applicability of the UVLM in wake steering and its potential for analyzing the wake effects experienced by multiple wind turbines in a wind farm.

A high-efficiency internal combustion engine using oxygen and hydrogen

The work proposes the design of a novel internal combustion engine developed to benefit from the availability on board of cryogenic pressurized hydrogen and oxygen to serve a mechanical propulsion system. The rationale behind the design, as well as the results of simulations, are presented. The two-stroke engine uses direct injection of hydrogen, oxygen, and water, and downstream power turbines, to deliver power densities above 220 kW/L, zero-emission of nitrous oxides and carbon dioxide, and efficiency above 0.70. The novel engine could be an alternative to fuel cell electric propulsion systems, delivering in addition to better hydrogen fuel efficiency, also reduced weight and space of the complete propulsion system.

Effect of atmospheric stability on the dynamic wake meandering model applied to two 12 MW floating wind turbines

AbstractCurrent global analysis tools for floating wind turbines (FWTs) do not account for the combined effects of atmospheric stability and wakes from neighboring turbines. This work uses the mid‐fidelity dynamic wake meandering model, together with turbulent wind fields generated based on stable, neutral, and unstable atmospheric conditions, to study the low‐frequency content of the global responses of two semisubmersible FWTs separated by eight rotor diameters. Incoming wind fields based on the Kaimal spectrum and exponential coherence model, the Mann spectral tensor model, and a time‐series input‐based turbulence model are used. The respective input parameters for these models are fitted to high‐fidelity large eddy simulation data.In unstable, below‐rated conditions, meandering leads to an increase in the yaw standard deviation of the downwind turbine of almost three times larger than the upwind turbine. Deficit and the upwards wake deflection affect the mean pitch and yaw, especially for the below‐rated wind speed scenario. The mean pitch of the downwind turbine is reduced up to half the mean pitch value of the upwind turbine, whereas the mean yaw changes direction due to the enhanced effect of shear. The effect of meandering on the structural loading is highest on the standard deviation of the tower‐top yaw moment of the downstream turbine, which increases more than 2.2 times compared to the upwind turbine value. Based on these findings, atmospheric stability affects wake deficit and meandering which in turn have a profound effect on the low‐frequency global motions and structural response of floating wind turbines.

Strategy for mitigating wake interference between offshore vertical-axis wind turbines: Evaluation of vertically staggered arrangement

Wake interference between wind turbines potentially decreases the power output of offshore wind farms. While the staggered arrangement shows promise for mitigating this interference, the understanding of the impact of vertically staggered setup remains ambiguous, especially in the case of vertical-axis wind turbines (VAWTs). Also, the difference between horizontally and vertically staggered arrangements needs to be clarified. This study aims to shed light on these ambiguities, systematically evaluating the effectiveness of vertically staggered arrangement in mitigating wake interference between offshore VAWTs. High-fidelity computational fluid dynamics simulations are used to investigate the wake interference between two vertically staggered MW-class VAWTs. Typical separation distances (3D ≤ LS ≤ 7D) and different vertically staggered distances (−0.25H ≤ LV ≤ 0.75H) are considered, where D and H denote the turbine diameter and blade span length, respectively. Employing the Taguchi method, the effectiveness of horizontally and vertically staggered arrangements is compared, and the relatively optimal layout of the turbine array is identified. The results show that the vertically staggered arrangement significantly improves the power performance of the downstream turbine, e.g., the power output increases by 75.96% when LS = 7D and LV = 0.25H. When dealing with smaller LS (e.g., 3D), a negative LV is suggested, which reduces the tower cost while preserving the performance improvement. The vertically staggered arrangement has a greater impact on wake interference than the horizontally staggered arrangement. In terms of the relatively optimal layout, the power output of the turbine array is increased by 46.27%. These findings would contribute to the layout design of offshore wind farms.

Wake characterization of tidal turbines in the Pentland Firth using vessel-mounted ADCP measurements

Knowledge of the extension and velocity deficit induced by tidal turbine wakes is crucial for the optimization of tidal farm layouts, as the wake induced by an upstream turbine may substantially affect the power and loadings of a device located downstream. The MeyGen project is the largest planned tidal stream energy project in the world: it aims to develop up to 398 MW of installed power in the Pentland Firth, Scotland, a site with current velocities reaching up to 5 m/s. During Phase 1 of the project, four 1.5 MW turbines were installed, providing a valuable opportunity to investigate turbine wake dynamics in full scale. As part of Phase 2, site characterization campaigns were carried out to plan the deployment of additional turbines, representing 28 MW of tidal power capacity. Hence, the influence of the existing devices on the downstream flow was investigated.
 This work introduces the method used for studying wakes downstream tidal turbines using vessel-mounted Acoustic Doppler Current Profiles (ADCP) measurements. During two spring tides, the data was collected using an ADCP Teledyne Workhorse 600 kHz configured for along-beam velocities recording with 1 m resolution, as well as a GNSS Hemisphere Vector V102. The aim of this study was to map the flow downstream the turbines already installed on site, in order to identify and characterize the wake. Cross-flow transect measurements were conducted at various along-flow distances downstream from the devices, and repeated several times for improved accuracy. The cycles of repeated transects were performed both at flood and ebb tides, with the turbine running or switched off to spot potential differences introduced by device operation. Mean flow velocity estimates in the cross-section were obtained from raw data using the location-based velocity solver developed by Vermeulen et al. (2014). The method provides improved flow velocity estimates from vessel-mounted ADCP measurements because it strongly reduces the spatial extent over which flow homogeneity must be assumed, thus decreasing the chances that this assumption will fail. The reduction of the volume across which homogeneity must be achieved may represent a significant advantage for the investigation of tidal turbine wakes. To the authors’ knowledge, this is the first time that this method has been applied to wake characterization in a tidal stream energy site.
 Outputs of the analysis include mean velocity magnitude and direction, as well as velocity deficits associated with the wakes. The mean velocity estimates obtained in each cross-section are compared for cases when the turbine was running or switched off. Findings reveal that vessel-mounted ADCP transects, coupled with the location-based methodology for velocity estimation, provide a powerful tool for tidal turbine wake characterization. The final presentation and paper will provide the results of this study.

Actuator-Line CFD Simulation of Tidal-Stream Turbines in a Compact Array

An actuator-line CFD model is used to simulate tidal-stream turbines acting alone or in a compact 3-turbine staggered array. CFD results confirm that the accelerated bypass flow between two proximal turbines in shallow water can enhance the power output for a close downstream turbine, with additional smaller effect of the upstream turbines’ operating point. Comparison with other authors’ experimental data in a narrow flume (IFREMER) and circular wave-current tank (FloWave) show some difference in relative loads between the turbines, possibly associated with ambiguity in overall blade pitch and difficulty in characterising onset flow in the FloWave tank. The accelerated bypass flow is persistent and largely established on the rotor plane of the upstream turbines, indicating how local array effects might be incorporated in simpler blade-element/momentum-theory design tools.

Experimental study on wake interference characteristics of horizontal-axis wind turbine

To investigate the flow field distribution under the interference of multiple wind turbine wakes, two different types of ground-based LiDARs were used to conduct wind measurement experiments in a wind farm. Taking the wake field distribution characteristics of two wind turbines as an example, combined with wind measurement data, the complex wake interference flow field is classified into three typical interference conditions: non-wake, partial wake and full wake. And the wake development characteristics under the above three wake conditions are quantitatively analyzed. The results show that under the condition of non-wake interference, the wakes are independent of each other, the wake recovery rate is basically the same. Under partial wake and full wake interference conditions, the wake center recovery rate of the downstream wind turbine decreases, and the wake recovery rate drops most obviously under the full wake interference condition.

Wake interaction between two spar-type floating offshore wind turbines under different layouts

Limited by layout space and manufacturing costs, wake interaction in offshore wind farms is inevitable and can have adverse effects on the performance of downstream wind turbines. To gain a better understanding of the wake interaction between Floating Offshore Wind Turbines (FOWTs), this paper conducts coupled aero-hydrodynamic simulations for two spar-type FOWTs under different layouts. The Unsteady Actuator Line Model (UALM) is used to analyze the unsteady aerodynamic loads of the wind turbine, while the hydrodynamic responses of the floating support platform are obtained using the Computational Fluid Dynamics (CFD) method. To predict the aero-hydrodynamic performance of the FOWT under combined wind and waves, an in-house CFD code developed at Shanghai Jiao Tong University (SJTU), called FOWT-UALM-SJTU solver, is utilized. First, grid convergence test and time step sensitivity study are performed to determine appropriate simulation parameters. Subsequently, numerical simulations of two FOWTs under tandem and offset layouts are conducted to investigate the influence of wake interaction on the performance of downstream FOWT. The dynamic responses of the FWOT, including aerodynamic loads, platform motions, and wake characteristics, are analyzed in detail. From the simulation results and discussions, several conclusions are drawn. Both platform motions and wake interaction contribute to an increased variation range of inflow wind speed experienced by the downstream FOWT, thereby exacerbating the instability of its aerodynamic loads. Under the tandem layout, the platform motions of the upstream FOWT and the downstream FOWT exhibit opposite effects on the aerodynamic loads of the downstream FOWT. Moreover, platform motions increase turbulence intensity in the wake region, accelerating wake velocity recovery and widening the wake width.

A data-driven machine learning approach for yaw control applications of wind farms

This study proposes a cost-effective machine-learning based model for predicting velocity and turbulence kinetic energy fields in the wake of wind turbines for yaw control applications. The model consists of an auto-encoder convolutional neural network (ACNN) trained to extract the features of turbine wakes using instantaneous data from large-eddy simulation (LES). The proposed framework is demonstrated by applying it to the Sandia National Laboratory Scaled Wind Farm Technology facility consisting of three 225 kW turbines. LES of this site is performed for different wind speeds and yaw angles to generate datasets for training and validating the proposed ACNN. It is shown that the ACNN accurately predicts turbine wake characteristics for cases with turbine yaw angle and wind speed that were not part of the training process. Specifically, the ACNN is shown to reproduce the wake redirection of the upstream turbine and the secondary wake steering of the downstream turbine accurately. Compared to the brute-force LES, the ACNN developed herein is shown to reduce the overall computational cost required to obtain the steady state first and second-order statistics of the wind farm by about 85%.

Benefits of individual pitch control on offshore wind turbine submerged in upstream wake

The present study investigates the benefits of an individual pitch control (IPC) strategy on a DTU 10MW with a floating platform subjected to an upstream wind turbine wake. The wind turbine wakes are generated via large eddy simulation, with which the dynamic response of the downstream wind turbines are calculated using OpenFAST with IPC engaged. Three wind speeds and six arrangements with varying longitudinal and lateral offsets of wind turbine pairs are investigated. Findings show that the large spatial variance of the wind velocity in the rotor-swept area leads to a higher fatigue load of out-of-plane blade-root moment (MOoP). The IPC is effective in suppressing the 1P peak of the power spectral density of MOoP, as well as the yaw fluctuation of the platform motion, which saw a maximum 20% reduction in the damage equivalent load of MOoP and up to 74% reduction in the platform yaw fluctuation in the staggered wind turbine layout.

A Simple Model for Wake-Induced Aerodynamic Interaction of Wind Turbines

Wind turbine aerodynamic interactions within wind farms lead to significant energy losses. Optimizing the flow between turbines presents a promising solution to mitigate these losses. While analytical models offer a fundamental approach to understanding aerodynamic interactions, further development and refinement of these models are imperative. We propose a simplified analytical model that combines the Gaussian wake model and the cylindrical vortex induction model to evaluate the interaction between wake and induction zones in 3.5 MW wind turbines with 328 m spacing. The model’s validation is conducted using field data from a nacelle-mounted LiDAR system on the downstream turbine. The ‘Direction to Hub’ parameter facilitates a comparison between the model predictions and LiDAR measurements at distances ranging from 50 m to 300 m along the rotor axis. Overall, the results exhibit reasonable agreement in flow trends, albeit with discrepancies of up to 15° in predicting peak interactions. These deviations are attributed to the single-hat Gaussian shape of the wake model and the absence of wake expansion consideration, which can be revisited to improve model fidelity. The ‘Direction to Hub’ parameter proves valuable for model validation and LiDAR calibration, enabling a detailed flow analysis between turbines. This analytical modeling approach holds promise for enhancing wind farm efficiency by advancing our understanding of turbine interactions.

Input wind speed forecasting for wind turbines based on spatio-temporal correlation

The scale and power of wind turbines are constantly increasing, resulting in more obvious wake effect in wind farms. Each wind turbine is affected differently by the wake, which increases the prediction deviation of input wind speed. Although Lidar can obtain accurate wind speed in advance, it may only be installed on part of the wind turbine in the wind farm due to economic costs. Accordingly, a method for predicting input wind speed of downstream wind turbines in real time with data from upstream wind turbines is proposed. Based on the wake effect, the method groups upstream and downstream wind turbines together to build a prediction model and improves prediction accuracy by mapping mechanical anemometers to Lidar. Firstly, matched upstream and downstream wind turbines are calculated by the WFSim model and the correlation between wind turbines. Then a Boost-CNN-GRU model based on an iterative approach over sliding windows is constructed to predict the input wind speed of the downstream turbine. The experimental results show that the prediction accuracy of the proposed method meets the engineering requirements with MAPE below 3%, MAE below 0.3, and RMSE below 0.4.