Data-Driven Wind Farm Control via Multiplayer Deep Reinforcement Learning

With large wind energy integration into power systems, wind farms begin to influence power systems in a much more significant manner. As wind energy systems utilize different generator technologies from the one utilized in the conventional power plants, the steady-state, transient and small-signal dynamics, as well as, power system stability will thus be significantly affected. The impact of wind energy systems on the power system dynamics and stability is thus of practical importance. As there is a significant increase in installation of wind turbines equipped with doubly-fed induction generator (DFIG) in recent years, a dynamic model of the DFIG wind turbine is firstly developed in this thesis. The model is validated against field measurement data, and the validation gives confidence about the accuracy and applicability of the developed model. DFIG wind farms consist of tens to hundreds of identical DFIG wind turbines increasing the complexity of the wind farm model and simulation time. A novel aggregation technique is developed in this thesis that incorporates a multiplication factor, namely mechanical torque compensation factor (MTCF), to the mechanical torque of the full aggregated wind farm model. The MTCF is initially constructed to approximate a Gaussian function by using fuzzy logic method. By optimizing the MTCF on a trial and error basis, less than 10 percent discrepancy is then achieved between the proposed aggregated model and the complete model. The proposed aggregation technique is then applied to a 120 MVA offshore wind farm comprising of 72 DFIG wind turbines and shows higher accuracy in approximating the wind farm dynamics as it appears at the point of common coupling (PCC) as compared to the full aggregated model. The proposed aggregated model computes faster than the complete wind farm model by 90.3 percent during normal operations and 87 percent during grid disturbances. To overcome the adverse effects due to the fluctuating nature of wind, two wind power smoothing methods are proposed using a fuzzy logic pitch angle controller for a smooth performance with a minimum drop in output power. One method performs partial smoothing with only 4.74 percent drop in the output power while the other method offers complete smoothing with a 8.28 percent drop in output power. The impact of wind energy integration on power system transient stability (PSTS) is studied quantitatively with the transient energy margin (TEM), which is calculated through the evaluation of the transient energy function (TEF). This study is carried out in two ways in the thesis. One is to analyse the impact of transient fault on the DFIG wind turbine as compared to SGs for different factors, like the fault clearing time, the grid coupling, the inertia constant and the voltage sag. The study reveals that transient stability of the DFIG wind turbine is hardly affected by the grid coupling, the inertia constant and the generator terminal voltage sag variations indicating its consistent transient performance within a wide range of these factors. The fault clearing time should be almost 11 percent faster for the system with the DFIG than the synchronous generator (SG). The other is to investigate the impact of the DFIG wind farm on the PSTS with the variation of different factors, which are the voltage sag, the fault clearing time, the load and the wind power penetration level. The study reveals that power systems integrated with DFIG wind farms are sensitive to transient events with high voltage sag, high fault clearing time, low load operation and high wind power penetration level. Machines at different locations individually possess distinct fault response and suffer from a large-scale power imbalance. As a result, reliable operation of DFIG wind farm integrated power systems demands upgraded equipment, such as advanced switchgear, fast breakers/isolators, efficient power reserve systems and advanced reactive power compensating device, etc.

Improving wind farm power output through deactivating selected wind turbines

Optimal design of wind turbine placement in a wind farm is one of the most effective tools to reduce wake power losses by alleviating the wake effect in the wind farm. In comparison to the discrete grid-based wind farm design method, the continuous coordinate method has the property of continuously varying the placement of wind turbines, and hence, is far more capable of obtaining the global optimum solutions. In this paper, the coordinate method was applied to optimize the layout of a real offshore wind farm for both simplified and realistic wind conditions. A new analytical wake model (Jensen-Gaussian model) taking into account the wake velocity variation in the radial direction was employed for the optimization study. The means of handling the irregular real wind farm boundary were proposed to guarantee that the optimized wind turbine positions are feasible within the wind farm boundary, and the discretization method was applied for the evaluation of wind farm power output under Weibull distribution. By investigating the wind farm layout optimization under different wind conditions, it showed that the total wind farm power output increased linearly with an increasing number of wind turbines. Under some particular wind conditions (e.g., constant wind speed and wind direction, and Weibull distribution), almost the same power losses were obtained under the wake effect of some adjacent wind turbine numbers. A common feature of the wind turbine placements regardless of the wind conditions was that they were distributed along the wind farm boundary as much as possible in order to alleviate the wake effect.

Researching on interaction between wind farm and power system, it usually uses simple model to describe the wind farm. The wind turbines in a wind farm always suppose as a coherent generator group, every wind turbines have identical wind speed. As wind turbine locates in the different site in the wind farm, the wind speed at each wind turbines will not change simultaneously. Wake effect is another factor that should take into consideration. So, it needs to get wind speed equivalent model. The equivalent model of wind speed in the wind farm is based on output of power. Gust wind and ramp wind are discussed. Wind farm will change gust wind into step varying wind and make ramp wind into different type wind. The arrangement of wind turbines will also affect equivalent model of wind speed. The result show that the length of wind farm will smooth the power output. It explains why the power output fluctuation of wind farm is smaller than single wind turbine. The onshore wind farm located in Dongtai Jiangsu province is proposed as an example. The equivalent wind speed model and equivalent impedance connected to grid are calculated. The results illustrate the impact of wind speed model on wind farm output.

This paper discusses a data-driven, cooperative control strategy to maximize wind farm power production. Conventionally, every wind turbine in a wind farm is operated to maximize its own power production without taking into account the interactions between the wind turbines in a wind farm. Because of wake interference, such greedy control strategy can significantly lower the power production of the downstream wind turbines and, thus, reduce the overall wind farm power production. As an alternative to the greedy control strategy, we study a cooperative wind farm control strategy that determines and executes the optimum coordinated control actions for maximizing the total wind farm power production. To determine the optimum coordinated control actions of the wind turbines, we employ a data-driven optimization method that seeks to find the optimum control actions using only the power measurement data collected from the wind turbines in a wind farm. In particular, we employ the Bayesian Ascent (BA) algorithm, a probabilistic optimization method constructed based on Gaussian Process regression and the trust region concept. Wind tunnel experiments using 6 scaled wind turbine models are conducted to assess (1) the effectiveness of the cooperative control strategy in improving the power production; and (2) the efficiency of the BA algorithm in determining the optimum control actions of the wind turbines using only the input control actions and the output power measurement data.

A numerical analysis of unsteady inflow wind for site specific vertical axis wind turbine: A case study for Marsabit and Garissa in Kenya

Wind farms use several wind turbines to generate electricity and provide a renewable source of energy. However, due to large forces as a result of hourly and seasonal variations in wind speed and direction, wind turbines experience stochastic loading that often lead to failures of wind turbine components such as the gearbox and generator. Wind turbine failures result in costly repairs and loss in revenue, while stochastic loading makes it difficult to predict the actual condition of a wind turbine. Consequently, it is challenging to schedule maintenance actions to avoid wind turbine component failures. In this paper, we present a stochastic discrete event system specification (STDEVS) model of a wind turbine component. In particular, we consider the wind turbine gearbox, which is one of the components that is often prone to failure, and model its stochastic degradation process. We implement a wind farm simulation with the STDEVS gearbox model and report on the impact of gearbox failures and maintenance scheduling on the performance of a wind farm over a 20-year period.

Abstract. Space–time correlations of power output fluctuations of wind turbine pairs provide information on the flow conditions within a wind farm and the interactions of wind turbines. Such information can play an essential role in controlling wind turbines and short-term load or power forecasting. However, the challenges of analysing correlations of power output fluctuations in a wind farm are the highly varying flow conditions. Here, we present an approach to investigate space–time correlations of power output fluctuations of streamwise-aligned wind turbine pairs based on high-resolution supervisory control and data acquisition (SCADA) data. The proposed approach overcomes the challenge of spatially variable and temporally variable flow conditions within the wind farm. We analyse the influences of the different statistics of the power output of wind turbines on the correlations of power output fluctuations based on 8 months of measurements from an offshore wind farm with 80 wind turbines. First, we assess the effect of the wind direction on the correlations of power output fluctuations of wind turbine pairs. We show that the correlations are highest for the streamwise-aligned wind turbine pairs and decrease when the mean wind direction changes its angle to be more perpendicular to the pair. Further, we show that the correlations for streamwise-aligned wind turbine pairs depend on the location of the wind turbines within the wind farm and on their inflow conditions (free stream or wake). Our primary result is that the standard deviations of the power output fluctuations and the normalised power difference of the wind turbines in a pair can characterise the correlations of power output fluctuations of streamwise-aligned wind turbine pairs. Further, we show that clustering can be used to identify different correlation curves. For this, we employ the data-driven k-means clustering algorithm to cluster the standard deviations of the power output fluctuations of the wind turbines and the normalised power difference of the wind turbines in a pair. Thereby, wind turbine pairs with similar power output fluctuation correlations are clustered independently from their location. With this, we account for the highly variable flow conditions inside a wind farm, which unpredictably influence the correlations.

During recent years, wind energy has moved from an emerging technology to a nearly competitive technology. This fact, coupled with an increasing global focus on environmental concern and a political desire of a certain level of diversification in the energy supply, ensures wind energy an important role in the future electricity market. For this challenge to be met in a cost-efficient way, a substantial part of new wind turbine installations is foreseen to be erected in big onshore or offshore wind farms. This fact makes the production, loading and reliability of turbines operating under such conditions of particular interest.

Large-eddy simulation of a utility-scale wind farm in complex terrain

<strong class="journal-contentHeaderColor">Abstract.</strong> With the increasing growth of wind farm installations, the impact of wake effects from wind turbines on power output, structural loads, and revenue has become a major concern. Hence, there is a need for precise simulation tools to facilitate efficient and cost-effective design and operation of wind farms. To address this need, we present HAWC2Farm, a dynamic and versatile aeroelastic wind farm simulation methodology that combines state-of-the-art engineering models to accurately capture the complex physical phenomena in wind farms. HAWC2Farm models each turbine in a wind farm using the aeroelastic turbine simulator, HAWC2, while sharing a large, high-resolution turbulence box to model atmospheric flow field effects at the farm level, and a modified Dynamic Wake Meandering (DWM) model for capturing wake interactions. The method is computationally efficient and provides valuable insights for wind farm design. It is validated using time series field measurements from the Lillgrund wind farm, including scenarios with yaw steering and a turbine shutdown. HAWC2Farm addresses the challenges of modelling the complex dynamics in wind farms, enabling more accurate, informed and cost-effective design and operation.

<strong class="journal-contentHeaderColor">Abstract.</strong> Offshore wind farms are more commonly installed in wind farm clusters, where wind farm interaction can lead to energy losses; hence, there is a need for numerical models that can properly simulate wind farm interaction. This work proposes a Reynolds-averaged Navier-Stokes (RANS) method to efficiently simulate the effect of neighboring wind farms on wind farm power and annual energy production. First, a novel steady-state atmospheric inflow is proposed. This inflow model is well suited for RANS simulations of large wind farms because it does not lead to the development of nonphysical wind farm wakes. Second, a RANS-based wind farm parametrization is introduced, the actuator wind farm (AWF) model, which represents the wind farm as a forest canopy and allows to use of coarser grids compared to modeling all wind turbines as actuator disks (ADs). When the horizontal resolution of the RANS-AWF model is increased, the model results approach the results of the RANS-AD model. A double wind farm case is simulated with RANS to show that replacing an upstream wind farm with an AWF model only causes a deviation less than 1 % in terms of wind farm power of the downstream wind farm. Most importantly, a reduction in CPU hours of 74.4 % is achieved, provided that the AWF inputs are known, namely, wind farm thrust and power coefficients. The reduction in CPU hours is further reduced when all wind farms are represented by AWF models; namely 89.3 % and 99.9 %, for the double wind farm case and for a wind farm cluster case consisting of three wind farms, respectively. For the double wind farm case, the RANS models predict different wind speed flow fields compared to output from simulations performed with the mesoscale Weather Research and Forecasting model (WRF), but the models are in agreement with the inflow wind speed of the downstream wind farm. The double wind farm case is also simulated with the TurbOPark engineering wake model. Similar wake shapes are reproduced by TurbOPark but the model predicts a larger wind farm wake magnitude compared to RANS and WRF. TurbOPark predicts much better results when its ground model is switched off and a wake expansion coefficient of 0.06 is used. The RANS-AD-AWF model is also validated with SCADA measurements in terms of wind farm shape; the model captures the trend of the measurements for a wide range of wind directions, although the SCADA measurements indicate more pronounced wind farm wake shapes for certain wind directions.

Wake-effect Minimising Optimal Control of Wind Farms, with Load Reduction

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

Mitigating wake effects between wind turbines is crucial for enhancing the overall output power of offshore wind farms. Therefore, optimizing turbine spacing and layout under turbulent conditions is essential. This study employs the NREL-5 MW wind turbine model to investigate the efficiency of a 3 × 3 wind farm. This research focuses on the influence of turbine spacing and layout on wake field distribution and output power characteristics under different turbulence intensities. A key innovation is the application of entropy production theory to quantify energy dissipation and wake recovery, providing a deeper understanding of the underlying mechanisms in energy losses. This research also introduces fatigue analysis based on the Damage Equivalent Load (DEL) method, revealing that staggered layouts significantly reduce cyclic loads and extend turbine lifespan. The results indicate that modifying the layout is a more effective strategy for enhancing the total power output of the wind farm, which proves to be more effective than altering the turbulence intensity. Specifically, staggered layout I (with a downstream stagger of 1.0 rotor diameter (D)) increases total output power by 28.76% (to 36.84 MW) and causes a 16.38% surge in power when the spacing increases to 5D. Expanding the wind turbine spacing mitigates wake interaction, resulting in a dramatic 59.84% power recovery for downstream wind turbines. The wind turbine’s lifespan is extended as a result of fatigue loads on the root bending moment being substantially reduced by the staggered layout, which alters the wake structure and stress distribution. The entropy production analysis shows that regions with high entropy production are primarily concentrated near the rotor and within the wake shear layer. The energy dissipation is substantially reduced in the case of staggered layout. These findings provide valuable guidance for the aerodynamic optimization and long-term operation design of large-scale wind farms, contributing to improved energy efficiency and reduced maintenance costs.