Control and optimization of wind turbine wakes for intermittency mitigation and power maximization using artificial intelligence

Abstract

Control and optimization methods for wake flow in wind farms using artificial intelligence (AI) have been developed. Specifically, deep-reinforcement-learning (DRL) based control method and particle-swarm-optimization (PSO) based optimization method have been implemented to maximize the generated wind power. Additionally, a genetic-algorithm (GA) based optimization method has been implemented to minimize wind power intermittency and maximize the generated power simultaneously. For an experimental domain, a multi-fan wind tunnel has been developed to generate a spatiotemporally varying wind flow, and an experimental wind farm has been developed to model a real-scale wind farm. For a computational domain, an in-house code using three-dimensional wake and wake-merging models has been developed to evaluate three-dimensional wind farm power. Reynolds-averaged Navier--Stokes (RANS) simulations have been conducted to analyze the generated power characteristics of wind farms. In the first part of the research, a DRL-based control method to take advantage of complex wake interactions in a wind farm is presented. Although the wind over a wind farm is changing, steady wind has been assumed in the most conventional methods for wind farm control. Under unsteady wind, the generated power of a wind farm becomes stochastic due to intermittent and fluctuating wind. To tackle the difficulty, a DRL-based method with which the pitch and yaw angles of wind turbines in a wind farm are strategically controlled is developed. Time-histories of the past wind and the predicted future wind are both utilized to identify the relation between the generated power and control. The present neural network is trained and validated using an experimental wind farm. The improvement in the generated power by the present DRL-based control method is demonstrated. In the second part of the research, higher annual energy production (AEP) is obtained by joint optimization which considers active yaw control (AYC) in the layout design stage. Although accurate representation of a non-centrosymmetric three-dimensional yawed wake is necessary for the joint optimization of a realistic wind farm, it has not been considered. Furthermore, non-convexity in the joint optimization becomes severe because the layout and yaw angles have to be optimized for all wind directions considering non-centrosymmetric three-dimensional yawed wakes, leading to a not globally but locally optimal layout. To tackle the difficulty, a PSO-based method that is capable of large-scale non-convex joint optimization is developed. In the present method, a farm layout is globally optimized with simultaneous consideration of yaw angles for various wind speeds and directions. The use of random initial particles which consist of the layout and yaw angles of wind turbines prevent from obtaining a locally optimal layout caused by non-convexity of the problem. The improvement in the annual energy production by the present simultaneously optimized layout is demonstrated. In the third part of the research, the wind power intermittency arising from turbulent wind fluctuations and wakes is mitigated by optimizing the layout. Conventional studies have used energy storages to improve the inconsistent power supply but cannot mitigate fluctuations of the actual generated power over time. Temporal fluctuations in the generated power are amplified in specific wind conditions (i.e., speed and direction) due to degraded power caused by wakes. To improve wind power intermittency, the spatial distribution of wake fields has to be designed to compensate for temporal fluctuations in wind conditions. In the present study, a GA-based multi-objective optimization method for a wind farm layout is developed to consider not only wind power intermittency but also the generated power. A new metric of annual wind power intermittency based on a transition probability of wind conditions is defined. The present GA-based method ensures that Pareto optimal layouts are evenly distributed according to multiple cost function values. The Pareto optimal layouts exhibit significantly reduced wind power intermittency and improved power generation compared to the conventional layout.