An Optimization Framework for Wind Farm Design in Complex Terrain

Abstract

Designing wind farms in complex terrain is an important task, especially for countries with a large portion of complex terrain territory. To tackle this task, an optimization framework is developed in this study, which combines the solution from a wind resource assessment tool, an engineering wake model adapted for complex terrain, and an advanced wind farm layout optimization algorithm. Various realistic constraints are modelled and considered, such as the inclusive and exclusive boundaries, minimal distances between turbines, and specific requirements on wind resource and terrain conditions. The default objective function in this framework is the total net annual energy production (AEP) of the wind farm, and the Random Search algorithm is employed to solve the optimization problem. A new algorithm called Heuristic Fill is also developed in this study to find good initial layouts for optimizing wind farms in complex terrain. The ability of the framework is demonstrated in a case study based on a real wind farm with 25 turbines in complex terrain. Results show that the framework can find a better design, with 2.70% higher net AEP than the original design, while keeping the occupied area and minimal distance between turbines at the same level. Comparison with two popular algorithms (Particle Swarm Optimization and Genetic Algorithm) also shows the superiority of the Random Search algorithm.

A new wake model and comparison of eight algorithms for layout optimization of wind farms in complex terrain

“…This model was first used in an idealized layout optimization problem, i.e., a wind farm with 25 turbines on a Gaussian 2D hill [33]. Recently, it was also employed in a more realistic optimization framework and tested on the design optimization of a real wind farm in complex terrain [34], which is, in fact, the same wind farm as described in Section 2. Random search algorithm, a metaheuristic algorithm developed specifically for wind farm layout optimization in [19], was used in this framework.…”

“…An adapted Jensen wake model was proposed by Feng and Shen for wind farms in complex terrain and applied in the layout optimization of a wind farm on a 2D Gaussian hill [33]. This wake model was later adopted in a design optimization framework that can consider real wind farms in complex terrain and tested on a wind farm with 25 turbines at a real complex terrain site [34]. Kuo et al [35] solved the layout optimization problem for wind farms in complex terrain by coupling computational fluid dynamics (CFD) with mixed-integer programming (MIP), in which the wake effects were approximated by CFD simulations iteratively.…”

Advances in Wind Farm Layout Optimization: Wind Direction Robustness and Wake Induced Asymmetric Thrust Load

“…These circles are situated in threedimensional space, where the heights depend on the terrain elevations, 𝐸 1 and 𝐸 2 , and the hub heights of the respective turbines, 𝐻 1 and 𝐻 2 , at points 𝑃 1 and 𝑃 2 , respectively. This suggests a "complex terrain" (i.e., various elevations), as tackled by Feng et al [49]. We simplify in terms of the downstream wind flow and assume that 𝑃 1 = (𝑥 1 , 𝑦 1 , 𝑧 1 ) 𝑇 with 𝑧 1 = 𝐸 1 + 𝐻 1 and 𝑃 2 = (𝑥 2 , 𝑦 2 , 𝑧 2 ) 𝑇 with 𝑧 2 = 𝐸 2 + 𝐻 2 .…”

Special Issue on Wind Turbine Aerodynamics

“…Sessarego et al [16] simulated wind turbine/farm flows in complex terrain using both Reynolds averaged Navier-Stokes with an actuator disc model and large eddy simulation with actuator line model and results were compared with detailed field measurements from two met-masts and SCADA (supervisory control and data acquisition) data. Feng et al [17] developed an optimization framework for wind farm layout optimization in complex terrain, which employs a CFD (Computational Fluid Dynamics) wind resource assessment tool, an engineering wake model adapted in complex terrain and an advanced wind farm layout optimization algorithm, and it was found that the framework can provide a better layout than the original layout.…”