Formulation and analysis of grid and coordinate models for planning wind farm layouts

In this paper, a comprehensive study of the effectiveness of the classical grid and coordinate models (CMs) in producing the optimal wind farm layout is conducted based on theoretical analyses and computational experiments. The wind farm layout planning with the grid model (GM) and CM is formulated as a combinatorial and a continuous optimization problem separately. Theoretical analyses prove that it is more complicated to solve CM than GM if the solution space of two models is searched exhaustively. In computational studies, the impact of advanced heuristic search methods on generating optimal wind farm layouts with GM and CM is analyzed. First, two models are solved with the multi-swarm optimization (MSO) algorithm, and CM, in general, produces better layouts, because swarm intelligence is inherently continuous and the flexibility of CM. To further evaluate the importance of selecting an appropriate heuristic search algorithm, the random key genetic algorithm (RKGA) is introduced to compare with MSO in solving GM. Results show that GM produces much better wind farm layouts with RKGA, which is inherently combinatorial. Computational results demonstrate that it is important to match the inherent suitability of heuristic search algorithms with the type of the layout planning models in the wind farm layout optimization.