Multi-modal optimization of offshore wind farm collection system topology based on nearest better most attractive particle swarm optimization

Abstract

The offshore wind farm collection system plays a crucial role in the development of offshore wind farms, reducing their significant construction costs as a key area of research. In complex and uncertain marine environments, solving strategies based on unique global optimization often fail to meet engineering design needs. This paper proposes an innovative solution for multi-modal optimization of the offshore wind farm collection system topology that provides designers with more decision-making freedom. In terms of the goal of minimizing cost, this research comprehensively considers capital cable costs, cable installation, and energy loss, taking into account the uncertainty of wind conditions. Additionally, the research combines the Weibull distribution model and the Jensen wake model to calculate the wind distribution. In terms of the optimization algorithm, a novel multi-modal algorithm, named Nearest Better Most Attractive Particle Swarm Optimization (NBMA-PSO), is proposed, in which the Prim algorithm is employed for population initialization, subpopulation grouping is achieved through a normalized difference representation and population balance strategy, and the iterative optimization is implemented through a two-stage PSO algorithm. Case analysis shows that compared with existing optimization algorithms, the proposed NBMA-PSO algorithm has better solving efficiency and stability and can effectively obtain multi-modal solutions. The proposed NBMA-PSO algorithm efficiently balances exploration and exploitation in offshore wind farm collection system optimization, demonstrating its capability to generate multiple high-quality solutions while reducing total costs by up to 4.1 % compared to traditional counterparts.