Abstract-Wind power is one of the widely used renewable resources and it is connected to power system steadily. In recent years, wind power is developed in the form of large-scale wind farm at offshore, which is composed of dozens or hundreds of wind turbines. In inner grid of wind farm, wind turbines are connected to each other through cable, and there are a wide variety of configurations depending on how to connect wind turbines. Due to difficult and expensive construction activity at sea, the problem to connect optimally wind turbines is very important. In order to solve the problem, this paper introduces a methodology based on the k-clustering algorithm, minimum spanning tree (MSP) algorithm and local search method. K-clustering is applied to divide wind turbines into k-groups, and MSP algorithm is used to link wind turbines in each group with the objective that total length of cables is minimized. Optimal configuration is determined by local search method which explores diverse combinations depending on the number of groups and the number of wind turbines in each group. The case studies show that the proposed methodology can be utilized usefully for designing inner grid of offshore wind farm.Index Terms-Offshore wind farm, inner grid, k-clustering algorithm, minimum spanning tree, local search method.
I. INTRODUCTIONWind power is a widely used renewable energy and is increased steadily. In particular, it is being developed as large-scale wind farm as the related technology and economics are advanced [1]. In order to gather more wind energy and to avoid environment problems and construction concerns such as securing the installation site, the recent wind power is developed at offshore. Compared with onshore wind farm, offshore wind farm (OWF) has some advantages, and it has also crucial disadvantages which are more expensive investment and maintenance costs caused by the difficulty of geographical access. Namely, since OWF operators have a considerable burden on investment, it is important to design optimal configuration for wind farm. Therefore, this paper investigates for optimal configuration of inner grid among components making up offshore wind farm. In the methodology, K-clustering algorithm divides wind turbines in inner grid into k-groups and then Minimum Spanning Tree (MSP) algorithm links wind turbines to each other based on the objective that total length of cables used for the connection of wind turbines in each group is minimized [2], [3]. Using local search method [4], the exploration is performed about diverse combinations made by the number of groups and the number of wind turbines belonging to each group. Alternatives generated by K-clustering and MST algorithms are evaluated in terms of total length of cables or total investment cost, and optimal configuration is finally selected. The rest of the paper is organized as follows; in Section II, typical compositions of OWF and layouts of inner grid are described. In Section III, the methodology including k-clustering, MST algorithms and local s...