Design optimization of offshore wind farms with multiple types of wind turbines. (1st November 2017)

• Record Type: Journal Article
• Authors: Feng, Ju; Shen, Wen Zhong
• Is Part Of: Applied energy. Volume 205(2017)
• Page Start: 1283
• Page End: 1297
• Publication Date: 2017-11-01
• Subjects: Offshore wind farm -- Non-uniform wind farm -- Layout optimization -- Design optimization -- Random search algorithm -- Levelized cost of energy; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2017.08.107
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: Non-uniform wind farm's design includes number, types and locations of turbines. The mixed-integer-discrete-continuous optimization is solved by random search. The extended random search outperforms the mixed-discrete PSO algorithm. Whether non-uniform designs better depends on the capital costs comparison. Better treat wind farms as non-uniform and consider multiple types of turbines. Abstract: Most studies on offshore wind farm design assume a uniform wind farm, which consists of an identical type of wind turbines. In order to further reduce the cost of energy, we investigate the design of non-uniform offshore wind farms, i.e., wind farms with multiple types of wind turbines and hub-heights. Given a set of different types of wind turbines with a different default hub height for each type, we can specify the design of a wind farm by the types of turbines, number of turbines for each type, and turbine locations. We consider the optimization of such design to minimize the levelized cost of energy, which is calculated using a capital cost model that covers the turbine cost and the balance of plant cost. An empirical wind turbine design cost and scaling model is utilized to model the cost of turbines with different sizes. Constraints on wind farm boundary, wind turbine proximity and total capacity are also included. We solve the problem with a newly developed extended random search algorithm and tested it in a realistic design optimization problem based on the Horns Rev 1 offshore wind farm in Denmark. The optimized non-uniform designs are compared with their uniform counterparts. We find that a non-uniform design can achieve a lower levelized cost of energy than its uniform counterparts, when the capital cost per MW is slightly lower for the smaller size turbine. Comparison with the mixed-discrete particle swarm optimization algorithm is also carried out for a non-uniform wind farm design problem with a fixed number of turbines, which shows the effectiveness and superiority of the proposed algorithm. Finally, the advantages and possible disadvantages of non-uniform design are also identified and discussed.