| dc.description.abstract | In recent years the global installed capacity of offshore wind has increased rapidly, due to the world’s green electricity demand. Increasing developments in the offshore wind sector has led to fewer possible locations for new wind farms, forcing the developers to move further and further away from shore. This shift from near-shore to far-shore wind farm locations, increases the complexity and costs of executing operations and maintenance, which can account for 25% of the production cost of power. The vessels and helicopters that are used to execute preventive and corrective maintenance activities are expensive, and a crane vessel can easily cost USD 40 000 per day. The potential savings in determining an optimal fleet size and mix for the execution of maintenance activities on an offshore wind farm are therefore considerable. Using a joint fleet for more than one wind farm, is a way of achieving savings in order to obtain cost-efficient projects. However, uncertain factors, such as turbine failures requiring corrective maintenance, vessel spot rates, electricity prices and weather conditions limiting the accessibility of the vessels, raise the need for decision support tools. In this thesis we investigate the possibility of using operations research to determine an optimal fleet size and mix for one or several offshore wind farms. The decisions to be made are how many vessels to acquire or rent in order to meet a given maintenance schedule, in addition to determining whether offshore station concepts are economically viable. Strategic decision support tools in terms of both a deterministic and a stochastic optimisation model are developed and will be presented. Based on different scenarios including turbine failures, vessel and helicopter spot rates, electricity prices and weather conditions, the stochastic model determines the optimal fleet size and mix that should be used to execute maintenance operations on one or several offshore wind farms. A computational study proves that the stochastic model is able to solve problems for real world wind farms with more than 400 wind turbines. Further, the value of the stochastic solution and the expected value of perfect information suggest that the stochastic model gives solutions that are fairly well hedged against possible future outcomes, and returns solutions that perform significantly better than the solutions from the deterministic model. In addition, the stochastic model is shown to have a great economical applicability, in terms of determining the willingness to pay for additional wave capacity of the vessels, the possible savings of using offshore station concepts and the potential savings in using a joint fleet on several offshore wind farms. The stochastic optimization model addressing the fleet size and mix problem for offshore wind is the first of its kind, and this thesis has proven that the model has a real world value in terms of being a great strategic decision support tool taking into account the inherent uncertainties of the problem. The model does, however, not consider the logistics of spare parts or the tactical day-to-day utilisation of the given fleet, thus further work is suggested on these issues. | nb_NO |
