Comparison of different techniques for offshore wind farm reliability assessment

The increase in wind power capacity installed worldwide has resulted in the necessity to include wind farms in power system reliability assessment. Especially, the installation of large offshore units connected directly to the transmission system has created the necessity of accounting for new aspects in the analysis. Various studies have been performed on the topic and many relevant factors of influence have been highlighted, but a complete assembly of all factors together has not been considered yet. This paper focuses on the comparison of different approaches for frequency and duration reliability assessments of offshore wind farms. The objective of the paper is to evaluate the most efficient technique, in order to represent a wind farm in the most realistic way and to perform a broad range of studies. Different aspects that influence the analysis are included in the paper, such as wind speed variability and randomness, system components, (wind turbines, internal grid cables and connectors to shore) failures and installation layouts. Two main approaches have been considered in this paper: One based on a sequential Monte Carlo simulation and one based on analytical methods with frequency and duration (F&D) analysis. The latter uses a probabilistic approach to define mathematical models of the system elements for calculating the output values, whereas the first model defines randomly the behaviour of each element for a number of sampled years and final results are evaluated as mean values. The computed results refer to the yearly output power and the capacity factor of the wind farm: These values, together with the required computation time and feasible future studies, are used in order to compare the efficiency and strengths of the two techniques. Furthermore, the possibility of distinguishing and evaluating effects of extreme wind conditions on the generation has also been used as a criterion in the comparison.