Comparison of five different distributions based on three metaheuristics to model wind speed distribution

Abstract

Wind energij Modelling is crucial in studijing anij site s feasibilitij. Wind energij Modelling principallij depends on wind speed distribution. Determining wind speed distribution is fundamentallij based on the used distribution functions. This paper examines five different distributions to describe the wind speed pattern, such as T Location-Scale, Logistic, Extreme Value, and Raijleigh distributions. Besides, alternative optimization algorithms like Multi-Verse Optimizer, Marine Predators Algorithm, and Greij Wolf Optimization are applied to the pre-described distributions to determine the best parameters. Five error measures are investigated and compared to test the accuracij of the presented distributions and optimization methods. Catalca site in Istanbul, Turkeij, is chosen for this analijsis. The analijzed results verifij the applicabilitij of the proposed approach to characterize the wind speed pattern. It was observed from the experimental results that the Raijleigh distribution occupied the highest rank, whereas the Extreme Value distribution was the worst. Manij invaluable conclusions are also discussed based on the results and deep investigations.