Comparative analysis of wind speed probability distributions for wind energy potential in Southwestern Nigeria

Abstract

This study investigates the wind energy potential of Lagos and Abeokuta in southwestern Nigeria using advanced probability distribution models. Conventional unimodal distributions such as Weibull, Gamma, Normal and Lognormal often fail to capture the bimodal and seasonal nature of wind speed in the region. To address this limitation, a Modified Bimodal Mixture Weibull Model (MBMW) incorporating a logistic sigmoid transition was introduced. The model adapts dynamically to changes in wind regimes, providing improved flexibility over the standard Bimodal Mixture Weibull (BMW). Wind speed data spanning 21 years (1995–2015) at 10 m height obtained from the Nigerian Meteorological Agency were analysed. Both BMW and MBMW models were fitted to the data using Maximum Likelihood Estimation (MLE), and their performance was assessed with Root Mean Square Error, Kolmogorov-Smirnov statistic, chi-square test, and information criteria. Results show that the MBMW consistently outperformed the BMW, producing lower errors and better alignment with empirical distributions. Energy estimates revealed significant differences between the two locations. Lagos, influenced by coastal wind systems, recorded an average power density of 566.61 W/m² and an annual energy yield of 4.96 MWh/m², while Abeokuta produced only 57.18 W/m² and 0.50 MWh/m². These findings revealed more energy potentials, which other studies and models could not unveil. It also confirmed Lagos as a viable site for large-scale wind energy exploitation, whereas Abeokuta is more suited to small-scale applications. The proposed MBMW framework demonstrates strong potential for improving wind energy assessment in complex environments and offers valuable insights for renewable energy planning in Nigeria.