Accurate prediction functions for turbine-scale wind energy resources from low-elevation measured data for Bangladesh

Abstract

This paper presents a highly accurate statistical modeling method for selecting prediction functions for turbine-scale wind energy resources in Bangladesh, which are not yet extensively covered in existing literature. The proposed resource modeling encompasses key turbine parameters, including turbine power, energy pattern factors, and wind energy outputs. The study, surveyed by the United Nations (UN), identifies prospective areas in mainland and coastal belt regions with the capacity to support commercial wind energy conversion systems for modeling purposes. As Bangladesh's current meteorological measurement facilities primarily collect low-elevation data (approximately 10 meters), the study employs prediction models and projection laws to estimate energy resources suitable for turbines of low-to-medium ratings (80 meters, less than 1 MW). The time-series probability distribution of available wind resources is analyzed in these potential regions, utilizing wind velocity prediction functions such as Weibull, Rayleigh, and Gumbel distributions. Their performances are compared against established statistical standards. Weibull factors are derived using graphical least squares (GLS) and modified maximum likelihood (MML) methods, and validated against parameter values reported in the literature. To enhance the analysis's coverage and accuracy, the Weibull function is expanded by incorporating the effect of output power ratio into its probability distribution. Wind power density (WPD) trends are confirmed through energy pattern factors, and a portable wind system model is employed to estimate the actual energy output at prospective locations, thereby increasing the comprehensiveness of the energy data modeling process.