Analysis of Wind Energy data At Two Different Heights

Abstract

The purpose of the study is to identify the best fitted distribution(s) to provide insight into the annual maximum Wind speed and wind Direction. The data of wind speed in daily time series format were gathered from the Pakistan Meteorological Department Karachi for a 29-year period (1990 to 2018). The measurements were recorded at 10m and 50m heights. First, we verified the fundamental presumption, such as the randomness of the observed data series using Run test. Additionally, the parameters of two parametric distributions, the Frechet, Weibull, Loglogistic, and Gamma distributions, were estimated. For estimation of parameters maximum likelihood and Bayesian method of estimations are applied. The performance of the candidate distributions was evaluated using the Kolmogorov Smirnove test at a 5% threshold of significance.

Furthermore, probability density function graphs were utilized to validate the wind speed data's behavior and theoretical framework. The resultant estimates depicted that for Wind speed all four distributions are best fitted with both methods but for wind direction only Weibull and Loglogistic are best fitted with both estimation methods. From the estimated result it is clear that with both methods (MLE and BE) only WD and LLD are best fitted distribution for wind speed and wind direction 10m and 50m heights.