Technical and financial analysis of 50MW wind farm at Gwadar, Balochistan

Statistical and computational models have been used worldwide by the researchers to forecast time series. This paper implements two techniques; Autoregressive Integrated Moving Average models and Neural Networks backpropagation algorithm to predict wind speed time series for Gwadar, a coastal city of Balochistan. Based on the wind speed forecast, power output has been estimated for a 50MW wind farm, choosing the appropriate turbine and optimum hub elevation for the installation of wind turbine. Furthermore, a financial study is presented of such a wind farm at Gwadar. For turbine selection, 317 turbines of various ratings and manufacturers were surveyed through software at different elevations - 60m, 80m, 100m and 120m. The results show that Neural Networks capture the trend in the wind speed time series more accurately than its statistical counterpart. Also, the optimum hub elevation for installing wind turbine at Gwadar is found to be 100m. Annual energy production under these circumstances is found to be 58.31 GWh/year.