Spatial Wind Speed Forecasting Using Artificial Neural Networks: Spatial Wind Speed Forecasting Using Artificial Neural Networks

Saif ur Rehman; Muhammad Shoaib; Imran Siddiqui; Shamim Khan; Syed Zeeshan Abbas

doi:10.46660/ijeeg.v11i4.297

Authors

Saif ur Rehman Department of Physics, Federal Urdu University of Arts, S&T, Karachi, Pakistan
Muhammad Shoaib Department of Physics, Federal Urdu University of Arts, S&T, Karachi, Pakistan
Imran Siddiqui Department of Physics
Shamim Khan Department of Physics Islamia College Peshawar, Peshawar Jamrod Road, University Campus Peshawar, Khyber Pakhtunkhwa, 25120, Pakistan
Syed Zeeshan Abbas Department of Physics, University of Karachi, Karachi-75270, Pakistan

DOI:

https://doi.org/10.46660/ijeeg.v11i4.297

Abstract

Spatial interpolation is a commonly used technique to simulate wind speeds in areas which are devoid of
such measuring devices. In this paper authors examine the applicability and efficiency of Artificial-Neural- Network
(ANN) formalism aimed at interpolating wind speeds in space domain. Additionally, the effect of the correlation
between the wind speed at target site and its correlated neighboring site is also examined in the present paper. Hourly
wind speed data set comprising of wind speeds recorded from April 2016 to August 2018 provided by Energy Sector
Management Assistance Program of World Bank is used for the study. The study is supported by including four
different wind speed measuring stations in Pakistan, namely, Tando Ghulam Ali, Umer Kot, Sujawal and Sanghar. Best
estimates from ANN model are obtained for Tando Ghulam Ali (MAPE= 7.37%) and worst estimates are observed for
Sanghar site (MAPE= 10.61%)

Spatial Wind Speed Forecasting Using Artificial Neural Networks