Innovative hybrid NARX-RNN model for predicting wind speed to harness wind power in Pakistan

The surge in wind energy utilization, particularly within regions characterized by consistent wind patterns, has attracted substantial attention. Pakistan has several wind corridors with potential energy extraction, primarily located in Balochistan and Sindh. This study aims to predict wind speeds and assess wind energy feasibility across Pakistan using a novel hybrid NARX-RNN model. We analyzed wind data from five distinct zones over seven years to identify the most promising site for wind energy. Our hybrid model demonstrated superior performance, achieving a 20 % reduction in RMSE and a 25 % reduction in MAPE compared to recent leading-edge methods. The model maintained robust performance across various wind speed ranges, with notable improvements in high wind speed conditions. We identified Sujawal as the most suitable location due to its consistently high wind speeds, minimal directional variation, and strong economic profile. This study provides valuable insights for sustainable wind energy production in Pakistan, offering a more comprehensive and accurate prediction model crucial for effective wind energy planning and operation.