The sporadic and unstable nature of wind speed renders it very difficult to predict accurately to serve various decisions, such as safety in the air traffic flow and reliable power generation system. In this study we assessed the autoregressive integrated moving average (ARIMA) and artificial neural network (ANN) models on the wind speed time series problem. Data on wind speed and minimum and maximum temperatures were evaluated. Wind speed was established to follow a time series that fluctuated around ARIMA (0,1,1) and ARIMA (1,1,1). The optimal ANN model was established at 10 hidden neurons. The performance indices considered all indicated that the ANN wind speed model was superior to the ARIMA model. Wind speed prediction accuracy can be improved to secure the safety of air traffic flow as well support the implementation of a reliable and secure power generation system at the airport.
- artificial neural network
- autoregressive integrated moving average
- Wind speed
ASJC Scopus subject areas
- Modelling and Simulation
- Computer Graphics and Computer-Aided Design