Short Term Load Forecasting for Transnet Port in East London South Africa

Energy consumption patterns at Transnet in East London South Africa varies stochastically. This is because of the transient weather patterns that exists at the harbour. It has become imperative to manage this load to save electricity costs and for future infrastructure development. Supply of electricity to the port requires accurate short-term load forecasting for efficient load management. Although researchers have recently proposed Artificial Neural Networks for short-term load prediction, many studies have not considered the quickly changing weather patterns that exists at the port. This study proposes a system architecture firstly with open- loop training using real load and weather data, and then a closed-loop network to predict the load as its feedback data. Nonlinear autoregressive exogenous model for load prediction was developed using mean squared error as a performance metric. To show the efficacy of the proposed model for load forecasting, the adaptive neuro-fuzzy inference system was used with the same data for short-term predictions. The minimum mean squared errors obtained for both NARX and ANFIS models were 0.0014156 and 0.0497, indicating that the NARX model is more accurate during the forecast of departmental loads. The Load forecast model was developed to implement management plans for internal load at the Transnet Port in East London South Africa.