A Comparison of Aggregate and Multi-Region Load Forecasting Models in Saskatchewan
Power systems in large geographic environments experience diverse weather phenomena. Due to spatial separation and economic diversification, load centres will exhibit differing electrical demand. This diversity of weather and electric loads proves challenging for load forecasters using a single aggregate model. In such systems, aggregate response of these load centres cannot be properly analyzed by a single load-weather model. Instead, the aggregate demand is best explained through multi-region modeling. This thesis describes the load and weather diversity within the control area of an electric utility in the province of Saskatchewan. A Similar Day model is contrasted against Artificial Neural Network (ANN) models based on both an aggregate and two multi-region systems. Results confirm the superior performance of the proposed multi-region load forecasting systems as compared to the two aggregate load forecasting models.