Improved Forecasting electricity demand using Taguchi experimental design

A good demand forecasting is an essential prerequisite to manage an electricity system. Over the years، many forecasting techniques of varying capabilities have been introduced. The precise extent of their influences، and the interactions between them، has never been fully clarified، although various forecasting factors have been explored in previous studies. Accordingly، this study adopts the Taguchi method to calibrate the controllable factors of a forecasting model. To this end an L9 (34) inner orthogonal array is constructed for the controllable factors of data period، horizon length، and number of observations required. An experimental design is then performed to establish the appropriate levels for each factor. At the same time، an L4 (23) outer orthogonal array is used to consider the inherited parameters of forecasting method as the noise factors of Taguchi method simultaneously. Through an eight steps pattern and using electricity data of Yazd province it is shown that the proposed model will improve quality of forecast and will reduce its expenses.