Development of software algorithm for energy modeling and forecasting

Abstract

"The problem of repetitive calculations involved in the comparative analysis of forecasted energy loads to determine the most applicable model out of a variety has been reduced by the development of a forecasting software algorithm done in this work. Three models and two test tools were catered for by the developed software at this preliminary stage. They are the linear, compound growth and cubic models while the test tools are the rank of correlation coefficient (ROCC) and the mean absolute percentage error(MAPE).The numerical modelling and forecasting procedures were presented and an algorithm flow chart was subsequently developed for the software. The programming language selected for the design of the software is MATLABJ because of its versatility in mathematical applications. On the completion of thesoftware, it was applied to an existing energy consumption case study available in literature to ascertain its reliability. The results obtained were in the range of the estimations found in literature. "