Load forecasting via detrending and deseasoning

Abstract

Load forecasting is a term usually applied to describe a process of estimation or prediction of future energy demand for a certain distribution grid or part of a grid. Large number of different methods and techniques used for load forecasting have been developed in the past and new and improved methods are regularly being reported in research literature. This paper describes one of traditional load forecasting approaches based on autoregressive moving average (ARMA) modelling of load demand time-series (TS). However, it reconsiders each step in this process and proposes some new procedures to improve and clarify the whole method. Effectives of described approach is demonstrated using energy consumption measurements recently recorded at substations in central London area.