Forecasting Electricity Consumption in Sulaimani Province by Using Spectral Analysis Approach

Abstract

In the modern world, with the extensive use of electricity in homes and businesses, studying and forecasting electrical consumption is crucial for its management and distribution. This research aims to forecast the electricity consumption in Sulaimani province using spectral analysis on time series data. For this purpose, hourly data within the time range of July to September 2022 has been utilized, and a new model has been proposed, consisting of a sinusoidal explanatory variable and a nonlinear quadratic variable. In this model, the most significant periodic cycles present in the time series under study were identified by employing the periodogram and Fourier expansion method, capable of identifying hidden cycles in the data. Subsequently, each of these periodic components was added to the model based on the principle of parsimony. The proposed composite model includes a quadratic component and two harmonic components with 24 and 12-hour periods. This model demonstrates notable efficiency and accuracy as it outperforms the classical Autoregressive Moving Average (ARMA) model based on the MSE=0.0009 rate and MAE=23.819 when compared