Short-term electricity load forecasting with Time Series Analysis

Abstract

Load forecasting plays a fundamental role throughout all segments of system health management for utility companies, including, but not limited to, financial planning, rate design, power system operation, and electrical grid maintenance. Recently, due to the deployment of Smart Grid technologies, utility companies' ability to create accurate forecasts is of even greater importance, especially in consideration of demand response programs, charging of plug-in electric vehicles, and use of distributed energy resources. In this paper, several time series Autoregressive Integrated Moving Average (ARIMA) and Seasonal Autoregressive Integrated Moving Average (SARIMA) models will be introduced for the purpose of generating forecasts of short-term load demand, at an hourly interval, based on data made available by the Electric Reliability Council of Texas (ERCOT). The case study which expands on the short-term data analyzed in [1] includes over 100,000 data points representing electricity load in Texas recorded over the past 14 years.