Short-Term Load Forecasting of 400kV Grid Substation Using R-Tool and Study of Influence of Ambient Temperature on the Forecasted Load

Abstract

Load forecasting helps to determine future electric load based on the historical data of the electrical system. Precise models for forecasting the electric power load are indispensable for the planning and operation of the utility. Load forecasting can also be used for load switching, demand-side management, identify and predict the energy consumption pattern to support an electric utility for future system operations. Many techniques can be used for load forecasting. The commonly used load forecasting technique is time series load forecasting. There are various factors which have an influence on the load and its forecast. Load forecasting using the time series ARIMA model for the forecast of the load is adopted in the present work. The Short-term is considered for the forecast i.e. one-day and one-week data along with the ambient temperature data as the independent variable. Forecast for four cases is carried out for determining the accuracy of the forecast. The error is expressed as Root Mean Square Error (RMSE). It is found that when the ambient temperature data is considered, the accuracy of the forecasted value increases as indicated by the reduction in RMSE. For one day load data without considering the ambient temperature it was found to be 6.009955 and when the ambient temperature of one day is considered it was 5.990357. Similarly, for one-week data without ambient temperature is was 4.68697 and with ambient temperature is was found to be 4.682843. The result obtained also prove that the exogenous data and number of data samples considered will play an important role in load forecasting accuracy using time series forecasting.