Effect of Real-Time Price on Short-Term Load Forecasting

Abstract

This paper presents a short-term load forecasting model using the back-propagation neural network (BPNN) model. The proposed model is based on data on loads and factors that directly affect electricity demand, such as temperature, humidity, load over time in the past, etc., collected from the electricity market ISO New England. In addition to the common factors, the article also considers a new factor: real-time price. The data used for training and forecasting are real-time data for three years from 2019 to 2021. The paper has shown that real-time price (RTP) significantly influences forecasting. The proof is that the Mean Absolute Percentage Error (MAPE) value of the predictive model without RTP data is 2.08%, and that of the model with RTP data is 1.44%. The paper also compares the performance of the training algorithms with each other to come up with an optimal algorithm compared to the proposed model. At the same time, the model is also applied to forecast a more extensive period, such as a week or a month, and has had positive results.