Application of ARIMA and 2D-CNNs Using Recurrence Plots for Medium-Term Load Forecasting

Load forecasting is beneficial for planning, operation, and control actions of power systems. Historical documented real-time load data can be utilized to predict the future load on power systems. In this regard the advanced artificial intelligence (AI) techniques for data analysis can be effective for medium-term load forecasting. The accuracy and computational cost of the model are key indicators for effectively forecasting power loads. In this paper, a recurrence plot (RP) time encoding, and 2D-CNN model is applied to a real-time Turkey load consumption dataset for making prediction and is compared with the autoregressive integrated moving average (ARIMA) model for the same data to demonstrate their effectiveness for medium term load forecasting.