Secondary Reserve Marginal Band Price Prediction with Classical and Machine Learning Based Techniques

As a consequence of the continuous growth being experienced by renewable energy systems, the role of the different ancillary services is becoming essential for the reliable operation of the electric system. This paper develops a methodology for estimating the secondary reserve marginal band price in the Iberian electricity market using four forecasting techniques: two classical models (ARIMAX and SARIMAX) and two machine learning models (Random Forest and Support Vector Regression). The methodology involves a Pearson correlation analysis and a Sequential Forward Selection algorithm to select the relevant model inputs and exogenous variables. A statistical data analysis is conducted to examine the temporal characteristics of the target variable and a data-preprocessing is performed for the proper implementation of the models. A grid search and a sequential division cross-validation are applied to determine the optimal parameters of the models. The performance is evaluated using three statistical metrics. The results show that the Random Forest model outperforms the other models, achieving the lowest MAE (2.40 €/MW) and RMSE (3.40 €/MW) values.