Short-Term Hybrid Probabilistic Forecasting Model for Electricity Market Prices

With the integration of new power production technologies and the growing focus on dispersed production, there has been a paradigm change in the electricity sector, mostly under a renewable and sustainable way. Consequentially, challenges for profitability as well as correct management of the electricity sector have increased its complexity. The use of forecasting tools that allow a real and robust approach makes it possible to improve system operation and thus minimizing costs associated with the activities of the electric sector. Hence, the forecasting approaches have an essential role in all stages of the electricity markets. In this paper, a hybrid probabilistic forecasting model (HPFM) was developed for short-term electricity market prices (EMP), combining Wavelet Transform (WT), hybrid particle swarm optimization (DEEPSO), Adaptive Neuro-Fuzzy Inference System (ANFIS), together with Monte Carlo Simulation (MCS). The proposed HPFM was tested and validated with real data from the Spanish and Pennsylvania-New Jersey-Maryland (PJM) markets, considering the next week ahead. The model was validated by comparing the results with previously published results using other methods.