EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

4189. Enhanced electricity prices forecasting from spillovers among commodities

Invited abstract in session TA-14: Enhanced statistical methods for energy challenges, stream Energy Markets.

Tuesday, 8:30-10:00
Room: 16 (building: 116)

Authors (first author is the speaker)

1. Giovanni Bonaccorsi
Management Engineering, Politecnico di Milano

Abstract

The deregulation of electricity markets has emphasized the importance of accurate daily electricity price forecasting (EPF) due to electricity's non-storability and the necessity for day-ahead bidding to align demand with production.
Accurate EPF is crucial for utilities and consumers, enabling effective bidding strategies, adjusting consumption to benefit from lower prices, and allowing producers to manage generation assets more efficiently.
The challenge in EPF arises from the complex dynamics of electricity prices, influenced by factors like weather conditions, business activity, and consumer habits, which introduce volatility, seasonality, and unexpected prices surges.
Despite the success of deep learning methods in the literature on EPF, recent global events such as the COVID-19 pandemic and the Russo-Ukrainian war have shown the need for more adaptable forecasting models.
This paper proposes an adaptive Long short-term memory (LSTM) model that responds to external shocks by monitoring a disruption signal based on the spillover of shocks across different consumer categories. This approach aims to enhance the accuracy of EPF by accounting for both direct and indirect effects of external shocks on electricity prices, using a dynamic threshold to update the forecasting strategy based on observed systemic changes.
We demonstrate the better forecast performance of our approach and provide an application with a model of energy storage arbitrage.

Keywords

Status: accepted

Back to the list of papers