Volume 44: Energy Transitions toward Carbon Neutrality: Part VII

Development of Forecasting Models for German Energy Prices and Carbon Intensity as a Decision Basis for Energy Flexibility Measures Jonas Wendt, Arthur Stobert, Jonas Valentin Schulte, Matthias Weigold

Abstract

Energy flexibility measures play a crucial role in the achievement of carbon neutrality. Renewable energy sources can only be prioritized if energy demand can adapt to the supply. For the targeted use of such measures, a better understanding of the energy markets and the affected systems is essential.
The availability of sustainable energy sources is highly dependent on fluctuating environmental conditions like solar radiation or wind speed. Combined with changing energy demand, this leads to volatility in energy prices and carbon intensity. To react to these fluctuations at an early stage, trends in electricity prices and carbon intensities are urgently needed in addition to weather forecasts, which are already available across the board. This gap shall be addressed by this publication, which presents a machine learning based tool to forecast electricity prices and carbon intensities beyond the German day-ahead market for the following 48 hours. Publicly available market data from the past five years was used to train the machine learning model, which achieved a mean absolute error (MAE) of 20.6 €/MWh for day-ahead energy prices during the first half of 2023. The tool forecasts carbon intensities with a MAE of 47.7 gCO2eq/kWh for the same period. The presented forecasting tool enables planning of energy-flexible operating strategies at an early stage and their implementation at industrial sites.
An air conditioning system as an exemplary industrial use case is used to demonstrate the relevance of the presented forecasting model in the context of energy flexibility. The utilization of the forecasting tool and the development of energy-flexible operating strategies resulted in potential savings of 12.33 % in operating costs and 9.94 % in carbon emissions.

Keywords forecasting models, energy trends, energy flexibility, carbon intensity, climate neutrality