Natural gas demand is a crucial factor for predicting natural gas prices and thus has a direct influence on the power system. However, existing methods face challenges in assessing the impact of shocks, such as the outbreak of the Russian-Ukrainian war. In this context, we apply deep neural network-based Granger causality to identify important drivers of natural gas demand. Furthermore, the resulting dependencies are used to construct a counterfactual case without the outbreak of the war, providing a quantifiable estimate of the overall effect of the shock on various German energy sectors. The code and dataset are available at https://github.com/bonaldli/CausalEnergy.

Natural gas is the most important energy source in Italy: it fuels thermoelectric power plants, industrial facilities and domestic heating. Gas demand forecasting is a critical task for any energy provider as it impacts on pipe reservation and stock planning. In this paper, the one-day-ahead forecasting of Italian daily residential gas demand is studied. Five predictors are developed and compared: Ridge Regression, Gaussian Process, k-Nearest Neighbour, Artificial Neural Network, and Torus Model. Preprocessing and feature selection are also discussed in detail. Concerning the prediction error, a theoretical bound on the best achievable root mean square error is worked out assuming ideal conditions, except for the inaccuracy of meteorological temperature forecasts, whose effects are properly propagated. The best predictors, namely the Artificial Neural Network and the Gaussian Process, achieve an RMSE which is twice the performance limit, suggesting that precise predictions of residential gas demand can be achieved at country level.

Forecasting natural gas demand is a key problem for energy providers, as it allows for efficient pipe reservation and power plant allocation, and enables effective price forecasting. We propose a study of Italian gas demand, with particular focus on industrial and thermoelectric components. To the best of our knowledge, this is the first work about these topics. After a preliminary discussion on the characteristics of gas demand, we apply several statistical learning models to perform day-ahead forecasting, including regularized linear models, random forest, support vector regression and neural networks. Moreover, we introduce four simple ensemble models and we compare their performance with the one of basic forecasters. The out-of-sample Mean Absolute Error (MAE) achieved on 2017 by our best ensemble model is 5.16 Millions of Standard Cubic Meters (MSCM), lower than 9.57 MSCM obtained by the predictions issued by SNAM, the Italian Transmission System Operator (TSO).