The Role of Weather Predictions in Electricity Price Forecasting Beyond the Day-Ahead Horizon

Abstract

The availability of historical data related to electricity day-ahead prices and to the underlying price formation process is limited. In addition, the electricity market in Europe is facing a rapid transformation, which limits the representativeness of older observations for predictive purposes. On the other hand, machine learning methods that gained traction also in the domain of electricity price forecasting typically require large amounts of data. This study analyses the effectiveness of encoding well-established domain knowledge to mitigate the need for large training datasets. The domain knowledge is incorporated by imposing a structure on the price forecasting problem; the resulting accuracy gains are quantified in an experiment. Compared to an "unstructured" purely statistical model, it is shown that introducing intermediate quantity forecasts of load, renewable infeed, and cross-border exchange, paired with the estimation of supply curves, can result in a NRMSE reduction by 0.1 during daytime hours. The statistically most significant improvements are achieved in the first day of the forecasting horizon when a purely statistical model is combined with structured models. Finally, results are evaluated and interpreted with regard to the dynamic market conditions observed in Europe during the experiment period (1 st October 2022 -30 th April 2023), highlighting the adaptive nature of models that are trained on shorter timescales. 4CONCLUSIONS .

Distributional neural networks for electricity price forecasting

“…Lastly, we use the Diebold-Mariano (DM) test that measures the statistical significance of the difference between the accuracy of the forecasts of two models (here, we use A and B to discern between them) [15,55,68]. Let us denote the vector of errors (here, the CRPS scores) of model…”

“…These conclusions might not be easy to derive based on Figure 2, however see Ziel and Weron [23], Sgarlato and Ziel [55] for more insights. was first investigated by Johnson [52] as a transformation of the normal distribution.…”

Day-Ahead Electricity Price Forecasting in the Contemporary Italian Market

“…The only widely used variable that we did not consider here is the weather variable, primarily because weather forecasts are typically available to companies under payment and are not easily accessible to researchers. However, recent research in [23] clarifies that short-term EPF models, including ours, usually incorporate weather variables indirectly. For instance, instead of directly considering solar irradiance, they might use photovoltaic generation, or replace wind speed with wind generation, and so on.…”

Forecasting the Occurrence of Electricity Price Spikes: A Statistical-Economic Investigation Study

“…However, in day-ahead markets, price spikes have become more frequent in recent years [11], too. This is likely attributed to the increased penetration of variable generation resources that carry an inherent production uncertainty [12,13]. More frequent and severe spikes in both types of market structures require market participants to be better prepared to hedge against the increasing volatility of electricity prices.…”