Probabilistic Forecasting of Temporal Trajectories of Regional Power Production - Part 1: Wind

TL;DR

This paper introduces a probabilistic modeling approach for forecasting wind power production at multiple time scales, utilizing weather forecasts and Bayesian inference to improve accuracy and calibration.

Contribution

It develops a multivariate Gaussian prediction model with temporal error correlation for wind power, incorporating weather forecasts and Bayesian methods, which is novel in this context.

Findings

The model provides calibrated, skillful forecasts for German wind power production.

Incorporating temporal error correlation improves forecast accuracy.

Different correlation structures are compared, showing their impact on prediction quality.

Abstract

Renewable energy sources provide a constantly increasing contribution to the total energy production worldwide. However, the power generation from these sources is highly variable due to their dependence on meteorological conditions. Accurate forecasts for the production at various temporal and spatial scales are thus needed for an efficiently operating electricity market. In this article - part 1 - we propose fully probabilistic prediction models for spatially aggregated wind power production at an hourly time scale with lead times up to several days using weather forecasts from numerical weather prediction systems as covariates. After an appropriate cubic transformation of the power production, we build up a multivariate Gaussian prediction model under a Bayesian inference framework which incorporates the temporal error correlation. In an application to predict wind production in Germany, the method provides calibrated and skillful forecasts. Comparison is made between several formulations of the correlation structure.