Accurate and reliable forecasting of photovoltaic (PV) power production is crucial for grid operations, electricity markets, and energy planning, as solar systems now contribute a significant share of the electricity supply in many countries. PV power forecasts are often generated by converting forecasts of relevant weather variables to power forecasts via a model chain. The use of ensemble simulations from numerical weather prediction models results in probabilistic PV forecasts in the form of a forecast ensemble. However, weather forecasts often exhibit systematic errors that propagate through the model chain, leading to biased and/or uncalibrated PV power forecasts. These deficiencies can be mitigated by statistical post-processing. Using PV production data and corresponding short-term PV power ensemble forecasts at seven utility-scale PV plants in Hungary, we systematically evaluate and compare seven state-of-the-art methods for post-processing PV power forecasts. These include both parametric and non-parametric techniques, as well as statistical and machine learning-based approaches. Our results show that compared to the raw PV power ensemble, any form of statistical post-processing significantly improves the predictive performance reducing the mean continuous ranked probability score (CRPS) by 11.1–14.7%. Non-parametric methods outperform parametric models, with advanced nonlinear quantile regression models showing the best results. Furthermore, machine learning-based approaches surpass their traditional statistical counterparts by around 2 percentage points in terms of the improvement in mean CRPS over the raw forecasts. • Seven statistical post-processing methods are compared for ensemble PV power forecasts. • Ensemble weather forecasts are converted to PV power using physical model chains. • All methods significantly improve the reliability of the ensemble forecasts. • Statistical calibration reduces CRPS by 11.1–14.7% compared to the raw ensemble. • Quantile regression neural networks outperform distributional regression methods.
Mayer et al. (Thu,) studied this question.