Back to AI News
arxiv

Explainable AI workflow for reproducible solar PV power forecasting

This paper reviews machine learning methods for solar photovoltaic power forecasting and proposes a reproducible workflow combining automated model search with explainable AI. It tests the workflow on an openly available dataset for a 16.32 kWp plant in Porto, Portugal, forecasting one day ahead from a 4 km numerical weather prediction grid. Over a full test year, the best models reached 11.6% nRMSE of installed capacity, about half the error of smart persistence, while conformal calibration improved nominal 80% prediction interval coverage from 64.5% to 81.7%. The study also finds that input feature choice affected accuracy more than algorithm choice, and that SHAP rankings shifted when correlated inputs were removed.

AI Brief

It provides practical, reproducible guidance on model selection, uncertainty calibration, and interpreting explainable AI for solar forecasting.

arXiv:2611.25894v1 Announce Type: new Abstract: Machine learning (ML) is now a standard tool for forecasting the output of solar photovoltaic (PV) plants, yet published accuracy figures are often hard to compare and harder to trust. Many studies validate with random splits, report no reference forecast, and present feature attributions without saying what they can and cannot show. This paper reviews the main families of ML methods for PV power forecasting and resource assessment, from tree ensembles to deep and foundation models, and sets out a reproducible workflow that combines automated model search with explainable AI. The workflow is then tested on an openly available dataset: a 16.32 kWp plant in Porto, Portugal, forecast one day ahead from a 4 km numerical weather prediction grid. Over a full test year, the best models reach a normalised root mean square error (nRMSE) of 11.6% of installed capacity, about half the error of smart persistence. The choice of input features changed accuracy far more than the choice among strong algorithms, and shuffled cross-validation was mildly optimistic. After conformal calibration, the nominal 80% prediction intervals covered 81.7% of observations, against 64.5% before. SHAP rankings stayed stable when the model was retrained on different periods but shifted when correlated inputs were removed, so attributions should be read as properties of a model rather than of the plant. The paper closes with practical lessons and open problems, including deployment on edge hardware and the effect of a changing climate.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!