An AI-Based Decision-Support Pipeline for Day-Ahead Photovoltaic Forecasting

Fariba Dehghan, Sebastian Stein, Vahid Yazdanpanah, Stephanie Gauthier, Masood Nazari
Proceedings of the Fourth UK AI Conference 2026, PMLR 348:52-67, 2026.

Abstract

Reliable photovoltaic (PV) forecasts can support low-carbon energy systems, but deployed sites may have only short and incomplete records. Physical and hybrid methods can be sensitive to weather inputs, calibration, and timestamp-alignment, while individual machine learning models may capture different parts of the forecasting problem. We study hourly day-ahead PV forecasting at a United Kingdom charging station using one year of inverter measurements, with 9.25% of hours missing. The pipeline checks timestamp-alignment, derives solar and clearness features, adds short-term weather context, and combines five complementary models using non-negative least squares stacking, with the combination fitted only on validation observations. We compare against smart persistence, a weather-scaled baseline that carries the previous day’s PV behaviour forward using target-day irradiance. With retrospective weather, the combined model reduces daylight normalised root mean square error (RMSE) by 31.2% under random day-fold evaluation and by 2.9% under rolling-origin evaluation, although the latter improvement is not robust across days. It also improves by 3.0% over the single model selected from validation performance. Replacing retrospective weather with a public product sampled at a constant 24-hour lead increases daylight RMSE by 13.1% and 4.2% under the two protocols, while retaining positive skill over smart persistence.

Cite this Paper

BibTeX
@InProceedings{pmlr-v348-dehghan26a, title = {An AI-Based Decision-Support Pipeline for Day-Ahead Photovoltaic Forecasting}, author = {Dehghan, Fariba and Stein, Sebastian and Yazdanpanah, Vahid and Gauthier, Stephanie and Nazari, Masood}, booktitle = {Proceedings of the Fourth UK AI Conference 2026}, pages = {52--67}, year = {2026}, editor = {Benford, Alistair and Büyükateş, Baturalp and Cabrera, Christian and Kiden, Sarah and Salili-James, Arianna and Zakka, Vincent and Zhou, Feng}, volume = {348}, series = {Proceedings of Machine Learning Research}, month = {29--30 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v348/main/assets/dehghan26a/dehghan26a.pdf}, url = {https://proceedings.mlr.press/v348/dehghan26a.html}, abstract = {Reliable photovoltaic (PV) forecasts can support low-carbon energy systems, but deployed sites may have only short and incomplete records. Physical and hybrid methods can be sensitive to weather inputs, calibration, and timestamp-alignment, while individual machine learning models may capture different parts of the forecasting problem. We study hourly day-ahead PV forecasting at a United Kingdom charging station using one year of inverter measurements, with 9.25% of hours missing. The pipeline checks timestamp-alignment, derives solar and clearness features, adds short-term weather context, and combines five complementary models using non-negative least squares stacking, with the combination fitted only on validation observations. We compare against smart persistence, a weather-scaled baseline that carries the previous day’s PV behaviour forward using target-day irradiance. With retrospective weather, the combined model reduces daylight normalised root mean square error (RMSE) by 31.2% under random day-fold evaluation and by 2.9% under rolling-origin evaluation, although the latter improvement is not robust across days. It also improves by 3.0% over the single model selected from validation performance. Replacing retrospective weather with a public product sampled at a constant 24-hour lead increases daylight RMSE by 13.1% and 4.2% under the two protocols, while retaining positive skill over smart persistence.} }
Endnote
%0 Conference Paper %T An AI-Based Decision-Support Pipeline for Day-Ahead Photovoltaic Forecasting %A Fariba Dehghan %A Sebastian Stein %A Vahid Yazdanpanah %A Stephanie Gauthier %A Masood Nazari %B Proceedings of the Fourth UK AI Conference 2026 %C Proceedings of Machine Learning Research %D 2026 %E Alistair Benford %E Baturalp Büyükateş %E Christian Cabrera %E Sarah Kiden %E Arianna Salili-James %E Vincent Zakka %E Feng Zhou %F pmlr-v348-dehghan26a %I PMLR %P 52--67 %U https://proceedings.mlr.press/v348/dehghan26a.html %V 348 %X Reliable photovoltaic (PV) forecasts can support low-carbon energy systems, but deployed sites may have only short and incomplete records. Physical and hybrid methods can be sensitive to weather inputs, calibration, and timestamp-alignment, while individual machine learning models may capture different parts of the forecasting problem. We study hourly day-ahead PV forecasting at a United Kingdom charging station using one year of inverter measurements, with 9.25% of hours missing. The pipeline checks timestamp-alignment, derives solar and clearness features, adds short-term weather context, and combines five complementary models using non-negative least squares stacking, with the combination fitted only on validation observations. We compare against smart persistence, a weather-scaled baseline that carries the previous day’s PV behaviour forward using target-day irradiance. With retrospective weather, the combined model reduces daylight normalised root mean square error (RMSE) by 31.2% under random day-fold evaluation and by 2.9% under rolling-origin evaluation, although the latter improvement is not robust across days. It also improves by 3.0% over the single model selected from validation performance. Replacing retrospective weather with a public product sampled at a constant 24-hour lead increases daylight RMSE by 13.1% and 4.2% under the two protocols, while retaining positive skill over smart persistence.
APA
Dehghan, F., Stein, S., Yazdanpanah, V., Gauthier, S. & Nazari, M.. (2026). An AI-Based Decision-Support Pipeline for Day-Ahead Photovoltaic Forecasting. Proceedings of the Fourth UK AI Conference 2026, in Proceedings of Machine Learning Research 348:52-67 Available from https://proceedings.mlr.press/v348/dehghan26a.html.

Related Material