Document Type

Article

Publication Date

2026

DOI

10.1109/ACCESS.2026.3698475

Publication Title

IEEE Access

Volume

14

Pages

87143-87156

Abstract

Solar photovoltaics (PV) are a major source of sustainable energy. Yet, their power output is highly sensitive to environmental variability, particularly solar irradiance, cloud cover, wind, and temperature. Accurate forecasting of PV power is essential for efficient grid integration and energy planning, especially in applications requiring reliable longer-term forecasting rather than one-step-ahead predictions. This study presents a PV power forecasting approach using Nonlinear Autoregressive models with Exogenous Inputs (NARX), integrating large-scale numerical weather historical data as exogenous variables. Although NARX models effectively capture temporal dependencies, they can become overly dependent on historical power values, reducing responsiveness to real-time weather changes. To mitigate these inertia effects, we propose an interpolation strategy that combines predictions from a NARX model and a purely exogenous model. Our composite design leverages the strengths of both approaches, enhancing physical realism and ensuring accurate power dropping during non-generating hours. Additionally, the framework incorporates error distribution modeling, where prediction residuals are characterized using a Laplace distribution, enabling the generation of structured forecast trajectories across different sunlight phases and supporting the construction of confidence regions. Multi-step forecasts are obtained through a recursive one-step-ahead strategy, updating predictions sequentially over the forecasting horizon. Model performance, assessed using RMSE, MAE, and Interval Score (IS), demonstrates that the proposed approach consistently outperforms benchmark models, including Long Short-Term Memory (LSTM), Support Vector Regression (SVR), Random Forest (RF), k -Nearest Neighbors (KNN), and traditional time series methods, in both accuracy and robustness, while incorporating uncertainty quantification to support operational decision-making in PV forecasting.

Rights

© 2026 The Authors.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 (CC BY-NC-ND 4.0) License.

Original Publication Citation

Castañeda, D. U., Abdullah, S., Morgan, J., Xu, H., Pokojovy, M., & Tseng, T. L. (2026). A composite NARXNN approach to photovoltaic power forecasting with integrated weather inputs and uncertainty quantification. IEEE Access, 14, 87143-87156. https://doi.org/10.1109/ACCESS.2026.3698475

ORCID

0000-0002-2122-2572 (Pokojovy)

Download

Share

COinS