Robustness of Deep Learning Models for PV Power Forecasting under NWP Forecast Errors: A Spatiotemporal and Physically Interpretable Analysis
2026-07-14 • Machine Learning
Machine Learning
AI summaryⓘ
The authors study how well different AI models can predict solar power output when the weather forecast inputs have realistic, complex errors. They create a testing method that simulates weather forecast mistakes while keeping the physics realistic, helping isolate how these errors affect predictions. They find that sequence-based AI models handle noisy weather data better than simpler models by relying more on past stable information. Their analysis helps guide choosing and evaluating models not just for accuracy but also for reliability and speed in real-world uncertain conditions.
photovoltaic forecastingnumerical weather predictionheteroscedasticitysequence modelsmachine learningPatchTSTGRUSHapley Additive exPlanations (SHAP)Integrated Gradients (IG)Pareto analysis
Authors
Dandan Chen, Yan Zhao, Xuepeng Chen
Abstract
Engineering use of AI forecasting models requires not only high nominal accuracy but also predictable behavior under uncertain inputs. In photovoltaic (PV) forecasting, this requirement is especially challenging because numerical weather prediction (NWP) errors are temporally correlated, state dependent, and physically coupled across variables. Existing evaluations, however, often rely on perfect forecast assumptions or simplistic perturbations that do not reflect these characteristics. This study presents a physically constrained robustness evaluation framework based on simulation, using virtual PV power as a controlled response variable to isolate the propagation of input uncertainty from confounders at the plant level. Six representative machine learning and deep sequence models, including PatchTST, GRU, N-HITS, and LightGBM, are evaluated under dynamic NWP perturbations with heteroscedasticity modulated by clear-sky conditions and Erbs reconstruction that preserves radiation consistency. The results show that sequence models provide stronger noise filtering and temporal resilience than a strong tabular baseline under medium to high disturbance regimes. SHapley Additive exPlanations (SHAP) and Integrated Gradients (IG) further support a feature reallocation tendency at the case level, in which predictive reliance shifts from corrupted future forecasts toward more stable historical observations and deterministic physical priors. A Pareto analysis of accuracy under clean conditions, robustness, and computational latency then translates these findings into engineering implications for robustness assessment and model selection under forecast uncertainty.