State transport routing improves short term photovoltaic power forecasts
State Transport Routing for Short-horizon Adaptation in Multi-horizon Photovoltaic Forecasting
Machine LearningArtificial Intelligence
Summary
Predicting solar power output accurately is tough because short-term trends don’t always continue over longer periods. The authors developed a method called state transport routing (STR) that adjusts existing solar power forecasts using recent power levels and their trends. STR smartly combines these signals during the first two hours without changing longer-term predictions. Tests on multiple data sets show STR consistently improves forecast accuracy for several advanced neural models.
What this means in practice
- •For energy grid operators: Improve short-term solar power forecasts to better balance electricity supply and demand using existing predictive models enhanced with STR.
- •For smart building managers: Enhance solar energy usage planning by integrating STR-refined forecasts into building energy management systems for more accurate short-term power predictions.
Authors
Xu Yuqing, Zhou Liguo, Sun Ze, Yu Lei, Jiang Mingming
Abstract
Recent power measurements provide valuable information for photovoltaic(PV) power forecasting, but directly extrapolating short-term trends can introduce substantial errors over longer forecast horizons. To address this challenge, we propose state transport routing (STR), a lightweight adapter that refines the predictions of a frozen forecasting model. STR combines the original forecast with two complementary trajectories derived from the latest measured power level and its recent trend. A horizon-conditioned router adjusts their contributions over the first 120 min, while leaving subsequent predictions unchanged. Experiments on four public PV datasets show that STR consistently outperforms a parameter-matched residual adapter. On PVDAQ, the same approach improves five neural forecasting backbones, reducing all-horizon normalized mean absolute error by 0.0201-0.2364 percentage points, with paired 95% confidence intervals excluding zero. No reliable improvement is observed for LightGBM. These findings demonstrate the potential of structured state adaptation to improve short-term forecasting across different neural architectures without retraining the underlying models or altering their longer-horizon predictions.