Firm Solar Power Forecasting With Photovoltaic Overbuilding and Battery Storage in Northeastern China

• Published in: IEEE transactions on industry applications pp. 1 - 11
• Main Authors: Gao, Qi, Yang, Dazhi, Yang, Guoming, Zhang, Hao, Liu, Bai, Chen, Yun, Shen, Yanbo, Xia, Xiang'ao
• Format: Journal Article
• Language: English
• Published: IEEE 2025
• Subjects: Batteries; Battery storage; Costs; Electricity; Firm power forecasting; Forecasting; Levelized cost of energy; Model chain; Numerical models; Overbuilding & proactive curtailing; Photovoltaic systems; Predictive models; Solar irradiance; Weather forecasting
• ISSN: 0093-9994; 1939-9367
• DOI: 10.1109/TIA.2025.3584293

Due to the intermittency and variability of solar irradiance, photovoltaic power is accompanied by high uncertainties, making it non-dispatchable. To overcome this inherent limitation, this work first considers a physical model chain to improve the conversion accuracy from weather forecasts to photovoltaic power forecasts. Then, a firm (i.e., error-free) power forecasting method is developed by leveraging battery storage together with the novel overbuilding & proactive curtailment strategy, which can, in principle, deliver effectively dispatchable solar power to the grid. From an optimization perspective, the firm forecast premium is introduced as the objective function to gauge the additional costs required to achieve firm forecasting. Case studies are conducted using relevant data for all major cities of northeastern China. It is found that the firm forecast premium over northeastern China ranges from 1.68 to 2.48, with respect to an oversizing ratio of 1.03-1.65, and a configured battery capacity of 2.28-5.13 MWh. Compared to the battery-only solution to firm up the forecasts, the overbuilding & proactive curtailment strategy brings on average a 32.74% premium reduction. Besides, the future price trends of PV and battery are considered to assess the current gap in achieving true grid parity. To examine the generalizability of firm forecasting, the sensitivity of the firm forecast premium to the choice of model chains and the distribution of firm solar resources in northeastern China are analyzed. Firm forecasting ultimately eliminates the error of solar power forecasts through tangible physical assets, avoiding grid-side penalties due to the need to manage the generation fluctuations at a macro scale.