Abstract
With the evolution of electricity market mechanisms and the increasing penetration of renewable energy, traditional empirical operation strategies for concentrating solar power (CSP) plants are becoming inadequate to meet the demands for system flexibility and economic efficiency. CSP-photovoltaic (PV) hybrid power plants have emerged as key support, effectively combining the regulatable thermal storage of CSP with the low-cost energy generation of PV. This study proposes an innovative data-driven stochastic model predictive control (SMPC) framework for optimizing the coordinated operation between CSP-PV hybrid plants and the power grid, and fully exploiting the regulation capability and proactive engagement. An integrated CNN-LSTM-Attention neural network is employed for solar irradiance and load forecasting, while a multi-timescale rolling optimization mechanism dynamically adjusts operational decisions based on real-time forecasting updates. This study employs the IEEE 30-bus system and a typical daily scenario as a test case to compare three operational modes adopted by CSP plants in China throughout their developmental stages: initially a rule-based strategy prioritizing nighttime generation over daytime output, followed by a guaranteed purchase strategy, and ultimately evolving into the proposed market-oriented dispatch strategy. Subsequently, the influence of three key sensitive parameters on the performance of the hybrid power plant is analyzed. The results show that the proposed data-driven SMPC strategy demonstrates superior economic performance and enhanced grid compatibility compared to rule-based strategy, while offering greater operational flexibility and proactive adjustment capabilities than guaranteed purchase strategy. It effectively reduces power system scheduling costs and significantly improves the utilization of renewable energy sources.
Pei Wang, Bangjie Hu, Kambiz Vafai, A data-driven stochastic model predictive control framework for optimal operation of CSP-PV hybrid system under grid coordination, Energy Conversion and Management, Volume 348, Part A, 2026,120653, ISSN 0196-8904, https://doi.org/10.1016/j.enconman.2025.120653
