Gain scheduling multi-model predictive control of the outlet temperature for the molten salt linear Fresnel collector

The molten salt linear Fresnel collector (LFC) exhibits operational characteristics such as pronounced time delays, significant nonlinearity, stringent constraints on molten salt flow velocity, and complex operating conditions. Addressing the limitations of existing control strategies—which fail to ensure stable regulation of the outlet molten salt temperature across diverse operating scenarios and thereby compromise the safe and reliable operation of solar thermal power plants—a gain-scheduled multi-model predictive control (GS-MMPC) algorithm is proposed. This strategy classifies three representative operating conditions based on the steady-state thermal energy collection (TEC) power of the TEC loop, identifies system model parameters at typical operating points using an optimized quasi-dynamic testing approach, and realizes globalized predictive model parameters by employing the dynamic average net TEC as the scheduling variable. Simulation results verify that the proposed method effectively maintains the outlet temperature within the target range, with steady-state deviations confined to ±2 °C under stable climatic conditions and controlled within ±10 °C during severe weather fluctuations. Additionally, the molten salt flow velocity is consistently regulated within the safety range of 0.5–2.5 m/s, thereby fulfilling the safety and stability control requirements of the TEC in concentrating solar power (CSP) plants.