Improved efficiency prediction of a molten salt receiver based on dynamic cloud passage simulation

A detailed dynamic model of a molten salt central receiver system including the control system has been implemented in Dymola and coupled with a raytracing software to analyze the dynamic behavior of the system during cloud passages. The simulations were conducted with time series of modified cloud camera data as well as synthetic cloud movements. As it turned out, especially clouds that shade a large portion but not the complete heliostat field cause massive fluctuations in the outlet receiver temperature when applying the control algorithm from Solar Two [1]. Additionally, occasional exceeding of the acceptable salt film temperature limit was observed. Therefore, some improvements of the control algorithm have been developed, which significantly reduce the number of such critical events while increasing the effective energy output of the receiver system. However, it has been shown, that only under consideration of local DNI distribution in the heliostat field, the simulations disclose the complexity of controlling a molten salt receiver. Finally, comprehensive simulations and analysis with different types of cloud passages revealed, that overall yield losses due to dynamic effects reach up to 4 %, but can only be considered by distributed DNI data input instead of scalar time series. In conclusion, high-resolution DNI mapping methods such as cloud-camera-based systems are required for more reliable yield predictions for molten salt power tower plants in the future.