Dynamic model reduction of a thermocline storage integrated in a micro-scale solar power plant

Concentrated solar power plants (CSPs) are one of the growing technologies thatwill help increase the share of renewable energy in the world’s electricity production.Coupling them with a storage tank allows for the storage of excess energy duringsunny periods to be reused during the day, hence improving the plant’s capacity factorand reducing the cost of electricity. Thermocline storage tanks are a very goodcompromise between cost and efficiency constraints, compared with other storagetechnologies. Nevertheless, powerful dynamic simulation tools are needed to modelefficiently the transients linked to the intermittency of the solar source. The aim ofthe proposed thesis is to contribute to the development of such tools.This paper first compares existing physical deterministic models of a thermoclinestorage tank and a parabolic trough solar field to reduced models over four referencedays. The deterministic models give accurate results with high simulation times,whereas the reduces models are fast, but loose some precision in the results. Someflaws of the simplified tank model are detected, and a third model of storage systemis designed. Based on the study of numerous charging and discharging processes, thelaw that characterizes the evolution of the thermocline is computed and integratedin the new model. This model is then validated over the same four reference days;the dynamic update of the height of the thermocline allows this new model to fitvery well any weather condition.The model developed has fixed dimensions and parameters, which limits its generality.As such, a fourth model of tank is developed, based on dimensionless numbers.This last model is validated in various conditions, and is therefore suitable to anysituation, with no constraint regarding weather conditions, geometry of the tank orworking fluid. The simulation time required by this model is between 75 and 180times less than that of the first complex model, and the robustness of the model isflawless, which makes it a very powerful tool. Finally, a new control strategy for thesolar power plant is assessed : it allows validation of the new model of tank in yet anotherset of working conditions, as well as investigation of advantages and drawbacksof one strategy over an other. An unexpected observation is that the thermoclineheight at the end of the day does not depend on the strategy used, even though theevolution is different in both cases. Some numerical issues that have been tackled tobring the model to a perfect robustness are also discussed.