화학공학소재연구정보센터
Applied Energy, Vol.111, 1120-1125, 2013
A thermo-mechanical model of packed-bed storage and experimental validation
Several new power plant technologies such as concentrating solar power plants (CSPs) or adiabatic compressed air storages (ACAESs) depend on heat storage systems as central plant elements. Where gaseous heat transfer media at elevated temperature levels are used, a regenerator-type heat storage is a particularly cost-effective solution. Though used in steel and glass industries today, a cost-effective adaptation to power plant applications is still an open issue. When designed as a packed bed, they offer large heat transfer area and numerous options for the use of low-cost inventory materials. However, such a packed bed design is prone to mechanical failures caused by the punctiform contacts, especially during thermo-cyclic operation. To reduce such risks, a simulation tool has been developed to investigate the thermo-mechanical behaviour during thermal charging and discharging process. As a modelling approach, the mechanical model of a packed bed was coupled to the thermal model of a regenerator storage. The mechanical model describes the mechanical state of each individual particle, based on the discrete element method (DEM). The coupling of the equations through the thermal expansion of the particles then allows to calculate the thermally induced forces and the resulting particle movements during the cyclic storage operation. The implementation of a time-step control speeds up the computation significantly. For model validation, a test rig was developed and erected. Main elements are a thermal storage with an inventory of 1.1 tons and air heaters providing a heat rate of 30 kW for an operation at high temperatures. High-temperature force sensors have been developed and been integrated into the storage walls to measure the local mechanical particle-wall contact forces under static and cyclic conditions. The present paper outlines both the model and the experimental investigation performed for model validation. (C) 2013 Elsevier Ltd. All rights reserved.