Efficient modeling of adsorption chillers: Avoiding discretization by operator splitting

Abstract

Reliable and computationally efficient models are crucial to improve the performance of adsorption chillers. However, modeling of adsorption chillers is challenging due to the intrinsic process dynamics. Currently, adsorption chillers are most often represented by 1-d, lumped-parameter models that use lumped models for all components but resolve the heat exchangers in one spatial dimension. Still, accurate simulations require fine discretization leading to poor computational efficiency. To increase computational efficiency, here, an alternative modeling approach is proposed that avoids discretization by applying operator splitting.The benefits of the proposed modeling approach are demonstrated in two real-world case studies. First, the resulting adsorption chiller model is calibrated and validated with experimental data of a lab-scale one-bed adsorption chiller: The proposed model retains the accuracy of a state-of-the-art model while increasing computational efficiency by up to 70%. Second, the proposed model is applied to a commercial two-bed adsorption chiller with excellent accuracy. Finally, the validity of model is tested by varying the ratio between the overall heat transfer coefficient and the heat capacity rate of the fluid in the heat exchangers. The proposed model is shown to be well suited for the conditions present in most adsorption chillers.