Development of a Hybrid First Principles-ANN Model for the Steam Hydrator in a Calcium Looping Process

Abstract

One of the promising technologies for isolating carbon dioxide from a mixture of industrial flue gases is the calcium looping process. This process involves a reversible reaction between sorbent Calcium Oxide and Carbon Dioxide. Because sorbent loses its activity after multiple cycles, hydration step was proposed, which is another reversible reaction where deactivated sorbent is treated with steam to form Ca(OH)2, which undergoes the backward reaction to give back the regenerated sorbent. Blamey et al. (2016) developed a shrinking core model based on which, studies were carried out on a small experimental reactor. This paper aims at developing a hybrid model by combining first principles model and an ANN model for improved prediction of the conversion in hydration of calcium looping process and to scale it up for optimal operations. The hybrid model is tested for various combinations of training variables and data sets with respect to temperature and cycle number and it is found that the hybrid model indeed gives better results. The performance prediction of Hybrid modelling is compared to the individual performance prediction of the ANN model and First principles approach.