Economic model predictive control for packed bed chemical looping combustion

Abstract

This study presents an economic model predictive control (EMPC) scheme for the packed-bed reactor (PBR) design of a chemical looping combustion (CLC) system. This scheme is enabled by a pseudo-homogeneous process model, which accurately represents the macroscopic process behaviour, and enables economics to be optimized online. Both CLC stages are optimized whereby the oxidation stage considers energy generation and inert gas use, while the reduction stage economizes CO2 production and fuel use. To understand the behaviour of the EMPC, the scheme is tested on a CLC plant represented by a multi-scale process model. Both stages are tested under varying initial inlets, energy prices, carbon prices, and fuel prices. The optimal policy for the oxidation stage is found to be that of maximal peak temperature, where the EMPC results in revenue while a standard NMPC controller may result in losses. The optimal reduction stage behaviour is highly dependent on carbon and fuel prices whereby a trade-off between CO2 selectivity and throughput are observed. The EMPC approach is found to be economically superior to advanced regulatory control with up to an order-of-magnitude and ∼33% economic improvements in the oxidation and reduction stages, respectively. This study represents a step forward towards the adoption of the intensified PBR CLC process as an emerging and viable technology for heat generation with inherent CO2 capture.