Surrogate Model Based Optimal Synthesis of Solid Sorbent Carbon Capture

Concerns about climate change have led countries to set ambitious targets for the reduction of anthropogenic greenhouse gas emissions and a number of initiatives have been set up to reduce CO2 emissions1. Carbon capture, utilization, and storage technology is an essential route to achieving a meaningful reduction of global CO2 emissions in the context of continued fossil fuel use in the power sector. For example, the U.S. Department of Energy (DOE) has initiated the Carbon Capture Simulation Initiative (CCSI) to accelerate the development of carbon capture technology. Solid sorbent based technologies are being used to demonstrate the capabilities of the computational tools being developed under this initiative. Obviously, process synthesis, which has had a conspicuous impact on the development, design and operation of (petro) chemical processes2, will play a vital role for the development of a cost-effective carbon capture process. In this study, a superstructure optimization formulation for the optimal synthesis of a carbon capture system has been established, various reactors including fluidized bed reactors and moving bed reactors together with different potential topologies have been included in the optimization formulation. Since detailed first principle models of the reactors are computationally intractable for large scale superstructure optimization, we have developed the Automated Learning of Algebraic Models for Optimization (ALAMO) approach to generate a set of low complexity algebraic surrogate models of different reactors to reduce the computational burden associated with reactor models developed in commercial process simulation packages. Through solving the above formulated mixed integer nonlinear optimization problem which aims at minimizing the increase in cost of electricity due to capture while achieving 90% capture target the potential equipment configurations, interconnections and design/operation levels of the selected equipment can be identified. This presentation will demonstrate the application of this superstructure-based process synthesis approach for a solid sorbent-based carbon capture system.