(591b) Modeling, Simulation, and Control of a Countercurrent Polymer-Based Water-Gas Shift Membrane Reactor for Process Intensification

Great strides have been made in the design of intensified processes due to their potential for increased efficiency and footprint reduction. An emerging challenge that comes with intensification corresponds to the control problems that arise when a process is intensified. Efficiency is crucial to the success of intensified unit operations, but such units can be more difficult to control than a more traditional setup. This challenge was highlighted in past DOE and NSF reports that identified as important next steps for process intensification and modularization: (i) the “limitations of current control systems for operation of process-intensive systems” (DOE, 2015); (ii) control problems due to “operating with highly fluctuating feedstock conditions” (Vlachos et al., 2017); and (iii) the need to “identify the relevant bottlenecks” in control of intensified processes (Vlachos et al., 2017).

The identified issues result in the loss of the degrees of freedom for control due to less unit operations and increase in the number of phenomena occurring simultaneously. This work seeks to use Model Predictive control (MPC) to assist in the control of a membrane reactor unit and to demonstrate a novel framework for simulating intensified process units. In this presentation, a recently developed non-linear MPC (NMPC) control strategy is implemented on a countercurrent water-gas shift (WGS) polybenzimidazole (PBI) membrane reactor modeled using the SimCentral simulation platform (from AVEVA Group plc.). The novel framework is demonstrated for simulating intensified process units and to explore various MPC strategies from the classical Quadratic Dynamic Matrix Control (QDMC) to a novel modified Sequential Quadratic Programming (SQP)-based NMPC.

For the application, a WGS membrane reactor is introduced with the purpose of treating syngas from a steam methane reforming or gasification process as a part of a more complex energy system. The membrane reactor model is developed using SimCentral, an equation-oriented process simulator from AVEVA. The membrane model is produced by discretization along the length of the reactor resulting in a set of submodels in series. Each submodel allows for the simulation of heat exchange, membrane permeation, reaction, or any combination of the three phenomena. This submodel structure enables a customizable, easy-to-implement unit model for future applications in the design of modular systems. In this case, a WGS kinetic model is provided by Choi & Stenger (2003) and the membrane material and model selected is from Radcliffe et al. (2016). The membrane reactor operation and associated PID controllers are simulated in SimCentral, while the classical MPC controllers and the NMPC framework developed by He and Lima (2019) is run in MATLAB which communicates to SimCentral through an OPC connection. The NMPC used in this work is a modified SQP-based NMPC. In this approach, utilization of relaxed step acceptance conditions allows for faster convergence and handling of high-dimensional systems with ease.

The MPC controllers are implemented to manipulate the set points of the sweep gas flowrate and the flow of steam into the reactor feed for quality control. The developed control framework is applied to set point tracking and disturbance rejection scenarios to identify the benefits and limitations of each implementation when used for an intensified process. The closed-loop results for NMPC when addressing the control issues will be discussed along with the identified bottlenecks in the control of the intensified process.

References

Choi, Y., & Stenger, H.G. (2003). Water gas shift reaction kinetics and reactor modeling for fuel cell grade hydrogen. Journal of Power Sources, 124, pp. 432-439.

Department of Energy (2015). Advanced Manufacturing Office: Process Intensification Workshop. Report of conclusions from a meeting sponsored by the Department of Energy. Alexandria, VA.

He, X. and Lima, F.V. (2019). A modified SQP-based model predictive control algorithm: application to supercritical coal-fired power plant cycling. Submitted for Publication.

Radcliffe, A.J., Singh, R.P., Berchtold, K.A., & Lima, F.V. (2016). Modeling and optimization of high performing polymer membrane reactor systems for water-gas shift reaction applications. Processes, 4(2), 8.

Vlachos, D., Ierapetritou, M., Dauenhauer, P., & Hock, A. (2017). Modular Manufacturing Workshop. Report of conclusions from a meeting sponsored by the National Science Foundation and the Department of Energy. Arlington, VA.