(257b) Investigating the Impacts of Time-Varying Operation on Equipment Fidelity

Economic model predictive control [1] is an optimization-based control design that has received significant attention in recent years due to its ability to compute control actions which explicitly optimize the process economics and may induce time-varying operation of the process states. Despite many reports of the benefits of EMPC with respect to an instantaneous measure of process profit rooted in the dynamics of the chemical process system (e.g., [2]), an avenue that has not been explored is whether the time-varying operating policies computed by an EMPC may impact process equipment fidelity through mechanisms such as fatigue, and if so, whether this would impact capital costs and still make EMPC a competitive control design from an economics standpoint. The power plant cycling literature suggests that time-varying operation of processes can result in costs, such as increased maintenance costs [3], which need to be considered in the design of an EMPC.

In this work, we begin investigating the types of issues that equipment may face under time-varying operation of a nonlinear process utilizing EMPC by performing the closed-loop computational fluid dynamics modeling of a steam methane reforming tube from [4] but coupling this simulation with finite element analysis (FEA) to analyze how transients induced by controllers impact stresses in the reforming tube walls. We begin by investigating the differences between the stresses in the walls at steady-state and under the transients involved when the process is initialized off steady-state but driven to a steady-state by a classical (i.e., proportional-integral) control design that enforces steady-state operation. We then analyze the closed-loop process under EMPC and seek to better understand how differences in the stresses in the walls are related to the control designs with the purpose of proposing EMPC designs with constraints which encourage the EMPC to choose control actions which are economically-optimal but do not cause unnecessarily significant damage to the process equipment.

[1] M. Ellis, H. Durand and P. D. Christofides, “A tutorial review of economic model predictive control methods,” Journal of Process Control, 24, 1156-1178, 2014.

[2] M. Ellis and P. D. Christofides, “Optimal time-varying operation of nonlinear process systems with economic model predictive control,” Industrial & Engineering Chemistry Research, 53, 4991-5001, 2014.

[3] N. Kumar, S. Paterson, K. Coleman, C. Lee, D. Agan and S. Lefton, “Power plant cycling measure – evaluating historical cycling to model future grid operations,” In IEEE Power Energy Society General Meeting, pages 1-5, Vancouver, British Columbia, Canada, 2013.

[4] L. Lao, A. Aguirre, A. Tran, Z. Wu, H. Durand and P. D. Christofides, “CFD modeling and control of a steam methane reforming reactor,” Chemical Engineering Science, 148, 78-92, 2016.