(82b) Fast Optimal Control of Exothermic Packed-Bed Reactors Via Reduced Order Models
AIChE Annual Meeting
2017
2017 Annual Meeting
Catalysis and Reaction Engineering Division
Modeling and Analysis of Chemical Reactors
Monday, October 30, 2017 - 8:20am to 8:40am
Utilizing volatile renewable energy sources (e.g., solar, wind) for chemical production systems
requires a deep understanding of their dynamic operation modes.
Taking the example of a methanation reactor in the context of power-to-gas applications,
a dynamic optimization approach is used to identify cooling trajectories for time optimal
reactor start-up which simultaneously inhibits distinct hot spot formation. Therefore, we developed
a dynamic, two-dimensional model of a packed-bed tube reactor for carbon dioxide
methanation which is based on the reaction scheme of the underlying exothermic Sabatier
reaction mechanism [1]. However, dealing with large-scale nonlinear dynamical process models
(derived from PDEs) often leads to many computational diculties. Facing this issue,
snapshot-based model order reduction techniques, such as proper orthogonal decomposition
together with the discrete empirical interpolation method (POD-DEIM) generate considerably
less complex models, featuring a lower number of states and, furthermore, guarantee
an acceptable model error [2].
We illustrate the applicability of POD-DEIM to the methanation rector [3] and show that
for dynamic optimization the reduced order model outperforms the full order model in terms
of CPU time, memory cost, and even feasibility. Thus, these surrogate models show a vast
potential for future PDE based online control applications (e.g., NMPC).
requires a deep understanding of their dynamic operation modes.
Taking the example of a methanation reactor in the context of power-to-gas applications,
a dynamic optimization approach is used to identify cooling trajectories for time optimal
reactor start-up which simultaneously inhibits distinct hot spot formation. Therefore, we developed
a dynamic, two-dimensional model of a packed-bed tube reactor for carbon dioxide
methanation which is based on the reaction scheme of the underlying exothermic Sabatier
reaction mechanism [1]. However, dealing with large-scale nonlinear dynamical process models
(derived from PDEs) often leads to many computational diculties. Facing this issue,
snapshot-based model order reduction techniques, such as proper orthogonal decomposition
together with the discrete empirical interpolation method (POD-DEIM) generate considerably
less complex models, featuring a lower number of states and, furthermore, guarantee
an acceptable model error [2].
We illustrate the applicability of POD-DEIM to the methanation rector [3] and show that
for dynamic optimization the reduced order model outperforms the full order model in terms
of CPU time, memory cost, and even feasibility. Thus, these surrogate models show a vast
potential for future PDE based online control applications (e.g., NMPC).
References
[1] Bremer, J., Ratze, K.H., and Sundmacher, K. CO2 Methanation: Optimal Start-Up
Control of a Fixed-Bed Reactor for Power-To-Gas Applications. AIChE Journal (2016).
[2] Benner, P., Gugercin, S., and Willcox, K., A Survey of Projection-Based Model Reduc-
tion Methods for Parametric Systems. SIAM Review (2015).
[3] Bremer, J., Goyal, P., Feng, L., Benner, P., and Sundmacher, K., POD-DEIM for
Ecient Reduction of a Dynamic 2D Catalytic Reactor Model. Computers & Chemical
Engineering (2016).