(343d) Multi-Scale Modeling and Design Optimization of an Industrial Hydrogen Production Plant with High-Resolution PSA and Steam-Methane Reformer Models

Hydrogen, a widely used and important feedstock for many industrial chemical processes, is typically produced at high purity through steam-methane reforming (SMR) processes. While steam-methane reforming can be very energy-intensive and is subject to catalyst deactivation, it is still the most widespread hydrogen production process and accounts for nearly half of the world production of hydrogen [1]. Such processes have very high theoretical energy efficiencies, with some elaborately integrated plants approaching 95% energy efficiency (accounting for export of steam), but the majority operate well below their design efficiencies due to non-ideal behaviors unaccounted for during design and operational optimization [2].

An SMR-based plant closely follows the general “reaction-separation-recycle” chemical process layout: furnaces with catalyst-packed tubes serve as the reforming reactors to produce syngas, from which hydrogen is subsequently separated by pressure swing adsorption (PSA). The carbon monoxide and unreacted methane are recycled to provide fuel for the furnace. Inclusion of more detail and relevant physical phenomena in the reaction and separation system models can help bridge the gap between optimal predicted and actual plant behavior [3]. However, there are many numerical and data-collection challenges related to simulating spatially distributed hydrogen reforming furnaces and PSA units. In our previous work [2], we established a detailed, physics-based SMR furnace model and showed that it could be seamlessly incorporated into a complete hydrogen plant model for optimization by using pseudo-transient models [4]. Using this extended model, the optimization of an industrial hydrogen-production plant improved the overall thermal efficiency by ~1.5%, which translated into a savings of $1.14 MM/year.

While a detailed reactor model was successfully employed in our previous work, the PSA unit was modeled as a component splitter with fixed recovery and product purity. This assumption clearly excludes the key relationship(s) between reactor furnace design and product separation efficiency. In the present work, we rely on a recently developed pseudo-transient approach for PSA modeling and optimization to seamlessly integrate a detailed, distributed (semi-continuous) PSA model with the other (continuous) units of the hydrogen plant. The full multi-scale process model with distributed reaction and separation unit operation models allows for hydrogen recovery and product purity to be optimized by altering the PSA parameters, and for the PSA design to be simultaneously specified and considered in the plant-wide efficiency.

[1] T da Silva Veras, TS Mozer, D da Costa Rubim Messeder dos Santos, A da Silva Cesar. Hydrogen: Trends, production, and characterization of the main process worldwide. International Journal of Hydrogen Energy, 42(4):2018-2033, 2017.

[2] A Kumar, TF Edgar, M Baldea. Multi-resolution model of an industrial hydrogen plant for plantwide operational optimization with non-uniform steam-methane reformer temperature field. Comp. Chem. Eng., 107:271-283, 2017.

[3] RC Pattison, C Tsay, M Baldea. Pseudo-transient models for multiscale, multiresolution simulation and optimization of intensified reaction/separation/recycle processes. Comp. Chem. Eng., 105:161-172, 2017.

[4] RC Pattison and M Baldea. Equation-oriented flowsheet simulation and optimization using pseudo-transient models. AIChE Journal, 60(12):4104-4123, 2014.

[5] C Tsay, RC Pattison, M Baldea. A pseudo-transient optimization framework for periodic processes: Pressure swing adsorption and simulated moving bed chromatography. AIChE Journal, 2017. Doi:10.1002/aic.15987