Multi-scale analyses of cycled industrial-scale packed-bed adsorbers

Doktorsavhandling, 2021

Separations processes account for 10%–15% of the total energy consumed in process industries worldwide. In such separation processes, it is common to use activated carbon in packed-bed adsorbers, to remove undesired substances from a process stream. This is particularly useful for the removal of harmful volatile organic compounds (VOCs) in a multitude of settings and in an ever-growing, billion-dollar industry. Much research effort has been expended in understanding and improving the use of activated carbon for VOC removal. In this context, numerical modeling has become an increasingly useful tool as computers become more powerful and faster.
Most of the currently used numerical models describe laboratory-scale environments, where the circumstances regarding the state and operation of the adsorbers are well-controlled. However, very little work has been done on this topic for industrial scale, with real-world operational cycles and bed states. Since there are major differences between the industrial operation of packed-bed units and laboratory-scale controlled environments, the applicability and performance of numerical models for real-world industrial settings need to be investigated.

In this work, a one-dimensional (1D) numerical model is formulated for an industrial-size adsorber and compared with real-world, industrial temperature data from a biomass gasification plant operated in Gothenburg, Sweden. The end-goal of the model is an improved understanding of the requirements for a successful numerical model of real-world, industrial conditions. This is done to facilitate the design and optimization of packed-bed setups for industrial conditions before construction of the actual facilities, as well as to characterize and improve units that are already operational. The model is also used to study how best to simulate industrial conditions and how to use steam as a regeneration medium for temperature-swing adsorption (TSA) operation.

To improve packed-bed adsorbers, a detailed three-dimensional (3D) numerical study is also performed on the material packing structure. Here, the flow through a bed section is studied and packings with different particle shapes are compared using the Lattice Boltzmann Method (LBM).

The results show that major trends in the industrial data are captured, while some aspects of the dynamics of the real process are not well-described. This is due to the complex composition of the product gas from biomass gasification, and limitations associated with the adopted modeling for steam and water. The results also show that in order to simulate industrial cases, the cycling procedure used in industry should be incorporated into the model, so as to account for the different adsorption mechanisms that emerge during cycling. Finally, it is shown that packing the bed with spherical (rather than cylindrical) particles reduces the pressure drop across the bed.