Deep Conversion in Packed Bed Reactors

A graphic depicting the concept of a Packed Bed Reactor.

Deep Conversion in Packed Bed Reactors

Research Activities:

1.) Experimentally and numerically characterize the packing of particles in packed bed with particle geometries and vessel geometries as parameters;

2.) Model fluid flow and species transport in packed beds using CFD methods;

3.) Model the packed bed as a “reactor”, incorporating the local voidage distributions and fluid flow;

Packed bed reactors are omnipresent in oil refining, chemical manufacturing, and other processing industries. In this kind of reactor, solid catalyst particles are held stationary in a large packed bed, and gas or liquid is made to flow through. Typical dimensions of these industrial packed beds have a diameter of about 5 to 10 meters. As the phases move through these columns containing one or more reactant species, they percolate into the catalyst particles which are typically 1-3 mm in size. The reacting species in them diffuse inside the particles, which occurs at the “active sites” of metals within the porous particles (pore sizes in the range of 1-10,000 nm). Product species diffuse back into the flowing phases and eventually flow out at the exit of the column as the “product flow”. Packed beds are mostly applied in hydrotreating for refining oil. Viscous crude oil or its derivatives flow concurrently with hydrogen gas and get chemically treated to produce clean fuels.

When pellet catalysts are loaded into these beds, they won’t pack uniformly into the vertical columns. The intervening space between the particles creates a bed void fraction that is variable point-to-point. Another issue is the relative orientation of non-spherical particles, which depends on the way the particles have been dumped into the column. This causes variations in the flow of the fluid phases and wetting patterns when both liquid and gas phases flow together in these packed reactors.

The local packing density variations, related variations in local fluid flow, and the reaction of chemical species lead to variations in the exit conversion of chemical species. In many situations, these variations are small relative to the average performance of the reactor. However, these small variations in packing distribution have a dramatic impact on overall reactor performance in deep processing applications. Deep desulfurization of diesel fuel is one of these applications, which reduces sulfur content in diesel from around 20,000 ppm to less than 10 ppm (which is a requirement for Euro 6 fuel in Europe or Bharat VI fuel in India). Other major petroleum processing applications include second-stage hydrocracking and resid hydrotreating applications. The latter is a modern process for converting heavy and tarry oils into lighter grade fuel-ready hydrocarbons.

The following problems need to be solved:

• In what way do the particles pack as they are dropped into the column to form the bed? What is the role of particle size, particle shape, particle properties, wall properties, and hopper design parameters? What is the role of the method of packing, such as the devices used for loading the particles into the column?

• Given that a certain bed is formed, how do we mathematically characterize the three-dimensional arrangement of catalyst particles?

• How can the simulated packing structure with appropriate experiments be validated?

• Given that a certain arrangement is formed, how can the fluid flow and the wetting be modeled?

• What do simulations at the particle dimension scale reveal about the characteristics of the flow?

• Given that we establish the third point, how does the arrangement affect reactor performance? Based on this, recommendations on how to pack the bed, so that optimal reactor performance is obtained in the end, could be made.

To answer these questions, simulations for packing of particles in packed beds with particle geometries and vessel geometries as parameters will be developed. The void fraction in beds with a known arrangement will be characterized experimentally. For the experimental data and the numerical models, the packed bed structure will be characterized mathematically. Furthermore, the fluid flow and species transport in packed beds will be modeled using CFD methods. Finally, the packed bed will be modeled as a reactor, incorporating the local voidage distributions and fluid flow.

These studies will result in the ability to quantify and characterize three-dimensional bed structures, which can be used as inputs to CFD models for tracking the local fluid flow. It will also help in understanding what kind of inlet boundary conditions lead to what kind of packing pattern. Based on this, design recommendations for improving the packed bed reactor can be made. The aim is to relate the way of packing to the formed bed structure through the performance of the bed as a reactor. Additionally, a deeper understanding of species transport in different void fractions of catalyst particle beds and a reduction in overall costs associated with achieving emission standards is desired.