(83a) Further Analysis of a Simulation of an Axial-Flow Gas Induction Impeller

This is an extension of the work presented at the North American Mixing Forum's Mixing XXIII meeting in the summer of 2024.

Hoffman et al. (2022) presented performance data for a gas induction impeller whose design differed from those previously discussed in the literature in two important ways. First, the impeller generates axial flow rather than radial, and second, the gas is induced through large openings in the impeller rather than small holes. The second difference leads to the induced gas forming bubbles that are significantly smaller than the openings in the impeller as opposed to forming bubbles that are similar in size to the openings at which they are generated.

Laboratory scale experimentation for the design, development and optimization of this class of impellers is more challenging than other types of gas handling impellers. Typical methods for measuring power can be impacted by the gas measuring method (because seals are required for many of the techniques) and methods for measuring the gas flow can result in a pressure drop in the system that would not be present in normal operation (this is even true using most non-contact measurement methods). The problem of measurement and validation becomes even more difficult at the production scale. These types of impellers are typically used with gasses that that are dangerous (flammable – e.g. hydrogen) or corrosive (e.g., fluorine, chlorine, bromine), and methods for measuring the gas flowrates in such setups are prohibitively expensive. In these cases, performance can be inferred from other measurements (e.g., reaction rates, concentration response). A validated computational simulation method will allow for faster optimization of these impellers and assist in the design of industrial scale systems.

Simulation of multiphase flows is inherently complex. Many gas-liquid simulations use Lagrangian particles to simulate gas bubbles in the system. This is a useful simplification of the physics, most notably when the gas entering the systems is already in a bubble form (e.g. a ring sparger or sintered sparger). This allows break-up and coalescence kernels to be used for bubble-bubble and bubble-fluid interactions. For gas induction simulations, the Lagrangian representation alone is insufficient to model the system. An immiscible fluid model is necessary because the gas phase is continuous until it exits the impeller. Once the gas exits the impeller, the simulation resolution must be such that the bubbles of gas can be resolved. This quickly becomes computationally intractable because bubble sizes can become quite small compared to the tank size. To address this issue, a conversion method from the Eulerian immiscible fluid model to a Lagrangian particle can be used. This allows simulation of the induction process and bubble transport be accomplished.

In this study, M-Star CFD will be used to simulate gas induction at two separate scales, a 0.597m (23.5 inch) vessel and a 1.52m (60 inch) vessel. Parameters to be investigated include ungassed power draw, power draw while inducing gas, and induced gas flow rate, all of which will be compared to experimental data. Given that the impeller of interest pumps axially, the effect of pumping mode (i.e. – down-pumping or up-pumping) on these parameters is also of interest.

References:
Hoffman, Shannon M., Eric E. Janz, Kevin J. Myers, and Nicholas A. Brown. 2022. “Performance Characteristics of an Axial‐Flow Gas Induction Impeller.” The Canadian Journal of Chemical Engineering 101 (3): 1371–86. https://doi.org/10.1002/cjce.24556