Oxygen mass transfer in CFD
In the process industry, oxygen is sometimes required for the reaction. In these cases, air or pure oxygen are blown into the system. Estimating the volume of gas supplied is obviously based on the process requirements, but also on the utiliatsation rate. This utilisation rate is a function of how much of the available oxygen can be transferred given the operating conditions. The question then is, can the rate of oxygen transfer into the liquid be estimated using CFD for process applications? Industries where this is important include mining, waste-water, petroleum and food and beverage.
Consequences of insufficient mass transfer
During the design, a required mass transfer coefficient is determined based on the process conditions. A mixer can then be sized appropiately This mixer would be sized based on correlations such as the van Riet equation for mass transfer with a Rushton turbine. We have found through our own CFDs that there is a decrease in the mass transfer achieved when scaling up. This is not expected since the correlations are typically based on lab scale tests only.
During the design, a required mass transfer coefficient is determined based on the process conditions. A mixer can then be sized appropiately This mixer would be sized based on correlations such as the van Riet equation for mass transfer with a Rushton turbine. We have found through our own CFDs that there is a decrease in the mass transfer achieved when scaling up. This is not expected since the correlations are typically based on lab scale tests only.
Oxygen transfer into the liquid is notoriously difficult as oxygen is only slightly soluble in water. Estimating oxygen transfer is also difficult due to the complex dynamics and the different scales starting at the scale of turbulent eddies to bubble sizes and finally to the scale of the reactor.
Insufficient mass transfer will limit the reaction and therefore the production efficiency which ultimately negatively affects production yield. Conversely too much mass transfer may mean that production capacity is lost or that resources are being wasted. Therefore, correctly estimating mass transfer is a must from the design stages all the way through to operations.
CFD as a solution
CFD provides a means to estimate the mass transfer in a reactor. CFD has the advantage of being able to evaluate multiple configurations rapidly at the required industrial scales. In test work, it is generally only possible to measure the average mass transfer which makes optimisation difficult. CFD however can provide an average mass transfer coefficient in addition to the local mass transfer coefficients throughout the reactor. When local details can be understood, dead spots can be eliminated and flow velocities can be optimised to improve the mass transfer efficiency.
How is the mass transfer coeficient calculated in CFD?
Mass transfer is difficult. CFD is difficult. For this kind of problem, both of these difficult fields are combined. Fortunately, we have a deep understanding of that can help. We have tested our methodology against literature to ensure that the simulation produces reasonable approximations of the mass transfer rate. From this validated simulation as a base we can extend to other problems. Once the simulation is set up and run, Higbie, Dankwerts and Frossling mass transfer coefficient transfer models are used to estimate the local mass transfer. These models are based on different parameters and therefore provide different mass transfer coefficients. This ultimately results in a range of mass transfer coefficients that can be expected. Within the simulation, bubble size and distribution are accounted for using population balance models that account for drag, breakup, bubble coalescence of different sizes. This approach has been shown to give excellent results.
Here is an example of a the mass transfer evaluated in a mixing application for two test scales. The smaller size is for a lab-test scale and the larger case is 10x times larger. We find it daunting that reactors have been scaled up from lab-tests without CFD when one considers the difference in size.
The image below shows the mass transfer coeficient distribution in the tank using the Dankwerts mass transfer model. This is just an example from a more detailed study. Click here for more detailed information. This figure shows higher mass transfer at the impeller tips which would be expected. Low mass transfer is expected below the impeller. In this case, it would be advantageous to improve the mass transfer characteristics in this region.
Mass transfer rates are improved with high relative velocities between the gas and water as well as turbulence.
Need our help?
Please get in contact with us so that we can understand your simulation needs. Mass transfer is complex and it is important to get all the correct information and requirement up-front. In order to carry out a simulation you would need to provide geometry files (step or stl files preferred), liquid and air properties, air flow rates. Of course more information will likely be required depending on the nature of the problem.
Have a look at the simple validation simulation that we performed here.
We look forward to hearing from you.