Learning Forum

[Richard_Marx]

Hello, I am trying to simulate a bioreactor in Ansys CFX 2019 R3 with sparging. The bioreactor has two liquid phases (water and oil) in which particle sizes of the oil droplets are important. In addition, the coalescence and breakup coefficients need to be tuned to match experimental results. I had been using a 5 group homogenous MUSIG model for the oil to understand the particle size, but now we need to add sparging to the reactor system. nPreviously I had used a similar setups for the air population balances (Air is the polydispersed phase with 5 homogenous MUSIG groups w/ Luo Svendsen Breakup Model and Prince & Blanch Coalescene. With just air and water it seemed straight forward that the degassing boundary condition was correct. Also, please note (I did not specify any vapor head space in the reactor as the default domain is all liquid. nI tried some initial runs with the air/oil/water system and of course both polydispersed phases see the degassing boundary condition as an outlet. I am looking for suggestions on how to set up my boundary conditions and or oil phase to properly model breakup/coalescence while not having the oil coming out the top of the bioreactor. Is an open boundary condition with head space added the best option here? Thanks in advance! 

[DrAmine]

Can you in concise way describe the oil phase and its morphology? Do you want to account several polydisperse phases? n

[Richard_Marx]

Thanks for the quick response Dr. Amine. nThe oil phase is corn oil and so we're interested in the particle size/distribution of it in order to understand it's impact on our cell population. nYes. I do want to account for several polydispersed phases (air bubbles from sparging and oil droplets). nThanks, nRich n

[DrAmine]

You will then require to tune the breakup and colascence kernels and perhaps add your own kernels. Moreover I have not seen any work where two secondary phases are modeled as polydisperse in the Eulerian Framework. Some approaches used A transport equation for interracial as alternative. Moreover the kernels assume that there is a single continuous phase. So what are now your questions?n

[Richard_Marx]

Okay. Thanks I will need to do some research into the best way to tune the kernels. nIs there any reason why you couldn't run two secondary phases in the Eulerian Framework?nI have moved to a pressure opening instead of a degassing outlet and initial results seem okay with the two population balances running. n

[Richard_Marx]

In addition, once the air population balance is complete, is there a way I could freeze the physics on the air bubbles and run just the oil population balance?nThe air will be the dominant force in the physics so if I were to get my air population balance correct, freeze the physics and introduce the oil, I don't believe the oil would impact the air bubble size. nThanks again for your help Dr. Amine! n

[DrAmine]

Okay. Thanks I will need to do some research into the best way to tune the kernels. Is there any reason why you couldn't run two secondary phases in the Eulerian Framework?I have moved to a pressure opening instead of a degassing outlet and initial results seem okay with the two population balances running.https://forum.ansys.com/discussion/comment/97854#Comment_97854

Expensive and requires a lot of closure laws. I generally use IMUSIG in CFX and that is working fine. Moreover I didn't work on cases requiring more than one polydisperse phase (IMUSIG A N phases) in CFX.nDegassing is actually only helpful if small amount of disperse phase should be removed.Regarding freezing certain Equations for certain phases: I am afraid this is not possible. One turn off Equations but all phases might be affected.n