Learning Forum

[dhy19944]

Hi，When I use vof-to-dpm model, since the fluid is non-Newtonian, the viscosity of the dpm particles must also be non-Newtonian. I tried to use expression to describe it as

IF(StrainRate<=0.1[s^-1],18[kg/(m s)],(IF(0.1[s^-1]

, and since the viscosity involves exponents, fluent suggests that there is a problem with the units

Can anyone help me with this error, or  tell me how to describe it using udf，thank you very much.

[Rob]

I'm not sure the DPM droplets do need a nonNewtonian viscosity unless the model is using the same material. I've run so many combinations of DPM/VOF/EWF that I've lost track of which model shares materials... 

You've set a parameter expression - I wonder if you're working in the wrong place. Use Named Expressions, 

 

[LI KUAN-LIN]

 

I am also using non-Newtonian fluids for VOF-DPM simulation. For the viscosity of the liquid, I use the Carreau Model, and when the viscosity of the particles is set to a constant value and VOF-DPM is turned on, the following dialog box pops up in the system. I think the viscosity of the particles also needs to use the viscosity properties of non-Newtonian fluids, but I am not sure how to change it.

 

[Rob]

It looks like some of the property checks are triggering. What else is switched on as my check here (laminar, Carreau model, VOF-to-DPM and DPM-to-VOF) are not showing that problem. 

[LI KUAN-LIN]

Thank you for your reply. I use the k-omega SBES turbulent model, while the Carreau Model uses functions related to shear rate.

I also tried to write the relationship between viscosity and shear rate as a UDF and import it into Fluent, hoping to represent the viscosity of particles using UDF. The UDF is as follows:

The fluid part can be successfully modeled using UDF. 

However, the following window pops up when modeling the particle part.

I think the UDF I wrote is only limited to fluid, so the particle cannot read this UDF. What I want to ask is, how can I modify the viscosity of the particle to match that of the fluid?

[LI KUAN-LIN]

The following are my case file and UDF to help troubleshoot errors. Thank you for your help.

https://drive.google.com/drive/folders/1g4h8TozejlCHBB4sUPy8Hu5ol9cqZ1zw?usp=drive_link

[Rob]

Staff are not permitted to download files, it's to avoid issues with US export law.

I'll check with a colleague. The DPM model technically doesn't have a cell thread, so the the UDF may not work in that form. 

[Rob]

And to follow up. Fluent does check the properties, and this was to ensure the set up was consistent when phase interaction models are used. As you've discovered, the downside is we can't force different material properties where it's sensible. I'll put an enhancement in as it's a useful check but needs an intelligent override feature. 

Rethinking the model. If a material is breaking up into droplets, just how nonNewtonian is it? Ie a shear thinning material is likely near the limits so can be considered as Newtonian with constant viscosity as it breaks up. What implication is there on the rest of the model if we assume that? 

[LI KUAN-LIN]

Thank you for your reply. First of all, I would like to express my deepest apologies for the discourtesy of uploading files without a clear understanding of your company’s regulations, causing you inconvenience.

 

Based on what you mentioned, I also believe that at the initial moment when the material breaks into droplets (primary breakup), which is when the VOF-to-DPM operation occurs, the fluid shear should be at its maximum. At this time, the particle’s viscosity can be set to infinite Shear Viscosity.

 

But if the droplet is successfully represented as a DPM particle, this particle will continue to break up (secondary breakup) according to related literature. Should the viscosity of this particle return to the use of non-Newtonian fluid viscosity at this time? In other words, we must know the shear rate of each particle (parcel) moving in the Eulerian phase in order to calculate the current viscosity through the Carreau Model. But what I am not sure about is how to calculate the shear rate of the particle. The above are my thoughts. Do you think they are correct?

Returning to my simulation, if I continue to use a non-Newtonian fluid to run VOF-to-DPM, is the current version of Fluent unable to operate? Or do your staff have other methods to troubleshoot the problems I encountered?

 

[Rob]

Don't worry about the files, other community members can download/help etc, it's just Ansys staff who aren't allowed to. 

As a DPM droplet is small relative to the cell how much shear effect are you expecting? I've not checked recently, but are the various break up models designed for shear thinning/thickening fluids? 

[LI KUAN-LIN]

Indeed, apart from some breakup models that require viscosity (Oh number), most breakup models are only related to the We number (without viscosity). If I only use breakup models related to the We number, can the problem I encountered be solved by the solution discussed above? (That is, when the VOF-to-DPM operation occurs, the fluid shear should be at its maximum. At this time, the particle’s viscosity can be set to infinite Shear Viscosity.)

[Rob]

If you can assume fixed viscosity then the problem goes away.

As an aside, I'm talking to one of the developers regarding shear & temperature based properties, but that's not something I can discuss further on this platform. 

[LI KUAN-LIN]

Thank you very much for actively addressing my questions. I hope the new version of Fluent next year can solve the problems we discussed.

[Rob]

Unless it's been fixed prior to our discussion it'll not be resolved in 2024R1.