I'd say that the necessity of two-phase flow simulation depends on the granularity of your CFD model. Typical 2-phase setups address differing phase behaviours, as e.g. the tyipcal distinction between liquid and gaseous, with accopanying slips of their material contributions in the computational cells.

If the sludge mixes well with the carrying water and exhibits an overall mostly uniform (mixture) behaviour I'd guess you are quite well off with a respective single phase assumption. But as soon as reality shows a distinct spatial separation, with regions of definitely differing state-describing variables (density, viscosity, ...) you may be probably better of with a two-phase system, as long as your geometries are prone to interact with the spatial distribution.

If there is no experimental data you may test a two-phase setup and watch out for respective potential different phase developments, but please bear in mind that such a projected result definitively needs a reality check! :-)

Cheers,

Petra

The sludge is a solid/liquid mixture, so you may well need to use a macroscopic two-phase flow model such as Comsol's Mixture Model Interface. This interface can take into account parameters like particle buoyancy and shear-induce migration.

Good luck,

Anthony 

If you use the module comsol. You can easily get the current modeling.

If you consider flow the well mixed. Can be modeled by assuming single-phase flow. But I do not recommend this method.

Dear Nikesh

Of course, a mixture of water and sediment (swasi) is a multiphase system (heterogeneous system) . However, depending on the required accuracy of the result and the fraction of fine particulate sediment (?) particles can be modeled her single-phase system. The General method may not be.

Lev

Конечно, смесь воды и осадка (всвеси) –это многофазная система (гетерогенная система) . Однако, в зависимости от требуемой точности результата и доли осадка мелких твердых (?) частичек можно моделировать её однофазной системой. Общего метода не может быть.

Левъ

Nikesh,

With your case, I presumed the homogeneous model can be applied:

It is assumed that the 2-phase-fluid is represented by a equivalent 1-phase-fluid with its properties derived from assumptions: equal flow for 2 fluids and thermal equilibrium established within the 2 fluids. Hence, the modified properties can be represented by:

1/rhoh=(x/rho1)+((1-x)/rho2)

1/miuh=(x/miu1)+((1-x)/miu2)

u=u1=u2=m_dot/(A.rhoh)

Otherwise, the separate flow model can also be used. However, with this model you need to know the interaction between 2 phase + flow cross section occupied by each phase.

Hope this helps,

Truong.

Thank you everyone for your insights. It was really helpful. 

In my case, there is not a lot of room for experiments since it is a huge water reservoir and the pipes are all already installed. All that I can do is install pressure sensors before and aft of the valve. I can also take mixture samples from an opening at the other end of the pipe. So far, the mixture sample shows that there is only a maximum of 2% of sludge in the mixture, the rest is water. I cannot really tell if the sludge is uniformly mixed with water inside the pipe. With time, the sludge in the reservoir gets emptied and the mixture contains less and less sludge. 

My approach so far has been to install pressure sensors before and aft of the valve and measure delta P at various valve opening angles. I can then use the measured Pressure before the valve as Inlet BC for my CFD. For fluids, I could take sludge samples every 5 second and measure its density and viscosity experimentally and use that resultant mixture as a new fluid for upto about 10 simulations at 5 second interval each. This system is also equipped with a flowmeter which can measure flow rate during that interval. 

How much accurate is that going to make, I certainly won't be able to say until I get some results out of it. But for now I am stuck at thinking which approach "Single phase or Multi phase" flow I should take! 

Hello Nikesh,

> ... only a maximum of 2% of sludge in the mixture ...

In that case forget about 2-phase flow, IMHO. Anything peculiar that would result from your custom adaption of a flow modeling would be so minor and therefore meaningless compared to any other cause that influences the flow regime in practice.

If you intend to identify a potential valve blocking I'd expect you can get out an arbitrary answer from a 2-phase modeling depending on your freely assumed (read: not practically verifiable) conditions.

Cheers, Petra

Petra, 

Thank you for your answer! This is something that I had in mind too but was not able to verify. 

So, in that case do you think would it also be appropriate to discard the 'less than 2%' sludge mixture and compute simply using water as operating fluid? 

------------------------------------------------------------

Truong Phung,

Is there any chance you could direct me or send me any papers or references that contain the expressions for the homogeneous model that you mentioned? 

I'd appreciate it!

Thanks,

Nikesh. 

Nikesh,

if you can assume a sensible mixing the water specific constants should do without any effective modifications. If there is a sludge settling at the bottom and your valve is prone to "see" this sludge with a comparably/markedly higher concentration at the relevant points in the investigated duct there would be an effect if it does not really mix with the water in the outflow. But to justify such an approach addtional experiments and insights would be needed, to characterize this behaviour and to set sensible initial/boundary conditions for the simulation.

If you do not have any access to such, any assumption for the simulation is just a wild guess, and you will be well in the domain of "colourful fluid displays", short CFD. :-)

Cheers,

Petra

Multiphase is the best!

It's based on your view point and the situation of the mixture. Two phase flow scheme is used when two different fluids have interaction and this interaction is important to us. If they are well mixed, it is not that necessary to use two-phase flow scheme.

Nikesh,

The fact  is that the paper from which i got the model dated back to 1983 and I don't have the hard copy in hand. However, I have extracted the the title of that paper as well as the author name from my old ref list:

-Title: Component pressure loss during two-phase flow

-By: Fairhurst C

-From: International Physical modelling of multi-phase flow (1983), page 1-24.

Regards,

Truong.

I agree with Fatemah, above.  It will depend on the degree of mixing of the two fluids, which will be a function of temperature, velocity and other factors.  If the two are intimately mixed and flowing rapidly, then you can use single phase flow simulators.  If the valve sees periods of one fluid then the other (blobs, globs or slugs of the sludge interspersed with water), then you'll have to use two-phase simulators.

Two phase will help yu identify the critical situations listed above -- which probably sparked the investigation in the first place!

A guess would be to modify the mixture with the measured properties. Other measurements should give decent boundary conditions. Two-phase models are still insufficient in their capabilities and may produce surprising results and  unnecessary headache.

First I would like to know how many componets are in either phase.

the droplet  or solid size?

do they form mixtrures that are stable or segregate?

under which conditions do they separate?.

The answers will tell you if they cvan be treated as single or multiphase.

Hello,

As mentioned above the characteristics of the flow are critical for determining the approach you are going to follow in Fluent.

I found a couple of publications that might help with your dilemma. The authors talk about different flow features and mathematical models, but the take home is relating the mixing characteristics to your model.

Additionally the last link is a guide to Multiphase flow in ANSYS that might also be useful to you.

http://www.ncbi.nlm.nih.gov/pubmed/18821237

http://pubs.acs.org/doi/abs/10.1021/ie50374a025

http://aerojet.engr.ucdavis.edu/fluenthelp/html/ug/node871.htm

I AGREE WITH DAVID SUNDIN. HOWEVER WE NEED TO USE

SOME AVERAGE VALUE FOR QUANTITIES LIKE DENSITY.

Thank you all. 

Thank you Juan Proano Aviles, those papers were very helpful. I'm trying to dig deeper now. 

Hello Carlos A. Grattoni,

Thank you for your answer.

In my case, there is water and sludge. I don't exactly know the actual content of the sludge. But it behaves more like mud (or soil). It is not completely soluble either. Just like mud, it segregates from water and gets deposited at the bottom. 

There is a 8m deep water reservoir. and sludge is deposited upto about 3m at the bottom. There is a pipe installed at the bottom. This pipe is connected to a valve and actuator. When the valve is opened, sludge flows along with water. We have taken samples from a hose at the other end of the pipe and found that the sludge content is just upto about 2% max, rest is water. 

This is what I'm trying to simulate. Pipe inlet -->valve-->pipe outlet. But don't know where to begin......!

Thanks!

Please note that, if the slugs are not almost every where and are not homogenized and mixed well, this approach (mixture method) only gives very approximated results. So it depends how accurate you want your results to be.

If the slurry is 2%, there is no way to use 2 phase model.Try to get as correct viscosity and density as possible and use the simple single phase model. We proceed in the same way when simulating water slurry and its distribution with a rotary atomizer in the flue gas desulf absorber with even 10% slurry with perfect agreement with experiments. 

@ Moha Taeibi-Rahni :: Thank you for your answer! My concern has been because the maximum sludge content in the mixture is only upto 2%, so I'm interested to know if single phase would work too! 

@ Miroslav Jicha :: That is exactly what I had in mind. Thank you very much. However, I was wondering if there is any journal or papers regarding any such cases. If you have any, I'd be grateful if you forwarded me some! 

Thank you! 

To Nikesh

I attached one of our papers that might be useful for you.