Modeling pipe flow in Hypermesh

Rahul_P1

Hello,

I have a project to model laminar water flow in a simple pipe. The result I should obtain after the fluid travels a certain distance inside the pipe has been well documented in theory to appear in this parabolic form:

I began by making a pipe in Inventor with the diameter calculated from the Reynolds Number formula as R_e= (rho*velocity*diameter)/viscosity for R_e<20,000 for laminar flow. My inventor model is as shown below:

I imported my model into hypermesh and tried following tutorial CFD-1200 at the following link:http://www.altairhyperworks.com/(S(3fu2zyrlbyi03xcofiue25jd))/hwhelp/Altair/hw11.0/help/hwtut/hwtut.htm?hypermesh_tutorials.htm But when I obtained my results on hyperview, they didn't look anything like the predicted theoretical parabolic result. My result is shown below:

Rahul_P1

As mentioned in the previous comment, I was not obtaining the results that the theory predicts. So I realized that one of my major mistakes is that I was defining the annular region to be the 'inlet component' as has been defined in tutorial-1200 (http://www.altairhyperworks.com/(S(3fu2zyrlbyi03xcofiue25jd))/hwhelp/Altair/hw11.0/help/hwtut/hwtut.htm?hypermesh_tutorials.htm). This does not make sense to me since the fluid element has to be one that flows through the pipe itself and not through the annular region.

Hence, my question is, what do I define as my 'Inlet component' in hyperview? Do I have to create a seperate element in Inventor with the same length as my pipe and insert into the pipe, which I can then call as my 'inlet (fluid) component'? To sum up:

Rahul_P1

Mukherja,

I see that this is a hollow pipe and not a solid pipe, this has no effect on what the inlet will be, the inlet is decided based on your fluid. so inlet will be the surface mesh of the fluid at the inlet region.

when you set the annular region as the inlet then you are saying the fluid flows between what you have labelled as inner wall and outer wall. Is this what you need to simulate?

You also say the fluid should flow through the pipe and not the annular region
'This does not make sense to me since the fluid element has to be one that flows through the pipe itself and not through the annular region.'
In which case the inner wall surface will be the only wall, you will have to solid mesh the inner region of the pipe (enclosed by your inner wall) as the fluid, currently there is no solid mesh there.

you see in this image from the tutorial you are following, the flesh coloured cfd tetramesh component is the fluid, if you generate face elements there i.e. surface elements (which will be just a circle covering the fluid at the point where the fluid enters (or leave for outlet) will be the inlet.

Rahul_P1

mukherja,

Here are the vector plots showing parabola front in both annular flow through a hollow pipe and normal flow through the pipe.

Also as a side note, flow is decided to be laminar when Re < 2300

Rahul_P1

Hello Rahul,

First of all thank you for clarifying the concept that the annular region IS the region that should be labelled as the inlet when I am surface meshing. I was confused about it initially, but it makes much more sense now thanks to your clarification. So I tried redoing the project with Re<2300 and a hollow cylinder, labelling the annular region as inlet, exactly as the tutorial 1200 suggests. My final hypermesh model looked like this:

Then I exported it into acuSolve and followed Tutorial 1800 with the following changes:

1. Under Problem Description I set my Turbulence Equation to 'Laminar'
2. I put in an initial velocity of 0.5 m/s in the y dir since the positive y axis pointed into my inlet.

My surface manager looked as follows:

But then when I tried to open acusolve it I got an error message as shown below:

Also, in an earlier run with Re=10000 when I had actually managed to reach the hyperview step, I did not obtain the parabolic curve as shown in my previous post ( I just got some small velocity swirl at the inlet and nothing along the pipe). Do you think that was because I had chosen too high a Re or because I have to do something different in acusolve than what tutorial 1800 suggests (along with the modifications I made).

Once again, many many thanks in advance for going over my lengthy doubt and providing me feedback. No one else at my college knows how to use this software so I really rely on this support forum. 

Rahul_P1

mukherja,

The undefined element error is because there are two different kinds of elements in your inlet and outlet collectors, (trias and quads) that were generated by your two different fluid elements (tet4 and wedge6)

These inlet and outlet containers you have separated as two different inflow and outflow containers, you have a tria 3 inflow and also a quad 4 inflow etc

They should be combined into one Inflow and one Outflow.

Rahul_P1

Hello Rahul,

I did what you suggested and it finally did work now. I realized that I had problems with the two different types of elements. My current final result on hyperview looks as below:

However, my final goal is to obtain the parabolic shape when the velocity flows through the entire pipe and not just the annuluar region as shown below:

So my question is, how do I go from defining the fluid to flow through only the annular region like I have done as per the tutorial, to the defining the fluid to flow through the entire width of the pipe. Do I need to create a seperate cylinder in inventor for the fluid element and then insert it into the pipe? Or do I have to do something like solid meshing the inner region as you suggested in the first answer to my question? The pipe is hollow through the middle.

Thank you

Rahul_P1

mukherja,

Solid mesh represents the fluid, so solid mesh has to be present wherever there is fluid, since you want to simulate the flow through the pipe you need to solid mesh the interior region of the pipe