[lbergmann]

I am modeling a valve with five inlets. I want to create a large body of water that connect all the inlets so that I can have water flowing from the body into the valve.

Later on I want to calculate the flow through each of the inlets. My thought was to create a named selection of the surfaces. I can do taht by having one part with two bodys, but then I have problems with having the water flow from one body to the next.

Any sugestions on this?

[ghorrocks]

I do not understand your question. A diagram would help.

But if you want to know the flow rate in 5 ducts a simple way of doing it is to put an interface in each duct (5 separate interfaces), then you can put a monitor point on the flow rate of each interface.

I think you can also prescribe planes in V15 so you do not need interfaces but I have never used this so am not sure about it.

[brunoc]

From what I understand, what you'll have is a duct that connects to the five inlets. So your five inlets will no longer be inlets, they'll just be internal surfaces. Your true inlet will be at this other body. Is that it?

If so, then just generate your mesh keeping those two bodies separated and create a mesh interface. Even if the mesh is the same, you'll still be able the create a 1:1 interface and then use one side of these interfaces to check the mass flow.

Another option is just to create the named selections and try to use something like massFlow()@REGION:nameOfValveInlet. The only thing here is that I'm not sure about where the face normals you point at, so you'll have to check that yourself, since it will directly affect your mass flow values.

Personally, I'd go with the first suggestion.

[lbergmann]

I am terrible at describing it.

But yes, my five inlets will no longer be inlets but will be internal surfaces. I managed to slice the geometry and select the surfaces.

My problem the arose when I did the post processing. I defined the inlet as a mass flow rate of 2 kg/s. Through the outlet I get 1.98 kg/s (this is due to meshing on the outlet.) But through my 5 "inlets" I get all together 8e-08 or almost 0 kg/s. I do not understand why.

I am not experienced in using CFX at all. I am not sure how to create interfaces?

[brunoc]

This is what I meant about the face normals.

My guess is that both sides of your surfaces are inside the same NS, so CFX sees two surfaces whose normals point at each other. Think of it as having an interface an then adding the massflow from both sides of that interface. The result will be 0.

Try the other suggestion I gave you above. It should be easy to change your setup and you won't even need to rerun the whole case; just run it using your current result as an initial guess. It will converge in a few iterations (if not right at the first one).

[lbergmann]

I dont think I get that either.

I have attached part of the geometry so you can see it. Basically fluid flows in through the outher area, through the small pipes and into the inner volume where it will flow out though one end. http://imgur.com/3NO8FPz

The lines that is seen in the middle of the pipes are where I managed to split the part into two bodies. I have one part and two bodies for this.

When I open up the results in CFD post and use the calculator there the mass flow thorugh those areas that the splitting of the geometry created is practically zero event though there is a flow both in and out.

[JuPa]

Originally Posted by lbergmann

I am terrible at describing it.

Give us a picture then.

[ghorrocks]

Your picture does not show much. Try showing a 3D view, also things like streamlines also help.