[lostking18]

Hi, there

I’m a bit confused about the mass flow definition in CFX. I know that when the surface norm is opposite to the flow direction, the mass flow rate calculated by ‘ massFlowInt(1)@A1 ’ is negative. But what is the definition of the surface normal direction in CFX? Also why some of the mass averaged quantities, such as temperature, computed by ‘ massFlowAve(Temperature)@A1 ’ is negative? If it is also because the surface norm is opposite to the flow direction, why sometimes the ‘ massFlowInt(1)@A1 ’ has different sign with the ‘ massFlowAve(Temperature)@A1 ’?

Thank you in advance!

[Opaque]

ANSYS CFX uses the following convention:

Incoming quantity is positive,

Outgoing quantity is negative,


Incoming increases content, Outgoing decreases content.

If you want to formulate it via the Normal direction:

Opposite directions --> incoming

Same directions --> outgoing

[lostking18]

Hi, Opaque:

Thanks for your reply. I think your definistion works when at the boundary of the computational domain. But what if it is the surfece inside the compuatational domain? Supposing a 45 degress axial cut plane in the middle of the compuational domain, it has both incoming and outcoming flow. So I think it really depends on how we define the normal direction of the surface. I'm not sure how this is defined in CFX.

[lostking18]

More specifically, I found at area A1, the mass flow computed by:
massFlowInt(1)@A1
always have different sign comparing to
areaInt(Normal X*Velocity u*Density + Normal Y*Velocity v*Density + Normal Z*Velocity w*Density )
I don’t know what is the reason for that difference.

[Opaque]

A surface in space w/o reference between interior and exterior does not have a well defined normal. It is whatever you want to make it.

Say, you take a plane in space defined by a Normal, and a point away from the Normal and it looks identical in space to the one using -Normal.

The Normal is arbitrarily defined. The important issue if that both results can only be different by the sign, and that the Normal is uniformly defined over the whole surface.

[lostking18]

Originally Posted by Opaque

A surface in space w/o reference between interior and exterior does not have a well defined normal. It is whatever you want to make it.

Say, you take a plane in space defined by a Normal, and a point away from the Normal and it looks identical in space to the one using -Normal.

The Normal is arbitrarily defined. The important issue if that both results can only be different by the sign, and that the Normal is uniformly defined over the whole surface.

Hi, Opaque, from my observation, it seems that the norma is pointing to the convex side of the curve.