[thijs1909]

Hi everyone,

I am currently setting up my case to study fully-developed flows in tubes with a helix structure. This means straight tubes with a spiral at the wall. I want to study influence of the helix parameters on the flow/pressure drop.

The Reynolds number of my flow is Re=5000 which is turbulent. I want to compare the CFD results of two RANS models and a LES model with experiments. To minimize computational cost, I want to minimize my spatial domain by using periodic (cyclic) boundary conditions. In the minimized domain I will specify two or three spiral pitches in the domain to study the effect of those spirals on the flow.

I started easy with just the laminar case (Hagen-Poiseuille) to validate my results. I use simpleFoam with periodic boundary conditions such that my outlet will be my inlet. For that I have to specify a pressuregradient and my mass flow will be calculated. Or I specify a mass flow and my pressure gradient will be calculated. I tried to find some examples but I cannot find a way to specify a fixed pressuregradient. Is this possible in OF 5.0 using fvOptions???

What I did find is specifying the mass flow (bulk velocity) using meanVelocityForce in fvOptions. It gave me a correct velocity profile compared to my analytic solution but a strange profile for the axial pressure drop, but I thought it had to do with a coarse mesh. I refined the mesh, but now I get very bad convergence and thus a bad sol, which does not fit my analytic solution.

From an experiment point of view, what is better or smarter to do, create an experiment and control the mass flow rate and read the pressure drop or control the pressure drop and measure the mass flow rate. I would go for the first option.

[thijs1909]

With a finer mesh, this is the result I obtain. Osscilatory behavior for the pressure, but a correct result for the velocity profile.

The velocity profile is corresponds with u(r)=2*Uavg(1-(r/R)^2) where I specified Uavg = 0.2m/s in fvOptions using meanVelocityForce

[thijs1909]

I figured out that in my log file a pressure gradient of dpdx= 259.611 is calculated. This value is correct if I use u(r)=-1/4nu *dpdx (R^2-r^2) and evaluate it a u(r=0)=0.4m/s I get a dpdx of 256, which is close to the one in my log file.

Why do I not see it in my paraview?

[gu1]

Originally Posted by thijs1909

Hi everyone,

I am currently setting up my case to study fully-developed flows in tubes with a helix structure. This means straight tubes with a spiral at the wall. I want to study influence of the helix parameters on the flow/pressure drop.

The Reynolds number of my flow is Re=5000 which is turbulent. I want to compare the CFD results of two RANS models and a LES model with experiments. To minimize computational cost, I want to minimize my spatial domain by using periodic (cyclic) boundary conditions. In the minimized domain I will specify two or three spiral pitches in the domain to study the effect of those spirals on the flow.

I started easy with just the laminar case (Hagen-Poiseuille) to validate my results. I use simpleFoam with periodic boundary conditions such that my outlet will be my inlet. For that I have to specify a pressuregradient and my mass flow will be calculated. Or I specify a mass flow and my pressure gradient will be calculated. I tried to find some examples but I cannot find a way to specify a fixed pressuregradient. Is this possible in OF 5.0 using fvOptions???

What I did find is specifying the mass flow (bulk velocity) using meanVelocityForce in fvOptions. It gave me a correct velocity profile compared to my analytic solution but a strange profile for the axial pressure drop, but I thought it had to do with a coarse mesh. I refined the mesh, but now I get very bad convergence and thus a bad sol, which does not fit my analytic solution.

From an experiment point of view, what is better or smarter to do, create an experiment and control the mass flow rate and read the pressure drop or control the pressure drop and measure the mass flow rate. I would go for the first option.

Could you explain how you did it? ''In the minimized domain I will specify two or three spiral pitches in the domain to study the effect of those spirals on the flow. ''
I'm modeling a problem similar to yours.

[agustinvo]

Originally Posted by thijs1909

I figured out that in my log file a pressure gradient of dpdx= 259.611 is calculated. This value is correct if I use u(r)=-1/4nu *dpdx (R^2-r^2) and evaluate it a u(r=0)=0.4m/s I get a dpdx of 256, which is close to the one in my log file.

Why do I not see it in my paraview?

Normally, this dpdx is in the direction of your channel flow. Take the value of two different slides of pressure, and compare the pressure drop you have