Learning Forum

[Filip Krsnik]

Hello.

I just started using fluent about 2 weeks ago and I need some help regarding simulation results of venturi tube. I have an assignment where I have to compare theoretical calculation values to the values from cfd simulation. I watched some tutorials which helped me to get started. I created geometry of venturi tube in design modeler and created a mesh with inflation to update to fluent. I've set material to water, inlet as velocity inlet, outlet as pressure outlet, stationary walls. 

Then following different tutorials I tried different combinations of solution methods and viscous turbulent models, but for all results my question is the same. 

When I try to calculate pressure difference between inlet and throat using values from simulation (either by clicking on pressure contour or using surface integrals), that pressure difference is always bigger than pressure difference i get from theoretical calculations. That also means when i use that result to calculate mass flow, it is bigger than theoretical one, and coefficient of discharge gets over 1. 

I understand that it has something to do with turbulent flow and viscousity losses, but I don't know how to take into account and calculate those losses and, in the end, get some valid comparison between my results.

I've tried to search my answer through a lot of litterature regarding venturi tubes and turbulent flow, but haven't had any luck, so I'm looking for help from someone who has much better understanding of Fluent and fluids than I have.

I'm in a bit of race against time with this assignment and I'm looking forward to your answers!

[Nikhil Narale]

Hello, 

The analytical equations make certain assumptions and simplifications, such as ideal flow conditions and neglecting frictional losses. On the other hand, CFD simulations take into account a more comprehensive set of factors, including viscous effects, turbulence, and flow characteristics.

Just for the purpose of validation, you can assume inviscid flow in your CFD simulation, disregarding the realistic nature of the flow (Viscous Model > Inviscid). Then the results may get closer to the analytical results. 

[Filip Krsnik]

Thank you for answering my question. After simulating inviscid flow I am still geting higher results than analitical in terms on pressure difference and mass flow rate. Is there any other factor I should take into account?

Also, I would still like to use data from viscous turbulent flow simulation and compare it to analitical results, but I don't really know how to explain what are all the factors that make simulation results different that theoretical.

Thank you for the help, I really appreciate it.

[Nikhil Narale]

"I would still like to use data from viscous turbulent flow simulation and compare it to analitical results"

Additionally, the analytical equation for the venturi meter, derived from Bernoulli's theorem, assumes irrotational flow (no rotational or swirling motions) and laminar flow. The complex flow behaviors are well captured by the Navier-Stokes equations, which are solved using CFD.

[Nikhil Narale]

Hello, 

 

Before we discuss anything further, could you please let me know if your solution is converging? Additionally, it would be helpful if you could provide a snippet of the residuals. Thank you!

[Filip Krsnik]

This is for realizable k-epsilon model using hybrid initialization and running 500 iterations. All residuals are set to converge at 1e-6. Coupled scheme with all disccretizations set to second order.

[Nikhil Narale]

 

The residuals look good. Have you sufficiently resolved the boundary layer? Can you share a mid-section of your mesh?

On a side note, the pressure values obtained from Ansys Fluent are typically gauge pressures (relative to a reference pressure). FYI.