Higher deviation of CFD results from experimental data
 

I am simulating an ROV ducted propeller (underwater) to study thrust force behaviour with the changing duct angle.

There is a significant discrepancy between the experimental thrust force and the results obtained from our computational fluid dynamics (CFD) simulations for the T200 thruster. Our experimental data, based on the technical details provided by Blue Robotics, indicates an expected thrust force of approximately 14.5 N at 1900 rpm.

However, our CFD simulations yielded a significantly lower thrust force of approximately 5.7 N.

Simulation details,

Ansys CFX was used.

I created two domains, stationary and rotary. I used a steady state with a frozen rotor approach.

Inlet velocity is zero. wall and outlet were set as openings. k-epsilon model was used.

1)What may be the reason for this deviation?

2)Are my boundary conditions correct for this type of application?

Thank you.

Looking at your RMS values, they look not converged (check the FAQ in this Forum for convergence problems) . I would check for the Max Residuals Nodes, given in the Output file. Also, from what I can see, the mesh might have too big expansion ratios from stationary to rotating domain. This should be checked as well.

Here is the FAQ Daniel mentioned: https://www.cfd-online.com/Wiki/Ansy...gence_criteria

You should have a look at the flow in the post processor to check the flow is doing what you expect. This is because I think your boundary condition setup is weird - you have an inlet with zero flow velocity (then it is not an inlet, is it?) and a wall with a opening pressure condition (then it is not a wall, is it?).

First look Global. Did you monitor the imbalanses and the masses through all boundaries Are these correct? Is the amount of mass and momentum going in equal to what is going out (Staitonary and rotating). This should be okay first before looking into details.