[vaibhavkhanna]

Hey, I am quite new to CFD programming and was writing my own Matlab code for 2-D flow between parallel plates to get more familiar with the simple algorithm which I will be using for my thesis. I am trying to simulate the entrance development region which results in the parabolic velocity profile.
My mesh size is 250x10, Length of the plate is 0.5m, and distance between plates is 0.01m.
Pressure under relaxation is 0.001-0.1 for different simulations shown in the per attached.
The issue I am facing with my simulations is that the continuity residuals (L2 norm of mass imbalance) for different Re and different fluids all seem to oscillate about a fixed value instead of decreasing as the iterations proceed. The termination criteria for my simulation is never met.
Here are some of the residual plots from my simulation. The velocity values are oscillating about the right solution (the centreline velocity is 1.5 times the inlet velocity) but the residuals do not decrease with iterations.

https://drive.google.com/file/d/16fb...IUvoA7YRG/view

[tas38]

Originally Posted by vaibhavkhanna

Hey, I am quite new to CFD programming and was writing my own Matlab code for 2-D flow between parallel plates to get more familiar with the simple algorithm which I will be using for my thesis. I am trying to simulate the entrance development region which results in the parabolic velocity profile.
My mesh size is 250x10, Length of the plate is 0.5m, and distance between plates is 0.01m.
Pressure under relaxation is 0.001-0.1 for different simulations shown in the per attached.
The issue I am facing with my simulations is that the continuity residuals (L2 norm of mass imbalance) for different Re and different fluids all seem to oscillate about a fixed value instead of decreasing as the iterations proceed. The termination criteria for my simulation is never met.
Here are some of the residual plots from my simulation. The velocity values are oscillating about the right solution (the centreline velocity is 1.5 times the inlet velocity) but the residuals do not decrease with iterations.

https://drive.google.com/file/d/16fb...IUvoA7YRG/view

This issue in monitoring residual convergence may be a result of monitoring "absolute" rather than "scaled" residuals. Your plot of the residuals for three different fluids suggests as much to me.


I suggest you modify your code to monitor scaled/normalized residuals of the form:





Also, using pressure relaxation on the order of 0.01 or 0.001 is overly small. If this is necessary, it may be a result of some bug/error within the code for which you are compensating.