I took a closer look at my CFD results.
And here's why it needs more work (and what you should know)...
Mesmerising isn’t it?
It’s not the animation but the changes in the fields over the time steps…
It can almost put me in a trance!
Imagine what this would look like if the geometry was a lot smoother, refined and with a higher-density mesh…
And this is often the reason it’s so easy to get carried away with the simulation results.
I’ve learnt from experience that CFD solvers are great at producing plausible results if setup with a little effort.
And I thought I’d pause and write this up, highlighting the issues at this stage for other Engineers following along.
What was observed:
A strong central jet forming and separating into the two immediate runners
The model used was Laminar with an Upwind scheme for increased stability.
This was expected, so to speak.
Features that are not so good (and not immediately visible):
The flow is being spread artificially - the Upwind scheme is first-order (stable but not so accurate).
The outer runners are starved from air - not great as an engine manifold (but good for an academic-style CFD investigation.
The flow regime for the physics is actually turbulent (Re ~13000).
Over the next few weeks I’m planning to improve this particular study and share the details with you right here.
The reality of running simulations
This goes beyond OpenFOAM.
It’s about running simulations in general.
As mentioned, I’m an experienced Engineer and I’ve always been skeptical of taking things at face-value. I take cause and effect quite literally!
I can’t get myself to trust pretty pictures.
But I can increase the trust in the results by following specific, objective steps that removes the subjectivity from the investigation, naturally boosting my confidence in it.
Physics solvers and computers are capable - very capable, regardless of the inherent errors they display.
No one can dispute the utility in the valuable insights they provide, especially an Engineer that has used them to directly optimise products designed for the real world.
Engineering discipline with a verification and validation mindset needs to be top-of-mind if we want to continue using simulation tools as part of the modern development process.
I share this because I believe tools should be assistive to the Engineer and not used as crutches, casually or carelessly.
See you soon. 👋
Nasser