OpenFOAM v9: First Contact

Hey there,

It’s Robin from CFD Engine and it’s OpenFOAM release week (again) – this time it’s Version 9 from The OpenFOAM Foundation.

I haven’t used the Foundation version for a while (since v6 I think) but it’s nice to have an idea of what’s going on over the fence – it’s not FOMO, it’s professional curiosity.

So this is my release note teardown, like the recent OpenFOAM v2106 one, the things that caught my attention, piqued my interest or surprised me – the things I think you might be interested in.

There are a boatload of multiphase, thermophysical, heat transfer, reaction & particle-related changes that I have no business commenting on so, if that’s your bag, you should definitely check out the full notes.

Breaking Good?

This flavour of OpenFOAM is going through a redesign & things are getting broken, intentionally.

Old complex solvers are being made redundant, dictionaries are changing purpose, content & location, functionality is being split, combined & streamlined across the board.

Some bits are backwards compatible, some not so much, but change is in the air and this could be a painful upgrade.

Hopefully it’s a worthwhile one that’s fit for the future 🤞

Here are my high (& low) lights…

DrivAer Tutorial

There’s a new external aero / simpleFoam tutorial featuring the fastback version of the DrivAer model (a popular automotive benchmark geometry). I look forward to borrowing this one in the near future 🤫

One thing that caught my eye here was the use of command-line options in the Allrun to change the way the case is run.

For example, you can use:

./Allrun -mesh <S|M|L|XL> to change model size, from 440K cells (S) up to ~200M cells (XL) – it uses refineMesh to refine your blockMesh before running snappyHexMesh.
./Allrun -cores <nCores> to change decomposition – it uses foamDictionary to set the numberOfSubdomains in decomposeParDict to whatever you’ve specified.

It’s all done within Allrun, using simple(ish) bash scripting. Scripting & OpenFOAM are a powerful team, but things get tricky if the command-line isn’t your comfy place, so it’s great to have worked examples like this to learn from.

`transformPoints`

There’s a nice usability change to transformPoints which allows you to specify a sequence of transformations using a single string on the command-line.

Do this translation, then this rotation, then this scale, etc etc.

Applies to surfaceTransformPoints too 👍

These tools are now quite different between .com & .org versions, both having features that it would be nice to see in the other one – hey ho.

`functionObject`

This release includes some big (breaking?) changes to the way functionObjects are packaged & used.

Most functionObjects no longer need a file containing the function definition. You can now specify the most common settings inline when you invoke the function, e.g.

postProcess -func faceZoneAverage(name=f0, U)

It might not look like much, but check out the giant controlDict in the new example case to get a fuller picture.

Seems neat – contrast it with the collection of FO files in the ESI treasure trove that do broadly the same thing – I like it.

I’d have to extract them into a separate file though, that controlDict scares me 🙈

R.I.P. `fvOptions` ⚰️

This feels like a big one – fvOptions is being split into fvConstraints & fvModels in the system & constant directories respectively.

The fvModels part of this change also makes several complex solvers redundant.

For example, reactingParcelFoam & sprayFoam are being replaced by reactingFoam, with their particle models now specified in fvModels.

There seems to be some backward-compatibility (for now) but this signals the direction of travel & will probably break things / open up new opportunities, in equal measure.

Bits & Bobs

streamLines now include an age field by default;
We now have Function2 which is twice as good as Function1 🤭
There are a couple of new turbulence model variants for kklOmega & kOmega;
TIL that the version in the FoamFile header refers to the version of the file format not the OpenFOAM version (it’s even less useful than I thought) & it’s now optional.

Thanks Again

Any OpenFOAM release represents a huge amount of work, so thank you to The OpenFOAM Foundation & it’s sponsors for continuing to maintain, develop & release such an amazing CFD toolkit & for giving us options 🙏

There is A LOT of other stuff going on in v9. I recommend taking 10mins to go through the release notes yourself. I find them a little tricky to parse, but you’ll may well find something that I’ve missed, specific to your CFD.

Let me know what you discover.

Until next week, stay safe,