[Gerry Kan]

Howdy folks:

I am having trouble recalling the name for the type of meshing, where the geometry is discretized like objects in Mindcraft. I called it 'rasterized' grid, but is there another (i.e., official) name tor this?

Thanks in advance, Gerry.

[sbaffini]

I always used stairstep, but for some strange reason openfoamers use castellated... if I have understood correctly what you mean

[aerosayan]

voxel based grids
https://en.wikipedia.org/wiki/Voxel


PS : Also it's called "Minecraft".
There are a lot of videos and papers on the voxel tech used in Minecraft.
You can search for those terms.

[Gerry Kan]

Thanks Folks:

Now I have the chance to look it up, Ferziger and Peric used the term "stepwise approximation," but I guess "voxelation" gives the proper nuance in this case, as it implies fixed spacing (exactly like Minecraft).

Thanks again, Gerry.

[Gerry Kan]

Originally Posted by sbaffini

... for some strange reason openfoamers use castellated...

Paolo:

A side note on "castellate" ... I find it also strange, as the word came from "castello" or "castellum" (which I suspect you already know), and it has nothing to do with the topology of the mesh cells.

In this context, I suppose means to create fortification, i.e., building a castle,
or in the case of OpenFOAM, the action of packing very fine cells around the boundary of the solution domain during the early stages of snappyhexmesh.

Gerry.

[sbaffini]

Originally Posted by Gerry Kan

Paolo:

A side note on "castellate" ... I find it also strange, as the word came from "castello" or "castellum" (which I suspect you already know), and it has nothing to do with the topology of the mesh cells.

In this context, I suppose means to create fortification, i.e., building a castle,
or in the case of OpenFOAM, the action of packing very fine cells around the boundary of the solution domain during the early stages of snappyhexmesh.

Gerry.

Thanks Gerry, good to know.

For some reason I kind of associated it to how outer castle walls appear at the top, if you know what I mean _|-|_|-|_|-|_|-|

[Gerry Kan]

Originally Posted by sbaffini

_|-|_|-|_|-|_|-|

Paolo, ha ragione! I forgot that's also called a castellation! Gerry.

[andy_]

Originally Posted by sbaffini

I always used stairstep, but for some strange reason openfoamers use castellated... if I have understood correctly what you mean

In the late 70s/early 80s in my company we used the term castellated for the version of the structured Cartesian grid CFD code that blocked out individual cells. Our terminals were Textronix 4010/14s and so we didn't use terms like pixels and voxels until later. The code development was shared with a group at Imperial College. The various CFD groups at IC largely shared techniques and terminology. Henry and co of OpenFOAM largely grew out of Gosman's CFD group in the Mech Eng. department. This may be why they use the older term but it is speculation on my part.

[sbaffini]

Originally Posted by andy_

In the late 70s/early 80s in my company we used the term castellated for the version of the structured Cartesian grid CFD code that blocked out individual cells. Our terminals were Textronix 4010/14s and so we didn't use terms like pixels and voxels until later. The code development was shared with a group at Imperial College. The various CFD groups at IC largely shared techniques and terminology. Henry and co of OpenFOAM largely grew out of Gosman's CFD group in the Mech Eng. department. This may be why they use the older term but it is speculation on my part.

Might be speculation, but the information seems quite spot on to me. Thanks

May I ask if that is a term that sounds more natural in the current English language with respect to, say, stairstep (which, indeed, sounds more natural in its italian translation)? Is there a difference with American English?

[Gerry Kan]

Originally Posted by andy_

Henry and co of OpenFOAM largely grew out of Gosman's CFD group in the Mech Eng. department. This may be why they use the older term but it is speculation on my part.

Andy:

I didn't know is that Weller's Gosman lineage; I never really looked beyond quoting has paper for OpenFOAM, to be honest. In some way that means OpenFOAM has the same lineage as Star-CD.

Gerry.

[andy_]

Originally Posted by sbaffini

May I ask if that is a term that sounds more natural in the current English language with respect to, say, stairstep (which, indeed, sounds more natural in its italian translation)? Is there a difference with American English?

Stairstep is an American word not an English one. It is not a word I would ever use and, indeed, I had to look it up to check. Staircase is an English word for a flight of steps although simply stairs is perhaps more common. Castellate/castellation is still reasonably widely used English terminology but I suspect like many English words usage is declining. Don't know how widely used it is in America.

[Gerry Kan]

Originally Posted by andy_

Stairstep is an American word not an English one. It is not a word I would ever use and, indeed, I had to look it up to check. Staircase is an English word for a flight of steps although simply stairs is perhaps more common. Castellate/castellation is still reasonably widely used English terminology but I suspect like many English words usage is declining. Don't know how widely used it is in America.

To be honest (and as a self-appointed authority on American English*) I have never heard of "stairstep" to describe such a mesh before; this is why I asked the question for an official term in the first place. Nevertheless, it does contain the necessary imagery.

And for "castellate" I have only heard in conjunction in this context with OpenFOAM. I guess Andy put some historical note behind this, which was interesting to know. Otherwise I had to look it up for exact definition. In both variants of the English language it means either "to create a castle" (my definition from above discussion) or "to build battlements on existing walls" (Paolo's definition).

Gerry.

* - Modesty forbids.

[andy_]

Originally Posted by Gerry Kan

I didn't know is that Weller's Gosman lineage; I never really looked beyond quoting has paper for OpenFOAM, to be honest. In some way that means OpenFOAM has the same lineage as Star-CD.

To some extent. Gosman himself developed within Spalding's CFD group at Imperial before forming his own. On the turbulence side he will have drawn on the work of Launder and students before he left IC after a major row with Spalding. The origins go back to the early 60s at IC and even further at Los Alamos which was probably the most active in the early days.

CD was David Gosman and Raad Issa's commercial venture based in Shepherd's Bush launched in the late 80s. The code would almost certainly not have been shared with their research activities at IC/Mines but the methodology would have been strongly correlated.

Henry and co were in David Gosman's group in the 90s. I presume this was FOAM but the group had become less forthcoming about what they were doing compared to earlier. I presumed Henry was a PhD student but might have been wrong since he doesn't seem to have written up. In truth I don't know what was going on within the group or the status of the software being developed. The commercialisation of FOAM via Nabla happened more than a decade after CD and well after STAR-CD had become a leading engineering CFD code particularly in the automotive field. I presume FOAM wasn't a success commercially hence the release of OpenFOAM.

[Gerry Kan]

Originally Posted by andy_

... before [Gosman] left IC after a major row with Spalding.

Andy:

Thanks for the background. Somehow you can substitute Gosman's name with any other and the sentence would still apply. RIP, Brian.

I remember dealing with a Nabla rep back in the early 2000s. However I haven't seriously worked with OpenFOAM until relatively recently. It has come a very long way, and the open source model has really worked for them.

Gerry.

[aero_head]

Originally Posted by Gerry Kan

Andy:

Thanks for the background. Somehow you can substitute Gosman's name with any other and the sentence would still apply. RIP, Brian.

Who is Brian?

Edit: Thanks to Andy for clearing that up below.

[andy_]

Originally Posted by Gerry Kan

Thanks for the background. Somehow you can substitute Gosman's name with any other and the sentence would still apply. RIP, Brian.

I wrote it badly. Spalding had been misusing university facilities to support his commercial activities which was recognised by pretty much everyone in the department. Launder, a nice chap, eventually submitted an official complaint in order that something could start to be done about it. Then things turned very nasty. I don't know the details because it was before my time but Launder felt unable to continue in the department and left for the US (height of the brain drain) before returning to Manchester.

Brian Spalding was a smart but rather weird and untrustworthy individual. There are plenty of stories about him including a couple of my own encounters.

Originally Posted by Gerry Kan

I remember dealing with a Nabla rep back in the early 2000s. However I haven't seriously worked with OpenFOAM until relatively recently. It has come a very long way, and the open source model has really worked for them.

I dabbled once or twice in the early days but found it hard to get past the nonsense and gave up. I need to take another look.

In what way has the open source model worked for them? CD-Adapco was bought for $970 million and so I presume you are not referring to financially. The code likely has a lot of users because it is free but how many are valuable users in the sense of generating income or moving things forward? (I am asking not stating that they are few).

[Gerry Kan]

Andy:

You didn't write badly. The nuance is exactly what I was referring to. Even though I personally have nothing to do with the person, product, or company, I have heard grapevines along that line, though not to the degree you described.

On OpenFOAM, the 'open' part has opened up a whole new cottage industry for people who provide their services as consultants and developers. While it costs nothing to use OpenFOAM, private companies pay relatively large sums for support and maintainence, not unlike other commercial codes. I personally know of large OEMs who have relatively large, dedicated in-house teams for OpenFOAM support and development. I cannot say if the original purveyors of this (ESI, OFF, and others) also bring in the same kind of money as ANSYS or Siemens, but their market share is, to say the least, not insignificant.

A colleague once told me, the difference between commercial codes, e.g. Star-CCM+ and OpenFOAM is like McDonalds and chop suey restaurants (or tikka masala, for folks from the Land of Hope and Glory).

Gerry.