PDA

View Full Version : sharing FE meshes on ISB web site



Marco Viceconti
06-17-1996, 06:57 PM
Dear Colleagues,
a few days ago I have posted the following message. It never came
out on biomch-l; probably it was lost in cyberspace (or better, since we
are changing the mail server, I made a mess).
Thus, I am posting it again. IF you receive it twice, sorry me.

***********
A while ago I have posted the a message about the idea of sharing finite
element meshes through the ISB web site. I have got about 20 replies,
although I have stressed that anybody interested should have drop a line,
just to say "me too!". Thus, the first result of my enquiry is that non so
many biomch-l subscribers are:
- interested in FEM
- interested in sharing messages

The majority of the reply are positive about the usefulness of sharing
meshes, although some of negative votes are very well motivated.

In my message I made a few specific questions; let me try to condense the
answer:

1 - Are you willing to share with others FE meshes you had developed? Do
you feel the mesh a shareable result or an extension of your personal
skilfulness, which you don't want "sell out"? If positive, how many models
you'd like to share, of which type and for which purposes?

Majority are willing to share meshes, although some express doubts about
usefulness and many indicate the reference geometry as a more useful shared
object.


2 - Do you think a FE mesh can be, under certain conditions, re-used for
other purposes or you feel every specific problem needs a tailored mesh?
More generally, does it makes sense to share FE meshes? If positive, for
which purposes (only educational, test replication, etc.).

Educational and research purposes got equal scores. With my surprise, test
replication got small support. My concern with validation of numerical
studies seems not shared. Some see shared meshes as a template or a
starting point for further refinements.


3 - In case you share a mesh, would you like to be protected against
commercial use of you mesh? If positive, do you feel the agreement-password
system used by the NIH for the Visible human of by us for the Standardized
=46emur (http://www.cineca.it/prometeo/) would be enough?

Most agreed with; some stressed the need for cross-citation rules.


4 - Now for the users: if this collection would be available, would you use
it? How frequently, and for which purposes (a base for my studies, I don't
want waste time meshing, educational only, etc.).

Here I got a wide spread of answers. Some indicated application limited to
student projects, other said it would be a way to save time a focus on the
"core business".


5 - Last, I promise. If you are really positive, in which format could we
store them? Patran neutral file would be general enough?
=46or most patran format is fine. Again some stressed the usefulness of
geometry rather than meshes. Some indicated STEP or IGES, while others
called for translators or multi-format.

What can I say more? It is hard to reach a final point; there is some
interest, but also some concerns. Furthermore, test replication does not
seems to be a major issue (I personally disagree). We have some experience
with managing the Standardized Femur Program" site, and although it is a
single object, it takes a lot of effort. I am not sure if the replies I
collected can justify the voluntary effort involved. I'd like to hear
comments from other colleagues and from BIOMCH-L people.

Marco Viceconti



**********ORIGINAL MESSAGES********
Dear Colleagues,
following some public and private discussions on the possibility to share
bio-related FE meshes through the SIB new web site, I agreed with Ton van
den Bogert to post the following enquire:

1 - Are you willing to share with others FE meshes you had developed? Do
you feel the mesh a shareble result or an extension of your personal
skillness, which you don't want "sell out"? If positive, how many models
you'd like to share, of which type and for which purposes?

2 - Do you think a FE mesh can be,under certain conditions, re-used for
other purposes or you feel every specific problem needs a tailored mesh?
More generally, does it makes sense to share FE meshes? If positive, for
which purposes (only educational, test replication, etc.)

3 - In case you share a mesh, would you like to be protected against
commerical use of you mesh? If positive, do you feel the agreement-password
system used by the NIH for the Visible human of by us for the Standardized
=46emur (http://www.cineca.it/prometeo/) would be enough?

4 - Now for the users: if this collection would be available, would you use
it? How frequently, and for which purposes (a base for my studies, I don't
want waste time meshing, educational only, etc.).

5 - Last, I promise. If you are really positive, in which format could we
store them? Patran neutral file would be general enough?

Of course, this is also a way to understand how many ISB members are
potentially involved with FEM; so if you are one of those, please find a
second to reply, event just to say "no time". As obivous if only five
person are interested, we can exchange our stuff each other without web
sites.

Marco Viceconti

************************************************** ***********

Dear Marco,

Yes, I'm interested. I'm only just beginning to use FEA, but I'll be
willing to share my nonhuman primate phalanx meshes. I'd also be interested
in seeing how others construct theirs. Count me in.

Cheers,
Brian G. Richmond
Doctoral Program in Anthropological Sciences, SUNY Stony Brook, NY 11794-436=
4

************************************************** ***********
=46rom: Gangming Luo
Subject: Re: Collection of FEM meshes for the ISB website
To: Marco Viceconti
In-Reply-To: Your message of Sat, 16 Mar 1996 15:49:57 +0100

I am very interested in sharing FEM meshes.



**************************************************
=46rom: Michael McPartland Subject:
RE>Collection of FEM meshes for the ISB website To: Marco Viceconti


Marco,
=46irst, thanks for taking the time and effort to get this project going.

I am relatively new to the FEM world and have just begun teaching a
course(s) at George Washington University which require investigations into
=46EM modeling.
I don't think I can offer much technical advice nor meshes at this time.
However, I do feel that the development of this database/wedsite and the
information maintained therein could be highly beneficial in both
strengthening the student educational process as well as improving
productivity of researchers.

I can therefore say that it would be a database/website that I would make
use of, and hopefully give back to, in the future.

Best regards,
Michael


Michael D. McPartland, Ph.D.
Instructor
National Crash Analysis Center
George Washington University







**************************************************
=46rom: "Gregory S. Klopp"

Dear Marco,

I have no problems with sharing a mesh but doubt its usefulness. As a rule,
the effort it takes to understand a borrowed complex mesh is often
comparable to the effort it takes to develop your own anyway. But, have you
considered this: If you decide to use a Patran neutral file, why not share
the hyperpatch model rather than the mesh. The model file would be much
smaller (so faster downloads) and be easier to eyeball for usefulness to a
given user's problem. The actual meshing is automatic anyway and the user
could do any customizing that he needs at this step. This would give you
the geometry you need and get around some of the tricky problems that can
erupt when using someone else's mesh. By the way, when I build a model, I
tend to develop it using hyperpatches as sort of a super element anyway and
then mesh these as required by the specific problem. That was my $.02 and
good luck.
**************************************************
=46rom: "Steve Boyd"

Marco,

In reply to your message, I am doing a small amount of FEM. From my limited
experience, I personally would feel uncomfortable to 'trust' a mesh that
someone else has developed for my own research. In short, I think that
sharing geometry would be more useful. This would save a lot of research
time required to determine geometry, and then the researcher could
concentrate on developing their own optimized mesh for their study.

Sincerely,
Steve Boyd

**************************************************
=46rom:

Marco,

1. I would be willing to share meshes I have created. I have created
several meshes of the proximal femur. These have been and are currently
being used for probabilistic FEA of joint replacement components.

I think of more interest to people would be the availability of the
geometry without the mesh - either in a format such as IGES or patran
neutral files.

2. I think meshes can be used for other purposes. Most likely they would
have to be modified, some extensively, some not so much. The availability
of a mesh may provide a good starting point for your own specific research
problems.

3. Yes. I believe any meshes shared in the scientific community should be
used for research purposes only. The agreement - password scheme sounds
good.

4. I would use it when the need arose. Generating meshes of anatomical
structures can be very time consuming. If at least the geometry were made
available, one could concentrate his efforts on creating a mesh that suits
his specific needs.

5. Patran neutral file format is widely recognizable and would be a good
starting point. If possible, other file formats could be made available
also.


Dan Nicolella
----------------------------------------------------
Daniel Nicolella
Southwest Research Institute
(210) 522 - 3222 DNicolella@swri.edu




**************************************************
=46rom: Wafa.Skalli@paris.ensam.fr (Wafa Skalli)

no time at present to respond because there are many questions and they
need reflexion time but I express my interest and will give response
further

Bye!
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3 D=3D=3D=3D=3D
Wafa Skalli
Laboratoire de Biomecanique
ENSAM, 151 Bd de l'hopital
75013 Paris
Tel: 44 24 63 68 Fax 44 24 63 64
email: wafa.skalli@paris.ensam.fr
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3 D=3D=3D




**************************************************
=46rom: "PLOEG HEIDI" (by way of

Dear Collegues,
following some public and private discussions on the possibility to share
bio-related FE meshes through the SIB new web site, I agreed with Ton van
den Bogert to post the following enquire:

1 - Are you willing to share with others FE meshes you had developed?

Ans: Probably not.

Do you feel the mesh a shareble result or an extension of your personal
skillness, which you don't want "sell out"?

Ans: A good mesh represents alot of time and money.

If positive, how many models you'd like to share, of which type and for
which purposes?

2 - Do you think a FE mesh can be,under certain conditions, re-used for
other purposes or you feel every specific problem needs a tailored mesh?
More generally, does it makes sense to share FE meshes? If positive, for
which purposes (only educational, test replication, etc.)

Ans: I think there are only a few applications which can share meshes (e.g.
educational) but sharing of solid models and material properties (or CT
data) would be more useful.

3 - In case you share a mesh, would you like to be protected against
commerical use of you mesh? If positive, do you feel the agreement-password
system used by the NIH for the Visible human of by us for the Standardized
=46emur (http://www.cineca.it/prometeo/) would be enough?

4 - Now for the users: if this collection would be available, would you use
it? How frequently, and for which purposes (a base for my studies, I don't
want waste time meshing, educational only, etc.).

Ans: I am interested but I am skeptical about how useful they will be.

5 - Last, I promise. If you are really positive, in which format could we
store them? Patran neutral file would be general enough?

Ans: Yes. A Patran neutral file would be good. In the future perhaps STEP.

Heidi Ploeg Heidi.Ploeg@Sulzer.CH
Biomechanics Tel: +41 52 262 6826
Sulzer Orthopedics Fax: +41 52 262 0167
P.O. Box 65
CH 8404 Winterthur
Switzerland
**************************************************
=46rom: Kathryn Jane Grande

>1 - Are you willing to share with others FE meshes you had developed? Do
>you feel the mesh a shareble result or an extension of your personal
>skillness, which you don't want "sell out"? If positive, how many models
>you'd like to share, of which type and for which purposes?

Not yet - maybe AFTER my dissertation is published. And it depends on the
okay from my advisor.

One model, of the aortic root and valve. I don't want to disclose the
purpose right now.


>2 - Do you think a FE mesh can be,under certain conditions, re-used for
>other purposes or you feel every specific problem needs a tailored mesh?
>More generally, does it makes sense to share FE meshes? If positive, for
>which purposes (only educational, test replication, etc.)

I'm really not sure how useful my mesh would be to others, it's rather
tailored. I would rather see educational use only (but is there any way to
regulate that)?



>3 - In case you share a mesh, would you like to be protected against
>commerical use of you mesh? If positive, do you feel the
>agreement-password system used by the NIH for the Visible human of by us
>for the Standardized Femur (http://www.cineca.it/prometeo/) would be
>enough?

I am not familiar with that system, but I do think a password system should
be the start of the security system.


>4 - Now for the users: if this collection would be available, would you
>use it? How frequently, and for which purposes (a base for my studies, I
>don't want waste time meshing, educational only, etc.).

I would use it for educational purposes, to help myself and my students
learn about biological FE applications.


>5 - Last, I promise. If you are really positive, in which format could we
>store them? Patran neutral file would be general enough?

I agree, a general format would be best. I don't know which technique for
general storage shoudl be used.

Good luck! Please post your responses.

__________________________________________________ _________________________
Jane Grande--grande@u.washington.edu--http://weber.u.washington.edu/~grande
Bioengineering Doctoral Student, University of Washington 206-543-3159 off
Cardiothoracic Surgery, Box 356310, Seattle WA 98195-6310 206-543-0325 fax

**************************************************
=46rom: "Anthony J. Petrella" Subject: Re:
Collection of FEM meshes for the ISB website

Marco:

With regard to your enquire about FE meshes on the net:

1. I would be willing to share meshes that I developed. Although I feel a
certain amount of selfishness about this initially, I think it is becoming
more important to disseminate information in the biomechanics community. I
would primarily be working on anatomic geometries of the lower extremities.


2. I think each problem probably requires a fairly unique mesh, but one
advantage I see to sharing is that I can compare my results to yours -
maybe with totally different constitutive models - and consider how
important the material definition is for modeling an implanted knee
prosthesis, for instance.
I would probably use the meshes for research, but I could see great uses in
the classroom also.

3. I don't have any experience with this, but I would like some protection f=
rom
commercial use.

4. I would use the database for research - maybe on a regular basis.

5. Patran neutral sounds fine. Maybe IGES would be appropriate?

Regards,
Tony
--
-------------------------------------
Anthony J. Petrella
Department of Mechanical Engineering
University of Pittsburgh
petrella@sprite.me.pitt.edu

**************************************************
=46rom sherebrin@uwovax.uwo.ca:

message was posted to the wrong address and cannot be attached; he is
basically negative about sharing meshes.
**************************************************

=46rom: drl10@psu.edu (David Lemmon)

Dear Dr. Viceconti:

In response to your questions regarding sharing of FE meshes:

>1 - Are you willing to share with others FE meshes you had developed? Do
>you feel the mesh a shareble result or an extension of your personal
>skillness, which you don't want "sell out"? If positive, how many models
>you'd like to share, of which type and for which purposes?

I am willing to share meshes I have developed with qualified users that I
know will employ them appropriately and who will share results. A means of
publishing FE meshes will enable some researchers (who may not have meshing
software or skilled personnel) to conduct what may be important and
beneficial research. It also gives meshing experts the opportunity to treat
their work as "publications" in an academic setting. Of course some type of
protocol for giving credit to the source of a FE mesh will need to be
established.

>2 - Do you think a FE mesh can be,under certain conditions, re-used for
>other purposes or you feel every specific problem needs a tailored mesh?
>More generally, does it makes sense to share FE meshes? If positive, for
>which purposes (only educational, test replication, etc.)

Sharing meshes would allow many researchers to study the same problem and
be able to compare and contrast results on a level field. It is important,
however, that the mesh is associated with a specific application and set of
properties, conditions, and solution parameters. As we know, the mesh is
only one component of a finite element model. There are also materials
properties, physical properties, loading conditions, boundary conditions,
contact/interface conditions, constraint equations, solution techniques,
and many others. One mesh can be used in an infinite number of ways with
widely varying results. There are also properties of the mesh itself. What
order of elements are used, and how "regular" are the elements? Also, a
given mesh may work well with one solution program, but not work well with
another. A mesh that is initially used in a linear, small-displacement
statics problem may be unsuitable for use with nonlinear materials and
geometry. In summary, I believe there needs to be an understanding among
the users that a given mesh must be applied to the same problem with
similar properties, conditions and solution methods. When a given mesh is
used with a new set of conditions, its validity must be established for
that application. If all these are taken into consideration, I believe
meshes are shareable.

>3 - In case you share a mesh, would you like to be protected against
>commerical use of you mesh? If positive, do you feel the
>agreement-password system used by the NIH for the Visible human of by us
>for the Standardized Femur (http://www.cineca.it/prometeo/) would be
>enough?

Researchers in an academic setting should probably accept the fact that
meshes they have relinquished to the public domain may be used for
commercial purposes. Personally, I would not contribute a mesh unless its
distribution would be traceable. I would also like it to have been
published in a journal before contributing it, so there is no question as
to its source. In addition, I would attach strict guidelines as to its use
and a disclaimer of responsibility for its misuse. An agreement-password
system is one way to accomplish this.

>4 - Now for the users: if this collection would be available, would you
>use it? How frequently, and for which purposes (a base for my studies, I
>don't want waste time meshing, educational only, etc.).

I would certainly try to obtain a good, proven mesh that is available in
the public domain. In addition, it would be good if contributors also
provide the base geometric data used to develop the mesh in their
respective preprocessing software. This data might be even more useful than
the mesh itself, allowing for specific variations to be made provided that
the end user has access to and experience using that preprocessing
software.

>5 - Last, I promise. If you are really positive, in which format could we
>store them? Patran neutral file would be general enough?
If one considers a "mesh" to mean nodes coordinates and element
connectivity, any text format would do, and I'm sure Patran neutral is as
good as any. User-written format conversion programs could also be
contributed by users who use the data regularly with a specific solver.
I hope my comments are helpful, and look forward to future developments in
sharing FE meshes.
_________________________
David R. Lemmon
Systems Coordinator
The Center for Locomotion Studies
Penn State University
10 Intramural Building
University Park, PA 16802

Phone: (814) 865-1972
=46AX: (814) 863-4755
**************************************************
=46rom: Rich.Hart@tulane.edu (Rich Hart)


Hi Marco:

>Dear Collegues,
>following some public and private discussions on the possibility to share
>bio-related FE meshes through the SIB new web site, I agreed with Ton van
>den Bogert to post the following enquire:


This is a good idea.

>1 - Are you willing to share with others FE meshes you had developed? Do
>you feel the mesh a shareble result or an extension of your personal
>skillness, which you don't want "sell out"? If positive, how many models
>you'd like to share, of which type and for which purposes?

Yes. I have a 3-D mandible and perhaps a 3-D femur and 3-D canine radius
and ulna models.



>2 - Do you think a FE mesh can be,under certain conditions, re-used for
>other purposes or you feel every specific problem needs a tailored mesh?
>More generally, does it makes sense to share FE meshes? If positive, for
>which purposes (only educational, test replication, etc.)

Some can be re-used directely. Others may be a useful template for further
refinements, etc.


>3 - In case you share a mesh, would you like to be protected against
>commerical use of you mesh? If positive, do you feel the
>agreement-password system used by the NIH for the Visible human of by us
>for the Standardized Femur (http://www.cineca.it/prometeo/) would be
>enough?


I'm not worried about commercial use, except limiting liability and having
the original work referenced.

>4 - Now for the users: if this collection would be available, would you
>use it? How frequently, and for which purposes (a base for my studies, I
>don't want waste time meshing, educational only, etc.).

Yes, particulary for student projects, etc.


>5 - Last, I promise. If you are really positive, in which format could we
>store them? Patran neutral file would be general enough?

Patran neutral file may be the best compromise, particularly if we could
have public domain translators available.


>Of course, this is also a way to understand how many ISB members are
>potentially involved with FEM; so if you are one of those, please find a
>second to reply, event just to say "no time". As obivous if only five
>person are interested, we can exchange our stuff each other without web
>sites.

>Marco Viceconti


**************************************************
=46rom: chho@me.chalmers.se (Christian H=F6gfors )

Definitely interested.
Christian

**************************************************

Date: Thu, 9 May 1996 11:27:45 -0400
X-Sender: mtn2@pop.cwru.edu
Mime-Version: 1.0
To: Marco Viceconti From: mtn2@po.cwru.edu
(Michael Nilsson) Subject: Re: Collection of FEM meshes for the ISB website


>1 - Are you willing to share with others FE meshes you had developed? Do
>you feel the mesh a shareble result or an extension of your personal
>skillness, which you don't want "sell out"? If positive, how many models
>you'd like to share, of which type and for which purposes?

I am currently working on a mesh of the rabbit knee, 2D thorugh the medial
condyle, including cortical, subchondral, cancelous bone, and cartilage. It
is used for my masters thesis and as soon as i'm done, hopefully in june or
july, i would be happy to publish it on your site.


>2 - Do you think a FE mesh can be,under certain conditions, re-used for
>other purposes or you feel every specific problem needs a tailored mesh?
>More generally, does it makes sense to share FE meshes? If positive, for
>which purposes (only educational, test replication, etc.)

I am sure models can be used for test replication, but also, for example in
my model, many parametr studies can be performed, more than one individual
can accomplish for one particular work.
I weould have been very happy , if this model already would have existed.
it would have saved me a lot of time.


>3 - In case you share a mesh, would you like to be protected against
>commerical use of you mesh? If positive, do you feel the
>agreement-password system used by the NIH for the Visible human of by us
>for the Standardized Femur (http://www.cineca.it/prometeo/) would be
>enough?

it is enough..


>4 - Now for the users: if this collection would be available, would you
>use it? How frequently, and for which purposes (a base for my studies, I
>don't want waste time meshing, educational only, etc.).

>5 - Last, I promise. If you are really positive, in which format could we
>store them? Patran neutral file would be general enough?

could multiple storage be available? abaqus is used a lot also in this
filed and therefore I would suggest mutiple storage or a translation
routine.

**************************************************

--------------------------------------------------
MARCO VICECONTI (lk1boq74@icineca.cineca.it=
)
Laboratorio di Tecnologia dei Materiali tel. 39-51-6366865
Istituti Ortopedici Rizzoli fax. 39-51-63668=
63
via di barbiano 1/10, 40136 - Bologna, Italy

Tiger! Tiger! Burning bright in the forest of the night,
what immortal hand or eye could frame thy fearful symmetry?
--------------------------------------------------
Opinions expressed here do not necessarly reflect those of my employer