cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Summary: Mechanica Loop

Highlighted
Newbie

Summary: Mechanica Loop

I still haven't resolved the problem. Yes, I am in WF2. I narrowed it
down to the intersection of two features. It'll mesh with either one
suppressed, but not with both active. Nothing weird about their
intersection, and the shade in the AutoGem mesh mode doesn't coincide
with these two features, which is weird. So I'm taking all the advice
into account, but I cannot go independent mode. I'm just not
experienced with indep mode.

There are Geom checks, but its been working with those. I haven't had
much success on this model with adjusting the aspect ratio and other
mesh settings, but I only gave it a couple of tries.

I'll get there...Thanks for all the responses!

--Scott


Original Message:

I've been running various design iterations on a model, basically adding
a simple rib here and there. Everything has gone well for the last
half-dozen modifications until last night. The latest one goes into the
AutoGem sequence and loops at the "Creating Solid Elements". Attached
is from the .stt after msengine crashed last night...As you can see the
% complete goes up and down as it iterates as well as the # of elements.
Any suggestions?
Thank you.

There are C1 discon's and edge angles < 0, but I checked and all the
other models had the same issues.

Searching for existing solid elements inside the volume ...
Found 0 existing solid element(s) inside the volume.
Pre-processing boundary features ...
Optimizing boundary point locations ...
Detecting thin features on the volume ...
Pre-processing volume features ...
Creating solid elements ...
99.4% complete after 2.0 min (27063 elements).
96.9% complete after 4.0 min (14113 elements).
95.9% complete after 6.0 min (15428 elements).
[...snip]
94.5% complete after 52.0 min (16739 elements).
95.3% complete after 54.0 min (17578 elements).

11 REPLIES 11

RE: Summary: Mechanica Loop

I would advise seeing if you could have PTC tech support take a look at
your file. I know at some point in the real near future, they plan on
eliminating independent mode and I have a real concern that issues like
yours might not be readily solvable from their integrated environment.
It might provide them useful clues to another type of problem they will
face by eliminating independent mode.

Regards,
-mark



I still haven't resolved the problem. Yes, I am in WF2. I narrowed it
down to the intersection of two features. It'll mesh with either one
suppressed, but not with both active. Nothing weird about their
intersection, and the shade in the AutoGem mesh mode doesn't coincide
with these two features, which is weird. So I'm taking all the advice
into account, but I cannot go independent mode. I'm just not
experienced with indep mode.

There are Geom checks, but its been working with those. I haven't had
much success on this model with adjusting the aspect ratio and other
mesh settings, but I only gave it a couple of tries.

I'll get there...Thanks for all the responses!

--Scott


Original Message:


I've been running various design iterations on a model, basically adding
a simple rib here and there. Everything has gone well for the last
half-dozen modifications until last night. The latest one goes into the
AutoGem sequence and loops at the "Creating Solid Elements". Attached
is from the .stt after msengine crashed last night...As you can see the
% complete goes up and down as it iterates as well as the # of elements.
Any suggestions?
Thank you.

There are C1 discon's and edge angles < 0, but I checked and all the
other models had the same issues.

Searching for existing solid elements inside the volume ...
Found 0 existing solid element(s) inside the volume.
Pre-processing boundary features ...
Optimizing boundary point locations ...
Detecting thin features on the volume ...
Pre-processing volume features ...
Creating solid elements ...
99.4% complete after 2.0 min (27063 elements).
96.9% complete after 4.0 min (14113 elements).
95.9% complete after 6.0 min (15428 elements).
[...snip]
94.5% complete after 52.0 min (16739 elements).
95.3% complete after 54.0 min (17578 elements).

RE: Summary: Mechanica Loop

MessageI hope they do not get rid of independent mode or I may be forced to
use something else. There are a whole bunch of problems that cannot be
easily solved if at all in integrated mode currently. I'm not one of those
old dogs that will not change. I use and teach integrated mode most of the
time, but there are still too many times you hit the wall, and independent
mode is the only way to get the job done. My customers pay me to get the
job done.

---Jim Holst

RE: Summary: Mechanica Loop

Jim,

I see this exactly the same way ! PTC: I see also that it is a big
mistake to integrate only with Pro/E; please use Interfaces or other
ways so that we can work also with other CAD systems; Pro/M is to strong
to throw it away !!!!!

With best regards

Stefan Reul




Jim Holst schrieb:

> I hope they do not get rid of independent mode or I may be forced to
> use something else. There are a whole bunch of problems that cannot
> be easily solved if at all in integrated mode currently. I'm not one
> of those old dogs that will not change. I use and teach integrated
> mode most of the time, but there are still too many times you hit the
> wall, and independent mode is the only way to get the job done. My
> customers pay me to get the job done.
>
> ---Jim Holst
>

RE: Summary: Mechanica Loop

What Jim said.

Randy Speed
President and CEO
Speed Consulting, LLC
(972) 938-0490 ph (972) 937-2319 fax
www.speedconsulting.com

RE: Summary: Mechanica Loop

Last time I saw John Buchowski (product line manager of Mechanica) in real life he assured the audience that independent mode would not disappear before all necessary functionality is built into integreated mode. He is aware of the issues. See also http://www.ptc.com/solutions/asktheplm/mechanica/index.htm

All we need to do is make sure we make ourselves heard as long as these issues still exist, so that they do not mistake our radio silence for an agreement with the disappearance ;-)

Patrick Asselman

P.S. He also said Motion would no longer be developed and probably disappear alltogether, which was more of a shock to me at the time. Are we the only ones using this? The new integrated dynamics options are "nice" but highly insufficient for real dynamics simulations, in my opinion.

>

RE: Summary: Mechanica Loop

I've been hearing noise about retiring Independent mode (aka Razna Mechanica) for the last 10 years, and we're not there yet. One thing that should be realized is that Integrated mode has been conceptually designed as a tool for design engineers, while Independent mode is an analyst's tool, and there are some major philosophical differences between the concepts. So simply moving ALL the functionality from Independent to Integrated may hurt the design engineer folks who like the simplicity of the Integrated and won't want it complicated, and my bet is that PTC's not going to do just that.
In the end, it's going to boil down to a simple equation -- the cost of supporting Independent (software-wise) vs the revenue/profit it brings in. Once PTC feels that the cost gets substabtially greater, they may decide to simply kill the Independent and leave it's users high and dry -- business is business, happened before... Luckily, Independent mode seems to be 'frozen in time', so it doesn't look like its support costs a lot. And, secondly, I completely agree the Independent users should make as much noise as possible to let PTC know the software is still very much used, and liked, and if it gets killed those users might elect to switch to a competing analyst's tool rather than to Integrated. That's my 2 cents.
I.

p.asselman@chello.nl wrote:
Last time I saw John Buchowski (product line manager of Mechanica) in real life he assured the audience that independent mode would not disappear before all necessary functionality is built into integreated mode. He is aware of the issues. See also http://www.ptc.com/solutions/asktheplm/mechanica/index.htm

All we need to do is make sure we make ourselves heard as long as these issues still exist, so that they do not mistake our radio silence for an agreement with the disappearance ;-)

Patrick Asselman

P.S. He also said Motion would no longer be developed and probably disappear alltogether, which was more of a shock to me at the time. Are we the only ones using this? The new integrated dynamics options are "nice" but highly insufficient for real dynamics simulations, in my opinion.

>

RE: Summary: Mechanica Loop

To All,

First, I'd like to say that I have read some very good suggestions to help
solve this problem. It's nice to see there are plenty of experienced users
involved in this exploder! I would like to add some of my own observations
for everybody's consideration. First, I would like to address software
development. As a current member of the PTC/User Simulation Technical
Committee, I can tell you that efforts are ongoing to make sure any
functionality that may still be "unique" to the independent mode of
Mechanica make their way into integrated mode. At the same time, PTC can
only be aware of necessary functionality to bring into integrated mode if
they receive the feedback. Participating in the Sim TC is the most direct,
effective way to provide good feedback to PTC. Please contact me if you are
interested in becoming a member of our TC. We have discussed the original
issue that Scott posted during our most recent meetings, and development is
taking place to add more specific feedback for failed meshes. When PTC
receives models that will not mesh, they like to add them to their list of
"difficult" models to mesh, if appropriate, to help develop more robust
code.

Now an observation from the "end-user's" standpoint related to Scott's issue
(I am unaware of the history of the model Scott had to use and am making
recommendations based on the typical process, or work flow, I've seen most
often): Almost every time I have discovered meshing issues, save two that I
can think of to date, it is because I have been given something other than
native Pro/E model or a Pro/E model which was created with poor modeling
practices - I know what you're thinking if you've made the models yourself -
"Chris, are you telling me that I make poor models?". The answer maybe
"yes" - unknowingly. It is very easy to make a model that "looks" good.
But some of them can be very complicated and have unusual situations when
features are moved, resized, added or suppressed. There used to be a time
when the FEA analyst created their own geometry inside the FEA package
itself (pre-Mechanica?) - some still do and have little or no meshing
issues. Since Structure uses p-element technology, geometry taken directly
from the Pro/E model can most always be used without defeaturing. This
creates the notion that it is possible to take models created by anybody
(read; less experienced users), apply all the appropriate BCs, and then
solve the model. I work with people who came from other companies where the
most competent CAD operator was assigned to create FEA models so that clean,
usable geometry was available for FEA work. To avoid these issues, I would
recommend using two techniques: 1) create all your features in the same
order that they would be manufactured and 2) use "Model Check" to make sure
there are no "small edges", "sharp edges" and "buried features". "Small
edges" and "buried features" cause most every meshing problem I've seen. I
have found, in general, that people still create CAD models for the primary
purpose of making a paper print to give the shop floor. Good modeling
practices have still not been established, across industry as a whole, to
make sure these models are also of the proper fidelity for analytical work.
As tools like Structure and Mechanisms are used more often, this issue needs
to be addressed more completely. We as a community need to challenge our
engineering departments (management) to improve the skill set of those
making CAD models that are passed to us so we may focus on what we do best -
create the simulation model and provide accurate results, not track down
modeling issues and repair them as needed ..... well, it's nice to dream
anyway :-)

Sorry for the lengthy note, but this is a topic I've seen come up on a
regular basis, internally and externally, and felt it may help the community
to address it.

Regards,

Chris

Christopher Kaswer
Sr. Analytical Engineering Spec.
The Timken Corporation
Automotive - Powertrain
59 Field Street
Torrington, CT 06790-1008
(860) 626-2426 - voice
(860) 496-3625 - fax
-

RE: Summary: Mechanica Loop

> We as a community need to challenge our
> engineering departments (management) to improve the skill set of those
> making CAD models that are passed to us so we may focus on what we do best -
> create the simulation model and provide accurate results, not track down
> modeling issues and repair them as needed .....

I strongly disagree!

It should be the software which adapts to the needs of the users, instead of the users adapting to the needs of the software. But for some reason software developers always turn it the wrong way around (not just the PTC).

Pro/M is supposed to be easy to use because of the P-elements, but in the mean time there are other FEA codes out there with traditional meshing which give fewer problems than Mechanica. (Although I must admit I am biased here because we are still stuck in v2001 for Mechanica while we have a recent version of Abaqus).

Best regards,
Patrick Asselman

RE: Summary: Mechanica Loop

Ken,
It's Dassault, not PTC, that bought ABAQUS...

Kenneth E Taylor <-> wrote:




The situation might resolve itself as PTC have bought ABAQUS so us analysts
may have the best of both worlds in the future


Ken T