Subject: Re: FEA theory needed for FEA software?
From: Paul Skoczylas
Date: Thu, 12 Dec 1996 08:14:38 -0700
Time to throw in my two cents...
Anyone who woud try using ANSYS or any other FEA software without
testing a few simple cases for which analytical solutions exist is
asking for trouble. (And the simple cases should use all the same
element types your real problem uses.) I've taken a course on finite
element theory, and I think a little time reading the manual and a few
days working on simple problems with known solutions would be more
useful to getting good results out of ANSYS (or other FEA).
A computer is a tool in the hands of a good engineer, and is dangerous
in the hands of a bad one...
All of which reminds me of a story a co-worker told me. Here at C-FER (
http://www.cfer.ualberta.ca ) we produce a software package to assist in
design, analysis and troubleshooting of progessing cavity pumps
installed in oil production wells. The software is not designed to work
in gas wells, although two-phase flow is considered, up to a certain
extent. The program is smart enough to know when there's too much gas
for it to calculate a valid result. A while back, a pump vendor called
my coworker to complain that our program wouldn't work in a case where
their in-house software would. It turned out that the case in question
had a very high gas-oil ratio, and their program almost certainly gave
incorrect results. The scary thing is that if a purchaser had trusted
the vendor (or the vendor's software) to do the calculations for them,
they would have lost hundreds of thousands of dollars on that one
application.
My point: You must know the limitations of whatever software you are
using, because often the software doesn't know its own limitations.
-Paul
Subject: hydraulic/vacuum manifold problem
From: allen@macnet.vpharm.com (david allen)
Date: Thu, 12 Dec 1996 10:21:39 -0500
I'm having a problem with a manifold system. I have 4 lines (1/16" I.D.)
attached, through valves, to a waste container which is under vacuum.
these lines are used to draw waste solutions from discreete containers.
when all 4 lines are pulling liquid, they work OK. when one of the lines
goes dry, the others loose suction and/or start backflowing. using check
valves will stop the backflow but won't prevent loss of suction and
flow. is there a way to keep the other lines flowing when one goes dry?
I've thought of trying a 4 channel peristaltic pump, but, that's a major
redesign of the system and I'm not sure the manufacturer would support it.
tia
da
Subject: Re: FEA theory needed for FEA software?
From: jmsche01@starbase.spd.louisville.edu (James M. Scherer)
Date: 12 Dec 1996 15:05:23 GMT
Lee Harding (lee.harding@autodesk.com) wrote:
-trimech: AutoFEA 3D from ANSYS has been doing this for a while in the
-Autodesk Mechanical Desktop™. So enough of the commercials.
-
-To address Mr. Rizzo's posting: We continue to chase that proverbial fleece
-because we are engineers. As such, we are not in the habit of concluding
-that a problem can't be solved simply because it hasn't been solved.
-
-To re-pose the question in a completely inappropriate form: Do you need to
-know the theories of friction and cleavage-plane slipping to use a pencil
-on the back of an envelope? No, but it doesn't hurt.
We chase that fleece because computers are a hell of a lot cheaper than
engineers.
Anyone can drag a pencil across a sheet of paper and see that it creates a
line. Structural analysis using finite element methods is not that simple.
Does ANSYS claim that it has software that will automatically produce a
valid and accurate analysis of any problem? Does the entire structure have
to be modeled in the computer? Do all the constraints and forces have to be
entered? Just because you have solid models of the components doesn't mean
software can create a reasonable finite element model of the system.
You would need a mighty huge computer to model some problems. If you knew
enough FE theory, you could create a simplified representation of the problem
that would actually fit in the computer.
I'm not knocking ANSYS; they have some powerful code, but software is a
tool, not a substitute for engineers who know what they are doing. Even
ANSYS has its limits. This approach is fine for routine and inherently
compact problems, but beyond that, you're up the proverbial creek without
a paddle.
Mitch
Subject: TO ALL EQUIPMENT MANUFACTURERS
From: Herbert
Date: Fri, 13 Dec 1996 02:19:28 +0800
A newly setup engineering company in China is looking for supppliers of
medical, scientific and inspection equipments. Interested parties please
contact hawk@hknet.com and send your product information to
Pansource Industrial Limited
Room 1407, Chang Chen Building
4, Si You Xin Road, Guangzhou
Guangdong Province, China
Postal Code : 510600
Tel : 020-87612888-1407
Look forward to hearing from you soon.
Subject: Re: hydraulic/vacuum manifold problem
From: Charles Coughran
Date: Mon, 09 Dec 1996 18:55:35 -0800
david allen wrote:
>
> I'm having a problem with a manifold system. I have 4 lines (1/16" I.D.)
> attached, through valves, to a waste container which is under vacuum.
> these lines are used to draw waste solutions from discreete containers.
> when all 4 lines are pulling liquid, they work OK. when one of the lines
> goes dry, the others loose suction and/or start backflowing. using check
> valves will stop the backflow but won't prevent loss of suction and
> flow. is there a way to keep the other lines flowing when one goes dry?
> I've thought of trying a 4 channel peristaltic pump, but, that's a major
> redesign of the system and I'm not sure the manufacturer would support it.
Depending on the details of the application, you can design a reverse
acting foot valve with a buoyant ball. When liquid covers the foot valve,
the ball floats away from the intake. When there is no liquid, the ball
seals the intake. Take a look at air purge valves for ideas.
--
Charles Coughran
ccoughran@ucsd.edu
Subject: Re: FW: Warning of a New Internet Virus
From: "Gerald P. Roston"
Date: Thu, 12 Dec 1996 13:07:39 -0500
Katarina Weinberg wrote:
>
> This is critical and valuable Information. Please pass it on.
>
> Subject: A New Internet Virus
>
[crap deleted]
This is an old hoax and is simply not true.
--
Gerry Roston
Cybernet System Corporation, 727 Airport Blvd
Ann Arbor, MI 48108
(313) 668-2567 FAX (313) 668-8780
gerry@cybernet.com
The opinions expressed are mine and do not reflect the official position
of Cybernet Systems
Liberty is to faction what air is to fire, an aliment without which it
instantly expires. But it could not be a less folly to abolish liberty,
which is essential to political life, because it nourishes faction than
it would be to wish the annihilation of air, which is essential to
animal life, because it imparts to fire its destructive agency. James
Madison
Subject: Re: SEARCH ENGINES
From: colonel@engsoc.carleton.ca (Rob Russell)
Date: 12 Dec 1996 19:45:21 GMT
In Brain Fart <32B06D0D.1CD0@vtt.fi>, Tapio Leino wrote:
: How could there be? There is no possibility to save any
: keywords with the HTML-page! This, however, should be the
: job of the producer of the page and nobody else! The
: inventor of the old HTML-document system and Mosaic just
: didn't see any need for this then!
Ever heard of META tags? They do just that. To put Mechanical
Engineering and Engineering as keywords for a page, insert the following
tag within your segment:
Furthermore, if you don't include the following tag, search engines will
use the first few lines of text from the page as a description of the
page:
If you'd like an example of a page that implements these tags properly, go
to http://www.cses.carleton.ca
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Robert Russell \ / Rob_Russell@Ottawa.Com
Carleton University x--o8o--x 2nd Year, Mech & Aero Eng.
http://wabakimi.carleton.ca/~rrussell
Director of Publications, Carleton Student Engineering Society
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
"No harm was done to my mental well-being." -- erdem@engsoc
Subject: Re: hvac for small house
From: waiwhite@zip.com.au (Ian A. White)
Date: Thu, 12 Dec 1996 20:46:50 GMT
In article <32a8aae0.869765913@newshost.ccsi.com>,
larry@ccsi.com (Larry Marshall) wrote:
> I am building a small one bedroom house (640 sq ft) in central TX.
> Central A/C & Heat seems like overkill. Am considering a wood stove
> for heating but would like recomendations for A/C
Certainly look at a reverse cycle air conditioner (heat pump). You
will need to consider your local winter temperatures, and depending on
how cold it gets, you might need backup heating. Having said this,
the unit I have installed advises you to clear snow from the outdoor
unit (it is a split system) to maintain operating efficiency, so it is
fully intended for operation in such environments. Even if you do
install backup heating, from an energy consumption point of view, it
would still pay you to use the heat pump for your normal heating, only
going to the backup system when conditions demand it. Why? Well with
an up to 300% efficiency it is hard to go past it.
I have just replaced my 15 year old heat pump with a new one, and the
Australian Standard ratings for cooling are conducted at 7 degrees C
outside temperature and a 20 degrees C inside temperature. The
manufacturer also states that the operating range is from -10 deg C to
46 deg C. The newer control systems are also far superior. For me
one overriding factor is safety. There is a much reduced fire risk
compared to other radiant and combustion heating systems.
From an energy, and cost point of view, local figures indicate that
wood stove heating is the least efficient form of heating. Even the
newer "improved" ones have major problems with pollution. Some of our
worst winter pollution is as a direct result of the use of wood
heaters. It is now difficult to get permission to install one in
certain local government areas.
Regards,
Ian A. White, CPEng
waiwhite@zip.com.au
WAI Engineering
Sydney 2000
Australia
Junk e-mail will be returned, as is, to the sender or the sender's host system.
Subject: Re: hydraulic/vacuum manifold problem
From: dave lawson
Date: Thu, 12 Dec 1996 13:18:02 -0800
david allen wrote:
>
> I'm having a problem with a manifold system. I have 4 lines (1/16" I.D.)
> attached, through valves, to a waste container which is under vacuum.
> these lines are used to draw waste solutions from discreete containers.
> when all 4 lines are pulling liquid, they work OK. when one of the lines
> goes dry, the others loose suction and/or start backflowing. using check
> valves will stop the backflow but won't prevent loss of suction and
> flow. is there a way to keep the other lines flowing when one goes dry?
> I've thought of trying a 4 channel peristaltic pump, but, that's a major
> redesign of the system and I'm not sure the manufacturer would support it.
>
> tia
>
> da
Your problem is that the pressure drop that occurs across the fluid
between containers is lost when air fills one of the lines. Since the
resistance to flow of air is so very much lower than for the liquid,
only air is drawn through and the pressure drop does not exist.
You have two obvious choices;
a) Increase the vacuum source flow (more cfm) so that when only air is
being drawn, so much will be drawn that sufficient pressure drop will
occur.
b) put an orifice in each line so that when one fills with air the
pressure drop is still maintained. This will significantly reduce fluid
flow rates.
Of course, if you raised the source containers above the collecting
container then the flow would not reverse at any time, but then you
wouldn't have needed the pump in the first place .
If you used a weighted float system in the source containers that might
work also. Similar to the feed control system used on furnace mounted
drum humidifiers or carburetors but working with a reversed control
sense. When the float lifts based on sufficient fluid level, the valve
opens to admit fluid. When the level drops, the valve closes. As long
as the valve is located below the lower fluid level, you draw only fluid
and maintain a constant pressure drop.
Dave
Subject: Re: active vibration control
From: John Thomas
Date: Thu, 12 Dec 1996 19:57:45 -0800
Diane Carlson wrote:
>
> Hi. I'm a sophomore at the University of Illinois Urbana-Champaign. I
> currently belong to the Society of Women Engineers and I am part of Team
> Tech. In Team Tech, the group is given an interdisciplinary problem by
> industry to solve. We solve the problem, write a paper, and present it
> at Engineering Open House here at school and then again at SWE
> nationals. This years project is to develop an active vibration control
> system for industry standard PC 3.5" disk drives. I was wondering if
> anyone knew of good papers, journals, or home pages with info on AVC or
> if anyone could send me more info on this topic. Any help you could give
> me would be much appreciated.There is a good collection of papers from an ASME proceedings on active
noise and vibration control from 1994(?).
Subject: Re: FEM / FEA for Composites?
From: billmcc
Date: Thu, 12 Dec 1996 15:43:45 -0800
Christopher Wright wrote:
>
> In article <32AEDDD0.33ED@compuserve.com>, 100015.132@compuserve.com wrote:
>
> >We are looking for finite element modeling (FEM) and finite element
> >analysis software systems which can handle or are specialized for fiber
> >reinforced plastics / composites. What is the most used system? What
> >runs on PCs, what on Apple Macintosh? What other platforms are there?
>
> COSMOS/M has this capability and runs on Windoze and the Mac. The Mac
> installation is good for a vertical market program--a little quirky for us
> die-hard Mac bigots, but 100% better that ANSYS on WinNT, which I've also
> used. The COSMOS modeller is better than ANSYS, but the post processing
> not as good. You might also want to check out LapFEA which is the old
> MSC/pal.
>
> Christopher Wright P.E. |"They couldn't hit an elephant from
> chrisw@skypoint.com | this distance" (last words of Gen.
> ___________________________| John Sedgwick, Spotsylvania 1864)
> http://www.skypoint.com/subscribers/chrisw
I would add a note of caution on the above: If you are at all interested
in the prediction of buckling loads (or are interested in ply by ply
analysis) none of the above code in my experience will do a good job.
All of them have been demonstrated to miss low order modes (the ones
that count). The only two general purpose commercial codes that do a
pretty good job (but don't just believe what you see - confirm it) on
most of the low order modes are Mark's and ABAQUS. A pretty good Mesher,
I hear, is FEAmap. I hear HYPERmesh works pretty well too but a little
more difficult to use. I am not aware of any codes that will give you
the ply-by-ply analysis.
Bill McEachern
Subject: Re: FEA theory needed for FEA software?
From: Pierre Hubsch
Date: Mon, 9 Dec 1996 09:08:53 GMT
The answer is yes,
You need at least some basic understanding of how FEA works to
get meaningful results. That said, with some of the better software
packages, you need very little theoretical understanding of what's
going on, because they include error-control tools (err, actually,
to apply these tools, it is better if you know what you are doing).
This observations is however resticted to standard problems.
If a non-standard situation arrises (which happens all the time
if you don't know what you are doing), you will need some insight.
That will get you to the book shelve and you will learn some of
the FE theory - EASY. There are also plenty of things that standard
packages won't do, so you have to code them yourself, which also
calles for some of the theory. You may be surprised though how little
you actually need.
An ech haett dat och op letzenburgich kennen schreiwen ...
Pierre
______________
Pierre Hubsch Dept of Bio-Medical Engineering
Technion - Israel Institut of Technology Haifa 32000
Israel tel 792-4-8294124
