engineering fundamentals Discussion Forum
Directory | Career | News | Standards | Industrial | SpecSearch®
Home Membership Magazines Forum Search Member Calculators






List Recent Topics | Start a New Topic

<< Previous Message No. 78 Next >>
Author: canucme
Time: 01/19/00 18:14 PST
This is a reply to message no. 67 by chrisoreilly
Reply | Original Message | New Topic | List Topics | List Messages on This Topic
Current Topic:
Area of Geometrical Shape
Here's a potentially quick and definitely dirty idea:

If you only have to change one variable to generate your different areas (which sounds like the situation for each of your 2 cases -- width constant, and height constant), AND, the area of your shape changes smoothly with this variable, then you can find your areas with a quick approximation scheme.

The idea is to use ProE to calculate the area of your shape for a few values of the controlling variable.  Pick values that are near the extremes of the variable, and values that are in between.  For example, if your controlling variable is called WIDTH, and it can range from 10 to 100, then try picking the values WIDTH = {10, 20, 30, 40, ..., 100}.

You will now have a handful of datapoints where you know the area (from ProE) for a given value of your controlling dimension.  Plot these against each other on an x-y plot in Excel (or equivalent). If your area is indeed smoothly varying, then you should see a nice smooth looking curve connecting your data points.  Use Excel to curve-fit this function.  Congrats: you now have a function that will give you the area for ANY value of your controlling dimension.  You can use this function to crank out 1500 measurements in no time, and all it cost you was using ProE to measure 10 to 20 distinct cases (and not 1500).

Note that this will only work if your area function is smoothly varying.  Otherwise, your curve-fit won't look too good.  You can improve the curve fit by taking more "data" with ProE (evaluating more points), but eventually this approach loses its appeal if you have to take 100 datapoints or more.  After all, you said that you didn't want to use ProE to crank out each area separately.

Give this idea some thought.  Perhaps it could work for you.  As a lazy person myself, I sure would try to make this method work.  Good luck.

[ List Replies to This Message Only ]

Home  Membership  About Us  Privacy  Disclaimer  Contact  Advertise

Copyright © 2018 eFunda, Inc.