HyperSizer picks the candidate design with MAXIMUM weight.

[iankim]

Hi to all who read this message...

I attached two sizing results and I don't understand them.
There are 25(5x5) candidate designs as I set 5 permutations for 2 variables -
Top Face and Web.
And HyperSizer recommends 25th design with maximum weight
even though it has 0.5277 MOS.
I thought it would pick the design with the weight somewhere between min and
max weight and the MOS would be slightly over 0.00.
What should be the problem???

I chose orthotropic material for this project w/o layup definition.

P.S. One more question is that how I can select a couple of failure criteria
among so many criteria in Failure tab.
In failure tab, there are many failure criteria and I want to choose just one or two
among them for my optimization calculation so that minimum MOS would be the
output of calculation w/ not all of criteria but just those criteria I chose.

[Phil]

It may be that your sizing variables are course enough so that reducing any variable would cause a negative margin.  For example, your thicknesses range from 6.39 to 12 mm with 5 permutations.  So your possible thicknesses are:

6.39
7.79
9.20
10.60
12.00

It is possible that going from a thickness of 12 to 10.6 could cause a negative margin.  Look on the failure tab to determine your minimum margin to see what could be driving this.

For turning on or off failures, simply go to the Failure tab and click on the hatched boxes in the Available Failure Analyses frame to turn failure modes on and click again to turn them off.  You can turn on or off all failure analyses under Limit or Ultimate but clicking the Toggle button.  You can return to default values by clicking the "Analysis Options..." button and selecting "Get Database Defaults..." from the resulting pop-up menu.

Please let me know if this helps.

[iankim]

Thanks for your speedy and kind reply, Phil.

After reading your message, I increased the upper bound from 12 to 120
and permutation from 5 to 50 for another testing.
And it showed the design with max weight(120) again and MOS is now over 100.0.

I suppose something is wrong with initial setting or in setup form and it affected
the optimization routine.

Can you guide me to how to solve this trouble?

[Phil]

Can you check the a and b lengths on the FBD tab of the Sizing form?  Are a and b lengths very low?  There is a check in HyperSizer for uniaxial panels that ensures that the b width (also called the Y span) is greater than 3 times the stiffener spacing.  Try increasing your a and b widths to something greater than 500 mm and re-size to see if you get a more reasonable result.