In this presentation we look at five steps any engineer can do to make their FEA results more trustworthy and useful, based on CAPINC’s years of experience helping customers get better returns on their Simulation investments.

1. Shuvom Ghose
800-424-2255 x1110
shuvom@capinc.com
www.capinc.com
The Summer of Simulation #3:
5 things you
should do before
showing your
FEA results to
anyone

2. is ’s leading
reseller of SolidWorks CAD and
tools, and
Stratasys 3D printers.

3. Slides prepared by CAPINC Engineer:
Who makes work
with your company’s specific
Shuvom Ghose

4. © CAPINC
• There will always be people who
doubt your FEA results (and FEA
use in general).
• Don’t give the doubters any
more ammunition.
• Do all you can to solidify your
FEA results before showing
them to people.

5. 1. Ask the right question.
2. Do a sanity check. Any check.
3. Make sure the mesh doesn’t matter.
4. Test a second failure condition.
5. Document your assumptions.
© CAPINC

6. Titanium:
Ti-6AL-4V
Fixed at 4 corners
8000 lbs.
Total upward
force
If you were given this aircraft
bracket to analyze, what would you
ask of FEA first?
Many folks might ask: “What is the maximum stress in the
bracket? Within, say, +/- 10 percent?”
(Bracket model courtesy of user “optimal_aj” from GrabCAD.com)

7. Is
Ti-6AL-4V
the closest alloy?
Is Fixed a realistic
constraint for
these bolts?
Is 8000 lbs.
really the right
force? What if
the load is at an
angle?
But if you ask about the absolute
stress, now EVERY aspect of your
study is called into question:
Is this mesh fine
enough for this
complex geometry?

8. Worse than that, now you’ll get into 8000 lbs.?
the “Percentage Game” with your
doubters. You’ll say:
“Given all these questions, the max
stress is probably accurate within
15%.”
Fixed?
Ti-6AL-4V?
Mesh fine
enough?
“15 percent? HA! You’ll need
accuracy better than +/- 5 percent to
convince ME!”

9. Fixed
Upward force
But instead of
maximum stress, what
if you asked:
“We’ve got two
competing bracket
designs. Which one is
better?”
© CAPINC

10. Then, with just a 1
minute FEA study, you
learn that the lower
bracket is:
77% Heavier
16% Worse
displacement in
the pull direction
© CAPINC
(Same load, mesh settings,
fixtures, material)

11. 77% Heavier
16% Worse
displacement
And the best part is, all those
doubts we had with the absolute
study fall away, since they are
the same for both relative
bracket studies!
This means doubters can’t object
to the results!
© CAPINC
“Um…”
(Same load, mesh settings,
fixtures, material)

12. 77% heavier
16% worse Z disp.
160% heavier
50% better Z disp.
216% heavier
34% better Z disp.

13. At CAPINC, we call this the “A vs. B comparison”. Use it to make
your FEA cycles faster, since EVERY design decision can be boiled
down to a choice of: “Do we do this? Or this? Or neither?”
© CAPINC

14. 1. Ask the right question.
2. Do a sanity check. Any
check.
3. Make sure the mesh doesn’t
matter.
4. Test a second failure
condition.
5. Document your assumptions.
© CAPINC

15. Returning to absolute
numbers, with 8000 lbs.
upward force on the lugs, the
maximum reported stress in
the part is:
3.6e5 psi.
© CAPINC
Is that way too much? Way too
little? In the ballpark? If we’ve
never seen a shape like this, how
can we tell?

16. Is your max stress
right next to one of
your restrained,
rigidly fixed faces?
Then the number
is fake.
The fixed face distorts the stress around it, since one
end of those mesh nodes can’t move at all, leading to
more stress on the element. To get a more accurate
number, we’d have to change the fixed restraint, or
model the bolts and plate they’re attached to.

17. © CAPINC
Using a Design Insight plot (found
in all levels of SolidWorks
Simulation), we can see that the
back legs carry most of the stress
when this bracket is pulled upward.
Blue= loaded

18. 8000 lbs.
Total upward force
Stresses ~
4e4 psi?
© CAPINC
Using an Iso Clipping plot (found in
all levels of SolidWorks Simulation),
we can see that the stress in those
loaded back legs is around 4e4 psi.
Is that way too much? Not enough?
How can we tell?

19. © CAPINC
And if you look at those back legs and
squint a little, that back leg is sort of a
cylinder about 0.5” diameter:

20. © CAPINC
8000 lbs.
So what would happen if we
took a cylinder of that size
and put 8000 lbs. of load on
it?

21. 8000 lbs. upward force
on test coupon
Stresses = 4e4
psi in 0.5” dia
section
8000 lbs. Total upward force
on bracket
Stresses ~
4e4 psi in
~0.5” dia
leg
With results in the same magnitude (e4 psi), we can conclude that our
initial stress results are… NOT INSANE. That’s all we can tell.

22. 8000 lbs. upward force
on test coupon
4e4 psi
© CAPINC
“You’re estimating the
stresses in your part
with a cylinder under
tension?
I can do what with a
simple hand calculation!
Why do we need some
fancy, expensive FEA
package to accomplish
that!?!
Back in my day…”

23. © CAPINC

24. On-line Hoop Stress Calculator:
http://www.engineersedge.com/calculators/hoop-stress.
htm
© CAPINC

25. © CAPINC

26. • What’s the MOST this number could be?
• What’s the LEAST?
• What is the range of your plot’s legend?
• Are deformations still “Elastic”, i.e. are
they under 0.2% strain? (Linear solvers
assume all deformations are elastic.)
© CAPINC

27. 1. Ask the right question.
2. Do a sanity check. Any
check.
3. Make sure the mesh doesn’t
matter.
4. Test a second failure
condition.
5. Document your assumptions.
© CAPINC

28. 31,000
mesh
elements
© CAPINC
How many mesh
elements are
‘enough’?
126,000
mesh
elements

29. One big mesh element is like one big spring, pretty stiff…
© CAPINC
Many mesh elements are like many small springs,
added all together, they bend more while staying in
the elastic zone…

30. © CAPINC
Tip displace:
5.14e-8 in
Tip displace:
5.27e-8 in

31. © CAPINC
Increasing mesh density
Result
‘Real’
answer
What ‘should’ happen as
you increase mesh
density

32. © CAPINC
Mesh elements
Z displacement,
(in)

33. © CAPINC
Mesh elements
Z displacement,
(in)
Mesh elements

34. Ask the right question!
Mesh accuracy doesn’t
matter if the second A
vs. B design is heavier
AND worse in
displacement!
© CAPINC
77% Heavier
16% Worse Z
Displacement
1 minute FEA study

35. © CAPINC
Mesh elements
Your result number

36. 1. Ask the right question.
2. Do a sanity check. Any
check.
3. Make sure the mesh doesn’t
matter.
4. Test a second failure
condition.
5. Document your assumptions.
© CAPINC

37. © CAPINC
8500 lbs.?

38. © CAPINC
5000 in-lbs.

39. © CAPINC

40.  Things Breaking
 Things Bending too much
 Things Overheating
 Things Shaking at Nat. Freqs.
 Long, Slender Things Buckling
 Fatigue Failure
© CAPINC
Break
Bend
Burn
Buzz
Buckle
Fat
BBBBBFat!
We checked for:

41. © CAPINC

42. © CAPINC

43. SolidWorks Simulation Professional:
Break and Bend Buzz
© CAPINC
Burn
Fatigue Optimize
Buckle

44. 1. Ask the right question.
2. Do a sanity check. Any
check.
3. Make sure the mesh doesn’t
matter.
4. Test a second failure
condition.
5. Document your assumptions.
© CAPINC
BBBBBFat!

45. Restraints
Material Properties
© CAPINC
Loads
Mesh
Contact Conditions
Geometry
Simplification

46. Restraints
Material Properties
© CAPINC
Loads
Mesh
Contact Conditions
Geometry
Simplification

47. © CAPINC
• Restraints
• Mesh • Geometry
Simplification
• Contact Conditions
• Loads
• Material Properties
BIGGER
assumptions
Smaller
assumptions

48. Appendix F: Assumptions
• Bracket variant 142A was restrained with fixed faces
© CAPINC
at each of its bolt corners.

49. Section 2: Assumptions (BEFORE results!)
• Bracket variant 142A was restrained with fixed faces
(green) at each of its bolt corners:
© CAPINC

50. Section 2: Assumptions
• Bracket variant 142A was restrained with bolt faces only allowed
rotational motion (green cylinders) and bottom faces only allowed
sliding motion (orange planes):
© CAPINC

51. 1. Ask the right question.
2. Do a sanity check. Any
check.
3. Make sure the mesh doesn’t
matter.
4. Test a second failure
condition.
5. Document your assumptions.
© CAPINC
BBBBBFat!

52. © CAPINC
• 3 rules to do material
selection right
• 10 designs by hand,
100 by FEA
• 5 things to do before
anyone sees your FEA
results
http://www.capinc.com/events/webinars/recorded-webinars

53. If you want cutting edge or
, contact
(salesinfo@capinc.com)
If you want more info on the FEA
concepts presented, contact:
(shuvom@capinc.com)