When post processing strain energy, NX allows you to display averaged or
un-averaged results. If I want to calculate total strain energy, which value
should I use?
Solution
To calculate total strain energy, you would want to use the un-avaraged strain
energy results. Using averaged strain energy makes no sense in this case. This
is easy to prove using a simple test model and comparing the NX results to the
hand calculation. Take the simple case of a prismatic bar measuring 1 x 1 x
36" clamped at one end and loaded axially with 1000 lbf. Assuming a Young's
modulus of 2.9E7, we get the following for total strain energy:
U = ((F^2)(l))/((2)(A)(E)) = (1000 lbf)^2(36 in)/((2)(1 in)^2(2.9E)) = 0.62
lbf-in
Solving this model in NX and summing the strain energy we get the following:
Unaveraged = 0.0620 lbf-in
Averaged = 0.905 lbf-in
This clearly shows that the un-averaged result is correct.
To avoid this confusion it is probably best to post process the elemental
strain energy result set instead of the element-nodal results. This will
provide a single total strain energy value for each element, avoiding
the averaging issue entirely.
Notes