Is it impossible to set perfectly no relative motion using /INTER/TYPE2?

gslee

Hi,

 

In figure, two 1T shells are fixed on one edge. (Gap between 2 shells is 1.1)

<?xml version="1.0" encoding="UTF-8"?>

Interface between top and bottom shells is modeled by using /INTER/TYPE2.

To set no relative motion, setting in type2 is default.

 

Concentrated load is excited at node number 7 on top shell. Load curve and scale are as follows. (The load has effect like Impact.)

 

I think that relative displacement between node 7 and 128 is perfectly zero because I use type2 default.

However, in x,y,z direction, relative displacement occurs.

 

It looks like almost zero but I want perfectly no relative motion. 

How can I set that?

 

Another question is 

 

Physically the plate is excited by impact, acceleration of plate is decayed

 

I don't use damping, so I think acceleration of some node is decayed to steady state response.

 

But acceleration of node 7 is like this. Response is increased. Why???

 

 

Thanks

 

<?xml version="1.0" encoding="UTF-8"?>

Simon Križnik

Hi,

 

FEM is a numerical technique used to determine the approximated (not exact) solution. Absolutely no relative motion is not possible due to the numerical error (truncation or round-off errors). A slight error due to round-off in one element is likely to result in appreciable error due to the cumulative effect of these errors throughout the domain. This error (0,01%) is insignificant when compared to other sources of error like discretization and formulation errors. In engineering practice, we are happy if FEM results match experiments within 10% due to material and geometry variability and experimental measuring inaccuracies.

gslee

Thank you so much.