Ivar KJELBERG
COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
13 août 2009, 09:09 UTC−4
Hi
In structural you select the Point Settings Load tab and add a force "F".
Note that you cannot add masses in bulk models (only with beams and shells).
It's just as easy to add a force along an edge in the Boundary Settings Load entry, but you must normalise to the length. sometimes you vant to have an uneven distributed load then you can write F*6*s*(1-s), as a parabolic distribution with maximum at the centre (note that (int(0,1,s*(1-s))=1/6, therefore the "6*"
Furthermore, in 2D the results are given per length (thickness) and by default this is 0.01[m] hence you have a multiplication by 100 of your forces if you are not carefull.Always check "reaction forces" and "tensor derived forces" for discrepancies before using the data of a calculation
Pls note that in 3D its slightly more complex as you have a double integrand and the progression of the two s's are not warrantied, as COMSOL says in their doc
Hope it helps
Ivar
Hi
In structural you select the Point Settings Load tab and add a force "F".
Note that you cannot add masses in bulk models (only with beams and shells).
It's just as easy to add a force along an edge in the Boundary Settings Load entry, but you must normalise to the length. sometimes you vant to have an uneven distributed load then you can write F*6*s*(1-s), as a parabolic distribution with maximum at the centre (note that (int(0,1,s*(1-s))=1/6, therefore the "6*"
Furthermore, in 2D the results are given per length (thickness) and by default this is 0.01[m] hence you have a multiplication by 100 of your forces if you are not carefull.Always check "reaction forces" and "tensor derived forces" for discrepancies before using the data of a calculation
Pls note that in 3D its slightly more complex as you have a double integrand and the progression of the two s's are not warrantied, as COMSOL says in their doc
Hope it helps
Ivar
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
13 août 2009, 09:39 UTC−4
If i have a bulk type, disk shaped, with two small anchar at the edge, what should i do then?
If i have a bulk type, disk shaped, with two small anchar at the edge, what should i do then?
Ivar KJELBERG
COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
13 août 2009, 09:53 UTC−4
Hi there
If you are using 3D solid structural "smsld" mode: then you use Point Settings Load and you add your force(s) in x,y,z or the projection of the correct components, or you define your own coordinate system ...
for edges start by having an even distribution ("s" concept is for next step, if really needed)
Now if you are in the time domain you must set a condition F*(t>0)*(t<1) on the time to turn "on/off" your force. I propose that you implement this by smoothening it with one of the heaviside function (i.e. flsmhs) , otherwise you will see your solver take long.
See the "guide" documentation, search for "heaviside", these are COMSOL functions, can be used from matlab too (if needed and if the comsol link is running)
there are a few examples of time dependance impulse in the doc somewhere, I would suggest to start by them, to locate them run the MS indexer on yor doc, or just open the three structural pdf docs and use the pdf search. These things are slighly complex (in the beginning) so just do not give up, once you master it its really great, and you will not find such possibilities in other programmes
good luck
Ivar
Hi there
If you are using 3D solid structural "smsld" mode: then you use Point Settings Load and you add your force(s) in x,y,z or the projection of the correct components, or you define your own coordinate system ...
for edges start by having an even distribution ("s" concept is for next step, if really needed)
Now if you are in the time domain you must set a condition F*(t>0)*(t