Menu

Discussion Forum

New Discussion
Filter

Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Apply a point load on arbitrary point

Posted Jun 14, 2011, 10:57 a.m. EDT Version 4.1 3 Replies

Bruno Rango
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi, I have a beam element and I want to apply a point load. How can I apply the load on an arbitrary point of the geometry?

Thanks,

Bruno.
3 Replies Last Post Jul 5, 2011, 1:38 a.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 14, 2011, 4:33 p.m. EDT
Hi

First of all Point loads, as well as Edge loads in 3D are singularities, hence are NOT good practice, but sometimes you can use the results if you do NOT analyse in detail the local stress concentration.

If you do not have a point I'm not sure how to make one, what I normally do is to use a surface load (edge load in 2D) and distribute the force over the area/edge length i.e. with a Gaussian. This gives you a concentrated load, but no pure singularities, if the Gaussian is larger than a few mesh elements. THe CoG of the Gaussian can then be moved over the surface/edge with a parametric sweep
Hi First of all Point loads, as well as Edge loads in 3D are singularities, hence are NOT good practice, but sometimes you can use the results if you do NOT analyse in detail the local stress concentration. If you do not have a point I'm not sure how to make one, what I normally do is to use a surface load (edge load in 2D) and distribute the force over the area/edge length i.e. with a Gaussian. This gives you a concentrated load, but no pure singularities, if the Gaussian is larger than a few mesh elements. THe CoG of the Gaussian can then be moved over the surface/edge with a parametric sweep
Bruno Rango
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 27, 2011, 10:17 a.m. EDT
Thank you for your answer.

I tried to do what you said but I don´t know how to use the function (interpolation) in the edge load label. I tried with: 1000[N/m]*int1(x[1/m]) but it don´t works. How should I do?

Other question: Is it posible to get a point graph without defing a node in that point?

Thank you!

Bruno.
Thank you for your answer. I tried to do what you said but I don´t know how to use the function (interpolation) in the edge load label. I tried with: 1000[N/m]*int1(x[1/m]) but it don´t works. How should I do? Other question: Is it posible to get a point graph without defing a node in that point? Thank you! Bruno.
Yannick Loosli
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 5, 2011, 1:38 a.m. EDT
Hi Ivar,

This Gaussian distribution trick to simulate point load is great!


I have however a question:
How can I really control the total load acting on my beam?

A simple example is a beam of meter fixed at both ends with a point load actin in the middle (let’s say 1000N).

I define a normalized gaussien by
FCT= (1/sqrt(2*pi*c^2))*exp(-(x-b)^2/(2*c^2))

X = curvilinear coordinate
b = point where the peak of distribution is located
c = std deviation of the distribution

The integral of FCT = 1. However if the std deviation is small enough, the peak is larger than one. Meaning that the “point load” acting on the beam center have force larger than 1000N (if F =1000*FCT). So the sum of all the “point loads” created by the Gaussian distribution is then substantially higher than the wished value.
I have plotted the reaction force vs c: the curve increase up to maximum (1000N per fixed point!) and then decrease….

I’m totally lost with this outcome. I’m sure the element size has smth to do with this. But my principal problem is still how to define properly my Gaussian in function of the element size to obtain a displacement field similar to the one obtained by COMSOL with a point load.

How to you proceed to circumvent this issue?


I have attached my model so that you can have a closer look.

Thank you for your feedback,

Yannick
Hi Ivar, This Gaussian distribution trick to simulate point load is great! I have however a question: How can I really control the total load acting on my beam? A simple example is a beam of meter fixed at both ends with a point load actin in the middle (let’s say 1000N). I define a normalized gaussien by FCT= (1/sqrt(2*pi*c^2))*exp(-(x-b)^2/(2*c^2)) X = curvilinear coordinate b = point where the peak of distribution is located c = std deviation of the distribution The integral of FCT = 1. However if the std deviation is small enough, the peak is larger than one. Meaning that the “point load” acting on the beam center have force larger than 1000N (if F =1000*FCT). So the sum of all the “point loads” created by the Gaussian distribution is then substantially higher than the wished value. I have plotted the reaction force vs c: the curve increase up to maximum (1000N per fixed point!) and then decrease…. I’m totally lost with this outcome. I’m sure the element size has smth to do with this. But my principal problem is still how to define properly my Gaussian in function of the element size to obtain a displacement field similar to the one obtained by COMSOL with a point load. How to you proceed to circumvent this issue? I have attached my model so that you can have a closer look. Thank you for your feedback, Yannick

Attachments:

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content