Discussion Forum

Hi

that is (no frequency change) because you add a "load" in Newtons, and no "mass" on your points/plates.
If you add a new thickness to the central small plate you will see a difference, and of you add a point mass you will see a change too, but a "point mass" is no "point load".

Take care to separate a "point load" coming from i.e. a "body force" suc as due to gravity acting on a mass, attached to your memebrane (with it's own defined volume, density hence an integrated own mass)

Basically forces/loads are ignored for an eigenfrqeuency analysis, only effects acting on "iomega*rho" will do

The easiest (as I cannot find Point mass in Shell physics bizarre, forgooten ?) is to make a simple beam model and add a point mass and then look at the equation used, and transpose to shell, something like

Weak constraits:
-mass*(u*test(u)+v*test(v)+w*test(w))*shell.iomega^2

--
Good luck
Ivar

Hi,

Thanks a lot Ivar. I am extremely sorry to disturb you once again with this simple question.

I am clearly able to put the weak constraints in comsol 3.5a and got the result. However, I am not able to know how to implement weak constraints in comsol 4.0a. Will you please spend some of your time to guide me how to do this ?

Thanks in advance,

Kiran

Hi

forgotten about that, have to check, as the weak constraints are not all 100% up to date I seem to remember, I'm already in 4.1 ;) will try to have a look tonight

--
Good luck
Ivar

Hi,

This should be straighforward in 4.0a: Right-click Shell and select Points>Weak Contribution to add a Weak Contribution node where you can enter the weak expresion.

The likely reason that you have not found it is that you have to activate the settings to show more options.
If you select "Show More Options" from the Model Builder window's View menu (the downward pointing triangle at the top of the Model Builder window) you get access to some advanced settings and additional properties such as weak contributions and weak constraints.

Best regards,
Magnus Ringh, COMSOL

Hi Ivar,

Once again, Thanks a lot for your help.

Now, I am able to solve my problem. I have just changed the thickness of that surafce to add the mass of the additional body by doing some calculation, which gave me a correct result. By increasing the thickness of the surface, we are indirectly adding some weight on the plate, that's why it gave a nice result.

With Regards,

Kiran

Hi

As Borger said above you must activate the settings, furthermore, if I remember right (I do not have 4.0 here) you could only add one dof at the time, so you need to add three times u, v, w separately. I fellto remember his was corrected in 4.0a.

Now there is another way to cheat ;) you add a truss or beam physics, on the point (or if not possible on a single point add it on a line ending at te point). Then set the E,nu,rho such that he additional (line) stiffening is neglected, and add a point mass. One thing you need to ensure that you have the same dependent variable names:

_u_=(u,v,w) and NOT the default _u2_=(u2,v2,w2)

another thing, if you mix 3D and shell/beam be aware that you need to knit in the rotational dof of the shell beam truss to the 3D solid, which by default does not have any rigid body rotations defined. Futhermore, there seem to be an issue on the materials, you cannot, today up to 4.1 mix a material on a 3D and a 2D physics, you need to call in the material data twice, the error messages are somewhat cryptic to understand, I havnt had time to fully catch what is wrong, but something is not fully clean, therefore I tend to avoid mixing 3D and beam/shell/truss for the time being

I understood from support that this is being looked after for a future release

--
Good luck
Ivar

Hi

finally I found back my 4.0a ;) here is an example, the frequency changes but I havent checked the results

I added the mass as 5 (default kg) in the weak point expression, but you can as well use a parameter mp=5[kg] and replace the "5" by "mp" in the weak equation

And its not a "Shell point load" but a "point mass" for eigenfrequecy analysis ;)

--
Good luck
Ivar

Hi Ivar, since you have mentioned about point mass earlier, kindly I would like to ask your help on how to model distributed point masses in Comsol. I want to add the masses across a plate to increase the irregularities in the system.

Thanks,
Khairul

Hi

in COMSOL you BC field entries accept functions, not only variables, and in fact in COMSOL most "variables" are in fact implicit "fields" depending on the spatial and temporal arguments (x,y,z,t) or r ....

You can add, on a boundary (i.e. surface in 3D) a local mass density by adding in boolean equations for the localisation of the mass such as

rho*((x-x0)^2+(y-y0)^2)<R0)

such an approach is extensively used in the level set method in the chemistry model

--
Good luck
Ivar

Ok, Thank Ivar for the quick response . I will try it first.

Ok, Thank Ivar for the quick response . I will try it first.

Where exactly I need to put the equation what exactly the variable means? Sorry for troubling you but I'm new with comsol interface. Normally I used lsdyna for modelling.I tried here and there but nothing works. Is it in the weak constraint?

I'm using comsol 4.2 and my model construct in 3-D dimension. I found that there is 'point mass' option for beam and trusses in 2-D modelling but not in 3-D.

Hi

in the plate example on the top thread model you shoud add a "added mass"
Loads are ignored for eigenfrequency analysis

With the added mass use the same Total mass in all three directions. You could also add a surface mass density

Check all the right click menus and submenus on your nodes (note that these changes dynamically when you change main physics settings and or solver settings, so you might need to reclick over the same node to discover new items, if you have changed some of the basic settings of your physics

--
Good luck
Ivar

Thanks Ivar, I've tried the method. I noticed the added mass is across the surface. But what I want to find is point masses located on top of the plate.
Now I used poinwise constraint. I've selected several point and changed the expression to the one at the top of the thread.
-mass*(u*test(u)+v*test(v)+w*test(w))*pzd.omega^2. Is this correct?.

Plus, in plotting the graph, how to change energy density (J/m^3) into energy (J)? It will be very helpful if you can show me how you do it as well.

Thanks,
Khairul.

Hi

you can go into the equations, but normally you can add point masses, at least in 4.,2a
And you have the RIGID Boundary

Sorry I see in 3D PZD there is no pointwise added mass (because its a severe "singularity" and causes often solving issues and loss of local precision)

your equation looks correct for the same mass in all three directions and in frequency domain solver mode, to be added as a Weak Pointwise Constraint. You can work by similarity from the equation view, indeed

--
Good luck
Ivar

I think I had my answer now..
Thanks Ivar, all the best to you too..

Hi Ivar,

Regarding the total energy.
I've checked the forum and in my point of view, none have clearly describe about pzd.W_int (total energy)

my questions are:
1)how to use this variable in Comsol 4.2a?
2) what mechanical energy flux means? is it the same as kinetic energy?
3)In frequency domain analysis, if i add the strain energy (integrated over the volume) and mechanical energy (integrated over the volume and divided by 2*pi*freq), is it the same as total energy?
4) Is it possible to find strain energy of the pzd material, because so far I can't find the way..

Appreciated if you can reply. thanks

Hi

I agree I cannot really find the exact formula for W_tot in the doc, so one need to turn on the equation view and dig into the equations. What I do know is that it is a global variable integrated over the full model

--
Good luck
Ivar