Discussion Forum

Hi,

I'm having a lot of trouble defining a load function. The funtion I want to model is:

latex.codecogs.com/gif.latex?F_%7B0%7D%5Chat%7Bz%7D%5Ccdot%20%5Cleft%5C%7B%5Cbegin%7Bmatrix%7D%20-%5Cint%20%5Cleft%5C%7B%28x-x%27%29%5E2%20+%20%28y-y%27%29%5E2%20+%20%28z-z%27%29%5E2%5Cright%5C%7D%5E%7B%5Cfrac%7B3%7D%7B4%7D%7D%20dx%20dy%20%26%20z%3Ez_%7Bmax%7D%26%20%5C%5C%20%5Cint%20%5Cleft%5C%7B%28x-x%27%29%5E2%20+%20%28y-y%27%29%5E2%20+%20%28z-z%27%29%5E2%5Cright%5C%7D%5E%7B-1%7D%20dx%20dy%20%26%20z%3Cz_%7Bmax%7D%20%5Cend%7Bmatrix%7D%5Cright.

(sorry for the length of the link, it would seem the forum does not support tags for images)

The input for the load is:

if(z>zmax, -F_0*intop1( 1/( (dest(x)-x)^2 +(dest(y)-y)^2 +(dest(z)-z)^2 )), F_0*intop1( ( (dest(x)-x)^2 +(dest(y)-y)^2 +(dest(z)-z)^2 )^(3/4)))

in the z-direction (x and y both 0) where the integral (intop1) is defined over the source boundary (a surface)

When I solve with this force it acts like it is not there. I've tried a lot of test functions:

F_0*intop1(((x-dest(x))^2+(y-dest(y))^2+(z-dest(z))^2)^(-1))
works (model responds to the presence of this force)
F_0*intop1((((x-dest(x))[1/m])^2+((y-dest(y))[1/m])^2+((z-dest(z))[1/m])^2)^(-1))
works too
F_0*intop1((((x-dest(x))[1/m])^2+((y-dest(y))[1/m])^2+((z-dest(z))[1/m])^2)^(1))
doesn't work (no response)
F_el*intop1(([m^-2](x-dest(x))^2+[m^-2](y-dest(y))^2+[m^-2](z-dest(z))^2)^(3/2))
doesn't work

I think the problem lies somewhere in how comsol handles units. If the unit is too exotic, it treats the load as if it were not there. It also does not give an error. I find this very illogical, but it's the best exlanation I could come up with.

Please help me!