Discussion Forum

Hi

Well for me, if E is the electric field, in Comsol notation for EC/ES physics we have:

Ex_... = -dV/dx , Ey_... = -dV/dy , Ez_... = -dV/dz,

So if you want the elelctric field gradient we need one more deivative, and Exx .. is not in the list

so I do not see any other way than

Exx_... = d(Ex_...,x) , Exy_... = d(Ex_...,y) ...

(replace _... by your application id)

by using the special operator d( variable_name , space_variable_name )
see Guide.pdf p163 and thereon (V3.5a)

and from thereon it should be straight forward, or have I missed an important point ?

Have fun Comsoling
Ivar

Yes, it is correct. But in Comsol E is -Vx, that mean electric field is the negative of potential gradient. I don't now how to write the gradient of electric field. The electric field in comsol have the notation normE, and the gradient of electric field to x then it's normEx?

Thanks!!

Doesn't diff() work it for you? As an example, diff(normE,x) would return you the x-derivative of electric field's norm.

By the way, in case it is needed, normE is only an absolute scalar value and not the Electric field itself.

Hi

Sorry for the confusion, you are perfectly right, (I was at home and had no direct access to Comsol to check, and I'm the one claiming one should always verify what one is doing ;)

Hope I have corrected it now

Ivar

So,
In Comsol language:
If E on x is Ex_...=-Vx then the derivatives of Ex is Exx_... = d(Ex_...,x) . I don't understend how to write correct this derivatives in Comsol language, because I don’t have succes to obtain results.

With diff(normE,x) it not return the x-derivative of electric field's norm. I don't now where it's the problem.

Thanks!

Hi

Simply do a surface plot your X field, something like

Ex_...

then adapt by typing your formula to

d(Ex_...,X)

check the upper/lower cas of x, it depends on you application mode and and you variable names

its d(,) and not diff(,) (as in Maple), check you guide.pdf on the special operators, thre are several that one should learn to use in there

Good luck
Ivar

Ostahie,

Adding to what Ivar said, check out what names your electric field variables are pre-assigned in the specific application mode you are working. I usually look it out in the post-processing window for selecting which field value to plot. For e.g., if you are using 2D boundary mode analysis module for TM waves, then there is a pre-defined tangential electric field variable called tEy_rfwb. If you want to take its derivative along y-direction, just write diff(tEy_rfwb,y)

The reason why you might be getting error is that there is no such variable as normE. I mentioned it only as a case example and not because it actually exists in any of the application modes.

Hope this helps.

Shakeeb,

normE_emdc=sqrt(abs(Ex_emdc)^2+abs(Ey_emdc)^2), where Ex_emdc=-Vx an Ey_emdc=-Vy.

If the derivatives of normE_emdc on x is dEx=diff(normE_emdc,x) , and the derivatives of normE_emdc on y is dEy=diff(normE_emdc,y) then gradient of electric field is gradE=sqrt(dEx^2+dEy^2).
My reasoning is correct?

Thanks!!!

So far as I see, the equation of gradE you specify is just the 'norm' of the gradient of electric field's 'norm' :)

If this is the quantity that interests you in the specific circumstance you are working in then that is fine. Otherwise, I don't think it is a standard practice to call this formulation of gradE as the gradient of electric field.

Please remember that gradient operator works on scalar fields to produce a vector field which provides the measure of how the 'scalar' field varies in different spatial directions. Since electric field is a vector field, what generally seems more meaningful to me is to specify the gradient of various components of electric field.

I am trying to plot ∇E^2.
For Electric field distribution of my design (in r, phi,z) coordinates, i used the following equation:

sqrt(es2.Er*es2.Er+es2.Ephi*es2.Ephi+es2.Ez*es2.Ez)

How do i plot the 'Gradient of Eelctric field square' if that is what it should be called i.e. ∇E^2.

The quantity should have units of V^2/m^3.

Let me know.