Discussion Forum

1. There may be important details in your model that could be involved, so I encourage you to upload it to the forum.
2. If by "electric field norm," you are simply referring to the customary norm of the electric field, then for a typical time-domain study, the field would be real (as opposed to complex) and so its norm would simply be its absolute value (i.e., its vector magnitude, which is the square root of the sum of the squares of its x, y, and z components), at any particular point in space and time (i.e., this would in general have dependence on both space and time.)
3. Your phrasing of the question suggests that you are expecting some kind of time and/or spatial integrating, summing, or averaging of E when computing the "norm"?? But unless you add such a feature to the computations (such as via post-processing), that isn't going to happen. If you do add something like that, I encourage you to call it a different name, to avoid confusion with what Comsol calls the "norm" of the field.

Hi Robert,

Thank you for your response! I have attached my model. My physics and electrics are pretty rusty; please excuse any misunderstandings. The goal is to determine the electric field strength at a point (in my model, the point is placed in between the 2 electrodes). I'm trying to test the effects of varying pulse parameters on the electric field strength (i.e. number of pulses, total pulse duration, etc.). I had assumed that the electric field strength and electric field norm are the same as the units are the same (V/cm), is it safe to make this assumption?

1. As I said, the norm is the absolute value of the vector field. You have arranged a configuration driven by a square wave, so that your field flips its vector direction back and forth, but its absolute value doesn't change (except briefly, when it goes through the transitions.) If you replace your surface plot of ec.normE with (for example) ec.Ex, you will see that the color changes with the sign of Ex, as you step through the different times. Likewise, if you replace ec.normE with ec.Ex in your middle-of-the-model point-graph plot. That's just the x component, but you could look at y separately if you want. Or you could make an arrow surface plot of (ec.Ex,ec.Ey).
2. I don't see any deliberate time-dependent physics occurring in your model, except for your driving function at terminal #2. As such, you are in-effect just computing the equivalent of a series of static, dc (aka, constant) currents. If I am correct, then any pulse parameters you change such as duration, number of pulses, etc., (at least, in the model you created) are just plain irrelevant, since every electrical quantity in your model changes and scales linearly and in-synch at every instant, apparently depending only on the potential to terminal #2 that you have applied at any particular moment. Now, if you want the physics to actually depend on the history of the problem (such as upon how many pulses have gone before) then you had better include that in your physics equations somewhere. For example, suppose the material properties (e.g., conductivity) depend on how much current has flowed through it lately, for example, because it might heat up over time and the conductivity might depend on its temperature. But you don't have any thermal physics being modeled (no thermal study) in your problem. So I'm guessing you have a lot more to set up here, if you want this to represent something of real interest to you.

I found that conductivity changes with electric field and the data fits into a sigmoid curve (conductivity in the y-axis, and electric field in the x-axis). I'll try to incorporate that into the model.

And thank you so much for your help, it was very helpful and has saved me a lot of time!