Discussion Forum

Hi,

I want to use Linear Extrusion to map the data from Wave Optics (2D, ewfd, Mode Analysis) to a parameter in structrual mechanics (3D, solid mechanics, frequency domain). Specifically, I need to map the lowest-order electric field distribution in 2D to a surface of 3D bulky material (with the same cross section).

What I confuse is that the differentiation of linear-extructed parameters is always zero, no matter how the function inside linext(..) varys to the derivative. Please see below example that I have already tested in COMSOL:

• e.g. the electric field always varys in x direction in lowest-order electric field of the fiber: d(abs(comp.ewfd.Ex),x)!=0, so that we can non-zero the linear extrusion after the differentiation: linext(d(abs(comp.ewfd.Ex)x))!=0, which very makes sense.
• However, when we do linear extrusion before differentiation, the function: d(linext(abs(comp.ewfd.Ex)),x) turns to zero in all mapping area.

That is concerning for further simulation because somehow, the dynamics of the latter physics require differentiation of parameters that were linear-extructed from the pervious simulation result, which means the differentiation could be zero during the whole latter study (I have not proved it but really concerning this potential issue).

Besides, I am also curious how linear extrusion maps the data if the source from previous simulation has multiple modes (e.g. mapping data from mode analysis with 4 desired number of modes)?

I hope I could express the issues clearly. Feel free to ask any doubts/ unclear info of my statement and looking forwards to your reply!

Thanks, Fenghao