Issues with PML and Varying Angle of Incidence in HFSS
时间：03-30
整理：3721RD
点击：
Hello,
I want to model a periodic unit cell in HFSS with a plane wave excitation in the XZ plane with varying angles of incidence. I have started by modeling an air box (20mm x 20mm x 187mm) with PML boundaries on the +/- z sides of the box and Master/Slave boundaries on the other sides. For the top and bottom faces (+/- y), I set the phase delay to 0. For the +/- x faces, I set the phase delay as a function of frequency: (2pi*freq*a/c)*cos(theta), where a is the x length of the box, c is the speed of light and theta is the angle of incidence measured from +x to -x. For the plane wave, I use k = (cos(theta), 0, sin(theta)) and E0 = (sin(theta), 0, -cos(theta)). I run the parametric study with theta = 0, 45, 90, 135, 180 deg for 3GHz. The unit cell would be observed between 8 and 10GHz.
The simulations complete, but the resulting fields are not as expected. It appears that k and S are parallel, but E0 and the E field are not in the same direction, and H is not as it should be. For theta=0, k starts along +x and E0 starts along -z, but after the run E is along +z and H is along -y. The values are also not constant along x, but are constant along z. The same goes for theta=180; k starts along -x and E0 starts along +z, but after the run E is along -z and H is along -y. For theta=90, E is sinusoidal in the XZ plane, H is sinusoidal in the YZ plane and P-vectors are along +z. All three increase in magnitude along +z.
Any ideas as to what might be going wrong here? For a simple box of air, I would expect uniform fields in the directions of excitation. Please help!
Many thanks!
I want to model a periodic unit cell in HFSS with a plane wave excitation in the XZ plane with varying angles of incidence. I have started by modeling an air box (20mm x 20mm x 187mm) with PML boundaries on the +/- z sides of the box and Master/Slave boundaries on the other sides. For the top and bottom faces (+/- y), I set the phase delay to 0. For the +/- x faces, I set the phase delay as a function of frequency: (2pi*freq*a/c)*cos(theta), where a is the x length of the box, c is the speed of light and theta is the angle of incidence measured from +x to -x. For the plane wave, I use k = (cos(theta), 0, sin(theta)) and E0 = (sin(theta), 0, -cos(theta)). I run the parametric study with theta = 0, 45, 90, 135, 180 deg for 3GHz. The unit cell would be observed between 8 and 10GHz.
The simulations complete, but the resulting fields are not as expected. It appears that k and S are parallel, but E0 and the E field are not in the same direction, and H is not as it should be. For theta=0, k starts along +x and E0 starts along -z, but after the run E is along +z and H is along -y. The values are also not constant along x, but are constant along z. The same goes for theta=180; k starts along -x and E0 starts along +z, but after the run E is along -z and H is along -y. For theta=90, E is sinusoidal in the XZ plane, H is sinusoidal in the YZ plane and P-vectors are along +z. All three increase in magnitude along +z.
Any ideas as to what might be going wrong here? For a simple box of air, I would expect uniform fields in the directions of excitation. Please help!
Many thanks!