[建模]CST MWS微带线仿真时候的疑问？

请教一下，微带线仿真时，可以使用open 作为边界条件么？ cVr+Wp7K#|  
我使用open边界条件时，发现S12在低频处出现明显损耗，求解原因。 ovi^bNQ  
u
K ,W  

图片:截图01.jpg

:V_UJ3xf  
.. U9 mK^  
mwI7[I2q  

未注册仅能浏览部分内容，查看全部内容及附件请先 登录 或 注册

可以。 ;K%/sIIke  
Z&P\}mm  
我又来算命了：楼主你能确定端口模式是正确的吗？

我又运行了一些模型，结果还是这样子。 Re %dNxJ=  
hefang老大 ，帮忙看下，模型里面哪个地方有问题啊！[attachment=70090][attachment=70091]

仿真使用的模型： +oq<}CNr{  
http://pan.baidu.com/s/1dDEW8jJ

heliu0017: 6(HJYa  
仿真使用的模型： rO1.8KKJ  
http://pan.baidu.com/s/1dDEW8jJ jqoU;u`  

:!,.c$M  

? 5hwz  
这次先说结论吧。CST只建议在仿真远场属性的时候才使用open (add space)边界，如果只是观察S参数的话应该尽量使用闭合边界。不过，这并不表示仿真微带线不可以使用开放边界。唯一需要注意的是波导端口被建立在边界内部的时候被称为internal port，在绝大多数情况下这种波导端口必须添加shielding才能得到“正确”的结果。 Aq/wa6^%  
IRI<no  
楼主的模型： 9x9E+DG#(  

图片:Model.jpg

-ktYS(8&  
^}GR!990  
open边界，6个方向设置了2的surrounding space。 \<y#$:4r<8  
a+\s0Qo<  
使用新的网格模式，只设置一个对称面，其余不做任何修改，得到的S参数： a\I`:RO=<Z  

图片:S.jpg

R"PO@v  
uRw%`J4H  
S21： ;Y(~'KF  

图片:S21.jpg

$I/RN  
UX'tdB !A  
S21的情况和楼主类似。原因就是internal port在0频不能正确工作。给两个端口都添加Electric shielding：
2Xfy?U  

图片:Shielding.jpg

q.lh  
An !i  
得到的S参数： Je` w/Hl/U  

图片:S with Shielding.jpg

/bqJ6$  
h3A|nd>\  
可以看到S11在0频处比原先低了将近40 dB。可想而知S21也肯定有改善： z7lbb*
Xe  

图片:S21 with Shielding.jpg

;nf}O87~  
V0:db  
假设使用闭合边界会怎样呢？把边界条件设置为conducting wall，取消surrounding space： 2f-Or/v  

图片:Model Closed Boundary.jpg

'}NH$ KA  
k${F7I(Tb  
波导端口不需要electricl shielding，得到的S参数： QOF'SEq"k  

图片:S Closed.jpg

9, 792b  
2{#=Ygb0  
0频处和上面一致，接近20 GHz出现谐振。S21： >[g'i+{  

图片:S21 Closed.jpg


N/AP8  
5{8,+ Z  
低频处没有楼主的问题。楼主可以自行计算闭合边界是否在19.47 GHz处引发了腔体谐振（box resonance）。 *Kw/ilI  
DRi/<  

---

!myF_cv}'  
结论：仿真非远场问题尽量使用闭合边界。如果一定要用开放边界，内部波导端口需要设置电/磁屏蔽。 SCMZ-^b  
TC'^O0aZ_  
CST MWS帮助文件《Waveguide Port Overview》： {*4Z9.2c*  
H0mDs7  
sk`RaDq@;  
Ensure shielding DZ|/#- k  
rHgrCMW  
8m)E~6  
Especially for inner waveguide ports is important to apply an electric or magnetic shielding to ensure a consistent integration of the waveguide boundary into the calculation domain. In case of a hexahedral mesh, port shieldings are modeled as a tube in direction of field imprint, with a length of two grid cells. For tetrahedral meshes this extrusion is not necessary. Please not that the boundary condition can be influenced by a port shielding. Also, conductors might be short-circuited by an electric shield. Generally, a electric boundary condition or PEC material  is given higher priority compared to a magnetic shielding. Furthermore, the solver detects whether the port contains an outer region (as in the corners of a coaxial line). If this is the case, shielding the port will not have any effect and is therefore not applied. kc:2ID&
 
oRCD8b?  
CST MWS帮助文件《Waveguide Port》： (%;D& ~%o  
|n&EbOmgf  
Ensure shielding: If the port mode domain is not already electrically shielded (as e.g. in a coaxial waveguide) this box can be checked to activate a special treatment of the boundary:electric: the boundary of the waveguide port is treated as a PEC shielding frame. {pB9T3ry]  
magnetic: the boundary of the waveguide port is treated as a PMC shielding frame. Note that conductors in the port shield will not be removed and might cut through the PMC shield.Leaving this box unchecked will apply the default treatment for port boundaries of the used solver. ,[Z;"wE  
CST MWS帮助文件《Special Solver Parameters - General》： rYr.mX  
Ensure shielding for all ports: If this check button is enabled, a shielding frame of PEC or PMC type surrounding each waveguide port region will be added. This option causes higher reflections at ports. Therefore, it should only be used if a calculation has become unstable. Note that this setting will be ignored for already electrically shielded waveguides (e.g. coaxial waveguides). Leave this button unchecked to apply the default treatment for waveguide port boundaries. cNqw(\rr  
2E3x=  
收兵回营！ g/soop\:  

学习~~~~多谢hefang大大。