Optimizing Feed Network of an antenna for S11 and Farfield Gain using CST

[denizduran]

I am trying to create an Hybrid Metasurface Antenna Array. My S11 and farfield pattern did not come out as desired. I know this problem is due to my feed network not being precisely correct. My feed system consists of 3 wilkinson power dividers that are not symmetrical. Predicting the phase difference between the paths are very hard. How can I optimize for S11<-10db for frequency range 4.2GHz-6.2GHz? I have also attached my CST file and the antenna design article I am trying to replicate. Any help will be appreciated.

 

[vfone]

Your simulated far field directivity gain (at 5.1GHz, where your simulated S11 = -15dB) is much lower than in the paper (4.6dBi vs about 11dBi)
So the diference between your S11 vs frequency and paper S11 may not be only the feeds.
Could be also differences related to array characteristics, especially mutual coupling between elements or substrate differences, characteristics that may affect the return loss of each antenna element.
However, due to changes in surface waves, those array types using grounded elements, gets higher mutual coupling between antenna elements.