There is no single answer to this and it depends on what you are attempting to demonstrate.
- The efficiency of coupling a Gaussian beam coupling into the entire structure?
- The efficiency of the structure once the signal is in the photonic crystal structure.
If it is the Gaussian coupling you place the source outside of the structure and the transmission you are measuring is what you get. However, you should realize that a Gaussian beam will not couple into this structure with anywhere near 100% efficiency so you should not be expecting a Transmission of 1.
If it is just to demonstrate that the structure efficiently propagates the signal you could leave your setup as is and then use 2 observation lines. One is your transmission line left as is while the other should be the same and located towards the beginning of the structure. What you could then do is export the power spectra from the two and manually normalize the output spectrum by the input spectrum. I have done a quick test of this and you get normalized powers around 1 after some post processing.