Firstly, I have no experience on EBG simulation. However, you should note the following comments:
1. In your cst model, you didn't discretized your geometry. In the "Mesh-Global mesh properties-Lines per wavelength", and the local mesh properties of PEC and Rogers materials should be set up.
2. I don't know the reason why you set only one port. Generally, 2 ports are needed.
3. For boundary conditions, PEC+PMC for periodic boundary is right. So why use PEC for Zmin? I think it should be open. And a proper value for the distance between the port and the side of the cell should be used.
I know this is a very old post, but just for reference in this thread here's my suggestion.
Ripples can be due to accuracy setting in CST transient solver, if you set your accuracy to -40 or -50 you'll get better results (excitation signal will decay better causing less problem when doing Fourier transform)
Another solution might be looking at your mesh settings for increased accuracy (adaptive mesh refinement).