Dear Javier,
Thanks for clarifying your approach.
It is difficult to answer your questions because you have supplied a lot of information and are using more than just NREL-developed codes and models. Neither do I have the time to assess all of all your steps in detail. So, instead of answering your direct questions, I’ll try to give provide general information.
In general,the linear model should follow the nonlinear well for small perturbations about the linearization point. If your linear model is not tracking the nonlinear model well for small perturbations, I suspect there is a problem somewhere in the solution or comparison. It may help to simplify the model as much as possible in order to track down a problem (e.g., start by eliminating the generator azimuth DOF).
MBC is useful for rotors with 3 or more blades. MBC does not eliminate the periodicity, it simply transforms the rotating states of a model into the nonrotating frame (a linear model of mixed rotating and nonrotating states can then be represented in the nonrotating frame). For example, a 1p-cyclic load in the rotating frame contributes to 0p and 3p loading in the nonrotating frame. Much of the periodicity of the mixed frame linear matrices is captured by the MBC transform. Further azimuth-averaging of the lightly-periodic MBC-transormed matrices has been shown to be quite adequate, such that direct periodic techniques are likely not needed for 3-bladed rotors. See, for example, this reference: Stol A, Moll HG, Bir G, and Namik H. “A Comparison of Multi-Blade Coordinate Transformation and Direct Periodic Technicques for Wind Turbine Control Design.” 47th AIAA Aerospace Sciences Meeting Including The New Horizons Forum and Aerospace Exhibition, 5–8 January 2009, Orlando, FL. AIAA-2009-479. January 2009.
MBC is very useful to the design of independent blade pitch controllers. Again, see the reference above or review the papers it references.
Perhaps someone with more experience with the application of linear models and/or MBC can provide further guidance.
Best regards,