I did an attempt in the past to obtain the behavior in the subject. I’m not even sure if the theoretical approach was correct.
Basically I would test FAST in a hardware-in-the-loop environment where the external bladed dll should serve as a proxy to the real world sensors. I didn’t had enough time to investigate this deeply, but I remember I was stuck on how to feed aerodyn wind with real wind data.
Anyway, supposing that a FAST run is quicker than real-time and the execution of the external DLL code relatively quicker than the integration step, the simulation should always be ahead of the real clock, then a delay loop in the DLL will slow it and sync to real.
Is this correct?