Faulty Synthetic eddy method with checkpoints

If you restart a simulation with an SEM boundary, it will produce exactly the same velocity profiles over a period of time as it did from the previous restart. This produces very regular, unrealistic flow patterns downstream of the inlet (see attached). It is particularly noticeably over short restarts.

This has issues for generating large eddies, as the eddy lengthscale is limited by the simulation time between checkpoint restarts. (A necessary part of running on ARCHER)