Fix problem where ConvertToDiffractionMDWorkspace was failing with an error
when source workspace had units of dSpacing
Problem occurred in ConvertToMDMinMaxLocal specifically which was called as
child algorithm from ConvertToDiffractionMDWorkspace
Have populated DIFC with default value of nan to align it with behaviour for
l2. Monitors and spectra without detectors get this default
This stops the code hitting a validation check in dSpacing::validateParams
which was checking for non-zero DIFC.
Would be nice to tidy up the logic around monitors so that don't pass nan
parameters into unit classes and expect nans back but just doing cautious
fix for now so can sort out for release