This past summer I spent a huge amount of time developing a UV-cured coating. Now that the lab work is done, we've started scaling it up.
You might initially think that the scale-up should be easy since there are none of the nasty surface/volume ratio scaling that normally occurs in reactors and all the implied heat/mass transfer/buildup issues that go with it. This is simple exposing a thin wet coating to a UV lamp for a short period of time.
But even in something as simple as this, there are a mess of potential problems. First off, the curing lamps are not even the same. Different manufacturers, different bulbs (and output spectra), different wattages and therefore different heat outputs. UV curing is more than just absorption of a UV photon. That is merely the initiation step. The remaining polymerization, crosslinking and termination steps are all subject to temperature conditions that (generally) do not plague the photoinitiation.
The coating methods are not the same either - in the lab I used a notched bar, but the pilot line has a type of roll coater. The low viscosity of the formulation that worked so well in the lab became problematic for the coater, so all the plans of looking at different coat weights ended up as something for the next round. We took what we could get.
Nonetheless, it appears that the formulation cured quite well, quite a bit faster than in the lab, although it does have a gamut of testing to run. Thank goodness we didn't have to look into other curing options: multiple exposures, different reflectors, bulb distance,...