Is your complete system, from generation to load, completely underground?
If not (which is the most likely answer), then there is always a chance of a lightning strike on an overhead conductor, that will transfer its way into the cables that feed your transformer, hence it is completely valid to test an indoor transformer with a lightning impulse waveshape. The voltage may be slightly under the operating point for spark gaps or surge arrestors at the start of the cable, but high enough to cause a problem in the insulation that you don't know about, because you didn't test it. And the rate of rise for a lightning impulse is much faster than that of a switching impulse, which can be very onerous, depending on your insulation structure.
You may well be performing an interturn test and a bulk insulation to earth and the other windings test, but that is a 50 or 60 Hz test, which is fine for a 50 or 60 Hz applied voltage. But during the life of a transformer, the insulation will not only see this power frequency, but also lightning shape and switching shape impulses, which are affected greatly by the capacitance of the insulation structure. This may produce overvoltages in certain areas of the insulation, which will not be overstressed at power frequency, and may cause failures that you don't expect because you assumed that they are not relevant.
Ausphil