The problem is, you don't know they're "not structural" until you're able to accurately measure the depth of the scratch. Few, if any, of us have the requisite equipment to do so. In fact, in the big airplane world, the way we measure scratch depth most of the time on metals is to blend out the scratch, and then measure the remaining thickness ultrasonically (again, $$$ equipment). Then if the remaning thickness is within the tolerance of the engineering as designed (information we don't know without consulting Van's engineers) you're good to go. If not, a new stress analysis is required to show that positive margins still exist in that exact location at the worst-case load condition (and we don't know what the worst case is without the engineers).
In other words, a statement that damage is "not structural", without accompanying rationale of exactly how that conclusion was formed, is completely without merit.