how many parameters from the real test do you have to compare to the simulation? with just one you'll have a hard time knowing if it is a setup problem or a sim vs physical problem. but 150% is pretty far off, i'd guess setup unless there is something about your prop that causes the KE method to break down. do you have a reseller you can work with on this? to get the developers to check your application?
that being said, you haven't told us much about your setup or your physical test to make sure that they are the same. this is the key in this investigation.
then the next is to make sure that you've done everyhting you can to eliminate numerical issues. for example you're asking about the comp domain, have you confirmed that the results don't change when you make it bigger? have yo ufollowed all the recommendations in the KB for rotating regions? have you confirmed a mesh independent solution?