sumpnz
Member
Justin, even in engineering 5% differences is often threashold for saying that the difference matters. For example, if you have a complex system, say an AMRAAM, or a Tomahawk and you determine by analysis that the first bending mode of the missile will occur at, say, 200Hz, and you do a test and find that the actual first bending mode occurs at anywhere from 190-210Hz, the analysis model is considered to be a pretty darn good representation of the actual missile. If you can match within 5% on the second and third bending modes, and the first torsional mode the model is considered accurate enough to predict responses to vibration inputs, and hence calculate fatigue lives. It's not often that you can get any analysis model to match that well however.
If you have a model that matches to within 2%, and you make changes that drop that difference to 1.32% (66% reduction), your boss will yell at you for wasting time and money on a meaningless effort.
Edit: Disclamer - those frequency numbers were pulled out of the air and have no bearing on real missiles.
If you have a model that matches to within 2%, and you make changes that drop that difference to 1.32% (66% reduction), your boss will yell at you for wasting time and money on a meaningless effort.
Edit: Disclamer - those frequency numbers were pulled out of the air and have no bearing on real missiles.