Abstract
It is common to establish shock and vibration design and/or test criteria for equipment mounted on a structure by computing a conservative upper bound for the spectrum of the dynamic load induced response of the structure based upon predicted or measured spectra at various points. This task is usually accomplished by one of five procedures, namely, the computation from the available spectra of a simple envelope, a normal tolerance limit, a distribution-free tolerance limit, an empirical tolerance limit, or a normal prediction limit. These five procedures are reviewed and illustrated using the power spectra computed from vibration signals measured during the lift-off of a large launch vehicle at 12 locations in a structural region where equipment will be mounted. The results are compared and the merits and liabilities of the various procedures are discussed.