Comparison of fatigue effects in simulated and actual earthquakes

Abstract

One of the unresolved aspects of simulated earthquake testing concerns the amount of low-cycle fatigue produced by a series of single-frequency-vibration tests compared to random-motion excitation. The recommended sinebeat vibration input using 5 cyc/beat at each measured equipment natural frequency is considered conservative because it can produce more fatigue damage than the most severe earthquake being postulated.

A computer-aided analysis has been made of simple mass-spring-damper systems representing the uncoupled structure and equipment response to well-known acceleration-time histories available for both the El Centro-1940 and San Fernando-1971 earthquakes. In the proposed testmachine simulation, the comparable sine-beat-vibration input was adjusted to duplicate the building-floor response at the base of the equipment, which was represented as a series of oscillators geometrically spaced in the 1 to 25-Hz range of damaging earthquake effects. Both large- and small-strain ranges have been compared, but these values are assumed for convenience and cannot be used as actual seismic effects in specific pieces of equipment.