NONLINEAR EXPERIMENTAL RESPONSE OF PRIMARY-SECONDARYSYSTEMS UNDER REPEATED SEISMIC EXCITATION

摘要

Repeated ground motion was observed in several recent earthquakes in California, Taiwan, Romania,Turkey and Japan. In these earthquakes the main structure or Primary system (P-system) may remainundamaged on the first or second excitation in the seismic sequence, yet exhibit a nonlinear response onfurther excitations. It is important to study the response of the supported non-structural equipment in suchevents. Non-structural components or Secondary systems (S-system) are crucial for the functioning oflifeline facilities such as hospitals and power plants. The most critical case of concern of the coupledPrimary-Secondary system (PS-system) is the near tuned case; causing quasi-resonance. Detuned responsecan be of concern in cases where at least one of the two subsystems yields in the seismic event. Inprevious analytical and experimental studies on coupled PS-system linear and nonlinear response; little orno concern was given to the effect of torsional coupling as a result of P-system stiffness eccentricity, masseccentricity of the P-system or S-system location eccentricity relative to the Centre of Mass (CM) of the Psystem.These parameters are present in most real structures due to functionality requirements and spaceoptimization. In this shake-table research program, the nonlinear torsionally coupled PS-system responseis characterized. The research program also serves as a means of calibrating the analytical solution of thetransient response. The PS-system model is excited by a synthetic seismic sequence. The P-system and Ssystemare modeled as lumped masses. P-system eccentricity is induced through static or dynamiccoupling; whereas S-system eccentricity is induced through varying the location of attachment relative tothe P-system CM. Torsional interaction and response amplification of the coupled PS-system arecharacterized and quantified. Experimental findings show that the torsional yielding of the primary systemhas significant implications on the de-amplification of near tuned secondary system response. The locationof the S-system affects its peak response amplification. A detailed analytical model is developed tosimulate the experimental behavior.