Reliability of Real-time Hybrid Simulation for Performance-based Design of Structures with Rate-dependent Seismic Devices

摘要

Real-time hybrid simulation divides a structural system into experimental substructures and analytical substructures. The experimental substructures are physically tested in the laboratory and the analytical substructures are modelled by a finite element program. It thus enables the complete structural system to be considered during the test and provides an economic and efficient experiment technique to evaluate seismic performance of structures with rate-dependent devices and their performance-based design procedure. Real-time hybrid simulation however often involves actuator control error when the desired displacements are applied to experimental substructures. This brings research needs for reliability analysis of experiment results so that real-time hybrid simulation can be taken full advantage for structural engineering research. This paper presents a discussion on existing variables that may be used for reliability analysis of real-time hybrid simulation results, including maximum tracking error, root-mean-square error, tracking indicator and energy error. The potential application of these variables for reliability analysis are evaluated through real-time hybrid simulation of a two-story four-bay steel moment resisting frame with magneto-rheological fluid dampers in passive-on mode. The experimental results show that these existing variables have both advantages and disadvantages and the energy error has the best potential. Research towards reliability analysis using existing variables is demonstrated to be necessary to advance current real-time hybrid simulation practice.