The 14 th World Conference on Earthquake Engineering October 12-17, 2008, Beijing, China A NEW HYBRID RELIABILITY ANALYSIS METHOD: THE DESIGN POINT - RESPONSE SURFACE - SIMULATION METHOD M. Barbato 1 , Q. Gu 2 and J.P. Conte 3 1 Assistant Professor, Department of Civil & Environmental Engineering, Louisiana State University, 3531 Patrick F. Taylor Hall, Nicholson Extension, Baton Rouge, Louisiana 70803, USA. 2 Engineer, AMEC Geomatrix Consultants Inc. 510 Superior Ave., Suit 200, Newport Beach, CA, 92663, USA 3 Professor, Department of Structural Engineering, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093 - 0085, USA. Email: mbarbato@lsu.edu, qgu@ucsd.edu, jpconte@ucsd.edu ABSTRACT : Classical reliability methods such as First- and Second-Order Reliability Methods (FORM and SORM) have been important breakthroughs toward feasible and reliable integration of probabilistic information and uncertainty analysis into advanced design methods and modern design codes. These methods have been successfully used in solving challenging reliability problems. Nevertheless, caution should be used in the applications of these methods since their limitations and shortcomings in terms of applicability and accuracy are known and documented. Current research trends highlight the importance of structural reliability analysis methodologies that are able to provide improved estimates of the failure probability without excessive increase in computational cost when compared with ordinary FORM/SORM analyses. In this work, a new hybrid reliability analysis method, denoted as Design Point – Response Surface – Simulation (DP-RS-Sim) method is proposed and illustrated. This method innovatively combines the design point (DP) search used in FORM/SORM analyses with the response surface method and appropriate simulation techniques. The need for this combination has emerged from the results obtained through visualization of the limit state surfaces (LSSs) typically used in finite element reliability analysis. In particular, the visualization results show that these LSSs are often highly nonlinear in the neighborhood of their DPs. As application example, the time-invariant reliability analysis of a reinforced concrete frame structure subjected to horizontal pushover loads is considered. DP-RS-Sim-based estimations of the probability of limit state exceedance (expressed in terms of displacement thresholds) by the benchmark structure are compared with FORM, SORM, crude Monte Carlo and Importance Sampling results in terms of accuracy and computational cost. It is shown that the new DP-RS-Sim method can provide accurate failure probability estimates at low computational cost compared to other structural reliability methods. KEYWORDS: Structural Reliability, Limit State Surface, Response Surface Method, Monte Carlo Simulation, Nonlinear Finite Element Analysis. 1. INTRODUCTION Classical reliability methods such as First- and Second-Order Reliability Methods (FORM and SORM) have been significant breakthroughs toward feasible and reliable methods for integrating probabilistic information and uncertainty analysis into advanced design methods and modern design codes. These methods have been widely used with success in solving challenging reliability problems. Nevertheless, caution should be used in their applications since limits and shortcomings in terms of applicability and accuracy are known and documented, e.g., (1) existence of multiple design points (DPs) (Der Kiureghian and Dakessian 1998; Au et al. 1999), (2) nonlinearity of the limit state surface (LSS) due to non-Gaussianess of the input process for random vibration problems (Der Kiureghian 2000), (3) nonlinearity of the LSS due to nonlinearity in the system (Barbato 2007). Accuracy of FORM and SORM approximations is strongly dependent on the nonlinearity of the LSS defining