ABSTRACT
Engineering structures involve uncertainties associated with initial imperfections in the structural systems, resistance deterioration, and increasing loads. Thus, probabilistic safety assessment of engineering structures has become increasingly important in recent decades. A cable-stayed bridge is composed of stay cables, girders, and towers. The potential failure modes are bending failure of girders, strength failure of cables, and displacement failure of girders. The failure of a cable may not lead to failure of the entire bridge. A new adaptive support vector regression (ASVR) method is proposed for structural system reliability assessment. Compared with traditional support vector regression (SVR), the proposed ASVR method involves two updating procedures to estimate structural limit state functions (LSFs). Two verification examples involving a nonlinear LSF and a geometrically nonlinear suspended structure are provided to illustrate the accuracy and the efficiency of the ASVR method. A pre-stressed concrete cable-stayed bridge is utilized to demonstrate the applicability of the proposed method. The main failure sequences of the cable-stayed bridge are identified.
