Efficient approach for calibration of load and resistance factors in the limit state design of a breakwater foundation

Abstract

This study proposes a facile procedure to overcome extensive computational burdens in Monte Carlo simulation (MCS)-based load and resistance factors (LRFs) calibration. Accordingly, a reliability analysis (RA) was conducted on a surrogate set of the constructed response surface functions (RSFs) rather than direct application to the actual implicit problems. Owing to the simplicity of the RA approach, the LRFs were then calibrated using an optimization framework. An adaptive boundaries algorithm was proposed to effectively reduce the number of iterations in the optimization. To elucidate the proposed procedure, four examples: two on slopes and two on foundation stability of breakwaters under earthquake loading were investigated. The results revealed that the surrogate RSFs set that was constructed for numerous failure surfaces not only accurately evaluated the system RA of the actual problems but also allowed for a relatively easier calibration of the LRFs within the optimization framework. In addition, it was found that in the calibration of the LRFs, the RSFs need to be reconstructed to properly account for the variation in the sampling points. The successful convergence of the proposed optimization model took only tens of minutes instead of dozens of days required using the basic. This validates the efficiency and accuracy of the proposed procedure.