THE APPLICATION OF ACTIVE LEARNING KRIGING IN DETERMINING THE RELIABILITY OF BRIDGE COMPONENTS

Abstract

The objective of this research was to develop a framework of analysis using Active Learning Kriging Monte Carlo Simulation (AK-MCS) to assess and optimize the reliability calculation of reinforced concrete bridges components. The methodology consisted of developing a computer code to perform AK-MCS analysis to calculate the reliability index of bridge girders and piers, verify the accuracy of the code by conducting a sensitivity analysis, and optimize AK-MCS analysis by balancing the accuracy and efficiency. The computer code was developed using MATLAB and its accuracy was verified by conducting 810 AK-MCS analyses (15 bridge girder and pier configuration x 54 unique AK configurations, where the latter refers to the set of correlation, regression, and learning functions). The verification analysis results indicted the sensitivity of the solution efficiency (run time and number of training points) to the choice of the AK configuration.