Development of Concrete Bridge Rating Expert System ( BREX ) in Japan

Abstract

The aim of the present study is to develop a concrete bridge rating expert system for deteriorated concrete bridges that was constructed from a hierarchical neural network in order to carry out fuzzy inference and machine learning. The proposed system evaluates the performance of concrete bridges on the basis of a simple visual inspection and technical specifications. The neural network applied in the study facilitates refinement of the knowledge base by the use of the Back-Propagation method, and it prevents the influence mechanism of the system from becoming a black box. In this study, the training set for machine learning is obtained from inspection of actual in-service bridges and questionnaire surveys of bridge experts. Furthermore, comparisons between the diagnostic results of bridge experts and those of the proposed system are presented in order to demonstrate the validity of the system’s learning capacity.