Abstract:
Condition assessment of reinforced concrete bridges is a complex subject imbued with uncertainty and vagueness. This complexity arises from numbers and relations of different kindof problems in reinforced concrete bridges. Condition assessment requires vast knowledge of the behaviour of reinforced concrete structures subjected to different phenomena such as excessive loading, environmental effects and chemical attacks. This requirement can be achieved through an expert system, which may represent human expertise. This study present an Artificial Neural Network (ANN) assisted crack rating system for RC bridges' girder which improves the effectiveness of crack condition rating. The ANN system was developed as an alternative to traditional crack rating methods. The rules for the ANN system were constructed from expert knowledge, technical books and inspection results of 5 different RC bridges. The results obtained by ANN model show high correctness ratio conformity to crack rating obtained by the traditional inspection methods.
