ABSTRACT
ABSTRACT The paper deals with the life-cycle performance prediction of deteriorating Reinforced Concrete (RC) structures by means of Artificial Neural Networks (ANNs). A three-layer ANN is developed and trained to capture the overall system performance based on limited amount of information related to local damage of some components, typically obtained from the results of visual inspections. The training datasets are formed to incorporate the results from several inspections carried out over given observation time intervals and to accommodate predictions over the remaining structural lifetime. The proposed ANN is applied to the life-cycle seismic capacity assessment of a three-story RC frame under chloride-induced corrosion.
