Abstract:
Climatic conditions, load, fatigue, aging and other factors causes a de-terioration in civil infrastructures. As a consequence, repair and maintenance work actions are needed, being the former considered as more expensive than the latter ones. Indeed, an accurate method for measuring corrosion is a fundamental prerequisite for the detection of damaged areas and for planning an effective re-pairing of concrete maritime structures. In this article a comparation between two surrogate models, Markov Chains and Neuronal Networks, is presented and ap-plied to predict the results of corrosion sensors of an infrastructure data set. The proposed methodology benefits from current monitoring practice and have the objective to develop a modular decision support system for the integrated asset management, taking into account operational, economic and environmental cri-teria. The results could contribute to the possibility of adapting these degradation models to aggressive environments and repaired structures, thus generating ac-curate maintenance strategies, and reducing costs. This methodology is part of the ongoing study “GIIP- Intelligent Port Infrastructure Management”.
