Reliability analysis of concrete structures with neural networks and response surfaces

Purpose - To research the feasibility in using artificial neural networks (ANN) and response surfaces (RS) techniques for reliability analysis of concrete structures. Design/methodology/approach - The evaluation of the failure probability and safety levels of structural systems is of extreme importance in structural design, mainly when the variables are eminently random. It is necessary to quantify and compare the importance of each one of these variables in the structural safety. RS and the ANN techniques have emerged attempting to solve complex and more elaborated problems. In this work, these two techniques are presented, and comparisons are carried out using the well-known first-order reliability method (FORM), with non-linear limit state functions. The reliability analysis of reinforced concrete structure problems is specially considered taking into account the spatial variability of the material properties using random fields and the inherent non-linearity. Findings - It was observed that direct Monte Carlo simulation technique has a low performance in complex problems. FORM, RS and neural networks techniques are suitable alternatives, despite the loss of accuracy due to approximations characterizing these methods. Research limitations/implications - The examples tested are limited to moderated large non-linear reinforced concrete finite element models. Conclusions are drawn based on the examples. Practical implications - Some remarks are outlined regarding the fact that RS and ANN techniques have presented equivalent precision levels. It is observed that in problems where the computational cost of structural evaluations (computing failure probability and safety levels) is high, these two techniques could improve the performance of the structural reliability analysis through simulation techniques. Originality/value - This paper is important in the field of reliability analysis of concrete structures specially when neural networks or RS techniques are used.