Probabilistic Analysis of Redundant Systems by the Identification of Multiple Failure Modes
Adelano Esposito1; Marcos Daniel de Freitas Awruch1; Herbert Martins Gomes1
1 Federal University of Rio Grande do Sul
doi:10.20906/CPS/COB-2015-0284
Resumo
The failure of a real structural system can be described by a set of interlinked failures or by the failure of independent components. So far, this type of problem has been handled in a simple fashion, not taking into account several existing failure modes, especially in redundant systems. The simplification of the real problem, overlooking some failure modes, may hinder the calculation of reliability. Bearing that in mind, a method for the probabilistic analysis of the system is proposed, using an improved version of the (HDM) Hyperspace Division Method platform, which allows the detection of multiple failure modes. Besides, the use of a chi-squared distribution allows for a directional integration procedure to calculate the failure probability of each mode in a more accurate manner than in conventional approximation and simulation methods. The key objectives of this study are: to identify the sequence of failure events that make up failure modes at ultimate and serviceability limit states in redundant structures by considering physical and geometric nonlinearities; to calculate the failure probabilities for each mode; and to compare the efficiency and accuracy of the proposed method with other existing methods and with Monte Carlo simulations. Results demonstrate that the proposed method is efficient at identifying failure modes, including those neglected by the existing approaches.
Palavras-chave: reliability; failure modes; structural analysis; redundant structures