Edited by: B.H.V. Topping and M. Papadrakakis

Keywords: stochastic structural analysis, parallel processing, advanced simulation, structural analysis software.

Significant progress has been made recently in the rational treatment of uncertainties in large finite elemtn (FE) models of complex structures [1], but the development of the corresponding software is still lagging behind [2].

The present contribution addresses methodological, conceptual and practical aspects related to software development for structural reliability analysis. From the methodological point of view, advanced simulation methods are emphasized, since they are most robust with respect to the number of uncertain parameters in the numerical model and hence more suitable for the analysis of large FE models of complex structures. With respect to the implementation, particular weight is attached to the interaction of the FE software and the uncertainty analysis software. Furthermore, the utilization of parallel computing is presented as a key requirement on any reliability analysis software.

The above issues are discussed and elucidated in the context of a software tool for uncertainty and reliability analysis, COSSAN^TM, which has been developed by the authors and their coworkers. In this code the interaction with third party FE tools has been accomplished both by integrating the uncertainty analysis into the FE code and by creating a standalone application which interfaces with the FE code in a non- or little intrusive fashion. The features included in the COSSAN^TM software include reliability analysis, stochastic finite element analysis, optimization, sensitivity and robustness analysis, as well as tools for parallel processing.

In order to demonstrate the application of the software, two examples from structural engineering are presented, namely a satellite structure under harmonic excitation and a multi-storey building subjected to stochastic ground acceleration.

In summary the paper emphasizes that there is no plausible motivation to forego the possibilities offered by systematic uncertainty analysis in structural engineering, since all ingredients - i.e. the theoretical basis, the efficient algorithms and the computing resources - are in place.

[1]

C.A. Schenk, G.I. Schuëller, "Uncertainty Assessment of Large Finite Element Systems", Springer-Verlag, Berlin/Heidelberg/New York, 2005.

[2]

M.F. Pellissetti, G.I. Schuëller, "On general purpose software in structural reliability", Structural Safety, 28, 3-16, 2006.

Return to the contents page
Return to the book description