Non Deterministic Approaches

 

James L. Beck^*, Gerhart I. Schuëller^† AND Héctor A. Jensen^#

Robustness in the prediction of structural response is an essential requirement of structural analysis for quality assurance. Errors in predictions can arise from many possible sources, such as uncertainty and variability in the structural boundary conditions, loads and material properties, and incomplete model capabilities. During design and operation, the acceptable, or desired, structural performance is usually defined in terms of specified response quantities not exceeding certain limits, such as interstory drift limits or floor acceleration limits. To assess the structural performance, it is important that robust predictions are made that treat all the uncertainties, from modelling future loads to modelling the structural behaviour. This leads naturally to the concept of robust reliability which quantifies the confidence that the analyst has that the design, when constructed, will meet the design specifications. Maximization of the robust reliability could be used, for example, to search for an optimal robust design among a feasible set of designs.

This mini-symposium will focus on appropriate methods to define, calculate and utilize the concept of robust predictions of structural response; for this purpose, the quantification of the various sources of uncertainty could be based on Bayesian probability, fuzzy logic, possibility logic, interval arithmetic, or infogap analysis. The goal is to present methods that can provide the structural analyst and designer with additional information about how well the performance specifications will be met, which can be used to improve the design and hence the quality of the constructed system. Some possible topics include approaches and algorithms for defining and calculating robust predictions and reliability, especially those quantifying the effects of modelling uncertainties, new algorithms for searching for an optimal robust design, and methods for reassessing robust predictions and reliability by monitoring the response of a structure during its operation.

^*California Institute of Technology, Mail Code 10444, Pasadena, CA 91125, USA, jimbeck@caltech.edu
^†LFU, Institute of Engineering Mechanics Technikerstrasse 13, 6020 Innsbruck, Austria, EU, Mechanik@uibk.ac.at
^#Univ. Tecnica Federico Santa Maria, Dept. of Civil Engineering, Casilla: 110V, Av. Los Placeres 401, Valparaiso, Chile, hector.jensen@usm.cl