Publication detail

Optimization of reliable beam-column construction and its statistical verification

KOKRDA, L. HRABEC, P.

English title

Optimization of reliable beam-column construction and its statistical verification

Type

presentation

Language

en

Original abstract

The aim of the paper is to present recent research advances in the area of optimum and reliable engineering design problems. In this paper, reliability-constrained optimal Bernoulli beam construction design problem is formulated.The problem combines different design elements and is further complicated by exogenous random parameters. The corresponding mathematical model involves ODE-type constraints that are next modelled by a finite element method (FEM) based discretization. The probabilistic constraints concerning reliability are reformulated by using a suitable penalty transformation. Hence, the stochastic programming scenario-based approach is chosen to obtain robust enough solutions. The developed solution algorithm involving preprocessing and postprocessing steps has been implemented in Matlab system. Various computational results based on different input data for considered test cases, which were consulted by experts in civil engineering, are presented and discussed. Selected results were verified by simulation-based procedures and empirical confidence regions for chosen beam's design reliability levels have been constructed.

English abstract

The aim of the paper is to present recent research advances in the area of optimum and reliable engineering design problems. In this paper, reliability-constrained optimal Bernoulli beam construction design problem is formulated.The problem combines different design elements and is further complicated by exogenous random parameters. The corresponding mathematical model involves ODE-type constraints that are next modelled by a finite element method (FEM) based discretization. The probabilistic constraints concerning reliability are reformulated by using a suitable penalty transformation. Hence, the stochastic programming scenario-based approach is chosen to obtain robust enough solutions. The developed solution algorithm involving preprocessing and postprocessing steps has been implemented in Matlab system. Various computational results based on different input data for considered test cases, which were consulted by experts in civil engineering, are presented and discussed. Selected results were verified by simulation-based procedures and empirical confidence regions for chosen beam's design reliability levels have been constructed.

Keywords in English

reliable engineering design, probabilistic constraints, penalty function, FEM

Released

12.12.2018