Publication detail

Inverse reliability problem solved by artificial neural networks

LEHKÝ, D. NOVÁK, D.

Original Title

Inverse reliability problem solved by artificial neural networks

Type

conference paper

Language

English

Original Abstract

An efficient inverse reliability analysis method is proposed to obtain design parameters in order to achieve the prescribed reliability level. The inverse analysis method is based on the coupling of an artificial neural network and a small-sample simulation method of the Monte Carlo type used for efficient stochastic preparation of the training set utilized in artificial neural network training. The calculation of reliability is performed using the first order reliability method. The validity and efficiency of the approach is shown using numerical examples taken from the literature as well as from civil engineering computational mechanics for both single and multiple design parameters and single and multiple reliability constraints.

Keywords

Design parameters; First order reliability methods; Inverse analysis methods; Inverse reliability analysis; Inverse reliability problem; Multiple reliability constraints; Reliability level; Training sets

Authors

LEHKÝ, D.; NOVÁK, D.

RIV year

2013

Released

16. 6. 2013

Location

New York, USA

ISBN

978-1-138-00086-5

Book

Safety, Reliability, Risk and Life-Cycle Performance of Structures and Infrastructures

Pages from

5303

Pages to

5310

Pages count

8

BibTex

@inproceedings{BUT107532,
  author="David {Lehký} and Drahomír {Novák}",
  title="Inverse reliability problem solved by artificial neural networks",
  booktitle="Safety, Reliability, Risk and Life-Cycle Performance of Structures and Infrastructures",
  year="2013",
  pages="5303--5310",
  address="New York, USA",
  isbn="978-1-138-00086-5"
}