Global structural optimization considering expected consequences of failure and using ANN surrogates

Authors:
Wellison José De Santana Gomes;André Teófilo Beck
Affiliations:
Department of Structural Engineering, São Carlos School of Engineering, University of São Paulo, 13566-590 São Carlos, SP, Brazil;Department of Structural Engineering, São Carlos School of Engineering, University of São Paulo, 13566-590 São Carlos, SP, Brazil
Venue:
Computers and Structures
Year:
2013

Citing 15
Cited 0

Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks

Neural Networks
Neurocomputing

Neurocomputing
Introduction to the theory of neural computation

Introduction to the theory of neural computation
Adaptive filter theory (3rd ed.)

Adaptive filter theory (3rd ed.)
Nonlinear positional formulation for space truss analysis

Finite Elements in Analysis and Design
A probabilistic neural network for earthquake magnitude prediction

Neural Networks
Uncertain linear structural systems in dynamics: Efficient stochastic reliability assessment

Computers and Structures
Non-stationary response of large, non-linear finite element systems under stochastic loading

Computers and Structures
A survey on approaches for reliability-based optimization

Structural and Multidisciplinary Optimization
A FEM procedure based on positions and unconstrained vectors applied to non-linear dynamic of 3D frames

Finite Elements in Analysis and Design
Efficient strategies for reliability-based optimization involving non-linear, dynamical structures

Computers and Structures
A novel RBF neural network with fast training and accurate generalization

CIS'04 Proceedings of the First international conference on Computational and Information Science
Reliability sensitivity estimation of linear systems under stochastic excitation

Computers and Structures
General purpose software for efficient uncertainty management of large finite element models

Finite Elements in Analysis and Design
Training feedforward networks with the Marquardt algorithm

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

The literature is filled with structural optimization articles which claim to minimize costs but which disregard the costs of failure. Due to uncertainties, minimum cost can only be achieved by considering expected consequences of failure. This article discusses challenges in solving real structural optimization problems, taking into account expected consequences of failure. The solution developed herein combines non-linear FE analysis (by positional FEM), structural reliability analysis, Artificial Neural Networks (used as surrogates for objective function) and a hybrid Particle Swarm Optimization algorithm, which efficiently solves for the global optimum. Optimization of a steel-frame transmission line tower is the application example.