Design optimization applied in structural dynamics

Didem Akcay-Perdahcioglu; Andries de Boer; Peter van der Hoogt

Design optimization applied in structural dynamics

Didem Akcay-Perdahcioglu, Andries de Boer, Peter van der Hoogt

Faculty of Engineering Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic

105 Downloads (Pure)

Abstract

This paper introduces the design optimization strategies, especially for structures which have dynamic constraints. Design optimization involves first the modeling and then the optimization of the problem. Utilizing the Finite Element (FE) model of a structure directly in an optimization process requires a long computation time. Therefore the Backpropagation Neural Networks (NNs) are introduced as a so called surrogate model for the FE model. Optimization techniques mentioned in this study cover the Genetic Algorithm (GA) and the Sequential Quadratic Programming (SQP) methods. For the applications of the introduced techniques, a multisegment cantilever beam problem under the constraints of its first and second natural frequency has been selected and solved using four different approaches.

Original language	Undefined
Title of host publication	1st International conference on Artificial intelligence for industrial applications.
Editors	T Tiskarna
Place of Publication	Czech Republic
Publisher	ICAI
Pages	45-51
Number of pages	6
ISBN (Print)	51
Publication status	Published - 19 Feb 2007
Event	First International Conference on Artificial Intelligence for Industrial Applications - Ostravice, Czech Republic Duration: 19 Feb 2007 → 23 Feb 2007

Conference

Conference	First International Conference on Artificial Intelligence for Industrial Applications
Period	19/02/07 → 23/02/07
Other	19-23 February 2007

Keywords

METIS-245282
Sequential Quadratic Programming
Structural dynamics
Genetic Algorithms
Design Optimization
Backpropagation Neural Networks
Finite Element Method
IR-58892

Access to Document

Akcay07design

Cite this

@inproceedings{26e633dee6894dd3851c9f8297dac83e,

title = "Design optimization applied in structural dynamics",

abstract = "This paper introduces the design optimization strategies, especially for structures which have dynamic constraints. Design optimization involves first the modeling and then the optimization of the problem. Utilizing the Finite Element (FE) model of a structure directly in an optimization process requires a long computation time. Therefore the Backpropagation Neural Networks (NNs) are introduced as a so called surrogate model for the FE model. Optimization techniques mentioned in this study cover the Genetic Algorithm (GA) and the Sequential Quadratic Programming (SQP) methods. For the applications of the introduced techniques, a multisegment cantilever beam problem under the constraints of its first and second natural frequency has been selected and solved using four different approaches.",

keywords = "METIS-245282, Sequential Quadratic Programming, Structural dynamics, Genetic Algorithms, Design Optimization, Backpropagation Neural Networks, Finite Element Method, IR-58892",

author = "Didem Akcay-Perdahcioglu and \{de Boer\}, Andries and \{van der Hoogt\}, Peter",

year = "2007",

month = feb,

day = "19",

language = "Undefined",

isbn = "51",

pages = "45--51",

editor = "T Tiskarna",

booktitle = "1st International conference on Artificial intelligence for industrial applications.",

publisher = "ICAI",

note = "First International Conference on Artificial Intelligence for Industrial Applications ; Conference date: 19-02-2007 Through 23-02-2007",

}

TY - GEN

T1 - Design optimization applied in structural dynamics

AU - Akcay-Perdahcioglu, Didem

AU - de Boer, Andries

AU - van der Hoogt, Peter

PY - 2007/2/19

Y1 - 2007/2/19

N2 - This paper introduces the design optimization strategies, especially for structures which have dynamic constraints. Design optimization involves first the modeling and then the optimization of the problem. Utilizing the Finite Element (FE) model of a structure directly in an optimization process requires a long computation time. Therefore the Backpropagation Neural Networks (NNs) are introduced as a so called surrogate model for the FE model. Optimization techniques mentioned in this study cover the Genetic Algorithm (GA) and the Sequential Quadratic Programming (SQP) methods. For the applications of the introduced techniques, a multisegment cantilever beam problem under the constraints of its first and second natural frequency has been selected and solved using four different approaches.

AB - This paper introduces the design optimization strategies, especially for structures which have dynamic constraints. Design optimization involves first the modeling and then the optimization of the problem. Utilizing the Finite Element (FE) model of a structure directly in an optimization process requires a long computation time. Therefore the Backpropagation Neural Networks (NNs) are introduced as a so called surrogate model for the FE model. Optimization techniques mentioned in this study cover the Genetic Algorithm (GA) and the Sequential Quadratic Programming (SQP) methods. For the applications of the introduced techniques, a multisegment cantilever beam problem under the constraints of its first and second natural frequency has been selected and solved using four different approaches.

KW - METIS-245282

KW - Sequential Quadratic Programming

KW - Structural dynamics

KW - Genetic Algorithms

KW - Design Optimization

KW - Backpropagation Neural Networks

KW - Finite Element Method

KW - IR-58892

M3 - Conference contribution

SN - 51

SP - 45

EP - 51

BT - 1st International conference on Artificial intelligence for industrial applications.

A2 - Tiskarna, T

PB - ICAI

CY - Czech Republic

T2 - First International Conference on Artificial Intelligence for Industrial Applications

Y2 - 19 February 2007 through 23 February 2007

ER -