A neural network approach to structure damage assessment

Author

Faravelli, L. ; Pisano, A.A.

Author_Institution

Dept. of Struct. Mech., Pavia Univ., Italy

fYear

35765

fDate

8-10 Dec1997

Firstpage

585

Lastpage

588

Abstract

Presents a method for damage detection in multi-bay planar truss structures. A neural network is trained by transfer functions of the structural system. The approach allows one to avoid all the problems which characterize the techniques based on system parameter identification. The neural network architecture and size, the choice of the learning rule and of the corresponding parameters are discussed. The neural network approach is able to uniquely identify the damaged element in almost all of the investigated cases

Keywords

failure (mechanical); failure analysis; learning (artificial intelligence); neural net architecture; parameter estimation; structural engineering computing; transfer functions; damage detection; learning rule; multi-bay planar truss structures; neural network architecture; neural network size; neural network training; structure damage assessment; system parameter identification; transfer functions; Actuators; Aerodynamics; Artificial neural networks; Feedforward neural networks; Finite element methods; Monitoring; Neural networks; Parameter estimation; Signal processing; Transfer functions;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Information Systems, 1997. IIS '97. Proceedings

Conference_Location

Grand Bahama Island

Print_ISBN

0-8186-8218-3

Type

conf

DOI

10.1109/IIS.1997.645426

Filename

645426