Research of composite health monitoring based on KPCA and GA-TWSVM

Aiming at the problem of composite anomaly detection and health monitoring, the improved twin support vector machine(TWSVM) with kernel principle component analysis(KPCA) is applied to aircraft composite health monitoring. Firstly, model of uniplanar multi-electrodes was partitioned into equal area...

Full description

Saved in:
Bibliographic Details
Published in2017 Prognostics and System Health Management Conference (PHM-Harbin) pp. 1 - 6
Main Authors Bao-yin Zhang, En-sheng Dong, Gang Guo
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2017
Subjects
Capacitance
Composite materials
Electrodes
genetic algorithm
Genetic algorithms
health monitoring
Kernel
kernel principle component analysis
Monitoring
Support vector machines
twin support vector machine
uniplanar capacitance sensor
Online AccessGet full text

Cover

More Information
Summary:Aiming at the problem of composite anomaly detection and health monitoring, the improved twin support vector machine(TWSVM) with kernel principle component analysis(KPCA) is applied to aircraft composite health monitoring. Firstly, model of uniplanar multi-electrodes was partitioned into equal area units with FEM so that data was acquired enough. Secondly, KPCA was used to select the dimension of feature vectors and short training time, then the best features were found by TWSVM optimized with genetic algorithm, called GATWSVM for short. Finally, mesured data of three typical composite were sent to KPCA-GATWSVM for verification. After verification of the simulation and the measured data, the effective certificate of general KPCA-GATWSVM was applied to the health monitoring of aircraft composite material. The classification accuracy of the simulation data reached 93.8%, and result of the measured data reached 100%.
ISSN:2166-5656
DOI:10.1109/PHM.2017.8079148