Document Details

Document Type : Thesis 
Document Title :
STRUCTURAL DAMAGE LOCALIZATION USING ARTIFICIAL NEURAL NETWORKS
تحديد أعطاب الهياكل باستخدام الشبكات العصبية الاصطناعية
 
Subject : Faculty of Engineering - Department of Mechanical Engineering 
Document Language : Arabic 
Abstract : Structural Health Monitoring (SHM) involves implementing a damage detection method. This work uses Subspace Identification (SubID) and Artificial Neural Networks (ANN) to detect a notch type damage in a thin aluminum plate. Introducing a damage reduces the stiffness of the plate and alters the modal parameters of the structure. By analyzing the vibrational data from healthy and damaged plates, the health-status of the structure can be determined. Damage detection is approached in three steps: data acquisition, feature extraction and statistical model development. Vibrational data from healthy and damaged plates are measured by using an accelerometer. A series of band-pass filters were developed to remove the accompanying noise. The natural frequencies were identified as a suitable feature for damage detection. SubID was implemented to extract the modal (natural) frequencies. These features were used as the input to the ANN for a group classification. The results from the ANN provided the health-status of the plate. The extracted frequencies by using SubID were verified by comparing them to the frequencies in the power spectrum of the filtered signal. The differences in the frequencies from healthy and damaged plates were successfully used to detect the notch-like damage on the plate with a success rate of 85%. Keywords: subspace, artificial neural network, accelerometer, vibration, structural health monitoring, frequency 
Supervisor : Dr. Kashif Saeed 
Thesis Type : Master Thesis 
Publishing Year : 1436 AH
2014 AD 
Co-Supervisor : Dr. Khalid A. Alnefaie 
Added Date : Thursday, November 27, 2014 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
ثابت يونس نالاكاتNalakath, Sabith YoonusResearcherMaster 

Files

File NameTypeDescription
 37575.pdf pdf 

Back To Researches Page