Date Approved

12-31-2003

Embargo Period

5-9-2016

Document Type

Thesis

Degree Name

M.S. in Engineering

Department

Electrical & Computer Engineering

College

Henry M. Rowan College of Engineering

First Advisor

Mandayam, Shreekanth

Subject(s)

Gas pipelines; Nondestructive testing

Disciplines

Electrical and Computer Engineering

Abstract

Nondestructive evaluation (NDE) plays a vital component in the operation and maintenance of large infrastructure such as gas transmission pipelines, nuclear power plants, aircraft, bridges and highways, etc. As this infrastructure continues to age it is essential that the inspection techniques reliably and accurately predict the integrity of these systems. No single NDE method is capable of inspecting all types of anomalies and extracting all required information – a combination of methods must be used and the resulting data fused. Moreover, newer systems that are developed are often made of composite materials that include metals and dielectrics. One interrogation modality cannot be used to inspect such components for reliability – multiple tests are always needed.

This thesis presents a technique that can be used to fuse data from multiple sensors that are employed in modem NDE applications, specifically in the in-line inspection of gas transmission pipelines. A radial basis function artificial neural network is used to perform geometric transformations on data obtained from multiple sources. The technique allows the user to define the redundant and complementary information present in the data sets. The efficacy of the algorithm is demonstrated using simulated canonical images and experimental images obtained from the NDE of a test specimen suite using magnetic flux leakage, ultrasonic and thermal imaging methods. The results presented in this thesis indicate that neural network based geometric transformation algorithms show considerable promise in multi-sensor data fusion applications.

Share

COinS