• Login
• Admin

Le, D. Van-Khoa (2022) A complete online SVM and case base reasoning in pipe defect dection with multisensory inspection gauge. PhD thesis, University of Nottingham.

Preview

PDF (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader Available under Licence Creative Commons Attribution. Download (12MB) | Preview

Abstract

An in-line inspection (ILI) robot has been considered an inevitable requirement to perform non-destructive testing methods efficiently and economically. The detection of flaws that could lead to leakages in buried concrete pipes has been a great concern to the oil and gas industry and water resource-based industry. The major problem is the difficulty in modeling the detection of cracks due to their irregularity and randomness that cannot be easily detected.

Consequently, the use of an advanced modality system has emerged. Common defects detection systems favor non-destructive testing methods, which utilize specific sensory data. Only a few systems focus on fusing different types of sensory data. Moreover, the decision mechanism in this system required heavy-power consumption sensors with the configuration from the expertise domain. In addition, the outcome of the decision system is a consequence of rule-based settings rather than a mixture of learned features. This work covers the study of defect detection of non-destructive testing methods using fusion inspection sensors, light detection and ranging (LiDAR), and Optic sensors. The studies on ILI robots are reviewed to construct an efficient gauge. The prototype robot has been designed and successfully operated in a lab-scale environment.

Ultimately, the study proposed a replacement for the standard expert system - in the branch of the CBR system, which is the crucial contribution of this thesis. Recent developments in Case-based Reasoning systems (CBR) have led to an interest in favoring machine learning (ML) approaches to replace traditional weighted distance methods. However, valuable information obtained through a training process was relinquished as transferring to other phases. As a result, the complete SVM-CBR system in this thesis concentrates on solving this gap by presenting an effective transferring mechanism from phase to phase. This thesis proposed a full pipeline integration of CBR using the kernel method designated with support vector machine. SVM technique is the primary classification engine for the combined sensory data. Since the system requires a learning SVM model to be invoked in every phase, the online learning mechanism is nominated to update the model when a new case adjoins effectively. The proposed full SVM-CBR integration has been successfully built into a pipe defect detection. The achieved result indicates a substantial improvement in transferring learning information accurately.

Item Type:	Thesis (University of Nottingham only) (PhD)
Supervisors:	Chen, Zhiyuan
Keywords:	pipe defect detection system, water resource-based industry, fusion inspection sensors, vector machine, pipeline
Subjects:	Q Science > Q Science (General)
Faculties/Schools:	University of Nottingham, Malaysia > Faculty of Science and Engineering — Science > School of Computer Science
Item ID:	69346
Depositing User:	Le, Dinh
Date Deposited:	24 Jul 2022 04:40
Last Modified:	24 Jul 2022 04:40
URI:	https://eprints.nottingham.ac.uk/id/eprint/69346

Actions (Archive Staff Only)

Edit View