Student:	Yiğit MAHMUTOĞLU
Supervisor:	Assoc. Prof. Dr. Kadir TÜRK
Department:	Elektrik-Elektronik Müh.
Institution:	Graduate School of Natural and Applied Sciences
University:	Karadeniz Technical University Turkey

Title of the Thesis:	TWO DIMENSIONAL POSITIONING OF LEAKAGES IN UNDERWATER NATURAL GAS PIPELINES WITH PASSIVE ACOUSTIC METHODS
Level:	Ph.D.
Acceptance Date:	15/3/2019
Number of Pages:	127
Registration Number:	Di1306

Summary:

In this dissertation, passive acoustic based methods have been proposed to locate

leakages in underwater natural gas pipelines (UNGPs), taking into account three different

underwater acoustic channels, which are single (direct) path channel, time-invariant

multipath channel and time-varying multipath channel. Simulation studies were performed

to demonstrate the success of the methods.

First of all, the nature of the acoustic pressure signal to be generated by leakages that

can occur in UNGP is investigated. According to this analysis, it has been shown that the

leakage based acoustic pressure signals are narrowband, and the leak hole diameter, bubble

diameter and the acoustic signal strength at the leak hole can be calculated by using

frequencies of the received signals and required parameters. Secondly, passive acoustic

localization methods based on the techniques of received signal strength and received signal

strength difference for positioning the leakages in the UNGPs have been proposed. As the

underwater acoustic channel model, a single path channel with direct path signal is used.

Thirdly, the fingerprint method has been proposed to locate leakages in the UNGPs in time

invariance multipath channel, where the received signal is composed of direct path signal

and reflected signals from the seabed and surface. Finally, the signal strength based curve

fitting method and the fingerprint method have been proposed to locate leakages in the

UNGPs in a time varying (due to moving sea surface, underwater currents, etc.) multipath

underwater acoustic channel. As a result of the simulation studies, it has been seen that the

fingerprint method has the best positioning performance.