Aristotle University Of Thessaloniki (AUTH)
Application development on embedded systems: localization and size estimation of leaks in metallic pipelines
Abstract
dc:descriptionThis doctoral thesis deals with the design and implementation of certain applications on embedded systems, concerning the localization and size estimation of leakages in metallic pipelines. The occurrence of leaks in pipeline networks is a serious issue both in the industrial world and in urban applications. For this reason, many researchers over the years have designed monitoring systems for the timely and reliable detection and localization of leaks in pipeline systems. So, a detailed study of the related literature was performed. Since the non-invasive localization and size estimation methods designed in this thesis rely on the vibroacoustic signals appearing in a pipeline when a leak is present, a study of the physical mechanisms along with the mathematical description of the vibroacoustic waves was conducted and the types of sensors that can be employed for their detection were examined. Moreover, the mathematical tools and the signal processing techniques utilized in the developed methods in this thesis are presented. At first, based on the related literature and experimental data from actual leak measurements, several acoustic leak localization methods were designed. The identification of the leak position in these methods relies on either the Time-Difference-Of-Arrival (TDOA) of the leak acoustic signals reaching the sensors, which are mounted on the pipe, or the attenuation that a vibroacoustic wave undergoes during its propagation. Moreover, for the development of the methods, the stochastic (non-deterministic) nature of the leak signals was studied, along with some physical phenomena related to the acoustic wave propagation, such as the dispersion, the frequency-dependent attenuation and the resonance. On the other hand, the acoustic method for the estimation of the leak size that was developed combines the information about the leak location with the multiple linear regression technique and it classifies a leak into one of the three available categories (i.e. “small”, “medium” and “large” leaks) according to its size, thus providing an idea about the severity of the situation. The leak localization and size estimation methods designed in the present thesis were developed in LabVIEW and Matlab software, as well as in the C++ programming language. Finally, some of the developed methods in this thesis were implemented (in C++) on a low-cost embedded system. This system consists of two nodes, each containing an acoustic sensor, an A/D converter, a GPS module and the Raspberry Pi 4B hardware platform. On this platform, the accuracy of the developed methods, the execution time and the memory consumption were examined and certain trade-offs were studied based on the configuration of the main parameters of the employed embedded system.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kousiopoulos, Georgios-Panagiotis
- Κουσιόπουλος, Γεώργιος-Παναγιώτης
Subjects
dc:subject × 18- Ενσωματωμένα συστήματα
- Αισθητήρες
- Ακουστικά σήματα
- Δίκτυα σωληνώσεων
- Embedded systems
- Sensors
- Acoustic signals
- Pipeline networks
- Φυσική
- Φυσικές Επιστήμες
- Εφαρμοσμένη φυσική
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Λογισμικό (software)
- Physical Sciences
- Natural Sciences
- Applied Physics
- Computer and Information Sciences
- Software
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/52740