Aristotle University Of Thessaloniki (AUTH)
Έλεγχος και βελτιστοποίηση συνθηκών λειτουργίας φυγοκεντρικής αντλίας, με χρήση δικτύου αισθητήρων και ρύθμιση στροφών
Abstract
dc:descriptionThe purpose of this study is to evaluate and compare advanced methods of operational state recognition and early fault detection, as well as control and optimization methods for the operation of centrifugal pumps. There are various methods that can detect the operating conditions of the pump either in real-time during operation or by inspecting the pump parts at a standstill or even after disassembling them. Additionally, simulations and models have been developed for the accurate prediction of a pump's operating condition based on a predetermined set of parameters. The most modern trend is the use of data collection sensors and suitable signal analysis algorithms. The use of computational methods and artificial intelligence for signal analysis has been very successful in detecting operating conditions in different types of pumps. Despite the widespread use of centrifugal pumps, there are still questions and problems that have not been extensively studied. Mechanical failures (in bearings, impellers, or the shaft), incorrect hydraulic parameters (operation without water, pipe failure, operation under cavitation, under-pressure or over-pressure), failures in the drive systems (electrical faults, internal combustion engine failures), and other issues, make the control, early recognition, and correction of operating parameters of the pumps a quite complex endeavor. The experimental research during this dissertation focuses on the use of vibration and sound sensors outside the pump casing to collect data in real-time and analyze it using modern computational methods such as artificial neural networks. The setup was designed to be as close as possible to real irrigation systems. Common equipment and low-cost materials were used with the aim of creating a future commercial tool for early detection of the operational status of a centrifugal pump that allows for preventive maintenance and proactive control. The experimentation was limited to electrically driven single-stage centrifugal irrigation pumps, as they are the most widespread type of pumping machinery in the agricultural sector for irrigation. The aim of this work is as practical an approach to the subject as possible, regarding the simplicity of the methods of taking measurements, the simulation of real operating conditions of the pump, but also the minimization of theoretical calculations for the extraction of safe conclusions. The results of the experimental measurements and their analysis are particularly successful and prove that, particularly with the use of vibration sensors, the recognition of the simulated pump operating conditions is successful. With minimal pre-processing of the signals and their analysis using computational methods, it is possible to develop a tool based on portable low-cost devices, which can diagnose in real-time and early on, damage conditions of the pump or suboptimal operation, and even intervene in its operation preventively or correctively, provided there are suitable speed control systems.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Karagiovanidis, Marios
- Καραγιοβανίδης, Μάριος
Subjects
dc:subject × 14- Φυγοκεντρικές αντλίες
- Προβλεπτική συντήρηση
- Μηχανική μάθηση
- Αισθητήρες δονήσεων και ήχου
- Centrifugal pump
- Predictive maintenance
- Machine learning
- Vibration and sound sensors
- Άλλες Γεωπονικές Επιστήμες
- Γεωπονικές Επιστήμες και Κτηνιατρική
- Γεωπονική μηχανική
- Other Agricultural Sciences
- Agricultural and Veterinary Sciences
- Agricultural Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/56971