Dublin City University
Detection of thin deformable membranes in pipe flow: Development of an inexpensive and robust embedded solution with Electrical Impedance Tomography
Abstract
dc:description.abstractPump clogging is caused by material accumulation around pump impellers. This can have a significant and negative impact on power consumption and reliability. Thin flexible membranes in the flow are often responsible for total and partial blockages in wastewater pumping applications. These membranes are usually found in sewage water but an efficient solution to help monitor them has not yet been reported. The research presented in this thesis aims to assess the feasibility of developing an inexpensive and effective embedded solution for the detection of these membranes. The potential application of the research results includes pump monitoring for better maintenance and smart real-time pump control for clogging prevention. The research has focused on Electrical Impedance Tomography (EIT). Recent available low-cost EIT systems have demonstrated high accuracy but find it challenging to deliver fast reconstructions of the conductivities within the Region of Investigation (ROI) and are usually evaluated under static conditions and limited fluid volume. To be suitable for smart pumping applications, the solution needs to be sufficiently robust to work in a dynamic and noisy environment with a large fluid to solid fraction if it is to detect thin immersed membranes. Electrode mispositioning or pipe deformations errors, which are known to influence the quality of reconstructions in EIT, must be considered. In this research, multiple EIT hardware and software configurations are evaluated to determine optimal solutions for conditions found in an industrial pumping environment. A novel reconstruction method is proposed. It allows fast operation and requires low computational resources while being capable of dealing with noise and system errors. It differs from traditional approaches based on Finite Element Methods (FEM) and relies instead on a non-iterative estimation of significant changes in the measured potential within sub-divisions of the ROI. The results indicate that the solution can detect thin immersed objects of at least 30 cm long at a sampling rate of 50 fps in pipe flow with an area averaged velocity of 4.5 m/s.
Degree
thesis:*- Name dc:type.qualificationname
- phd
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- Dublin City University
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Torrecilla Rubio, Rafael
Rights
- Language dc:language
- en