Document Type

Theses, Masters

Rights

Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence

Disciplines

Electrical and electronic engineering

Publication Details

Thesis submitted for the qualification of Master of Philosophy, School of Control Systems and Electrical Engineering, DIT, May 2012.

Abstract

This research project was set up to monitor, on an ongoing basis, the condition of the end windings and their support structures of a 288 MVA 2-pole synchronous generator with a known end winding vibration problem Excessive vibration at the end windings was caused by the natural frequencies of individual end bars (local) and of the entire end winding structure (global) being at or close to the magnetic forcing frequency of 100 Hz. Resonant vibration such as this has been a cause of major failures in machines of the same type in the past, resulting in significant down time ranging from a few weeks to a number of months, with the obvious implications in terms of cost, generator availability and revenue loss. The project covers the installation of an end winding vibration monitoring system, subsequent analysis and testing in order to attempt to lower peak vibration levels, a low tune modification of the end winding support structure (carried out by the OEM) in order to remedy the natural frequency issue, subsequent data analysis and development of regression models to allow the prediction of vibration levels based on plant data and the implementation of the models on PI ACE (Advance Computing Engine). A webpage, available to all plant personnel via the company intranet was then created to display the actual measured vibration data against the model predicted values and other relevant plant data. With the issue of excessive vibration due to resonance having been remedied by the Original Equipment Manufacturer (OEM) the focus of the project shifted somewhat towards providing an easily interpreted, easily accessible method of monitoring the condition of a now relatively healthy machine well beyond the conclusion of this project. The development of the regression models and web page was seen as the best way of achieving this.

DOI

https://doi.org/10.21427/D7N61J

Download


Share

COinS