Improving Reliability of Screw Compressor by Condition Monitoring

Please enable javascript in your browser.

Talebi Ahooie, Elahe | 2015

385 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 48207 (08)
University: Sharif University of Technology
Department: Mechanical Engineering
Advisor(s): Behzad, Mahdi
Abstract:
Screw compressors are one of the most applicable compressors in marine engineering. Producing of compressed air for ignition of ship’s disel engine, air conditioning, cooling of food and producing of Nitrogen are applications of this compressor in marine industry. So, improving reliability of these compressors is so important.First of all, componenets of screw compressors, failure modes and different ways for improving reliability of this compressor are studied. The best way to improve reliability of rotating machinery is condition monitoring and so this way was chosen to be studied. In this approach, by using CM data and failure history data of a high pressure water descaling pump are used for verification. The best algorithm, takes two continues TBF and their CM data as inputs and delivers the next TBF and CM data. In each algorithm, there is artificial neural network with three hidden layers. The best algorithm has maximum mean error of 7.65% and minimum mean error of 1.83%
Keywords:
Reliability ; Condition Monitoring ; Neural Network ; Screw Compressor ; Remaining Useful Life ; Forecasting

No TOC