Machine Diagnostics Workshop 12 June 2017
12 June 2017, ILEC Conference Centre, London, UK
|Monday 12 June|
(in particular variable speed and load with wind turbines)
Author: R B Randall, Emeritus Professor, University of New South Wales
Full-time delegate £400 +VAT
Full-time student £200 +VAT
|Background:||In recent years there has been an upsurge of interest in monitoring variable speed machinery, stimulated by the increased use of wind turbines for renewable energy generation. Another potential application is to mining and marine vehicles with planetary gears driving the wheels or propellers. Conventional data collectors and analysers cannot really deal with this situation, so it is necessary to add new diagnostic tools to the machine diagnostic system. This workshop describes a number of very powerful signal processing techniques, many only recently developed, which overcome a large number of the diagnostic problems. They could be included in a supplementary computer-based analysis system, processing time records gathered on field data collectors, or directly in permanent monitoring systems.|
|Course content: ||Discussion of the signals produced by faults in gears and bearings, both localised and extended. |
Basic signal processing – Fourier transform and variants – Hilbert transform and demodulation – Time synchronous averaging (TSA) – Cepstrum analysis for diagnosis and editing – Cyclostationary analysis; cyclostationary signals of order 1 (CS1) and 2 (CS2) – Time/frequency analysis by STFT, wavelets, wavelet packet analysis, Wigner-Ville distribution and Wigner-Ville spectrum.
Separation of deterministic and random signals and application to separation of gear (CS1) and bearing (CS2) signals – Pros and cons of different methods.
Optimisation of bearing diagnostics by envelope analysis – Advantages of Hilbert transform techniques in frequency domain – Advantages of squared envelope – Use of kurtogram and/or spectral coherence to optimise demodulation bandwidth.
Advantages of phase demodulation-based order tracking (arbitrary resolution, iterative improvement, etc) – Covering a wide speed range using overlapping windows – Blind determination of unknown tooth numbers in a gearbox – Determination of instantaneous speed from a reference frequency signal, either tacho or from response acceleration.
Comparison of Teager Kaiser energy operator (TKEO) by real-time time domain processing versus post-processing using Hilbert transform techniques in the frequency domain (but ideal zero phase shift filtering and differentiation) – applications to gear and bearing diagnostics and advantages of frequency domain Hilbert techniques.
Separation of forcing function and modal response effects in variable speed operation – Use of cepstrum to make this separation in the time/frequency domains before order tracking to the rotation angle/order domains – Coping with the effects of passing through fixed resonances in variable speed operation – Different methods of processing of gear signals (CS1, fault information primarily dominated by forcing functions) and bearing signals (CS2, fault information typically carried by resonances).
All techniques will be illustrated by numerous examples.
|08.30 – 09.00||Welcome and introductions|
|09.00– 10.00||Vibration signals produced by gears and bearings|
1. Gears – Fault information basically deterministic and carried by harmonics of shaft speeds.
2. Bearings – Fault signals stochastic (approximately second order cyclostationary, CS2), even though repetition frequencies roughly follow shaft speeds. Fault information typically carried by fixed resonances independent of speed.
|10.00 – 10.15||Tea and Coffee Break|
|10.15 – 11.15||Basic signal processing|
Fourier transform, Hilbert transform, demodulation – TSA – cepstrum analysis – cyclostationary analysis (CS1 and CS2) – time/frequency analysis.
|11.15 – 12.00||Separation of deterministic and random signals|
Separation of gear and bearing signals by a variety of methods – enhancement of bearing signals using the kurtogram and spectral correlation.
|12.00 – 13.00||Lunch|
|13.00 – 14.00||Optimised bearing diagnostics|
A semi-automated method for a wide range of situations and speeds, based on envelope analysis of an optimum demodulation band after separation of bearing signals.
|14.00 – 15.00||Order tracking|
Phase/time map – angular resampling – advantages of phase demodulation method – covering a wide speed range using overlapping windows.
|15.00 – 15.15||Tea and Coffee Break|
|15.15 – 16.00||Teager Kaiser energy operator (TKEO)|
Comparison of real-time techniques in the time domain with Hilbert techniques in frequency domain – application to speed determination and gear and bearing diagnostics.
|16.00 – 17.00||Gear and bearing diagnostics in variable speed operation|
Separation of forcing function and modal response effects using ceptstrum (before order tracking) – different processing of gear signals (CS1) and bearing signals (CS2).
|17.00 – 17.30||Group discussion|
Click here to download a booking form.
The British Institute of Non-Destructive Testing, Midsummer House, Riverside Way, Bedford Road, Northampton NN1 5NX, UK.
Tel: +44 (0)1604 438300; Fax: +44 (0)1604 438301; Email: email@example.com