DETECTING AND PREDICTING DETECTING

Transcription

1 3/13/28 DETECTING AND PREDICTING MW WIND TURBINE DRIVE TRAIN FAILURES Adopted for Wind Power Management class by Andrew Kusiak Intelligent Systems Laboratory 2139 Seamans Center The University of Iowa Iowa City, Iowa Tel: Fax: David Clark Director, Turningpoint/Commtest Causes of vibration MISALIGNMENT IMBALANCE MECHANICAL LOOSENESS RESONANCE BEARINGS 9% of machine vibration is attributed to these 5 issues THE ABOVE PROBLEMS MANIFEST THEMSELVES IN THE CONDITION OF THE ASSET TYPICAL SENSOR LOCATIONS Basically monitor what fails or what is expensive when it fails 1

2 3/13/28 3 COMMON EXAMPLES OF MW WTG ISSUES 1. MISALIGNMENT 2. BEARINGS 3. PLANETARY GEARBOX EXAMPLE # 1 Misalignment of a Megawatt class wind turbine Gearbox to generator Shows as a high peak at the running speed First vibration signature shows.5 amplitude in a velocity measurement (before) Second shows.25 amplitude after alignment (after) BEFORE.5 High amplitude at running speed AFTER.25 2

3 3/13/28 EXAMPLE # 2 BEARING FAULT IN A MEGAWATT CLASS WIND TURBINE COMMON BEARING ABREVIATIONS BPFO (outer race defect) BPFI (inner race defect) FTF (cage defect) BSF (rolling element defect) There are 2 ISO standards for alarms and 1 widely accepted 3 year study also used for alarming AMPLITUDE = SEVERITY PROGRESSION OF THE FAILURE TYPES OF VIBRATION MEASUREMENTS INDICATE THE FAILURE PROGRESSION OF THE COMPONENT. VELOCITY AND DEMODULATED MEASUREMENTS GIVE AN APPROXIMATE TIMELINE How do I interpret the results? Look for peaks at known bearing fault frequencies in both the normal vibration velocity spectra and the demodulated spectra. No peaks in either spectrum: Condition is good, use as a baseline for future comparisons. Peaks appear in Demod only: Early warning indication that defects exist (or the bearing needs lubrication). Peaks appear in Velocity and Demod spectra: Plan replacement at next maintenance period. Peaks appear in Velocity spectra only, combined with a rise in the Demod noise floor: Replace the bearing now! 3

4 Cursor A: orders Hz in/s deg O/All.37 in/s -pk 8/3/27 8:11:3 AM O/All.37 in/s -pk RPM orders O/All.376 g rms 8/3/27 8:1:57 AM O/All.376 g rms RPM orders 3/13/28 WIND TURBINES ARE VERY EASY TO COMPARE TO EACH OTHER TO DETERMINE A FAULT Alarm Types - Envelopes - Vertical - Vel Freq 12 CPM VIBRATION ANALYSIS INTERPRETATION IS REDUCED TO TRAFFIC SIGNALS WITH ALL MAJOR CONDITION MONITORING SUPPLIERS O/All.251 in/s -pk in/s -pk orders 9/1/1998 7:25:2 AM O/All.251 in/s -pk RPM GOOD(.3) vs. BAD(.3) Generator Bearing Generator - GNDE - Horizontal [Tach] - Vel Freq 2 Hz [Tach] Generator - GNDE - Horizontal [Tach] - Acc Freq 2 Hz [Tach] 8/3/27 8:11:3 AM 8/3/27 8:1:57 AM MEGAWATT GENERATOR BEARING looks like this: NU224E.TUP2 FTFI in/s -pk g rms There are 2 ISO standards for alarms and 1 widely accepted 3 year study also used for alarming 4

5 3/13/28 VIBRATION SIGNATURE IS LABELED WITH BPFI (INNER RACE DEFECT) P load - down wind bracke - Horizontal - Demod (2-1kHz) 12 CPM [Tach] 12/14/27 1:33:47 PM Machine Note (12/14/27 8:18:19 AM) The generator run at full speed and full load. the test was done with the test rectifer. Please consider no fan on the last two tests..14 Cursor A: CPM orders.13 g O/All.71 g rms MEGAWATT CLASS GENERATOR BEARING looks like this: g rms , 4, 6, 8, 1, 12, CPM 12/14/27 PM g rms RPM 1:33:47 O/All GEARBOX PLANET BEARING looks like this: 5

6 i 3/13/28 in/s - pk BPFI -2.1% 617 BPFI -2.1% Gearbox - Planet - Vertical - Demod (2-1kHz) 1 Hz 12/22/25 3:1:51 AM Cursor A: 3 Hz 617 BPFI -2.1% 617 BPFI -2.1% 617 BPFI -2.1% orders.6 in/s O/All.322 in/s -pk EXAMPLE # 3 MW WIND TURBINE GEARBOX GEAR ISSUE There are 2 ISO standards for alarms and 1 widely accepted 3 year study also used for alarming GEAR MESH FAULT FREQUENCY = GEAR TOOTH COUNT x GEAR TOOTH COUNT x RPM Hz 12/22/25 3:1:51 AM O/All.322 in/s -pk 72.6 RPM PLANETARY GEARBOX SIGNATURES Power (in/s -pk).35 Gearbox - Planet - Vertical - Vel Freq 15 Hz 12/22/25 5:54:38 AM Alert: Low er Harmonic Bearing Frequencies Danger: Higher Harmonic Bearing Freq Overall.22 Cursor A: Hz orders.1 in/s Cursor B: Hz orders.3 in/s.2 Diff: Hz orders -.7 in/s.18 O/All.79 in/s -pk n/s -pk Hz 12/22/25 5:54:38 AM O/All.79 in/s -pk 72.6 RPM Data captured on stage 2 mesh on the planet at 2:54 p.m. Again, the problem appears to be with very early stages of gear wear on the carrier. No other faults were noted. As levels of vibration in all these tests seem quite low, and there is no substantial alarming; it is safe to conclude that we are seeing only early stages of wear in the gearboxes that require no immediate action, but warrant more frequent monitoring. 6

7 3/13/28 WIND TURBINE MODELS USED IN THE EXAMPLES ( in no particular order) VESTAS V-8 GE 1.5 Clipper 2.5 PREDICTING USING VIBRATION Know before wind season starts Know before the warranty expires Know what needs to be fixed Know when it needs to be fixed Know what parts need to be ordered Know if it s an up tower repair or crane call Know if it was rebuilt or installed properly 7