Acosutic signatures of gear defects using time-frequency analyses and a test rig

This page in under construction

Test rig

 
 
 
 
 

It consists of two gear trains with the same ratio mounted back-to-back to ensure power recirculation through two separate shafts rotating in opposite directions. Power is dependent on the rotational speed and a torque precontraint applied to both gears at standstill in the following way:  loosen one of the semi-flexible gear couplings tapered on one of the shaft while other couplings are fastened (1) lock  the gear train farthest from the loose coupling in the torque transmission line (2) apply a static torque at one end of the loose coupling (3) fasten the loose coupling back: the two shafts then transmit torques in opposite directions (1) finally unlock the gear train (2) Both gear trains are the driven by a variable-speed ac motor (4) whose power cpmpensate for mechanical losses of the rig. Acoustic responses are measured by accelerometers mounted on the casing of the gear trains next to the ( healthy) bearings supporting the gears.  An optical keyway is mounted on both shafts to index meshing gears properly. It is sensed by a Hamatsu photocell generated one TTL pulse/rev for each shaft.  Accelerometers and tops are connected to National Instrumenst AT oe PCI MIO-1 daq card with a special software acquiring acosutic respnses of the acceleromerts at equally spaced angular positions of the shafts (corresponding at least at twenty repsonses per tooth meshing interval) counted from the optical keyways.

Spur gears

The gear ratio for the spur gear setup is 53/27. Gear module is   while the distance between shaft center is 200 mm.

• Noise generated by sliding tooth profiles. Remember that involute gears roll when profiles are in contact along the pitch circle and slide otherwise. This particularly holds true with spur gears that can periodically stop generating  high frequency noise when tooth profile roll on each other at the meshing frequency. Abrasion and lubrication play a major role in this pattern.
• Sliding may be modulated at the  rotational frequencies of shaft when these run eccentrically like teeth with variable addenda and dedenda could behave. •Pitch, profile (watch end and tip relief, if any) and helix errors can excite many frequencies as shocks would do. This can be watched both in time (spikes) and time-frequency domains (vertical rays showing up at the resonances of casing and accelerometers).
• If gears are mounted on roller bearings, the cage motion of these bearings can modulate the transmission path of the acoustic waves  to the outboard accelerometers monitoring the acoustic waves (located after the outer ring)
• Another fault that can generate typical signatures is tooth cracks that can behave somewhat like a pitch (transmission error).

Some pdf documents to download

Just click for a free (!) executable Virtual Instrument and data files to train in joint-time-frequency analyses (JTFA) with real gear train signatures:  give a try with a 9 Mbytes ftp download to go through...from the next web page..

.

Will contact the ULB specialist on this topic?