The kinemantic scheme with the measurement locations and several representative spectra and timewaves are in the attched file. The dominant vibrations are in the horizontal direction of measurement while at the same time I noticed also high levels of axial vibration at the end bearing of the gearbox.
The frequency spectra are dominated by the harmonics of the gearmesh and there can be see some modulation with the pinion's speed. Gears are new and from other manufacturer. Compared with the previous gears on this cement grindng ball mill, new gears have a larger modulus and different number of the teeth but with the similar gearmesh - now 61.8Hz instead of 57Hz before. The old gears which were worn had smaller vibration values - around 12 mm/s while now we have, with the new gears, around 28 mm/s with gearmesh frequencies and their harmonics... The girth gear pinions were aligned according to their standard procedure and within limit values.
The axis of a pinion shaft is well misaligned compared to the output shaft of the gearbox - I tried to measure the alignment with lasers but I couldn't get the laser beam in my targets... Anyhow between theese two shafts is the 7 meter long shaft with gear couplings which, they say, can compensate for some misalignment.
Also when cement grinding ball mill in operation large movements of this shaft are visible. It would be logical to do the alignment between the gearbox and a pinion but it is not possible to move the motor and a gearbox. Now, when the cement grinding ball mill is started, vibration values are around 14 mm/s, first 4 harmonics of the GM, and then for about half an hour they go up to 28 mm/s - in horizontal direction while in the vertical and axial direction there is no such change.
There are some other guys included in this problem and some have an opinion that this could be related to the resonance. They did some FEM calculations and got some natural frequencies close to forces frequencies. I did several bump tests on a girth gear and on a long shaft and I did not record any of the natural frequencies near the GM and its harmonics.
Also the peaks in the spectra are very sharp and I suppose that if excite the resonance there should be some visible haystack beneath the frequency component. Also if there is a natural frequency excited: 1. I suppose that there would be a dominant peak and not harmonics. 2. The situation would be similar in the case with the old gears.
When you look at the time averaged timewaves recorded immeadiatelly after the start and when the cement grinding ball mill is warm - they are different. Both of them show 4 impacts in one gearmesh cycle, but they occur at different locations in the gearmesh cycle. Also there is one gear specialist involved and according to his calculations and measurememnts a tooth interference is very possible.If we get lower vibrations the it confirms with an idea of interefernce but if not...?
