Started 4 months ago with failing diodes (one or two) every 2 weeks (exciter/rotor/stator were megged and deemed ok), then every week (AVR was replaced with a spare) , then every day, then a a rewind of exciter rotor because it burnt (one of the phases overheated and shorted diodes were found). After rewind (4 days), remanent field was very hard to restore and sensing winding before turning on AVR sometimes showed 25V instead of usual 210V for this unit.
Synchronization is supposedly ok. Vibration levels are ok but operators feel it's different from say 1 year ago. Air is polluted with sawdust.
Most of the trips were over current or loss of field excitation. Logging system shows a reactive power positive peak (two times normal output) followed by a near zero down peak. Both peaks have a duration of less than 2 seconds, current shoots up during positive peak and usually trips. Peaks are more common during day time rather than night.
Specialists are suggesting a High-pot test (we don´t appreciate the idea of additional damage), a new rotor, or a new generator.
Our next step should be rotor removal and inspection. Any suggestions for a more precise diagnostic and less downtime?