in our gas turbine bearing-3 metal thermocouple BTJ32A has started spontaneous hunting from 113 deg C normal to a high value of 129 degC and more and to a lower value of 108 deg C; It all started after a high value of 118 deg C in BTJ31A and later shifted to BTJ32A. This shifting of "leads" by these thermocouples was happening for an year. The high value of BTJ32A then went to a second-wise hunting at a higher loads of 70MW and above and less than 50MW.. in between it is steady and not giving a hunting,, Also the BTJ31A is showing a steady about 113 degC. There is no spare thermocouple in bearings as per the new design of GE. Does the bearing is affected with laminar to turbulent oil flow during the steady BTJ32A - load margins. Does any of you faced the bearing damage after a similar event. Does thermocouple will have such a fast transients of reading?

Could your expertise show some lights..

 

It may be noted here that the type of hunting observed is up and down the base value (steady value at a separate load band of 50 to 70 MW) on a time of each second..

 

What have you done to try to troubleshoot the problem?

What were the results of your effort(s)?

What is the #3 bearing drain temperature? Has it increased or decreased or remained the same at similar loads over time?

Does the lube oil header temperature remain relatively constant during the hunting?

Does this only occur on loading after starting, or does it also occur on unloading during shutdown?

Do you know what kind of bearing is in the #3 bearing housing--a journal bearing or a tilting pad bearing?

What have the vibrations done on the #3 bearing? Do they increase or decrease when the hunting occurs?

 

> What have you done to try to troubleshoot the problem?

This is a tilting pad, with load between pads. We have checked the thermocouple connections; terminations found okay though ground was detected in both thermocouples 31A and 32A. 31A was showing ground and some times not. mV measured was found correct as per the calculations, but drastic hunting was not observed in multimeter (not catched by multimeter, though +/- 0.1 mV variation was observed); Machine is running and hence we could not enter load compartment for inspection of thermocouple up to seal gland. Oil analysis for babbits were negative. Early history was taken after last bearing inspection, where Bearing pad aft-east side pad was found broken earlier, after which alignment was done (where BTJ32A is installed). After that some time s 31A was reading higher than 32A,and after some days 32A was showing more than 31A. This rotation of more reading went for some days...Afterwards 3A started reading higher temperatures, and later 32A has started reading high...it continued till it gives "fast jumping up & down readings"..

> What were the results of your effort(s)?

as explained above

> What is the #3 bearing drain temperature? Has it increased or
> decreased or remained the same at similar loads over time?

about 66 deg C.. no major deviations,,

> Does the lube oil header temperature remain relatively constant during the hunting?

yes ,... around 55 deg c

> Does this only occur on loading after starting, or does it also occur on
> unloading during shutdown?

machine is running... but on the interval of 55 to 75 MW, it is showing steady readings of 113-114 deg C ... No "fast jumping / hunting"

> Do you know what kind of bearing is in the #3 bearing housing--a journal
> bearing or a tilting pad bearing?

It is a tilting pad..5 pads 2 at bottom... GE Frame9E..Bearing-3

> What have the vibrations done on the #3 bearing? Do they increase or decrease
> when the hunting occurs?

It is not affected....

Primarily the doubt is whether the thermocouple will behave "fast hunting in seconds" in any case if it is healthy.. or is it faulty..If it is faulty why the high temperatures are recorded on Btj31A (other pad) earlier.. Is there any history of such cases in other plants...Is the steady temperatures on specific load gaps can be attributed to increased clearances or a change of oil flow from laminar to turbulent or anything like this...How much is the effect of metal to metal rubbing in Lube oil drain temperature, in self aligning tilting pad bearings ..?

 

By the way, a "jumping up and down" condition is often referred to as fluctuating.

A Speedtronic panel can accept either grounded or ungrounded T/C inputs, so that shouldn't be the problem.

You say that the pad where the problem T/C is located was found to be "broken". Does that mean that the pad, with the T/C embedded in it, was replaced with a new pad with an embedded T/C? (I'm presuming the T/Cs are embedded; you haven't said if they are or aren't, and I've only seen one very old machine (a Frame 5) which had removable bearing metal T/Cs.)

From the information provided, I'm also presuming that the problem with the fluctuating readings started after the "'broken" pad was found and after the alignment was done. Please confirm or explain further.

One of the things you haven't said you've tried is to exchange the two pairs of T/C leads at the Speedtronic T/C input terminal board. If the fluctuating reading was now seen on BTJ31A and BTJ32A was now stable, you could say the problem was NOT in the Speedtronic turbine control panel. If BTJ32A remains fluctuating when you exchange the two pairs of T/C leads, then it may be that the problem is in the Speedtronic panel.

By the way, are there any Diagnostic Alarms related to the fluctuating T/C reading or the card or I/O Pack the T/C is configured in?

If the readings change position when you exchange the pairs of T/C leads, there may still be some problem with the T/C extension wiring between the junction box closest to the T/Cs and the Speedtronic panel.

And, if the "broken" pad was replaced with a new pad with an embedded T/C, then it's entirely likely that the T/C connections weren't properly made up, though that seem unlikely for the fluctuating conditions you are describing.

You say the vibrations haven't changed and don't change with load, so that would seem to indicate that all is well with the oil flow and the bearings.

I have no experience with laminar oil flow versus turbulent oil flow, but I do recall from university that there should be no metal-to-babbit contact in a bearing because there should be an oil film covering the entire bearing surface, providing lubrication as well as cooling to the bearing and journal. There should also be an oil wedge under the shaft (I would imagine on a tilting pad bearing there should be one on each pad!) where it's compressing the oil as it's trying to roll up the bearing face(s).

Me, I'd try to rule out the Speedtronic, the wiring, and any electrical conditions. Have you checked the tightness of the screws of the terminations?

In my mind, I wouldn't think that the shaft, spinning at 3000 RPM, and the subsequent oil wedges and films could change temperature that quickly to cause the fluctuations you are describing. It's almost like there's some vibration causing an electrical termination to have poor contact. Vibrations do change slightly with load, and another thing that happens is that the bearing and coupling shift with changes in metal temperature (and remember, the #3 bearing is in the gas turbine exhaust!).

I wish I could be more help; but I've never encountered a problem like this. T/Cs are pretty simple; they usually fail in an open condition, but sometimes just before they fail they are "intermittent" and do fluctuate, but not just at certain times (or loads, as in your case). And, I'm not a petrochemical/oil engineer; just a lowly controls engineer. Hopefully, someone else may have another suggestion or two.

Again, if the problem didn't start until after the "broken" pad was repaired/replaced, at which time it would seem the T/C was likely disconnected and re-connected at a minimum, then I would be suspecting around the workmanship of the reassembly. I would be doing everything to eliminate every possible other source.

As a last thought, many times there are spare T/C extension pairs run between the junction boxes and the turbine control panel. You could try switching to a spare T/C extension pair for the problem circuit to see if that has any effect.

Please write back and let us know what the results of your efforts are.

 

Dear CSA thanks.. I am writing more details as per your queries..

> You say that the pad where the problem T/C is located was found to be "broken".
> Does that mean that the pad, with the T/C embedded in it, was replaced with a new
> pad with an embedded T/C? (I'm presuming the T/Cs are embedded; you
> haven't said if they are or aren't, and I've only seen one very old machine (a
> Frame 5) which had removable bearing metal T/Cs.)

The pad was having babbit pieces in positrion during last inspection, after which new pads weere used. Infact prior to the earlier isnpection there were a fluctuation to a higher reading of 120 degC but stayed for about 10sec and come back to 108 degC..After the pad replacement it has run for some time, and then the type of hunting in seconds was observed..The thermocouple is not embedded into the Babbit, but is the basemetal, around 1 to 1.3 mm below the babbit-basemetal interface..The MI sheathed thermocouple is inserted into a hole given on backside of the tilting pad, fixed with spotwelded shims and adhesive ouside the hole.

> From the information provided, I'm also presuming that the problem with the
> fluctuating readings started after the "'broken" pad was found and after the
> alignment was done. Please confirm or explain further.

As given above , the readings were normal, for more than 1-1/2 years, after which the fluctuations started..Infact before the high frequency probematic fluctuation, the thermocouple in other pad has read a 6 deg C rise, and then got normalised..

> One of the things you haven't said you've tried is to exchange the two
> pairs of T/C leads at the Speedtronic T/C input terminal board. If the
> fluctuating reading was now seen on BTJ31A and BTJ32A was now stable, you
> could say the problem was NOT in the Speedtronic turbine control panel. If
> BTJ32A remains fluctuating when you exchange the two pairs of T/C leads,
> then it may be that the problem is in the Speedtronic panel.


sorru I have not mentioned this..TWe have ruled out markV issues.. the cable pair for the MarkV wiring from field was changed.. and same phenomenon was happening..in BTJ32A..As it has shown some relation with load, we have not suspected the MarkV..

> By the way, are there any Diagnostic Alarms related to the fluctuating T/C
> reading or the card or I/O Pack the T/C is configured in?

None were there...

> If the readings change position when you exchange the pairs of T/C leads,
> there may still be some problem with the T/C extension wiring between the
> junction box closest to the T/Cs and the Speedtronic panel.

All terminations were checked and found okay.. MarkV side wiring was replaced and found BTJ32A is showing the fluctuations...But drastic changes in 1 Sec....!!!

> And, if the "broken" pad was replaced with a new pad with an embedded T/C, then it's entirely likely that the T/C
> connections weren't properly made up, though that seem unlikely for the
> fluctuating conditions you are describing.

----- snip -----

> Me, I'd try to rule out the Speedtronic, the wiring, and any
> electrical conditions. Have you checked the tightness of the screws of the
> terminations?

Yes....found okay..

> In my mind, I wouldn't think that the shaft, spinning at 3000 RPM, and the
> subsequent oil wedges and films could change temperature that quickly to cause
> the fluctuations you are describing. It's almost like there's some vibration
> causing an electrical termination to have poor contact. Vibrations do change
> slightly with load, and another thing that happens is that the bearing and
> coupling shift with changes in metal temperature (and remember, the #3
> bearing is in the gas turbine exhaust!).

\ like a casing slippage...But no conclusions were drawn..

> I wish I could be more help; but I've never encountered a problem like this.
> T/Cs are pretty simple; they usually fail in an open condition, but sometimes
> just before they fail they are "intermittent" and do fluctuate, but not
> just at certain times (or loads, as in your case). And, I'm not a
> petrochemical/oil engineer; just a lowly controls engineer. Hopefully, someone
> else may have another suggestion or two.

----- snip -----
> the junction boxes and the turbine control panel. You could try switching
> to a spare T/C extension pair for the problem circuit to see if that has any effect.

The recent supplies are coming with simplex type only.. No more Duplex is suplied..naturally so we are nt able to get any mV / continuity on the second cable pair there..

> Please write back and let us know what the results of your efforts are.

I will update when results are known...
Thanks..

 

>Dear CSA
We had a chance to open up the case again...

Surprisingly we got two issues:

(i) On the thermocouple side, the yellow wire in aft left side loaded pad, which was having the 32A thermocouple has three 0.5 mm length insulation damages at three location.Tye red wire of the K type was intact and no cuts, 'bruises' exposing the conductor were observed. I think these cuts does not have any relation to the rapid hunting on loads happened, unless the red wire is also grounded momentarily somewhere. The insulation tests of cable and the thermocouple itself, after the lube oil cutoff, was found excellent. Does lube oil can cause a continuity path to ground? I don't think so...

(ii) The loaded bearing pad was having cracks on the same place on Babbit, where the 32A thermcouple is located. Crack depths could not be ascertained, as no NDT tests were carried out. There were also some alignment issues. Load on the bearing also seems to be very high considering the cracks.

Bearing replacement and load reduction adjustments were done.

This is the case. Though there were confusions with the thermcouple, this leads us that the hunting/spiking happened in "seconds" period, can also happen in such dynamic cases.

Any comments..? Thanks..

 

Thanks for the information, RK.

A quick search using your preferred Internet search engine would tell you that oil can be a conductor. Some mineral oils are used as insulators in transformers, but in general, oil can conduct electricity, just not very well.

I have troubleshot grounds and found many a "soft" ground caused by the presence of oil in junction boxes.

As for the effects of grounds on T/C circuits, Speedtronic panels can operate with grounded or ungrounded T/Cs in any combination.

I would presume you will be replacing the bearing pad and T/C combination, and hopefully it will be mechanically reassembled properly and the alignment will resolve the loading issues, and you will see a stable feedback.

It will be interesting to see how the signal behaves after reassembly.

 

Taken to service ... temperatures are okay; New bearing; New thermcouples, New cables.... going smooth!!

 

Great news!

Best of luck going forward!