GT Loss of flame trip

J

Thread Starter

Joy

February- 2006
One of our Gas turbine (Mark V , 6581B) tripped on loss flame . Following alarms logged in the trip history.
Trip history showing
FSGR suddenly dropped from 18 % to 0.6%,
FPG2 dropped from 16.24 Bar to 10.3, then 6.6
FSG changed from 45 % to 43 % then 53%
Exhaust Temp TTXD 1,4,7,10,13,16 readings were in the trip history was showing minus value.

0433-12:19:05.406- True-Water cooling temp sensor fault
0110-12:19:05.468- True-Compressor Inlet T/C disagree
0419-12:19:05.468- True-Lube Oil tank T/C #1 fault
0257-12:19:05.468- True-Gas fuel temp high spread
0366-12:19:05.468- True- Air Atomizing temp measurement fault
0060-12:19:06.781- True- Loss of flame trip

We have replaced SRV servo valve and <R> core TCQA card.

Last September this unit tripped on loss of flame alarm and that was due to trip coil 20FG failure.

April-2006
Same unit tripped with exactly same alarm and trip history.
Checked 20FG coil, terminations, stroked SRV , everything working fine.
Replaced TCQA <R> core again. Now the Unit is running normal.

What could be the root cause of these trips ?
Is there any ribbon cable problem?
Any body knows the mark V card interconnecting ribbon cables GE part number ?

Earlier this unit was affected lots of signal traffic problems, so we disconnected local HMI from the <C> , so only one multiunit Remote HMI is connected to <C> core now.
After this modification following alarms not appeared so far.

DCC BMS: OUT OF MEMORY
DCC UNABLE TO GET DPM BUFFER FOR XMIT
DCC DPM: NO BMS MEMORY FOR ISR
DCC DPM: INVALID DESTINATION ADDRESS

How to make RC filters in the solenoid circuits?

Kindly analyze these problems

Kindly share your experiences and suggestions.

Thanks in advance
Joy
 
Dear JAS,

Unit is tripping during normal operation. Sir, We are really confused with all these alarms, Is it any hardware problem or software issue???

Regards,
Joy
 
This appears to be a re-post of an event first posted in Nov, 2005 (http://www.control.com/1026216284/index_html), and in a similar post after that.

Have you verified the polarity of the servo current be applied to each of the three coils of the SRV electro-hydraulic servo valve? It appears that there is something which is "interrupting" the T/C readings for <R> processor. If <R> were to drive its servo current positive (to close the valve) and one of the other servo coils had the wrong polarity applied to it, then the combined servo currents would act to remove the flow of high-pressure hydraulic fluid from the SRV actuator and the SRV would slam shut--and a 'LOSS OF FLAME' trip condition would occur and be detected and annunciated.

Another troubling thing previously reported was the "TCQA <unsed> #143" Diag. Alarm, which seems to indicate a mismatch between the TCQA Card Configuration File used during the PROM upgrade (TIL 1480-2) and the one required for the version of PROM installed on the TCQA card. It's not normal for "<unused> #nnn" Diag. Alarms to be annunciated unless there is some kind of card configuration file/PROM version mismatch.

It has been requested to report any EXISTING Diag. Alarms prior to the trip condition, but that information has not been provided.

Something seems to be "interrupting" the T/C data in <R> processor. The TCQA card has been replaced; TIL 1480-2 has been implemented (at least partially--the "<unused #nnn>" Diag. Alarms are troubling and could be an indication of something--GE should be consulted about this, especially if these Diag. Alarms DID NOT occure before the TIL was implemented!!! and this was suggested before, also...); the ribbon cables have been visually examined.

The TBQA card on <R> is where all the T/Cs for <Q> are terminated. They are divided on the terminal board into groups of 15 (for <R>, <S>, and <T>) and then ribbon cables carry the signals to the control processors. Have you tried replacing the TBQA card on <R>? You could even just swap it with the one on <C>--it's a lot of work, but if the problem follows the card, then that card could be replaced with one from spares).

GE has also advised people to use a non-conductive grease on the ribbon cable connections; has this been used on all the T/C ribbon cables and on all the ribbon cables on the TCQA in <R>--on both ends of the ribbon cables connected to the TCQA? (Use the signal flow diagram of the Case Layout Wiring Drawings to trace all the ribbon cables.)

Another idea might be to replace the TCQC (or whatever card is located in Loc. 4 of <R>), and/or the TCPS of <R>, since the power for all the I/O cards associated with <R> comes from the TCPS in <R>.

It's most likely a hardware issue, or a card configuration file issue. It has happened more than once during a PROM upgrade, that one processor (out of three) was not properly upgraded (one MUST make a detailed procedure for each and every processor and card and PROMset being upgraded, and each and every step MUST be checked off when exchanging PROMsets...failure to do so can have very unexpected results).

Another thing which good to do during major maintenance outages is to, with the processor powered down, unseat and reseat all the ribbon cable connectors in each processor and core, one at a time, just to try to help the corrosion issue on the ribbon cable connectors. This author has been in MANY countries where ambients are very hot and humid, and the people who were working in the Control Compartments where the Mark Vs were located were NOT comfortable with air conditioning and would continually leave the doors open. So, the commissioning personnel would lower the A/C (air conditioning) thermostat settings even LOWER. Then, during the night when nobody was working in the Control Compartment, everything would get bloody cold, and the workers would come in in the morning and throw the compartment doors open and the humidity in the air would condense on the metal doors and printed circuit cards. Sometimes, it would look like it had rained after an hour or so of this in the morning!

Failed or failing air conditioners can also cause problems in the Control Compartment. The purpose of the air conditioners (which usually also have heating elements in them) is to CONTROL HUMIDITY and keep the electronic components cool. Not so cold that moisture condenses, but no so hot that humidity can work its way into cable connections, etc. If you see evidence of rust on motor starter contactor coil pole faces, it's a good sign that there is or has been a humidity control problem at some time. Are the air conditioners working properly in the unit which is experiencing the problem? Have they been working properly?

You need to try to find out what's different with that unit/site. Again, VERIFY the SRV servo coil polarities--one quick way to check is to report the servo currents for all three processors for the SRV when the unit is running stably. If the P2 pressure transducers are calibrated properly, the servo currents for all three processors should each be APPROXIMATELY -2.67%, +/- 1.33%, or from -1.33% to -4.00%. If you have LARGE mismatches in servo currents between the three processors (i.e., one which has a positive value and the other two are more negative than expected, or the magnitudes of all three are just much higher or lower than expected) then provide the values of P2 pressure being seen by each of the three processors (from the Prevote Data Display). The easiest way to see all three servo currents is to use the AutoCalibrate display just to view the LVDT info and servo current info. You can even quickly print a "snapshot" of an instant in time for all three processors, rather than one at a time using DIAGC (which may or may not be properly configured for the PROM versions installed in the Mark V!!!).

Let us know what you find, please.

markvguy
 
From the data printed on the outside of the RC filter GE commonly used across 125 VDC solenoids connected to Mk V turbine control panels:

BEC-9636
RC1971
.5MFD+/-20%
250VAC-600VDC
220OHM1/2WATT
104X122CA002

(The last number at the bottom is GE's Part Number.)

markvguy
 
Top