MARK M KIIGROUNDF FAULT CTIF TRE FICAT TION
Int troduct tionFauji Fertilizer Com mpany Limited has two Ur
plants nea Sadiqabad, d rea ar ctRahimYarK KhanPakistan.
Electricalpow wersourcetot thisAmmonia Distric Urea complex,aret
twoGasTurbin nesTG701& TG702,servin ngrelentlessly as a b
backbone of th plants for t past 30 ye he the ears. MARKII i
installed on is TG70 TG701 tur 01. rbine has had Speedtronic and
125VDC power supply groun faults since 2003. These ground fault
caused man mysterious nd e e ts ny outag gesontheturb
bine.Historyda ataanalysisrev vealsatleasto onetrippingor emerg
gency shutdow each year that can be attributed to these ground wn
faults. Attempts for the detection and rectifica r n ation of these
ground faults eral turnaroun with no s nds success. This year in
plant were made in seve 09,thislongaw waitedtargetw wasaccomplish
hed. turnaroundOct200 papergivesab briefoverview ofthesequen
nceofthejobs conductedto Thisp find b both the groun faults one by
one. It gives an insight t the hidden nd to proble ems that engi
ineers came a across while dealing with M MarkII control system of
General Electric, insta m alled on the G Turbine. It also unveils
Gas some ingenious ideas that help ped us to tac ckle with the
grounding
lem. probl
BackgroundTG701 MARKII control system has three possible sources
of ground faults namely Speedtronic Ground Fault, 125VDC Ground
Fault and Thermocouple Ground Fault. The Speedtronic Ground Fault
and 125VDC Ground Fault were hampering the turbines smooth
operation and functioning for the past seven years. The latest
tripping of TG701 occurred in September just one month before the
turnaround 2009 as a direct effect of 125VDCGroundFault. Ground
faults are extremely dangerous to any
typeofelectricalorelectronicssystems.Besides
outagestheycancausesparkinganddamageto hardware.
Groundfaultsareannunciatedwhenoneofthe two poles, positive or
negative, is grounded. The fault should be addressed in first
opportunity as grounding of second pole will
resultinsparkingandtrippingofthemachine. Based on above, the
overhaul of Speedtronic MARKII System and rectification of both
groundfaultswereplannedintheTA2009.
Oncethegroundfaultwasremovedthevoltage shiftswerenormalized.
SNO1 2 3 4 5 6 7 8 9 10 11 12
SUPPLY TERMINALP28 P28,PCOM
REGULATOR VDC28.03
SGDD VDC28
P28,GND PCOM,GNDP12 P12,ACOM
23.68 4.311.99
23.6 4.5411.9
P12,GND ACOM,GNDN12 N12,ACOM
7.52 4.3912.04
7.5 4.4611.99
N12,GND ACOM,GNDP5 P5,DCOM
16.3 4.65.08
16.15 4.34.92
P5,GND DCOM,GND
0.53 4.56
0.62 4.44
Table:1SpeedtronicVoltageLevelswithGroundFault
Ground fault indication was consistently persisting on the
Firstout Panel before turnaround.Itwasduetotheactuationofboth,
Speedtronic SGDD card and the 64D relay for
125VDCgroundfaultdetectors. Ground fault rectification activity
started with completeoverhaulingandcleaningofMARKII control system.
Following modules were dismantledandcleaned:
SpeedtronicPage:Speedtronicpageor1Lpage contains control and logic
cards of the system, 126 in number. All cards were taken out
checked,cleanedandinsertedback.
SystemMaintenance
Preliminary numeric and photographic sample
dataofMARKIIcontrolsystemwascollectedat
differentphasesofTG701shutdown.Thedata was later used for ground
fault analysis and comparativestudy,oncethefaultwasremoved. This
data helped us in determining the Speedtronic ground fault as
permanent fault and the 125VDC as intermittent fault. Thermocouple
ground fault was ruled out because the thermocouple ground fault
detectorcardSGDEwasnotonalarm. The Speedtronic voltage levels under
fault conditionareshowninthetablebelow.Aclear
abnormalshiftof4.5VDCinallpowersupplies
andCOMswithrespecttogroundishighlighted.
Fig1:SpeedtronicPageCards
Ontherearsideofthispagethousandsofwires
areconnected.Alltheseconnectingwireswere checkedandcleaned.
Fig2:SpeedtronicPageRearView Fig4:BulkPowerSupplyModule
FinalRegulatorModule:Allpowersupplycards
(+28VDC,+12VDC,12VDC,+5VDC,BusMonitor card) of Final Regulator were
taken to instrument shop, checked and overhauled. The
supplyconnectionsatthebackofregulatorrack were cleaned and lugs
were changed where required.
Black Start Inverter: Black Start Inverter was
dismantledfromthecontrolpanelandtakento
electricalshopformaintenanceandcheckingof agingcomponents.
Fig3:BlackStartInverterInstallation
Fig5:FinalRegulatorSupplyCards
Power Supply Module: Bulk power supply module was overhauled in
electrical shop. Its power switching cards EPSU and EPSV were found
in bad condition covered with carbon dust. These cards were
cleaned, checked and reinstalled. One capacitor was replaced as it
wasshowingzeromicrofarads.
Source Selector Module: Source selector was removed from the
cabinet, overhauled in the shopandreinstalled.
All cards and relays on this panel were also
checkedcleanedthoroughly.
Fig6:SourceSelector
Speedtronic Relay Page:All70relaysonrelay page were taken out,
cleaned and checked for contact & coil resistance. Relays with
higher contact resistance were replaced. Cable
connectionsatthebackofRelaypagewerealso cleaned. Mechanical
actuation of relays was alsochecked.
Fig8:RemoteAnnunciatorModule
Remote Annunciator Panel: MarkII UCR panel was also cleaned and
checked for any faulty instruments. Fused bulbs and lamp holders
werereplaced. All above modules were extremely dirty and had carbon
dust deposition. They were removed from their location for the
first time andwereattendedinworkshop. Overhauling of 10 local
annunciator cards and 17 receptacle cards was done at site.
Terminal strips HCTB, LCTB, and RTB were also cleaned. Twelve Input
Isolator cards were overhauled. Onecardwasalsoreplacedafterwards.
All fused bulb and lamp holder of local and
remotepanelswerealsochanged. Once the maintenance was completed
system wasreinstalled,integratedandpoweredup.
MARKIICalibrationFig7:RelayPageandRemoteAnnunciator
RemoteAnnunciatorModule:TheModulewas taken to the instrument
shop for overhauling.
Calibration of different controls and protection
loopswasperformed.Checking&calibrationof power supply, Speed
control loop, acceleration control loop, temperature control loop,
PCD bias adjustment, overtemperature, speed
protectionandstartupcontrolwasperformed.
therewasnogroundingonanyofthe125cards pulledoutof1Lpage.
Inthenextstepallpowersupplycables28VDC, +12VDC, 12VDC, 5VDC, PCOM,
ACOM and DCOM were disconnected on 1L page. These
are130powersupplycablesdisconnectedfrom
theverticalandhorizontalbusesof1Lpage.
Disconnectionsequenceofpowersupplycables was from horizontal bus
one by one from bottomtotop. Fig9:SpeedtronicCalibrationSetup
SpeedtronicGroundFaultSpeedtronic ground fault was persisting on
SGDDgrounddetectorcardandonthefirstout panel. It was necessary to
isolate wiring and hardware step by step while observing the
statusofgroundfaultindication. All 126 cards on Speedtronic 1L page
were pulled out one by one. Each card has 51 pins. Ground fault
indication kept persisting till the lastcardwaspulledout.
There was a problem in disconnecting power supply of the first
row of 1L page. This action would have disconnected the SGDD card
from power supply whereas its ground fault indication was important
to us in troubleshooting. A technique was adopted to overcome this
problem. SGDD card was powered up outside the 1L page. External
biasing was provided to the card through the wires directly from
Main BusBar.Thisstephelpedusalotincatchingthe ground fault. 1L page
was then completely disconnected from Main Bus Bar. Yet ground
faultIndicationpersisted,rulingouttheground faultin1Lpage.
Fig10:1LPageCards Fig11:SGDDCardWired&PoweredUpOutsideits
Slot
IntheendonlySGDDgrounddetectorcardwas
leftonthe1Lpageandgroundfaultindication was there on the card. This
confirmed that
Once the card was functioning all the power
distributioncablesofregulatedsupplies28VDC, 12VDC, 12VDC, 5VDC,
PCOM, ACOM, DCOM
were disconnected one by one from the main busbar.
traced and it led us to the fuel gas pressure transmitter96FG.
To conclude the grounded part, 96FG was disconnected from the area
JB while 12VDC still connected to main bus bar. SGDD card
indication remained reset. The situation was
verifiedseveraltimes.
Fig12:MainBusBarsof28VDC,+12,12,+5&COM
In the above step REMAN module, Local annunciator module,
thermocouple processing module and synchronizer module were
disconnected.SGDDgroundfaultindicationstill
persisted.Atotalof235cableswereremoved while isolating all above
modules from Main BusandtheirrespectiveTerminalBlocks.
Fig14:SGDDCardIndicationReset
Voltage level of complete system was rechecked for analysis.
Voltage shift in all regulated supplies and COMs normalized as
highlighted in the table. No effective voltage
betweensupply,COMandgroundwasfound.It
wasconcludedthatSpeedtronicgroundfaultis removed from MARKII
system. Complete system was reassembled. All cables/modules
wereconnectedonebyonetodetectanyother possible cause of Speedtronic
ground fault. Finally1Lpagewasreconnectedtothebusbar. SNO1 2 3 4 5
6 7 8 9
SUPPLY TERMINALP28 P28,PCOM P28,GND
REGULATOR VDC28.03 28.08
SGDD VDC28 28.1
PCOM,GNDP12 P12,ACOM P12,GND
.7211.99 11.86
.7211.9 11.83
ACOM,GNDN12 N12,ACOM N12,GND
.7312.04 12.3
.7311.99 12.15
ACOM,GNDP5 P5,DCOM P5,GND
.765.08 4.9
.735.02 4.89
Fig13:ThermocoupleModule&Annunciator
10 11 12
DCOM,GND
.75
.71
After disconnecting about 50 more cables we found the suspected
cable on 12VDC distribution circuit. The cable was physically
Table2:SpeedtronicVoltageLevelswithoutGround Fault
No ground fault appeared in reassembling the
system.SGDDcardwasreinstalled.
125VDCGroundFault125VDC Ground fault was categorized as an
intermittent ground. It appeared occasionally on 64D ground
detector relay and would reset immediately. The data pertaining to
125VDC powersupplywascollectedforanalysis.
Therewasanabnormalshiftin125DCvoltage levels indicating a Ground
Fault. This type of Ground Fault is hard to find as it is subject
to circumstances. SNO FUSEFU12 FU7 FU6 FU5 FU4
TERMINALFU12+,GND FU12,GND FU7+,GND FU7,GND FU6+,GND FU6,GND
FU5+,GND FU5,GND FU4+,GND FU4,GND
ASFOUND
Fig14:FaultyFuelGasTransmitter96FG
1 2 3 4 5 6 7 8 9 10
New transmitter 96FG was installed in the area. During a
thorough examination of faulty 96FG in the workshop the black
signal cable of 96FG was found showing only 1.7 resistance with
ground. The Ground Fault persisted even when the suspected cable
was desoldered from the PCB board. Finally, it was concluded
thatthesignalterminalandgroundterminalof
PCBboardof96FGwereshort.
VDC 128 12 132 1.2 132.5 1.55 133 .09 132.6 .07
NORMAL VDC70TO60 70TO60 70TO60 70TO60 70 TO60 70TO60 70TO60
70TO60 70TO60 70TO60
Table3:125VDCVoltageLevelwithGroundFault
During one of its random appearances the Ground Detector Relay
64D remained actuated for some time. Immediately, all the fuses
were pulled out one by one starting from FU7. The ground fault
normalized as soon as FU4 negative supply fuse was pulled out. The
grounding appeared again immediately after
thereinstallationofthefuseFU4. FU4 fuse supplies power to all of
the input isolator cards ISOL. All these cards and their respective
circuits were checked. Some variation in the voltages with ground
was observedonISOLSIIE0Gcard.
Fig15:New96FGInstalled
The 125VDC ground fault was finally found in the cables entering
the MARKII cabinet from bottom. These cables were routed over a
channelsupport.Onthetopofthesecablewas a horizontally bolted plate.
These cables were pinched in between the support and the top metal
cover plate. Any weight / load on the upper plate would press these
cables against thesharpedgesofsupport. Fig16:125VDCFuses
The voltage level on SIIE0G card was found between 56 to 48 VDC.
These inputs were related to Boiler damper and Bypass damper logic
and were coming from AIU4 cabinet. Some pairs of this multicore
cable were also terminatedonRTBandHCTB1. This multicore cable was
disconnected from both ends the turbine and AIU4 side in UCR
rackroom and cable was meggered and were found healthy.
Furthermore, AIU4 logic was
alsoanalyzedandcableswerephysicallytraced.
MeggeringofareacablesofBoilerDamperand
Fig17:Topplatetakenasidetoseethecablesoverthe Bypass Damper was
also performed. Although verticallymountedsupport the variation was
noticed in voltage levels but
evenafterdisconnectingtheAIU4cabinetfrom
Twooutofsevenpairsofthiscablewereinuse. Speedtronic system 125VDC
ground fault One of the cores had a major cut while others appeared
regularly. Hence it was obvious that hadminorcuts.
AIU4didnotparticipateinthegroundfault. Ground fault checking
procedure was also incorporatedinroutinechecking/calibrationof
switches, limit switches, fire detectors, probes,
FDs,SparkPlugetc.Specialformatsweremade
foreachinstrument.Grounddetectorrelaywas still energized after all
these checks were performed. In subsequent step SOVs were checked
for the ground fault. 20FD liquid fuel SOV was found grounded and
rectified immediately. Same procedure was repeated for liquid fuel
purge SOV 20PL and fuel clutch SOV 20CF. Still 125VDC Ground Fault
persisted in the system indicatingmultipleGroundFaults.
Fig18:CutontheGroundedCable
The cables were then insulated, protected and properly routed to
their respective locations. The cables which were not being used in
the
systemwereseparatedandadequatelysecured forfutureuse. The Ground
Fault indication on Firstout panel reset immediately as soon as
125VDC ground fault was rectified. Voltage levels normalized
aftertheremovalof125VDCgroundfault. SNO1 2 3 4 5 6 7 8 9 10
Power supply switching protection system is available on bulk
power supply hence it is at lesser risk than the Speedtronic. Still
the 1L page, 1R Page, Local Annunciator, and Remote
Annunciator,ThermocoupleProcessingmodule, Flame Detector Module,
Spark Plug ignition Module, synchronizer and Relays could have
directlyaffectedbygroundingoftheotherbus.
FUSEFU12 FU7 FU6 FU5 FU4
TERMINALFU12+,GND FU12,GND FU7+,GND FU7,GND FU6+,GND FU6,GND
FU5+,GND FU5,GND FU4+,GND FU4,GND
ASLEFT VDC67.2 64.3 67.3 64.7 67.2 63..93 67.0 64.7 67.13
64.77
NORMAL VDC70TO60 70TO60 70TO60 70TO60 70TO60 70TO60 70TO60
70TO60 70TO60 70TO60
AcknowledgementWe whole heartedly thank the guidance and
directions of our seniors who paved the way
towardsasuccessfulcompletionoftheproject. Tanvir Vaqar is senior
department managerElectrical & Instrument at Fauji Fertilizer
Company Limited (FFC), Sadiqabad Pakistan. He holds BE degree in
electrical engineering. Over the past 27 years he has been working
in the area of electrical
powergenerationanddistributionnetworks,control & protection
systems design, modification and maintenance.
Table4:125VDCVoltageLevelwithoutGroundFault
RecommendationsOnce the Ground Fault is removed from the system,
more responsibility comes on our
shoulders.Followingareoursuggestions. 1.
WhiledetectingaSpeedtronicgroundfault
theSGDDcardcanberemovedfromitsslot andbiasedfromMainBus 2. While
biasing SGDD card externally never
insertanotherSGDDcardinslot`1L0Tasit
wouldgiveambiguousresults.Useonlyone SGDDcardtodetectGroundFaults
3. 125VDC ground detector relay should be cleaned and checked
before starting to searchforthegroundfault.
AsgharBhuttaisSectionHeadofInstrumentIatFFC.HeholdsBEdegreeinElectronics.Hisareaof
expertiseisutilitiescomplexcontrolsystems.He
wasinvolvedintheerection&commissioning
ofUreaplantexpansionunitin1993.
AuthorsAdeel Imtiaz is an Instrument Engineer at FFC, Pakistan.
He holds M.phil and M.Sc degree in electronics. His current
assignment is Waste Heatboilersandturbinecontrolsystems. Usman
Ghani is a Trainee Instrument Engineer at FFC, Pakistan. He holds
BE degree in Electronics. Currently he is working on
DistributedControlSystem.
ConclusionGround fault has existed in the TG701 in
differentformsconsistentlyfrom2003to2009.
Negative12VDCsupplyof96FGwasgrounded. If the positive 12VDC supply
cable was grounded for even a moment, it could have caused severe
results. By design there are no protection fuses in Speedtronic as
in 125VDC circuit.