Battery Ground fault on a GE 9OO1EA Gas Turbine


Thread Starter


We have GE 9001EA gas turbines, and we are suffering from intermittent 125 VDC ground.

The DC ground is existed as soon as the turbine get 40% of its speed (1200rpm)-exactly when the flame is appeared-, and stays in its situation. But the ground fault disappeared as soon as the flame is disappeared.

-What can the reason of this battery ground?

-Which devices that come into play at this phase?

NB: I verified the clutch coil of heavy fuel pump and the isolation was good.

Waiting for your help please.

we have two 7ea machines also faced the same problem and still problem persisting.

I mean DC ground fault alarm is appearing sometimes we are monitoring through trend in mark 6 both positive and negative.duration of ground fault is within seconds(4-10) only.

Please see the below findings in rectify DC ground fault but as i mentioned still there is some problem in cables we are trying how to rectify problem

* Machine tripped on diagnostic alarms from base load . It was observed that pressurizing fan, Aux. hydraulic oil pumps and ratchet system were not running. forced the required signals to put machine on ratcheting.

It was found that F#25 fuse(nominal rating-1.5 A, 250 Volts, Short Circuit Rating-100A) was blown on PDM (Power Distribution Module) card of MK-VI, which caused the loss of power to 1D3, 1D4, 1D5 and 1C5 boards.

* After further checking it was suspected that some of DC loads were grounded.

* Gas Turbine door switch 33DT11 wire insulation found in bad condition. All six door switches power supply disconnected from MK-VI side to avoid DC grounding. Replaced the blown fuse F#25 on PDM and all alarms got reset. Forcing removed and machine prepared for startup.

* Compressor bleed valves stroking carried out and found normal.

* start command initiated to Unit-1 at Unit-1 again tripped on same fault at 3535 rpm. Contacted to G.E. for further diagnosis. G.E. was also in our direction of locating ground in system but exact area was not suggested by G.E. They additionally suggested replacing the PDM card of was replaced

* As decided, all the cables connected from Mk-VI cards 1D3, 1D4, 1D5 and 1C5 to field instruments were meggered with 100 V megger to check the IR values of field devices. The cables for following field devices were showing low IR value.<pre>

The cable from Compressor Bleed valve-2 limit switch to MARK VI panel including ceramic connector at Turbine Compartment end was replaced by laying down another temporary cable.

* Remaining cables disconnected from MARK VI panel cards and from field devices as they were for indication purpose only.

* Haz Gas Detectors (5A, 5B in Turbine Compartment and 9C in GVM) were forced by C&I.

* GT#1 was standstill during this period as PDM card was isolated for above purpose.

Finally its common wire was not in circuit as it was out with other grounded wires.

GT#1 tripped at 11:23 hrs at 3561 rpm with same diagnostics alarms appeared earlier.

* C&I noticed that three fuses blown off (F#21,F#23 and F#25) on PDM this time. VCMI Communication card also failed this time.

* Safety Document issued to work on MARK VI panel for rectification of problem with PDM card and meggering cables to field instruments.

* Almost 500 cables were meggered by C&I during the day to measure IR values of field instruments. IR value of 4 more cables of control circuit observed low during this measurement.

* GE TA arrived at site at 16:30 hrs.

* Received 3 VCMI cards from JGP at 21:00 hrs.

* VCMI card was replaced and configured by GE TA.

* Following alarms appeared during Unit start up.
23:03;19:307 G1 1 Q 0435 <T> SLOT 6 VAIC DIAGNOSTIC ALARM
23:03:19:307 G1 1 Q 0458 <Z> VPRO DIAGNOSTIC ALARM
23:03:19:307 G1 1 Q 0548 <T> SLOT 7 VAIC DIAGNOSTIC ALARM
23:06;12:792 G1 0 Q 0458 <Z> VPRO DIAGNOSTIC ALARM
* 6 EXHAUST Thermocouples went to open circuit status while other Thermocouples were reading 100 Deg C. Informed to GE TA and C&I to check in the field.
* GT tripped at 23:17 hrs at 3528 rpm on same diagnostic alarms.

* It was informed to GE TA and it revealed that GE TA was analyzing the problem based on the information that GT tripped at different speed at all three times. However all the time list of events / alarms were submitted to GE which were sufficient enough to have full data required for analysis. GE TA confirmed the same with trends on MARK VI.

* GE TA collected required data and sent to GE Engineering for further diagnosis.

* F#25 fuse in PDM card was replaced and GT ratcheting started.

* C&I and Electrical continued further checking of cables in field and MARK VI panels in PEECC#1 based on the events occurring above 95% speed of GT. During checking, C&I observed imbalance in DC voltage in the system.

* SRV and Gas Control valves coils checked by C&I for any damage.

* Further checking of grounding in system was taken up in Accessory, Turbine & Excitation compartments. All cables in turbine compartment JB were seen in bad condition. Again bleed valve limit switch cable was found ground. Used spare available core. There was high potential of inter-turn shorting of different signal cores in same cables inside turbine compartment.

* Further grounding was found in excitation cable from GT excitation compartment JB to Generator Exciter damaged (insulation open). Picture pasted hereunder shows the status of cable inside conduit. The same was repaired.

During machine start-up lot of dips in DC voltage were seen on trend which shows still persisting of ground in the system finally achieved baseload

Difficulties Observed in Troubleshooting:

DC ground fault was not continuously persisting and number of grounds were many. Moreover ground was appearing for very short period as well as only during the tripping time (intermittent) hence it was difficult to locate. Difficulties were faced in locating these ground faults as number of cables is huge (480 nos). Removal and re-connection of each control cable on each card was consuming lot of time.

Trends below show status of DC voltages and observations w.r. to groundings. Split between the +ve & -ve values are clearly showing the status at different periods.

Grounds in Turbine control system (DC system) pull down voltage of DC control system. This results in loss of turbine control system and hence tripping of machine with diagnostic alarm.

Single ground fault in DC system will not affect the normal operation of system but when this turns to multiple grounds; it will pull down the voltage and will result in loss of control.

Bleed valve and Excitation system comes in line at around 90 % speed. whenever these loads were getting energized, due to multiple grounds in these loads, turbine control system DC voltage was dropping and hence machine was tripping with diagnostic alarm.
Dear Kumar,

I have never seen so many field cables grounded at the same before. Do you have a heat leak problem anywhere near your JB?

Kumar, you say you have two machines. Are both machines experiencing the same problems? Or just one?

And when did these problems start? How long have they existed?

I have seen a similar problem on a Mark VI installation once before, and it was very difficult to troubleshoot and find, but it was eventually found and resolved. The problem was ultimately traced to problems that had occurred during commissioning, and one of the processor racks had developed an internal fault. Further, there were multiple grounds on the positive leg of the 125 VDC supply (in the exciter compartment, in the generator protection panel, and on one of the solenoid outputs (20VG-1, as I recall)) and when a second ground developed on the negative leg of a solenoid supply (20PG-2, as I recall) this caused all manner of problems, including fuse-blowing and T/C problems (through the processor rack internal fault).

It's important to clarify some of the statements made in Kumar's posting. First, a Speedtronic panel can withstand multiple grounds <b>on EITHER the positive OR the negative leg of the DC power.</b> In other words, there can be two, three, or 14 grounds on <b>EITHER</b> the positive <b>OR</b> the negative leg and the panel will not be adversely affected.

<b>However</b>, if there are grounds on <b>BOTH</b> the negative and positive legs, that's effectively a short directly across the battery! There could be nine grounds on the positive leg, but if there is also one ground on the negative leg that's, again, just like putting a short directly across the battery.

The second thing to be aware of is that grounds on <b>any device powered by the 125 VDC source being used to power the Speedtronic panel will be detected and reported by the Speedtronic panel.</b> In other words, if the same 125 VDC battery that is used to power the Speedtronic panel is also used to power a fire protection system, and the Emer. L.O. Pump, and the Emer. Seal Oil Pump, and the excitation system, and the generator breaker close- and trip circuits, and the Generator Control Panel, then a ground or grounds in any of those circuits and devices will be detected by and reported by the Speedtronic panel. <b>The Speedtronic panel cannot distinguish between devices connected to the Speedtronic and devices not connected to the Speedtronic.</b> So, a 125 VDC battery ground alarm annunciated by the Speedtronic is not limited to devices connected to the Speedtronic.

In other words, if you're confident that you have properly checked all of the devices connected to the Speedtronic panel and have not found any grounds, then you need to continue to investigate ALL other devices and circuits connected to the 125 VDC battery that's being used to power the Speedtronic panel.

The third thing to be aware of is that devices and instruments connected to the Speedtronic that are not directly powered by 125 VDC are not included in the ground detection. For example, servo-valves, LVDTs, mA transmitters, passive- or active speed pick-ups, etc., are not included in the 125 VDC Battery ground detection circuit. This is because the 125 VDC is converted into the necessary voltages required by these devices, and this conversion (in the Speedtronic power supplies) isolates them from the 125 VDC battery. So, it's virtually pointless to troubleshoot or meggar wires and cables of devices that are not directly powered by the 125 VDC battery.

So, what devices that are connected to the Speedtronic are directly powered by the 125 VDC battery?

- Discrete (contact) inputs

- Solenoid outputs (20CB-1, 20VG-1, 20PL-1, etc.)

- Trip solenoids (20FG-1, 20FL-1, etc.) connected to the protective cards of the Speedtronic

This should make troubleshooting grounds easier--the fact that one doesn't have to be concerned about LVDTs and servo's and mA transmitters. There should be fuses and switches in the Power Distribution circuitry to isolate the 125 VDC power to the discrete (contact) input terminal boards and the solenoid output terminal boards, as well as to the protective terminal boards (in a Mark VI, that would be the TREG and TRPG). If the ground is present when the unit is not running, it should be possible to isolate the discrete (contact) inputs, and then the solenoid outputs, and if the ground goes away during either of these tests one can then further isolate them to individual terminal boards (the 125 VDC power is usually jumpered to multiple terminal boards, so they cables can be disconnected at the terminal boards one at a time to see if the ground goes away).

As for Kumar's problems, it seems there are multiple grounds, on both the positive and negative circuits. Where these grounds are is anyone's guess at this point. A LOT of devices are energized and de-energized when the unit reaches 95% speed, not just the excitation system and 20CB-1. Exhaust frame blowers start, sometimes compartment vent fans start, pumps are shut down, solenoids are energized and de-energized. One has to review the application code in the Mark VI to be certain, but each circuit could easily be troubleshot.

One of the most common places for grounds to occur after commissioning is in junction boxes which are outdoors, and are exposed to ambient conditions (heat, humidity, rain, etc.). Very often, poor construction practices result in conduits being run into junction boxes from the top (which can allow moisture to enter through the conduit penetration). Rain can run down conduit, and then through the penetration into the junction box.

A lot of times the wrong junction boxes are used for some external locations.

I've visited several sites that had steam packing leaks that were dripping continuously and directly onto junction boxes that weren't properly installed and sealed, and when opened they were completely full of water.

Another very difficult ground to find involved some pressure switches in the discharge of the exciter rectifier cooling fans. These switches were connected to discrete inputs to the Speedtronic panel, but were very difficult to access. They were found to have gotten wet from a heavy rain storm and poorly sealed compartment enclosures.

Sometimes, the vibration switches of Cooling Air Fans are connected to discrete (contact) inputs of the Speedtronic panel, and water/moisture can get into these junction boxes if not properly installed or sealed.

At one site, the site's building fire detection system was powered from the same 125 VDC battery that was used to power the Speedtronic panel. The ground was found on a remote switch in the building. This building fire detection system was NOT connected to the Speedtronic panel, but a ground in the circuitry that was directly powered from the same 125 VDC source was causing a ground to be annunciated on the Speedtronic.

At another site, a long-time ground in a generator field flashing circuit was found to be causing problems during start-up.

Grounds have also been found in transformer protective devices (over-pressure relays; level switches; etc.) where the conduit and wiring was not properly installed and sealed. These devices were powered by 125 VDC from the Generator Control Panel (which is usually on the same circuit as the Speedtronic turbine control panel).

A lot of solenoids can be energized and de-energized when the turbine is not running, including the trip solenoids connected to the protective terminal boards of the Speedtronic (20FG-1, 20FL-1, etc.). Sometimes, it's necessary to force them all on, one at a time, leaving each one forced as the next one is forced, to find a problem.

When there are multiple grounds on a panel, it's always a good idea when disconnecting wires or unplugging cable connectors to leave them disconnected or unplugged as you continue checking for grounds, until all the grounds are found. Then, they can be plugged back in or re-connected, one by one, continuing to monitor the DC for any other grounds.

Troubleshooting grounds is not an easy task, but it is not impossible, either. One has to understand how grounds are detected, and how power is distributed on a GE-design heavy duty gas turbine with a Speedtronic turbine control panel.

As for Kumar's problems, I would not be surprised to find (as Kevin Goh has alluded to) that some of the exhaust T/C cables have been damaged by heat (again, related to poor construction practices). It's very distressing to hear about problems of this magnitude, but even more distressing to hear about all the time and effort being expended to no avail.

Someone needs to sit down and come up with a comprehensive troubleshooting plan to be executed with the unit at rest and off cooldown, isolating all the various systems and circuits supplied by the 125 VDC Battery being used to power the Speedtronic panel, and be able to disconnect cables and wires and open isolation switches as necessary. (This will likely require that many motors be shut down in order to prevent some from starting and scaring people, even causing harm.) And then that plan needs to be executed, logically and methodically, with all the results noted so that any questions which might come up afterwards can be answered (what was done; what the results were; etc.). Discrete (contact) inputs and solenoid outputs are powered through the J12x cables of the PD core; solenoid outputs are powered by the J8x cables; cables J7A and J7W are used to power the trip solenoids through the TRPG and TREG, respectively (all of this is gleaned from the Power Distribution section of the Mark VI System Guide, GEH-6421, Vol. II). In addition, sometimes J17, J18, J19, and J20 are also used for some solenoid output circuits.

Finally, there are published limits to the voltage excursions that can be tolerated by the Speedtronic panel. If a short (such as could be caused by multiple grounds on both legs of the DC system) causes the battery voltage to drop below about 90 VDC, then there is circuitry to shut the Speedtronic power supplies off to protect the panel.
Thanks for this information

The alarm (battery ground) is Annunciated just after flame appearing and stay even if we get 95MW or more, and disappeared only and just after disappearing of flame.

we have:
- mark 6e
-heavy fuel turbine of 100MW
-electrical motor for cranking 88CR
-and electrical motor for gearing 88VG

Can we do some test when the turbine is running (for example forcing some contact at cimplicity and disconnect it from the mark 6e card)? because I think that's difficult to do it when not running because of alarm disappearing.

Thanks for replying me

Without being able to review the application code and I/O configuration of the Mark VIe at your site, it's very difficult for us to comment on what could be causing your problem.

When did this problem begin? After a maintenance outage? After a heavy rain storm? Has it been occurring since commissioning?

There's something unusual here that I can't define. Firing a Frame 9E/EA at 40% speed is a big flag which indicates something unusual for me. The fact that a Mark VIe is being used on this turbine probably indicates that the unit is under warranty, so have you contacted the packager to see if they can come and assist with finding the ground?

If you're absolutely certain that the alarm is coming at the instant that flame is established and clears when the flame goes out, I would suggest looking at the logic signal L28FD and determining which outputs are driven by L28FD. In other words, which outputs (discrete, contact outputs, and solenoid outputs) are actuated when L28FD changes logic state.

Sometimes compartment ventilation fans are started and stopped based on the presence of flame in the turbine. If this is the case, then there are usually pressure switches sensing either fan discharge pressure or a slight vacuum in the compartment. Many times grounds appear when switches change state.

You mentioned before that flame is established when the unit is at 40% speed. That's pretty high for most Frame 9E/EA units, but then there might be something we're not aware of with respect to your turbine and how it's operated. Many times, 40% speed drives one of the speed sensing relays, like possibly L14HA. Many devices and motors are switched on or off based on speed levels.

Many times ground appear also when solenoids are energized or de-energized. If you're very certain that the ground is occurring at about the same time flame is established and clears when flame goes out, then the fuel stop valve is one good circuit to examine. 20FL-1 is usually the liquid fuel stop valve solenoid, and it would be opened just prior to firing and establishment of flame, and it would be closed just prior to extinguishment of flame. This solenoid is usually connected to the TRPG and TREG boards.

I don't think there's a lot that can be done while the unit is running without increasing the possibility of tripping the turbine. You can force the stop valve solenoid when the unit is not running, and you can force the positive leg and the negative leg of the DC supply to the solenoid independently if the device is connected to both the TRPG and TREG. You don't need hydraulic pressure to just energize and and de-energize the solenoid to check for a ground.
Thank you verty much CSA for your last reply!

For the "when" question: this began after a days of rain and not since commissioning. I'm sorry because I was wrong when I told that the firing is at 40% (it was at 16% so I'm sorry)

For the (discrete, contact outputs, and solenoid outputs) which are actuated when L28FD is changes logic state we tested but no alarm appears after.

For the 20FL-1, we forced it and no ground.
You're most welcome, and you likely have narrowed your troubleshooting with your answer.

If the problem started after a period of rain, then it's most likely there is moisture in some switch housing or junction box nearest a switch housing that is OUTSIDE of any turbine- or accessory compartment. The likely culprits would be compartment vent fan switches (ones that sense fan discharge pressure or compartment "pressure" (negative or positive pressure)), or cooling water module vibration switches or cooling water pump discharge pressure switch(es).

Many times exterior conduit is incorrectly installed, meaning that conduit entries are made into the tops of junction boxes rather than the sides or bottoms. Water is more likely to enter a JB around the conduit penetration if it's at the top of an enclosure instead of on the side or bottom. Also, many unscrupulous individuals use conduit as ladders when climbing around compartments and either break the conduits or severly crack them around unions or JB entries. Sometimes unsuitable JBs are used for exterior applications, or they are left open or not fully closed and allow water to enter the enclosure/box.

But, it the problem started after a period of rain, then it's likely somewhere outside of a compartment, on some field device or instrument that's not been properly installed or not properly re-installed after a maintenance outage or exposed to water from poor conduit connections, or something like that.

Most battery grounds are in external devices; I would say the overwhelming majority of them are.

If you're looking for an "easy" way to troubleshoot grounds--forget it. There's not. Just get out and identify all the exterior JBs and devices (switches--temperature, pressure, limit, etc.) and go through them and eliminate them one by one, until you probably find the culprit.
Thank you again CSA for your replies, recently we found the problem in the clutch 20CF (even if we forced it before and no result). A wire wasn't isolated (not good).
So this thread is resolved
Thank <b>YOU</b> for the feedback!

"Feedback is the most important contribution!"(c) here at It's what makes forums like this work for years after the thread is resolved because others can read the thread and see that there was resolution and help from the information provided!