Hi,

We have a GE Frame 9E Gas Turbine on our plant. The control system is Mark VI. The generator is Air cooled. Our lube oil system contains the following main pumps

Shaft Driven Main Lube Oil Pump
Auxiliary (AC) Oil Pump
DC Lube Oil Pump [Emergency Lube Oil Pump 88QE]

The system has been working fine for some time but we have been facing the problem that when the DC Lube Oil Pump cuts in due to Lube Oil Header Pressure Low [L63QALX] it does not stop on Auto. There is no manual push button to put the pump on manual and stop it. The DC Lube Oil Pump keeps running and it has to be stopped by forcing the signal [L63QALPUL --> Lube Oil Header Pressure Not Low Pulse to Stop 88QE].

I might add that the Lube Oil Header pressure normalizes as soon as the AC Lube Oil Pump starts (AC supply is restored) and the Lube Oil Header Pressure Transmitter and Pressure Switch [L63QA2L] are also working fine. The two pressure switches installed on the Generator Side of the Lube Oil are also normal and have shown no abnormality.

We suspect that the pulse that is supposed to stop the DC Lube Oil pump is rendered ineffective by the signal [L4QELATCH] that goes high as soon as the DC Lube Oil Pump starts and resets after 30 minutes.

Can anybody please help me in understanding the sequence of DC Lube Oil Pump for GE Frame9E Mark VI system.

The DC Lube Oil Pump is working normally in nearly every other respect, it starts during the Startup sequence as it should for 30 seconds and then stops as it should, it starts when its control knob is moved to the Test Position and stops when the knob is returned to 'Auto' position, it starts for its weekly test and stops BUT the problem only comes when it cuts in due to loss of AC Power.
.
.
.
Any ideas, solutions, suggestions would be welcome.

 

Thanks for the information you provided in the post; it helps! We still need some more information for clarity, though. Please answer all the questions below when you respond.

Have you discussed this with GE? What have they said?

The reason for low L.O. Pressure is not exactly clear. You say you have a shaft-driven Main L.O. Pump, so if you lose AC (which seems quite unusual also!) why does the L.O. pressure drop?

So, you lose AC when you're running, producing power, when the Main L.O. Pump should be providing L.O. flow/pressure, and the unit continues to run, the AC (Auxiliary) & DC (Emergency) L.O. Pumps start, and the DC won't shut down? Please clarify this for us.

Does the unit trip and is it coasting down when the Emer. L.O. Pump starts? If so, at approximately what speed does the Emer. L.O. Pump start?

Because if the Main L.O. Pump is a positive displacement pump and if it is working properly it has a pretty strong output flow and pressure down to a very low turbine shaft RPM, sometimes the Emer. L.O. Pump won't start until the shaft speed drops to approximately 100 RPM (which freaks some people out during testing during commissioning!).

The sequencing you are describing I have never seen before, so I would have to look at it to understand it and try to determine it's intent. There have been a lot of changes to GE "standard" sequencing in the last few years; some good, some not so good. And that's just my personal opinion, and I apologize to anyone who doesn't want to hear my personal opinion.)

In the day (when I was in the field), the Emer. L.O. Pump was started when the pump discharge (Main or Aux) dropped below the level of a pressure switch. This was done to try to ensure the Bearing Header pressure (which was less than pump discharge pressure, controlled by the Bearing Header Pressure Regulator) stayed high for as long as possible. In other words, the Emer. L.O. Pump would start as pump discharge pressure was decreasing, but was still above bearing header pressure.

What pressure is 63QAL sensing: pump discharge pressure, or bearing header pressure?

There was a deadband in the switch that was used as a permissive in the Emer. L.O. Pump control circuit, and I believe it was something like 4-6 psi (or approximately 0.4) barg, but I may be wrong about the exact value of the deadband, and the logic you are describing doesn't seem to be like what I saw for years.

The purpose of the deadband being "high" was to prevent the Emer. L.O. Pump from cycling on and off, which is not desirable at all.

There was also some sequencing that would cycle the Emer. L.O. Pump for a couple of minutes every few minutes if the shaft reached zero speed and there was no Aux. Pump Discharge pressure (because of a likely loss of AC or Aux. Pump or Motor failure) just to keep L.O. flowing to the bearings on a periodic basis for cooling.

I'm wondering if the rocket scientists at GE have written logic to automatically start the Emer. L.O. Pump whenever there is a loss of AC, regardless of the pump discharge or bearing header pressure. (For many years the philosophy at GE was to keep critical and emergency functions like this simple and easy to understand. That was because a lot of the people who designed the circuits and schemes had field experience and knew that most times simple and uncomplicated was best. GE has hired a lot of engineers recently, a lot of them will little or no field experience, and they have written some unnecessarily complicated sequencing, and this may be one of those situations. And it may not; we still don't have all the information about the circumstances and devices and sensing locations.)

I'm also wondering if they have written sequencing to keep the pump running for some period of time just to ensure it doesn't cycle. Is there any L27MC1N and/or L27MC2N permissives in the ELOP sequencing? (Not that their presence would mean the sequencing was written to do one thing or another; just that someone was considering loss of AC for some reason when they wrote the sequencing.)

I think the sequencing is not completely understood, and it may be overly complicated. Without being able to examine the application code for the Mark VI at your site, it's very difficult to say for certain.

You are correct, though, in that GE purposely doesn't put an OFF push-button on the Emer. L.O. Pump. Many years ago they did, but they don't any longer. (Usually, changes like this are made when there is a catastrophic failure, so this makes one wonder what the root cause for the change actually was.)

In my experience on the units I have worked on (which include more Frame 9Es than I would have liked to have worked on) the Emer. L.O. Pump would shut off as soon as there was sufficient Aux. L.O. Pump discharge pressure and the pressure switch was actuated again. Sometimes that wouldn't happen until after a OFF/ON cycle when the turbine had reached zero speed without Aux. Pump discharge pressure had completed, but it would automatically shut down.

If the unit is at zero speed and you want to shut down the Emer. L.O. Pump after the AC is restored and the Aux. Pump is running, you can just open the Emer. L.O. Pump's MCC breaker for a couple of seconds and then re-close it. (The shaft is not spinning so it shouldn't hurt anything, unless there is some kind of turning gear that is rotating the shaft, which we also don't know from the information provided.)

 

CSA... perhaps GE may have purposely eliminated the DC-Driven L.O. pump's "Stop PB", as a result of the infamous Blackout of '65!

The blackout was exacerbated by loss of "Big Alice", the Louis Allis 1000 MW generator located in NYC, which at he time was the largest generator in the world. It was tripped off-line. Unfortunately, the during roll-down the DC L.O. pump was unavailable and the generator's bearings were "wiped!" Subsequent investigation couldn't determine if the "Stop PB" was inadvertently actuated or not!

As a consequence some systems implemented logic so that if the pump started automatically it couldn't be tripped manually during roll-down unless the Shaft-driven L.O. and AC-driven L.O. pumps were on-line.

Regards, Phil Corso

 

Thank you very much for the detailed reply, I really appreciate your help.

The scenario is as follows

The unit trips due to opening of 220 kV breakers (some sort of electrical fault in the grid). As a result the 11 kV and 380 V Bus Bars go dead (loss of AC). The Shaft RPM start to drop, at about 90% RPM the Auxiliary Lube Oil Pump (Emergency Bus Bar-AC) should cut in, but as the AC bus is de-energized so the AC Lube Oil pump does not start due to which the lube oil header pressure starts to drop. The Lube Oil Header Pressure Low switch is actuated (at 2.8 bar) and it sends a start command to DC Lube Oil Pump, which starts as it should. Please note that the DC Oil Pump starts when the Lube Oil Header Pressure drops, it does NOT start due to a loss of AC Power (in a way it does as the Lube Oil Pressure drops because the AC Lube Oil Pump is unavailable due to loss of AC Power).

The Emergency Diesel Generator cuts in and AC supply is restored within 1 to 2 minutes of Blackout. As the AC supply is restored the Auxiliary Lube Oil Pump starts and the Lube Oil Header Low Pressure switch Resets (3.1 bar) {at this point I believe a one second pulse L63QALPUL is generated that should turn off the DC Lube Oil Pump}, BUT the DC Lube Oil Pump keeps running. The shaft RPMs at this stage are quite high (more than 1000 rpm). The RPMs continue to decrease and the Generator Lift Lifting Pump (Jacking Oil Pump) and the Turning Gear cut in on Auto at their preset RPMs.

At this point the unit should be ready for startup but we do not get the “Ready to Start” status on the HMI. We believe that this is due to the fact that the DC Lube Oil Pump is still running along with the Auxiliary Lube Oil Pump. When we shut down the DC Lube Oil Pump, by forcing L63QALPUL, the GT HMI starts showing “Ready To Start”, and we are able to initiate the startup sequence.

The pressure switch L63QAL2 and the Lube Oil Header Pressure Transmitter are both mounted on the common discharge of the AC (88QA) and DC Lube Oil Pumps (88QE). The deadband of the pressure switch is 0.3 bar and the deadband for the pressure transmitter is 4 psi {configured in Mark VI}.

The ELOP sequencing does contain L27MC1N. If we back trace the L63QALPUL {the pulse that shuts down the ELOP} signal we find that the pulse will be generated if EITHER of the following conditions is met,

1. BOTH the Lube Oil Header Pressure Transmitter and Lube Oil Header Pressure Switch [L63QAL2] sense normal pressure.

2. L14HM is ‘0’ {shaft speed is below 285 rpm} AND L27MC1N [AC Bus Voltage Normal] becomes ‘1’.

If either of the condition is met, the signal L63QALPUL goes high for one second to shut down the ELOP. If both these conditions are met one after the other then two separate pulses will be generated.

At the moment we suspect that these pulses are generated within the 30 minutes window when the ELOP is latched ON due to the timer (which we strongly suspect has been put there to avoid the cycling of the ELOP) and that is why these pulse signals are unable to stop the ELOP. After the latching logic is over (30 min) the ELOP continues to run because the stop pulses have come and gone.

This seems rather illogical and I’m quite certain that there is something in the sequence that I’m unable to comprehend, because it doesn’t make sense for the DC Lube Oil Pump to keep running after the Lube Oil Header Pressure has normalized. There has to be an anomaly here but I’m unable to put my finger on it.

We have not discussed with GE yet, we will have to if this issue continues.

 

Correcting an error in my 28-Oct-10 (23:21) post about the '95 Northeast Blackout...

The 1,000 MW generator, affectionately called "Big Alice", was manufactured by Allis-Chalmers, not Louis Allis!

Phil

 

One of my main points was that the output of the shaft-driven Main L.O. Pump should be high enough when the shaft speed is above 1000 RPM to keep the bearing header pressure above the low-low trip setpoint. I've done tens of trip-from-CRANK speed with the Aux. L.O. Pump disabled to ensure the Emer. L.O. Pump will start when the pressure drops, and the Main L.O. Pump output is always high enough to keep the header pressure near normal until the shaft speed gets down to approximately 100 RPM. This is what FREAKS people out--they "think" the ELOP should come on immediately, but it's "saved" (saving the battery capacity" from starting until it's absolutely necessary), and also in doing so the intent again is to prevent cycling of the ELOP.

My other main point is that this tripping shouldn't be happening. Someone from the plant (management; owner) should be all over the utility to understand and prevent this from occurring. This is not normal to have to worry about your emergency systems because of nuisance trips. I know plant managers and owners who would be writing letters and getting lawyers involved, telling the utility that any catastrophic damage that could be traced to one of these events would be attributed to the utility and they would be responsible for any costs, including damages and lost generation. (And that's not just in North America any more. The litigious nature of business in general is being raised all around the world, to the benefit of almost no one except lawyers.)

You should be discussing this with GE. You may have discovered an unintended operational glitch with their new-fangled, (apparently) overly complicated sequencing. Large corporations are slow to react to complaints, except when you tell them this is the Emer. L.O. Pump, and it's affecting availability and reliability. Then, they're a little (notice I said "little") quicker to respond.

Forcing logic to toggle the output to shut down the Emer. L.O. Pump in this case seems to be the only thing to do until you get some advice or assistance or are directed to do something different by GE. You have discovered how to stop the ELOP, and you should make a temporary Operating Procedure that all operators (and technicians and plant supervisors) are aware of that details what to do and when in the event these unusual trips keep occurring or until you get some other direction or come to some understand or agreement with GE.

As our esteemed contributor, Phil Corso, said there is inherent "risk" in having a latching ELOP STOP push-button, and I don't think GE is going to change their policy on that. But you might get them to reconsider the sequencing.

But as has also been said before, if you own the machine and it's not being warranted by the packager, you are free to modify whatever you want however you want. You just have to take the responsibility for doing so should anything unanticipated occur.

I would also recommend you create a DDR (I think this can be done with later versions of Toolbox for the Mark VI, but I may be wrong about that) and/or edit one of the Trip History capture blocks to include as much data related to this event and condition as possible to try to gain a better understanding of the sequence of events. GE will be grateful, and so will you.

 

CSA... What did I say that led you to conclude I "said there is inherent "risk"...?

I never suggested a solution to Sarmad's dilema. I only proffered some history that could explain GE purposely omitting an OFF push button, especially in light of your "catastrophic failure" comment.

In fact, if asked for a solution, I would have requested a logic timing-chart!

BTW, except for the rise in New York City's birth-rate, the "Blackout" was indeed a catastrophic failure!

Regards, Phil

 

we are operating 4 GT's in the refinery and in the many others i have seen in other refineries the logic is the same as said here , the L4QEX signal latches once the LO pressure drops below 0.42 and it is not possible to switch off the EOP without forcing the signal. but the force need not be a permanent one , you just have to force it off , then take off the force signal , during this time the latch is reset and if the lube oil pressure is normal then the EOP will not start. we have been doing this as a standard procedure for the past 3 years lol. and yes BHEL reply (the vendor for GE gas turbine in India ) for this is that it is standard "GE design".

 

If Unit Bus Bar Supply not available or 6.6Kv not available, DC pump will be forced ON by latch relay in DC lube oil pump breaker.

 

I agree with CSA. The only way to solve this is to talk to GE. Goodluck with that, I suggest be prepared to make it an SOP like "processValue" says. The only cubicle in the PEECC without a stop button is the cubicle for the ELOP. It is better to keep the pump running, rather than to have it turned off by someone who doesn't know what he is doing. It happens... a lot.

Also, the DC ELOP is timed to turn off automatically UNLESS there is an MCC power failure (which does happen in your case, essentially you come out of a black out every time the 220kv supply trips).

But what I don't understand is that why don't you have emergency power available while the grid is out? Why doesn't the emergency diesel generator cut in on time? Your site is facing an "emergency blackout" every time. What happens if, God forbid, the DC ELOP fails? It seems that it is an operational regime to rely on the DC ELOP during every 220kv Line tripping on your site!

To my mind, that is the bigger issue. Don't get your bearing pads wiped like the folks in NY 95

 

Thank you all for sharing your experience and knowledge. We are investigating the problem in detail and hope to find an alternate solution (to logic forcing).
.
.
.
@CSA: Can you please elaborate on the possible reasons for the sudden drop in Lube Oil Pressure as the turbine speed starts decreasing. While the decrease in pressure is understandable (the shaft rpm are decreasing), should it not decrease gradually, proportionally with the rpm (or maybe not, maybe it was designed to be like this).

The trend of the Lube Oil Header Pressure Transmitter shows a sudden drop in pressure from 75psi to approx. 28 psi within a second(DC Lube Oil cut in set point is 42 psi) as the unit trips, resulting in the ELOP starting. Any comments?
.
.
.
@Process Value: Thanks a lot for sharing your experience. The "standard GE design" appears more like a bug to me but I'm only a novice in this field. Maybe the programmer who designed this sequence had something in mind that I'm unaware of right now.

 

Emergency Power is available. EDG cuts in within 1 minute of the tripping and AC Lube Oil Pump also starts within two minutes. The PROBLEM is that the DC Lube Oil Pump should stop when the Lube Oil Header pressure normalizes, but, it does not. Also as CSA pointed out, the Lube Oil Header pressure should not drop so much (within 2 minutes of GT Tripping) that the DC Lube Oil Pump has to start in the first place.

 

Sarmad

are you sure you are giving the right values here , the MOP/AOP pressure might be 75PSI but the lube oul pressure will not be that high , it is about 22-30PSI(max). It is my belief that you are looking at the MOP/AOP discharge pressure and not at the LO header pressure. similarly the auto cut in for DC lo pump is around 0.42 kg/cm2 or 6 PSI.

The DC LO pump is a low discharge pressure pump , you will not have your regular LO pressure while the DC LO pump is running. In our refinery , the DC Lo pump will generate around 0.84 kg/cm2 of pressure while running. check your PID for the actual discharge pressure of DC lo pump. and yes one more thing , if there is a huge pressure difference between the DC LO pump pressure and the LO header pressure , then there is a problem with the LO bearing header pressure regulating valve (VPR-2).

please check and reply so that we can help you better.

 

You need to look at the Piping Schematic (P&ID) for the L.O. System and look at the various components which might be not functioning properly. It's a complete mystery why the Main L.O. Pump discharge pressure is dropping so quickly. It's a positive displacement pump and the pressure is controlled with a relief valve. It would seem the pressure when there is load and the unit is running is fine, and that kind of rules out the relief valve.

Trip oil pressure comes from the combined L.O. Pump discharge header (with the exception of the Emer. L.O. Pump, if I recall correctly). There is supposed to be a check valve installed in the Trip Oil header, and there is supposed to be an orifice drilled into the check valve. The Check valve is usually a swing check valve and it's closed during normal operation, with oil flowing through the orifice to establish Trip Oil pressure (it's a static system). (That's why it usually takes a short period of time for Trip Oil pressure to rise when starting or transferring fuels because oil has to flow through a small orifice.) When the unit is tripped, pressure upstream of the swing check is suddenly relieved and the swing check valve opens to relieve Trip Oil pressure very quickly.

For the pump discharge header pressure to drop so quickly on a trip, there must be oil "excess" oil flowing when there is a trip. Somewhere. I'm not saying it has to be through the Trip Oil system check valve, but that's one possibility. Without being able to see the Piping Schematics for your turbine I don't know what kind of generator you have and where Bearing Lift Oil pressure/flow comes from in the system. It could be that if the generator has Lift Oil that a solenoid is opening somewhere, or a check valve isn't working correctly, or something like this.

Without being able to see the unit and the drawings for the unit, it's very difficult. But pressure usually drops when the flow gets too high.

I also spoke with a former colleague and he told be about a TIL (Technical Information Letter), 1420, that talks about a 30 minute "latch" for the Emer. L.O. Pump on loss of AC just to ensure cooling flow to the bearings. So, that might be part of the sequencing.

Please let us know what you find.