startup fuel flow excessive

A

Thread Starter

AHMED RASHAD

gas turbine mark 5, before fire, it tripped by startup fuel flow excessive

At 20%

1- fuel oil bypass valve not moving (0%) and this mean 100% flow.

2- demand still -25% (0%) because 20FL SIGNAL STILL NOT COMING.

3- SERVO CURRENT FEEDBACK STILL HIGH (70)

NOTE
I MEASURE THE VOLTAGE -8 FOR ALL THREE COILS BEFORE START
can you please till me where can i find the problem cause
 
H
Change the stop/speed ratio valve-SRV moog servo valve and try the unit again. Verify there is trip oil establish as the turbine goes to cranking speed.
 
When did this problem start? After a recent maintenance outage? If this occurred after a maintenance outage, can you describe what work was done to the liquid fuel system during the outage?

Did the problem start after the liquid fuel bypass valve's servo-valve was replaced?

How often does this particular machine operate on liquid fuel?

By looking at the sequencing in the Mark V, you can determine that the Liquid Fuel 'Start-up Fuel Flow Excessive' alarm is generated when the flow divider feedback exceeds some level.

Unfortunately for most Frame 5s it's not possible to see the valve stem to know what position the valve is in, and there are not LVDTs to be able to know precisely what position the valve is in. It is a bypass valve, so it should be open during shutdown (which is zero flow to the turbine) and closes to increase or maintain fuel flow to the turbine.

Positive servo current to the servo-valve is supposed to decrease or shut off the flow of fuel through the bypass valve. It's very interesting that you measured -8 VDC at each of the coils when the reference was -25% (which is the "shutdown" reference, meaning the Mark V is telling the bypass valve not to allow any flow, negative flow, in fact!) to the turbine. If you were measuring the voltage with the proper polarity of the voltmeter leads you should have read approximately +8 VDC with a reference of -25%.

If the liquid fuel stop valve (VS1-1, controlled by 20FL-1) was not energized and the Mark V tripped the turbine on excessive fuel flow, then it would seem that the flow divider feedback to the Mark V somehow "spiked" and caused the indicated liquid fuel flow-rate to exceed the trip value.
 
That's it!

The solution to every problem is to replace the Moog servo-valve, and of course re-calibrating the SRV.

Even if it's in a completely different fuel system, it always works.

 
A

ahmed rashad

> When did this problem start? After a recent maintenance outage?

1-after long time of outage and also i made change for filter inside
* THE UNIT USE FUEL OIL ONLY

2-i changed again with repaired one(1gpm)
Result : bypass again not move (0%)

3-i do not have new so i use servo for igv(5gpm)
Result : bypass valve not move (0%)

my FIRST question please
you said while shutdown the bypass should be open
but the position now (0%)IS THIS MEAN OPEN?

SECOND:
while i was CRANC i made force 20FL to open
then the unit tripped with fuel excessive again
IS THIS NORMAL?

THIRD: MY STUFF TOLD ME IN NORMAL
THE BYPASS GO FROM 0% TO FULL
THEN IT MAKE CONTROL
IS THIS RIGHT?

AND I DO NOT UNDERSTAND THIS :
> If the liquid fuel stop valve (VS1-1, controlled by 20FL-1) was not energized and the Mark V tripped the turbine on
> excessive fuel flow, then it would seem that the flow divider feedback to the Mark V somehow "spiked" and caused the
> indicated liquid fuel flow-rate to exceed the trip value.
 
You have told us this is a Frame 5, and *most* Frame 5s don't have LVDTs on the Liquid Fuel Bypass Valve (VC3-1). So, how are you certain the valve is at any particular position, particularly since the valve stem of most GE-design Frame 5 heavy duty gas turbines is not visible to an observer. There is no pointer/scale attached the valve stem; the stem can't be seen; and without LVDTs or some other means of monitoring the position of the valve stem/plug it's virtually impossible to say for certain what position the valve stem/plug is in.

If the LFBV (Liquid Fuel Bypass Valve) has LVDTs the signal name for position is usually FSL. On many Speedtronic turbine control systems, FSL is a "valid" signal name even if there are no LVDTs (the whole "max case" thing makes things needlessly complicated and confusing). And if there are no LVDTs then this signal will not change, from a probably constant state of zero.

The liquid fuel system of most GE-design heavy duty gas turbines calculates a liquid fuel flow-rate reference and then adjusts the servo current output to make scaled feedback from speed pick-ups on the liquid fuel flow divider equal to the flow-rate reference.

What are the signal names your STUFF [sic] is monitoring (since you don't seem to be monitoring anything yourself)?

In general, the fuel flow-rate reference signal name is FQR and FQROUT. In general the scaled liquid fuel flow-rate feedback signal from the liquid fuel flow divider speed pick-ups is FQL or FQL1 or FQLM or FQLM1 or something similar to these.

FQR (the reference) and FQL (the feedback) get summed to produce the servo-valve output current. If the feedback is equal to the reference, the output will be effectively zero mA (to which the null bias value is added). If the feedback is greater than the reference then the servo-valve output current will be more "positive" to try to reduce the actual liquid fuel flow-rate. If the feedback is less than the reference the the servo-valve output current will be more "negative" to increase the actual liquid fuel flow-rate.

There have been reported cases where there is electrical "noise" on the one or more of the liquid fuel flow divider speed pick-up wires. This noise can be interpreted as liquid fuel flow, and if the noise level is great enough then the servo-valve output current will never change to increase the liquid fuel flow-rate.

There have been some cases where the mechanical wear of the liquid fuel flow divider internals is so great that the mechanical vibrations from the turbine shaft and accessory gear and Aux. L.O. Pump (and Aux. Hyd. Pump, if present) can cause the toothed wheels of the flow divider to vibrate. Newer Speedtronic panel speed pick-up input circuits are extremely sensitive in order to be able to better sense zero speed, but this has caused problems with noise and worn flow divider internals.

Another problem which has been reported here on control.com is the gap between the speed pick-up and the toothed wheel. It seems newer Speedtronic panels don't require a very small gap, and that a small gap can cause vibrating toothed wheels to send a signal which the Speedtronic interprets as fuel flow even when there is not fuel flowing.

So, this may be what's happening at your site: the speed pick-ups on the liquid fuel flow divider are causing the Speedtronic panel to believe that liquid fuel is flowing when in fact there is no liquid fuel flowing, and possibly even that the false flow-rate exceeds the flow-rate reference. There might be noise on the flow divider speed pick-up wiring, or the divider toothed wheels may be vibrating because of mechanical wear or even damage, or the speed pick-up gap may need to be increased.

You need to determine what the Speedtronic believes the liquid fuel flow-rate is when the unit is cranking/purging (when the Liq. Fuel Stop Valve is closed--when 20FL-1 is de-energized). If there is indicated flow when the stop valve is zero, and if the indicated flow spikes (momentarily jumps) high enough it could cause the unit to be tripped on excessive fuel flow when in fact there is no fuel flowing.

If, during firing (when the Liq. Fuel Stop Valve is opened) the servo-valve output current never changes and you can see the Speedtronic indicates liquid fuel flow and the indicated liquid fuel flow-rate is greater than the reference, then there is some kind of problem with the flow-rate feedback/inputs. (To establish liquid fuel flow the servo-current needs to go at least slightly negative, under normal conditions, in order to get the bypass valve to close to force fuel to the turbine.)

If, during firing, you use the selector valve at the liquid fuel flow divider you can determine if there is actually liquid fuel flowing or not. The pressure of each liquid fuel nozzle (as read by rotating the selector valve handle through positions 1 through 10) should be greater than the liquid fuel check valve cracking pressure, which is usually a value approximately 100 psig, sometimes higher, but not usually lower. In order for fuel to flow the pressure must be higher than the cracking pressure of the liquid fuel check valves, and if it's not higher than the check valve cracking pressure then it's pretty likely there's no liquid fuel flowing.

Use the Liq. Fuel System Piping Schematic (P&ID) to located all of the components of the liq fuel system on the Accessory Base and in the Turbine Compartment.

You say you have a Mark V; use the VIEW2 utility to capture data for the liq fuel flow-rate reference and liq fuel flow-rate feedback and the servo-valve output current and analyze the data.

Was the liquid fuel flow divider worked on during the last mechanical outage? If so, what work was done?

When was the last time the liq fuel flow divider was serviced/refurbished?

What are the gap distances of the speed pick-ups on the liquid fuel flow divider?

Please write back to provide the answers to these questions, and to let us know how the troubleshooting progresses.

If you (or your STUFF [sic]) is not familiar with using the VIEW2 utility, refer to the Mark V Maintenance Manual, then I suggest you bring someone to site who can be of help in sorting out the problems.
 
A

ahmed rashad

hello sir

thank you very much for your effort

> You have told us this is a Frame 5
****************i said mark 5
note : our unit frame 6

> If the LFBV (Liquid Fuel Bypass Valve) has LVDTs
****************we have only ruler indicator

> What are the signal names your STUFF [sic] is monitoring (since you don't seem to be monitoring anything yourself)?
****************very good remark

> So, this may be what's happening at your site: the speed pick-ups on the liquid fuel flow divider are causing the
> Speedtronic panel to believe that liquid fuel is flowing when in fact there is no liquid fuel flowing,
******************verygood
in sensor manual gap 0.007 + or - 0.002
i remember i decrease the the gap from
15 actual to 10 in but the unit run after that
what is your opinion

> may be vibrating because of mechanical wear or even damage
******************this is new flow driver

> You need to determine what the Speedtronic believes the liquid fuel flow-rate is when the unit is
> cranking/purging (when the Liq. Fuel Stop Valve is closed--when 20FL-1 is de-energized)
******************as i told you i force 20FL
while cranking and the unit tripped with the same alarm and i asked you is this normal ?
please i need reply for this question

> You say you have a Mark V; use the VIEW2 utility to capture data for the liq fuel flow-rate reference
**************-25%

> fuel flow-rate feedback
till 70%

> servo-valve output current
in normal current feedback go from 70 to 0
in our case go from 70 to 78

the history for this unit always load fluctuation then we change servo valve.
and may be we found the fluctuation still found
 
So, you have a method for seeing if the valve stem and plug is actually moving. Does it move during firing (when 20FL-1 is energized and VS1-1 is open)?

As for your question about whether or not the unit should be tripped on excessive fuel flow during cranking/purging when 20FL-1 is not energized (or is forced to logic "0"), well, sometimes GE sequencing isn't so logical. Without being able to see the sequencing in the Mark V at your site it's impossible to say for certain, but I have seen instances where the start-up excessive liquid fuel flow-rate detection was enabled prior to firing and if the feedback from the flow divider speed pick-ups exceeded the trip setpoint the unit would indeed be tripped--even if 20FL-1 was not energized.

So, while it doesn't seem "logical" the unit should be tripped on excessive fuel flow when there shouldn't even be fuel flowing (when 20FL-1 is de-energized and VS1-1 is closed), at some time the engineers surmised that if there was excessive flow-rate feedback from the flow divider when there shouldn't be any flow-rate at all that the unit should be tripped and the cause investigated and resolved.

It's not quite clear to me exactly the sequence of events with regards to the changing of the speed pick-up gaps. You say the flow divider is "new"; I've seen people call refurbished equipment "new" when in fact it's not, and it was not refurbished by the OEM (Original Equipment Manufacturer). Lots of companies say they can refurbish equipment to like-new specifications, but can't. We've had several reports of this on control.com, most recently with regard to electro-hydraulic servo-valves used on GE-design heavy duty gas turbines.

Was the Mark V provided with the original Frame 6 installation, or was it a retrofit replacement for an older Speedtronic control system? I've never seen a gap setting of less than 0.010 inches for a liquid fuel flow divider, but I haven't seen every Control Specification in existence. And, if you're not referring to the GE Control Specification for the gap setting, then that could also be a problem.

You have been given a course of action: Use one of the VIEW utility applications to monitor the turbine speed and liq fuel flow-rate reference and servo output current and liq fuel flow divider feedback, and the status of 20FL-1 and 20CF-1. If you see liquid fuel flow feedback during the period PRIOR to the opening of VS1-1 by the energization of 20FL-1, then there is some problem with the indication coming back from the liq fuel flow divider speed pick-ups. Period. Full stop. End of discussion.

A liq fuel flow-rate reference of -25% is typical for any time when there should NOT be liquid fuel flowing. Usually there is a few second delay between the time firing begins (20FL-1 is energized) and the time the liq fuel flow-rate reference is "released" from -25%, but within 15-30 seconds after 20FL-1 is energized it should start ramping positive from -25% to the required fuel flow-rate reference (the ramping is to prevent a step-change and excessive liquid fuel flow).

If the feedback is not high enough to trip the unit BUT it is higher than the liquid fuel flow-rate reference, then the LFBV will never close when 20FL-1 is energized to open VS1-1--because the Mark V believes there is more fuel already flowing than required. And, if the pressure at the flow divider outputs to the ten combustion cans/fuel nozzles is not greater than the cracking (opening) pressure of the liq fuel check valves then there will no liq fuel flowing in the system.

You have all the necessary information--per the information provided--to troubleshoot the problem and let us know what you find.

The problem could be that the twisted, shielded pair cables being used to transmit the signals from the liq fuel flow divider speed pick-ups to the Mark V are not properly shielded (the drain wires are not properly grounded).

The problem could be mechanical vibration from other equipment on the Acc. Base causing the flow divider to vibrate and because the gap distance is too small the speed pick-ups are sending erroneous signals to the Mark V.

It could be there is air in the lines or in the fuel filter(s) that's causing or contributing to the problem.

You have introduced a new condition (the unstable operation after servo change-out). Let's get one condition troubleshot and resolved at a time. (I would ask: Why is it necessary to replace servos so "often"? I've been on 20 year-old machines with the ORIGINAL servos on the LFBV!) When you replace the servo-valves, do you verify servo current polarity before attempting to start or run the machine?

Lastly, when the unit is running at 20% speed, the High-pressure Liquid Fuel Pump is also running at 20% of rated speed. So, to get the same amount of fuel to the fuel nozzles as is being called for at, say 100% rated speed, the LFBV must be closed more at 20% speed than at 100% speed. (The HP Liq Fuel Pump is a positive displacement pump.) It's not uncommon for the LFBV to nearly close during firing because the fuel pump is not running at rated speed so it's not producing rated pressure or flow.
 
A

ahmed rashad

Hi sir,

good news may help

1-I INCREASE THE GAP TWO TIMES (LESS THAN HALF TURN) BUT SAME PROBLEM

2-I increased the time to trip from 2sec to 4sec

3-the unit succeed to startup and load BUT

*********THE FUEL FLOW READING IN THE BEGINNING ((ONLY)) SPIKE
THEN GOING TO BE NORMAL AND NOT SPIKE AGAIN

*********THE LOAD IN THE BEGINNING FLUCTUATING
THEN NORMAL

######
MY FIRST QUESTION I ASKED YOU MANY TIMES BUT TILL NOW NO CLEAR REPLY BECAUSE YOU DO NOT UNDERSTAND ME:

THE UNIT RUNNING CRACK

then

I FORCED 20FL TO OPEN (ENERGISED)
AND OF COURSE THIS MEAN 20CF ENERGISED

THE LFBV NOT MOVED (0 POSITION)FROM RULER
THIS NORMAL
BECAUSE REF -25%

THEN
TRIPPED BY EXCESSIVE FUEL
AND THE FB READING HIGH

IS THIS NORMAL ???

DOES 0 POSITION IN THE RULER MEAN FLOW TO TURBINE OR TO PUMP

IF FLOW TO TURBINE
AND THIS ACCEPTABLE BECAUSE 0 POSITION
OF LFBV MEAN CLOSED (FLOW TO TURBINE)

HOW CAN THE DESIGNER MAKE
WHILE SHUTDOWN AND IN STARTUP THE FLOW TO TURBINE?

IF FLOW TO PUMP
HOW CAN FLOW EXCESSIVE WHILE THE FUEL NOT TO TURBINE?

I SUSPECT NOW AS YOUR SUGGESTION
NOISE RIDE THE ACUTAL SIGNAL
BECAUSE
THIS UNIT ALWAYS SUFFER FROM FLUCTUATION
OF THE FEEDBACK SIGNAL

ANOTHER REMARK AND I WANT YOUR OPINION
BEFORE,
WE FOUND THE CONNECTION FOR FLOW DRIVER MOTOR NOT TO ROTATE THE FLOW DRIVER
I MEAN
IF YOU ROTATE THE MOTOR BY HAND IN THE DIRECTION SAME LIKE ELECTRICAL CONNECTION
THE FLOW DRIVER WILL NOT ROTATE
AND
IF YOU ROTATE THE MOTOR BY HAND IN THE DIRECTION OPPOSITE TO ELECTRICAL CONNECTION THE FLOW DRIVER WILL ROTATE
AND THIS MEAN NO NEED FOR THIS MOTOR

DO YOU THINK IF WE MAKE THE ELECTRICAL CONNECTION TO ROTATE THE FLOW DRIVER MAY CAPTURE THE ROTATION IN THE BEGINING???

AND IN THIS CASE HOW CAN WE SOLVE THE FLUCTUATION
OF FEEDBACK SIGNAL DURING RUNNING????

AGAIN NOISE
THIS IS VERY DIFFICULT TO CAPTURE
GOODBYE SIR
 
Dear Ahmed,

Stay away from the SRV servo...:) while working on LFBV....

What does the GT use for liquid fuel? Diesel, Nafta or Jet B maybe? Does your machine have also fuel additive skid (required for Nafta and JET B fuel)?

1. The spike you are seeing is probably not a real spike but, its part of the sequence that close the LFBV just before firing command in order to check that the speed probes are functioning and opening the check valves! The FQL signal must be seen by the turbine control panel prior firing and your "stuff: is most probably right about the sequencing logic. If the "spike" last longer than 1.5 second, its not a spike.
0% means bypass and 100% means to the flow divider.

2. Can you tell us the liquid fuel pressure? Can you monitor if this pressure is dropping during cranking speed, just before firing?

To be continued.....

Good Luck...

A. Oztas
 
1) We would need to see the data (VIEW2 data; RealTime Plot trend; etc.).

2) This is good; let the execssive fuel flow condition exist for longer allowing more unburnt fuel to enter the turbine so that if flame is established it will be established in many places all at once.

3) We would need to see the data (VIEW2 data; RealTime Plot trend; etc.).

I have responded to this question in my last response. If you are CRANKing the unit (meaning the unit is in CRANK mode, not FIRE or AUTO mode) and you energize 20FL-1 by forcing the logic signal for the output, then the fuel reference is never going to change from -25%. CRANK mode does not initiate the purge timer/permissive for allowing fuel to flow.

If the unit is in FIRE or AUTO mode and you are forcing 20FL-1 open prior to the expiration of the purge time, then liq fuel reference is not going to increase above -25%.

I have never seen a pointer/scale/"ruler" on a Frame 6 LFBV, so I can't say. <b>THE LFBV CLOSES TO INCREASE FUEL FLOW-RATE TO THE TURBINE. IT IS CALLED A "BYPASS" VALVE FOR A REASON. IF YOU WOULD LOOK AT THE LIQUID FUEL PIPING SCHEMATIC (P&ID) YOU WOULD CLEARLY SEE THIS.</b>

The bypass valve should be 100% OPEN prior to the energization of 20FL-1, and then ramp closed to cause fuel to flow to the flow divider and then to the fuel nozzles a few seconds after 20FL-1 is energized <b>and the purge timer permissive is satisfied</b>.

Does the scale/ruler indicate OPEN or CLOSED? Or just 0 and 100? When the unit is not running or not firing, the LFBV should be in the FULL OPEN condition, or 100% OPEN, which is equal to 0% flow to the flow divider and fuel nozzles.

<b>Per the Liq. Fuel Piping Schematic (P&ID), the LFBV is in parallel with the High Pressure Liquid Fuel Pump--or, in other words, it connects the output- and input ports of the pump. When the LFBV is OPEN, ALL of the liquid fuel flowing through the pump is recirculated back to the suction of the pump--meaning, there is NO flow to the flow divider and the fuel nozzles.</b>

As the LFBV closes when the fuel pump is running, less fuel is recirculated back to the suction port of the pump, which means (because the pump is a positive displacement pump) that the portion of fuel not being recirculated has to go to the flow divider and fuel nozzles.

There are check valves at each of the fuel nozzles. It usually requires approximately 100 psig (refer to the Device Summary for your unit for the exact cracking pressure, or usually the value is stamped on the body of the check valve) fuel pressure in order for the check valve to open and fuel to flow into the nozzle/combustor.

So, the LFBV must close sufficiently to increase the output pressure of the HP liquid fuel pump to cause the pressure to exceed liquid fuel check valve cracking pressure in order for fuel to flow to the nozzles and into the combustors where it can be ignited.

I have never taken a flow divider with a "helper" motor apart, but I'm told, and I believe, that there is a clutch between the motor and the flow divider shaft, so that the motor can be used to help the flow divider shaft start turning but that if the required fuel flow-rate means the flow divider shaft should spin faster than the motor is turning the flow divider shaft it is not restricted by the motor. And, even when the "helper" motor is used, it only seems to run for a short period of time, meaning a couple minutes or less.

If the "helper" motor is not turning the shaft in the proper direction when it's running, that that could be a very large problem. As I said, they are usually DC motors, so the polarity of the supply voltage is critical. You said something to the effect that you found the "helper" motor not connected. What exactly did you mean? Not electrically connected? Is there an output from the Speedtronic control system to run the "helper" motor?

If you are burning a "light" distillate that is not "cold" (the fuel temp is not less than approximately 15 deg C), then the "helper" motor may not even be required. But that's just an educated guess. Most every unit I've worked on that had liq fuel flow divider "helper" motors did NOT use them if they only burned HSD or #2 distillate. They only seem to be required if the fuel has a high viscosity (such as HFO, Heavy Fuel Oil), or the ambient can be cold enough to cause the distillate (light) fuel oil to have a high viscosity unless it's heated before being supplied to the turbine.

As for the noise issue, if you are able to establish flame and have stable operation, then it's not clear there is noise on the flow divider pick-up wiring--or the speed of flow divider is such that it's canceling out the noise when fuel is flowing. You need to look at the data you have from the VIEW utility, or use an oscilloscope or some other means of monitoring and recording the flow feedback during purging and firing.

Capturing data with the VIEW utility applications on the Mark V operator interface is the best thing you can do for yourself. I think you are grasping at straws right now (looking for ANY possible problem, and all possible problems). You seem to have been changing servos without knowing if the servos needed to be changed. I wonder if you have verified the polarity of the servo currents being applied to the servo each time you changed a servo (polarity is important!). Servos get improperly blamed for all manner of ills.

To Ali Oztas' comments, if the liq fuel supply pressure up to the Liq Fuel Stop Valve (VS1-1) is not stable during firing or during operation, then the speed/load is not going to be stable. This is where air in the lines can be particularly problematic. And, there is usually a Liquid Fuel Forwarding Piping Schematic (P&ID) to be able to know all of the components in the liquid fuel "supply" (forwarding) system--upstream of the liq fuel stop valve.

The liquid fuel system--all of it--is a very tightly interconnected system. Most of the operation parameters are not controlled by the Speedtronic, or even monitored by the Speedtronic. They have to be manually adjusted and monitored and checked when there are problems. But, without referring to the Piping Schematics (P&IDs) it's very difficult to understand how things all have to work together.

In the Manuals provided with most GE-design heavy duty gas turbines (regardless of packager) there should be some System Descriptions. Use them, in conjunction with the Piping Schematics (P&IDs) to learn and understand each of the systems.
 
Top