GT Tripped in Low Liquid Fuel Pressure Trip

B

Thread Starter

barindra75

Frame-V gas turbine is dual fuel type (NG & HSD). It was running in NG 100%. We tried to change over the fuel from NG to HSD. After selecting HSD, liq fuel stop valve opened and HSD on skid pump clutch engaged. After 15 sec turbine tripped in low liq fuel pressure (63FL-2).

Before taking the HSD in line, leak attending job carried out and line was drained from HSD stop valve to burner nozzle portion. Before HSD stop valve the HSD pressure was stable at 4.5 kg/cm2 and there was no dip in the pressure. This pressure is maintained by HSD forwarding pump with a pneumatic control valve.

As per logic, the on skid HSD screw pump suction side pressure switch (63FL-2) set point is 0.73 kg/cm2 with time delay 15 sec.

HSD forwarding pump flow capacity is 30 m3/hr with discharge pressure 9 kg/cm2.

My question is,
1. I am confusing about the reason for which this pressure sustained less than 0.73 kg/cm2 for a prolonged time for 15 sec?
No leak / pressure drop found later on the same line. The line volume may be hardly 1 m3. Low pressure sensed after stop valve opening.

2. When the turbine tripped, Gt was running at 100% gas, liquid fuel was not in line till then. Why the turbine tripped on that condition? Only de-clutching of the HSD screw pump was sufficient to save the pump, the system could have continue to run on GAS?

Please enlighten.
 
barindra75,

What was the exact Process Alarm message the Speedtronic turbine control system annunciated when the turbine "tripped"?

What happened to the load during the transfer attempt and right before the unit "tripped"?

Without being able to see the actual liquid fuel piping at your site, as well the P&IDs for your turbine and auxiliaries, and the programming in the Speedtronic turbine control panel, it's impossible to say too much about the hardware and what might have happened for sure. However, read on.

It's not usual for a Speedtronic turbine control system to be programmed to trip the turbine on low fuel supply pressure when it's running at rated speed. And, whether or not the Speedtronic tries to revert to the running fuel when there's a low fuel supply pressure on the fuel being transferred to is not standard on all or even most machines.

So it's really hard to say exactly what the software did or was supposed to do.

Usually, during a transfer there is a 30-second period before the flow-rate of the fuel being burned on starts being ramped down (decreased) and the flow-rate of the fuel being transferred to starts ramping up (increasing). And, the fuel transfer (when the fuels are actually changing) is usually programmed for 30 seconds also. And if I recall correctly an automatic transfer on low fuel supply pressure can't occur unless the unit is fully on one fuel or the other--so not while a transfer is in progress. The way the Speedtronic programming works is that, once begun, a transfer has to be completed and then if there's a low pressure on the running fuel it would initiate another transfer which takes about a minute to complete. If the Speedtronic was programmed to allow for Mixed Fuel Operation AND a knowledgeable and well-trained operator were at the controls and watching what was happening, the operator might be able to stop the transfer and then start a transfer back to the first fuel--but that would take a knowledgeable and well-trained operator and an appropriately configured Speedtronic. And a good bit of luck, too. And I've never seen a Speedtronic programmed to do that automatically; never, ever.

It's possible that the Speedtronic was programmed to trip the turbine on low liquid fuel pressure--not likely, but possible. The 15-second delay actually seems a little long to me, but it's there so that "momentary" pressure/flow problems will not result in nuisance or erroneous alarms. Yes; the High-Pressure Liquid Fuel Pump needs liquid fuel for lubrication, but that's not generally the reason for the low liquid fuel pressure switch at the pump suction (which is where I presume and suspect where the sensing line for 63FL-2 is located).

BUT, the most important thing to remember about a properly programmed Speedtronic turbine control panel is: <b>EVERY</b> condition that results in a turbine <b>trip</b> is SUPPOSED to have a Process Alarm and text message to indicate why the turbine tripped. <b>EVERY</b> trip condition. It may not be obvious, but it's almost always there (annunciated). There is the odd--and unfortunate--Speedtronic that was not programmed to adhere to this philosophy, but it is not common. And, it can be corrected in the field (depending on the vintage of Speedtronic).

So, that's why the Alarm Printer is SO important. And why every operator should write the Process Alarm text message(s) on <b>EVERY</b> turbine trip in the Operations Log Book--so that detailed troubleshooting can be done later. Some sites even require that the Process Alarm printout be stapled or taped into the Operations Log Book (which is a GREAT idea).

(I'm waiting to hear why the Alarm Printer isn't or wasn't working.... There's always a justification for this. When there really isn't. Think about it. When a turbine trips, a LOT of things are usually happening at the same time--especially in a combined-cycle power plant, but even in a simple-cycle power plant. That paper Alarm Printout is <b>SO</b> necessary for understanding what happened and why--especially in a case like this when there's a seemingly contradicting information. It's just not logical to operate a plant without a working Alarm Printer, and without referring to those printouts after the Control Room and the plant returns to a calm condition and everyone wants to understand what just happened and how it might be prevented from happening again.)

Now, for the hardware. You said that a fuel leak downstream (after) the Liquid Fuel Stop Valve had just been repaired, and that the leak was near the fuel nozzle ("burner"). Well, that means there was likely a lot of air in the piping--and tubing--downstream of the VS1-1 (the Liquid Fuel Stop Valve), probably more piping and tubing than whatever the 1 m3 volume you're talking about (which is 1000 liters, or approximately 22 gallons, by the way). After VS1-1, there's usually a small High-Pressure Liquid Fuel Filter, then the Liquid Fuel Flow Divider--with a manual selector valve and gauge assembly--then ten (10) individual tubes which run under the inlet and up and into each of the fuel nozzles and combustors. But, it would certainly seem that <b>if</b> there was pressure--and flow--that even if all of the volume downstream of VS1-1 would have filled pretty quickly, probably in just a few seconds after VS1-1 was opened and the High-Pressure Liquid Fuel Pump started turning because the pressure downstream of the pump will quickly build up to the liquid fuel supply pressure.

It's not clear where the gauge/transmitter was sensing liquid fuel supply pressure. Usually, there are gauges at the inlet to the Accessory Compartment, at least one gauge on the Accessory Gauge Panel, but Position 12 (if I recall correctly) is the pressure at the inlet to the High-Pressure Liquid Fuel Pump. That's the only reading that I would trust to say that there was truly pressure "in the liquid fuel piping" <i>at the suction of the High-Pressure Liquid Fuel Pump</i>--which is where there MUST BE pressure--and flow, of course.

Getting back to the repair, unless a formal, written procedure is used to isolate the liquid fuel system and then to de-isolate the system, it's very common for overzealous people to close valves and then for them to not be opened. But, it is always the responsibility of the lead outage person to personally walk down all of the system to be sure that all valves are in the proper position before returning the turbine and auxiliaries to the Operations Department. And, I would say the Operations Department should also accompany the outage lead on that system walk-down, and/or to walk the system down personally before starting the turbine.

So, without more information--a lot more information, particularly the Process Alarm text message(s) from the "trip"; and without being able to see the P&IDs, at least, and the programming in the Speedtronic, it's really not possible to say too much more than: It certainly seems like even if there was truly pressure <i>at the suction</i> of the High-Pressure Liquid Fuel Pump there was insufficient flow to allow the liquid fuel to maintain flame.

That's about all that can be said based on the information provided. There's simply not enough information to be of help. If you want to understand--or if you want help in understanding--why a turbine tripped, it's necessary to know what Process Alarms were annunciated at the time of the trip.

(I was hoping someone else would respond and ask some questions to get more information to provide a better response. But, it's not my year. Yet.)
 
barindra75,

>After VS1-1, there's
>usually a small High-Pressure Liquid Fuel Filter, then the
>Liquid Fuel Flow Divider--with a manual selector valve and
>gauge assembly--then ten (10) individual tubes ...

and

>Position 12 is the pressure at the inlet to the High-Pressure
>Liquid Fuel Pump. That's the only reading that I would trust
>to say that there was <b>truly</b> pressure "in the liquid fuel
>piping" <i>at the suction of the High-Pressure Liquid Fuel
>Pump</i>--which is where there MUST BE pressure--and flow,
>of course.

"Position 12" refers to the position of the manual selector valve on the Liquid Fuel Flow Divider gauge assembly.

Also, 0.73 kg/cm2 is only about 11 psig--which is VERY low liquid fuel pressure. I believe the Liquid Fuel Forwarding Skid pressure regulator that you referred to is supposed to be set at something like approximately 4 barg (about 60-65 psig).

The Liquid Fuel Check Valves at the fuel nozzles (burners) usually have a cracking pressure of about 100 psig--which is above the liquid fuel supply pressure of approx. 60-65 psig. 10 psig would be very low pressure. Have you verified the operation and isolation valve position for 63FL-2?

Usually if there's a 63FL-2, there's a 63FL-1, which is upstream of the Liquid Fuel Stop Valve, and should also alarm if the liquid fuel supply pressure from the forwarding skid discharge pressure is low. Is there a 63FL-1 Low Liquid Fuel Supply pressure switch, and, if so, was it also alarming about low pressure?

Also, there is a limit switch, 33FL-1, on VS1-1. Usually, this limit switch drives logic signal L33FL1C to indicate the Liquid Fuel Stop Valve is fully closed--which is a permissive for a 'Ready to Start' indication for most GE-design heavy duty gas turbines which burn liquid fuel. Sometimes there are two limit switches on VS1-1, the other being 33FL-2 driving signal L33FL2O, and the main purpose of 33FL-2 is to provide indication when VS1-1 is fully open. Does the unit have 33FL-2 driving L33FL2O, and, if so, do you know if it was indicating VS1-1 was open when the 63FL-2 was indicating low pressure?
 
B
>The Liquid Fuel Check Valves at the fuel nozzles (burners)
>usually have a cracking pressure of about 100 psig--which is
>above the liquid fuel supply pressure of approx. 60-65 psig.
>10 psig would be very low pressure. Have you verified the
>operation and isolation valve position for 63FL-2?

Isolation valve position checked and verified. It was ok.

>Usually if there's a 63FL-2, there's a 63FL-1, which is
>upstream of the Liquid Fuel Stop Valve, and should also
>alarm if the liquid fuel supply pressure from the forwarding
>skid discharge pressure is low. Is there a 63FL-1 Low Liquid
>Fuel Supply pressure switch, and, if so, was it also
>alarming about low pressure?

L63FL-1 switcg is not present in Mark-V logic. However there is pressure transmitter at upstream of stop valve and is monitored by plant DCS system. during the incidence this pressure was smooth and normal at 4.5kg/cm2.

>Also, there is a limit switch, 33FL-1, on VS1-1. Usually,
>this limit switch drives logic signal L33FL1C to indicate
>the Liquid Fuel Stop Valve is fully closed--which is a
>permissive for a 'Ready to Start' indication for most
>GE-design heavy duty gas turbines which burn liquid fuel.
>Sometimes there are two limit switches on VS1-1, the other
>being 33FL-2 driving signal L33FL2O, and the main purpose of
>33FL-2 is to provide indication when VS1-1 is fully open.
>Does the unit have 33FL-2 driving L33FL2O, and, if so, do
>you know if it was indicating VS1-1 was open when the 63FL-2
>was indicating low pressure?

L33FL1C limit switch of liquid fuel stop valve is logically AND ed with l63FL-2 that means low liquid fuel trip will initiate only if L63FL-2 is low and stop valve open feedback is positive.
 
barindra75,

What were the Process Alarms at the time of the "trip"?

Without the list (times and text messages) of the Process Alarms just prior to, during and after the "trip" there's no more help that can be provided.
 
Top