Hi All,

I know this issue would seem to be discussed in other threads before but i really could not find an answer to my questions in threads i found.

Site have a GE frame 51P unit which only operated on Liquid fuel/Standard combustion system (non DLN). unit originally was been equipped with Mk II control system but two years ago it was been upgraded into MK VIe.

recently, unit tripped on exhaust over temperature while running on base load. we did checked the LF bypass valve manually by actuating it by air and we replaced the FD, servo VV, magnetic pick ups, welded FD to eliminate vibration and incorrect MPUs readings.
later we tried to start unit for more than once and it was tripping on fail to ignite or overtemperature.

GE was contacted and they did send a TA, he did change the null bias current (without following calibration procedures) he started with a default value then unit tripped on overtemperature, he changed the value and again unit tripped on Fail to ignite. lastly after the last change, unit started with no problem.

Truly, we spent almost over two weeks troubleshooting this unit as we tested almost everything (check valves, FD, LF Pump, Forwarding system, Filters,..)and TA only spent just 3 hours and solved it!!!!

Now i have some questions hopping that someone could kindly help answering it:

1. in such trips (over temperature and fail to ignite) when can one be sure that it is related to null bias current vale. how to be sure about it? what are the exact symptoms?.

2. in units like frame 51P that doesn't have motor driven hydraulic pump. how to calibrate/calculate null bias current for LF bypass valve????? i believe the followed try and error method is not a recommended one.

3. How could the null bias value have that much impact on the unit behavior?? we are talking about a very near values. i mean does a servo with null bias of 2.5 behaves totally different when it changed to 2.6???

Finally, your contribution is highly required and highly appreciated.

Thanks to all of you.

 

Hi, Gemi,

This is the second post you've made providing very little in the way of useful information in the original posting and without much data to use to analyze the problem/event. In the 9FA post, you've admitted that there was a failure to make copies of Trip Log files and there's not much data to go on; Historian data can be pretty lacking as you've also admitted.

You said the unit does not have an AC motor-driven Aux. Hydraulic Oil pump. And you've also said you used air to actuate the hydraulically-actuated Liq Fuel Bypass Valve. Most sites without an Aux. Hyd. pump crank the unit to obtain hydraulic pressure for troubleshooting and maintenance operations requiring hydraulic pressure. Other sites have fashioned some tubing and connections to use the Hydraulic Ratchet Pump to supply pressure to the hydraulic system. Still others have built a portable AC motor-driven pump that they can connect to the Hydraulic system to provide pressure/flow for troubleshooting and maintenance activities.

But, I've never heard of a site using compressed air to stroke the LFBV. (Until today, that is.)

We have no idea what transpired before the Exh. Overtemp. Trip. We don't know that the exhaust temperature spreads were. We don't know how the Liquid Fuel Forwarding (supply) system was working. We don't know how the Liq. Fuel Pump Clutch was working. We don't know how the Liq. Fuel Flow Divider was working. When they installed the Mark VI did they properly configure the resistors in the Liq. Fuel Pump Clutch circuit for the coil being used on the machine (if any were required; depending on the coil, they may not be, but because you mentioned the unit previously had a Mark II it may need them)?

Sounds like you took a "shotgun" approach to the problem, replacing and welding and stroking and all. And, all of that introduced air into the lines. Describe how you flushed the air out of those lines prior to attempting to fire the machine? (Specifically, length of time, methods, equipment, results, etc.)

Using compressed air in the hydraulic lines of the Liq. Fuel Bypass Valve also introduced air into those lines and the actuator itself. I'm not even sure how well the servo-valve would operate (I don't think it would, so you probably removed it, didn't you?) with compressed air. Can you be certain that all of the air was removed from the LFBV (Liq. Fuel Bypass Valve) actuator? Please detail the method you used to purge the LFBV actuator of air prior to installing the servo valve and applying hydraulic pressure to the actuator?

We also don't know if a TMR Mark VI was installed on the unit in place of the Mark II, or if the Mark VI is a SIMPLEX unit. (It makes a difference in the amount of null bias which is required, if the unit has two- or three servo coils.)

Because the typical LFBV used on most GE-design heavy duty Frame 5 gas turbines doesn't have LVDTs, it would probably be <b>very</b> difficult to observe the effects of any change of null bias current value(s). Now, there might be an odd Frame 5 unit that has LVDTs on the LFBV, but it would definitely be the exception rather than the rule. (Some older units with variable displacement pumps had LVDTs, but you didn't say the unit had a variable displacement liquid fuel pump; you specifically said it has a LF bypass valve, which, again, would not typically have LVDTs.) And, to really observe the effects of changing the null bias current on a Liq. Fuel Bypass Valve without LVDTs, one would likely need to see Liq. Fuel Flow Divider feedback and LFBV servo current during operation (when there was stable fuel flow).

And <b>two more questions</b> about null bias currents. When the unit was being controlled and protected by the Mark II, how was Hydraulic pressure established? And was it ever necessary to "calculate/calibrate" null bias current values with the Mark II? (I believe the Mark II had potentiometers to add the null bias to the servo-valve output current, and I believe the Control Spec.-System Adjustments had some mention of making sure the null bias was correct.)

When there is air in the Liq. Fuel lines (not even mentioning the hydraulic LFBV actuator just yet), it's pretty common for there to be some pretty tough times starting the turbine. It's very common for there to be 'Excessive Fuel Flow' alarms (possibly even trips), but I'll wager those were forced out (weren't they?), and sometimes exhaust overtemperature trips, and exhaust temperature increases without flame detection, and/or flickering flame detection, and/or white smoke coming from the exhaust, and loss of flame trips, and sometimes, just plain failures to ignite--until all of the air is purged from the system. But, we didn't hear about any of that.

And, let's not forget the air in the LFBV hydraulic actuator. If there was air in that actuator that caused instability, along with air in the Liq. Fuel system, that's not going to make for very easy starting.

The problem was more likely "resolved" by the removal of all air from the LFBV hydraulic actuator and the Liq. Fuel Flow Divider and individual Liq. Fuel lines up to the nozzle.

If you won't crank the machine to establish hydraulic pressure, then it's pretty safe to assume the machine wasn't cranked to perform some (or at least even one) "false fires" (with the ignitors disabled) to purge the air from the liq fuel system up to the fuel nozzles. (How would you know the air was all gone? When the flow divider feedback was stable and steady streams of liq. fuel could be seen coming from the false start drains.) And, this would also require some cranking to purge the turbine and exhaust of unburnt liquid fuel prior to enabling the ignitors to fire the machine. But, it doesn't sound like this was done based on what we've been told. (It's even possible to think of a way to purge the air from the Liq. Fuel system without cranking the unit (though it would require some materials and labor).)

Which brings us full circle to the lack of information provided. And now you're posting that our reply is highly "required". Highly required.

Highly required.

Indeed.

Here are some recommendations based on the information that has been provided:

Get some training for your technicians.

Develop some standard operating procedures so that important data is collected and isn't overlooked or lost.

Look into alternative methods of establishing Hydraulic pressure for troubleshooting and maintenance activities.

 

Hi CSA,

Firstly, i would like to apologize for using the expression "USING AIR TO ACTUATE THE LFPB VV" I did not mean that we used air to actuate as i believe it can not be !!

we used air just to stroke LFPB VV into Fully closed and Fully Open positions and that is to make sure the valve is not sticking and as a flushing at the same time. and we usually remove the servo during this process.

Secondly, regarding air in LF Piping. we have a procedure that is divides Purge process into two stages:

(1) From Forwarding skid to upstream LF Stop VV by using vents on Skid and Filters. of course Forwarding pump is running all the time.

(2) Then we Open stop valve by force, and vent thru HP Filters High vent point.

i would like to state that, this procedure is always being followed before any startup even if there is no problem. and during the last issue, we used to apply this procedure for at least two cycles before any startup.

Thirdly, if less information is provided in this thread this is because i have more generic questions. remember, we do not have any issues any more. i just wanted to fully understand the whole relation between Null bias value and all LF combustion issues.

but regarding our case here:
1. we did fully upgraded Mk II into TMR MK VIe.

2. servo is 3 Coils controlled by PCAA board.

3. LFPB VV is not equipped with LVDT it is only MPU is used as a feedback.

and regarding the first overtemperature trip:
Unit was running on Base load without any issues as spreads were in normal operating ranges. forwarding skid was supplying pressure with normal pressure (we keep operation Logs and it is hourly based) only before the trip the flow from FD started to fluctuate for short period then unit tripped.

and as usual, since MK VIe upgrade we always had a noisy MPUs from startup to almost firing speed. so this is the reason we welded the FD as MPUs noise could be the reason for the trip.
and please remember, we tried to troubleshoot the problem step by step (i mean we started by changing the servo with another one of the same PN. then we had startup and unit tripped. as a result we moved to the next possible cause and so on...).

and please also keep in mind that, in my first statement i did not detailed all troubleshooting i just stated briefly as IT IS A GENERIC THREAD AND NOT A SITE/UNIT RELATED.

Fourthly, some of us here used to Crank the unit to establish Hyd pressure then start to actuate/calibrate the LFPB VV. but this method have a very serious defect that, the LF PP is directly connected to the shaft and at Crank we have to open the LF stop valve to maintain PP seal/lubricate (it is self seal) and with actuating the Bypass valve, fuel will be injected into the turbine. with old units like these there is always a high possibility of false start drain blockage piping and hence Fire in next startup...
we also tried to use the ratchet pump, but it is not designed for such long cycles of operation so it keeps overloading all the time after short periods of operation and we have to wait till it cools down before start it again.

again, my purpose from this thread was to generally understand the relation between Null bias and all LF combustion issues and how to accurately calculate the best optimum value.

Thanks for you reply it is really appreciated.

 

One of the purposes of adjusting the null bias would be to make the actual value being controlled by the servo-operated device be closer to the reference value. So, for a LFBV without LVDTs, the actual value would be flow divider feedback and the reference value would be liq fuel flow-rate.

When LVDTs are involved, it's much easier to see the effects of changing the null bias. Without LVDTs, you would need to monitor actual fuel flow-rate versus flow-rate reference to see the effect of any change in null bias.

Now, if the firing FSR was marginal to begin with, and the null bias was increased, that might make firing a little easier (presuming that increasing the null bias had the effect of increasing fuel flow).

I've never been a fan of modifying the default value of null bias for non-DLN machines. There are, literally, thousands of machines (with both conventional combustors and DLN combustors) running perfectly fine with default null bias current values.

The purging process you described will only purge the air out of the system upstream of the flow divider. If you replaced the flow divider, there would likely be air in the lines between the high-pressure liquid fuel filter and the flow divider, and the flow divider, and downstream of the flow divider in the individual fuel lines to each fuel nozzle in each combustor. So, venting through the HP fuel filter would not eliminate the air from the piping downstream of the filter. Sure, some air might make it's way up there, but with no place for the liquid fuel to flow (the check valves prevent flow because of their cracking pressure, around 100-120 psig, or thereabouts) the fuel won't move very far downstream of the HP fuel filters during your purging process.

The LFBV is OPEN when it's calling for zero flow, recirculating the output of the pump back to the input of the pump. It's a "bypass" around the pump (hence the name). So, when the Liq. Fuel Stop Valve is opened with forwarding pressure upstream of the LFBV, if the LFBV is at the zero flow position it's full OPEN and fuel will flow straight around (no pun intended) the High Pressure Liquid Fuel Pump, as long as there is a path for the fuel to flow.

So, if the fuel lines to each fuel nozzle were disconnected, and some flexible hoses temporarily connected to the disconnected fuel lines, and the hoses were run to some catch basin or even temporary barrels to catch the fuel oil, then just by starting the forwarding pump, opening the forwarding stop valve, and opening the Liq. Fuel Stop Valve (which generally requires Hydraulic pressure) then fuel would flow to/around the pump (it would be lubricated) through the LFBV, through the HP fuel filter, through the flow divider, and through the individual lines, and out of the temporary hoses. When a stable flow of fuel was seen coming out of each hose, the Liq. Fuel Stop Valve could be closed, the temporary hoses removed and the fuel lines reconnected to the nozzles. There would be a small amount of air very close to the nozzles, but not very much air at all. One could still vent through the HP (and LP, if present) filters with forwarding pressure, and any high point piping vents.

If air is used in the LFBV actuator, I don't know how it would be purged, unless a flushing block was installed in place of the servo and Hydraulic pressure/flow was used to flush the actuator of air, and then the servo re-installed. (GE used to provide flushing blocks for the Liq. Fuel Stop Valve actuator, the LFBV actuator, and any other servo-operated actuator for flushing during commissioning. They were usually lost after that, but were supposed to be turned over to the Customer for future possible use.)

Hence, my contention that all the GE TA did was get lucky on his visit to your site. If he made incremental changes to the null bias (you indicated changes of 0.1%, which is 0.001 mA, since 10 mA equals 100%), then it's likely the changes were ineffectual at best, and all of the air was finally purged from the both the Liq Fuel system and the LFBV actuator and the unit fired.