Hello, it has been a while.

In normal operation on the other 3 GE 7ea Gas Only Simple Cycle units on-site, the TTRF1 would raise through 1950 deg F on up past 1985 deg F through the transfer to Premix Steady State.

The other morning during a cold start (<20 deg F), one of the units here was unable to reach Premix Steady State after a few attempts even once warmed up. This particular unit was ramping up in the normal fashion, Primary, Lean-Lean Pos, Lean-Lean Neg, and then onto Premix Transfer @ 1950 deg F. The issue we are seeing is that rather than having the TTRF1 temperature continue to raise from 1950 past 1985 deg F, this unit was plateauing around 1950 deg F. This temperature dip caused the unit to indicate "Modeless" for a moment and flash back to Lean-Lean Neg mode, eventually leading to an Extended Lean-Lean condition.

All of the unit instrumentation was reading correctly and well in line with the other 3 units that were started successfully alongside this one. (CTDA, CPD, afpap, afpcs, afpbd, afpep, etc.) Going back and looking at historical data, I can see that since 2023 this unit has barely made it to Premix Steady State with TTRF1 hovering just past 1955 deg F during Transfer during cold weather starts. (Fall of 2023 was the last C.I. with GE)

Once in Premix Steady State, this unit has excellent NOx and CO emissions levels while fired around 2055 with <70 deg F spreads at Base in the Summer and Pre-Selected 84MW in the Fall/Winter.

Any feedback or insight into this condition is appreciated.

Thank You,

Bob Grave

 

Could the transfer purge valves be leaking by? Not the actuators but the valves themselves?
Also, the SRV actuators are known to get a flat spot in actuation that causes them to "stick" when moving very slowly if they haven't been serviced, usually caused by the dither setting in the servo.
Anyhow, you can check for a "stick" in actuation by performing an offline calibration of the SRV. Then, with the calibration mode in MANUAL, stroke the valve to about the position it is in during transfer and step it 0.5-1% at a time and see if demand starts to distance it's self from feedback. Sometimes it will not show up when moving the valve 5% or more but when you move it very slow, it will stick and jump to setpoint.
If it is sticking, you can usually stroke it open and closed a few times to flatten it out temporarily and schedule a time in the near future to swap out the SRV.

 

Hello.

I was able to close the purge valves and pressure test the space between them with 80psi. The pressure held for quite some time.

For the SRV, looking at the trend during startup and transfer, FSGR was smooth, opened proportionately with P2 pressures, and matched the other units in startup at that time.

This is a head scratcher.

 

Good checks.
What about the flame scanners? Are the primaries showing 0 flame during the transfer or are they flickering?
Also the gas valve demand vs feedback during the transfer? TTRF1 is slow to increase so maybe something holding the fuel back.

 

We were able to find a fairly substantial leak coming from two of the can to CDC flanges on #4 and #9. Checking torque values on those can bolts today.

Looking through the past trends from the past attempts, the flame detectors were operating properly along with gas valve demand vs feedback. Ultimately, the TTRF was just too low to proceed to Premix <1975 deg F. (L3FXPM1) The thought process is that the loss of air was a contributing factor to the lower TTRF.

Thanks again for the ideas. I will update later today hopefully.

 

@bgrave_1,

What Process Alarms were active prior to the combustion mode change, and what Process Alarms were annunciated during the aborted transfer?

Any related Diagnostic Alarms?

USUALLY when a DLN-I machine goes into a combustion mode transfer, particularly Lean-Lean to Premix, the Mark* initiates automatic loading to ensure the machine increases load (fuel flow) to try to ensure the machine goes through the transfer fully. (Sometimes when a Pre-Select Load Reference is set that is "in the middle" of the automatic loading after the transfer the Mark* will unload the machine which can cause a transfer from Premix to Lean-Lean--and the Mark* gets blamed for not being able to maintain the Pre-Selected Load Reference. Of course, the operator is at fault because they selected/set such a load reference that DLN-I isn't capable of maintaining under the current machine operating conditions. Of course, the current machine operating conditions vary and that requires adjusting the Pre-Selected Load Reference to something slightly higher but it usually takes operators months, even years, to understand this. So, the Mark* continues to be blamed for an operator error.)

Personally, I've a lot of experience with DLN-I and I have never seen this particular problem. 20 deg F is officially cold for a machine in my opinion. Does the machine have IBH? Is the IBH control valve working properly? If the actuator is pneumatic, is there a possibility some moisture in the tubing supplying the air to the actuator is frozen? Does it have a low point drain? Is the air pressure regulator upstream of the actuator unfrozen and working properly?

Are there any CTD T/Cs not reading properly ("large" difference between T/C reading)? CTIM T/Cs working properly with no differences?

What is the perversion of the Mark*?

 

2001 7EA units here.
These units have Mark VIe Migrations from 2019. No diagnostics during the attempted transfers, only Modeless for a moment, re-igniting Primaries, and flashing back to Lean-Lean Neg Mode until it times out to Extended Lean-Lean.

This TTRF issue of not continuing to raise past 1950deg F in transfer happens in Pre-Select 84MW or Base Load select. These units start transfer around 37-39MW so we are well outside of that crossover when we select 84MW.

The Inlet Bleed Heat valve is working correctly with demand and feedback tracking correctly with downstream bleed heat line pressures and inlet air temp increasing once the valve opens at FSNL and throughout the attempted transfer.

CTD numbers are within 1-2% and track almost exactly the same as the other units here.

Thanks again.

I will update once we evaluate the leaks and hopefully address them.

 

After some additional guidance, I was able to check the partload curve and it was set just a bit too cold. All 4 points were raised roughly 15 degrees and we were able to get through transfer into Premix.

Thanks again for all of the suggestions.

 

@bgrave_1,

Thank you for the follow-up!

I presume this was after a "recent" DLN-I tune and that this machine was set differently from the other two.?.?.?

I'm also wondering if the 'Mark* VIe Migration' you are referring to was one of the several (per)versions of the Mark* V to Mark* VIe Migration or ???

Anyway, just curious.

And again, thanks for the follow-up. It's what really sets Control.com apart from other World Wide Web forums.

 

Sorry for the delayed response. The 7EA units here have an Auto-Tune software that does not revert back to baseline GE tunes between runs. The previous run on this machine was on a roughly 50 deg F higher temperature day. The unit had ran for several days with that cooler automatic tune.

One of the drawbacks we are seeing is that many of the 3rd party Auto-Tune software systems only tune one direction and don't seem to remember their starting values or associate a different tune with hotter or cooler inlet temperature days.

The adjustment that was made to the TTKGVI curve on this unit solved the Transfer issue while cooler out. We also tightened up the Auto-tuning curve deadband to hopefully prevent this error in the future.

Thanks again

 

@bgrave_1,

Thank you very much for the information. I was wondering if the control system had AutoTune, and now I know. I think AutoTune works well for machines that run more or less continuously, or start every few days. But, peakers that run infrequently are quite well-suited for AutoTune capabilities. And now I understand why!

Happy Holidays--and thanks again very much for the feedback and information!

 

@bgrave_1,

Is the AutoTune of which you speak an OEM version or an aftermarket supplier's?