GE Frame9e tripping during ramp down at around 2000 to 2300 RPM. while coasting down on these RPMs SRV fluctuations between command and feedback seems obvious and difference between command and feedback become around 0.5 to 0.8%. P2 pressure also tries to maintain, due to completion of 8 minutes from shutdown command till this fluctuation, machine gets tripped. during this phase SRV position is around 14 to 18%. Trend and alarms also attached.

 

@ali arsalan,

Please tell us if this problem just started, OR if it started after a "calibration" of LVDTs (SRV and/or GCV)????

Does the machine normally run on natural gas fuel?

Does it normally shut down without problems when running on natural gas fuel?

What is the model of the Mark* turbine control system?

What is happening to the GCV during this period when the machine speed is decreasing at a very slow rate?

Has the source of the natural gas fuel changed recently? Is the natural gas fuel being burned now different--LHV and/or HHV or methane content?

 

Please tell us if this problem just started, OR if it started after a "calibration" of LVDTs (SRV and/or GCV)????
The problem has started without any calibration.
Does the machine normally run on natural gas fuel?
Yes
Does it normally shut down without problems when running on natural gas fuel?
Yes previously but not since last 8 10 months
What is the model of the Mark* turbine control system?
Mark6
What is happening to the GCV during this period when the machine speed is decreasing at a very slow rate?
While machine is ramping down, SRV position till 20 remains normal while going further down a difference between cmnd and fb of around 0.5 to 0.8 of srv starts coming. This continues for 5 to 6 minutes untill completion of 8 minutes and machine gets tripped. Srv normally keeps on fluctuating between 15 to 17%
Has the source of the natural gas fuel changed recently? Is the natural gas fuel being burned now different--LHV and/or HHV or methane content?
All normal we have checked these all things.

Note: servo was replaced by the team but null biased was not perform, servo was replaced by the team due to same issue. I was not part of the team that time. I have recently joined. I am doubted may be null biasing can do some betterment. Need suggestion plz thanks.

 

Regarding your GCV question, gcv is coming down normally as you can see in the trends but when srv starts fluctuating gcv keep on standing at 11%, but as per the history previously normal shutdowns were a routine. No cmnd fb error of gcv found, can ne scene in the trends

 

@ali arsalan,

I travel a lot and I work unusual hours and today I don’t have a computer to look at the “trends.”

I neglected to ask if the machine has conventional combustors or DLN-I combustors.

This problem usually results from either a change in GCV LVDT feedback calibration OR a change in fuel make-up OR an issue with SRV controlling P2 pressure properly (presuming no one mucked with start-up/shutdown fuel control parameters (Control Constants).

You started the difference between SRV command and feedback was only around 0.5-0.8 percent without stating if the position was erratic (above and below the command) or stable above the command or below the command. The SRV moves to whatever position is required to make the actual P3 pressure (usually signal fpg2) equal to the reference (usually signal FPRG). The primary purpose of the SRV is to be the gas fuel stop valve. The secondary—and very important—function is to control P2 pressure upstream of the GCV based on machine speed.

If the machine speed slowed down and possibly stalled then too much fuel was flowing into the machine to allow the speed to decrease. This could be because the P2 pressure was too high, which could be because of a problem with the P2 pressure transmitters or a problem with the SRV or the SRV actuator. Again presuming the start-up/shutdown Constants haven’t changed.

You mentioned something about the machine not having shutdown on gas fuel for 8-10 months. Was it running on a different fuel or just idle during that period?

 

P3 pressure is stable.
P2 starts hunting when srv starts hunting.
Fb error is mostly on the above side of command. For example if command is 15 fb is 15.5. That might be the reason we got a null bias of less value that is 19.5
Regarding shut down for 8 to 10 month I meant to say 8 to 10 months before the shutdowns were normal.
Turbine always runs on gas fuel.
It is a non dln machine with 1 gcv.
We are taking a shutdown after 2 days. Do you think null biasing and saving the new value could do the job for us.

 

Your replies will be highly helpful. Thanks alot

 

@ali arsalan,

Is the Mark* VI a TMR or SIMPLEX control panel?

Do you know if the application code running in the Mark* VI determines the null bias current automatically or is the null bias current value a static value that is entered in the regulator configuration?

A value of 19.5 for null bias current is extremely unusual. I don’t recall if the application code null bias current calculation had any limits on the value it calculated.

The first thing I would recommend when the machine is shut down (after the gas fuel supply is isolated and locked out and the gas fuel pressure upstream of the SRV is vented) would be to manually open and close the SRV using the Autocalibration function to monitor how smoothly the valve moves and to observe the servo current as the valve is moving. I would recommend opening/closing the valve pretty slowly in increments of 1% or 2% at a time during this testing and monitoring the SRV servo current outputs of all three processors (presuming the Mark* VI is a TMR panel).

This should help determine if there is a problem with either the servo valve OR the SRV actuator (I’m presuming the SRV is hydraulically operated). The three servo currents should be relatively stable and equal and have the same polarity. If the SRV actuator is sticking or binding, especially around the position it was at when the instability during the trip occurred) that could indicate a problem with the hydraulic actuator or the servo. If the three servo currents are VERY different and have different polarities that could suggest a problem with the servo valve. If the null bias current is calculated by the application code running in the Mark* VI that could also be suspect.

The null bias current of the regulator should rarely be anything other than 2.67, plus or minus 1.3 (meaning the null bias current specified in the servo regulator should be a value between 1.3 and 4.0). When the SRV is maintaining a stable position (during testing) the individual servo currents should be at or near the null bias value specified in the regulator configuration. And 19.5 is a wildly out of range number.

If the o-rings in the SRV actuator are leaking that could explain excessive servo current required to maintain a steady position or be indicative of servo valve problems. But currents of 19.5 to maintain stable position are really difficult to imagine. Especially if all three processors are at 19.5.

I hope to get a chance to look at the data you attached to the original post later today after I get some food and sleep and a shower and I can use my computer to scroll through the data. But something is unusual about this problem or its description and I can’t imagine what it is at the moment. 19.5 is such an unusual value of null bias current.

 

Null bias value 1.95 not 19.5

 

I am sorry it was a typo error null bias value was found 1.95. Plz suggest accordingly. Extremely appologetic

 

@ali arsalan,

1.95 is still kind of an unusual number--UNLESS what you are referring to as "null bias current" is the value of servo current (per processor) when the machine is running an the Mark* VI is properly controlling P2 pressure. And that's NOT what it typically referred to as null bias current. Null bias current is the amount of current that is "added" to the servo current output to overcome the "failsafe" spring tension in the servo valve. The amount of current one sees when the machine is running at stable load and speed should be approximately equal to the null bias current value (normall between 2.67, +/-1.3)--but under real operating conditions is never exactly equal to the null bias current value specified in the servo regulator configuration.

The data you supplied is once every six seconds--that's really poor resolution data for troubleshooting a problem like this.

Can you just post the raw data, please? (Sorry; my eyes and brain just aren't ready for jumping around the pages to try to get a good look at the low resolution data.)

By eight minutes into the shutdown the speed should have decreased a LOT more than just slightly more than 50%; that's what the timer does--it "monitors" the shutdown and if it takes longer than a preset time then it just shuts the fuel off. So, the problem here is even though the P2 pressure isn't really stable the fuel flow to the machine isn't decreasing enough to cause the machine speed to decrease as it should. The problem could be that the fluctuating P2 pressure is making fuel flow be higher than it should be even if the GCV position is stable and as per the shutdown schedule, but by itself the 1.95 servo current value isn't going to do that--and it's not going to make the SRV fluctuate (oscillate).

During shutdown after breaker opening typically the Mark* has three or four values of FSR that it selects as the machine speed decreases. The values decrease in magnitude as speed decreases--and if the amount of fuel actually flowing into the machine doesn't decrease then the machine speed won't decrease.

As was written before, the SRV IS NOT a position control loop--it's a pressure control loop. The Mark* moves the SRV to WHATEVER POSITION is required to make the actual P2 pressure (fpg2) equal to the P2 pressure reference. If the LVDT feedback from the SRV LVDTs isn't as it should be then the Mark* could be trying to hit a moving target. OR, if the P2 pressure transmitters aren't working correctly (the pressure feedback isn't stable when it should be) then again the Mark* could be trying to hit a moving target.

I would add a check of the P2 pressure transmitter operation and calibration to the tests to be performed during your upcoming shut down.

The Mark* turbine control system for GE-design heavy duty gas turbines and devices like the gas control valves and the IGVs uses a very different type of control scheme than most other OEMs use. When the device (in this case the SRV) is at a stable position the servo output currents from should be slightly negative--by a value which is very near (in magnitude) to null bias current specified in the servo regulator. Most OEMs use a control valve driver that is proportional to the amount of pressure or flow or position--but NOT the Mark*. So, just to say a servo current output is 1.95 at ANY time during operation doesn't tell us anything about the reference pressure/flow/position or the actual pressure/flow/position.

Anyway, that's all I have time for; it's time for food and sleep because in about six hours, I got to be back on site for more hurry up and wait (which is what will be written on my tombstone-HUAW (Hurry Up And Wait). If I had even 5 cents for every hour I've spent hurrying and waiting I would be a rich man--VERY rich.

Tchau!

 

Thanks for such great responses. Well we have already verified the P2 related transmitters and all are normal. I would like to ask you do we still need to put the new null bias value since it has not been saved after last servo replacement. secondly what other things do you suggest. do you suggest any thing with GCV or SRV to be checked.

 

@ali arsalan,

I really don't have any idea what you are calling a "null bias current value."

I don't know where you obtained the "null bias current value" you are asking about saving.

For the majority of GE-design heavy duty gas turbines the typical null bias current value of 2.67 is the correct value for almost all machines with TMR Mark* control systems. And, the limits of adjustment--if necessary--are between 1.3 and 4.0.

Again, for the majority of GE-design heavy duty gas turbines it is not necessary to change the default null bias current value of 2.67. AND, it's only necessary in SOME cases. If you or someone at your installation is reading the Control Specification and thinking they need to adjust the null bias current value using the procedure in the Control Specification, you or they are wrong. WHENEVER a servo valve is replaced the default null bias current value should be entered--and saved--in the servo regulator configuration before running the machine. IT IS NOT NECESSARY TO USE AUTOCALIBRATE TO DO ANYTHING WHEN A SERVO VALVE IS BEING REPLACED--FULL STOP. PERIOD. AUTOCALIBRATION IS ONLY NECESSARY WHEN REPLACING LVDTs OR AFTER LVDTs WERE REMOVED AND RE-INSTALLED WHEN REFURBISHING OR REPLACING THE ACTUATOR/VALVE ASSEMBLY. FULL STOP. PERIOD.

I am referring, of course, to new-in-the-box servo valves from the manufacturer. I AM NOT referring to refurbished or repaired servo valves (many companies providing this service for servo valves do not have the proper equipment for adjusting or verifying the adjustment of the failsafe spring).

The ONLY occasion when it necessary to adjust the null bias current to some value other than 2.67 to something between 1.3 and 4.0 is when the MEASURED PHYSICAL POSITION IS NOT EQUAL TO THE CALIBRATED LVDT FEEDBACK--and to arrive at that conclusion it is necessary to use some external position measurement (for example, for an SRV it would be something like a dial indicator with a magnetic base) to compare against the calibrated LVDT feedback. Small differences or a couple of thousands of an inch are no big deal--and for SRVs LVDT position IS NOT CRITICAL!!! (Because the SRV is not a position control loop--it is a pressure control loop and the Mark* will move the SRV to WHATEVER position is necessary to make the actual P2 pressure equal to the P2 pressure reference. So accuracy of the actual position versus the indicated position from LVDT feedback is not critical. Full stop. Period.

If someone has adjusted the null bias current value to something other than 2.67 for a device and the servo valve has been determined to be malfunctioning and needs to be replaced the null bias current value in the servo's regulator configuration should ALWAYS be set to 2.67 before beginning operation. (And, AGAIN: It is never necessary to perform an Autocalibration when only replacing a servo valve. Autocalibration is only necessary when something has been done to affect the output of one or both LVDTs (such replacing one or both LVDTs, or some physical change in the stroke of the actuator, etc.).

I would suggest verifying the calibration of the GCV LVDTs using the manual function of Autocalibrate.

I would suggest verifying the calibration of the SRV LVDTs--not so much looking for errors in the LVDT calibration, but looking for problems with the actuator or servo valve in maintaining the reference position (such as leaking o-rings in the hydraulic actuator, or worn internal areas in the hydraulic actuator, or sticking operation of the hydraulic actuator--things like that). If the amount of servo current required to maintain a reference position increases (in the negative direction) from what is required for the majority of other positions then that would likely be a problem with the actuator or the servo valve. (Negative servo current opens the valve--against the very large spring which is ALWAYS trying to close the valve! If the servo current increases in some area of the valve stroke and is normal in all other that suggests a problem with the actuator or the valve assembly (sticking; binding; etc.).

If the lube oil hasn't been tested for a while it would be strongly recommended to have it tested by a reputable company. Dirty lube oil is the biggest cause of servo valve issues. (Remember, GE-design heavy duty gas turbines use lube oil as the hydraulic fluid for actuators.) So, oil cleanliness is very, Very, VERY important to reliable long-term operation of servo valves.

It's also very common for the "last chance" filters at hydraulic actuator manifolds to rarely, if ever, be changed. If they get really dirty and don't get changed they can rupture and send dirt and contaminants into the servo valve and hydraulic actuator.

Lastly, the issue here is: Why is the fuel flowing to the machine during shutdown not decreasing enough to allow the machine speed to decrease? A properly tuned and operating GE-design heavy duty gas turbine burning natural gas fuel during a normal fired shutdown should decelerate to about 20% speed without losing flame in any of the combustors with flame detectors at which point the Mark* will close the SRV (and any auxiliary gas fuel stop valve) to shut off the gas fuel flow, extinguish flame in the combustors and the machine will continue to coast down to Cooldown Operation. So, something is preventing the GCV from reducing the fuel flow as it should. It could be the SRV assembly (servo valve, or hydraulic actuator, etc.). That's what you need to determine: Why isn't the gas fuel flow decreasing as it should? Could the gas fuel make-up have changed? Could there be a problem with GCV hydraulic actuator or valve assembly or even the GCV servo valve? (Again, dirty oil doesn't discriminate--it contaminates ALL servos pretty much equally!)

Please write back to let us know what you find and how you resolve the problem. Tripping at part speed (around 50%) isn't ideal, but it's a lot better for the machine than tripping at higher load from rated speed. I might suggest taking data during the fired shutdown prior to the outage, and maybe do another couple of STARTS and shutdowns to get as much data (HIGH SPEED data!) as you can. And then leave time after the outage and before the machines needs to running to do another couple of tests to see if the problem is resolved. AGAIN--tripping at low speed is better than tripping from rated speed and load. Much better. The main purpose of running the machine down to low speed while burning gas fuel is to try to reduce the machine internal temperatures gradually instead of just shutting the fuel down at 95% speed (as was done on older vintage machines). So, a few tests during fired shutdowns aren't the worst thing for the machine. They should help you troubleshoot and hopefully find the root cause(s)--because sometimes there are more than one cause contributing to problems like this. Just run the machine up to FSNL, wait 10 or 15 minutes if the machine was started from cold, and then just initiate a normal STOP and gather the data as the machine decreases speed--good, high speed data, preferrably.

Best of luck!

 

@ali arsalan,

If you aren't familiar with using Trend Recorder (sometimes called Trender) in ToolboxST it's an extremely handy and useful troubleshooting tool. I'm going to suggest some signals to be monitored using Trend Recorder/Trender that were missing from the data you sent in the original post.

fsgr (the LVDT feedback from the SRV)
fsg (the LVDT feedback from the GCV)
FPRG (the P2 pressure reference the SRV is trying to maintain--it changes as TNH changes)
fagr (the median value of SRV servo current)
fag (the median value of GCV servo current)
FQ (the fuel flow rate; sometimes it's FQG for gas fuel)
FSR (Fuel Stroke Reference)
FSR_SD (Shutdown Fuel Stroke Reference)

These signals, along with the signals included in your original data file, would be very helpful.

The default capture rate for Trend Recorder/Trender is very fast and would be most helpful in troubleshooting. It makes for a large .trn file, which you will need to export to .csv files (using the export function in Trend Recorder/Recorder) to attach to posts in Control.com in order for us to analyze the data (unfortunately, there is no "universal" Trend Recorder/Recorder file viewer and one has to have a running copy of ToolboxST in order to view .trn files directly--hence the need to export the data from the trend file to .csv files for others to view and analyze (using Excel or some other spreadsheet app/program).

ONE of the MANY nice features of Trend Recorder/Recorder is that one can use the same file and capture many tests/data sets in one file (of course the file gets very large--but that's not a difficult problem to deal with). So once you get the Trend Recorder/Trender file set up with the signal names you can use those signal names multiple times when testing to capture data for troubleshooting.

YOU CAN'T TRIP THE TURBINE OR CAUSE ANY PROBLEMS WITH MACHINE OPERATION USING TREND RECORDER/TRENDER!!! I can't stress this enough--it's a very benign and very useful feature of ToolboxST. And one can practice using Trend Recorder/Trender any time when the machine is running, starting up or shutting down. In fact, it's a very good practice to have trends of good start-ups and good shutdowns to troubleshoot problematic start-ups and shutdowns in the future. They can also be useful for just analyzing and getting familiar with looking at data from operations. Play with Trend Recorder/Recorder. IT WON'T TRIP THE TURBINE!!! It's a very useful and relatively easy feature to use of ToolboxST.

I don't know if you have or are familiar with Trip History Displays, but when looking at a Trip History Display one is using Trend Recorder/Trender. In fact, that's another way people can get familiar with and practice using Trend Recorder/Recorder, by looking at Trip History files. (NOTE: Trip History files are stored on the HMI hard drive for a limited period of time. When a machine trips for any reason, it's an extremely good practice to automatically open the Trip History Display and the save it to another folder/location on the HMI for future reference. Create a folder on the MS-Windows desktop for Trip History files, and save them there. Usually, the date/time the Trip History file was created is coded in the file name. If your operators or operations supervisors don't already open and save Trip History files when the machine trips--for any reason!--they should start very soon.

If you have questions, just ask. I will be traveling so I will only have access to my phone (I don't like to use my computer in airports and cafes). But I can answer most questions on my phone.