Flow Divider Magnetic pick ups

W

Thread Starter

Wajid Hussain

WE have 1 Gas Turbine Frame 9E (3000 rpm) Mark VI controlled and i work as a Trainee Engineer at the plant. I did not understand the reason/facts behind the problem faced recently. And that is Flow Divider Magnetic Pickups were hunting repeatedly during Start up and Shutdown.Because of this issue Loss of Flame was seen on HMI and Turbine got trip.

I might be not providing full information regarding this issue but I'll provide the information what so ever is required.

Please do lead me to the solution.
Thanks
 
Dear Mr. Wajid,

I have faced the same for GE Fr.6 Machine. The flow divider's (Rotary type) magnetic pickup %rpm reading hunting during start-up.
I have observed that this is happening because of vibrations caused due to the starting means diesel engine. The vibration causes, a slight movement inside flow divider which is getting picked up by magnetic pick up sensor.

I could not sort this out. But during system commissioning, fine tuning of the setting for 'flow excessive trip' was accordingly adjusted to avoid unnecessary trip.

Mounting of flow divider could also add to this.

Regards,
Anish
 
Wajid,

I experienced similar problem on our fr6 mkVI. The magnetic pickup 77fd reading was severely hunting only before the ignition.

I found it is due to interference from starting motor. We utilized 6 kv ac motor. If the starting motor was stopped then the problem was disappear. MkVI acquires the frequency (or period) on the waveform of 77fd to determine the speed, so small signal interference is enough to cause the reading to be hunting. After ignition the interference was still there but too small compared to the actual magnetic pickup signal.

This problem leads me to the shielding issue on the 77fd cable system. I will work on it on the next outage. Maybe you can start from the cable shield as well. Or if you have oscilloscope or scopemeter your investigation will be much easier, because seeing visually the signal is a good advantage.

Hope it helps,
Riko
 
Wajid Hussain,

In an effort to improve the zero speed detection capability of Speedtronic turbine controls (especially the Mark VI and Mark VIe versions) GE has modified the algorithm used in the speed detection input circuits of both the turbine shaft speed pick-ups and the liquid fuel flow divider speed pick-ups, which has lead to a <b>LOT</b> of problems for both turbine shaft speed pick-ups and liquid fuel flow divider shaft speed pick-ups.

In the past, liquid fuel flow divider speed pick-up sensing was critical at low flow-rates (low flow divider speeds) during starting and accelerating. So, one of the ways the older Speedtronic panels (which used analog circuits) improved the low speed sensing capabilities of the liquid fuel flow divider speed pick-ups was to decrease the gap between the flow divider toothed wheel and the speed pick-up face to 0.010 inches, as opposed to 0.050 inches for the turbine shaft speed pick-ups (which are usually the exact same speed pick-ups as those used on the liquid fuel flow dividers).

So, if the liquid fuel flow divider shaft speed pick-ups are gapped at 0.010 inches, try increasing the gap to as high as 0.040 inches. It usually helps--a lot. The improved speed detection algorithms in use in the Mark VI and Mark VIe can still detect low speed at the larger gaps but don't seem as affected by mechanical vibrations transmitted from starting motors or hydraulic pumps at zero liquid fuel flow or low liquid fuel flow-rates such as during starting and accelerating (and even shutdown). The toothed wheels in the liquid fuel flow dividers can be affected by mechanical vibrations transmitted from other components. Even changing the stainless steel tubing used for conduits for the speed pick-up wiring to flexible stainless steel hoses can have a positive impact. Some sites have even gone so far as to mount the flow divider on rubber disks and use flexible metal hoses to connect the output of the flow divider to the manual selector valve (extreme, but it does work if the gap is left at 0.010 inches--which it doesn't need to be at on newer, digital Speedtronic turbine control systems).

This problem is especially prominent on older turbines that have been retrofitted with Mark VI and Mark VIe turbine control systems. Some of the teeth on older liquid fuel flow divider toothed wheels are very thin and under a large speed pick-up face will cause even more problems. Even the type of cut of tooth of the toothed wheels can exacerbate the problem. Roper offers improved toothed wheels with wider teeth and better bearings which also help alleviate the problem of mechanical vibrations transmitted from starting motors and hydraulic pumps, etc.

Typically, shielding of the wire is not an issue (as long as shielded wire was used!).

Please write back to let us know what you find and how you resolve the problem.
 
W

Wajid Hussain

We have checked the signal and shield wire of magnetic pick ups and found normal. During our last start up what we have observed that fqrout (Liquid Fuel Bypass valve Servo command) was stable and magnetic pick up feedback fql1 was continuously hunting. Due to this Servo current and valve were hunting. This all happens when Turbine reaches at FSNL. After synchronization with grid it (fql1) started getting better but the response became smooth at 40MW load.

Our turbine starts on Liquid fuel then changes over to Gas after 40% load due to low Btu gas.
 
Wajid Hussain,

The liquid fuel supply piping must be clear of all air; this includes all the fuel filter canisters as well as all piping.

Also, many liquid fuel forwarding systems have pressure regulators (some don't) and these pressure regulators generally require some adjustments to get the pressure to be stable.

If the supply pressure is unstable, then the control valve will be continually trying to respond to changes in pressure which will cause changes in flow. So, stable liquid fuel supply pressure is very important.

Stable flow at higher loads will usually occur when there is air in the piping and/or the liquid fuel supply pressure is not stable at low loads.

Please write back to let us know how you resolve the problem.
 
I faced the problem again after servicing was done on diesel engine side. The machine tripped on Liquid control fault during acceleration.
One of the flow divider magnetic pick-up is steady, while the other hunts to very high levels.

But the same pick-up reads steady when the machine reaches self sustaining speed (diesel engine disengaged). Further on, at machine loaded condition both the magnetic pick-ups are ok.
 
We had the same problem whenever the gas turbine (fr 5 -Mk Vie) is running on gas fuel also. i.e the signal (fdbk-77fd1,2) has a non zero value and varying up to certain values, approximately 10HZ. this caused alarm of (liquid fuel bypass trouble) and inhibit transferring the fuel to liquid fuel if this value increased above a certain constant (3) for 10 sec.

then i checked the overall loop from the sensor to the MK VIe panel and found it OK. Then i realized that the main reason of this reading is the vibration of the turbine at accessory gear. and you can find if you noticed this feedback from the HMI while some hitting on the sensor with his hand.

so you have to increase the protection margin to avoid any fake trips.
 
V

Vhendrayawan

Good day,

I have similar problem in my gas turbin, hunting flow HSD (value of fql_pr). Problem indentification only 1 signal (fql_pr1) have problem, other signal fql_pr2 OK. Our analysis from this problem, possibility root cause due to fly wheel gear condition (not fix gear condition by acc gear vibration).

Please for information if you have standard backlash fly wheel gear flow divider 9E (Roper), please share to me.

please for advice
My Best Regard
 
Vhendrayawan,

I can't address the allowable run-out specifications, but I do have a couple of suggestions. (Roper has always been very helpful when I've contacted them for assistance; I suggest you contact them for their recommendations on when maintenance should be performed, and possibly what the allowable run-out specifications might be.)

Based on the signal names you typed I would guess the unit has a Mark VI or Mark VIe. Both of them have very sensitive speed pick-up input circuits, and just about ANY looseness in the flow divider bearings/gears can cause problems. If you're only having issues with on speed pick-up, have you checked the gap of the pick-up, and have you checked all the terminations between the speed pick-up and the turbine control panel? And, have you checked to make sure the speed pick-up wiring is properly shielded, and that the shield drain wire is terminated at one end only? (There are several good threads on control.com with some "drawings" about acceptable and unacceptable shield drain wiring practices, available by using the 'Search' function at the top of every page.)

Finally, because of the highly increased sensitivity of the Mark VI/VIe speed pick-up inputs, many sites have had good success increasing the speed pick-up gap to as high at 0.040 inches (approximately 1.02 mm). This will only help for so long if the bearings/gears in the flow divider are worn. (Flow dividers, like all other mechanical pieces of equipment, do require periodic maintenance and refurbishment.)

Please write back to let us know how you fare!
 
V
Dear CSA,

I'm sorry for late reply.

I already check possibility related this problem:

1. Gap sensor pickup and flywheel gear -> 0.01 inch (manual book standard /general electric)

2. backlash flywheel gear -> 5mm from diameter flywheel (not fix condition during vibration -> but we dont have std related this phenomena)

3. Check output cable -> terminal -> I/O module Mark VI -> result not found abnormality (not founding noise etc)

in our plant using Mark VIe (after retrofit from Mark IV, old plant)

My Best Regard
 
Vhendrayawan,

Thank you for the feedback.

I purposely mentioned the sensitivity of the Mark VI/Mark VIe speed pick-up inputs (including those of the liquid fuel flow divider) because it's a real problem for older machines upgraded/retrofitted with newer control systems.

In the Mark IV era, the liquid fuel flow divider speed pick-ups were adjusted very close to the toothed wheels of the flow divider because it was necessary to get a good signal at low RPM when the unit is firing and accelerating. You will notice that the turbine-generator shaft speed pick-ups are gapped at 0.050 inches, and the liquid fuel flow divider speed pick-ups were gapped at 0.010 inches.

I personally have been to Mark IV-to-Mark VIe Migration sites where simply starting the Aux. Hyd. Pump (after the Aux. Hyd. Pump was started, of course) caused so much vibration of the thirty year-old liquid fuel flow divider that the Mark VIe thought liquid fuel was flowing when it was not (the unit was at zero speed).

The Customers (plural) SWORE the liquid fuel flow divider vibration was never a problem with the Mark IV (we never witnessed operation of the unit with the Mark IV before the Mark VIe was installed. The speed pick-up gap was checked and found to be 0.010 inches, per the Mark IV Control Specfication/Device Summary. At two of the three sites the liquid fuel flow divider had NEVER been refurbished or replaced in the thirty-plus years of operation since commissioning.

We ran twisted, shielded pairs outside the compartments, on the ground, to eliminate the possibility of electrical noise on the cables from being run in conduit with other signals. That didn't resolve the problem. I could take a large combination end wrench and simply tap on the liquid fuel flow divider and cause enough vibration to make the Mark VIe think liquid fuel was flowing (with both the original cabling and the temporary cabling).

That's when we adjusted the gap to be as high as 0.037 inches and the erroneous problem of liquid fuel flow when the unit wasn't running were solved. Subsequently, I spoke with Roper Industries and learned there is a LOT involved with using passive (inductive) speed pick-ups on the small diameter toothed wheels such are used on liquid fuel flow dividers, including the diameter of the speed pick-up face versus the width and diameter of the toothed wheel, as well as how the toothed wheel is made (whether is stamped or hogged or machined). GE uses the exact same speed pick-ups (3/4-inch, if I remember correctly) for BOTH turbine-generator shaft speed sensing AND liquid fuel flow divider speed sensing. This probably isn't the ideal size for the liquid fuel flow divider, given the width and diameter of the toothed wheel is so much smaller that the turbine-generator shaft's toothed wheel, which is also machined (the toothed wheels of the liquid fuel flow divider are usually stamped or hogged, which makes the tooth edges a little non-uniform which can also impact speed sensing).

All three sites I visited with the liquid fuel flow feedback issues refused to believe that the problem was with the liquid fuel flow divider condition because the Mark IV didn't have any problems, and all the problems started when the Mark VIe was installed to replace the Mark IV. So, we increased the speed pick-up gap until the problem went away and advised the Customer, in writing, that unless the flow divider was refurbished it may be necessary to further increase the gap in the future if the problem arises again before the liquid fuel flow divider was replaced or refurbished.

And that was that. I know, anecdotally, of many other sites which have had to increase the liquid fuel flow divider speed pick-up gap to solve the vibration-induced problem.

Because of the greatly increased sensitivity of the speed pick-up inputs to the Mark VIe, it is recommended to increase the gap of the liquid fuel flow divider speed pick-up <i><b>if the flow divider can't be refurbished to reduce backlash.</b></i> Roper also has kits specifically to help with erratic speed feedback (wider toothed wheels and more uniform teeth edges).

Please write back to let us know how you fare in resolving the problem.
 
V
Dear CSA,

thanks for reply, and many thanks for suggestion. I'll implement your suggestion in my next outage.

but if gap increased to 0.04 inch, its possibility make other risk?
example: miss fuel calculation during start up (lower speed), etc.
because from datasheet sensor (my site using Ai Tech 70085-1010-405)
if gap increased (0.01 to 0.04) output sensor in low speed will be ignored.

Regard,
Veri
 
Vhendrayawan,

I don't know how the sensor measurements in the graph are taken, or what the sensitivity of the measurement device ("tachometer") is, which would have a great effect on the measured output.

I don't recommend going to 0.050 inches, but I can tell you that the same sensors work pretty well in other applications at 0.050 inches, on the same Mark* (Speedtronic) turbine control systems.

The real problem here is the condition of the liquid fuel flow divider. Refurbish it to manufacturer's specifications, and it's pretty likely the entire problem will go away. Not entirely likely, because--again--the sensitivity of the Mark VI/VIe inputs have been greatly increased. So, you may still find it necessary to increase the gap to avoid any problems.

This has been done with great success on many units. It's called field engineering. If you try to start the unit and there's no flow registered on the HMI, then you've increased the gap too much.

I would strongly recommend a test, with the spark plugs disabled, of the change before trying to start and actually run the unit. Configure the Trend Recorder to monitor the liquid fuel flow divider speed pick-ups, and fuel flow, and LFBV position (if the LFBV has LVDTs). De-energize the circuit that feeds the ignitors (spark plugs), put the unit in FIRE mode, and initiate a START. The unit will go through a purge and firing sequence and fuel will actually be admitted to the turbine--but because the ignitors/spark plugs are disabled, it will not establish flame. (Of course, you want to do this on a cold turbine, not one which has just been shut down.) The Trend Recorder will capture fuel flow feedback, and, if there's none, but you see white vapours coming out of the exhaust stack, you will know that fuel has indeed flowed into the turbine and that the gap is too large for the pick-ups to register.

The turbine control system should annunciate an alarm, "FAILURE TO START" but that is to be expected, right? After all, the ignitors/spark plugs were disabled. The unit should continue to operate, cranking, and that will help to purge any liquid fuel out of the turbine and exhaust. I suggest that if there was fuel flow to the turbine, as evidenced by EITHER flow divider feedback, OR by the presence of unburned white vapours emanating from the exhaust stack, that you let the unit continue to crank for 5 or 10 minutes or so before clicking on STOP.

If you are satisfied that the fuel flow was correct, all you need to is leave the unit in FIRE and cranking for 5-10 minutes while you re-apply power to the ignitor/spark plug circuit, then simply select AUTO and the unit will open the liquid fuel stop valve and admit fuel to the turbine and the spark plugs will ignite the fuel and off the unit goes!

Again, the small gap was necessary for older analog circuits that were not very sensitive to low flow-rates (slowly spinning toothed wheels). But, because the Mark VI/VIe inputs are now much more sensitive than the Mark IV the small gap is not applicable. The commissioning person should have known that.

Please write back to let us know how you fare!
 
Top