Mark5 Speed Problem

S

Thread Starter

saha

Hello everybody,

In our installation we have a two shaft GE gas turbine (PGT10). The control system is Mark V simplex architecture (<R> core only).

The HP shaft has got three speed sensors (Magnetic pick up), called TNH1, TNH2 and TNH3 connected to <R> via the QTBA card.

During the start up sequence (from0% to 65% HP speed) all the three values TNH1, TNH2 and TNH3 in MK5 HMI are identical. However, above 65% HP speed, TNH1, TNH3 increase normally to full speed and TNH2 stays at 65 % and then drops below this value down to 39%

What we did?

1) We checked the wiring, changed the speed sensor the problem still persists.

2) We have used the function generator to simulate the HP shaft speed. We have connected the function generator to TNH1, TNH2 and TNH3 in parallel directly in the QTBA card (this eliminates the sensor and/or wiring problem), then we increased the frequency from zero all the way up, we noticed the same phenomena, the TNH2 can not follow TNH1 and TNH3 when the speed is above 65%.

3) We have changed the QTBA cad but the problem still the same.

Has someone encountered this problem or can someone help us to solve this problem?

Thanks
 
Hi saha,

When did this problem begin?

Or is this something you just noticed?

How is it adversely affecting the turbine operation?

What Diagnostic- and Process Alarms are being annunciated when the unit is above 65% TNH?

Are these same HP shaft speed pick-ups also connected to the <P> core at PTBA-1 & -2, -5 &-6, and -9 & -10?

What happens to the TNH_OS signals (in the Prevote Data Display) when TNH2 stops following TNH1 and TNH2?

What does TNH do when TNH2 stops following TNH1 and TNH2?

I don't know why three speed pick-ups would be required for a SIMPLEX application for sequencing in <R>; only for the <P> protective core.

It's not unheard of for units equipped with mechanical OS bolts (we don't know if yours is or not) to jumper the same speed pick-up inputs to both the primary processor(s) and the protective processors. I've seen some units with only two passive speed pickups and an OS bolt jumper the same two speed pickups to <R> and <P>, and then jumper one of the <P> inputs to the third <P> input (because there *MUST* be three speed pick-up inputs to <P>).

The QTBA card is just a terminal board, and Mark V terminal boards were designed to have few, if any, active components to hopefully reduce the likelihood of failure, requiring replacement.

Look in the Signal Flow Diagrams in Appendix D of the Application Manual, GEH-6195, for details of where the speed pick-up signal travels from the QTBA.

Have you tried using the frequency generator only on the TNH2 input and running the frequency up and down and observing the effects?

When you used the frequency generator did you disconnect the speed pick-up input wires from the QTBA (to reduce effects of loading on the freq generator output)?
 
Hello CSA,

Thanks for your reply,

Answering your questions:

The problem begun, after we did major inspection on the machine.

Since TNH is computed as the HIGH SELECT of (TNH1, TNH2, TNH3), the problem is not affecting the machine operation. The machine is running normally.

There is no diagnostic alarm; there is only the following process alarm:

L30TNH_DIFF XA-7745 HP speed signal differential

Yes, the HP shaft SPEED PICKUP also connected to the <P> core.

TNH_OS signal follows TNH1 and TNH3 all the time even when TNH2 stops following TNH1 and TNH3.

When TNH2 stops following TNH1 and TNH3, TNH continue following TNH1 and TNH3 because TNH is the HIGH SELECT (TNH1, TNH2, TNH3)

Yes, our machine is equipped with Mechanical Over Speed bolt.

We have tried using the frequency generator on all inputs TNH1, TNH2 and TNH3.

When we used the frequency generator, the field wiring were completely disconnected from the QTBA card.
 
Well, I would guess it was something that was done during the major inspection outage that caused the problem.

Have you traced the speed signals from the QTBA to the <R> core using the Signal Flow Diagrams in Appendix D of the Mark V Application Manual, GEH-6195? I would suggest the next step would be to replace the next printed circuit in the circuit, but I don't recall what that is at this time. It's shown in the Signal Flow Diagrams in the manual.
 
Hi everybody,

We have the same problem in our 4 PGT 10 (Simplex) since we had Mark V and after upgrading some cards to Mark VIe, Three speed pickup are connected to QTBA for Hp shaft and Three for LP shaft. while troubleshooting we noticed that the problem is only in the R core and not in P core (the same speed pickups are jumped to PTBA) and since the same problem is in 4 machines we suspected a design problem so we looked in the GEH-6195D ( appendix D, page D-38) and we found that in the TCQC card (the TCQC is an old Mark V card) the circuits of the pulse rate inputs are not the same; the 1st, 2nd, 5th and 6th are similar and are made for speed pickups but the 3rd and 4th are made for liquid fuel flow, that's why maybe we lose the signal when frequency becomes high especially with the HP shaft of the PGT 10 which is 10800 Hz in TNH 100%, tnh 2 is connected in the 3rd input and tnl2 in the 4th input actually TNL 2 is ok with TNL 100% = 7900 Hz, bu if we inject frequency in the 4th input we lose the signal also in high frequencies.

The machines are running normally but with the alarm of HP speed differential.
 

Attachments

Hi everybody,

We have the same problem in our 4 PGT 10 (Simplex) since we had Mark V and after upgrading some cards to Mark VIe, Three speed pickup are connected to QTBA for Hp shaft and Three for LP shaft. while troubleshooting we noticed that the problem is only in the R core and not in P core (the same speed pickups are jumped to PTBA) and since the same problem is in 4 machines we suspected a design problem so we looked in the GEH-6195D ( appendix D, page D-38) and we found that in the TCQC card (the TCQC is an old Mark V card) the circuits of the pulse rate inputs are not the same; the 1st, 2nd, 5th and 6th are similar and are made for speed pickups but the 3rd and 4th are made for liquid fuel flow, that's why maybe we lose the signal when frequency becomes high especially with the HP shaft of the PGT 10 which is 10800 Hz in TNH 100%, tnh 2 is connected in the 3rd input and tnl2 in the 4th input actually TNL 2 is ok with TNL 100% = 7900 Hz, bu if we inject frequency in the 4th input we lose the signal also in high frequencies.

The machines are running normally but with the alarm of HP speed differential.
Hi

Machine running normally with "alarm " displaying is not a "machine operating normally"

What you mean by "lose the signal " in high frequencies?


Is all the configuration in: TCEA_ Card Definition is correct?

ControlsGuy25.
I have a read on Mark 5 manual instructions
 
Hi ControlsGuy25,

I mean by running normally that the problem didn't trip the machine at least, in P core all is Ok so no problem with TCEA in Mark V or with PMVP with mark VIe after upgrade.
The problem is with R core only.
I mean by "lose the signal in high frequencies" when we use frequency generator we read the value we generate in the HMI, in low frequencies the value generated appears normally but when we increase the frequency we notice that the value in HMI drops and then becomes zero.
Even when the machine starts we lose signal only when the HP accelerates and its speed is near to TNH 100%. (I don't remember in which speed exactly we lose the tnh2 )
 
I'm very confused; so, please help me to understand. I do understand why the <P> core requires three (3) speed pick-up inputs (because of <X>, <Y> & <Z>). But, if the Mark V turbine control panels are SIMPLEX I don't really understand why it's necessary for three speed pick-ups for <R>. In a SIMPLEX application, only two (2) speed pick-up inputs are required--that's what the asterisks mean on the drawing amine_nacer supplied.

Is this "HP Speed Differential Alarm" a Process Alarm or a Diagnostic Alarm?

If the unit had a liquid fuel flow divider, and it was a SIMPLEX, the second flow divider input would be at the input with the asterisk

I can envision that GE might have made a mistake with the PMVP card/software and maybe even made a mistake including the application code block that compares speed input signals (making the app code look for thee primary (<R>) speed pick inputs to compare with the three <P> core inputs). And, if this is a Diagnostic Alarm, I could understand that there was a software oopsie here. But, one would think that with the length of time the Mark V-to-Mark VIe Life Extension product has been out this would be fixed.

It's still possible there is some configuration in ToolboxST for the <R> speed pick-up inputs that isn't configure properly for a SIMPLEX panel. But, I don't recall any need for three speed pick-up inputs for a SIMPLEX panel's control processor (<R>) in the Mark V era. Perhaps this is something the designers of the Mark V-to-Mark VIe felt was necessary (to have three primary speed pick-up inputs to <R>). BUT, if that were the case, then one would think there would have been some changes to the firmware of the PMVP. (As I recall, the TCQC has no configurable values; it's strictly a signal "pass-through" card, doing some analog-to-digital and digital-to-analog conversions for other cards (in the Mark V that would have been the TCQA, and TCQB, if so equipped).)

I do see where the filtering on the FF1 and FF2 speed pick-up inputs is different than on the shaft speed pick-up inputs (HP1 and -2, and LP1 and -2), and that could possibly cause a loss of signal at high frequencies (one of the beauties of working on digital control panels is that one did not have to be an electronics expert--and I'm certainly not one because I didn't need to be; just to understand very simple rudimentary electronics for mA inputs and T/C inputs and such). BUT, this still doesn't explain the need for three (3) speed pick-up inputs to a SIMPLEX <R> processor. Does the app code choose the MIN value of the three inputs, or does it choose the MED value, or the MAX value? Do all three signals have individual signal names? (If I recall correctly in the Mark V SIMPLEX product, the two HP shaft speed pick-up inputs were HIGH-selected by the firmware on the TCQA card, and the high-selected signal was TNH or TNH1 (there was no TNH2--or THN3).

When describing alarms, it's most helpful to describe them fully--as either Process- or Diagnostic Alarms. As I recall, there are both kinds of alarms for speed pick-up inputs, since these are so critical.

This would be a really good question for GE. I imagine since these are controls upgrades on PGT-10 machines, they were done by GE Oil & Gas, out of Florence, Italy.?.?.? I have been told they are pretty easy to contact and get help from (email or phone).

Please let us know about the clarifications requested above, and how you fare in resolving this issue.
 
Top