Control.com - Nerds in Control

We are faced with a problem with a GE Gas turbine frame 3; control system is TMR mark 5. The problem is that we have abnormal fluctuations in SRV, GCV, intervalve pressure and the exhaust temperature. This is occurred when the machine was running and the FSR selection switch was in AUTO mode. However, when the FSR signal was switched to MANUAL mode while the machine was running, all the parameters become stable.

This problem occurred only after we have isolated the power to the Mark 5 panel (to do a maintenance job on other system), before the machine was running normally with stable parameters.

Has anybody had experience this problem?

THANKS

I guess by FSR Manual, you mean the FSR Gag control. Is this set at 100%?

As this is twin-shaft, I guess a mechanical drive, you need to be a bit careful. Instability like this can be driven from the load end. If you can stabilize the load side first, it would be easier to trouble-shoot. Can you put it in manual and even hold the 2nd. Stg. nozzle control in manual? Remember, the LP turbine drives the fuel, so instability on the load side will cause exactly the symptoms you have. If you then still get instability, it's probably coming from the fuel control side. Unfortunately it can be a bit difficult to know whether it is the SRV or GCV, instability in either produces similiar symptoms. The easy, but expensive fix, is to change both servo valves. If you have time and the shut-down window, change them one at a time, recal, and see if the problem goes away.

Good luck and report back.

Bob, it could be the 2nd stage nozzle, that can aslo be triggering the fluctuations in the fuel system. As per my expierience, at one or two installations of twin shaft machines, I had seen malfunction of 2nd stage nozzle causing fluctuations in gas turbine parameters. So I feel 2nd stage nozzle calibrations and feedback, physical & LVDT feedback need also to be looked at both at standstill and when turbine is running.

Thanks for your replay, The problem is solved.
Since the turbine was running on speed control (FSR=FSRN), I suspected the problem is on the speed control loop. After checking the parameters of the speed control loop I found two constant parameters (FSKNTC & TNKRNR) are not the same as the constant mentioned in the specifications. I have no idea why they have been changed.

The machine was running correctly without fluctuations before power isolation. This is because the correct constants were in the RAM of the RST Processors, and certainly were different from the values on the EPROM of the Processors. When we isolated the power, the correct RAM values were lost. And when the power was recovered to the Mark 5 panel (Processors Rest), the EPROM values (which are incorrect values) are transferred to the RAM. Hence, the machine is running with incorrect constant values which cause fluctuations on the turbine parameters.

Salutations

In a 2-shaft turbine it will be a bit time-consuming to correct a problem like this. We can provide many "guesses". In Mark V HMI (or <I>), there are some utilities that can collect turbine data at the frame rate of 32 Hz. Some of them are VIEW2, VIEW2T, VEIWPV etc. The maintenance manual gives detailed information to run these utilities. It is fairly straight forward. How quickly you can locate the fault depends on your knowledge of the logic and the selection of signals to be viewed.

As an example, you can start monitoring the currents through the three coils of the servo for the GCV. Or you can start from the different FSR signals - FSRSU, FSRN, FSRT, FSRMAN, FSRSD, etc.

To eliminate SRV from this scenario, you may want to collect the signals associated with SRV control and feedback. SInce everything is stable in manual mode, I would concentrate on FSRN and the associated logics.

In essence, your success depennds on the selection of signals and your patience.

From the descrition of the symptoms, it appears to me that the LP turbine speed pickups are the possible cause. But don't take my word blindly, follow the systematic data collection procedures.

A final note:

Better to unload the turbine, during the whole process. We don't want to run the turbine at the maximum load in a situation like this, right?

Good luck!!!