We are running Frame 6, Mark V GT , only with gas. Gas sensors 45HT 5 & 6 of turbine compartment ventilation duct has detected gas to high level (trip value). On the explosive level checking by lab at various location & reported ( NIL) & methane( 2ppm) we decided to run the machine inhibiting the sensors .We observed air leak from the outer cross fire tubes. It was also analyzed for explosive. Report was negative. Still we took shut down and replaced a gasket and sensors, cables checked and was intact. Restarted and everything was OK for 4 hours after start up. Again the same (45HT 5 & 6) sensors started indicating gas to trip values. Now 6 more outer cross fire tubes flange leak is observed. Turbine compartment temperature is maintained by opening the door. Thermal imaging shows 15 deg C rise in temp of outer cross fire tubes indicating a flow.

We like to hear any similar experience and action plan. We are running with inhibited sensers now.

 

Have you been doing any maintenance on the machine recently? I ask because we had similar experience after a bad overhaul and did have gas leaks at the can to fuel nozzle flanges and at the gas pigtail flanges which had the wrong gaskets fitted. but we could detect this using a portable gas analyzer. Are you sure your gas portable gas analyzer is OK? Sounds very dangerous to me if you really have gas leaks and combustion gas leaks and are running without gas detection!! We also had X-fire tube leaks like you, basically a bad overhaul!! It would be odd for both gas detectors to have failed this way, but anything is possible. Do you have any spares or can you calibration check your detector heads.

 

Recent no major job is done in last 1 1/2 years. Machine is due for M.I in sept 2010.

We have taken the analysis by portable gas detector and by Sample collected and analysed by Gas Chromatograph which can analyze in ppm. There is no reason to doubt the GC analysis and portable analyser. We have done more than 50 sample.

3rd gas sensor is showing only 0.3 and is not increasing, but varying. Can you share your action plans during the similar occasion.

 

As Bob Johnston suggests, when did the problem start? After a maintenance outage?

To Bob Johnston's reply I would add that I would be surprised if there was natural gas leaking from the cross-fire tubes. And, if there is any flow into or out of the cross-fire tubes, that in itself is not a good condition to be operating the machine.

I'm not saying that natural gas can't leak from the cross-fire tubes, I'm just thinking it wouldn't be very likely, but that's just a scientific guess and represents my opinion and not fact. I would more suspect that if hazardous gases (natural gas) are(is) present in the compartment that it would more likely be coming from leaking flanges in the fuel gas manifold or nozzle connections. But this is just my opinion.

I would also add that I have come to learn that hazardous gas sensors are themselves very sensitive. They are sensitive to dirt, oil (even oil from ones hands!), and oil vapors. If the covers or the beads beneath the covers get dirty or contaminated they can erroneously indicate the presence of hazardous gases.

If the sensors aren't covered during maintenance activities, and aren't properly maintained (cleaned, replaced when dirty or contaminated), they can be very problematic. I have seen vapors from L.O. leaks (excessive L.O. leaks that weren't properly dealt with in a timely fashion that caused pools of L.O. in the bottom of the turbine compartment) cause erroneous alarms, and even contaminate the sensor covers/beads. The temperatures in the turbine compartment and the circulation of air can cause the vapors to be "deposited" on all surfaces, including the sensors.

You say you've had "lab" analysis of the compartment environment and have confidence there is no hazardous gas in the compartment. If that's indeed true and you are firmly confident there is no explosive atmosphere in the turbine compartment, then you have no gas leaking from the cross-fire tubes. And, most likely the problem is with the sensors and/or ambient conditions in the compartment, or with the "care" taken during maintenance outages when working near the sensors.

 

Thank you, for a technical diagnosis and a prompt reply.
We had adopted many methodologies to identify the leak or any after combustion. We had used the explosive meter, we had analyzed the GT turbine compartment air at different location with gas chromatograph, we had used thermal imaging and also pyrometer to identify varying range of gas leak and temperature rise. In none of the methods we could find even traces of gas or oxidation tendency.

Finally we are coming to a conclusion that the gas detection by sensors may be due to the high temperature of the hot air from cross fire tube leak along with the dust accumulation , (thanks for your views)
Based on these conclusion we are deciding to run the machine till M.I which is due in September 2010. Any feed backs are still welcome