H
Tripping:
The GE MS5001 Turbine tripped on high seismic vibrations trip signal generated by Bentley Nevada 3500 (velometers). Bentley Nevada 330500 velometers are installed on the north end of the generator and on the south end (inside Load Gear compartment).
OBSERVATIONS:
Upon troubleshooting, following observations were made:
- Both Seismic probes had spiked at the same time
- Even when GT was started on cranking, the probes showed spikes reaching alarm and trip limits (trip was bypassed on seismic and activated on proximity probes)
- All observed spikes were of the same amplitude and appeared simultaneously on both probes
- Bentley Nevada card self diagnostics were run and turned out fine. Nevertheless, it was replaced with a new card for the troubleshooting duration
- When Adre was connected, it showed spikes on all 4 proximity probes of the generator at the same time and the amplitude was almost twice that being displayed on Mark-VI. This i believe can be justified by the fact that Adre has a different ground and consequent reference point for reading the buffered signal from Bentley Nevada while Mark-VI and BN shared the same instrument ground for reference. It is to be noted that where random spikes were observed on other probes too, they weren't simultaneous (down to the second time value) and of the same amplitude as was the case with generator proximity probes
- The generator proximity probes and generator seismic probes share the same JB
- JB was grounded with the generator body
- Seismic probes shielding cables were grounded ultimately on the instrument ground bar inside Mark-VI cabinet
- Proximity probes shielding cable was connected at the Mark-VI end on the ground bar but open (yet fully taped so they're not exposed and completely insulated) at the JB end
Possible Causes:
From doubting an issue with Bentley card the complete focus was shifted to grounding issue when the GT was started on cranking and it was observed that both seismic probes spiked at the same time and with the same amplitude.
Seismic probes were temporarily completely disconnected from the system and their wires removed from all ends. All 4 proximity probes showed spikes occasionally on cranking mode (visible only on Adre and not Mark-VI). When their shielding cable was opened from the instrument grounding bar, the spikes disappeared from that specific proximity probe but continued on other probes.
To make sure it wasn't static charging, the grounding brushes were observed and were found to be in perfectly healthy condition. Is there any chance an excessive Gaussian field may still exist, and if so, is there any way to test it without major parts removal? (Otherwise I'll have it scheduled for the next outage).
The spikes aren't real because the probes located on generator and gearbox shouldn't be spiking at the same time. Moreover, the 1X value on Adre doesn't change. Also, shielding cable isn't being grounded with the JB or body at any place.
I still believe it s an grounding issue and something is causing looping on the instrumentation grounding bar and I'm working on checking the shielding cables of all 24VDC signals in the field to see if any is being shorted with the JB body and causing this.
Query:
What other possible causes can there be? The turbines shut down for a very limited time so a prompt response will be greatly appreciated. Can there be any other contributing factors?
The GE MS5001 Turbine tripped on high seismic vibrations trip signal generated by Bentley Nevada 3500 (velometers). Bentley Nevada 330500 velometers are installed on the north end of the generator and on the south end (inside Load Gear compartment).
OBSERVATIONS:
Upon troubleshooting, following observations were made:
- Both Seismic probes had spiked at the same time
- Even when GT was started on cranking, the probes showed spikes reaching alarm and trip limits (trip was bypassed on seismic and activated on proximity probes)
- All observed spikes were of the same amplitude and appeared simultaneously on both probes
- Bentley Nevada card self diagnostics were run and turned out fine. Nevertheless, it was replaced with a new card for the troubleshooting duration
- When Adre was connected, it showed spikes on all 4 proximity probes of the generator at the same time and the amplitude was almost twice that being displayed on Mark-VI. This i believe can be justified by the fact that Adre has a different ground and consequent reference point for reading the buffered signal from Bentley Nevada while Mark-VI and BN shared the same instrument ground for reference. It is to be noted that where random spikes were observed on other probes too, they weren't simultaneous (down to the second time value) and of the same amplitude as was the case with generator proximity probes
- The generator proximity probes and generator seismic probes share the same JB
- JB was grounded with the generator body
- Seismic probes shielding cables were grounded ultimately on the instrument ground bar inside Mark-VI cabinet
- Proximity probes shielding cable was connected at the Mark-VI end on the ground bar but open (yet fully taped so they're not exposed and completely insulated) at the JB end
Possible Causes:
From doubting an issue with Bentley card the complete focus was shifted to grounding issue when the GT was started on cranking and it was observed that both seismic probes spiked at the same time and with the same amplitude.
Seismic probes were temporarily completely disconnected from the system and their wires removed from all ends. All 4 proximity probes showed spikes occasionally on cranking mode (visible only on Adre and not Mark-VI). When their shielding cable was opened from the instrument grounding bar, the spikes disappeared from that specific proximity probe but continued on other probes.
To make sure it wasn't static charging, the grounding brushes were observed and were found to be in perfectly healthy condition. Is there any chance an excessive Gaussian field may still exist, and if so, is there any way to test it without major parts removal? (Otherwise I'll have it scheduled for the next outage).
The spikes aren't real because the probes located on generator and gearbox shouldn't be spiking at the same time. Moreover, the 1X value on Adre doesn't change. Also, shielding cable isn't being grounded with the JB or body at any place.
I still believe it s an grounding issue and something is causing looping on the instrumentation grounding bar and I'm working on checking the shielding cables of all 24VDC signals in the field to see if any is being shorted with the JB body and causing this.
Query:
What other possible causes can there be? The turbines shut down for a very limited time so a prompt response will be greatly appreciated. Can there be any other contributing factors?
