Armstrong Energy, a 625-MW, dual-fuel, simple-cycle peaking facility in Shelocta, Pa, is equipped with four 7FA.03 gas turbines. It is owned by LS Power and operated by IHI Power Services Corp. Brent Yatman is the plant manager.
Unit axial vibration monitoring
Challenge. Plant personnel desired a method for monitoring axial vibration levels during all phases of operation and for recording the results in the site’s plant information (PI) system. With known thrust-bearing issues in the fleet, as indicated on the 7F Users Group forum, the site was determined to closely monitor vibration to prevent abnormal wear and mechanical stress on unit components.
Armstrong relies on a Bently Nevada 3500 panel with two axial probes mounted on T1 and T2 turbine tilting-pad journal bearings and two radial probes mounted on both of the generator’s elliptical journal bearings. There are also two thrust position probes mounted on T1 and one keyphasor probe on the speed pickup wheel.
Solution. Plant personnel installed jumper wires from the thrust-position inputs to spare channels on the same Bently Nevada 40M monitor card. This enables simultaneous monitoring of axial vibration and rotor position without installing an additional probe or probes. Jumping the input signal did not affect the gap voltage; therefore, plant personnel felt confident this configuration would be successful.
The additional inputs were connected to radial vibration channels and associated with the keyphasor input to provide 1× and 2× vibration levels. By programming the inputs in this fashion, it became a simple procedure to route these signals into the Mark VI Speedtronic panels for display on the HMI.
Alert and alarm levels were arbitrarily set as identical to previously installed components because there was no guidance on these levels.
Results. Now that the plant had readings available on the HMI, it was time to verify them. This was accomplished by performing a rotor-position test (bump test). The rotor was pushed against the inactive face on the thrust bearing using hydraulic rams and jacks and a dial indicator was mounted on T1 bearing housing and set to zero.
The rotor was then forced against the active face and the dial-indicator reading was compared to thrust-position reading both at the Bently panel and on the HMI. The bump test was performed three times with repeatable results.
Additionally, each time the rotor was forced against a thrust-bearing element, hydraulic pressure was raised to 5000 psig and then lowered to 2500 psig in accordance with the OEM’s methodology. With three sets of repeatable results identical to the readout of the thrust probe at the panel and on the HMI, the site was assured that there was no issue with their set up.
The units then were operated, and vibration and position levels observed. Again, there were no issues with readings when compared against prior levels.
Plant personnel, guided by Peter Margliotti, next added the new point to the PI system and can now view both real-time axial vibrations and trend levels over time to identify possible issues before catastrophic failure.