“Cedar Bayou Generating Station, located near Baytown, TX, consists of three gas and fuel oil-fired super-critical steam generators producing a combined output of 2,340 megawatts.”
Reliant Energy is an international energy services and delivery company with approximately $48 billion in annual revenue and total assets exceeding $31 billion, based in Houston, TX. The company has nearly 25,000 megawatts of power generation capacity in operation in the U.S. and is one of only three companies to rank among both the five largest power marketers and the five largest natural gas marketers in North America.
Cedar Bayou Generating Station, located near Baytown, TX, consists of three gas and fuel oil-fired super-critical steam generators producing a combined output of 2,340 megawatts. Predictive maintenance (PdM) technologies used at Cedar Bayou are oil analysis, vibration, thermography and ultrasonic inspection. The oil analysis program was initiated in 1997 using operations and plant laboratory personnel to pull samples on targeted critical equipment. Samples were taken at random as plant workload permitted and then sent to an outside lab for testing and analysis. In 1998, a single craft person was selected to oversee the program and to ensure samples where taken in a timely manner due to sampling inconsistencies in the past.
Reliant demonstrated its belief in the value of the oil analysis program in 1999 with the formation of a central analysis group and the installation of onsite testing facilities. The onsite facility focused mainly on basic analysis such as particle counts, moisture content by Karl Fischer titration and viscosity, with the equipment screening capability added by creating and examining filter patches onsite. Basic equipment installed at each plant included the following:
Training was required for the specialist assigned to the oil analysis program. The cost of basic equipment and personnel training initial startup cost for onsite testing at each plant is estimated to have been about $25,000 to $30,000 dollars. Other testing required was conducted at the company’s central laboratory. Sampling was done on a time-interval basis according to equipment criticality within the overall production system. Equipment was then grouped into routes that were scheduled to ensure timely completion.
On January 9, 2001 routine samples were taken as part of a normal start-up route for unit #2 main turbine. Once the route was complete, the samples were taken to the onsite analysis laboratory for basic analysis, consisting of a pore blockage particle count, absolute viscosity and moisture content. The particle count test on the #2 main turbine bearing oil return line sample showed a dramatic increase in particle count, having jumped from a normal cleanliness level of 14/11 to 22/19, and was well over the critical alarm limits.
A second sample immediately taken for test verification confirmed the earlier results. An exception test for iron, triggered by the high particle count, was performed using a ferrous contamination alert probe. The ferrous test showed that of the particles counted in the 10-micron range, 99.9 percent of them were ferrous particles. This result prompted an onsite filter patch examination. It was manually examined using the laboratory microscope at 10X and 100X magnification. The examination revealed a very large concentration of metal particles providing additional proof that there was a major problem occurring with that unit. The large concentration of iron particles and the absence of nonferrous metals such as tin and lead in the above test (primary metals in a Babbitt bearing) prompted us to immediately rule out a bearing problem. The absence of increased bearing vibration and temperature was not consistent with a bearing journal failure, yet the amount of wear particles and their relatively large size indicated that some part of the turbine was failing very rapidly.
Because the turbine rotor spins at 3,600 rpm, it was critical that the problem be found quickly and corrected prior to a catastrophic failure. Turbine vibration and temperature monitors did not indicate a problem. Plant management was notified of the problem and the recommendation was made to immediately begin filtering the reservoir to limit further damage to the turbine bearings by the iron circulating within the oil system.
Reservoir filtration was started using a 45-gpm vacuum dehydration filter skid with a six-micron polishing filter. Plant operations began an outside inspection of the steam turbine and generator set. A third sample was pulled and taken to the company’s material analysis group for analysis by Scanning Electron Microscope and Energy Dispersive X-Ray (SEM/EDX) to attempt to identify the source of the metal particles. Plant operations’ inspection discovered a bumping sound emanating from the turbine turning gear area, an indication that the turning gear may not have fully locked into the disengaged position. The turning gear uses an electric motor to slowly rotate the turbine rotor when offline to prevent the turbine rotor blades from sagging against the turbine case and seals. Plant operations suggested that this could be the source of the wear particles, but they could not offer a definitive explanation.
The turning gear was manually operated and fully locked into the disengaged position, which stopped the bumping sound. At the materials analysis lab, a patch was made using the oil from the third sample, and was placed in the electron microscope for analysis. Results of the SEM/EDX test showed the metal particles to be 98.8 percent iron, 0.69 percent chromium and 0.49 percent manganese (Figure 1).
Microscopic examination of the wear particles at 350X showed the wear mechanism to be severe sliding contact wear, which was indicated by the size of the particles and the striations seen on the wear particles (Figure 2).
A picture taken of the patch surface shows a very high concentration of the particles being >20-microns in size, which is also indicative of severe sliding contact wear (Figure 3).
The materials analysis group forwarded the test results to Dwayne Jenkins, the central PdM oil specialist, who consulted the steam turbine central maintenance engineers and the plant oil specialist. The group looked at the following information:
They concluded that the particles were most likely coming from the turbine turning gear pinion striking against the shaft bull gear because this was the only area on the turbine that could produce the amount and type of wear particles being generated. Plant operations agreed with the findings, believing that the turbine turning gear had not locked fully in the disengaged position when the unit was brought up. This left the pinion free to bump against the bull gear. Operations had corrected the problem and the two groups agreed that the turning gear should be inspected the next time the unit was taken out of service.
Filtration was continued on the turbine oil reservoir until the particle counts returned back within normal operating parameters at a cleanliness level of ISO 14/11. The filter skid was then shutdown and the unit was allowed to run for 24 hours without additional filtration. The reservoir was then sampled and tested for particle counts and moisture to ensure that the problem had been corrected and no more wear particles had been generated. Further samples were taken and sent to the company’s central lab to be tested for RPVOT (Rotating Pressure Vessel Oxidation Test) and other fluid properties to ensure that degradation of the lubricating oil, due to the large amount of wear particles in the system, had not occurred.
The unit was taken out of service a few weeks later, and inspection of the turbine turning gear showed that the pinion gear had been damaged beyond repair but was still usable until a replacement gear could arrive. Had no routine oil-sampling program been in place to identify the problem, the turning gear pinion would no doubt have continued degrading until it was destroyed. Not having the turning gear when the unit came down would have allowed the hot turbine rotor to sag, potentially causing further problems.
The iron particles in the oil would have continued circulating through the turbine bearings creating wear and generated more particles. This condition would have progressed until enough damage was done to the turbine bearings to cause the vibration readings to begin to rise, at that point the damage would have been much more severe.
It is easy to see that not catching a problem of this type could lead to millions of dollars worth of damage and downtime. General estimates show the numbers to be between $500,000 to $1,000,000 saved in downtime, material and man-hours had the problem progressed until vibration problems brought it to the plant’s attention. Onsite oil analysis provided quick recognition of an existing problem, allowing fast response actions to be taken to limit further damage to the equipment. Wear debris analysis coupled with the employees’ knowledge of the equipment’s metallurgy and design, along with plant operations’ knowledge of equipment systems, provided quick recognition of the problem component. This allowed the problem to be corrected, preventing further damage to the turbine systems. Catching a problem like this is a good example of how oil analysis is used as a tool to monitor machine health, but like any other tool, proper use is critical to achieve the desired effect.
The case presented above required the use of all the steps and avenues of information that are part of a properly working oil analysis program. Without them, the faulty turning gear would have been missed and further damage would have likely occurred. With them, the faulty turning gear was properly identified and corrected, and plant management was shown a prime example of the value of an effective oil analysis program.