The second annual meeting of the Generator Users Group (GUG), today organized under the Power Users Group umbrella, was conducted in parallel with the annual meetings of the Combined Cycle Users Group (CCUG) and Steam Turbine Users Group (STUG) in San Antonio, Aug 22-25, 2016.
Recall that the GUG was launched in late 2015—a collaborative industry effort initiated by IEEE Fellow Clyde V Maughan, president, Maughan Generator Consultants, with major support from NV Energy, which hosted the first meeting at its Beltway Complex and Conference, and Duke Energy. The steering committee formed to assure success, chaired by Duke’s Kent Smith, continues to guide the group.
CCJ ONsite’s coverage of GUG’s second annual meeting begins in this issue and concludes in next week’s publication. Summaries of user presentations, compiled by Maughan, follow after this introduction; presentations by consultants and suppliers are aggregated separately.
Two points to keep in mind as you read through this material:
The GUG’s mission is to provide a forum for owner/operators of electric generators at coal-fired, nuclear, and combined- and simple-cycle gas-turbine plants to share experiences, best practices, and lessons learned on design, installation, O&M, and uprate/upgrade. Expected outcomes are improved safety, maintainability, availability/reliability, and efficiency, as well as the transfer of industry knowledge from experienced engineers to those wanting to gain hands-on know-how.
Although CCJ ONsite focuses on serving gas turbine users, the editors suggest you not bypass the few summaries of nuclear and coal-fired plant experiences incorporated by Maughan because they offer valuable lessons learned to all generator users.
Several of the user presentations summarized here were made by members of the steering committee—Chairman Smith, John Demcko, PE, of Arizona Public Service Co, Dave Fischli, PE, of Duke Energy, and Leopoldo Duque Balderas of COMEGO (Mexico)—illustrating the depth of knowledge of the GUG leadership and the value of participation in this annual meeting. Follow the website for details on the 2017 meeting at the end of August as they are made available; first posting is expected in April.
User presentations made at the GUG’s second annual meeting and summarized in this issue are listed below; links provided enable quick access to the topics of greatest interest to you.
Users wanting to dig deeper into each of these areas can access the presentations of interest on the Power Users website. You must be registered to participate in the forum, a relatively simple process if you’re not already signed up.
As built, McGuire Nuclear Station was equipped with two 1450-MVA, 4-pole generators. featuring water-cooled stators. Within five years of COD in December 1981, both units had been de-rated by about 140 MVA because of operating problems that included end-iron overheating.
There have been numerous maintenance events over the years, including these:
Stator rewedged in 1998.
Field rewound in 2007 because of shorted turns and thermal sensitivity.
Increased oxide fouling in the stator bars.
Three chemical cleanings conducted in a 10-year period.
Cracking of aluminum shields in the main lead box.
Three elevated EL CID indications of about 170 mA each that were slowly increasing between inspections.
A life-extension study was performed in 2010 and it was decided to reuse the generator fields on both units because they had been rewound recently. Relative to the two stators, engineers decided to do the following:
Purchase a replacement stator with the winding designed for 1550 MVA.
Convert parallel rings from hydrogen- to water-cooled.
Upgrade the stator cooling-water (SCW) system to include alkalizer injection for controling pH to address oxide fouling problems.
Replace the HV bushings with 1550-MVA capability.
After installing the new stator on one unit, the old stator was shipped to the OEM’s factory and rewound with upgrades for use in the other generator.
Site acceptance tests of the new stator included ultrasonic measurement of flow in stator cooling-water hoses, EL CID testing of core, and DC hipot. During the site acceptance test the EL CID test failed, with readings as high as 169 mA detected. A loop test conducted to validate EL CID readings was terminated within three minutes because it failed to meet test temperature criteria. Core repairs on the new generator were required.
Numerous problems were identified with the hose water flow on the SCW system, including near-stagnant flow in six parallel rings. The latter is particularly important in that immediate gross overheating of the rings likely would be accompanied by complete winding failure. Corrections of the hose problems identified were complicated and expensive in dollars and outage time.
Issues in hydrogen-cooled machines
The serious problems discussed in this presentation involved a nine-year-old hydrogen-cooled generator rated 193 MVA, 13.8 kV. Its bar design is shown in Fig 1. During a routine maintenance outage, engineers found all the tube-to-copper resistance readings satisfactory, generally above 1000 ohms, with the exception of the bottom tube in the top coil, slot 22, which read 81.8 ohms. The ends of this tube were cleaned and dried and mica inserted, but the reading did not improve.
OEM guidelines were the following:
Less than 500 ohms, investigate.
Less than 100 ohms, change the affected bar.
Further investigations—including the removal of series blocking and groundwall insulation—did not reveal the cause of low resistance or improve the low value. The bar was removed from the winding and sent to a non-OEM laboratory for investigation. No additional understanding was found by the initial investigations. Electrical tests revealed the short likely was located from 7% to 10% of the way from the exciter end. Stripping of the groundwall revealed severe localized overheating locations.
After disconnecting the copper grounding strip shown in Fig 2, the tube-to-copper resistance value increased to more than 50 Mohms. This is a complex bar cross section, not well understood. It is clear that the low resistance values were caused by gross overheating, burning (which resulted from circulating currents caused by a short between the copper strip and a strand), and the resulting carbonization of insulation components. However, the root cause of the short remains unknown.
Leopoldo Duke Balderas has many years of powerplant O&M and engineering experience
Connection-ring vibration monitoring system
Intermountain Power Project (IPP) went online in 1986/1987 with two hydrogen-cooled generators rated 991 MVA, 26 kV. By the 1990s, leaks had started in the generator windings and both units usually failed leak tests after 1994. A global strand header repair was performed on Unit 2 in 1996 and on Unit 1 in 1997. Leaks continued even after the epoxy repair, with the leaks usually found at joints in the connection rings (Fig 1).
Plant management decided to rewind both generators, including replacement of the connection rings and use of the OEM’s new vertical strand header braze procedure. Unit 2 was rewound in fall 2010, Unit 1 in spring 2011. In December 2011, Unit 1 suffered a massive winding failure, attributed to a failed bolted joint in the neutral connections in the dome. (A similar joint that had not failed is shown in Fig 4.)
Immediately after Unit 1 failed, Unit 2 was taken offline. The flexible connections were examined and arc indications found after only six months of service (Figs 2, 3). The replacement connection rings included additional bolts on the lower tang (Fig 4). But the root cause of Unit-2 arc indications and of Unit 1 failure was use of improper bolting techniques for the stainless steel bolts.
IPP personnel remained concerned with whether the problem had been fixed—in particular because there was no advance warning for the failure. The OEM recommended installation of fiberoptic vibration probes and this was done in January 2012; vibration started to increase in January 2013. Analysis and interpretation of data gathered have not provided definitive results— partially because of little industry experience on connection-ring terminals; plus, weak technical support.
A second vibration probe system has been installed on Unit 1 and a second system will be installed on Unit 2. The trends of the second systems will be compared with the output from the present probes.
Rotor field ground indications
Grounds in two different excitation systems were discussed by John Demcko, who has an exceptionally broad and deep background in power generation equipment: those with brushless excitation systems and those with collector rings and brushes. The two case studies presented illustrated the challenging complexity of excitation-system diagnostics and maintenance. When we speak of a field ground, the speaker said, what we really mean is an excitation system ground.
The first ground considered was on a brushless system. In spring 2006, a modern brushless field ground detection system, Accumetrics Inc’s earth fault resistance monitor (EFREM), was installed on Combined Cycle Unit 4 at the company’s West Phoenix Generating Station (Fig 1). It replaced an OEM system that never worked properly.
With the Accumetrics system, the field resistance to ground is monitored offline, as well as online, and is telemetered to the plant DCS. Six months after installation, the field ground alarm came in solid while the unit was offline. Since the alarm occurred during heavy rain, engineers decided to dry the system before drawing any conclusions. In 3.5 hours, resistance increased from 12 kohms to 20 Mohms. Better waterproofing and caulking of possible water ingress points thus far has been effective in mitigating the issue.
The second ground was on a collector/brush system. During a normal startup, generator voltage did not build up to nominal rated value. Investigation provided some startling information. First, the field itself had a ground. But there was a station-battery ground as well. The combination of the two grounds allowed part of the field excitation current to bypass a portion of the field turns.
While the field was being rewound, resolution of the station-battery ground was pursued. It was found at a taped connection joint left lying on the steel-deck floor (Fig 2). Over many years, the taped insulation had worn away, resulting in a hard ground.
Experience with Alstom air-cooled generators
System-wide, Southern Company has seven Alstom air-cooled generators rated 313 MVA, 21 kV. There have been significant maintenance and operational issues on these units, including the following:
Stator phase-connection conductor fatigue.
Stator endwinding voltage grading deterioration.
Stator spring-plate fatigue.
Stator side-filler migration.
Stator frame plate weld failure.
Field retaining-ring insulation deformation.
Field slot-liner cracking risk.
Field-winding pole-connector fatigue.
Only the last was discussed in this presentation.
Initial awareness of the problem came in April 2012 coincidental with the follow-up inspection of a phase-bar blocking modification. Prior fleet-wide inspections offered no clear evidence of pole-connector fatigue. But review of previous inspection photos with a focus on the probable crack-initiation areas showed signs of possible initiation (upset metal). Follow-up inspections over the last four years have shown all previous “possible initiation” sites to have definitive cracks with propagation in progress. Photos shared to illustrate the problem included those here labeled Figs 1-6.
Inspections in April 2013, April 2014, and July 2014 on the unit with most advanced fatigue condition revealed continuing crack propagation on both pole (redundant) connectors. Repair was implemented in December 2014.
Engineers concluded there is a definitive correlation to start/stop cycles. Inspection data show the rate of crack propagation to have some consistency for a given pole connector but it clearly varies from one connector to another. It is expected that pole-connector replacement eventually will be required on all seven generators.
Generator end-plate indications
The Altamira II combined-cycle plant was notified in October 2014 of some findings in the generator-rotor end plates that had occurred at other Mexican plants with similar air-cooled generators. Two months later, inspections at Altamira II revealed several cracks on the rotor end plates for its two gas turbines and steam turbine (Figs 1-3).
The OEM strongly recommended not running the units in this condition because of the risk of catastrophic failure. Before restart, the OEM recommended replacing the rotor end plates at both ends of all three fields. Removal of the retaining rings would be required to do this, and based on the OEM’s experience, destructive removal was likely. To avoid destructive removal of the retaining rings (spares availability was a major concern), the end plates were removed destructively, without touching the retaining ring.
Root cause of this fleet problem was stress corrosion cracking. Recommended preventive action included replacement end plates made of an improved material, application of anti-corrosive paint on the end plates, and the elimination of tapped holes for the rotor baffle assembly around the inter-pole center (the lower set of holes, most easily seen in the center photo).
Oil-intrusion events are a fact of life at Duke Energy, a large utility with hundreds of generators in service. To learn more about how others in the electric power industry deal with oil intrusion, GUG Chairman Smith and his Duke colleagues conducted an informal survey. More than three-quarters of those surveyed said they have oil-intrusion concerns.
Next question: “What level of oil intrusion do you consider a concern?” Responses varied:
Less than 10 ml weekly concerned no one.
Between 10 and 20 ml weekly concerned 26%.
From 20 ml weekly to 10 ml daily was of concern to 37% of those surveyed.
From 10 ml to 100 ml daily got 21% of the respondents concerned.
The remaining 16% were not concerned until intrusion exceeded 100 ml daily.
Smith next summarized the following recent oil-intrusion events in the Duke fleet:
Operator error was blamed for the pumping of more than 3000 gal of oil into the generator bushing box on a large water-/hydrogen-cooled machine.
During startup, the oil detraining tank on a hydrogen-cooled unit was overfilling and the bypass valve had to be manipulated to control tank oil level. When the unit was being removed from service for repair, tank level increased and oil was pushed into the machine.
A large hydrogen-cooled unit consistently required oil clean-up from the machine’s belly but it had no liquid detector alarms. Although there’s no known impact to date, stator re-wedging will be needed.
Oil intrusion investigation on a large water-/hydrogen-cooled generator is ongoing. Intrusion was noted during the last rewind; the end-bell mating surfaces were not as flat as expected.
Cleaning was required in each of the four cases cited above. It ranged from minor to extensive and given the complicated internal complexities of the generator, cleaning may never be finished in some cases. Corrective actions, sometimes ineffective or incomplete, can include the following:
Enlarge flex-seal grooves and add additional pumping locations.
Re-pump flex seals.
Machine end bells to achieve better mating surfaces.
Replace TiteSeal™ compound with Flex Seal®.
Install drain holes in bushings to drain oil from the cooling path of the bushings.
Replace seal-ring springs.
Correct piping deficiencies.
Stator ground in a GE 390H generator
The subject generator is connected to the steam turbine serving a large F-class combined cycle. Plant began commercial operation in 2004 and ran reliably until 2014, accumulating nearly 45,000 operating hours and 1000 starts. During a startup in fall 2014, the unit tripped on volts/hertz and ground relays.
Visual inspection of the external components revealed no problems, but A phase was grounded. This is a three-circuit winding, and two of the three circuits were grounded. The OEM recommended a full rewind, and each bar was checked using a Megger™ before being removed. The bottom bar in slot 1 was found grounded, with its insulation heavily cut by outside space block (OSSB) migration inward.
Many other bars showed damage from OSSB movement (left and center photos). The core damage at slot 1 is shown in the right-hand photo. Burning is much greater than would be expected from the >5 amps of a single ground and probably resulted from core lamination shorting.
The repairs performed included loosening belly bands, rounding the corner of the compression ring (flange), replacing OSSBs one at a time, adding a punching with master bond coating, reinstalling the compression ring, compressing the core to a higher level (2000 ft-lb increased to 2500), retightening of the belly bands, and rewinding the generator with all-new bars.