OEMs tell users their gas turbines have critical parts with finite lifetimes and that replacement of these parts may be necessary to assure reliability and safety moving forward. Rotors are a primary target of this initiative. Today, the only way to determine if an ageing rotor is in sufficiently good condition to continue operating is to disassemble it and to nondestructively examine individual wheels, bolts, etc.
For readers unfamiliar with the Frame 6, the fleet was launched in 1978 by General Electric Co with the 31-MW MS6001A. To date, more than 1000 units have been shipped worldwide by the OEM and its manufacturing associates for either 50- or 60-Hz power generation service. Units serve in simple- and combined-cycle configurations and typically are found in base-load and cyclic-duty operations.
Industry shorthand refers to the engine as the “6B,” only nine “A” models having been built. The first “B” that went into operation 31 years ago reflected a 5.7-MW increase in output over the first “A” and a 170-deg-F increase in firing temperature to 2020F. After four model upgrades, the PG5681B offered today is rated 41 MW at a firing temperature of 2084F. This design is now 12 years old. From the beginning, the single-shaft, two-bearing unit has employed a 17-stage compressor and three-stage turbine.
A focal point of the 26th annual Frame 6 Users Group meeting, held the last week of June at the Hyatt Regency Greenville (SC), was rotor lifetime assessment. Several owner/operators active in the group have engines north of 150,000 hours and wanted to know what they should be doing to meet the intent of the OEM’s Technical Information Letter (TIL) 1576, issued in 2007.
The document requires rotors with 200,000 equivalent operating hours or 5000 equivalent starts (whichever comes first) to undergo a comprehensive inspection. Hours-limited rotors that pass inspection, with or without rehabilitation or replacement of critical parts, can be certified for extended service (50,000 or more hours). The OEM’s current position reportedly is that rotors having accumulated 5000 starts are at end of life no matter how good they might look.
The group’s steering committee, co-chaired by Jeff Gillis of ExxonMobil Chemical and Sam Moots of Colorado Energy, incorporated into this year’s program a comprehensive and balanced look at rotor life management. It included the following:
• A formal presentation by the OEM on the subject.
• A special GE roundtable on compressor and gas-turbine life management.
• A user’s experience with the OEM’s rotor lifetime assessment process.
• A third-party service provider’s methodology for, and experience in, conducting rotor end-of-life inspections.
The user speaking on his company’s experience with GE’s lifetime inspection service began with a backgrounder on the equipment inspected and the scope of the inspection. His plant’s three 6Bs were commissioned in 1989 and had accumulated 180,000 fired hours in base-load service on natural gas using steam injection for NOx control. First step in the inspection process was complete rotor disassembly. Next came nondestructive examination (NDE) of CrMoV steel wheels installed in the turbine and aft compressor (rows 14-17), and the distance piece. Inspection efforts focused on dovetail slots, rabbets, wheel bores, and bolt holes. Ground rules specified that all indications found had to be evaluated by the OEM’s engineers. Those evaluations might allow continued use as-is or require blending, repair, or component replacement.
The user purchased a new rotor and installed it in Unit 1. The original Unit 1 rotor was installed in Unit 2 after inspection and refurbishment. The Unit 2 rotor, in turn, was installed in Unit 3 after it was inspected and refurbished. A decision regarding the disposition of the original Unit 3 rotor is pending. The speaker said the OEM’s schedule for the end-of-life (EOL) assessment typically is nine weeks, broken down as follows: two to three weeks for de-stack and clean-up; two weeks for the EOL inspection, three to four weeks for reassembly assuming only minimal machining is required to correct any minor deficiencies identified.
The user next presented inspection results for one unit to give his colleagues a view of the types of NDE techniques employed and the level of detail sought. Here’s what he said:
• Visual inspection and replication. Pitting was found in the bore from the compressor 16th stage through the second-stage turbine wheel, but it was not life-limiting. No microcracking or variations in grain structure were identified. Corrective action: Bore surfaces were reconditioned/honed to eliminate stress concentrations and bore fatigue damage.
• Magnetic particle inspection of each component produced no findings.
• Hardness readings all were within the OEM’s serviceable tolerance ranges; no concerns.
• Ultrasonic testing of bore surfaces of all components covered in the scope of work produced no limiting findings. Some indications were found imbedded into the distance piece. They were considered “birth defects.” Fracture assessments determined no anticipated propagations to failure within 750 starts—decades of useful life for a base-load unit.
The bottom line: GE certified this rotor for 750 total starts and 300,000 equivalent operating hours.
The third-party view on GT end-of-life inspections was provided by the team of Paul Tucker, president, First Independent Rotor Service of Texas (FIRST), Humble, Tex, and Gary Hensley, president, Veracity Technology Solutions, Tulsa. Tucker and Hensley were the principals in the first third-party consortium to conduct EOL inspections of GE frames following the release of TIL 1576. Some believe Tucker, who worked in GE rotor shops for more than two decades before striking out on his own, and Hensley may have done their first Frame 7 EOL inspection before the OEM.
The results of that work were reported at the CTOTF Fall Turbine Forum in 2007. The 7C rotor inspected came from a 1 x 1 STAG 100 combined cycle installed at Arizona Public Service Co’s West Phoenix Generating Station in 1976. It had more than 6000 actual starts at the time of the inspection and is still in service today. The rigorous EOL inspection program—including visual, magnetic-particle, ultrasonic, and eddy-current NDE, as well as metallurgical and dimensional verification—revealed no reportable indications. This led the inspection team to conclude, based on inspection-process fidelity, inspection methodology/criteria experience, and the excellent inspection report, “that it is more than reasonable to assume no defects would grow and propagate into anything near critical flaw size in the next major inspection interval.”
Tucker and Hensley discussed ongoing work in EOL inspections, which at the moment, focuses on several Frame 5s. Tucker pointed out that while he and Hensley had not yet conducted a lifetime assessment of a Frame 6 engine, this machine is similar in many respects to the Frame 5. For example, the compressors are very similar in design, both engines operate at similar speeds and temperatures, etc. The major difference between the two, Tucker said was in the turbine section: The Frame 5 has two stages, the Frame 6 three—like the Frame 7.
Inspections conducted by FIRST/Veracity to determine if an engine is fit for duty include the following:
• 100% three-dimensional boresonic inspection for internal flaws.
• Eddy-current inspection for creep.
• Hardness and replication for grain structure.
• Critical diameters measured for bore shrinkage.
• Finite element analysis (FEA) when flaws are identified.
Regarding the last point, FIRST and associates have developed rotor-component fracture models to perform FEA analyses when needed. Tucker said that FIRST’s critical flaw size and fatigue life estimations enable the company to analyze flaws with its models to see if flaw characteristics are detrimental to machine operation. He said the models can determine if specific flaws will fail, and if so, when in terms of hours and/or starts.
For the Frame 5P EOL inspections currently being conducted in FIRST’s shop, Tucker and Hensley are checking compressor wheels for rows 13 to 16, the distance piece, and both the HP and LP turbine disks (Figs 1-6). In one of the machines, a flaw cluster in the HP turbine wheel was identified during the phased-array ultrasonic inspection (Fig 7). FEA analysis suggested that this wheel not be returned to service.
 |
 |
 |
 |
 |
 |
 |
The most recent flaw found by FIRST/Veracity—coincidentally, at the time of the Frame 6 meeting—was in the 16th-stage compressor wheel of a rotor with 4800 starts. FEA analysis showed it to be “non-detrimental.” Tucker pointed out that the material used for compressor wheels is more forgiving than that used in the manufacture of turbine disks. At this point he stressed, “What you might find is not always bad,” mentioning that some users he speaks to believe a material defect means the part must be scrapped. Not true, Tucker continued, it’s not a “gloom-and-doom” thing.
Inspections conducted by FIRST/Veracity nominally take five days; however, analysis of any flaw that might be found can add a couple of weeks or more to the schedule. It was at this point that Tucker suggested to the Frame 6 owner/operators that they do an EOL inspection earlier rather than later—during a major or when you’re going to replace compressor blades, he said. Having a baseline condition assessment enables users to better manage the lives of their rotors. Plus, if it appears that one or more components will have to be replaced in the future, knowing earlier allows owner/operators to plan for that eventuality and to order refurbished or new parts on a standard delivery schedule.
A question from the floor: If you do not find defects during an EOL inspection, how long will the rotor last? Tucker responded by saying that creep life is the user’s greatest concern and asked rhetorically, “What factor of safety is built into the hours and starts numbers specified in TIL 1576? 2x?” In closing, Tucker mentioned ongoing work by FIRST/Veracity aimed at developing a new protocol to enable specific rotor-in-place tests for EOL evaluations. If that work is successful as expected, it would eliminate the need for unstacking the rotor in a shop unless a flaw is found or other work is required.
