501F user discussions, presentations cover from air inlet to exhaust

The compressor session that launched the user portion of the 2019 501F meeting touched on several topics of interest to the group, including the following:

    • Complications with the re-installation of outlet guide vanes (OGV) surprised many attendees. The OEM installed the airfoils backwards, something most plant personnel didn’t think was possible. A clue that something was amiss: Only 7 MW was gained from work done during the outage, half what the sister unit at this 2 × 1 combined cycle gained the previous year with essentially the same work scope.

A flashback occurred during tuning, at just under 60 MW, and the unit was shut down. Borescope inspection revealed pieces missing from the OGV and that the upper half of the OGV ring was installed incorrectly. The OEM was said to be modifying its design so it would not be possible to install the OGVs backwards.

Damage was considerable—including some to the first-stage vanes. Rocket tips in one basket burned up because of the flashback. Some pieces of hardware went downstream and were found ahead of the rotor air cooler. The user sharing this experience supported his impromptu presentation with a series of significant photos. 

    • Another user reported liberation of a section of a Row 5 blade, but the root-cause analysis (RCA) was inconclusive.

    • Attendees generally agreed that coating the compressor was beneficial.

    • Representatives from plants in the South said that with evap cooling and/or wet compression you’re going to have problems removing compressor blades, possibly in all rows, and this may be a job for a machine shop. Recommendation was to remove the compressor blades at the second major. A user said the OEM agrees with this approach and would provide refurbished blades for units covered by a long-term service agreement. An additional fee will get you new blades.

    • Borescope plugs can be difficult to remove. A Chesterton anti-seize product was recommended by one attendee, to prevent galling of stainless.

    • A user representing one of the largest 501F fleets in the country said his company was keeping compressor blades in service for at least three majors, adding that half of its units had run at least 100k hours with no failures reported. It was said there was no way to NDE for fatigue impact.

Hexavalent chromium has been a gas-turbine-outage hot topic for the last two years or so given the focus on personnel safety. Surprising, perhaps, because the welding of chromium-rich piping and boiler components has been ongoing for decades with personnel protection and safety always a top priority. If you want to learn more about hex chrome, contact any of the OEMs that participated in the 2019 meeting. Each has published guidelines and procedures for removal and disposal of the yellow material.

A user described his plant’s experience with hex chrome during a recent hot-gas-path (HGP) inspection. The insulation provided by the OEM for the gas turbines was replaced during the outage. Residue (dust) containing hex chrome was found under the insulation in the exhaust-bearing tunnel on one unit.

An environmental services company was engaged to properly remove and dispose of the residue. Process involved establishing a regulated work area, air monitoring, removal of the dust using wet decontamination techniques and hand tools, and final testing (OSHA ID-215M) to confirm airborne concentrations were below the Permissible Exposure Limit. Results showed less than 0.0017 mg/m³, well below the so-called Action Level of 0.25.

Row 1 vanes got some air time with a user calling attention to a new design with larger cooling holes less likely to plug and cause burn-up of the critical part. This development may be particularly important to users wanting 32k longevity. The speaker said the new design had been operated for 8000 hours at his plant with a borescope inspection confirming success. He said vanes of both the old and new designs could be integrated in the same row.

Four-way joint leakage is topic of interest at many user-group meetings. An attendee reported on the use of Deacon putty rope as a possible solution. The product was injected to the joint area via a false bolt hole and then cured as instructed. It helped but did not stop the leak.

Combustible-gas detection upgrades were prompted by a rapid gas expansion incident in the electrical package that damaged the DCS and the exterior wall of the motor control center, among other things. The speaker reminded attendees that the lower and upper explosive limits for hydrogen and methane, respectively, were 4% and 74% for the former and 5% and 15% for the latter. Gas concentrations outside those limits are either too lean or too rich to support combustion. Explosion-proof components are a must in areas where gas can be released or can accumulate.

Exhaust cylinder and manifold replacement on one gas turbine at a 2 × 1 combined cycle was part of an 11-week major inspection that included rotor lifetime inspection, replacement of HGP and CI (combustion inspection) parts with those from a third party, and generator and starting-package inspections. Operational stats in round numbers: 1650 equivalent starts, 93,000 total fired hours, and nearly 22,000 turning-gear hours.

Exhaust system issues prompting the replacement of both cylinder and manifold included cracking of the diffuser (severe), strut shield, aft static seal, and struts. The plant owner elected to switch OEMs for the new exhaust components and their installation. The illustrations and photos incorporated into the presentation would be of value to someone planning a similar project. 

Torque tube and air separator were replaced by an alternative OEM when the rotor undergoing a comprehensive inspection was in its shop (see item immediately above). The original air separator was replaced with one of bolted design. More detail is available in the MHPS section elsewhere in this issue. 

Important to note is that there was no problem with the existing torque tube and air separator. They were replaced with no schedule impact to mitigate risk while the rotor was in the shop (a requirement for replacement), given the problems experienced with the torque tube and air separator from the original OEM. The owner had experienced a torque-tube failure at another plant.

A third torque-tube failure was revisited by another user, this one on a starts-based Model FD (DLN combustor) with a nominal 1400 equivalent starts and 9500 equivalent baseload hours. Following a long run, plant was informed of a vibration increase (exhaust bearing) by the owner’s M&D center. Corporate engineering and the plant agreed the unit had to be shut down for inspection.

All major gas-turbine components were checked—including torsion bars, turbine support structure, and trunnions—but no connection to the vibration increase was identified from that work. The 2× component of vibration showed an increase in both magnitude at operating speed as well as shifts in both magnitude and speed during coast-down. These findings were consistent with known rotor structural failure signatures. Experts said an increase in the 2× component would “confirm” a torque-tube crack.

The owner’s engineers met with another user who had experienced a torque-tube crack to compare vibration signatures from both incidents. They were similar. Data and vibration signatures were included in the presentation.

A test plan was developed by the OEM and owner and a restart was attempted. The unit failed to reach full-speed/no-load and was shut down. The rotor was de-stacked at the OEM’s shop and a 22-in.-long crack found; the air separator was fine. Torque tube was replaced in-kind because the OEM did not have a new design available at the time.

Generator stator frame cracking had been an ongoing issue with one of the plant’s three generators and the presenter from previous years was back at the podium again in 2019. Continuity of expertise is important in years-long investigations.

Review: In October 2015 dusting on endwindings was identified during a routine inspection. While investigating the cause of the dusting, plant personnel discovered 12 cracks in several frame locations. The OEM requested further inspections the following January and March. Three additional cracks were documented.

During a maintenance outage in July 2016, the lower half of the generator was inspected and areas that had not been examined previously were checked. The result: Five additional undocumented cracks were found in the frame rings, plus one in a baffle ring. A few months later the upper half of the generator was re-inspected and three more indications were found. There was no NDE so the indications could not be confirmed as cracks.

By the end of November 2016, the AeroPac I had accumulated more than 56k baseload hours and nearly 2500 equivalent starts. Seventeen upper-half and 13 lower-half cracks had been recorded. To mitigate the problem, stop-crack holes were drilled. 

The OEM suggested installing 44 accelerometers on the machine to help in finding the root cause of the cracking.

Ahead of the 2019 user-group meeting, the OEM revealed that the crack-stop holes had indeed stopped the cracks. It also recommended performing intermediate partial inspections as outage schedules permitted. Plus, the RCA concluded low-cycle fatigue initiated the cracks and high-cycle fatigue promoted their propagation.

The OEM proposed a repair plan earlier this year. Attend the 2020 meeting of the 501F Users Group for a report on the results of that effort.

GE experience. The 501F is unique among user groups in that all of the world’s major frame gas-turbine manufacturers participate in meaningful fashion. Perhaps the most important question on the minds of owner/operators coming into the 2019 meeting was this: How would GE perform on the overhaul of an engine with which it had no native experience?

While a sample of one is of little statistical significance, Mexico’s GPG Company, familiar to many readers, helped answer at least some questions its colleagues had with an information-rich presentation of more than 50 slides.

GPG has broad gas-turbine knowledge for a relatively small company (four plants). Here’s a rundown:

    • Hermosillo, 1 × 1 combined cycle powered by an Alstom GT24 engine.

    • Naco Nogales, 1 × 1 combined cycle powered by a Siemens 501G.

    • Tuxpan, two 2 × 1 combined cycles powered by Mitsubishi 501F3s.

    • Norte-Durango, 2 × 1 combined cycle powered by Siemens 501FD3s.

Its Norte-Durango and Tuxpan units are covered by GE performance LTSAs and include both planned and unplanned maintenance for gas and steam turbines and auxiliaries, generators and auxiliaries, main steam valves and actuators, and control systems.

GE delivered on its performance upgrades at Tuxpan with an increase in combined-cycle generating capability of 8.3% and a heat-rate improvement of 2.6%. Numbers were similar for Norte-Durango.

User-group meeting attendees generally do not get the level of detail provided by GPG’s engineers. They spelled out what they believe to be the vendor’s strengths and weaknesses and where improvement is needed. One example: New connection rings for the generators did not fit (design error), delaying outage completion by 28 days.

Posted in 501 F&G Users Group |

Comments are closed.