Balanced program of user-only discussion forums and presentations by owner/operators and outside experts focuses on member O&M concerns

Frame 6 gas turbines (GTs) began generating kilowatt- hours in 1978. Since that time, nearly a thousand engines have been installed worldwide. Design features have evolved significantly over the years—two different models of the 6A and seven of the popular 6B— to accommodate evolving market needs, such as increased generating capability and lower emissions.

This means there are many subjects to cover at a Frame 6 Users Group meeting. For example, owner/ operators with the latest engines firing at the highest turbine inlet temperatures need to know about the coating/repair technologies available to extend the lives of hot-gas path (HGP) parts; those with the older machines generally have more interest in upgrades to boost performance.

The Frame 6 Users Group celebrated its 20th anniversary in mid June at The Buttes (a Marriott Resort) in Tempe, Ariz. It was a high-energy meeting, driven from the front of the room by Co-chairs Larry Flashberg and Jeff Gillis with a big assist by “lifers” John F D Peterson and Charlie Zirkelback and several other long-term members in the audience. Q&A and general discussion were non-stop over the two-and-a-half-day program.

Steering committee
The Frame 6 Users Group steering committee expanded in 2006 to add expertise in key areas. Five of the six members elected in 2005 returned for another year and are shown together in the group photo along with newcomer Kyle Todt (bottom row, right). Scott Berry, Zahi Youwakim, and Brian Walker are in the top row (l to r), while Co-chairs Jeff Gillis and Larry Flashberg are to the left of Todt.
Company affiliations of the returning committee members:

  • Flashberg, plant engineer, Saguaro Power Co.
  • Gillis, senior staff engineer, ExxonMobil Chemical Co.
  • Berry, plant manager, Anderson and Richmond powerplants, Indiana Municipal Power Agency.
  • Youwakim, utility plant engineer, Huntsman Petrochemical Corp.
  • Walker, manager of maintenance, Foster Wheeler Martinez Inc.

Thumbnails of the new members
Todt is plant operations engineer for Chevron Global Power Generation’s Sycamore Cogeneration Plant in Bakersfield, Calif. The Univ of Missouri-Rolla grad (mechanical engineering) is responsible for planning and executing maintenance projects/outages, and assisting in operations troubleshooting and special projects, at four cogeneration facilities—each with one Frame 6B.

Also joining the steering committee this year are Tim Elgin and Frank Moreno. Elgin is general manager of the Pittsfield (Mass) Generating Co LP, formerly known as Altresco, which was one of the first combined-cycle plants in New England when it went commercial in September 1990. He has been involved in all phases of Frame 6 engineering, operations, and maintenance—Pittsfield has three engines—since initial startup. The facility, managed by PurEnergy Operating Services, ran base-load until 2004; it now is a merchant peaking plant. Elgin is a 1984 graduate of the Massachusetts Maritime Academy (marine engineering).

Moreno is with ExxonMobil Chemical in Beaumont, Tex. More biographical detail, as well as information on how to participate in the group’s activities and discussion forum, is available at http://www.frame6usersgroup.org.

But it takes more than good floor leadership and knowledgeable and loyal members to make a meeting successful. A top-notch technical program developed by the steering committee (sidebar), a comfortable venue and amphitheater-style meeting room, and knowledgeable vendor representatives participating in the exhibition are very important as well.

Also, don’t underestimate the value of a relaxed social program, excellent food, and well-organized logistics for keeping a smile on everyone’s face. Kudos for this effort go to Conference Coordinator Wickey Elmo of Goose Creek Systems Inc (GCSI).

The primary goal of the user organization is to provide members a forum for open dialog and exchange of information conducive to improving O&M practices related to the Frame 6 series of engines. It also allows owner/operators to interface with the manufacturer—GE Energy, Atlanta—regarding generic fleetwide issues.

These objectives are accomplished at the annual meeting through useronly roundtable discussions. There were 13 roundtable discussions at the 2006 conference covering the following subject areas:

  • Unit operating history.
  • Compressors.
  • Maintenance execution, including LTSAs.
  • Combustion section.
  • Environmental issues, DLN.
  • Accessory and load-gear issues.
  • Black start.
  • Controls.
  • Generator and excitation.
  • Component repairs, parts evaluation.
  • Turbine section
  • Power enhancement.
  • Inlet system, including cooling.

Information on group activities of a non-technical nature pertinent to both members and other interested parties is provided at www.frame6usersgroup.org. Remember that meeting attendance is your access to the value-added Frame 6 network. Pencil in on your calendar now that the 2007 conference will be at the Wyndham Jacksonville Riverwalk, Jacksonville, Fla, June 11-14; watch the website for details on registration and exhibiting/sponsoring as they become available. About 100 Frame 6 users are expected to participate. Contact Elmo at 704-753-5377 or info@frame6usersgroup.com with specific questions.

Be confident that there will be participants at the upcoming meeting from generating facilities with the same equipment and duty cycle that your plant has. This makes the conference an ideal place to network. Others certainly have experienced the same issues as you, and have had the experts tell them, as they probably have told you, “Never saw this problem before.”

Some of the demographic information collected during the 2006 meeting may be helpful for justifying to management your participation in Jacksonville. Here are a few significant facts:

  • Forty different generating companies were involved, including one from the UK and one from Nigeria. Ownership included 20 chemical plants, six refineries, six independent power producers, and five regulated electric utilities. Several produce thermal energy in addition to electricity.Most attendees represented several engines; one user was responsible for 10.
  • About three-quarters of the audience represented base-load machines; most of the remainder, peaking.
  • More than 90% burned natural gas alone or in combination with distillate or a gaseous process byproduct.
  • Regarding hours of operation, the fleet leader (at the meeting) had accumulated more than 175,000. Six engines represented were over the 170,000-hr mark. About onequarter of the users present were over the 100,000-hr mark.
  • Majority of the users present injected steam into their GTs; 10 injected water.

A review of the suggested discussion topics submitted by the Frame 6 member community in advance of the meeting offers an idea as to the broad scope of the user information exchange. Here’s a sample:

  • How to conduct a 100,000-hr compressor inspection.
  • Making GTs black-start and faststart capable.
  • Precision clearances during outages.
  • Vibration analysis.
  • Fourth major: What are people doing different?
  • Vibration on the generator outboard bearing.
  • Wheel-space temperatures, loss of performance.
  • Fuel-nozzle pluggage.
  • Turbine-bucket run time and refurbishment.

Program overview

The roundtable discussions are sandwiched between formal user and vendor presentations. Alternating between open forums and prepared presentations keeps the program lively and delegates fresh.

The first formal event was the Users Welcome Reception, the evening before the opening session on June 13. However, the golfers showed up earlier to participate in a tournament hosted by Rick Parker of Zokman Products Inc and Lee Wood of Wood Group Gas Turbine Services.

The first foursome teed off at the ASU (Arizona State Univ) Karsten Club shortly after noon. Temperature was 113F and there was precious little shade. By the time everyone had completed 18 holes no one could remember—or seemed to care—who had the lowest score or the longest drive, or was closest to the pin.

Day One was long. Flashberg and Gillis started promptly at 7:30 am and finished at 4:30, the formal program interrupted only by a first-rate buffet lunch at the picturesque Top of the Rock restaurant and two coffee breaks. Then came the vendor fair and reception (too much good food again), which lasted until 8 pm.

Day Two was pretty much the same, except for the ending. Replacing the vendor fair was a fun western reception and barbeque dinner at the rustic Pinnacle Peak Patio somewhere in the Arizona desert. Leave it to Elmo to come up with something different and provide yet another opportunity for relaxed networking (montage above).

User-only roundtables

Unit operating history

First thing that Flashberg and Gillis do after making some opening remarks is to get everyone comfortable and participating in the discussion. What better way to do this than to go around the room and ask each attendee to introduce himself/herself and then to characterize the various plants/equipment represented.Several delegates revealed that they’ve been attending Frame 6 meetings since the end of the 1980s, attesting to the valuable exchange of ideas, practices, and procedures year after year. All but a handful of attendees were responsible for machines that had accumulated more than 50,000 service hours. Fewest service hours numbered 500, for a peaker that was started only about 100 times in its first three years of operation.

One user said his combined-cycle plant had over 3700 starts in 12 years of service. A hot start takes 90 minutes at that plant; a cold start, 150 minutes. A few other users said they had more than 1000 starts on their machines; two said they start more than once a day.

Experience in firing a range of fuels is to be expected with this group because a half, or more, of the participants work at refineries and chemical plants and they have access to byproduct gases from process units. Several who blend byproduct gas with natural gas stressed the need to hold the mixture within Wobbe limits for reliable, safe combustion. This typically means restricting the percentage of process off-gas in the mixture to about 15%; olefins were said to cause burner fouling. Fuel systems for burning two gases concurrently are addressed in detail by Peterson in the user presentations section of this report.

About one-third of the participants were responsible for machines predating 1990; another third had engines built in the 1991-1995 timeframe. About a third of the group was operating at a turbine inlet temperature of 2020F; no one lower. Highest TIT was 2080F and there were several at that temperature.

Mark IV and Mark V control systems each were installed on about a third of the engines represented by the attendees. Half a dozen users were evaluating a Mark VI retrofit; another was considering a non-GE control system. Typical reason for conversion: Can’t get the expertise needed to maintain the Mark IV.

Compressor discussion

With everyone primed and ready to participate, the meeting shifted into high gear with a roundtable discussion on compressors. Flashberg called for the first question, reminiscent of a country auctioneer asking for the first bid. It was one sure to generate considerable discussion: “Are any plants using fogging or wet compression experiencing compressor problems?”

Reason for asking: The user had to refit the filter house, located at a chemical plant, with a stainless steel liner after only three years of service. Silencers took the big hit, he said, because holes were punched after the galvanizing process thereby exposing carbon steel to the corrosive environment. Galvanized coating just peeled off.

First response came from a merchant power producer in California who said his plant had been doing fogging and wet compression for five years problem-free. Another user with an evaporative cooler at the GT inlet said after 15 years of service he too was seeing accelerated wear of silencers and trash screen.

Much discussion ensued on experience with equipment supplied by the various manufacturers of fogging/wet compression systems, nozzle arrays, droplet size, water quality, positive drainage of filter houses, etc. Sensitive subjects like this require your attendance at the meeting to gain an accurate perspective from the discussion. As noted earlier, the 2007 conference will be in Jacksonville, June 11-14; plan to be there.

Consensus of the group: Inspect the inlet air house thoroughly each outage to see what work is required. Be sure to check the trash screen for rubbing wear where the horizontal and vertical wires cross. Plan to replace screens and silencers during the second major, or the third, depending on plant location. Suggestion: Specify Nimonic 50 wire instead of stainless; it resists fretting/rubbing wear. Place your order early; the material can be a long-lead-time item (couple of months).

Another user recommended MIG or TIG welding for any work on the filter house. Reason: If welders use sticks and fail to do a proper job of slag removal that material will wind up in the compressor. Yet another idea for protecting the compressor: Order interlocking media for your evap cooler and get rid of caulking strips. It prevents dirty air from bypassing the filter.

Change filter media every major. Keep in mind that industry experience with glass media generally is not positive. Paper is fine, several owner/operators reported, if you don’t use demineralized water. It causes the glue to disintegrate.

Proper fire protection was stressed when welding on or near the filter house. Put a fire blanket over the filters and be sure to saturate the evapcooler media before work begins. When replacing the trash screen, remember that the easiest way to do this may be to take off the front of the inlet house and lower it to the ground.

2006 Peterson Award to Zirkelback

The John F D Peterson Award, named for the first recipient, is presented by the Frame 6 Users Group in recognition of extraordinary contributions to the organization and to the gas-turbine-based sector of the electric power industry. C E (Charlie) Zirkelback received the award for 2006 at the annual meeting from Steering Committee Member Brian Walker and Peterson (photo).

Zirkelback is recognized worldwide for his expertise in the maintenance and performance assessment of rotating machines. His work at Union Carbide Corp from 1966 to 2003 focused on continuous improvement of equipment availability, reliability, safety, efficiency, and production capacity. Zirkelback has lectured and written extensively on the following subjects associated with large steam and gas turbines, compressors, generators, and pumps: specification, evaluation, installation, preventive/predictive maintenance, protective monitoring, failure analysis, troubleshooting, repair, and remedial design.

Throughout his career, Zirkelback has been active in several organizations dedicated to gas turbines—including, the ASME Industrial Gas Turbine Institute, Frame 6 Users Group, and the Gas Turbine Users Association. He remains active in the industry and is currently president of Z-MechTech Inc, a consultancy (zirkelcm@cableone.net, 361-552-5252).

A formal vendor presentation on compressor washing by Hugh Sales of Gas Turbine Efficiency, Houston, followed the morning coffee break and generated more compressorrelated questions. How long to online wash daily without detergent? Sales said daily washing for about three minutes was typical. How often must spray nozzles be replaced? Sales replied that erosion was rare but nozzle replacement during a major probably was a good idea.

There were several other questions regarding droplet size, water temperature, water pressure, etc, for optimum cleaning. Guidelines suggested by Sales are included in the later section on vendor presentations.

After Sales left the room, useronly discussion continued. Some of the points made:

  • A daily compressor rinse with demineralized water is beneficial because aerosol salts do go through the filters and you want to wash them off IGVs (inlet guide vanes), blades, and vanes to maximize the life of those parts.
  • Air-purge the wash water system after use. Piping and other metal parts can deteriorate if they remain in contact with demineralized water.
  • Group consensus: Offline wash quarterly, online wash daily in summer. A couple of users reported using detergent for their online wash one day in seven because local air quality is adversely impacted by a process-plant setting, heavy auto traffic, etc. Caution was advised, however: Online washing with detergent can blind flame scanners. If you blind two, the unit will trip.
  • Experience with various detergents was part of the discussion. Users suggested running tests on both water- and solventbased formulations to identify the best product for your particular machine/site conditions.
  • Factor wash-water effluent disposal into your evaluation of alternative detergents. This can be a major consideration if you do not have a plant-wide wastewater treatment system. Keep in mind that treatment is different for water- and solvent-based chemicals. Also, factor into your decision the chemical concentration required. One user reported that his plant was able to reduce the chemical concentration by more than half and achieve the same level of cleaning. Advantages included easier rinsing, reduced water usage, smaller waste stream.
  • Periodically check to see that inlet-house drains are open when water washing. Any puddles forming on the filter-house floor may be sucked into the compressor in slugs which can cause the outlet thermocouple spread to exceed limits and trip the unit.

Change of subject. Question: “I have a 14-yr-old machine in baseload service. Compressor efficiency has decreased by several percent. What might be the cause?” A few users suggested that it might not be a compressor problem, but rather a leaking bleed valve. Another said the following; “Check your compressor blades. When they wear thin you can increase efficiency but after that rounding occurs and efficiency drops.” Others said it’s all about tip clearances and to check them; when excessive, there’s financial justification for reblading the unit. Yet another opinion: “Shot-peen blades for a performance boost.”

Next question concerned use of non-OEM compressor blades. Two comments, both “no problems encountered.” Helpful hint for repair of cracked compressor casings: User reported that a crack running between two borescope ports was fixed successfully using a commercially available metal stitching technique.

Re-shimming of stators on two machines undergoing major inspections was mentioned. The missing shims were said to have passed through both machines without damaging downstream parts. User opted for making its own shims and hammering them into place.

Two users experienced rotor “lockups” on trips from full load. The first said the ratchet wouldn’t “come on” and then modified that to say it actually did start but tripped on overload. When the machine cools, he added, you can turn the rotor. Problem diagnosis: Engineers noted that there’s no insulation on the GT proper, that the engine is insulated by the compartment. With cooling fans on after a trip, the top of the casing cools much faster than the bottom and the rotor locks up until the casing returns to thermal equilibrium. Solution: Turn off cooling fans on a unit trip.

Second user said it was 48 hours after a trip before their ratchet would work. Problem diagnosis: Exhaustframe leaks heated up the No. 2 bearing and there was slight rotor movement that caused a few compressor blades to hang up.

Maintenance execution, LTSAs

First thought on the minds of many attendees when this session started: The quality of work done under longterm service agreements has slipped and continues to slip. Several owner/ operators with LTSAs said outright that the expectation of having firstrate people available to conduct their outages was not being met. A couple went as far as suggesting that outage teams “were not qualified” to do the work required.

The frustrations experienced by many Frame 6 members were clearly evident in the snowballing interchange. One contended that GE field engineers don’t have approval authority for activities that you would expect they could approve. Control was from “the office,” he added. Another echoed: “Biggest problem with my last outage was the inability of GE people to make decisions impacting schedule.”

“Expense is up, quality down,” offered one user. Yet another brought into question the quality of subcontractors brought in by the OEM. The situation appears similar outside North America. A participant from the UK said for their DLN conversion, the parts came from America, labor from “somewhere else,” and the technical advisor from France. He rated the outage “a mess.” This must have been an accurate assessment: The OEM reportedly got hit with LDs (liquidated damages) and paid.

One attendee suggested that not having an LTSA might be an advantage. If you don’t get the people you want, he said, you can go to a competitor. Also mentioned during the discussion was the wide range of LTSAs and CSAs (Contractual Services Agreement) available. Someone asked if you opted for the “full LTSA,” would you get the GE “gold card” and preferential treatment? No answer to that one, but another user said some of the problems mentioned go away when you have enough engines onsite or in your immediate area to justify a dedicated CPM (contract performance manager).

Flashberg finally was able to tone down the emotion by recognizing an owner/operator with a sense of fairness who thought some of the issues were directly related to changes users often make to outage schedules, sometimes only a few weeks before the contract start date. When that happens, he said, you can’t expect to get the same people GE scheduled for your plant under the original plan. They have to be reassigned.

Another user offered that the more closely you define the scope of work for the outage, the better will be your experience. This is especially important for an ageing fleet. Attendance at Frame 6 meetings gives you a heads-up on what work to plan for. Build contingency into your schedule to accommodate what you might have to do. That’s better than trying to factor in a “major discoverable” after the outage has started.

Yet another idea with merit: Don’t try to shoehorn your maintenance outages into the traditional March to May timeframe. The GT-based power generation industry—now totaling more than 300,000 MW in the US alone, including the steamturbine component of combined-cycle plants—is far to large to expect that the lion’s share of the work can be done in just three months, or even six, if you include the fall as an alternative as many do.

Several users said they now schedule outages based on the availability of field-service personnel and shop capability, as well as on internal resources to support the effort. One reported that he now schedules outages for the summer to accommodate the needs of process. It’s the process that dictates to power, he added.

The group worked up a big appetite jawing about LTSAs and spent the last few minutes before lunch reflecting on maintenance execution. One complaint from several members located on the Gulf Coast was that hurricane repairs have ravaged powerplant budgets.

Staffing for unit support was briefly discussed. For one 6B, thinking was that a fraction of one mechanic was required, about one I&E technician, and a fraction of one operator. No one said exactly what the fractions were, however.

A couple of attendees initiated a discussion based on their belief that accountability and ownership are no longer part of the worker culture. Suggestion was to have signoffs on normal operator duties to bring accountability into focus.

Generator and excitation

All attendees had air-cooled generators as you would expect for a nominal 40-MW engine. Roughly two-thirds of the participants had GE machines, remainder Brush.

First question came from a user looking for help with a troubleshooting scheme for a solid-state exciter that had passed its 20th birthday. Problem description: Voltage goes “nuts” periodically. Voltage spikes to 1000 V, stays up for a short period, then drops back to normal. Cause of the problem is unknown. A couple of users suggested there might be a problem with circuit cards, or perhaps a wiring short. Certainly nothing definite, but a “lead” nonetheless.

Second problem: Two users independently reported intermittent local vibration at the front bearings of their machines. Vibration occurs early in the morning during summer months; when the temperature warms around mid day, it drops to normal. Defining characteristics: More frame vibration than shaft vibration, 60-Hz frequency. Inspections identified no problems and all clearances were at recommended numbers.

As for a possible solution, one user remembered a 10-yr-old stiffener mod and thought it might well be the answer. Note: A big benefit of user meetings is that long-term owner/ operators often have a deeper O&M knowledge base than the OEM’s field service personnel.

Regarding generator maintenance, most users recommend pulling the rotor during every major. Steps in a typical rotor overhaul are outlined in “Key steps in inspecting, reconditioning generator rotors,” p OH-75, 2007 Outage Handbook supplement inserted in the middle of this issue.

The need to have an owner’s representative onsite when the generator is being overhauled was stressed. Careless shop work and other horror stories dominated part of this session. Examples: Induction heater had insufficient capacity to allow removal of retaining rings, end blocks were modified and came loose in during high-speed balancing, etc. Shop visits are a necessary part of the vendor selection process.

Other discussion focused on a miniature air-gap inspection crawler developed by GE to check wedges, how to prevent lube oil leaks, and electrical tests to conduct periodically.

Combustion section

Discussion related to interval extension occupied a major part of this session. Seven in attendance said they have gone to 24,000 hours between inspections by implementing an extender kit. Another user said his plant opted for a partial extender kit (not interested in all aspects of the full extender kit) to get 16,000 hours. This is considered a “mini major” because first-stage nozzles won’t last more than 32,000 hours. Scope of work will vary with the outage.

A full Nimonic body for transition pieces (not just the “picture frames”) was recommended to reduce repair frequency. One user reported 48,000 hours on his engine’s Nimonic TPs without repairs. Two more notes:

  • Advanced coatings on first-stage nozzles contribute markedly to extended lifetime.
  • Exhaust-casing flex seals are failure- prone. Replace when you have the machine apart to keep exhaust spreads within spec.

Open discussion

On the last day of the meeting (morning only) attendees are invited to pick topics they’d like to discuss. The following thumbnails are based on that activity.

Fire protection. Most users have low-pressure CO2 extinguishing systems, some Haylon, a few high-pressure CO2, one a fogging system. The user selecting this subject cited the premature-discharge characteristic of one vendor’s offering. System is great at putting out fires, he said, but it was not fault-tolerant. Operational reliability is very important, also, because you don’t want your turbine tripping unnecessarily.

It seems that this particular manufacturer’s system is sensitive to heat. Recommendation is to use hightemperature wire and conduit and to keep the control wire away from high heat sources. Putting a roof over CO2 bottles and the control panels also was recommended for keeping direct sunlight off critical components. In another instance, rain shorted out one of the pull boxes and dumped the CO2. Now two pulls are required to initiate operation.

Annual testing of fire detection/ suppression systems was recommended. One user said they check each detector and conduct a “puff” test on CO2 because it verifies operability and uses only 2% of system capacity. Note that CO2 systems typically are sized for two full dumps plus 10%.

A suggested practice is to qualify a fire detection/suppression system when it is installed, but without dumping CO2. Check detectors, dampers, nozzles, etc. Operation of dampers, in particular, should be verified during all outages. One user experienced inadvertent closing of package dampers that caused the compartment to overheat on next operation and the unit tripped.

Lube oil. Discussion focused on the temperature of lube-oil to GT bearings. One in five said their supply temperature ranged from 110F to 120F, 40% said 120F-130F, and 30% reported 130F-140F; remainder operated at a higher temperature. About two-thirds of the attendees relied on fin-fan coolers for lube-oil cooling, remainder closed cooling-water systems. Getting lube-oil temperature to the recommended level in the summer when fin-fan coolers reject heat can be challenging in some cases. Booster cooling alternatives were suggested.

More on interval extension. One user favoring a 16,000-hr interval said you do a close inspection of the entire engine each outage but don’t pull the rotor, thereby mitigating risk. Run first-stage buckets to 48,000 hours (possibly 64,000 hours) and put them on the shelf. Philosophy is to operate parts as long as possible without compromising reliability.

Another owner/operator said they were comfortable with the idea of achieving 24,000-hr intervals with the 6B extender kit. Yet another said they run DLN-equipped machines 24,000 hours without ever conducting a combustor inspection. They just do the HGP at 24,000.

Regarding first-stage buckets, half a dozen users reported running 16- hole perimeter-cooled blades. Plan is to strip and recoat at 24,000 hours, run another 24,000 and toss. Another attendee said they wouldn’t make the decision to scrap until after metallurgical examination; buckets might just be able to run longer with a second repair. Inspection results for firststage buckets on a 2080F machine after 24,000 hours were good.

For the second stage, about half the crowd was running with six-hole buckets, the remainder seven holes. A couple of users hit the 72,000-hr mark and bought new blades.

User presentations

Formal presentations by owner/operators are well-received at user group meetings. Learning how one of your colleagues solved a nagging problem, advanced the state of the art, or developed a better solution than the OEM gives you a good feeling. In effect, you share in your colleague’s accomplishment. Plus, you never have to be concerned about the person at the podium wanting to sell you something.

There were several outstanding user presentations at the 2006 Frame 6 meeting (bullets below). Two presentations already have been developed into feature articles and are just referenced here to avoid double coverage; the complete articles are available at www.psimedia. info/ccjarchives.htm.

  • Changeout of compressor inlet guide vanes (IGVs) with rotor in place, Scott Berry, plant manager for Indiana Municipal Power Agency’s Anderson and Richmond peaking plants and member of the Frame 6 Users Group steering committee. Access “Replacement of damaged lower IGVs with rotor in place saves IMPA a million,” p OH-89, 2007 Outage Handbook supplement inserted in the middle of this issue.
  • Upgrading from Mark IV+ to Mark VI controls, Jeff Gillis, senior staff engineer, ExxonMobil Chemicals, Baytown, Tex, and cochair of the Frame 6 Users Group. Access “Upgrading controls to maximize performance, availability: Profiling a Mark IV+ to Mark VI conversion,” 2Q/2006, p 118.
  • Highlights of a third major at 133,000 fired hours, Jeff Gillis.
  • Unconventional dual gas fuel system offers increased fuel-mix flexibility, John F D Peterson, BASF Corp, Geismar, La.
  • Disaster recovery, black-start considerations, Zahi Youwakim, utility plant engineer for Huntsman Petrochemical Corp and member of the Frame 6 Users Group steering committee.

133,000 fired hours and counting

Who knows how long a durable frame engine like the 6B can operate before replacement is more cost-effective than repair. As noted earlier, 20 machines represented by owner/ operators at the user group’s 2006 meeting had passed the 100,000-hr mark, six were over 170,000 hours. Train 3 at ExxonMobil’s Baytown (Tex) chemical complex was within about 40,000 hours of the leader at the time of the conference.

Gillis discussed the third major for Train 3’s GT shortly after lunch on the first day of the meeting. It was during this outage that the 6B was converted from Mark IV+ to Mark VI controls, which was the subject of Gillis’ presentation on the next day (see reference in second bullet above).

Results of the Train 3 major accentuated the positives of the 6Bs: durable, reliable power producers when properly operated and maintained. Major action items for the outage beyond the controls conversion were turbine major, generator overhaul, recoating of the inlet bellmouth, casing alignment, repair of inlet filter house, installation of a non-pressurized second-stage nozzle ring, and new honeycomb second-stage shroud blocks.

Gillis started at the compressor end. His assessment of compressor condition: “Looked good.” See for yourself (Fig 1). Gillis said the compressor had been water-washed daily when operating (which was most of the time) and was crank-washed just prior to the outage.

The whole rotor was lifted from the lower casing half, secured in a shipping capsule, and trucked to the OEM’s shop in Cincinnati for cleaning and balancing. Glass-bead media was used, Gillis noted, because a CO2 blast doesn’t work well on the sticky deposits— an observation confirmed by several attendees—that were attributed to small oil leaks in the bellmouth area. Fig 2 shows that the compressor looked like new after cleaning.

Corrosion/erosion of the bellmouth dictated refurbishment of that component. A very hard two-part epoxy coating was selected for the job. Fig 3 shows that the repair vendor’s first attempt was unacceptable; note the orange-peel finish which would have adversely impacted efficiency. The bellmouth was returned to the shop where the first coating was sanded off and the job redone to the satisfaction of ExxonMobil. Corrosion also was found in the air-inlet filter house. Distressed metal was removed and new plate welded in.

The turbine section at the start of the outage is shown in Fig 4, after rotor overhaul in Fig 5. It, like the compressor was in good condition. First-stage nozzle ring was removed (Fig 6) and replaced with a spare (Fig 7). ExxonMobil’s three 6Bs at Baytown share four first-stage nozzle rings so there’s always one available. Shroudtip wear was in evidence on the second stage (Fig 8) but wasn’t so bad that it couldn’t be weld-repaired. Plant opted to replace the standard design with a honeycomb shroud (Fig 9). Varnish was found on the No. 2 turbine bearing (and on the No. 1 compressor bearing as well) and it was rebabbitted to eliminate the hot spot (Fig 10).

Cavitation damage was found on main-oil-pump gears and they were replaced. This typically is an action item each major for this machine.

Generator before and after overhaul photos are presented in Figs 11 and 12. Retaining rings were pulled, end turns checked, loose end blocks were addressed, and collector rings were ground and realigned.

Dual-fuel firing with two gases

Dual-fuel GTs firing natural gas as the primary fuel, with distillate oil as the backup, are common in the industry—particularly at electric generating plants built to supply the grid. In process plants, such as refineries and chemical plants, dualfuel turbines often operate on two gases rather than on gas and oil. This allows operators with access to process- byproduct gas streams to reduce their outside fuel purchases.

Peterson began by explaining that conventional dual-gas control systems use two separate sets of control valves, gas manifolds, and fuel nozzles, so that the two fuels mix only in the combustion zone. While this arrangement is proven and reliable, each fuel is limited on turndown. To illustrate: When burning both fuels, the minimum rate for either is about 30% of the total heat input, meaning the maximum for the other gas is 70%. To burn more of either gas would require shutting off one and burning 100% of the other.

However, as long as the two gases modulate between 30% and 70% (the total being 100%) these systems work well, he said. The 30% minimum flow limit is necessary to maintain the pressure ratio across the fuel nozzles in the proper range for safe and stable combustion. This pressure ratio, usually between 1.04 and 1.40, is the quotient of the absolute pressure of the fuel gas in the nozzle divided by the absolute pressure of the combustor.

In some cases, Peterson continued, more turndown is required— especially when the quantity of the second fuel is variable. One solution is to mix the two gases in a header after the control valves. The mixed fuel is then fed to a common set of fuel nozzles.

But the header solution is limited in that gases with large differences in Wobbe Index could never be burned in the same set of fuel nozzles over the full mix range. (To brush up on Wobbe, read “Improve GT operating flexibility, reliability with fuel-system mods,” 2006 Outage Handbook supplement to the 3Q/2005 issue of the COMBINED CYCLE Journal.)

Therefore, a second manifold, with its own set of fuel nozzles, is opened in parallel with the first when the pressure ratio across the first set approaches the 1.40 maximum (Fig 14). As soon as the second manifold is in service, the pressure ratio falls, but not so far that it reaches the 1.04 minimum. The deadband between single- and dual-manifold operation depends on the orifice areas of the two sets of fuel nozzles.

Peterson next offered a common scenario for a Frame 6B burning natural gas and a hydrogen-rich, low-heating-value byproduct gas: GT comes to base load on natural gas only, with a fuel-nozzle pressure ratio of around 1.25, well below the 1.40 maximum. The ring manifold pressure is roughly 215 psia, and combustor pressure (assumed to be compressor discharge pressure minus 2 or 3 psi) is about 175 psia.

The low-Btu fuel is introduced at a minimum 10% of the mix and the natural gas cuts back to 90%. The pressure ratio rises to perhaps 1.30, because of the lower heating value of the mix, but well below the 1.40 maximum. The percentage of byproduct gas is increased until the pressure ratio reaches 1.37. The mix at that point is roughly 85% natural gas and 15% low-Btu gas.

Then transition to two-manifold operation occurs, Peterson said, dropping the fuel-nozzle pressure ratio to around 1.10, well above the1.04 minimum. Once dual-manifold operation is established, the percentage of low- Btu gas can be adjusted anywhere between 10% and 90%, depending on composition.

There are a few additional things happening during dual-gas operation that require your attention, particularly when transferring from singleto dual-manifold operation, and back again. The second header, when not in use, must be purged with compressor discharge air to cool the nozzle tips. Then Peterson cautioned: If the byproduct fuel contains hydrogen, you must purge the second manifold with nitrogen prior to putting it in service and immediately after taking it out of service. Nevertheless, he said, the flexibility and turndown of a dual-gas system as described is worth the added complications of purges and control complexity.

Experience. The second Frame 6B installed at BASF’s Geismar Chemical Production Facility is equipped with steam-injected diffusion combustors and has operated successfully with two-manifold, dualgas system since its installation in 1998. A cooperative effort among GE Energy, Atlanta, engineering contractor Lockwood-Greene of Spartenburg, SC, and BASF personnel at Geismar, resulted in successful resolution of control issues during commissioning, leading to eight years of high reliability and availability for the machine.

A conventional dual-gas system with a 30/70 mix limitation would not have provided adequate flexibility in managing the low-heating-value fuel stream. The arrangement described, which permits mixes to 10/90, allows the safe and reliable combustion of a byproduct gas stream widely variable in flow and composition.

Disaster recovery

Lessons learned in getting industrial and commercial facilities back into operation following the disastrous hurricanes of 2005—Katrina in late August, Rita in mid September— have been a topic of discussion at virtually every user-group meeting this year.

Youwakim, one of many Frame 6 members from the Gulf Coast, began with an organized presentation, but the session soon turned into a wide open discussion among attendees. Ideas and lessons learned were flying back and forth so quickly it was virtually impossible to take notes. If you weren’t there you certainly missed a couple of ideas that could save your plant thousands of dollars— at least.

Youwakim said that perhaps the biggest problem was that water systems were inoperative. Gillis added that all cooling towers in his area were wrecked. Back to Youwakim: Thought the plant had back-start capability, but getting cooling water to key components for that start was a Herculean task.

Several users said their steam systems gave them the most trouble.

An O&M superintendent mentioned that they caught the batteries in time and recharged them with the emergency generator. Then they could power up the Mark IV control system, which is a prerequisite for GT restart. Generator had to be dried out using a small electric heater before it could be placed in service. Plant operated as an island for about a week before the utility could supply power again. But service reliability was not normal; the grid was shaky for another couple of weeks.

Berry suggested keeping about a week’s worth of fuel oil onsite in case gas is not available. Also, to be sure your UPS system/battery bank and circuitry are arranged to enable a startup. This means being able to deliver ac power to Mark IV controls, igniters, computer screens and printers, fuel-oil pump, etc.

Many of the ideas/suggestions presented at the Frame 6 meeting were the same or similar to those offered by Paul Terry, maintenance manager for RS Cogen LLC, Lake Charles, La, at the 501F User’s Group meeting last January. Terry’s remarks are summarized in the section “Preparing for, and recovering from, natural disasters,” which was included in the 501F report last issue (p 14). It’s worthwhile reviewing before updating the disaster-recovery plan for your plant. Another idea: Meet with a couple of other powerplant managers in your area who are served by the same utility to benefit from their experience.

Invited presentations

Owner/operators don’t have all the answers. That’s why a blue-ribbon lineup of invited presentations is critical to the success of a user-group meeting. The Frame 6 steering committee is particularly good at selecting subjects of importance to virtually everyone in the group and at finding speakers who can conduct a meaningful hour-long “workshop” without promoting their company’s products/services.

In Tempe, outside experts covered lube-oil (LO) systems, engine O&M, generator testing, compressor washing, borescope inspections, and evaluation of third-party hot-gas-path repairs. All were scheduled for, and ran, an hour, with questions—that is, except for the OEM’s Frame 6 clinic which took two.

Lube-oil refresher

It is important for power engineers to understand lubrication fundamentals— especially if they have responsibility for GT operation and maintenance. Keep in mind that high operating temperature is the leading cause of premature turbine oil failure in large frame engines. The drive for higher firing temperatures to increase efficiency has been the primary incentive for suppliers to develop more thermally robust turbine oils.

Likewise, the cyclic operating regime of most large frames, coupled with high bearing temperatures, can contribute to fouling of system hydraulics and adversely impact starting reliability. New formulations have been developed in the last few years to remedy this problem and to extend GT drain intervals.

The “refresher course” in GT lubrication conducted by Glen Sharkowicz, a technical advisor for Exxon- Mobil Lubricants & Petroleum Specialties Co, and his colleague, Joe Cervassi, was particularly valuable to those attendees who haven’t rethought lubricant selection in years and/or may be rusty on the proper oil analysis program to warn of varnish formation and other problems.

He began by reviewing the differ-ent types of base oils and additives that are mixed to create a finished lubricant, pointing out that the properties and quality of the base oil have a significant impact on thermal and oxidative stability, viscosity, and oil consumption.

Additives enhance base-oil lubricating qualities and lifetime. For example, oxidation inhibitors increase lubricant life, rust inhibitors prevent corrosion, foam inhibitors reduce foaming tendencies, and anti-wear additives increase load carrying capability.

Consistent and correct sampling

procedures are essential and should be performed while the control and bearing oil systems are in operation. Important considerations include sample location, sampling hardware, bottle cleanliness and flushing. The following sampling locations are listed in order of preference, but equipment configuration can limit location choices:

  • Bearing header return lines to the sump.
  • Sump dip sample.
  • LO pump discharge and upstream of system filters.
  • Sump bottom drain (as a last resort), but remember to flush well. It could take more than a barrel of oil to pull a representative sample.

LO analysis is a tool used to predict the performance of new or in-service oils. It should also be used to determine compatibility with other lubricating oils in top-add applications. Bear in mind that LO analysis provides a snapshot of certain measurable parameters; analysis intervals may be monthly, quarterly, and yearly on a test-by-test basis (sidebar). Data trends offer the most insight on equipment and lubricant performance.

Often, the most valuable and timely information on LO condition is right in your hand at the time of sampling. Don’t miss this opportunity to assess key performance parameters. The use of clear, clean sample containers permits quick and easy quality checks as identified below:

  • Color: Unusual and rapid darkening can indicate contamination or excessive degradation.
  • Odor: Sour-smelling oil also can indicate contamination or excessive degradation.
  • Air entrainment: Air bubbles in the body of the LO sample should clear in five minutes.
  • Foam: After a vigorous shake, foam from the surface should clear in 10 minutes.
  • Water: Sample should be transparent. If you cannot read printing through a clear sample container, then water levels above 300 ppm may be present.
  • Solids: Look for solids settling out as signs of external and internal contamination.

Several key tests used to assess LO condition are invaluable in preparing for your next outage (sidebar). Results of these tests will identify off-line action—such as oil replacement, addition of special additives, etc—that may be necessary to ensure reliable operation of your rotating equipment and associated hydraulic control systems.

Experts generally suggest that you monitor these parameters monthly, or at least quarterly:

  • Viscosity, ASTM D445. For more information on the American Society for Testing & Materials and its standards, access www. astm.org.
  • Water by Karl Fischer Titration, ASTM D1744.
  • Acid Number, ASTM D974.
  • ISO Cleanliness Code 4406. For more information on the International Organization for Standardization and its objectives, access www.iso.org.
  • CP Metals. This is the inductively coupled plasma technique for the determination of trace metals in solution.
  • PQ™ Index.

Annually, check your oil’s suitability for continued use by conducting these tests:

  • Viscosity, ASTM D445.
  • Rotation Pressure Vessel Oxidation Test, ASTM D2272.
  • Water by Karl Fischer Titration, ASTM D1744.
  • Acid Number, ASTM D974.
  • ISO Cleanliness Code 4406.
  • Rust, ASTM D665 A.
  • Demulsibility, ASTM D1401.
  • Foam, ASTM D892 Sequence 2.
  • ICP Metals.

Sharkowicz next recommended several “best practices” regarding LO systems, including the following:

  • Keep the circulation system on during shutdown.
  • Keep oil above 100F during shutdown.
  • Optimize cooling systems to ensure that oil is not overstressed thermally.
  • Run hydraulic control circuits through full range of motion possible—on shutdown and prior to start up, to flush out any debris.

Frame 6 engine O&M

Thumbnails of key lube-oil tests

Viscosity, ASTM D445

Viscosity is, perhaps, the most important characteristic of turbine oil because of the tight clearances in journal and thrust bearings. Changes in viscosity can result in unwanted repositioning of the rotor, both axially and radially.

Rule of thumb: Unless the oil has been contaminated or severely oxidized, viscosity should remain consistent over years of service. ASTM D4378-97 proposes a 5% change from the initial viscosity as a warning limit.

Rotation Pressure Vessel Oxidation Test (RPVOT), ASTM D2272

RPVOT warns of a loss in LO oxidation stability. Oxidation is driven by heat and exposure to contaminants such as water. As a turbine oil degrades, it forms weak organic acids and insoluble oxidation products that adhere to governor parts, bearing surfaces, etc. Severe oxidation is conducive to the formation of varnish on hot bearing surfaces that retard heat transfer and can overheat heat journals. Severely oxidized oils also can foul turbine control elements.

The accelerated oxidation test is an industry standard for identifying oxidation stability problems with in-service turbine oils. ASTM D4378–97, “Standard Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines,” identifies an RPVOT decline to 25% of the initial new-oil RPVOT value—coupled with an increase in Acid Number (AN)—as a warning limit. Many turbine OEMs simplify this by using the 25% of initial RPVOT without reference to AN. Some OEMs also specify a 100-minute minimum RPVOT. Waiting for an increase in AN could present the risk of turbine bearing seizure if the turbine oil cannot be replaced in a timely manner.

Fourier Transform Infrared Spectroscopy (FTIR) is an alternative analytical technique used to identify organics— such as organic acids formed by oxidation of LO. Efforts to revitalize severely oxidized turbine oil with oxidation inhibitor can put equipment at risk. Oil that has a RPVOT value below the OEM’s 100-minute minimum more than likely has lost its inherent base stock oxidation stability, making additive addition an impractical solution. Additives can temporarily boost RPVOT, but the diminished nature of the base stock may sharply reduce the time it would normally take to form heavy varnishes and sludges.

Water by Karl Fischer Titration, ASTM D1744

Testing for water is important for minimizing the risk of not detecting turbine-oil oxidation and rust formation. Excessive water will alter an oil’s viscosity (up or down depending on conditions). Studies also warn that water levels above 250 ppm in hydrogen-cooled generator windings may lead to stress corrosion cracking of generator rotor retaining rings. Water in turbine oil in warm storage tanks can promote the spread of microbial growth that will foul system filters and small-diameter gauge and transducer line extensions.

ASTM D4378-97 identifies 1000 ppm or 0.1% water as a warning level; however, some OEMs specify 500 ppm. In hydrogen-cooled generators, an upper limit of 250 ppm should be maintained.

Acid and Base Number by Color- Indicator Titration, ASTM D974-02; Acid Number (AN), ASTM D664

Sharp increases in AN may indicate contamination or a severely oxidized oil. Organic acids formed by oxidation can corrode bearing surfaces and should be addressed in a timely manner. ASTM D4378-97 offers guidelines of 0.3 to 0.4 mg KOH/g above the initial value as an upper warning level. Many oil analysts view an upward movement in AN as small as 0.1 as worthy of concern.
Dark-colored oils that cannot be analyzed by D974 because the color-indicator end point is obscured can be analyzed by D664.

ISO Cleanliness Code 4406

Turbine journal bearing clearances and hydraulic servo valve clearances dictate the need for clean oil. Excessive bearing wear and servo valve sticking can result if tight cleanliness standards are not maintained. An OEM average turbine oil cleanliness level is ISO 18/16/13 or a National Aerospace Standard (NAS) 1638 cleanliness level of 7 is desirable. The three-range number ISO cleanliness code correlates to concentrations of particles larger than 4, 6, and 14 microns. Turbine OEMs offer specific guidelines on recommended cleanliness levels.

Rust, ASTM D665 A

Rust particles act as oxidation catalysts and can cause abrasive wear in journal bearings. Rust inhibitors are normally kept at proper levels through makeup and can plateout on metal surfaces for added rust protection. Rust inhibitors can impact water separation so field introduction of such additives generally is not recommended. In-service oil testing should be conducted with distilled water as identified in D665 A, not synthetic seawater (D665 B). ASTM D4378-97 considers a “light fail” as a warning limit.

Demulsibility, ASTM D1401

Water shedding characteristics are important to LO systems that have had direct contact with water. The ability to separate water by natural density difference and remove it through bottom drains will improve a turbine oil’s oxidation stability. Demulsibility can be compromised by excessive water contamination.

ASTM does not offer warning limits for demulsibility.Some turbine OEMs identify levels of 3-ml emulsion after 30 minutes on new oils. In-service oil warning limits of 15 ml or greater of emulsion in 30 minutes should serve as a fair warning limit. The impact of demulsibility depends on the system residence time and anticipated levels of water contamination. LO demulsibility can show failure in the lab, but with sufficient residence time, the system oil may shed water at a rate that does not impact turbine oil performance. Small sumps with lower residence times require better demulsibility performance than larger sumps.

Foam, ASTM D892 Sequence 2

A turbine oil sample may test for foam at a higher level than suggested by the OEM. But this typically presents no field foaming issues because of the low position of the LO pump suction. If the foam level in the turbine sump is six inches or less and does not overflow the sump or cause level monitoring errors, the turbine oil foam should not be a concern. LO at the turbine sump surface should show at least one clear area (no bubbles) and larger breaking bubbles should be seen at this interface.

ASTM D4378-97 offers warning limits of tendency 450 ml with a stability of 10 ml. Foam tendency is the foam volume measured in a graduated cylinder after five minutes of pushing air through the LO sample. Stability represents the volume amount after 10 minutes of settling time has elapsed. A foam stability of less than 5 ml is a good indication that foam bubbles are breaking and the turbine should not experience operational problems from foaming.

When addressing foam problems, cleanliness, contamination, or mechanical causes should be investigated before adding antifoaming agents in the field. Excessive additive addition can result in an even greater problem with increased air entrainment. Dirt is a leading cause of foam, so ISO cleanliness test should be conducted. Keep in mind that testing for foam should be conducted only when foaming presents an operational problem and for product compatibility testing.

PQ™ Index

The PQ is a sensitive magnetometer that measures the mass of ferrous wear debris in a sample and displays this as a PQ Index. This unit-less number is quantitative and can be trended with acceptable linearity over a wide range of ferrous debris concentrations and particle sizes. Measurement does not require that LO be removed from the sample bottle.

GE Energy’s Dr Roointon Pavri had conducted engine workshops for the Frame 6 Users Group in past yearsand done such good job that he was invited back yet again. Don’t be fooled by the “Dr” in front of Pavri’s name; he’s not an egghead, far from it. Pavri has intimate design knowledge of the Frame 6; plus, he understands how the machine is operated in the real world of power generation and what the needs of the user community are in terms of O&M guidance. Pavri was asked by the steering committee to review the machine’s performance history and to address the following subjects:

  • Wheel-space temperature.
  • Black-start capability.
  • Performance degradation and recovery.
  • Maintenance considerations/clearances.

Performance history. The Frame 6 has been in production for 28 years, starting life in 1978 as the MS 6431A, a 31-MW machine with a heat rate of 11,220 Btu/kWh (ISO conditions at base load on distillate oil with no inlet or exhaust losses). Exhaust temperature was 891F.

The 6B represented a significant design upgrade. It launched in 1981 as the MS6521B, capable of 36.7 MW and a heat rate of 11,120 Btu/kWh. Firing temperature of this engine was 2020F, an increase of 170 deg F over the 6A; exhaust was 1017F.

The current production machine is the PG6581B, which debuted in 2000, at 41.5 MW and 10,724 Btu/ kWh. Firing temperature is 2084F, exhaust 1016F. With almost 1000 units sold worldwide, the 6B is one of the most popular frames ever built.

Pavri compared the design parameters for the 6581 to those for several earlier models (6541, 6551, and 6561) introduced after 1986 and the machines attendees were most familiar with.

Next he addressed a user question that concerned upgrading a DLN 1 combustion system to DLN 1+ to reduce emissions. Pavri, who welcomes questions during his presentations, said going from DLN 1 to 1+ was a “leap,” requiring new transitions, liners, and fuel system. To his knowledge at the time such a conversion had not been done on a 6B but one or two had been done for the 7EA fleet. GE’s Rick Romero, also in attendance, said that the first 7EA converted to the DLN 1+ was running at 4.5 ppm NOx with single-digit CO.

Wheel-space temperature was of considerable interest to the group. Pavri began by saying that 99% of the time he hears of high wheel-space temperature, the perceived problem can be traced to improper installation or positioning of a thermocouple (TC). He pointed out that differences in readings between two TCs of say 50 deg F is impossible because cooling air and hot gas are rotating at about 30% of the shaft speed and the mixture essentially is homogeneous. But if both TCs are indicating high temperatures, then there could be a problem.

What did Pavri say to do if there’s a large temperature difference between two TCs: Change the high one or average the two, but don’t get upset. He put a couple of cutaway diagrams up on the screen to show how to install TCs to ensure accurate readings. If the TCs are not properly seated and they back out a fraction of an inch, he added, you will get a high reading. Also, make sure that you’re using the correct TC for this application.

Continuing, Pavri said that if you’ve had a high temperature since startup, there probably is less cooling air flowing into the wheel space than is required. If the temperature goes up gradually over time, seal wear may be allowing more hot gas into the space than there should be. Another possible reason: bucket creep. A sudden rapid increase in temperature could be caused by seal damage— such as a broken angel wing.

Black-start capability is of particular interest today because of (1) last-year’s storm experiences on the Gulf Coast and (2) global terrorist activities. Pavri reviewed setups for both diesel and motor starting. When considering diesel starting, be sure that auxiliaries requiring ac power are supplied separately. Otherwise, you’ll need a nominal 250- kW diesel/generator set with deadbus closure capability to pick up the auxiliary load. Here are some points to keep in mind:

  • Fuel. If gas is not available at proper pressure a separate compressor is required.
  • Ignition power, spark plugs. An uninterruptible power supply (UPS), which converts battery supplied dc power to ac, normally is available.
  • LO pumps. The dc emergency pump should be adequate.
  • LO cooling. Cooling-water pumps must be energized.
  • Diesel engine cooling. Cooling water must be provided to the diesel engine.
  • Cooling-water fans. Power must be supplied to these fans.

For black start by motor, all starting power must be supplied by a separate diesel/generator with dead bus closure capability. This typically demands a 1-MW set.

Performance degradation impacts both revenue and expense. The more performance degrades, the less power there is to sell. Also, as efficiency decreases, more fuel must be purchased to produce a given number of kilowatt-hours.

Pavri focused mainly on the compressor, which he said accounts for 70% to 85% of the power loss attributed to engine performance degradation. Air flow and efficiency drop because of corrosion, erosion, FOD (foreign object damage), rubs, fouling, and surface roughness. The first four are non-recoverable losses, last two are recoverable. Recoverable losses are defined as those that can be corrected without changing out parts.

Fouling is found primarily in the front stages of the machine because of the moisture present and it impacts air flow more than efficiency. Fouling is caused by airborne contaminants such as dust, sand, salts, hydrocarbons, pollen, insects, industrial chemicals, etc. Moisture comes from humidity, rain, fog, and the evaporative cooler if installed.

Compressor efficiency is influenced most by the condition of the machine’s later stages and units in base-load service have cleaner back stages than units that cycle. The amount of corrosion-product buildup in cycling machines depends in part on ambient conditions.

GE’s plan for minimizing performance degradation includes proper selection of (1) inlet air filters and (2) coatings, materials, and surface finishes for compressor blades, and (3) siting and orientation of air inlet louvers to avoid fouling from coal or other dusts, cooling-tower plumes, etc. Plant personnel must continually monitor performance parameters, corrected for ambient conditions, to advise when corrective action—such as filter replacement, online and offline compressor cleaning, etc—is necessary.

To read more on the subject turn to “Assessing GT performance degradation,” p OH-94, 2007 Outage Handbook supplement inserted in the middle of this issue.

The maintenance considerations/ clearances portion of Pavri’s presentation covered step-by-step the clearances that should be measured with precision during hot-gas-path and major inspections. By the time the second major rolls around, he said, clearances are virtually impossible to measure wi th precision because of casing distortion, foundation shifts, parts not repaired to spec, etc. Closing clearances must be evaluated on an individual basis against the next planned outage, history, etc.

Inclosing, Pavri reviewed quickly all of the TILs (Technical Information Letters) issued for the Frame 6B in 2005.

Compressor cleaning

Pavri’s comments on performance loss were the perfect introduction for Hugh Sales, a VP for Houstonbased Gas Turbine Efficiency, which specializes in compressor washing. Sales conducted a tutorial on the process variables that have greatest impact on washing efficiency, providing practical information for plant-level application.

The optimization variables for a water wash system that you should be familiar with are these:

  • Water temperature. Water that’s too cold is conducive to blade temperature variation, loss of cleaning efficacy, and the increased need for chemicals. Water that’s too hot leads to increased early vaporization and loss of fluid penetration. Plus, it presents a safety risk to operators. The “just right” temperature, Sales said, was about 140F.
  • Water pressure is optimum within the range of 750 to 1400 psig. The designer’s goal is to accelerate water to approximately airflow speed, not to increase blade impact. Water pressure within the range cited provides maximum dispersion across IGVs and the first stage, assures maximum penetration into the compressor core, minimizes the waste of water and chemicals (if used), and ensures penetration of the air-flow boundary layer on blades and vanes.
  • Droplet size. Sales said this was the most critical variable in terms of cleaning penetration and erosion (Fig 15). If droplet size is too small—less than 40 microns according to Sales—water never reaches the blade surface, it merely follows air flow and no washing is accomplished. If it is too large—over 180 microns—the droplets may damage blades and inertial forces that throw water to the outer case may occur.
  • Water volume depends on the size of the turbine. Sales quoted research and field experience that concluded the following: For many turbines, only 20% to 40% of the water typically used by an OEM’s system really is required for an optimal compressor wash—that is, assuming all other variables are optimized.
  • Nozzle design and placement. Spray pattern for a nozzle should be engineered after considering the other variables being optimized. If true optimization is accomplished, Sales said, you only need one set of nozzles to accomplish both online and offline washes. Also, you will require far fewer nozzles (perhaps only a third as many) than others might say you need, he continued. Typical retrofit applications, Sales added, normally call for replacing existing OEM offline nozzles with those capable of both online and offline washing.

Get Shidler (or Ginder)

Next time you and your colleagues are sitting around the break room with nothing to do, try a game of charades. It’s probably easy to get your partner to say “borescope” with a minimum of acting. But that’s not the word you want. You’re looking for the name of a person/company that does the visual inspection.

Since most plants have borescopes, you may hear words like Olympus or Karl Storz next. Both are well-known manufacturers. Keep digging. Chances are the next word you’re going to hear is “Shidler” or “Ginder,” that’s how well-known Rod Shidler and Rick Ginder are known in GT-based powerplants. Game over. But ask this question, too: “What’s the name of their company?” Chances are few, if any, in the break room will know it is Advanced Turbine Support Inc (ATS).

A successful borescope inspection requires both quality equipment and a technician who knows intimately the engine he or she is inspecting. The financial benefit of a complete and accurate borescope inspection to the owner/operator, began Shidler in his presentation before the Frame 6 Users Group: It allows condition based maintenance, which typically permits extension of service intervals specified for traditional hours- or starts-based maintenance programs.

An example Shidler offered on the value of bores coping for maintenance decision-making concerned inspections of two side-by-side units. Only one GT was scheduled for maintenance during the current outage; decision was based on equivalent starts and hours.

But that unit looked far better internally than the one scheduled for overhaul during the next outage, six months away. That the second GT would even make it to the next outage without serious damage was called to question. Plant management opted for condition-based maintenance, opening the machine in poorest condition and overhauling it; the unit in better “health” was rescheduled for its overhaul during the next outage.

This decision is what most would call a “no-brainer.” Perhaps it was. But the point of the case history was that if condition-based maintenance made sense in this situation, why not always?

A user asked Shidler why the plant didn’t know about the damage to the one unit before he got there. His response: Industrial frames are so rugged that major blade damage can go undetected by vibration monitoring and you wouldn’t know anything until a catastrophic failure—certainly not a comforting thought for those in attendance. For example, he continued, tip rubs can initiate radial tip cracks and the pieces liberated nominally can measure an inch by an inch—large enough to do downstream damage but not of sufficient mass to influence vibration readings.

Shidler spent about half of his time in front of the group showing the types of wear and tear that you can identify by borescoping the engine from the compressor IGVs (inlet guide vanes) through the last turbine stage (montage). These photos were selected from the ATS archives and include images from several different inspections involving several engine models. He also ran through the final report the company prepared for a 6B inspection which illustrates the value of periodic visual condition monitoring.

For more on the subject, read “Timely borescope inspection prevents turbine damage,” p OH-89, 2007 Outage Handbook supplement inserted in the middle of this issue.

Generator (system) testing

John Estes Jr, president, E2 Power Systems Inc, Littleton, Colo, had a great deal to say on the testing of generator systems, but it was tough for a mechanical engineer to keep up with his 78 slides in the hour time slot.

The beginning was easy enough. Estes encouraged users to develop a routine inspection and preventive maintenance (PM) program that would be implemented every 18 to 24 months during a regularly scheduled GT maintenance outage. Very important to keep good records, he said, to guide follow-on activities. These records should include operating events, as well, to aid in rootcause analysis when necessary. Such events might be relay trips, short-circuit faults, lightning strikes, full-load rejection, major VAR swings, unexplained alarms, etc.

Estes told attendees that their inspection/PM programs should have three parts. The first is a list of what to inspect/service. It took him three slides to present the more than two dozen items he thought should be included—from control logic to voltage regulator. So much for the belief of many that generators are relatively simple rotating machines requiring a minimum of attention. Nothing could be further from the truth. Listen to Estes for an hour and you could develop insomnia.

Second part is a list of what to include in your toolbox. In his, Estes had clamp-on ammeters, megger, oscilloscope, relay test set, dc power supply, cleaning supplies, and several other items. Third part of your plan—very important to reliability—should be how to identify the telltale signs of impending problems with an initial walk-down inspection using your eyes, nose and some “helpers”—such as mirror, borescope, and digital camera. Loose wires, rust, water leaks, oil leaks, excess heating, brush chatter, and the presence of wildlife (such as snakes, raccoons, squirrels) are some of the things to look for.

Program in hand, it’s time to begin work. Estes started with the generator field and suggested insulation testing with a 500-V megger to determine the integrity of the windings which will alert you to possible moisture and contamination problems. Then, check the resistance of fieldwinding copper with a digital lowresistance ohmmeter to identify the presence of bad connections. Verifying the condition of brush rigging and collector rings was next on the list, followed by a search for any shorted turns. Clues that shorted turns exist include higher-than-normal field amps to attain rated voltage, vibration, etc.

For the generator stator, be sure to do the following:

  • Test insulation with a 5000-V megger.
  • Check the resistance of statorwinding copper.
  • Inspect end windings for cleanliness, overheating, damage, etc.
  • Verify proper operation of generator heaters.

Estes covered in rapid-fire order the high-voltage neutral and main busses, neutral grounding, generator auxiliary compartment, lightning arrestors/surge capacitors, generator breaker, potential transformers, power potential transformer (excitation power source), protective relays, generator RTDs, meters, transducers, synchronization, alarms/annunciators, control sequencing, trip tests, fuse sizing, excitation/voltage regulators, and shaft grounding brush. He concluded with an overview of NERC (North American Electric Reliability Council) testing.

Third-party repairs

Hans van Esch of Turbine End-user Services Inc (TEServices), Houston, last of the invited speakers on the program, spoke to his area of expertise— metallurgical aspects of industrial GT component repair, focusing on third-party repair evaluation. Van Esch is a frequent contributor to the COMBINED CYCLE Journal, and the main points of his presentation are captured in these four recent articles, which are available at www. psimedia.info/ccjarchives.htm:

  • “Six steps to successful repair of GT components,” 2Q/2005, p 25.
  • “Selecting the appropriate vendor to refurbish parts for your turbines,” 2006 Outage Handbook supplement to 3Q/2005, p OH-7.
  • “Shops verify as-received condition of components with an ‘incoming’ inspection,” 4Q/2005, p 56.
  • “Verifying repairs, final inspection results,” 1Q/2006, p 50. ccj