Changing the way the Power Generation Industry buys parts.

CTOTF: NDE, ‘met’ analysis, engineering study critical for true rotor life assessment

If you read the Best Practices Awards section in the 1Q/2008 issue you know that John Love­lace retired as CTOTF chair after the Spring Turbine Forum in Savannah and was succeeded by Vice Chair Bob Kirn of the Tennessee Valley Author­ity.

If you read the first article in this issue you also know that the Combustion Tur­bine Operations Task Force has been reorganized and expanded in scope (p 2). Kirn announced this at the Fall Turbine Forum, held in Tuc­son, September 14-18.

What may be new to you is that Lovelace retired from Arizona Public Service Co (APS) in mid September and has reinvented himself as an independent turbine consul­tant (contact details to follow in the next issue after the office furniture is delivered). He won’t get bored waiting for the job that will jump-start his new career: Lovelace was recently appointed to the CCJ Editorial Advi­sory Board, along with powerplant water expert Dr Barry Dooley of Structural Integrity Associates Inc.

In addition, he volunteered to serve as curator for the fledgling Gas Turbine Historical Society (see sidebar in Dave Lucier’s “These baby boomers also deferring retirement,” elsewhere in this issue).

If you were unable to attend the Savannah meeting you missed several very meaningful presenta­tions and discussions. Summarized here are some important develop­ments from non-OEM experts for owner/operators of large frames—the GE 7F in particular.

The GE Roundtable section, orga­nized by John Gamble (he retired from TVA in June) and Dominion Energy’s Larry Rose, couldn’t have been better. Scheduled from 8 to 5, it ran nearly an hour long and only stopped then because participants were hungry. In sum, there were a dozen presentations by services pro­viders and users.

The Legacy Roundtable, which is allocated only four hours of program time at CTOTF meetings because there’s so much to pack into the four-day meeting, focused rotor life man­agement. Chair Steve Hedge of NRG Energy and Vice Chair Eddie Mims of Colectric Partners, with help from APS’ Scott Takinen and Lovelace, pulled out all the stops to give users the confidence to challenge OEM end-of-life limits in terms of starts and hours. Technical presentations by qualified experts and the utility case history profiled lend encourage­ment.

GE Roundtable

Chair: John Gamble, Tennessee Val­ley Authority

Vice Chair: Larry Rose, Dominion Energy

To arrive at a meaningful consensus on 7FA com­pressor issues, you’d probably have to interview scores of users. A far simpler approach is to call Advanced Turbine Support Inc (ATS), Gainesville, Fla, and ask Rick Ginder and Rod Shidler to summarize the results of their inspections. That’s what Gam­ble and Rose did for CTOTF’s GE Roundtable last spring.

ATS does hundreds of com­pressor borescope inspections annually and Ginder and Shi­dler know what most—if not all—the problems are, where to look for them, and what NDE (nondestructive exami­nation) tools should be used to confirm the presence of criti­cal indications. In addition, the two engine sleuths have a good feel for what the wear and tear found typi­cally means in terms of operational risk—particularly valuable informa­tion for merchant plants with “must-run” contracts.

At the Spring Turbine Forum and Trade Show, ATS presented a com­prehensive inspection plan for iden­tifying and managing S0-S5 stator-vane issues. First thing Shidler said was that a standard borescope inspec­tion isn’t intended to identify all the issues. That sent a chill through the audience of about 50 alert users.

In Shidler’s view, a comprehen­sive S0-S5 inspection, which takes about 12 hours, has two components: a high-resolution borescope examina­tion and in-situ NDE. This approach generally exceeds current insurance requirements so at least some would ask, “Why do it?” The answer is sim­ple: It makes good business sense. Over time, unit availability will be higher and maintenance costs lower.

Early results confirm this logic: By “stacking” inspection methods you don’t raise alarms and take a machine out of service for an indica­tion that looks like a crack but isn’t. Also, you’ll find indications hereto­fore missed that could result later in vane liberation and extensive down­stream damage.

Of the dozen or so units that had completed ATS comprehensive inspections prior to the spring meet­ing, Shidler and Ginder found two with cracks in stator vanes. For one machine, impact damage was the cause; for the other, it may have been high-cycle fatigue (HCF).

What Ginder considered “disturb­ing” about the former was that the crack appeared to have been initiated at a very small impact site—so small it’s virtually impossible to see in print which is why a photo is not included here. The engine had approximately 4500 hours of service and 350 starts when the indication was discovered. Serial numbers of the vanes were non-sequential with those identified with other failures.

Keep in mind that if you lose part of a stator vane in the first or second row you’re probably looking at a new compressor and an invoice asking for perhaps as much as $20 million. This is good to remember if your vanes are found in poor condition and an expert suggests replacing the lot. At $6 mil­lion it may be a bargain.

The value of a high-res bore­scope is that it can identify pitting from corrosion on or near the leading edges of S0 to S5 vanes that can’t be found with a standard instrument (Fig 1). Such pitting is closely linked to vane failure. Ginder said that it’s possible to borescope approximately 95% of the S0-S5 stator vanes.

The leading edges of S0 through S3 stator vanes (together with the first inch or so of airfoil surface on the suction side) are inspected using the latest eddy-current technology as well. S0 and S1 vanes, those particu­larly sensitive to even the smallest of pits and thinnest of hairline surface cracks, are checked 100% (Fig 2). Any crack indications are validated with dye penetrant.

S2 and S3 vanes are examined by eddy current based on results of the high-res borescope inspection (Fig 3). Typically, Shidler recommends eddy-current examination of about 50% of the S2 vanes, 30% of S3, because of the difficulty in accessing 100% of the vanes.

Ginder offered a couple of com­ments on pitting:

• If you find several connected pits in a row on one or more of your vanes—commonly referred to as in-line pitting—take immediate corrective action. In some cases vane replacement may be war­ranted.
• Surface pits sometimes are not as small as they appear. Many “cone out” as they go deeper. This means that blending onsite probably will not be successful. One example given: The OEM was prepared to dress vanes onsite and wound up replacing entire ring segments. At another plant, a pit was found within a pit.
• If your GT must be opened up for other work, don’t expect that someone can do a casual visual check for pitting with any degree of reliability. With pits of even minute size important, eye strain is debilitating without an aide such as dye penetrant—at a mini­mum.
• Horror story: Gross pitting on the trailing edge of an unflared vane resulted in liberation of the top 4 in. of the airfoil. Investigators believed that the cooling tower located just upstream of the com­pressor inlet, which used chlori­nated potable water as the heat-transfer medium, was the source of the problem.

Shidler talked about cracking. He said 7FA stator vanes appear very susceptible to cracking and that it doesn’t seem to take much of a stress riser to initiate failure. Shidler suggested that if blending of cracks is required on stator vanes located in the first few rows of the compressor, you should check the repaired airfoils within 25 starts. No matter how good the blending is, he continued, you have created an “area of susceptibil­ity” and cracks may initiate there.

Someone asked a question on HCF failures. Ginder began by saying as many as two dozen stator-vane fail­ures may have been attributed to HCF over the years. He suggested that the underlying cause might be long-term corrosion-product build-up that locks vanes into the ring segments that characterize rows S0 through S4. Such “lock-up,” he believes causes HCF near the tips of the vanes.

Shim warning. It seems you can’t have a conversation on Frame 7 compressors without mentioning shims. A user asked something like, “What are you seeing with respect to shims?” Shidler said for some units ATS is seeing shims migrate into the flow stream faster than he and Ginder would have expected based on years of inspection experience. One example: Shim protruding into the flow stream about 1/8 in.; next inspection, 27 starts later, it was out 1 in. In another case, the shim was “sticking out a bit” during a routine borescope inspection; 170 starts later it liberated.

Finally, an attendee mentioned “ring of fire” and that ignited anoth­er tangential dis­cussion. Recall that some in the industry attribute early-stage com­pressor corrosion problems to plant location—specif­ically to plants sited along the Gulf Coast from Texas through Florida. These units ingest humid salt-laden air which, of course, is cor­rosive.

One user barked, “That’s not entirely true. We have a couple of units in the Mid-Atlantic region that are three hours from saltwater and they suffer from similar corrosion.”

Another owner/operator added the following: “About an eight-hour drive north and 15 miles inland we have two units with S2 and S3 dis­tress after only 2000 hours of opera­tion. They sit across the fence from two E-class GTs supplied by another OEM. No problem with those.” Group consensus: The ring-of-fire root-cause argument cannot be supported with the facts available.

Vane pinning boosts GT reliability

Shidler’s warning on shim migration was the perfect segue to a presenta­tion on compressor vane pinning by Rodger Anderson, manager of gas-turbine (GT) technology for DRS-Power Technology Inc, Schenectady, NY. A former GT design engineer for what today is called GE Energy, he has spent the last several years analyzing compressor failures in the 7EA and 7FA engines and developing independent mitigation solutions.

Anderson may be known best by users for the stator-vane pinning technique he developed for both tak­ing the “play” out of loose square-base compressor blades found downstream of R4 in Frame 7s and for preventing the migration into the flow path of shims used to fill out vane rows.

For those unfamiliar with vane pinning and wanting to know more, access www.combinedcyclejournal.com/archives.html, click 4Q/2004, click 7EA Users on the issue cover, scroll to “Compressor blade issues.” Next, click 2Q/2006, click CTOTF, scroll to “GE Roundtable. . . .”

Anderson began by reviewing the dominant causes of compressor failures; they are listed below. Note that the first three typically are prob­lematic only if the affected blades and/or vanes are stimulated on reso­nance.

• Manufacturing defects and poor quality control.
• Foreign object damage (FOD).
• Erosion, corrosion, and other prob­lems linked to the operating envi­ronment.
• Operational issues—such as rotat­ing stalls and flow surges.
• Severe vane attachment wear leading to liberation or to vari­able spacing between the station­ary vanes and adjacent rotating blades. Latter results in blade excitation.
• Protruding shims, which excite rotating blades.
• Liberated shims, which damage downstream blades and vanes on impact.

Anderson then ran through a series of photos illustrating how the square bases for vanes downstream of stator row 4 (S4) can wear at the corners and cause casing groove wear, how heavy casing-groove wear results in blade looseness and rock­ing, and the now you see it/now you don’t S15 vane that went missing. Not pretty pictures, but very valu­able for showing users what kinds of damage to look for, and where, dur­ing planned maintenance outages.

Next, he put up an interesting graph showing typical vane-tip axial movement. The OEM limit (slightly more than a tenth of an inch) was compared to actual measurements before and after repair. According to Anderson’s data, vane-tip axial move­ment typically is about half the OEM limit, but he has found vanes that exceed the OEM limit by as much as one-third. After repair, move­ment is less than 5% of the limit. Another point worth noting: Vanes in the upper casing half (UH) typically move three times the distance at the tip than vanes in the lower casing half (LH).

More instructional photos fol­lowed: 7FB S17 vane failure after only about 100 hours of service, R0 leading-edge erosion, collateral com­pressor damage costing about $5-mil­lion to repair, protruding and bro­ken shims, rotor-blade leading-edge impact damage (possibly from a liber­ated shim), etc.

When inspecting your com­pressor sections, keep in mind that most missing and protruding shims are found near the horizontal joint at the LH left side looking downstream and UH right side. Anderson said that the counterclockwise rotation of the rotor produces aerodynamic forc­es that press the LH vanes against the horizontal joint on the right side and the UH vanes against the hori­zontal joint on the left side. Keeper bars in the UH at the joint prevent the vanes from sliding in the casing groove.

Compaction of the vanes at the horizontal joint (LH, right side; UH, left side) tightly pinch and generally hold the shims; however, on the oppo­site sides of both casing halves, the vanes and shims are loose. As shims wear and liberate the vanes loosen more, increasing the aerodynamic stimulus on nearby rotor blades.

Anderson said shims don’t just release into the flow path without warning. Rather, they slowly work their way out from between the indi­vidual square platforms supporting vanes downstream of S4. The migra­tion process depends on many fac­tors—such as number of engine stop/start cycles—and it may be a year or more from the time a borescope examination first identifies a pro­truding shim and it releases (assum­ing no corrective action has been taken). Note that shims can protrude nearly three-quarters of an inch into the flow stream before liberating.

While protruding shims do not damage by colliding with downstream components, they can produce poten­tially damaging flow disturbances at the tips of the downstream rotor blades depending on how much they stick out above the vane platform (Fig 4). Anderson pointed out that because shims are concentrated at the horizontal joints, protruding shims will stimulate rotor blades once or twice each revolution.

DRS commissioned a detailed har­monic analysis, starting with a finite element model to predict the natural frequencies and mode shapes for a typical R15 7FA rotor blade. The first tangential mode was found to be near 480 Hz—either eight or four times a once or twice per revolution stimulus from protruding shims. Anderson presented the results—right down to Campbell diagrams. Highlights:

• There is a very high probability of a single shim protruding halfway into the flow stream.
• There is a moderate-to-high prob­ability of two shims protruding halfway into the flow stream because of vane-compaction loca­tions 180 deg apart.
• Calculations indicate that a shim protruding for from six to 10 hours can produce stress cycles that exceed the endur­ance limit of some blades. Thus any potential blade stimulus identified dur­ing an inspection should be eliminated.

Oftentimes protrud­ing shims can be pulled out or ground off remote­ly by a specially quali­fied borescope inspection team. This obviously doesn’t solve the problem but a must-run unit can be returned to service until a planned shutdown is possible.

• Since the material and mechanical strengths of operating blades will be less than those for a smooth bar in air (basis for the investigation), there is a moderate probability that early 7FAs will experience high-cycle fatigue failures of rotor blades because of vane looseness, shim protrusion, FOD, and corro­sion pitting.

Most of this risk can be miti­gated by pinning, according to Ander­son. He offered these suggestions:

• Pin together forward-stage blade and ring assemblies to stabilize the end vanes in the ring segment and to prevent shims from pro­truding and liberating. Another idea is to cut these assemblies into smaller segments and then pin them together. The advantage of this approach is that the smaller segments are easier to remove than the OEM’s segments and plant staff is more likely to replace pitted, worn, and lower-efficiency blades as wear and tear becomes evident.

Also important: The two-stripe vibration mode stimulus from pro­truding shims in S0 through S4 is eliminated with pinned segments. Plus, stabilized vanes have less position variability to cause rotor-blade stimulation.

• Locking in place mid-stage shims eliminates the possibility of down­stream collateral damage as well as high-cycle fatigue fail­ures of downstream rotor blades.

Anderson claimed that pinning improves compres­sor section reliability and availability because com­pressor vanes are main­tained in the as-built con­dition. To date, DRS has pinned vanes in more than 50 7EA and 7FA compres­sors—almost equal numbers of each.

The 7EAs usually exhibit severe vane/casing wear in S5 and high numbers of protruding or missing shims in other downstream rows.

The 7FAs typically are charac­terized by either severe vane/casing wear beginning at S13 and/or pro­truding or missing shims in other upstream rows—including the blade and ring-segment rows S0 through S4. Many of these engines had severe damage downstream of the shim lib­eration locations (Fig 5).

Stator-vane repair options

Vane pinning described in the previ­ous section is one solution for dealing with hook-fit wear and migrating shims in compressor sections. There are others. John Yelincic and Fred Price of Allied Power Group, Hous­ton, offered several options available to owner/operators of Frame 7F GTs.

Yelincic, who made the presenta­tion, began by showing attendees the nature of the damage they were deal­ing with: hook-fit wear, which if not addressed could end with liberation of an airfoil (Fig 6); so-called butt-face wear caused by the “working” of adja­cent blades in contact with one another (Fig 7); and shim liberation (Fig 8).

One reason for the accelerated wear and tear in 7F stator vane rows downstream of S4 is that individual vanes with straight-land hook-fits are inserted into a radial-cut groove. Yelincic said that careful analysis shows that there are only three theo­retical lines of contact between each side of the vane base and the com­pressor case—and each line of con­tact is only 0.15 in. long.

Wear and tear can be reduced by increasing the contact area between the vanes and casing groove and by connecting several vanes together—the assembly being more rigid and capable of resisting aerodynamic forces than a single airfoil.

Flat or radial hook-fit? When Allied receives a partial row, row, or rows of vanes for repair, the custom­er has the option of refurbishing to the original flat hook-fit or to radial (curved) hook-fits for full contact with the casing groove.

Yelincic took a few minutes to run through the overhaul process. Incom­ing vanes are identified, cleaned, and inspected (visual, nondestructive examination, dimensional check). The owner receives a detailed report with recommendations. Most vanes are in repairable condition and most repairs are relatively minor—usually a little blending and straightening are all that’s required to remove small dents, dings, and tip rubs. Next, the vanes are ready for weld build-up in the hook-fit area, followed by final machining (Fig 9).

To minimize the amount of airfoil tip movement, which can be significant when stator rows are comprised of individual vanes, Allied uses a tie-wire (in reality, nominal 3/16-in.-diam rod stock made from a nickel superalloy) to join five or six vanes into an assembly (Fig 10). This option obviously requires an extra machining step, that to create the groove in the side of the platform to retain the tie-wire.

To shim or not to shim. The owner must also decide if vane rows will be reassembled using shims or not. Allied offers two solutions for the user to choose from, Yelincic said. First is a shim solution consisting of a precision-machined, tight-tolerance button installed on the shim that locks into a mating hole bored into the stator-vane platform (Fig 11).

The second option is for users want­ing to eliminate shims. To accomplish this objective, Allied builds-up weld material on the vane-to-vane con­tact sides of several platforms and machines them to the required new dimension. Recall that shims for Frame 7 machines typically range in thickness from about 40 to 80 mils. Think of the weld deposit as a “fixed” shim, one integral with the platform. The weld deposits on a few vanes are final machined in the field to assure tight fit-up.

Finish work for each option may include some or all of the follow­ing procedures as necessary: stress relief/temper heat treatment, non­destructive examination, shot peen­ing, and coating (if requested by the customer).

Yelincic finished with an experi­ence report. He said that the first vane rows with the tie-wire tech­nique were installed in April 2006. The machine had some rows of vanes with flat hook-fits, others with radial hook-fits. Regular borescope inspec­tions since job completion have been complimentary.

Correcting downstream deflection

Greg Gaul, engineering manager for Houston-based Leading Edge Turbine Technologies Ltd, walked GE Roundtable attendees through his company’s method for correcting downstream deflection (DSD) in GT nozzles.

He began by explaining the poten­tial problem facing owner/operators of Frame 6 and Frame 7 engines with cantilevered turbine nozzles made of cobalt-based superalloys: Deforma­tion of the stationary airfoils caused by combustion-gas loads and high temperatures.

DSD reduces axial clearances between stationary nozzles and rotat­ing buckets, compromises radial seal clearances, and causes high wheel-space temperatures. The generally accepted state-of-the-art for repair, Gaul said, is to deposit weld material on the hook-fit ID to rotate the nozzle segments forward.

While the weld-buildup method is reasonably straightforward, it does not attempt to restore the blade profile which suffers increased defor­mation over the nozzle’s lifetime. Dimensional restoration of the nozzle is possible, continued Gaul, by using Leading Edge’s elevated-tempera­ture dimensional restoration process called CorrectPath.

Here’s an overview of that pro­cess:

• Conduct incoming metallurgical evaluation. If satisfactory, proceed with steps below:
• Re-establish the forward hook location.
• Assemble part in fixture.
• Heat to elevated temperature.
• Move part to desired position.
• Allow metallurgy to stabilize.
• Remove from fixture.
• Conduct applicable nondestruc­tive examinations (NDE) to con­firm material integrity and the absence of material distress.
• Confirm desired dimensions using Leading Edge’s saddle fixture, which represents actual engine clearances.

Gaul then showed a series of pho­tographs to demonstrate the Correct­Path process on nozzles with more than 100,000 equivalent base-load hours of service, but metallurgy suit­able for continued operation. First, the vane segment is locked in a fix­ture that captures the outer hooks and provides reference points for nozzle profile adjustment.

Next, an induction coil is installed to distribute heat across the vane assembly in a manner that repli­cates actual fired-engine conditions (Fig 12). Heating is controlled using numerous thermocouples to moni­tor temperature and help regulate energy input to the coil (Fig 13). An insulation blanket reduces heat loss and helps maintain the proper temperature profile (Fig 14). When all areas of the segment are at the desired temperature, the inner side­wall is moved to its proper position, eliminating distortion (Fig 15).

The shop then verifies that pack­ing teeth, angled when received because of airfoil bending, are in the correct position and straight. Final metallurgical analysis is conducted to confirm suitability for continued service.

Gaul concluded leaving these thoughts:

• Schedule impact of adding Cor­rectPath to a typical nozzle-repair work-scope: Two to three days.
• CorrectPath corrects distortions that conventional DSD correction does not address.
• Nozzles with high operating hours and significant creep distortion can be restored to near new condi­tion.
• Process is controllable and repeat­able and can be customized to achieve any correction or over-correction desired.
• Extends nozzle life significantly.

Other presentations

There were several more presenta­tions at the GE Roundtable—two made by equipment/services pro­viders, half a dozen by users. Floor discussion was robust and continued well beyond the closing bell, which speaks to the quality of the program developed by Chair Gamble and Vice Chair Rose.

The two vendor presentations—one from Wood Group, the other from GTE (Gas Turbine Efficiency)—were general in nature and designed to bring users up to date on company direction and capabilities.

Wood Group began by identify­ing the frame engines it supports (essentially all but the most sophis­ticated machines made by Siemens, GE, Mitsubishi, and Alstom) and by listing the wide range of services it offers electric generators. Next, there were technology backgrounders on several types of defects found dur­ing overhauls, descriptions of repair techniques (TIG and laser powder welding, for example) and facili­ties, coating processes, failure-mode analysis, combustion-system lifetime improvement, etc.

Key points for users to take away included the following:

• Company is heavily invested in technical research to offer high-integrity solutions. In several instances, Wood Group has devel­oped and implemented repair pro­cesses in advance of the OEM.
• Achievements include solutions that offer lower life-cycle costs than those of competitors and the ability to repair otherwise con­demned parts to extract maximum value from existing assets.

GTE provided an overview both of its business and of its capabilities in combustion systems, fuel manage­ment, GT auxiliaries, instrumenta­tion and electrical controls, field ser­vices, and critical parts.

The company has grown dra­matically in the last two years and now is positioned to deliver techno­logically advanced integrated solu­tions for environmental, process, and asset optimization. Focus is on improved performance and avail­ability, fuel efficiency, restoration of optimal design, and parts life exten­sion to increase the profitability of GT owner/operators as well as to pro­mote a cleaner environment.

Products/services of particular interest to GT users include compres­sor water wash and power augmen­tation systems to improve unit performance, combustor nozzle repair and testing, and combustion dynamics monitoring systems.

Plant staffing was the focus of one plant manager’s presentation. He urged users to develop an action plan for replacing employees expected to retire over the next several years. Plan should include, he continued, the transfer of knowledge from senior to junior workers, providing new hires comprehensive technical train­ing, working with community and technical colleges to increase the size of the labor pool, and plans for prepar­ing promising employees for supervisory positions and for retaining those who matter most (Sidebar).

Safety expectations offered good ideas for virtu­ally everyone in attendance to take back to their respec­tive plants. Speaker began by asking “How good is good enough when it comes to safety performance?” He then charted a comprehensive course for perfection—doing everything right all the time.

Ratchet-mechanism and torque-converter issues filled about a half hour of presentation/discussion time. You really had to know your stuff to participate in this dialog. For a novice looking at a disas­sembled ratchet clutch and pawl teeth that failed because of excessive point contact, a 1000-piece jigsaw puzzle had to seem simple.

Varnish formation on IGV (inlet guide vane) servos was a topic intro­duced by a user in the audience. His experience was that a commercial system relying on balanced-charge particle agglomeration for minimizing varnish formation worked best when the lube/hydraulic oil was hot; the electrostatic alternative on cold oil.

R0 blade erosion. Another user in the audience took the microphone and started talking about R0 blade erosion. He said his company’s 7FAs were refitted with “regular” R0 blades—that is, non p-cut—and suf­fered erosion damage that required attention after only fogging for one summer. R0 blades in three of the company’s eight units revealed crack­ing.

Next, he expressed dissatisfaction with the OEM’s blending capabili­ties. Essentially, the procedure used on his units was to sand with 80-grit paper and polish with emery cloth. A complaint was that the treatment doesn’t run the full length of the blade. One blend removes 24 mils; after three blends the blades must be replaced.

Phase-bus fault. One reason user-group meetings are so valuable is that you have the opportunity to learn from forced outages others have experienced. Hopefully, you can avoid similar incidents by factoring lessons learned into the collective thinking at your plant.

Case in point was a phase-bus fault between a generator and its breaker that a proactive user shared with his colleagues. Details: One of a pair of single-shaft combined cycles at this generating facility tripped offline when the fault occurred, but the generator breaker failed to open. The sister unit fed the fault briefly before the affected unit’s generator protection circuit removed it from service. The grid also fed the fault briefly, until the generator step-up protection circuit isolated the grid from the fault. Result: Auxiliary power was lost to both units.

The steam turbine’s dc emergency lube-oil pump started, but failed to deliver oil to the ST and genera­tor bearings. Two turbine journal bearings (one a combination thrust/journal) and two generator journals were significantly damaged (Fig 16). High vibration caused by the bearing failures affected the Zurn coupling between the ST and generator and it was damaged beyond repair.

Note that the gas turbine did not sustain any bearing damage because it has a dedicated lube-oil system that continued to operate as designed throughout the event.

Root cause of the phase-bus fault was rainwater that entered the excit­er transformer cubicle through a damaged roof seal (Fig 17). Water traveled down the B-phase fuse sup­port and accumulated on top of the exciter transformer core. It then entered the transformer’s B phase through the insulated bracing, caus­ing a flash-over and the unit trip.

Lesson learned: Tough to do every­thing that should be done at any plant, especially in view of today’s staffing levels. But efforts have to be redoubled to avoid overlooking some­thing of importance that appears so very unimportant. At this desert loca­tion it was easy to defer roof repair. It doesn’t rain often, but sometimes when it does, water comes down in buckets—as in this case.

The emergency lube-oil pump’s failure to perform is a classic—some­thing that probably has occurred in several plants previously. The motor started and ran but the pump did not rotate. Reason: The coupling was improperly installed by contract personnel during the last overhaul. It failed at some point, allowing the pump shaft to drop away from the motor shaft. Lesson learned: You can’t have too many eyes checking the work of contractors during an outage.

While possible that the coupling failure occurred during this incident, it’s more likely that the routine test to verify pump operation (typically conducted monthly) did not include verification of developed head and lube-oil flow.

You just had to believe there were at least a handful of users in the audience who would call their respec­tive plants to verify the existence of foolproof test procedures and control-system logic. Further that their tests provide positive indication of pres­sure and flow sufficient to support turbine operation.

Others have faced similar chal­lenges in verifying the prepared­ness of emergency lube-oil pumps. A related idea can be found at www.combinedcyclejournal.com/archives.html, click 1Q/2008, click “O&M Best Practices Awards” on issue cover, scroll to p 32 and read “Logic changes enable testing, verification of dc lube-oil pump operation.”

Inlet-bleed-heat system inspec­tion. A user compiled a short pre­sentation based on results of an inlet-bleed-heat piping inspection recommended by GE Technical Infor­mation Letter (TIL) 1320. Recall that the inlet-bleed-heat system injects hot compressed air into the inlet air stream to prevent icing at the com­pressor inlet.

Goal was to inspect for stress cracking on (1) piccolo pipes where they attach to their respective pipe guides, (2) in the areas where the pipe guides attach to the floor, (3) on the duct floor, and (4) in the pipe guide. The unit had accumulated about 200 starts and 1000 actual operating hours at the time of the inspection.

The results: Stress cracks were found in five of the 19 piccolo pipe guides, just above where they are welded to the floor (Fig 18). The cracks were considered minor and did not traverse the entire width of the pipe guide; however, they did penetrate the entire thickness where observed.

The presenter said he was not aware of any root cause analysis to explain the cracking; further that the OEM’s “fixes” were not permanent (they cracked as well). He suggested a user-designed solution probably would be best and might involve reinforcement of the duct floor and a pipe-within-a-pipe retention arrange­ment.

The inspection team also noted some corrosion along the floor weld of the fogging-drain dam where some surface rust and paint distress was in evidence. The simple angle-iron dam was installed to prevent any large water droplets that form along the floor during fogging operations from being ingested by the compressor. Water collected flows via two drains to a water trap and from there to ground. Minor corrosion and deposits also were found on the inlet silenc­ers.

What TILs cost owners

Any experienced 7F user who has participated in a CTOTF GE Round­table, attended a 7F Users Group meeting, or read this journal regular­ly is keenly aware of the many issues that can adversely impact operation of the Model 7FA—an industry work­horse.

More than 700 of these machines are said to be operating at US sim­ple- and combined-cycle generating facilities, so it’s relatively easy to find several people at either of the meet­ings noted to listen to your thoughts on why the R0 compressor blades are prone to failure, how to reduce the potential for mid-compressor fail­ures, how to eliminate shim migra­tion, why the OEM doesn’t support fogging, etc. The list of issues appears endless, yet the Frame 7 continues to be the engine of choice for many power producers.

But not everyone. There is a grow­ing group of owner/operators express­ing resentment about having to pay for inspections and “upgrades” linked to what they believe are design defi­ciencies. One attendee at the CTOTF GE Roundtable said he heard that insurance companies have paid out more than $400 million in claims associated with the machine. But that’s just the tip of the iceberg, the user said. Compliance with GE’s TILs thus far is said to have cost owners well over $1 billion of their own money; one guesstimate is that the tab may already be as high as $4 billion, and counting.

That seems like an improbable number until you do some old-fash­ioned arithmetic. Consider the fol­lowing list of some of the “extra” costs associated with operating a 7FA:

• TIL 1132-2R1, VIG spring and thrust washers. Inspect x-gap clearances annually. Estimate: $85,000 annually.
• TIL 1315-2R1, Inner-barrel coun­terbore plugs for the Model 7241. Insertion of bore plugs requires lifting of the compressor discharge case. Estimate: $200,000.
• TIL 1389-1R1, Compressor rotor-blade erosion from water inges­tion used in evaporative cooling. Take dental molds and conduct fluorescent penetrant inspection after every 100 hours of opera­tion. Estimate: $200,000 for each inspection.
• TIL 1472-2, 7FA+e S17 vane replacement recommendations. Three additional shifts added to hot-gas-path inspection to replace S17 vanes. Estimate: $300,000.
• Annual borescope inspection and tip grinding by OEM. Estimate: $315,000.
• TIL 1562, Shim migration and loss. Borescope inspection for shim migration and shim work. Estimate: $360,000.
• TIL 1576, GT rotor inspection. Remove rotor, inspect in shop, refurbish/replace components; depending on rotor type, expect the entire process to take two to three months. Estimate: $8 mil­lion.

Important to note about the fore­going numbers: They are not “firm” figures that can be simply added and then multiplied across the board to calculate a total cost to the industry. The estimates were developed on an individual basis and without consid­eration of other factors. For example, a given task might require removal of a casing, but if the casing already was removed to address another issue, that cost would have to be sub­tracted from the estimate.

Also important is that replace­ment power costs factored into these numbers were calculated using aver­age replacement power costs in the western states. The actual replace­ment power cost would vary from location to location and by time of year. Startup fuel is natural gas at $7/million Btu.

Legal recourse? With today’s thin margins in competitive power mar­kets, open discussion of the foregoing material increases blood pressure and raises questions on possible legal action that would allow owners to recoup at least some losses beyond those covered by insurers.

One irritant is the consequential-damage disclaimers in equipment warrantees. Using OEM warranty logic, if a portion of a compressor blade were to liberate because of defect, repair/replacement of downstream components would be the responsibil­ity of insurers and/or the owner. Hard to find anyone in the user community who would think this fair.

A question invariably asked: “Why haven’t insurers and owners chal­lenged consequential-damage dis­claimers in court?” Well they have, and won in at least one instance known to the editors. Five years ago, the New York law firm Bruckmann & Victory LLP (www.bvlaw.net) successfully argued that a warranty issued is a warranty on the machine and damage that happens inside the GT is direct damage, not consequential (“Under­writers at Interest v Siemens West­inghouse et al,” NY Supreme Court Index No. 02/602100, order dated Dec 17, 2003 [Gammerman, J]).

The following is excerpted from a commentary published by Bruck­mann & Victory (B&V) in December 2003:

“A New York Supreme Court recently granted summary judgment to subrogated insurers seeking to enforce a Siemens Westinghouse gas-turbine warranty. This significant decision recognizes that a commonly used consequential-damage disclaim­er does not apply when a defective part causes downstream damage inside the turbine.” The details: A 501F engine had defective transi­tions that cracked. Liberated materi­al damaged turbine blades, incurring a $4.5 million loss.

The owner sought a warranty repair from the OEM. “Siemens replaced the defective transitions under warranty but not the turbine blades. Siemens asserted that its warranty against defects did not cover the damaged blades. Specifi­cally, Siemens argued that the tur­bine-blade damage constituted ‘con­sequential damage,’ which the sales contract disclaimed.”

The owner paid Siemens for the turbine-blade repairs and recovered $4.3 million from its insurers which retained B&V to enforce the warran­ty. “The Court agreed with insurers and entered a $5,525,000 judgment [damages plus interest]. In reaching its decision, the Court determined that the warranty covered the tur­bine in its entirety, not merely the individual parts. Thus, the Court found that the downstream blade damage was ‘clearly covered by the warranty’.”

Legacy Roundtable

Chair: Steve Hedge, NRG Energy Inc

Vice Chair: Eddie Mims, Colectric Partners Inc

At virtually every user-group meeting, representatives of plants with GTs approach­ing their second major inspection/overhaul, plus ones beyond that milestone, debate with the OEM the validity of its end-of-life limits for frame rotors.

Perhaps the biggest irritant to an owner/operator: Its OEM has decreed that a certain number of starts (or number of hours) triggers a very expensive rotor inspection to gauge “remaining life,” without providing the engineering rationale for those limits. Keep in mind that these are not actual starts and hours but rather the adjusted numbers cal­culated using equations provided by the manufacturer—ones undoubt­edly reflecting a considerable factor of safety.

For GE engines, company doc­ument GER-3620K, “Heavy-Duty Gas Turbine Operation and Mainte­nance Considerations,” accessible on the Web, states that where a specific interval has not been defined for an engine model, a rotor inspec­tion should be performed at 5000 factored starts or 200,000 factored hours, whichever comes first.

Interestingly, no adjustments are presented in that literature to reflect type of combustion system, turbine inlet temperature, and other factors that might impact rotor life. Thus rotors for the 7B and 7FA+e have the same inspection intervals.

John Scheibel, manager of GT technology for the Electric Power Research Center (EPRI), Palo Alto, Calif, mentioned in his presenta­tion before the Legacy Roundtable that some rotors have been retired based on “design life” rather than for cause, thereby squandering poten­tially significant remaining operat­ing hours. He added that OEMs fre­quently revise rotor life downward. Premature replacement and related outage costs for a given rotor have been known to run as much as $5 million, Scheibel said.

Dr Ashok Koul, PE, president of Life Prediction Technologies Inc, Ottawa, echoed Scheibel’s message when he spoke, suggesting that rotor disk “replacements only are necessary when their condition dic­tates and not when their design lives expire.” Koul described LPTi’s XactLIFE™ tool which allows the user to predict crack-initiation life of such GT components as blades, vanes, and disks at the microstruc­tural level (visit www.lifeprediction­tech.com for more information).

Gil Dean, president and princi­pal engineer, AccTTech LLC, Greer, SC, outlined for users a proven ana­lytical approach for determining the life remaining in turbine rotors. Dean should know, having spent three decades with GE, including a stint as engineering manager for rotor design and as the F-rotor task force manager.

He discussed the rolls played by, and the interactions among, the three core groups involved in the life-extension evaluation process: design, materials, and inspection. Dean stressed that any program without all three would be a failure or very lucky. He closed his presen­tation with the following observa­tions (in reality, a list of warnings and risks):

• The design lifetimes of rotors quoted by the OEMs are based on constant materials properties.
• Material service degradation, particularly with regard to frac­ture toughness, is a very impor­tant part of the analysis.
• There is no such thing as no defects.
• Wheel forgings were not defect-free when new in the 1970s.
• Structural analysis is required for NDE sizing, location, and dis­position.
• Critical flaw size is smaller after years of operation than it was ini­tially because of temper embrit­tlement.
• Damage-tolerance and risk anal­yses are required to determine the reinspection interval.
• Knowing how to dismantle a rotor is not the same as knowing how it works.

Report card

If your rotor “passes” the OEM’s inspection it can be certified for extended service. However, expect that critical parts will require reha­bilitation or replacement before this happens. Users have told the edi­tors that TIL 1576, “Gas Turbine Rotor Inspections,” published in 2007, says 7EAs can be certified for up to 50,000 additional factored hours of service, Frame 5s for up to 100,000—provided factored service hours was the trigger for the inspec­tion. These sources did not know of a comparable number for 7FAs.

To illustrate how quickly the life limits must have been assigned, con­sider that the 5000 factored-starts trigger for an inspection is exactly the same as the absolute end-of-life number specified by the OEM. This begs the question: Why would you inspect if the rotor already has been declared “dead”?

An OEM rotor inspection repre­sents a major commitment on the part of the owner in terms of unit unavailability. The last bullet point in the previ­ous section suggests two to three months. This is about the same time Siemens Energy suggests allowing for their Class IIB inspection which involves destacking of the rotor (access www.combinedcyclejournal.com/archives.html, click 1Q/2007, click “501D5/D5A” on the issue cover).

Such a prolonged outage has some users thinking about selling their superannuated rotor to the OEM and buying a reconditioned spare. After refurbishment, your old rotor becomes a reconditioned spare for another plant.

OEM alternative

But there’s another option: Use a qualified third-party to perform the inspection. A pioneer in this approach is Arizona Public Service Co. John Lovelace, who at the time of the meeting was both CTOTF chairman and a consulting engineer for the utility, said the cost for the inspection was less than what the OEM would have charged; plus, the time from when the rotor was removed until it was reinstalled was only a couple of weeks (the in-shop inspection work took just five days working one 12-hr shift per day).

The biggest benefit: The 7C rotor, removed from one of West Phoe­nix Generating Station’s STAG 100 combined cycles that were installed in June 1976, had more than 6000 starts (Fig 19). It would have been scrapped had the rotor been shipped to the OEM for inspec­tion.

Half of the Legacy Roundtable was dedicated to presentations on what the third-party team did to inspect the 7C rotor and what was learned. The team consisted of these principals (areas of expertise are noted):

• Rich Curtis, VP engineering, Eta Technologies LLC, Coventry, Ct (www.etatechnologies.com): met­allurgy.
• Gary Hensley, president, Verac­ity Technology Solutions, Tulsa, Okla (www.veracityts.com): non­destructive examination.
• Paul Tucker, president/CEO, First Independent Rotor Services of Texas (FIRST): mechanical/dimensional.

The trio developed a rigorous advanced-technology inspection pro­tocol to assess the condition of GE rotors operating beyond the TIL 1576 limits. Rotors passing the inspection would be returned to service for an additional major-inspection inter­val, thereby saving the owner the cost of a reconditioned spindle or a new one. The group is currently focusing on GE rotors but plans to expand its services to rotors made by other OEMs.

Curtis led with a presentation covering the inspection approach and metallurgical work conducted. He began by discussing some of the important considerations with respect to the West Phoenix rotor. For starts-based units, Curtis said, cyclic loading (low cycle fatigue, LCF) of the bore region is life-limit­ing. A metallurgist’s concern is that over many cycles, small “insignifi­cant” subsurface defects (for exam­ple, inclusions and “stringers”) theo­retically could link up and develop into a flaw of critical size. Worst case scenario if that happened: a wheel burst.

Thus the scope of work for a proj­ect like this must include a thor­ough volumetric inspection of the bore to identify any inter­nal defects that might exist. Other critical areas to inspect include through-bolt holes, bucket dovetail serrations, and the web/rim area (the region between the bolt circle and bucket attachment).

Inspection plan imple­mented. For the West Phoenix 7C, nondestructive examina­tion (NDE) of the three turbine wheels and two spacers (Figs 20-23) consisted of the following steps:

• Visual and magnetic-particle (MT) inspection of all surfaces.
• Semi-automated phased-array ultrasonic (UT) inspection of the bore and of the wheel web and rim areas.
• Semi-automated eddy-current (ET) inspection of all bolt holes.
• A 100% ET examination of the wheel-rim bucket dovetail serra­tions.

Metallurgical evaluation includ­ed a microstructural assessment of bore and rim regions, fore and aft, using the replica technique. Plus, hardness measurements of bore, rim, and dovetail regions, fore and aft.

The dimensional assessment: Measurement of rabbet fits, bolt-cir­cle faces, bores, and bolt holes; also, seal features on spacers and bucket attachment on wheels.

Results, recommendations. No reportable indications were identi­fied using the plan outlined above. Given that no defects had initi­ated or propagated in the first 6000 starts, Curtis and the other team members concluded “that it is more than reasonable to assume no defects would grow and propagate into any­thing near a critical flaw size in the next major inspection interval.” The recommended repeating the inspec­tion scope of work during the next major.

The conclusions, Curtis contin­ued, are based on inspection-process fidelity, inspection methodology/cri­teria experience, and the excellent inspection report. This certainly sup­ports the recommendation to run for another major-inspection cycle, he told the editors.

However, Curtis recommends the industry support independent analy­ses (possibly conducted by EPRI or an independent company such as AccTTech) of various rotor designs considering the large number of spin­dles that will be inspected in the com­ing years for life-extension purposes. The effort would be of value to the user community for several reasons, including these:

• An independent design review offers a “sanity check” on the OEM’s end-of-life limits. Keep in mind that the design tools avail­able to engineers today are vastly more sophisticated than those used to design rotors 30 years ago—even 20 years ago or less. It’s reasonable to assume that rotors now approaching end of life as the OEM interprets it may have very conservative factors of safety and substantial remaining life.
• Determine critical flaw sizes to establish UT inspection criteria.
• Assess critical material proper­ties.

Curtis concluded his prepared remarks by covering in detail how the metallurgical and hardness tests were conducted and the detail results obtained. Hensley followed with the specifics on each type of NDE proce­dure. Critical tools and techniques reflected their aerospace DNA, per­haps being used in the electric power industry for the first time.

To illustrate: A proprietary 128-element phased-array fixture was used to conduct longitudinal and shear-wave UT of the wheel-spacer center bores and of the three turbine wheels. More than 100,000 images were collected to assist in 3-D reconstruction of the volumetric composition of spacer/wheel bores and flats. No reportable indications were noted.

Tucker concluded the formal pre­sentations by describing the results of his mechanical and dimensional review. Dimensions to be measured were specifically selected to indicate changes that may have occurred in service—such as wear or growth. Measurements were compared to as-built values (deviations were record­ed) and assessed in terms of service­ability based on Tucker’s experience with these components spanning three decades.

Tucker pointed to the thorough­ness of his inspection, saying that traditional measurements of rotor components involve only a modest number of convenient locations. He made 10 times the number of mea­surements typically made at the repair-shop level. ccj