W251 Users Group: R1 compressor-blade failures dominate discussion at annual meeting

If there ever was a gas-turbine (GT) model in need of a user group it is Siemens Power Generation’s W251, known today as the SGT-900. There are only 49 of the latest engines (versions B10, which debuted in 1983, through B12A) thought to be in service and they are scattered worldwide—including the US, China, India, Canada, Korea, Pakistan, and Argentina.

Most of the nominal 50-MW frames are installed in industrial settings; about half are in base-load service. They are “workhorses” that generally operate under demanding conditions for owners who cannot afford “oats.” With so few units in the fleet, and today’s tight budgets, there are not many caregivers and very little “free consulting” is available. Users rely heavily on one another in time of need.

What comes across loud and clear in every “needs assessment” survey of SGT-900 owners is “price.” Upgrades are welcome if they pass demanding financial analyses, but when it comes to replacement parts and repairs, price is just about everything.

There are supporters for this frame, of course. They include the OEM and third-party suppliers of critical parts and services such as ETS Power Group Inc, Stuart, Fla; Liburdi Turbine Services Inc, Dundas, Ont, Canada; Trinity Turbine Technology LP, Iowa Colony, Tex; Sulzer Hickham Inc, Houston; and Stork H&E Turbo Blading Inc, Ithaca, NY.

Two-dozen users representing about 40% of the B10-B12A fleet attended the group’s 2007 meeting in mid April, which was held under the CTOTF umbrella for the first time. The Combustion Turbine Operations Task Force, the nation’s oldest GT user group, serves most frames and aeros in electric generation service rated more than 25 MW. By working through CTOTF Group and Conference Coordinator Wickey Elmo (wickelmo@carolina. rr.com), the 251 users eliminated their administrative workload and allowed Chairman Joe Mitchell and Vice Chairman Federico Kitzberger to focus exclusively on program content.

 Mitchell, the former facility manager for Newark Bay Cogeneration, represented India’s Spectrum Power Generation Ltd (sidebar, p 36). Kitzberger is the power generation manager for Argentina’s Capex SA, which operates five W251B11-12A engines at its Central Termica Agua del Cajon combined-cycle plant.

R1 blade failures

A portion of last issue’s report from the CTOTF Spring Turbine forum in Orlando was devoted to failures of first-row blades in GE Energy’s 7FA compressors (the so-called R0 issue). In another room at that same meeting, the W251 users were getting the details on four Row 1 failures in their fleet. Plants in Pakistan, New Jersey, China, and Korea all had suffered R1 failures at various points in their respective operating histories.

Coverage of the W251 incidents was not compiled into one session like that for the 7FA failures. Rather, there was a user-only discussion independent of a presentation summarizing a root-cause analysis (RCA) conducted on two of the four failures by a metallurgical and engineering R1 compressor-blade failures dominate discussion at annual meeting.

Failures of first-row compressor blades in four W251 (SGT-900) engines worldwide have caused significant damage team from Austin-based Mechanical & Materials (M&M) Engineering and Turbine Technology International (TTI), Rochester, NY. Plus, the OEM offered its views on the subject during the Siemens Day session.

The editors gathered information on the W251 compressor failures from different sources during the meeting and have presented it in one section below to provide continuity of subject matter.

This was Mitchell’s first meeting as chairman, but you never would have known it. He, like Kitzberger, who chaired the group the previous two years, is intimately familiar with the W251 and an excellent discussion leader. Both also know many of the participants and the histories of their respective units, allowing them to call into the dialog an “expert” when needed. There’s no waiting to see if someone will volunteer an answer or some advice. Easy to get caught napping at this meeting—if you dare.

The group was galvanized by the compressor failures and participants came together to help one another. Three of the four plants affected were represented, so the specifics of each accident were available first-hand. The general facts surrounding the failures offered no clues to link the incidents by cause. One machine had only 50,000 hours of service, two were approaching 100,000 hours; operating regimes differed; ambient environments and global locations varied.

One take-away from the user-only discussion, Mitchell said, was that attendees were ill at ease not knowing the exact cause(s) of the incidents. The prevailing theory: Surge or some other form of brief air-flow stall/reversal probably was to blame. Also of concern to the users were the risks faced by the fleet regarding spare parts. Vendors do not carry sufficient inventory to accommodate multiple users if more than one is hit with a failure at approximately the same time. Long lead times could be disastrous for some owners.

Having reliable third-party suppliers is particularly important to this group. The OEM, like any other for-profit enterprise, is limited by the amount of capital it can tie up in inventory. For this fleet, it generally has one or two spares of a given part in stock Replacements are ordered as the shelves empty and lead times rarely are less than several months.

But Siemens does have a spare rotor for the SGT-900 that it can exchange with a user’s to minimize outage time. The company’s rotor/component exchange program was described in the 501D5/D5A Users report in the last issue (access 1Q/2007 at www.combinedcyclejournal. com/archives.html).

During the compressor discussion, Stork was mentioned as an alternative to the OEM for new compressor blades, ETS for engineering of diaphragm repairs, and Trinity and Sulzer Hickham for making the diaphragm repairs.

Several weeks after the meeting, Mitchell said, a few users were so concerned about the potential for a failure at their plants, they ordered new first-stage compressor blades and would install them as soon as practical to minimize the risk of a liberation incident. One glance at the photo on p 33 certainly could expedite a purchase order. Note the missing airfoil at about the 4 to 5 o’clock position in the R1 set of 19 blades.

Newark Bay RCA

M&M Engineering’s Mark Tanner and TTI’s Robert Dewey presented the conclusions and recommendations of their root-cause failure analysis for the incident at Newark Bay Cogen and for one of the overseas plants.

Background facts first: Both GTs were rated 48 MW; Newark Bay was burning natural gas, the overseas unit blended gas; both had nearly the same number of equivalent operating hours (EOH), nominally 49,000; equivalent starts (ES) at failure were 1300 at Newark, 1717 overseas; Newark failed during startup, the other while operating at base load; neither exhibited any abnormal operating conditions.

Wet fluorescent magnetic particle (WFMP) examination revealed no indications for the US machine; however, there were six indications on five of the overseas unit’s blades. No pitting was observed on the Newark blades, but half of the blades on the other engine were heavily pitted (largest pit was 50 mils deep).

The more than two-dozen photos, photomicrographs, and charts shown were a valuable primer for every owner. Many users seemed confident that they could go back to their plants knowing what to look for and generally be able to make go/no-go operational decisions based on their observations. Also, to decide what additional testing, if any, should be conducted by NDE (nondestructive examination) professionals.

Failure location at Newark Bay was 7.9 in. from the bottom of the blade and 3.2 in. from the leading edge on the convex side; overseas 10.5 in. from the bottom and 2.1 in. from the leading edge on the convex side. The metallurgists also provided the size of the origin and striation spacing to assist in making comparisons if other R1 blades in the fleet were to fail.

Conclusions of the metallurgical analysis:

• Mode of cracking in both cases was high-cycle fatigue (HCF).
• Crack origin was a very small anomaly (6 × 4 mils) in Newark, a 16 × 6-mil corrosion pit overseas.
• Some corrosion was present in the anomaly, but it was not typical of that associated with chloride pitting. By contrast, corrosion pitting and stress corrosion cracking (SCC) was observed on several of the overseas blades; another cracked blade showed HCF emanating from SCC.

Other observations common to both units:

• Origin areas tested positive for chlorides.
• Mul t iple beach marks were observed.
• Blade mechanical properties and chemical composition were in-spec.
• For the Newark unit only, aluminum oxide particles were found embedded in blade surfaces.

What all the findings mean, said Tanner and Dewey, is that blades in both units failed from HCF on the convex side; beach marks indicate intermittent cracking. Further, the small crack initiation site on the Newark blade indicates that blade stress level played a major role in the failure.

The presenters then dug into the nitty gritty of the technical approach to the methodology structural modeling, explaining why they did what they did. Next came discussion of dynamic analysis, structural modeling, crack propagation analysis, fracture mechanics and damage tolerance, etc. Good “stuff” but undoubtedly difficult for many to absorb in one sitting.

The conclusions brought everyone to attention. Tanner and Dewey said the following:

• The materials and service regime were not accountable for cracking.
• Maximum steady stress and lowcycle fatigue were not a primary cause or major contributor to the failures.
• Neither was resonance.
• Nonsynchronous vibration was a principal source of HCF. The convex side is where flow separation is likely to occur, promoting deposition/ pitting. Plus, dynamic stress is high enough to initiate HCF.
• Problem is propagation, not initiation.
• Damage accumulation is intermittent and occurs at startup and shutdown.

Simply put, the R1 failures were caused by pitting in an area of high dynamic stress; both conditions are required for failure to occur. Tanner and Dewey also concluded that the R1 blades have marginal tolerance for damage.

Perhaps the most important slide was the last one: Recommendations. Tanner and Dewey said managing airfoil surface condition was more practical than redesigning the blade. They suggested replacing or refurbishing R1 airfoils before 50,000 hours of operation, which translates to no later than the first major.

When refurbishment is the option selected, owners must be certain the blades are pit-free (including nicks only a mil deep) on the convex side. If 10 mils of material or more is removed from the airfoil to eliminate the pits, be sure to verify that blade frequencies are in spec. Noted, too, was the possibility of identifying a suitable coating to limit pit formation. Perhaps the coating technologies offered as possible solutions to the 7FA R0 problem are worth reviewing (refer to the CTOTF report in 1Q/2007).

The Siemens R1 update, presented by Principal Engineer Dan Stankiewicz reflected the limited first-hand access to the information that the OEM had about the failures. Siemens metallurgists and engineers had not gained access to any of the failed blades prior to the Orlando meeting and its conclusions were based solely on photographs provided by the owners.

Yet the OEM’s researchers came to many of the same conclusions reached by the M&M Engineering/ TTI team. Stankiewicz observed that all of the failures exhibited similar characteristics but they did not occur at the same height on the airfoil. One blade broke much closer to the base of the airfoil than the others; even the three fractures that occurred closer to the tip were not at identical heights. The location of the break is important because it can help in isolating a specific fracture mode. Siemens used the latest analytical tools to conduct modal and harmonic- response analyses that suggested the incidents may have been caused by one or more of the following conditions: under-frequency operation, multiple compressor surges, chemical contamination resulting in incipient stress cracking, inlet icing, excessive overspeed. Review of operating records confirmed that various combinations of these off-design conditions had occurred at each site.

The OEM’s recommendations:

• Confirm that any chemicals used to wash the compressor, treat evaporative cooling water, etc, meet Siemens specifications.
• Consider coating airfoils when airborne contaminants, such as chlorides, are present.
• Inspect the compressor inlet immediately after significant upsets— such as surge, inlet icing, off-frequency operation, and excessive overspeed.

A visual check, followed by fluorescent or magnetic- particle inspection (sensitivity level III) is recommended. Pay particular attention to the convex (suction) surface of the airfoil. Stanki ewi c z c onc luded hi s remarks by mentioning that a R1 technical advisory is being developed to help guide decision-making by owner/operators.

Siemens Day

Chairman Mitchell said that rotor overhaul strategy for the ageing W251 fleet ranked a close second to the R1 compressor blade issue in terms of discussion time during the closed user sessions. Siemens recommends a Class IIB inspection at 100,000 EBH, he continued, and is telling users that operational risk is medium to high over extended operation if it is not performed.

Most owners are struggling to justify the benefit of such an overhaul, Mitchell added, considering the cost and time involved. In their minds, the OEM has not demonstrated convincingly that such an inspection at this interval will thwart a catastrophic failure. They point to several engines in the fleet that are beyond 120,000 EBH and have operated without incident.

Siemens’ program to eliminate or mitigate the risks of component failure from high-hours operation and to extend the lifetime of units from 100,000 EBH/3200 ES to 200,000 EBH/6400 ES was articulated by Chris Yager, senior engineer, Mature Frame Gas Turbine Service Engineering & Projects. He said that the objective of the lifetime extension program (LTE) was to provide customers with an engineered process for GT lifetime evaluation and extension in accordance with their operational objectives.

The LTE program, continued Yager, is an engineered process that has three main elements: generic frame review, unit-specific evaluation, and detailed inspections. Siemens design and service teams begin by compiling generic frame information. This includes identifying the latest analytical tools available for determining the life remaining in critical parts, such as the unit’s rotor and compressor discs; selecting the best materials available to maximize the life of replacement parts; reviewing a fleet service history of more than 3-million EBH to identify generic frame issues, etc.

A meaningful evaluation of the unit requires that the owner provide detailed historical O&M information— including hours, starts, trips, fuel, major events, abnormal operating conditions, inspections, repairs. Operation at off-design conditions is particularly important to the OEM’s engineers and metallurgists.

Some reasons: Conditions such as overspeed, high vibration, engine surge, and fast or slow starting increase the stress/strain on components; high firing and/or disc-cavity temperatures can change materials properties and increase their susceptibility to fracture; high hours on turning gear can reduce the tolerance of critical components to the wear and tear of normal operation.

Inspection. Critical non-consumable components/parts—including rotor, exhaust section, blade rings, inlet guide vanes, supports, and trunnions— must be thoroughly inspected. This includes the sampling and detailed analysis of materials. Inspection and testing of the non-rotor components would be done at an extended major. Note that consumable components—such as combustor hardware, blades, and vanes— are normally considered outside the scope of the LTE review.

Both time- and cycles-dependent ageing mechanisms are considered in the analysis. First group includes creep, alloy instability/loss of material properties, oxidation/corrosion, component distortion, wear, mechanical damage, and erosion; low- and high-cycle fatigue are in the second group along with wear attributed to turning-gear operation and low-temperature corrosion associated with extended downtime.

The Class IIB and supplemental LTE rotor inspections suggested at 100,000 EBH require top shop disassembly/ reassembly and NDE skills as well as specific design knowledge of the engine. General details are presented in last issue’s 501D5/D5A report, referenced earlier. Important to remember is that each LTE inspection is tailored to the unit’s service history and the owner’s future operational objectives and risk profile. The goal is truly “lifetime extension,” Yager interjected, not “making it to the next overhaul.”

In the final section of his prepared remarks, Yager reiterated the purpose of the LTE program: To help reduce the risk of catastrophic failure beyond the targeted design life. He then presented the conclusions of a sample rotor risk-assessment analysis conducted by Siemens that scored the probability of failure, severity of failure, and ability to detect an impending failure for various failure scenarios over extended operation beyond the targeted design life.

Multiplying the “scores” for each yielded a “risk level” for each failure scenario. With 13 potential failure scenarios identified, no inspection results in five being considered high risk and eight medium risk. A typical major inspection can improve this to four scenarios being considered high risk, seven medium risk, and two low risk. Note that the severity of a potential failure may warrant a “high” risk rating while the probability of occurrence is low because of limitations in detectability during a major.

Conducting the Class IIB and supplemental LTE inspections drops the risk level for all potential failure scenarios to “low” with scores that in some cases are significantly lower than for the major-inspection case. Two 251 rotors were undergoing Class IIB and supplemental LTE inspections at the time of the annual meeting. One inspection was being done at the OEM’s Hamilton (Ont) works, the other in Houston; details were not available.

Upgrade portfolio. Stankiewicz and Senior Engineer Robert Mozzoni collaborated on a presentation that identified many of the product upgrades available for the 251B10-12A models, explained the benefits of installing them, and suggested the optimum time for installation—for example, next HGP inspection. The upgrades discussed were the following:

• Disc-cavity cooling reliability package.
• Electrohydraulic IGV actuator.
• Electrically actuated throttle valves.
• Redundant instrumentation.
• Online compressor performance monitoring.
• Increase exhaust-temperature limit at high ambient temperature.

The disk-cavity reliability modification includes more effective sealing solutions (replacement of labyrinth seals in the exhaust face of the interstage seal housing with brush seals and rope seals between the R2 and R3 interstage seal housings and the vane segments) and use of a bypass valve and orifice scheme to better control the flow of cooling air into the disk cavity than was possible with the original fixed orifices. Latter helps to prevent high temperatures in the disk cavity and allows the engine to run reliably and less likely to auto-unload.

Replacing pneumatic IGV actuators with electrohydraulic actuators is relatively simple, requiring no changes to the inlet cylinder and IGV unison ring. Benefits can include improved control, accuracy, and repeatability; tighter control of exhaust temperatures at part load; and help in alleviating sticking issues.

Advantages of replacing pneumatically actuated throttle valves with electrically actuated valves can include improved accuracy and control and easier calibration. This can reduce fuel consumption from ignition to base-load operation. Some owners also can realize an increase in starting reliability because of fewer failed starts and trips initiated because of poor acceleration and over-temperature on hot starts.

Redundant instrumentation. All trip circuits to the control system— such as bearing oil pressure, exhaust backpressure, turbine overspeed, etc—are upgraded to permit two-of-three voting, to eliminate the possibility of a trip on single-point failure. Thus, if the failure of a device triggers a momentary false signal, a trip is not initiated without confirmation by a second device. Increased engine reliability and protection are the benefits for owners. If a single device is identified, it can be replaced online (without shutdown) via a mechanical lockout mechanism that allows continued operation of the other two devices.

System monitoring and control package measures/ calculates operating conditions in real time at all loads and IGV positions for comparison against baseline data. Information provided helps operators to maximize overall plant efficiency and guides decisions on work required during the next outage. Calculated metrics include heat rate, compressor mass flow and efficiency, turbine efficiency, exhaust pressure drop, etc. Exhaust-temperature control curve can be modified to increase power and efficiency when operating at high ambient temperatures.

Feature was included as a standard feature on many B12As and can be retrofitted on other engines during a combustor inspection without impacting schedule. Modifications to the exhaust manifold permit this upgrade by mitigating potential issues with support distress—such as deformation and cracking.

Additional mods and upgrades covered by Stankiewicz and Mozzoni in less detail included steam power augmentation, starting package replacement, emissions control using steam or water injection, island control, B10 to B12 upgrade, and improved R2 vane segments.

Transition-piece redesign. An upgrade accorded special status at the Orlando meeting was the new transition piece (TP) for the 251. Senior Engineer Bill Van Nieuwenhuizen reviewed the issues with the original TP as he has at the last few meetings of this user group:

• Exit mouth deformation which caused excessive blade-path temperature spread and NOx emissions. Forced outages resulted.
• Excessive panel temperatures— primarily associated with DLN applications which have combustion- basket exit cones to reduce CO emissions. Generally did not cause forced outages.
• Cracking at the ends of stiffening ribs. Generally did not cause forced outages. Van Nieuwenhuizen reminded the group of the new TP’s design features:
• Baskets canted to minimize turbulence in the hot gas stream, thereby reducing heat transfer.
• Exit-mouth outer seal replaced with an integral mounting-bracket seal assembly.
• Special cover plates to angle baskets.

He also mentioned that the only engine modification required to install the new TPs is relocation of igniters and flame detectors. Retrofit requires a total outage duration of 12 days with one 12-hr shift/day; this includes any necessary field modifications.

The new TP has been installed in three engines, continued Van Nieuwenhuizen, but still is in the design verification phase of its development. Comprehensive, heavily instrumented tests of new-design TPs installed in the first engine in November 2005 verified peak temperatures that were more than 400 deg F lower than measured for the previous design—as engineers predicted. Dynamic strain measurements showed no indications of resonance and the exit mouth remained dimensionally stable.

Through 1Q/2007, the first engine had accumulated more than 450 operating hours and 55 starts with the new TPs. Two crawl-throughs had been conducted, the latest last February. The report card: “Transitions in very good condition and dimensionally stable; inner seals free.” Also, an alignment change to correct minor fretting of crossflame tubes, observed during the first inspection in December 2005, was successful.

The second engine equipped with the new TPs had accumulated about 400 hours and 30+ starts by the end of the first quarter. A crawl-through after 42 hours and five starts identified no problems with transitions, new transition hardware, or crossflame tubes.

The latest unit equipped with the new transitions was retrofitted only a few weeks before the Orlando meeting. ccj

Independent voice of the gas-turbine-based generation sector

© 2014 CCJ Online, Inc. All rights reserved. | Published by PSI Media, Inc | 7628 Belmondo Lane | Las Vegas, NV 89128