Alloy Aging and Creep Damage

Turbine blades are manufactured from precipitation-hardened nickel-based superalloys that develop their high strength at high temperatures from gamma prime, Ni₃Al precipitates within the material. In the as-new condition, the size, density and morphology of the gamma prime precipitates are optimized for maximum strength (Figure 12–24a). However, during operation at elevated temperatures, the microstructure undergoes various changes. The strengthening gamma prime precipitates spheroidize, enlarge and join, resulting in a reduction in material strength (Figure 12–24b and c). Topologically close-packed (TCP) phases can also form resulting in a dramatic loss of ductility and reduction in creep strength (Figure 12–25a). Microstructural degradation takes place by diffusion mechanisms and the rate of degradation is therefore a function of both time and component temperature.

Due to the centrifugal loading and operating temperatures, the airfoil material also undergoes creep deformation. Creep is a diffusion-based mechanism correlated exponentially to temperature. Eventually, creep results in the formation of voids that link to cause creep cracking (Figure 12–25b). While the detection of creep voids is a reliable indicator of creep damage, in most alloys, voids only form after a large fraction of the creep life is exhausted and hence they may be absent even in significantly creep-damaged material. Prior to the formation of such voids, it is difficult to identify damage due to creep. Creep damage may be evident by reduced creep rupture life measured by mechanical testing; however, the effect is confounded with the aging effects.

Both creep and microstructural damage can be repaired through appropriate heat treatments. Knowledge of the material condition can be used to ensure the appropriate thermal processing is implemented to restore the material properties needed for continued service at minimal risk. Base material that exhibits creep damage requires hot isostatic pressing (HIP) at appropriate temperature and pressure to collapse creep voids that may have formed. Alloy degradation is corrected by heat treatments capable of fully restoring the microstructure and material strength.

As microstructural degradation occurs as a function of temperature, the condition of the base alloy provides a relative thermal map of the blade’s operating temperatures. Once a typical thermal profile has been benchmarked for a given blade design, abnormal engine operation affecting the operating temperature of the blade can be detected.

Examples include short-term excessive overheating, long-term overheating, and temperature profile shift.

The base alloy condition can also provide insight into the suitability of the service interval length. Gross microstructural damage or the presence of creep voiding can indicate an inappropriate service interval at the conditions exposed. In the absence of an identifiable cause, it would be recommended to shorten the service interval in future. Conversely, if only moderate base alloy deterioration is present, service interval extension could be considered.

Aluminide/MCrAlY Coating Deterioration in Clean Environments

Diffusion and overlay coatings provide oxidation protection for operating environments for which the base alloy material does not have acceptable resistance. Oxidation resistance of coatings is principally based on aluminum content. The high aluminum content existing as β-aluminide forms a protective aluminum oxide layer slowing down the oxidation damage process. This process is exhaustive and once the aluminum content is depleted below a critical level, the β-aluminide transforms to gamma prime. At this point, a mixed oxide is formed that does not provide the same protective nature.

Figure 12–27 displays examples of coatings after service with minimal, moderate, and complete β-aluminide depletion. In simplification, the coating protective life is based upon the amount of β-aluminide remaining.

Coating breaches can occur due to either complete β-aluminide coating depletion or thermal-mechanical coating cracks (Figure 12–28). Generally, coating failure is deemed when the base alloy sustains oxidation damage that is either irreparable and/or increases the risk of failure to unacceptable levels.

Optical examination of the coating is able to determine the amount of remaining β-aluminide in a coating, examine for cracks as well as oxidation damage to the coating and underlying base alloy. Data obtained from characterizing the coating/subsurface condition can be used to (a) assess if the coating provided adequate protection during previous service, (b) if inadequate, determine a more suitable coating based upon the damage mechanisms, and (c) assess if there is potential for service interval extension (incorporating other modes of damage).

On sets that the coating had failed to prevent damage to the base metal, the coating type, service interval, and operating conditions should be reviewed. In cases where the base alloy did not exhibit excessive alloy degradation, the coating may be defined as the primary life-limiting mechanism. In such cases, a coating upgrade may be beneficial.

Full depletion of the coating, combined with gross base alloy degradation, may suggest elevated operating temperatures or too long of a service interval at the operating conditions applied. If neither the coating nor any other form of degradation has approached their life limit, there may be the potential to increase the service intervals.

Airfoil Deposits

Deposit build-ups on airfoils can reduce efficiency or provide an indication of an upstream issue. Indirect damage to the airfoil can be sustained from cooling holes becoming blocked. The origins of deposits can range from ingested material, parts rubbing or upstream failures (melted and adhered metal).

Determining the composition of deposits building up on turbine blades can provide insight as to the material origins. Similar to impact damage, deposits can indicate issues with filtration, upstream component failures, or objects left within the gas path during maintenance. Adhered airfoil deposits generally do not directly affect the reparability of a blade set.

Internal Surface Condition

Due to the inaccessibility and nature of damage incurred within the internal cooling passages, internal damage to the base alloy cannot be repaired. The decision to continue servicing the parts should include an evaluation of whether the existing internal damage can be tolerated. As non-destructive testing cannot detect the majority of internal damage modes, to determine whether a set of internally cooled blades can be safely returned to service requires a metallurgical life analysis.

Repair Decisions and Development

The information obtained through metallurgical examination can be used to (a) identify the reparability of the remaining blade set, (b) identify repair procedures necessary for continued service, (c) determine if a change in the coating type would be beneficial and (d) determine the cause of indications or conditions noted during inspections.

Metallurgical evaluation is by far the most thorough way of characterizing the sustained damage and therefore can be used to develop repair processes customized to the part and ensure the appropriate restorations are completed.

Features of F Class Hardware

Advanced Coatings

The F class of engines makes more extensive use of coatings than previous generations of engines:

• F technology engines are the first industrial designs to use thermal barrier coatings (TBCs) on rotating airfoils. Previously, the use of TBC coatings in industrial designs was restricted to stator vanes and combustion hardware applied by air plasma spraying. To meet the demands of coating the more critical rotating blades, advanced techniques such as electron beam evaporation and high energy plasma spraying whose use had been restricted to aero components in the past are now used to generate more strain tolerant coatings (Figure 12–33).

• Coatings are applied to downstream stages that were uncoated in earlier designs. The coatings are the same type of MCrAlY coatings that were used in the first stages of previous engines.

• Internal coatings play a more critical role. The thin walls associated with the complex serpentine cooling passages result in higher surface temperatures on the cooling hole walls and therefore need protection against oxidation. The coating processes, however, are the same processes used in earlier industrial and aero designs.

Single Crystal and DS Materials

The first use of single crystal alloys in industrial gas turbines is found in F technology engines (Figure 12–34). While such alloys have been used for many years in aero-engines, the challenge of casting larger industrial components prevented their use in industrial engines until recently. The F technology engines also see much more extensive use of directionally solidified alloys, which saw use only in the first stage blades of some of the previous generation of industrial engines.

Repair Technologies

Rejuvenation

Vacuum heat treatments are traditionally used during the repair process to restore the microstructure, improve weldability, apply brazes and diffuse coatings. Significant metallurgical knowledge and experience are required to tailor the heat treatment steps to the superalloy being processed. To ensure that the creep, fatigue, and aging damage incurred during service are reversed as much as possible, and that the correct microstructure and mechanical properties are produced after all welding and coating steps have been completed, a full solution heat treatment, graphically illustrated in Figure 12–35, is essential.

The benefits of regularly performing such rejuvenation in combination of other repair steps have been well established on earlier generation engines. By restoring the microstructure through heat treatment on a routine basis, the life of the components can be extended well beyond the life at which they would otherwise need to be retired (Figure 12–36). For example, the use of rejuvenation at appropriate intervals for Frame 7EA first and second stage buckets can extend their life to more than twice that achievable through re-coating only (Figure 12–37). The economic benefit at a site operating four units was calculated to be approximately $5 million over a 5-year period.

DS and Single Crystal Alloys

The same principle of extending the life of hot section components by rejuvenation heat treatments has been applied to directionally solidified components as they have been introduced in the earlier industrial and aero engines. The heat treatments function in the same way to restore the gamma prime microstructure and related mechanical properties.

An additional concern for DS and single crystal alloys is the possibility that recrystallization will take place during heat treatment if the components have been subject to plastic deformation. These alloys develop their strength in part from controlling the orientation of grain boundaries or, in the case of single crystals, eliminating them all together. Figure 12–38 shows recrystallization of a single crystal alloy subjected to shot peening to introduce surface deformation. If such damage were to occur on an airfoil surface, for example, as a result of impact, the LCF and creep properties would be reduced. Rejuvenation of DS and single crystal alloys must therefore include examination of the etched airfoil surfaces for evidence of recrystallization after the heat treatments.

Repair Materials

Welding

Many repairs involve the replacement of missing material or repair of cracking. Welding is the most common method used to repair, join, or replace material in hot section components. The conventional means of depositing material is by fusion welding, which has been used for repair of airfoils since the 1970s. The process can vary from manual TIG to automated plasma- or laser-based systems with the best results obtained from tightly controlled integrated equipment.

With the more readily weldable alloys used in combustors and stator vanes, extensive weld repair of these components is possible, although distortion can become a problem. However, with the higher strength alloys used for rotating blades, the tendency of the base metals to crack is a serious limitation. For this reason, repairs are typically limited to the low stressed regions at the tip of the blades. Furthermore, in more recent engines, including the F class engines, oxidation resistance has also become an important issue.

The experience with the RB211 HP blades shown in Figure 12–39 represent a good example of how the use of superior welding and materials technology can be used to produce a longer life repair, by harnessing the use of higher oxidization resistant materials that are extremely difficult to weld. The same processes were subsequently applied to successfully restore the tips on the Frame 7FA and V84.3A blades discussed below.

Selection of weld filler materials is a balance of the properties required in the joint and the weldability needed. Ideally, welds would be performed with filler metal compositions matching those of the base material, which typically have strengths approaching those of the cast alloy. This can typically be achieved with the cobalt alloys used in many stator components and weaker nickel alloys used in combustors, because the alloys are fairly readily weldable. However, for higher strength nickel-based alloys used for rotating blades, weldability becomes a serious problem and most repairs are performed with weaker alloys, such as Inconel 625, that have good weldability. By limiting repairs to low stressed regions, this approach has been used to successfully extend the life of older generation blades.

In the 1990s, as firing temperatures increased, first in aeroengines and then in the F class engines, oxidation began to become a factor in weld alloy selection. The lower strength alloys used for conventional repair have lower aluminum content and are therefore less oxidation resistant. For example, new Rolls Royce RB211-24C high-pressure turbine blades (Figure 12–39) typically lose material from the shroud edges by oxidation after one overhaul cycle (24,000 hours), necessitating weld repair. However, experience with conventional weld repairs showed that the repaired shroud edges oxidized at an accelerated rate, resulting in their removal from service for further repair after 8000 hours (Figure 12–40).

To select an appropriate weld filler material for such oxidation limited applications, candidate alloys with higher aluminum content were compared to the base material and conventional repair materials by cyclic oxidation testing. The higher aluminum alloys have significantly better oxidation resistance. However, the weldability of these alloys is significantly less than the conventional materials and cannot be reliably applied by manual welding because of cracking problems.

It has been shown that automating welding processes for superalloys reduces their tendency to crack. The automated techniques were therefore developed for repairs with these oxidation resistant fillers. This approach was used to repair the RB211-24C high-pressure blades, with the result that the repaired blades last to the 24,000 overhaul interval in better condition than new hardware (Figure 12–41).

LPM™ Powder Metallurgy

In many cases, welding is not desirable because of excessive distortion or cracking considerations. For this reason, there has been considerable effort in developing non-fusion weld repair techniques. Over the past two decades, many wide gap and modified wide gap processes have been developed and employed in the repair and manufacture of hot section gas turbine components; G.E.’s advanced diffusion healing (ADH), Snecma’s rechargement per brasage diffusion (RBD), Howmet’s effective structural repair (ESR), Chromalloy’s surface reaction braze (SRB), Turbofix, Sermafill, and Liburdi Engineering’s LPM™. All of these processes use high and low melting components to limit the amount of melting point depressant and to help bridge large gaps. However, the patented LPM™ process is the only modified wide gap process capable of filling large cavities resulting from the removal of structural cracks or casting core holes. In particular, the LPM™ process has been used to rebuild components that are missing large amounts of material while maintaining very low levels of melting point depressants.

The LPM™ process is a multi-layer process incorporating high and low melting metal powders to effect sintering and controlled liquid phase bonding. The powders are mixed with organic binders to produce flexible tapes or clay-like putties. The component is then heated slowly in a vacuum furnace to burn off the binders and finally heated to and held at a temperature between the melting points of the high and low melt materials to promote liquid phase sintering, transient liquid phase solidification (isothermal solidification) and diffusion bonding (Figure 12–42). The resulting deposits show excellent mechanical strength, creep strength, and ductility when compared to conventional wide gap deposits, without the need for lengthy diffusion cycles.

Because of its ability to bridge such wide gaps, the LPM™ process can be used to repair very large, structural cracks as well as to build up material on worn surfaces for dimensional or throat area restoration. Since the liquid phase consolidation takes place while the part is at a uniform temperature in the vacuum furnace, there is no lateral shrinkage or distortion. Since its introduction in the early 1990s, the process has been used to repair and upgrade numerous hot section vanes, blades, combustors and transitions.

Experience with the repair of the frame 7EA nozzles demonstrates the effectiveness of the LPM™ process. After an initial 24,000-hour service interval since new, these FSX 414 cobalt alloy nozzles typically show extensive cracking especially around the leading edges, trailing edges and between the two airfoils on the outer shroud. The cracking is primarily caused by low cycle thermomechanical fatigue. The cracking at the leading edges is branched and extended through to the internal vane cavities. Some oxidation erosion and associated cracking also occurs on the outer and inner diameter shrouds between the airfoils. Because of the degree of damage, conventional weld repairs would either crack or result in excessive distortion of the nozzles.

In preparation for LPM™, the cracks are ground out and the eroded surfaces are blended back to unaffected base alloy. This leaves large holes ranging from 25 mm × 100 mm in size at the leading edges of some airfoils to areas larger than 50 mm × 50 mm requiring 0.75 mm build-ups (Figure 12–43). The large holes, which are typical of vane repairs where the structural cracks are removed, were filled with LPM™ putty that was sculpted to the original contour of the nozzle. The eroded areas were covered with a 1 mm thick LPM™ tape to restore the dimensions. Small cracks and impact damage were repaired with the LPM™ putty (Figure 12–44). During repair these areas were successfully repaired using MarM247-7 LPM™.

After the second service interval of approximately 24,000 hours (∼48,000 hours total operating time), the nozzles do not show the large thermomechanical fatigue cracks at the leading edges or between airfoils on the outer shroud surface formed in the new parts. This suggests that the LPM™ had improved the performance of these nozzles due to the improved strength and low cycle fatigue resistance of the LPM™ repair (Figure 12–45).

After 48,000 hours of service, the nozzles had small cracks on the shrouds, oxidation erosion on the internal and external surfaces, and thinned trailing edges. The thinned trailing edges were particularly of concern as the throat area between the airfoils had increased significantly due to material loss during service and repair. This increase in throat area can lead to harmonic imbalance and efficiency loss. As part of the second repair the small cracks and oxidation damage were blended in preparation for LPM™ application and the thinned trailing edges and oxidized areas were thickened by applying LPM™ tape to both surfaces. The cracks were repaired using LPM™ putty. After the second repair the nozzles were again returned to service. The nozzles have since been returned for an LPM™ repair at ∼72,000 hours. After repair, the nozzles will be returned for an additional service interval.

GE Frame 7F

First Stage Buckets

The first stage buckets in the GE frame 7F engine are manufactured from directionally solidified GTD 1 1 1 alloy. They feature a complex serpentine cooling scheme and are coated with an over-aluminized NiCoCrAIY coating on the external surfaces and a simple aluminide coating on the internal passages. The most typical damage requiring repair is oxidation and cracking of the tip (Figure 12–46) and general consumption of the coating life.

The repair workscope typically involves the following steps:

• Incoming inspection

• Strip coatings (internal and external)

• Fluorescent penetrant inspection

• Tip weld repair with oxidation resistant filler (Figure 12–47)

• Finish to contour

• Rejuvenation heat treatment

• Internal and external coating

• Final inspection

Second Stage Buckets

The second stage buckets on the frame 7F suffer from a similar problem as the 7EA buckets. In both engines, the shrouds lift due to the centrifugal forces acting on them, resulting in misalignment of the Z notch contact faces. The repair developed for this type of damage involves applying material by the LPM™ process to realign the contact faces and restoring the contacts surfaces by hard face welding (Figure 12–48). During this repair, it is possible to realign the shroud to lower the stresses and reduce the tendency for future shroud lift.

The typical workscope for repair of these buckets would include:

• Incoming inspection

• Strip coating

• Fluorescent penetrant inspection

• LPM™ reinforcement

• Weld hard facing

• Rejuvenation heat treatment

• Finish to contour

• Reapply coating

• Final inspections

Nozzles

While the geometry of the nozzles in the frame 7F turbine are different than those in the EA model, the materials and construction and general features are very similar. The repair of these components, therefore, is similar to that described for the 7EA components described above. The differences in geometry, of course, require different tooling (Figure 12–49).

The typical workscope for repair of 7F nozzles includes:

• Incoming inspection

• Remove cooling inserts

• Strip coatings

• Fluorescent penetrant inspection

• Welded and LPM™ repair

• Finish to contour

• Rejuvenation heat treatment

• Coating

• Final inspection

Combustors

Since the materials used in the manufacture of combustors are readily weldable, repairs are typically based on welding cracks and missing material. However, there are many applications where distortion may be an issue, such as when surfaces need to be built up over a large area, and in these instances, the LPM™ technique is more suitable. The principal difference between the F class combustors and earlier models is the increasing complexity in geometry. For this reason, having appropriate gauging and tooling is increasingly important (Figure 12–50).

A typical combustor workscope would entail:

• Incoming inspection

• Disassemble

• Strip coatings

• Fluorescent penetrant inspection

• Welded and LPM™ repair

• Finish to contour

• Rejuvenation heat treatment

• Coating

• Final inspection

Siemens V.84.3

Stator Vanes

The stator vanes in the V.84.3 design differ from those in the GE design in several ways. In addition to the obvious geometry differences, these vanes are manufactured from high strength nickel-based superalloys used only for rotating components in GE designs.

From our repair perspective, this means that repair by welding techniques will not be practicable. While welding is possible, the tendency for cracking to occur within the base material will limit the yield achievable by a welding process. By contrast, the LPM™ process avoids the stresses associated with the welding process making full yield repairs possible. Figure 12–51 shows trailing edge cracking in a V.84.3 second stage vane during the LPM repair process.

The typical workscope for repair of V.84.3 stator airfoils includes:

• Incoming inspection

• Remove cooling inserts (as required)

• Fluorescent penetrant inspection

• LPM™ repair

• Finish to contour

• Rejuvenation heat treatment

• Coating

• Final inspection

Alstom 11N2

Combustor Hardware

Because the combustor in the 11N2 engine is of a silo type, the components appear to have little similarity to those in a GE engine (Figure 12–52). However, in both instances, the components are manufactured from readily weldable materials, so the repair workscope is actually very similar. The components typically suffer from wear and cracking, which can be repaired by welding. Surfaces that are protected by a thermal barrier coating will have the coating removed and reapplied as part of the repair process.

The typical workscope for repair of 7F nozzles includes:

• Incoming inspection

• Strip coatings [as applicable]

• Fluorescent penetrant inspection

• Welded and LPM™ repair

• Finish to contour

• Rejuvenation heat treatment

• Coating

• Final inspection

Microstructure

The microstructure of the original IN738 LPM™ material contained elongated borides. The 100 μm to 500 μm borides were inter-granular and had length to width ratios of greater than 10:1. This microstructure resulted in lower ductility but tended to perform well in service.

Both of the newer MarM247 LPM™ compositions have small cuboidal boride particles. The 10 μm to 50 μm cuboidal particles are found intra-granularly and have a length to width ratio of about 1:1. As a result, these two new compositions show improved ductility and better fracture properties than the IN738 LPM™ material. The MarM247-7 composition has a slightly higher volume fraction of borides than the MarM247-3 but both have lower boride volume fractions than the IN738-based material. The borides in the MarM274-3 structure tend to be smaller than those in the MarM247-7 structure.

All LPM™ compositions show a good interface with the base alloy. LPM™ materials have been applied to the following alloy systems: GTD111, GTD222, Rene N4, IN738, IN939, IN625, MarM002, MarM247, Rene 80, Rene 125, Rene 142, Hastelloy X, CMSX 4, C1023, U500, U520, U700, U720, 304 stainless steel, 321 stainless steel, 347 stainless steel, ECY768, X40, X45, FSX414 and others. No needle-like topologically close packed phases (TCP), microcracks, or abnormal phases were found at the interface between LPM™ and any of the aforementioned alloys. The interface between LPM™ is generally a 50 μm interdiffusion zone that contains a coherent interface with the base alloy.

Microscopic examinations of large cross-sections of the original IN738 LPM™ and the newer MarM247-based compositions have shown that any pores tend to be smaller than 125 μm in size and total porosity levels are lower than 1%. As the LPM™ material is produced from fine powders, the resulting microstructure has a very fine grain size.

Composition

Due to the multi-layer nature of the LPM™ process the resulting deposits have a very low weight percent of boron, the melting point depressant used in the low melt component (Table 12–4). As boron is a very light element having an atomic mass of 11 compared to 59 for nickel, small weight percent changes in boron result in very large changes in the atom percentages. Most conventional wide gap processes contain 1.5 weight percent boron or more. The LPM™ materials contain less than 0.9 weight percent boron due to the deposits containing less than 30% low melting material. The small weight percent and subsequent atomic percent of boron in the LPM™ materials leads to low volume fractions of boride and better ductility than other brazing processes. The less than 30% low melting material also results in high levels of γ′ forming elements and subsequently excellent creep properties as is evident in the relatively large amounts of Al and Ti in the LPM™ material when compared to conventional wide gap materials.

Test Bar Manufacture

Most of the studies done on conventional and modified wide gap processes have concentrated on joint widths not larger than about 1.5 mm. These small joints widths allow for large amounts of boron to diffuse out of the joint into the base material. Often the joints are given long diffusion cycles of up to 24 hours to allow for the diffusion of the boron into the base alloy. Thus the data generated from these test bars is not representative of the mechanical properties of overlays or relatively thick repair joints formed on actual component repairs. It was felt that for this testing the butt joint should be 6.3 mm wide such that diffusion into the base alloy did not play a role in the joint properties at the centerline of the joint. The test bars with a 6.3 mm joint width should be representative of the bulk LPM™ properties and should also test the interfacial strength.

6.3 mm × 6.3 mm slots were cut down the center of several 50.8 m × 50.8 mm × 9.5 mm MarM247 blocks. These slots were filled with the MarM247 LPM™ materials and heat treated for not more than 2 hours at the maximum temperature reached during the alloy solution cycle. The blocks were then aged for 2 hours at 1120°C and 16 hours at 843°C to produce the desired γ′ precipitate structure. Stress rupture and tensile bars were cut from the blocks such that a 6.3 mm region in the center of the gauge was LPM™ (Figures 12–54 and 12–55). The bars were X-ray and fluorescent penetrant inspected. Test bars with defects larger than 500 μm were rejected.

Tensile Properties

Tensile test bars were tested at a variety of temperatures. Table 12–5 shows the ultimate tensile strength and reduction in area for each of the test bars. The MarM247-3 material shows slightly better ductility due to the small boride particles and the smaller volume percent of borides. The ductility of the MarM247 based LPM™ materials is only slightly lower than that for common nickel-based superalloys and is considerably higher than the levels for conventional wide gap brazing processes that generally show little or no ductility. None of the bars failed at the interface and all bars failed intergranularly as would be expected. The modulus for both materials was approximately 170 GPa.

Creep Properties

Stress rupture tests were performed using the bars described previously. The tests were conducted to ASTM E139. The stress time and temperature were varied to generate a broad range of Larson Miller parameter values. The temperature varied from 787–982°C and the stress varied from 5–35 MPa. The Larson Miller plot is shown in Figure 12–56. The average reduction in area was 4.2% for the MarM247-7 LPM™ and 7.3% for the MarM247-3 LPM™. These values are similar to those for cast MarM247.

Due to the fine-grained nature of LPM™ materials, they tend to have lower creep strengths when compared to large grained cast nickel-based superalloys. The LPM™ materials have creep strengths higher than conventional welding alloys such as IN625 for nickel-based superalloys and L605 for cobalt-based superalloys.

The creep strength compares more closely to fine grained wrought superalloys and cobalt-based superalloys. During the studies it was also found that the creep properties of LPM™ varied little with the composition. It has been hypothesized that the creep properties are highly dependent on the grain size. Studies to this effect will be conducted in the future by increasing the grain size through long diffusion cycles.

Low Cycle Fatigue

Low cycle fatigue test bars were produced as described previously and were inspected for surface defects larger than 0.50 mm. The testing was conducted at 850°C using a trapezoidal waveform at an R ratio of –1.0. The specimens were held at the maximum compressive strain for 2 minutes during each cycle. Failure was defined as the point at which the maximum stress amplitude decreased to 50% of the value at 100 cycles. It should be noted that the butt joint in the middle of the test bars produces an effective gauge length of only 6.3 mm and the strain is measured over the total gauge length. During testing the majority of the elongation occurs in the 6.3 mm joint. This means the strain in the joint area is much higher than that measured. The data reported in Figures 12–57 and 12–58 are for the total strain over the gauge length of the test bar. Testing indicated the fatigue properties are similar to that of cast MarM247 for MarM247-7 and likely slightly lower properties for the MarM247-3.

Superalloy LPM™ Example #1: General Electric MS7001EA Row 1 Nozzles

After the initial 24,000 hour service interval, these FSX 414 cobalt alloy nozzles show extensive cracking especially around the leading edges, trailing edges, and between the two airfoils on the outer shroud. The cracking was primarily caused by low cycle thermomechanical fatigue. The cracking at the leading edges was branched and extended through to the internal vane cavities. Some oxidation erosion and associated cracking had also occurred on the outer and inner diameter shrouds between the airfoils. Based on the degree of damage it was felt that conventional weld repairs would either crack or result in excessive distortion of the nozzles. It was decided the LPM™ was the best approach for repair.

In preparation for LPM™ the cracks were ground out and the eroded surfaces were blended back to unaffected base alloy. This left large holes approximately 25 mm × 100 mm in size at the leading edges of the airfoils of several nozzles and areas larger than 50 mm × 50 mm requiring 0.75 mm build-ups (Figure 12–59). The large holes, which are typical of vane repairs where the structural cracks are removed, were filled with LPM™ putty that was sculpted to the original contour of the nozzle. The eroded areas were covered with a 1 mm thick LPM™ tape to restore the dimensions. Small cracks and impact damage were repaired with the LPM™ putty. During repair these areas were successfully repaired using MarM247-7 LPM™.

After the second service interval of approximately 24,000 hours (∼48,000 hours total operating time), the nozzles did not show the large thermomechanical fatigue cracks at the leading edges or between airfoils on the outer shroud surface. This suggested the LPM™ had improved the performance of these nozzles due to the improved strength and low cycle fatigue resistance of the LPM™ repair (Figure 12–60).

After 48,000 hours of service, the nozzles had small cracks on the shrouds, oxidation erosion on the internal and external surfaces and thinned trailing edges. The thinned trailing edges were particularly of concern as the throat area between the airfoils had increased significantly due to material loss during service and repair. This increase in throat area can lead to harmonic imbalance and efficiency loss. As part of the second repair the small cracks and oxidation damage were blended in preparation for LPM™ application and the thinned trailing edges and oxidized areas were thickened by applying LPM™ tape to both surfaces. The cracks were repaired using LPM™ putty. After the second repair the nozzles were again returned to service. The nozzles are expected to return for LPM™ repair at ∼72,000 hours in early 2004.

Superalloy LPM™ Example #2: Siemens Westinghouse W501G Row 1 and 2 Core Closure

Liburdi Engineering’s LPM™ is specified for closure of the core holes on Siemens Westinghouse 501G row 1 and row 2 directionally solidified MarM247 turbine blades. The casting core holes in the 501G blades are necessary to retain the core during manufacture of the blades. These large core holes are difficult to close via welding due to cracking of the directionally solidified grain boundaries. Conventional wide gap brazing of the 25.4 mm × 6.3 mm casting holes is also not practical (Figure 12–61). LPM™ was selected to close the blade tips as it has excellent creep properties, very low porosity, and can close the casting holes without generating cracks.

LPM™ is applied as putty into the core holes such that each hole is slightly overfilled. The blades are then heat treated to consolidate the LPM™. After finishing to contour and machining the tip to nominal dimensions it is not possible to visually determine where the casting holes were (Figure 12–62). To date over 49 sets of row 1 and row 2 blades have been closed using this process. Figure 12–63 shows a blade tip after more than 8000 hours of service.

Wear Resistant LPM™

Wear resistant LPM™ (known as LZN™) compositions based on chrome carbide have been developed and used in the manufacture and repair of hot section gas turbine components. These materials are composites of hard particles in an oxidation resistant matrix. The wear resistant LPM™ was designed to replace plasma, high velocity oxyfuel (HVOF) and weld hard facing materials.

The major drawback of plasma and HVOF chrome carbide coatings is their poor mechanical bond to the base alloy. The LZN™ compositions form a metallurgical bond with the base alloy. This metallurgical bond consistently results in failure of the adhesive in the ASTM C633 adhesion test for flame sprayed coatings. The other drawback of plasma spray and HVOF coatings is the necessity for 90% spray angles near 90% to the surface. The LZN can be applied into slots on opposing surfaces and on surfaces that have no direct line of site for a plasma or HVOF gun.

In comparison to weld hard facing, the LZN™ material does not result in base alloy cracking and shows improved hardness. Weld hard facing materials often result in cracking at grain boundaries intersecting the face to be welded. In some components welding results in cracks at or beyond acceptable limits resulting in high levels of deviated or scrapped material. Due to the nature of the LPM™ process LZN™ does not result in cracking of the base alloy. During weld hard facing, the weld pool mixes with the base alloy reducing the hardness of the deposit reducing hardness values to be below 50 Rc (Rockwell “C”). The LZN™ materials have higher volume percents of carbide particles resulting in hardness values of 60 to 65 Rc. The volume percent of carbide is more than 55% of the deposit. Voids are typically smaller than 25 μm and the porosity levels are below 5%.

Because the LZN™ material can be used for hard-facing of shroud contact faces, transition piece outlets, floating seals, etc., it must be oxidation resistant. The composition of the material is shown in Table 12–6. The chrome carbide particles are expected to start decomposing above 815°C and it is felt that the oxidation resistant matrix should be stable to temperatures above 980°C. Oxidation testing at 870°C for 1000 hours showed a native chrome oxide scale 2 μm thick had formed on the deposit and no significant attack of the deposit had occurred.

The LZN material has been used to hard face the shroud contact faces on CMSX 4 single crystal turbine blades. The blades in question had no direct line of site for plasma or HVOF hard facing and the geometry of the contact face did not allow access for a gas tungsten arc weld hard face. As a result LZN was chosen to replace the conventional hard facing techniques. The LZN material is applied to the contact faces as an LPM™ tape material and is given appropriate heat treatments to consolidate it. The LZN material showed no abnormal phases or unusual microstructures at the interface with the CMSX 4 base alloy. The material was machined to contour using conventional aluminum oxide grinding wheels. After machining the contact faces did not show any cracking or spalling and the LZN had successfully hard faced the z-notch on these blades with a very difficult geometry.

Abrading LPM™

Abrading LPM™ compositions based on large angular hard particles have been developed for use on the tips of unshrouded turbine blades. Unshrouded turbine blade tips run against shroud blocks to ensure a tight air seal and subsequently high engine efficiency. Unfortunately the turbine blades are often worn by the shroud blocks as opposed to the shroud blocks being worn by the turbine blades. This results in the loss of material from the tips of the difficult to repair turbine blades. It also results in the removal of any protective coatings from the blade tips. A need was recognized for a harder material that would machine into the shroud blocks, preventing the blade tips from being worn away and prevent the blade tips from oxidizing.

The abrading LPM™ material is a mixture of large hard particles in an MCrAlY-based matrix. This material is made into an LPM™ tape and applied to the blade tip. The tape is contoured to the blade tip in the green state and is given an appropriate heat treatment for the base alloy. After heat treatment the abrading tip is contoured and ground to the appropriate tip height.

In summary, Superalloy LPM™ compositions based on IN738 and MarM247 have proven extremely successful for the repair and manufacture of hot section gas turbine components. These materials show excellent creep, tensile, and stress rupture properties for a brazing process. The materials have proven successful during service on a variety of components from different engines.

Hard particle LPM™ compositions used for wear resistance and abrading have also been developed and tested. The wear resistant LZN™ material shows very good hardness and can be used in wear applications where welding, plasma spray and HVOF hard facing methods do not work. The abrading LPM™ material can be used on unshrouded turbine blades where tight tolerances and wear of the tip shrouds into the blade tips is an issue.

Life Assessment

Based on the design philosophy described above, the residual life assessment encompasses stress analysis, operation, and repair history, quality control as well as material investigations of the service-exposed parts.

A preliminary step for plant life extension involves the compilation of all operating data such as the number of starts, hours of operation, and rapid temperature changes (e.g., due to load rejection or turbine trip, Figure 12–65).

The history of maintenance measures and the trends of findings of the most recent fact-finding on minor and major inspections must also be accounted for.

Following selection of the components to be considered, the corresponding original design criteria and corresponding analysis are used and updated if necessary.

Material investigation corresponding to the design criteria includes both non-destructive evaluation and destructive testing of components:

Crack examination yields preliminary indications of exceeding high-cycle fatigue strength. In the event that cracks occur due to cyclic loading, these can be assigned to loading based on the crack initiation site and crack propagation. If the blade is free of cracks, the design strength and the high-cyclic loading limits are used for further evaluation. An example of such investigations is described below.

Based on these material investigations, the original design and mode of operation of the gas turbine, and reusability of the component(s) in question are evaluated from a technical standpoint. Following consideration of the costs and benefits of measures such as requisite repairs, the final decision is made regarding continued use or replacement.

Stage 1 and 2 of the Turbine Blading

Rotor Blade Stage 1 (Tla1)

The original design of the Tla1 was developed in the late seventies. This blade already uses—compared to the state of the art in those years—an advanced serpentine type convective cooling channel. Turbolators have been added to the cooling configuration over the years. The cooling holes at the trailing edge had to be electrochemically manufactured (ECM) at that time. Since then, several steps of increasing turbine inlet temperature have been carried out. The original design revealed several findings during the refurbishment after 33,000 EOH of operation. This led to the decision to limit the service lifetime to one hot-gas-path inspection interval of 33,000 EOH. The findings were mainly:

• Considerable loss of wall thickness due to oxidation at the outer surface for the chromized type of this blade

• Significant internal oxidation especially at the trailing edge that caused certain cracking at the cooling holes in combination with cyclic operation due to LCF

• Occasional LCF cracking of the leading edge due to cyclic operation

Therefore, about 6 years ago this blade was redesigned with several features. Due to the progress in casting technology it was feasible to cast-in cooling outlet slots instead of the ECM manufactured holes. This measure resulted in a reduced concentration factor, thus significantly improving the low cycle fatigue strength. The outlet slots were provided with six rows of cylindrical pins to connect the suction side and the pressure side. These pins increased turbolance effects and enlarged the cooling surface for a more effective heat transfer.

The cooling of the leading edge was improved by a separate cooling channel with turbulators and a cooling air outlet at the blade tip. Additionally this feature made the airfoil more resistant against potential foreign object damage.

Additionally this blade was protected with the MCrAlY type vacuum plasma spray coating SICOAT 2231 on the outer surface.

A systematic life assessment program was carried out on service exposed rotor blades stage 1 after 33,000 EOH of operation to verify these improvements.

Samples from base load units revealed excellent coating behavior even at the leading edge, which is the position of the highest material temperature. It retains still usable coating and therefore potential for longer refurbishment intervals.

A unit being operated at intermediate duty with numerous start-stop cycles over one HGP inspection interval showed comparable excellent coating behavior. Only at the leading edge was the coating nearly consumed while other regions looked as good as new.

The leading edge of the upgraded design, opposite to the original design, showed no crack indications after comparable operating conditions. Also no notable internal oxidation was found with the upgraded design. This is of special interest for the life of the trailing edge. The trailing edge of the upgraded design was found crack free in all cases.

In addition a microstructural temperature evaluation by investigating the coarsening of γ′-phase was performed for the leading edge. In accordance with the computational analysis this investigation showed that the material temperature is about 50°C lower compared to the original design (Figure 12–66).

With the upgraded design the usable life of the Tla1 has been extended by 100% from one to two HGP inspection intervals including the refurbishment after 33,000 EOH.

Rotor Blade Stage 2 (Tla2)

The MCrAlY type SICOAT 2231 protective coating was implemented to the outer surface of rotor blade stage 2 as well. On samples taken from such blades after service, excellent coating and base material conditions have been investigated after one HGP inspection interval of 33,000 EOH in every case. One example in Figure 12–67 shows a still usable coating thickness of >150 m and no base material attack, as was the case with the chromized type. Therefore, with this enhanced coating, comparable to the Tla1 the usable life of the Tla2 has been extended by 100% from one to two HGP inspection intervals including the refurbishment after 33,000 EOH. The enhanced coating is not only applied on new blades but also available for the blades of the original design.

Stage 3 and 4 of the Turbine Blading and Hot-Gas-Path Casings

Stator blades in stages 3 and 4 undergo residual life assessment when reaching the 100,000 EOH milestone, as they have potential for continued service. Highly influenced by the operating and repair history and also depending on the material and design variant, the inner casing and the mixing casings are assessed for their residual life after 100,000 EOH.

In summary, life extension options for hot components during regular HGP inspections and after long operating periods 100,000 EOH were studied.

Repair methods and protective coatings have been continuously improved for those hot components that need regular life extension measures during hot-gas-path inspections.

The investigations for life extension of hot-gas-path components after the long operating period of 100,000 EOH have revealed that several hot-gas-path components have definitely reached the end of their usable life, while others, e.g., rear rows of the turbine blades, may be usable for one more 33,000 EOH period or even longer.

Plant life extension including hot-gas-path life extension of V94.2 gas turbines has been applied on 11 gas turbines V94.2. Also a large number of additional plant life extensions are due to take place in the coming years. While the basic statements for life assessment and life extension are valid for the whole V94.2 fleet, it must be emphasized that each gas turbine needs to be evaluated individually because of its individual operating conditions. Up to now, life extension projects involved the early versions of V94.2 with relatively low turbine inlet temperature (TIT_ISO) × 1000°C and with the original hot component design, while many of the future projects will involve a higher TIT_ISO > 1000°C for base load operation with different hot component design.

The most profitable plant life extension projects:

The best way is to evaluate these items in a joint project involving the electric power producer and the gas turbine manufacturer.

Gaseous Fuel

In general, gaseous fuels are considered more suitable than liquid fuels for continuous base-load operation. Fuel quality has an important role in performance degradation of gas turbines. The ideal fuel is a clean, dry gas, with a low sulfur content, which can originate from different sources: pipeline quality natural gas, vaporized LNG or LPG, and process synthesis gas.

Fuels containing traces of metal create deposits on the nozzles and blades, impacting GT integrity and performance. In one particular application, Ni deposits concentrated on the tip of the burner forced a burner redesign. In many other cases, the deposits plug the blades’ cooling holes and create areas of overheating, which eventually lead to cracks in the material. Special coatings on the blades are often used to minimize the corrosion. Another negative effect of deposits, as mentioned above, is accelerated thermal performance degradation. In the current market, with gas prices near record levels, increased demand, and strained supplies, there is little incentive for the producers to maintain a consistent gas quality. Traditionally, gas liquids, such as propane and butane, were removed and sold as chemical feed-stocks because their value was higher than in the fuel gas market. Not any more.

Some experts estimate that as much as one third of the gas produced in the US is no longer processed to remove liquid hydrocarbons. Therefore, the physical and chemical characteristics of pipeline gas are significantly different from the “original design fuel.” This “out of specification fuel” can create mechanical problems and accelerate the degradation process.

Liquid Fuel

Heavy-duty gas turbines are capable of burning a variety of liquid fuels, ranging from light petroleum distillates to heavy residuals. The combustion process of light distillates such as naphtha, kerosene, and No. 2 fuel oil is well understood and has been proven through millions of hours of operation. The degradation rates on these types of liquid fuels are predicted to be 20% higher than on natural gas fuel. The real challenges for GT design and operation come from the use of heavier ash-forming fuels. The combustion of these fuels can be extremely detrimental to GT life, due to corrosion, erosion, and fouling of the blades. Hot corrosion, caused by traces of elements such as sodium (Na), potassium (K), lead (Pb), and particularly vanadium (V) in the fuel, has a major effect on GT performance and the service life of the hot gas path. The number of equivalent operating hours (EOH), which serves as the basis for degradation and maintenance intervals, increases more rapidly based on fuel contaminant concentration, accounting for a factor of 1.5 for each operating hour (by certain estimates).

Water

The effect of water on GT degradation depends on its use and point of entry. At the inlet of the compressor, droplets of water generated by the fogging system may cause erosion on the blades and corrosion of the inlet ducting and plenum. Water and steam injection in the combustion system is used for NO_x reduction. On one hand, the increased hot gas mass flow causes a rise in the pressure ratio and power output. Thus, the mechanical loading of the blades is higher. On the other hand, the higher moisture content in the hot gas affects the heat transfer properties and consequently increasing the metal temperature of the blades. Finally, exposure to gases with higher moisture content reduces the wear resistance of the turbine blades' coatings. Therefore, a weighting factor on the number of operating hours should be applied as a function of the amount of water injected.

Marine and Offshore Environment

The airborne salt in different forms (aerosol, spray or crystal) is impacting the fouling of the compressor. High environmental relative humidity causes the salt to remain as supersaturated droplets. The database concerning the amount and form of the salt content indicates considerable variation and dependency on the wind speed on the level of contamination.

Recoverable and Non-Recoverable Degradation

The process of degradation cannot be avoided, but its impact on performance and hardware life can be mitigated by undertaking a number of preventive actions, starting with conceptual design, and continuing through the project execution and operation and maintenance activities. Here it is appropriate to introduce the definitions of recoverable and non-recoverable degradation. The recoverable degradation is considered the portion of the GT performance, which can be restored by a series of maintenance activities without opening the machine or replacing major parts. This degradation recovery is commonly achieved by conducting compressor washing and minor engine adjustments (such as resetting the inlet guide vanes). The remaining degradation is regarded as non-recoverable. Equipment manufacturers predict the non-recoverable degradation for the equipment in the form of degradation curves.

Figure 12–68 provides an example of recoverable degradation by online and offline compressor water wash. It should be noted that these notions of recoverable and non-recoverable degradation could be confusing, since after major overhauls, the GT performance can be restored to almost the level achieved in “new and clean” conditions.

Limits on allowable degradation, recoverable and non-recoverable from base load conditions, should be established with the equipment manufacturers. Operating below these limits increases the risk of compressor surge. In the project development phase, the impact of the recoverable and non-recoverable degradation, scheduled outages, etc., should be duly accounted for in the project pro forma analysis.

Fuel Treatment

During the initial design phase, actual specifications should be provided for fuel quality entering the gas turbine. If GT manufacturer fuel requirements cannot be met, a comprehensive precombustion fuel treatment must be developed and implemented, aimed at reducing the concentration of contaminants. It is highly recommended to implement for gaseous fuels an online, real-time, continuous composition monitoring system, based on gas chromatography. The pretreatment process for liquid fuels should involve chemical reactions and/or mechanical systems. Na and K salts are soluble in water and can be removed by the on-site treatment known as fuel washing. Other methods of fuel treatment include the use of high-temperature corrosion additives. To reduce the amount of expensive additives, a preliminary step for improving fuel quality entails using of commercially available mechanical fuel treatment skids composed of centrifugal separators, pumps, and sludge tanks. The devices can reduce the concentration of harmful contaminants (Na, K, S, and V). A typical two-stage system can reduce the Na content from 150 ppm to 1 ppm for heavy salts.

Air Filtering System

The air filtering system must have the capability to eliminate 99.95% of all particulates above 20 μm, since these particles are the main source for compressor erosion. Installations located in the desert, in coastal areas or near dust emitting sources (coal dust, cement dust, fly ash, and pollen) should be equipped with self-cleaning systems that have an effective dust removal system. The ambient air for many combined cycles located close to refineries contains air-borne heavy hydrocarbon matter that eventually could sinter on compressor blades and cause non-recoverable degradation. In the absence of an efficient and proper designed inlet system, the ingestion of humid and salty air at a seaside location may lead to erosion and corrosion in the compressor as well as in the turbine section of the gas turbine. Finally, it is recommended that an ambient air quality analysis be conducted during initial site development, as extremely useful information.

Degradation Curves

To cover the gap from the “new and clean” performance guaranteed by the manufacturer to the performance measured during the test, equipment degradation curves have been created. The predicted degradation curve, in logarithmic shape, is a function of operational hours. This trend represents the non-recoverable degradation, which cannot be restored without opening of engine covers to refurbish the components back to a new and clean condition. Such events occur during hot gas path inspection or major overhaul. This cumulative degradation curve represents the contributions of each individual engine components. Since degradation, as described above, is not simply affected by the number of fired hours, some GT manufacturers created a methodology where degradation is a function of EOH, a weighted number of hours accounting for different operating conditions: type of fuel, number of starts, stops, trips, sudden changes in load, and use of diluents. For example, a normal start adds 10 EOH and each hour operating with water injection is considered to add 1.5 EOH. An hour of peak load operation at increased firing temperature equals 3 EOH. Some manufacturers indicate that each trip could represent between 25 and 400 EOH, depending on the GT load at the time of the trip. The determination of the weighting factor for each of the operational conditions on the EOH value is done solely by the equipment manufacturers. It is assumed that these factors are determined on the basis of the fleet performance experience and inhouse modeling.

Operations and LTSA (Long-Term Service Agreement)

Assessing the GT degradation is important not only for acceptance performance tests but also for day-to-day operations. Asset managers/operators should carefully plan the operation and scheduled maintenance of the gas turbine, to keep the number of EOH as low as possible. However, in the current opportunistic power market, decisions on the way to operate the equipment are guided by commercial considerations rather than technical. To mitigate the uncertainty associated with maintenance costs, many owners enter in a maintenance contract often called a long-term service agreement (LTSA). An LTSA typically commits the OEM or non-OEM service contractor to provide, on a relative, fixed price basis, parts and maintenance services for gas or steam turbines.

Commercially, an LTSA should offer many advantages to owners, including the predictability of relatively fixed long-term costs, while keeping the OEM motivated through performance bonuses to provide support and ensure a predetermined level of performance (see Thomson, R., and J. Yost. “Avoiding Pitfalls in Service Contracts.” Turbo Machinery).

An LTSA must be clearly written and discussed, so that the parties involved do not have false expectations about the extent of its coverage. In many LTSA the validation of the actual GT degradation is an important element of the agreement. The suggestions and recommendations concerning the evaluation of performance degradation made above should be implemented in the LTSA documents and carried out in the field.

Engine Modeling

To simulate the blade- and disk-operating environment, an aero-thermal model for the power turbine was developed. The aero-thermal model is a two-stage turbine mean-line program with the gas path geometry constructed from reverse engineering actual hardware. The aero-thermal program calculates hot gas path temperature, pressure, and flow across the turbine stages using the published power turbine performance data at the rated conditions. The calculated aero-thermal results relative to the stage 1 blades were then used as the boundary conditions for the blade heat transfer analysis.

Disk Cooling

The operating temperature of the first stage power turbine disk is strongly dependent on the cooling air temperature inside the front disk cavity. The power turbine rim cooling system is a series of pipes, connectors, flexible hoses, drilled nozzle vanes and temperature probes. Cooling air is bled from the gas generator intermediate compressor stage to the power turbine first stage disk rim and nozzles. It is designed to protect the first stage turbine disk from the extreme heat of gas generator exhaust gases.

From the power turbine historical operating data, it was found that the measured disk cavity temperature is a function of power turbine inlet temperature (Figure 12–70). The average measured disk cavity temperature based on the typical power turbine load conditions was used in the analysis.

The case goes on to detail work that involved blade/ disk heat transfer calculations, peak stress, stress relaxation, creep rupture, probabilistic modeling, and then life expectancy calculations. Comparisons can then be made between different turbine disks. The calculations are too lengthy to repeat in this case.

In summary, an engineering approach has been developed to analyze real life expectancy of aged power turbine disks. The approach integrates mechanical design, analytical calculation, aero-thermal analysis, finite element modeling and probabilistic modeling of material properties. Applying the approach to a power turbine disk with 183,400 service hours, it was found that the disk could run for an infinite life assuming continued operation at the current conditions.

This conclusion is consistent with the result of a non-destructive examination conducted prior to the analysis. The dimensional inspection indicated no measurable plastic deformation per engineering drawings, while fluorescent penetrant testing found no cracking. The good dimensional stability after 27 years of operation suggests that it is unlikely the disk will fail due to primary stress related ductile bursting in the near future.

Since the power turbine inlet temperature and cooling air system have a significant impact on the disk metal temperature, and hence the life, operators should implement a fleet-wide monitoring program that would include the disk cavity temperature measurement and periodic inspections for any unforeseen damage mechanisms to ensure safe long-term operation of their power turbine disk fleet.