Turbo Machinery Dynamics Part 13 pptx

stresses at the same temperature load, making the segmenting network finer and individual cracks to become wider under cyclic loading. The amplitude and frequency of the acoustic signals during the heating phase decrease with the number of cycles. In-plane tensile as well as compressive stresses are present in the coating, depending on the depth. Parallel cracks occur because the coatings are applied in layers, and can grow because tensile stresses are present at their end. The formation of delamination cracks depends on the oven conditioning to a much greater degree than is the case with vertical and parallel cracks. Few delamination cracks occur on specimens that are not conditioned, even when subjected to a large number of cycles. Aging by cyclic laser irradiation does lead to compaction of the ceramic coating, but the relatively cooler bond coat results in only a slight growth in the oxide coating. Consequently, the cyclic laser aging promotes the growth of delamination cracks to a lesser extent than a preced- ing oven aging. Nickel-based alloys are used for aeroengine blades with a protective nickel-aluminide diffusion coating for oxidation resistance at high temperatures during operation. The effect of the presence of the coating on the operational life of single crystal superalloy blades is an issue of concern, in large measure due to the oxidation of the coating’s con- stituents. Under combined mechanical and thermal load conditions that mimic the strain- temperature behavior at critical locations in blades, the presence of an aluminide coating on SRR99 results in substantial life reduction at 0.7 percent mechanical strain (Bressers et al., 1996). Corroboration of the difference in life between the noncoated and Ni-aluminide coated samples is provided by Johnson et al. (1997). Cylindrical bars of the Ni-based superalloy SRR99 with the long axis oriented 10° of the <001> direction are used in the evaluation. The coating has two layers. The outer layer is a polycrystalline B2-NiAl (3–6 µm grain size) + g ′ of thickness 23 ± 2.5 µm. The inner subcoat diffusion zone (17.5 ± 2.5 µm thick) is followed by a continuous layer of g ′ (2 to 3 µm thick). The coated and uncoated samples are subjected to 0.7 percent mechanical strain and temperature cycle varying between 300 and 1050°C. Digitized light microscope images of the coating surface are obtained during the tests with a video camera. Incremental polishing of the images allows the characterization of the coating and of the cracks as a function of depth over selected areas. Images are then digi- tized for measurement to obtain density, spatial characteristics, and number of cracks. Energy dispersive spectroscopy is used to chemically define the surface. The cracks in the uncoated sample generally initiate at or near the surface following the formation and concentration of the oxides that may be described as a discontinuous oxida- tion fatigue process. Some cracks emanate from the top and base of the spikes, but the event is generally difficult to detect. After the test sequence, the cracks grow enough in size to permit identification and correlation with the earlier-detected indications. A simplified sequence of events for the crack initiation procedure may be sketched by The fraction of the oxidized area to the local regional area may be used as a gauge for damage initiation and growth studies. Both oxide spikes and cracks behave as rigid inclu- sions at the sample surface under compression. Since a strain gradient is formed between the inclusion and the alloy, both features contribute to discontinuous oxidation process on the surface. Figure 11.20 shows the experimental data, where contributing factors to the oxidation process are expressed by the Avrami type rate equation ξ = 1 − exp(–K⋅t n ) (11.4) where K = 3.27 × 10 −11 and n = 2.6. Values of n in this range indicate that the initiation rate Casting defect Oxide spike Cracks Oxidation fatigue Tensile failure →→ SUPERALLOYS FOR TURBINES 463 FIGURE 11.22 Surface event initiation rate (Johnson et al., 1997). FIGURE 11.23 Uncoated sample with oxide layer removed (left), oxide concentration area (right) (Johnson et al., 1997). FIGURE 11.20 Initiation rate of discontinuous oxidation and accumulation (Johnson et al., 1997). 464 FIGURE 11.21 Time sequence of surface changes in coated sample (Johnson et al., 1997). is decreasing, and the overall transformation is mostly two-dimensional. Parameter K defines the character of the mixed surface damage process under different conditions. A time sequence of images of the surface for the coated sample is shown in Fig. 11.21. A large number of bright particles are noticed around 1000 cycles, and may correspond to oxidation products of coarsened coating constituents. Few cracks are detected before 2000 cycles. Accumulation of initiation events of surface features larger than 50 mm for the uncoated and coated test specimens are shown in Fig. 11.22. Larger concentration of oxi- dation features may be considered a precursor to cracking. Initiation rate in the coated sam- ple surface changes considerably between 2000 and 7000 cycles, and tends to saturate well before the end of the test. The impact of the coating on fatigue life is most noticeable in the growth rate of cracks until eventual failure. Growth rates are similar for the coated and uncoated test pieces up to 5000 cycles, when the rate accelerates in the coated sample. In the coated piece, failure is encountered after 9469 cycles, while the uncoated piece is good for 21,820 cycles. This observation may be explained by the coalescence of crack events, leading to rapid extension of the major crack. In the posttest examination, surface oxide is removed to expose oxidation areas and cracks. Figure 11.23 shows the front surface at a depth of 80 µm and a section through an oxidation region. A central nickel oxide area forms to create concentric layers of oxidation products of Al, Cr, and Ta. In the coated mate- rial, three main layers of differing damage and microstructure are identifiable. At 24 µm depth, damage is chiefly in the form of oxidized cavities in the coating, composed mainly of Al rich and lesser extent of Ni compounds. At 37 µm depth, cracking predominates at the boundaries of Al and Ni oxides. The presence of coarse grain boundary particles sug- gests the coating may offer greater resistance to creep. As the subcoating is pene- trated, some cracks indicate association with substrate solidification defects. At 65 µm depth the continuous g ′ layer is passed, but the microstructure does not fully match that of the bulk portion. Etching revealed that the Ni and Al oxides in the coating are closely twined. 11.11 FIBER-REINFORCED CERAMICS FOR COMBUSTOR LINER Improved turbine efficiency can be achieved through higher temperatures at the inlet, and hence the emphasis on effective cooling and high-temperature resistant materials for the components of the hot path. Carbon- and silicon-carbide-based materials offer such a potential. But the advantageous qualities of monolithic materials in structural applications, such as Young’s modulus, thermal conductivity, and oxidation resistance, are partly offset by their relatively low fracture toughness. Additives and particulate reinforcements have been tried as toughening agents. Reinforcement with SiC or C continuous fibers increase the fracture toughness by one order of magnitude and fracture energy by two orders of magni- tude (Helmer, Petrelik, and Kromp, 1995). The hostile stress and temperature environment restricts the choice of toughening phases because of chemical incompatibility and thermal expansion coefficient mismatch with the matrix. MTU of Munich, Germany has researched the composite materials in an experimental investigation (Filsinger et al., 1997). The materials are selected from a group of C, SiC, and fiber-reinforced glasses. Under an oxidizing atmosphere, fiber degradation may be expected in the two-directional composites. Assuming a minimum tensile strength of 150 MPa to be sufficient for reliable operation of the component, the durability in a 1000°C environment SUPERALLOYS FOR TURBINES 465 would be only 8 min for the C/C material and 6 min for the C/SiC material. This clearly shows the need for effective external oxidation protection. The coatings influence the mechanical properties of the base material, and hence the tests are conducted for coated specimens. A cross-sectional view of the ceramic can form of combustor is shown in Fig. 11.24. Combustor walls are made of an inner layer of a hot-gas resistant composite, a middle layer of a flexible oxide fiber, and an outer metal casing. The flexible insulation in the radial direction and the spring-supported swirler in the axial direction allow for an almost unhin- dered thermal expansion of the ceramic flame tube. The insulation is intended to keep the fiber-reinforced material at an approximately uniform temperature and lower thermal stress level. Temperature is measured by thermocouples placed along the outer ceramic wall. Flanges on the outside of the combustion chamber provide access to the flame tube for mea- suring the radial temperature distribution with Pt-Rh-Pt thermocouples. The combustion chamber is assembled in a Klockner-Humboldt T216 gas turbine with a power output of 74 kW at 50,000 rpm. The nominal pressure ratio is 2.8, air-mass flow is 0.9 kg/s, and turbine inlet temperature is 810°C. The combustor dome is made of sintered silicon carbide. Since large holes are required for dilution at the end of the flame tube, the dome is separated from the flame tube by a nickel alloy spacer. The flame tube is 210 mm long and has an inner diameter of 144 mm. The thickness of the ceramic wall is 3 mm. Pressure, temperature, speed, and power are recorded during engine operation. Wall temperatures determine the thermal loading of the composite materials. Figure 11.25 dis- plays typical axial temperature distributions in the ceramic wall at different operating conditions. Peak temperature is 1050°C, occurring between the dilution holes in the mid- dle of the flame tube at nominal speed, with all four flame tubes displaying similar values. 466 MATERIALS AND MANUFACTURE Atomizer Swirler 3° 3° 10 mm r − r + r − r + Plane 1 Plane 2 Flame tube Oxide fiber material layer Outer metal casing Thermocouple FIGURE 11.24 Ceramic flame tube construction (Filsinger et al., 1997). Accumulated test time is limited to 10 h. Starting with operation at low values the thermal load is gradually intensified to a peak of 87 percent maximum load. Between the tests, the flame tubes are inspected visually for recording morphological changes by macrophotography. Hot gas profiles are measured for all operating conditions. All tubes withstood the thermal load under the oxidizing atmosphere without severe dam- age. One flame tube showed no damage on the inner surface, but on the outside a small chip of the chemical vapor deposition SiC coating is separated. This may have resulted from a mismatch between the localized thermal growth. The thermal load causes considerable dis- coloration, and may be the result of a layer of SiO 2 arising from the oxidation process. The thin amorphous glass layer reflects different wavelengths of the incoming light, depending on the thickness, and the surface appears with a rainbow of colors (Fig. 11.26). The test program is extended for the promising SiC/SiC composite combustion cham- ber, and is in operation for 90 h without indications of any damage. The unit has gone SUPERALLOYS FOR TURBINES 467 FIGURE 11.25 Temperature distribution on outer surface of flame tube (Filsinger et al., 1997). FIGURE 11.26 Flame tube after 10 h of operation (Filsinger et al., 1997). through a number of start/stop cycles that have the potential for developing critical loads because of the high temperature gradients. A regenerative twin spool ceramic gas turbine design aims to achieve thermal efficiency of 42 percent at turbine inlet temperature of 1350°C. Developed by Kawasaki Heavy Industries (Takehara et al., 1996), the lower pollutant emission and multifuel capability gas turbines are to be used in cogeneration systems. Some unique features include simple- shaped ceramic components and stress-free structures using ceramic springs and rings. Figure 11.27 provides a cross-sectional layout of the engine. 468 MATERIALS AND MANUFACTURE Compressor impeller Combustor liner Gas generator rotor Gas generator nozzle Recuperator Power turbine blade Power turbine nozzle FIGURE 11.27 Cross section of ceramic gas turbine (Takehara et al., 1996). Coil springs GGT wave rings & piston rings B A GGT nozzle wave ring PT nozzle wave ring FIGURE 11.28 Stress free support system (Takehara et al., 1996). A single can combustor and a high-pressure ratio recuperator are conventionally designed. A ceramic gas generator and power turbine nozzles and scroll are supported in the metal engine casing by elastic ceramic parts. Piston rings, also made of ceramic, are used for inner and outer seals. Wave rings are designed to absorb thermal expansion and dynamic displacements, illustrated in Fig. 11.28. The seals and rings are made of Si 3 N 4 . Nozzle assemblies are produced by binding segments with SiC fibers, which are converted into fiber-reinforced ceramic in the form of a monolithic ring. The nozzles are capable of withstanding elevated temperatures, provide adequate stress characteristics, and can be readily installed within the engine. A single-stage impeller provides compression ratio of 8:1 and flow rate of 0.9 kg/s with an adiabatic efficiency of almost 80 percent. A channeled diffuser provides for adjustment of inlet angle to the impeller’s discharge angle. A schematic drawing of the combustor is shown in Fig. 11.29. The ceramic liner is supported by coil springs to absorb relative ther- mal growth between the liner and the metal case. The combustor has a bypass line with a valve to control the ratio of air and fuel. Endurance testing at this stage of development of the turbine records 19 cycles with 94 accumulated hours. 11.12 CERAMIC COMPONENTS IN MS9001 ENGINE The advisability of implementing ceramic components in a utility-sized turbine in com- mercial service for power generation has been assessed by General Electric Company in conjunction with Tokyo Electric Power Company. The program calls for evaluating the performance of the ceramic combustion transition piece, stage 1 bucket, nozzle and shroud, and stage 2 bucket and nozzle (Grondahl and Tuschiya, 1998). A recent production MS9001FA gas turbine in a single-shaft advanced combined cycle mode of operation is specified as the baseline for the comparison. Primary performance evaluation study is conducted at a constant NO x emission level of 25 ppm in the exhaust from the turbine. Since the emissions are directly related to the SUPERALLOYS FOR TURBINES 469 FIGURE 11.29 Combustor arrangement (Takehara et al., 1996). combustion reaction zone temperature, the turbine inlet temperature (at the exit plane of the transition piece) is held constant in the analyses. Baseline compressor airflow is also main- tained constant. But the pressure drop in the combustion system and cooling air extraction from the compressor discharge are decreased with ceramic components, causing increased airflow through the combustor. This results in reduced fuel-to-air ratio, lower flame tem- perature, and less NO x . Hence, fuel flow and firing temperature are increased as necessary to maintain the level of temperature at the turbine inlet. The materials considered in the study are limited to monolithic ceramics. Monolithic silicon-nitride, SN-88, is the primary candidate for all the components, with its physical properties shown in Table 11.3, and a fracture map of the material strength capability pro- vided in Fig. 11.30. The peak application temperature of the ceramic is limited from oxi- dation considerations to 1315°C. The ceramic transition piece is placed in an outer metal casing, and has provision for impingement cooling. The metal shell supports the pressure difference between the inside and the outside of the system, and minimizes the leakage at the ceramic liner tiles. 470 MATERIALS AND MANUFACTURE TABLE 11.3 Properties of SN-88, Sintered Silicon Nitride Temperature (°C) Property 25 400 800 1200 1400 Density, g/cm 3 3.52 ———— Young’s modulus, GPa 300 300 300 290 280 Shear modulus, GPa 20 120 120 116 112 Poisson’s ratio 0.27 0.27 0.27 0.27 0.27 Thermal expansion coeff. × 10 –6 /K — 2.7 3.3 3.5 3.5 Thermal conductivity, W/m⋅K 71.1 33.5 25.1 20.9 — Specific heat, J/kg⋅K 669 1004 1213 1297 — FIGURE 11.30 Fracture mechanism map of SN-88 (Grondahl and Tuschiya, 1998). The attachment of the ceramic segments permits exchange of heat radiation between the outer shell and the ceramic. The exit seal leakage area is similar to a conventional design. The stage 1 shroud is cooled by air from the compressor discharge, and is subject to max- imum gas-path temperature below the lower-limit oxidation temperature limit of 1204°C for SN-88. The stage 1 nozzle uses film cooling from a number of holes in the airfoil and side- walls to efficiently reduce heat transfer to the metal by reducing the film temperature at the surface (Tsuchiya et al., 1995). The vanes also require impingement cooling. The vanes have sidewall thickness between 5 and 6 mm. The stage-2 nozzle uses cooling air from the com- pressor 13th stage. The first- and second-stage buckets are convectively cooled by air extracted from the 17th stage of the compressor. The bucket design is described by Terama et al. (1994), together with a discussion of the associated development effort. The gross combined cycle efficiency and improvement in the output relative to the base- line engine are shown in Fig. 11.31, using the minimum oxidation limit values for the ceramic as shown in Fig. 11.30. The results include benefits from the increased fuel flow and the firing temperature needed to maintain turbine inlet temperature and NO x emissions constant. Maximum gain is obtained from the first-stage nozzle vane and bucket, mostly because of the reduced cooling airflow with the ceramic design and the consequent increase in the flow through the combustor. Maintaining the fuel-to-air ratio for constant NO x also results in the large increment in the firing temperature. SUPERALLOYS FOR TURBINES 471 FIGURE 11.31 Cumulative performance benefits with ceramic components (Grondahl and Tuschiya, 1998). REFERENCES Bressers, J., Timm, J., Williams, S., Bennett, A., and Affeldt, E., “Effects of cycle type and coating on the TMF lives of a single crystal nickel-based gas turbine alloy,” Thermo-Mechanical Fatigue Behavior of Materials, ASME STP 1263, American Society of Testing of Materials, Philadelphia, Pa., pp. 82–95, 1996. Cheruvu, N. S., “Development of a corrosion resistant directionally solidified material for land based turbine blades,” ASME Paper # 97-GT-425, New York, 1997. Decker, R. F., Mihalisin, J. R., Transactions of American Society of Metals, Vol. 62, p. 481, 1969. Decker, R. F., “Strengthening mechanisms in nickel base super alloys,” Climax Molybdenum Company Symposium, Munich, Germany, May 1969. Dinis-Ribeiro, N., and Sellars, C. M., “Strength and structure during hot deformation of nickel base super alloys,” Super Alloys Conference, Araxa, Brazil, April 1984. Evans, A. G., Wang, J. S., and Mum, D., “Mechanism based life prediction issues for thermal barrier coating,” paper presented at TBC Workshop, Cincinnati, Ohio, May 1997. Fell, E. A., Mitchell, W. I., and Wakeman, D. W., “Iron & Steel Institute Special Report,” Vol. 70, p. 136, 1969. Filsinger, D., Munz, S., Schulz, A., Wittig, S., and Andrees, G., “Experimental assessment of fiber reinforced ceramics for combustor walls,” ASME Paper # 97-GT-154, New York, 1997. Fleischer, R. L., The Strengthening of Metals, p. 93, Reinhold, New York, 1964. Gegel, H. L., Prasad, Y. V. R. K., Malas, J. C., Morgan, J. T., Lark, K. A., Doraivelu, S. M., and Barker, D. R., “Computer simulations for controlling microstructure during hot working of Ti 6-2-4-2,” PVP Vol. 87, ASME Pressure Vessels and Piping Conference and Exhibition, New York, p. 101, 1984. Gell, M., and Duhl, D. N., “Processing and properties of advanced high temperature alloys,” Metals, ASM, Park, Ohio, p. 41, 1986. Grondahl, C. M., and Tuschiya, T., “Performance benefit assessment of ceramic components in a MS9001FA gas turbine,” ASME Paper # 98-GT-186, New York, 1998. Guimaraes, A. A., and Jonas, J. J., Metallurgy Transactions, Vol. 12A, 1655, 1981. Ham, R. K., “Ordered alloys: Structural applications and physical metallurgy,” Claitors, Baton Rouge, La., p. 365, 1970. Helmer, T., Petrelik, H., and Kromp, K., “Coating of carbon fibers and the strength of fibers,” Journal of American Ceramics Society 78:133–136, 1995. Hughes, S. E., and Anderson, R. E., Technical Report # AFML-TR-79-4146, U.S. Contract # F33615- 76-C-5136, 1978. Immarigeon, J. P. A., “The role of microstructure in the modeling of plastic flow in P/M super alloys at forging temperatures and strain rate,” Advisory Group for Aerospace Research and Development, Neuilly-sur-Seine, France, AGARD-LS-137, 4–1, 1984. Johnson, P. K., Arana, M., Ostolaza, K. M., and Bressers, J., “Crack initiation in a coated and uncoated nickel-base super alloy under TMF conditions,” ASME Paper # 97-GT-236, New York, 1997. Kempster, A., and Czech, N., “Protection against oxidation of internal coating passages in turbine blades and vanes,” presented at Power Gen Conference, 1998. Klarstrom, D. L., Super Alloys 1980, ASM, Metals Park, Ohio, p. 131, 1980. Kool, G. A., Agema, K. S., and Van Buijtenen, J. P., “Operational experience with internal coatings in aero and industrial gas turbine airfoils,” Proceedings of the ASME Turbo Expo, The Netherlands, Paper # GT-2002-30591, New York, 2002. Koul, A. K., Immarigeon, J. P., Dainty, R. V., and Patnaik, P. C., “Degradation of high performance aero engine turbine blades,” Proceedings of the ASM Materials Congress, Pittsburgh, Pa., pp. 69–74, 1993. Lackey, W. J., Report # ORNL/TM-8959, 1984. Leverant, G. R., and Kear, B. H., Metallurgy Transactions 1: 491, 1970. Luthra, K. L., and Wood, J. H., Metallurgical Coatings Conference Proceedings, Vol. II, Elsevier, San Diego, p. 271, 1984. McLean, M., “Directionally solidified materials for high temperature service,” The Metals Society, 153, 1983. 472 MATERIALS AND MANUFACTURE [...]... during the brazing (Fig 12 .13) The heat-treated test specimens are TABLE 12.2 Composition of Powder Metallurgy Material for Brazing Element Mar M-247 alloy Mar M247–7 Mar M247–3 Mar M247 wide gap IN 738 PM Ni Cr Co Mo W Ta Hf Al Ti Nb Y B 59.9 08.5 10.0 00.6 10.0 03.0 01.5 05.5 01.0 00.0 00.0 00.0 65.26 10.11 07.52 00.45 07.52 02.26 01 .13 04.14 00.75 00.00 00.00 00.87 58 .13 11.95 13. 00 00.42 07.00 03.00... wrought parts often arises from segregation, resulting in inconsistent and reduced thermomechanical response The powder-based process is then employed when cast or wrought components are not suitable Some intrinsic attributes of powder materials make them appropriate for turbine components A high rate of solidification results in smaller intermetallic particles and reduces the spacing between the particles,... especially in highly stressed parts that may be critical from fracture and fatigue considerations Thermomechanical processing is essential for neutralizing the ensuing defects The powders are required to have a very fine size The magnitude of the stress and metallurgical alterations during machining depend on machining parameters such as feed, speed, depth of cut, cutting tool material, part geometry, 475 Copyright... of castings is subjected to radiographic examination, particularly for aircraft engines 12.3 INVESTMENT CASTINGS The investment casting process has retained most of its features over the centuries in the making of items of jewelry, with innovations introduced in applications for gas turbine components made of superalloys A precise replica of the part is first produced in wax or a plastic polymer Compensation... to reduce loss of heat and to obtain a controlled rate of solidification After the casting is cooled, the shell and the core are separated from the metal parts mechanically and chemically, and also split from the cluster at the connecting ducts The parts then go through inspections, heat treatment, or densification by the hot isostatic press (HIP) method The pattern for the preparation of the mold must... in the slurry than immersed in a fluidized bed of the particles The binder then cures by chemical reaction for further applications The rate at which drying occurs must be controlled to avoid distortion A typical shell mold may be coated 5 to 10 times to develop thickness and strength The ceramic slurry is made of a finer grain then the subsequent particles Beside grain configuration, microstructural... in a vacuum to obtain proper bonding between the particles (Reichman and Smythe, 1970) The powders usually are spherical, and filtered through a fine mesh to reduce contamination Methods such as atomization by inert and soluble gas, the rotating electrode process, and centrifugal atomization are found to be practical for commercial production of the parts In the gas atomization method, the vacuum-refined... skimmed by the centrifugal action of the rotating electrode, and is split into atomized particles in the chamber A plasma arc may also be used instead of the electrode The common feature of all the methods is to provide for atomization of the melted alloy into a powder form free of contaminants The size of the average particle obtained from the rotating method tends to be larger than from the atomizing... differences in the physical aspects of the procedures The presence of oxidized particles developed during the atomization has repercussions on physical properties such as ductility and low-cycle fatigue life Leakage of air into the system is primarily responsible, and even low levels of incidence tend to reduce the minimum properties Hollow particles may also develop mostly from trapped gas or from shrinkage... grains may be several inches in length, and impart substantial strength to the material by deleting the transverse grain boundaries (Cockell and Boyce, 1985) Thermomechanical processing aims to neutralize the intrinsic and extrinsic defects arising from contamination Intrinsic defects occur during the powder making process from the argon pores and oxidized particles Extrinsic effects may occur from handling . Journal of American Ceramics Society 78 :133 136 , 1995. Hughes, S. E., and Anderson, R. E., Technical Report # AFML-TR-79-4146, U.S. Contract # F33615- 76-C- 5136 , 1978. Immarigeon, J. P. A., “The. modulus, thermal conductivity, and oxidation resistance, are partly offset by their relatively low fracture toughness. Additives and particulate reinforcements have been tried as toughening agents Ohio, p. 131 , 1980. Kool, G. A., Agema, K. S., and Van Buijtenen, J. P., “Operational experience with internal coatings in aero and industrial gas turbine airfoils,” Proceedings of the ASME Turbo