0

Manufacturing Materials and Metallurgy

1998;():V005T12A001. doi:10.1115/98-GT-370.
FREE TO VIEW

Stress relaxation and constant displacement rate tensile tests were performed on poly-crystalline GTD111 alloy material removed from General Electric MS6001B first stage combustion turbine buckets. Samples were examined in the standard heat treated condition, thermally exposed at 900°C for 5000 hours and from service run buckets. Creep rates of the material were measured and evaluated directly in terms of temperature capability at 850°C and 900°C. Stress relaxation tests done at 0.8% total strain indicated that the creep rate properties in the service exposed airfoil were an order of magnitude higher than the material properties in the standard heat treated condition measured in the root form. In terms of temperature capability, the creep rate properties of the service run airfoil material had decreased by the equivalent of almost 40°C.

The stress relaxation test method was demonstrated to be a very useful tool in quantifying the degradation of creep properties in service run components. Creep data that would require years to gather using conventional creep tests was generated in a few days. This now makes realistic life assessment and repair / replace decisions possible during turbine overhauls.

The test method’s unique ability to measure changes in creep rate over a large stress range, enabled the technique to distinguish between changes in creep strength due to (normal) microstructural evolution from the combined effects of microstructural evolution and strain related creep damage.

A method for estimating standard constant load creep rupture life from the stress relaxation creep rate data is also presented along with time-temperature parameter correlations. The data sets examined in this study indicate that creep rupture lives can be estimated within a factor of three from the stress relaxation data.

The information and analysis techniques described in this paper are directly applicable to metallurgical life assessment evaluations and the re-qualification of repaired General Electric buckets in Frame 3, 5, 6, 7 and 9 engine models.

Commentary by Dr. Valentin Fuster
1998;():V005T12A002. doi:10.1115/98-GT-371.
FREE TO VIEW

Turbine inlet temperatures have now approached 1650°C (3000°F) at maximum power for the latest large commercial turbofan engines, resulting in high fuel efficiency and thrust levels approaching or exceeding 445 kN (100,000 lbs.). High reliability and durability must be intrinsically designed into these turbine engines to meet operating economic targets and ETOPS certification requirements.

This level of performance has been brought about by a combination of advances in air cooling for turbine blades and vanes, computerized design technology for stresses and airflow and the development and application of rhenium (Re) containing, high γ′ volume fraction nickel-base single crystal superalloys, with advanced coatings, including prime-reliant ceramic thermal barrier coatings (TBCs). Re additions to cast airfoil superalloys not only improve creep and thermo-mechanical fatigue strength but also environmental properties, including coating performance. Re slows down diffusion in these alloys at high operating temperatures.(1)

At high gas temperatures, several issues are critical to turbine engine performance retention, blade life and integrity. These are tip oxidation in particular for shroudless blades, internal oxidation for lightly cooled turbine blades and TBC adherence to both the airfoil and tip seal liner. It is now known that sulfur (S) at levels < 10 ppm but > 0.2 ppm in these alloys reduces the adherence of α alumina protective scales on these materials or their coatings by weakening the Van der Waal’s bond between the scale and the alloy substrate. A team approach has been used to develop an improvement to CMSX-4® alloy which contains 3% Re, by reducing S and phosphorus (P) levels in the alloy to < 2 ppm, combined with residual additions of lanthanum (La) + yttrium (Y) in the range 10–30 ppm. Results from cyclic, burner rig dynamic oxidation testing at 1093°C (2000°F) show thirteen times the number of cycles to initial alumina scale spallation for CMSX-4 [La + Y] compared to standard CMSX-4.

A key factor for application acceptance is of course manufacturing cost. The development of improved low reactivity prime coats for the blade shell molds along with a viable, tight dimensional control yttrium oxide core body are discussed. The target is to attain grain yields of single crystal CMSX-4 (ULS) [La + Y] turbine blades and casting cleanliness approaching standard CMSX-4. The low residual levels of La + Y along with a sophisticated homogenisation/solutioning heat treatment procedure result in full solutioning with essentially no residual γ/γ′ eutectic phase, Ni (La, Y) low melting point eutectics and associated incipient melting pores. Thus, full CMSX-4 mechanical properties are attained. The La assists with ppm chemistry control of the Y throughout the single crystal turbine blade castings through the formation of a continuous lanthanum oxide film between the molten and solidifying alloy and the ceramic core and prime coat of the shell mold. Y and La tie up the < 2 ppm but > 0.2 ppm residual S in the alloy as very stable Y and La sulfides and oxysulfides, thus preventing diffusion of the S atoms to the alumina scale layer under high temperature, cyclic oxidising conditions. La also forms a stable phosphide.

CMSX-4 (ULS) [La + Y] HP shroudless turbine blades will commence engine testing in May 1998.

Commentary by Dr. Valentin Fuster
1998;():V005T12A003. doi:10.1115/98-GT-375.
FREE TO VIEW

The durability of protective coatings on combustion turbine blades and vanes is a critical issue in the power generation industry. Coating life usually dictates the refurbishment intervals for these components, and these intervals have generally been of shorter duration than desired by the operators of the equipment. Both MCrAlY and aluminide type coatings protect against oxidation and hot corrosion by forming a protective Al2O3 surface layer. Degradation of the coatings occurs by depletion of the aluminum content of the coating through interdiffusion with the substrate and through the formation and spallation of an external Al2O3 scale.

The results obtained in this study clearly show that the application of a thin interlayer of Ni-Re beneath the MCrAlY coating can significantly decrease the growth rate of the inner β-NiAl depletion zone. Order of magnitude reductions in the inner depletion zone thickness formed at 1000 hours were obtained with both the Ni-32 wt.% Re and the Ni-47 wt% Re interlayer coatings. Since formation of the inner depletion zone is believed to result from interdiffusion with the substrate, these results suggest that the Ni-Re interlayer provided a significant impediment to the inward diffusion of Al into the substrate.

Commentary by Dr. Valentin Fuster
1998;():V005T12A004. doi:10.1115/98-GT-391.
FREE TO VIEW

Chemical vapor deposition (CVD) is an old coating technology, but it was not successfully utilized to aluminize gas turbine hardware until recently (1989). In CVD aluminizing, the use of multiple, independently controlled, low temperature, external, metal halide generators combined with computer control of all process variables gives flexibility and consistent quality that is not possible with any other aluminizing process.

It has been shown that harmful coating impurities (such as sulfur and boron etc.) can be transported to a coating from a high temperature aluminum source in the coating chamber during aluminizing. Representative processes include: pack cementation, above the pack, SNECMA, and high activity CVD. In contrast, it has also been demonstrated that CVD low activity aluminizing removes harmful impurities (S, P, B & W etc.) from the coating during deposition. Furthermore, clean, low activity coatings (simple aluminide MDC-210 or platinum modified MDC-150L) have been shown to exhibit superior oxidation resistance compared to similar coatings made by other aluminizing processes. A second significant source of impurities in platinum modified aluminide diffusion coatings is electroplating, that is, plating bath components (S, P, CI, K, Ca etc.) are codeposited with the platinum, and these impurities can have either a beneficial (K&Ca) or a detrimental (S,P&Cl) influence upon the oxidation resistance of the product coating.

The results of investigations on the transport of impurities during aluminizing and electroplating, plus the influence of these impurities on oxidation resistance of the product coatings will be presented and discussed.

Commentary by Dr. Valentin Fuster
1998;():V005T12A005. doi:10.1115/98-GT-401.
FREE TO VIEW

Nickel-based superalloys have been used in the gas turbine hot section components for their outstanding mechanical properties at elevated temperatures. Increasing the alloy strength at high temperature is usually achieved at the expense of the alloy’s environmental stability. Oxidation and high heat flux could be limiting factors in the use of these alloys at temperatures above 1800°F. To help overcome these limitations, protective coatings can be applied to the alloy surfaces to provide oxidation and hot corrosion resistance. These coatings are applied to alloys which can be produced in various forms such as equiaxed, directionally solidified or single crystals with varying chemistries. Elemental additions such as hafnium, rhenium, etc. are added to promote the strengthening of these alloys, and could result in varying effects on the coatability and coating performance. This paper discusses the effects of various substrate elements on the processing and stability of diffusion platinum aluminide coatings.

Commentary by Dr. Valentin Fuster
1998;():V005T12A006. doi:10.1115/98-GT-403.
FREE TO VIEW

The tensile stress relaxation behavior of two NiCoCrAlY bond coat alloys was examined at several temperatures between 25 and 899°C (1650°F) and at 0.1, 0.3, 0.5, and 0.8% strain. One alloy was made from Praxair’s CO211 powder and served as the reference alloy, while the other was a Westinghouse-developed, oxide-dispersion-strengthened alloy. The specimens were loaded to the desired tensile strain at a constant strain rate, and the elastic modulus, yield strength, and yield strain were determined as a function of temperature for the two alloys using the stress/strain information from this loading segment. There was not a statistically significant difference in the high temperature elastic properties between the two alloys, although the oxide-dispersion-strengthened alloy tended to exhibit larger yield strengths. The relaxation data for both alloys were reduced into a form in which instantaneous stressing rate during relaxation was examined as a function of stress and temperature using an Arrhenius power-law model. The oxide-dispersion-strengthened alloy exhibited a larger stress exponent and activation energy than the reference alloy between 677–899°C (1250–1650°F), and was generally more creep resistant. The results from this study demonstrate that bond coat relaxation should occur during engine operation. Bond coatings fabricated from the oxide-dispersion-strengthened alloy have the potential to reduce residual stresses in the TBC ceramic top coating.

Commentary by Dr. Valentin Fuster
1998;():V005T12A007. doi:10.1115/98-GT-468.
FREE TO VIEW

Cyclic oxidation behavior of aluminide, platinum modified aluminide, and MCrAlY coatings has been investigated at three temperatures. Aluminide and platinum modified coatings were deposited on GTD 111 material using an outward diffusion process. CoCrAlY coating was applied on GTD-111 by Electron Beam Physical Vapor Deposition (EB-PVD). The oxidation behavior of these coatings is characterized by weight change measurements and by the variation of β phase present in the coating. The platinum modified aluminide coating exhibited the highest resistance to oxide scale spallation (weight loss) during cyclic oxidation testing. Metallographic techniques were used to determine the amount of β phase and the aluminum content in a coating as a function of cycles. Cyclic oxidation life of these coatings is discussed in terms of the residual β and aluminum content present in the coating after exposure. These results have been used to calibrate and validate a coating life model (COATLIFE) developed at the Material Center for Combustion Turbines (MCCT).

Commentary by Dr. Valentin Fuster
1998;():V005T12A008. doi:10.1115/98-GT-469.
FREE TO VIEW

Inconel 718 alloy is used extensively in aerogas turbines and this alloy is most difficult to machine and highly prone to dimensional instability after machining. Such detrimental phenomenon poses enormous problem in engine assembly and affect structural integrity. This paper highlights the systematic research work undertaken to study the plastic deformation characteristics of Inconel 718, effect of process variables on machined surface, subsurface and dimensional instability. Also illustrated the technique developed for simultaneous optimization of several process variables such as cutting speed, feed, depth of cut, rake angle and tool nose radius, to control the residual stresses and dimensional instability, within the acceptable tolerance band of the component. Prediction equations were developed for residual stress, dimensional instability, tool life, surface finish and material removal rate. Predicted data were validated experimentally. This paper also presents the qualitative and quantitative data on dimensional instability with specific case studies of jet engine components and clearly illustrates the approach followed to develop technique to control such detrimental effect.

Commentary by Dr. Valentin Fuster
1998;():V005T12A009. doi:10.1115/98-GT-478.
FREE TO VIEW

A life prediction method for combustion turbine blade coatings has been developed by modeling coating degradation mechanisms including oxidation, spallation, and aluminum loss due to inward diffusion. Using this model, the influence of cycle time on coating life is predicted for GTD-111 coated with an MCrAlY, PtAl, or aluminide coating. The results are used to construct a coating life diagram that depicts failure and safe regions for the coating in a log-log plot of number of startup cycles versus cycle time. The regime where failure by oxidation, spallation, and inward diffusion dominates is identified and delineated from that dominated by oxidation and inward diffusion only. A procedure for predicting the remaining life of a coating is developed. The utility of the coating life diagram for predicting the failure and useful life of MCrAlY, aluminide, or PtAl coatings on the GTD-111 substrate is illustrated and compared against experimental data.

Commentary by Dr. Valentin Fuster
1998;():V005T12A010. doi:10.1115/98-GT-488.
FREE TO VIEW

The use of single crystal (SX) nickel-base superalloys will increase in the future with the introduction of SX blades into large gas turbines for base-load electricity production. Prolonged periods of use at high temperatures may cause creep deformation and the assessment of damage can give large financial savings. A number of techniques can be applied for life assessment, e.g. calculations based on operational data, non-destructive testing or material interrogation, but because of the uncertainties involved the techniques are often used in combination. This paper describes a material interrogation (metallographic) technique for creep strain assessment in SX alloys.

Creep tests have been performed at 950°C on the SX alloy CMSX-4 and quantitative microstructural studies performed on specimens interrupted at various levels of strain. It was found that the strengthening γ′-particles, initially cuboidal in shape, coalesced to form large plates or rafts normal to the applied stress. The γ-matrix phase also formed plates. CMSX-4 contains ∼ 70 vol % γ′-particles and after creep deformation the microstructure turned itself inside out, i.e. the gamma “matrix” became the isolated phase surrounded by the γ′-“particles”. This can cause problems for computerised image analysis, which in this case, were overcome with the choice of a suitable measurement parameter.

The rafts reached their maximum length before 2% strain, but continued to thicken with increasing strain. Although of different dimensions, the aspect ratios (length/thickness ratio) of the gamma-prime rafts and the gamma plates were similar at similar levels of strain, increasing from ∼1 at zero strain to a maximum of ∼3 at about 1–2 % strain.

Analysis of microstructural measurements from rafting studies on SX alloys presented in the literature showed that the aspect ratios of the γ- and γ′-phases were similar and that at a temperature of 950–1000°C a maximum length/thickness ratio of about 2.5–3.5 is reached at 1 to 2% creep strain. Measurement of gamma-prime raft or (or gamma plate) dimensions on longitudinal sections of blades is thus a suitable method for high temperature creep damage assessment of SX alloys. This gives a considerable advantage over conventional Ni-base superalloys whose microstructures are usually very stable with respect to increasing creep strain.

Commentary by Dr. Valentin Fuster
1998;():V005T12A011. doi:10.1115/98-GT-508.
FREE TO VIEW

Recently developed INCONEL® alloy 783 (nominal composition of Ni-34Co-26Fe-5.4Al-3Nb-3Cr) is precipitation strengthened by Ni3Al-type Gamma Prime and NiAl-type Beta Phases. Due to its low co-efficient of thermal expansion (CTE), high strength, and good oxidation resistance alloy 783 has been specified for use in aircraft gas turbine components such as rings, casings, shrouds, and seals and has been considered for use in a number of other critical industrial turbine components.

In this study, commercially produced alloys 783, 718, and 909 were annealed and aged following recommended heat treatments. The materials were then isothermally exposed at 1100°F (593°C) for times up to 10,000 hours. At 1000 hour intervals, specimens of these alloys were removed from the furnace and subjected to room temperature tensile (RTT) and high temperature tensile (HTT) testing at 1200°F (649°C). The microstructure of as-produced and exposed materials was characterized using optical microscopy, Scanning Electron Microscopy (SEM), and Transmission Electron Microscopy (TEM). Variation in tensile properties with isothermal exposure time was correlated with the microstructure.

Commentary by Dr. Valentin Fuster
1998;():V005T12A012. doi:10.1115/98-GT-509.
FREE TO VIEW

Looking at HCF, the most critical component on the turbine is the blade.

At the present moment there is a need for a HCF criterion suitable for evaluation of stresses calculated with the Finite Element Method. Today there are some, such criterions e.g. Sines. However the criterion does not include important aspects such as the influence of; ‘geometrical size’, ‘stressed volume’ or ‘stress gradient’.

A better understanding of the HCF phenomena would give an increase in the design precision and an opportunity to increase the load on the blades.

A proposal is made for a new theory for HCF assessment. The input required by the theory is roughly the Haigh-diagram for a smooth test specimen.

The theory is a merger between the Sines criterion and the statistical approach made by Weibull. 3 basic assumptions are made; equivalent stresses are calculated according to Sines, ‘the weakest link assumption’ and finally that the fatigue limit for a smooth specimen has a statistical distribution.

Comparing the proposal to classical HCF assessment with Haigh-diagram, the following factors are incorporated:

1, stress concentration factor

2, fatigue notch factor (or notch sensitivity factor)

3, geometric volume dependence

4, different Haigh-diagrams for push-pull, bending and torsion

If the factors are incorporated partially or fully remains to be shown. It is also possible to give a geometrical quantification of the so called ‘stressed volume’.

Qualitative and quantitative tests have been made. Looking at them, the theory looks promising. However, it has to be tested further to get acquainted with any shortcomings of the theory.

Commentary by Dr. Valentin Fuster
1998;():V005T12A013. doi:10.1115/98-GT-510.
FREE TO VIEW

Conventionally cast GTD-111EA first stage buckets from GE Frame 5001P-NT engine after 42,000 hours of service, directionally solidified GTD-111DS Frame 5002C buckets after 49,000 hours of service, and conventionally cast IN-738 Frame 5002B buckets after 81,000 hours of service were evaluated before and after refurbishment. These buckets were coated with GT-29, GT-29 plus and RT-22 coatings respectively. Coating condition, microstructural degradation, tensile properties, and creep properties were evaluated in the service aged condition. Microstructure was assessed after the hot isostatic pressing (HIP) and solution heat treatment and again after the full refurbishment and recoating of the buckets. It was found that the GT-29 coating was breached 100% by oxidation in the 5001P engine after service. The GT-29 plus coating on the 5002C engine was breached very slightly and the majority of the coating was still in good condition. The RT-22 coating on the IN-738 bucket was stripped before it was received for evaluation at SwRI. Gamma prime growth, coalescence and agglomeration were found in all of the buckets. Continuous carbide network near the airfoil leading edge was found in the GTD-111EA bucket.

Significant improvement in the microstructural condition was observed after the refurbishment for the GTD-111 buckets. However, the gamma prime morphology was not normal for the IN-738 bucket. Both tensile and creep properties of the GTD-111 buckets showed significant improvement after refurbishment. However, for the IN-738 bucket little improvement on the mechanical properties was observed. The reasons for this and the correlation between the microstructure and properties are discussed in this paper.

Commentary by Dr. Valentin Fuster
1998;():V005T12A014. doi:10.1115/98-GT-511.
FREE TO VIEW

The surface metal temperature on the first row cooled buckets varies along the height of a bucket and among the buckets in a row. Two first-row GT 29+ coated buckets, from the same turbine, have been investigated for in-service material and coating degradation variation. The results showed that the γ′ particles were coarser and the width of the interdiffusion zone below the CoCrAlY coating was wider in one bucket than the other, indicating that the former bucket had seen higher metal temperature than the latter. Fewer uncertainties are associated with interdiffusion zone width measurements, and the change in zone width can be used to develop a time-temperature correlation for the subject bucket material and coating system. The uncertainties associated with γ′ particle measurements for time temperature correlation are discussed.

Commentary by Dr. Valentin Fuster
1998;():V005T12A015. doi:10.1115/98-GT-526.
FREE TO VIEW

The effect of thermal ageing at 870 °C for 8000 h in air on the microstructure/composition and mechanical properties (RT and 870 °C) has been studied in aluminized CoCrAlY coatings consisting of four layered structure (region I-IV) of advanced gas turbine blades. Thermal ageing led to a little oxidation/nitridation and a decrease in the Al content in a near surface region I. In a coating region II, coarse Cr rich σ precipitates formed during the thermal ageing. Thermally aged internal (III) and near interface (IV) coating regions showed extensive dispersion of σ and/or Al/Ni rich β/α eutectic precipitates. Small punch tests at RT and 870 °C in air have shown that the coating regions I and II of imaged and aged blades indicated easier formation of brittle cracks regardless of the composition change. The ductility of the regions III and IV at RT and 870 °C, and the low cycle fatigue life of the region III were reduced by the thermal ageing. The mechanical degradation at elevated temperatures in the aged coating regions III and IV is elucidated by taking into account the microstructure/composition evolution and environmental oxidizing effects.

Topics: Coatings , Ductility
Commentary by Dr. Valentin Fuster
1998;():V005T12A016. doi:10.1115/98-GT-547.
FREE TO VIEW

The 20.65 MW gas turbine experienced catastrophic damage. The failure occurred at the first stage buckets and resulted in damage of the all buckets of this stage. Five rotor disc grooves were also seriously damaged. Additionally, all second stage buckets, first and second stage nozzles, shroud segments, the No 2 bearing casing (turbine side), compressor moving blades, and other elements were damaged. Due to urgent power generation needs, it was decided to repair a seriously damaged stage 1 rotor disc in-situ, and replace all the other damaged parts.

The development of a propietary welding technology for the in-situ repair of the five damaged disc grooves without disc disassembly, and of in-situ disc grooves’ mechanized machining is fully described. The repair process included the removal of damaged grooves, method of groove restoration by welding deposition, stress relief and groove machining to recover their original geometry. After rotor disc repair and assembly, the rotor was put back into service. The approach to the repair of the rotor disc damage has been successful. It enabled significant reductions in expenditure on replacement parts and a reduction of outage time to be achived.

Commentary by Dr. Valentin Fuster
1998;():V005T12A017. doi:10.1115/98-GT-549.
FREE TO VIEW

A simple, yet useful procedure is developed to generate tool paths with global interference checking for five-axis machining of turbomachinery components with complex geometries. Based on the projected distance between the surface data and the cutter-axis of a cylindrical ball-end mill, interference between the surface of a workpiece and the cutter can be detected. Given the cutter contact points of the surface and the cutter’s size, it can produce the cutter location data without incurring interference through relatively rotating and tilting the workpiece. Applications of the developed approach to five-axis machining of centrifugal compressor impellers with thirteen and fifteen blades are illustrated to demonstrate the usefulness and reliability of the procedure.

Commentary by Dr. Valentin Fuster
1998;():V005T12A018. doi:10.1115/98-GT-550.
FREE TO VIEW

Nickel-based superalloys are some of the most difficult materials to weld. Exceptional aging characteristics result in susceptibility to strain-age cracking in the heat affected zone and in the weld itself. To overcome the latter situation, filler alloys are commonly used which are much less strain sensitive than the base alloy. Mechanical strength is sacrificed for ductility with these materials thereby limiting the weldable areas of gas turbine components to regions of relatively low operating stresses. Sophisticated automated techniques have been developed to join superalloys using fillers having similar or identical chemistries to the base alloys. These welding methods are expensive compared to manual processes and are limited when the application involves localized repairs in varied locations. This study demonstrates the successful application of a specially developed process for the manual GTAW joining of precipitation-hardening alloys with filler materials of similar composition and properties.

Commentary by Dr. Valentin Fuster
1998;():V005T12A019. doi:10.1115/98-GT-551.
FREE TO VIEW

The compressor and turbine blades of aircraft engines and gas turbines have to be overhauled at regular intervals. First, the worn out blades are welded manually or by a welding robot. After welding, the blades will normally be shaped by hand, using grinding bands. Automation of the shaping of the welded beads reduces costs and process time while giving higher accuracy and quality. The following repair technique has been found to allow automatic tooling of the weld on final domains of blades, via a chain of processes, which run on one, single standard, three axis machining center. 1) Digitization of blade cross sections below the welded bead; 2) Calculation of the nominal geometry of the final domains; 3) Calculation of the NC processing programs; 4) NC processing of the final domains of the blades. The complete repair for one blade (steps 1 to 4) takes two to three minutes. The 3D digitizing system captures the actual geometry and chucking position of the blades completely. Therefore, no CAD data or measuring data of a master piece are required and the free chucking position of the blades allows simple fixtures. The blades can be repaired without user interaction, individually or in a batch quantity.

Commentary by Dr. Valentin Fuster
1998;():V005T12A020. doi:10.1115/98-GT-559.
FREE TO VIEW

The economic concerns of developing a new repair procedure for a gas turbine for commercial production can be more challenging to overcome than the technical concerns. New repairs face a wide variety of economic and technical issues that branch into all phases of an operation. Coordination of a repair requires cooperation between engineering, production, finance, customer service, marketing, regulatory agencies, materials/purchasing, and senior management. A complete financial analysis of both internal and external expectations is required for a fiscally responsible project. This paper will examine the economic consideration underlying repair development and will include a discussion of the interaction and interests between the various functional groups.

Commentary by Dr. Valentin Fuster
1998;():V005T12A021. doi:10.1115/98-GT-565.
FREE TO VIEW

The use of oxide dispersion strengthened (ODS) superalloys for excellent high temperature oxidation resistance has been well established. This is achieved by the formation of a dense, slowly growing and tightly adherent alumina scale on Fe-Cr-Al and Ni-Cr-Al ODS alloys. The addition of oxide dispersion strengthening confers a structural capability for operation temperatures of up to 1350°C. Traditionally these materials have been used in relatively thick sections where the core “reservoir” of aluminium is adequate to provide a continuous replenishment of the protective oxide coating.

This paper examines the use of one of these materials, PM2000, in thin section structures. In particular it addresses the problem of using thin foils in the manufacture of a typical engineering structure, seal honeycomb. Thin foils in the thickness range 90 to 150μm were tested to explore the oxidation limit under extreme temperature conditions. The results indicate that material life is determined by foil thickness and aluminium content of the alloy. Comparison is drawn with existing typical materials for these applications e.g. Haynes 214.

Commentary by Dr. Valentin Fuster
1998;():V005T12A022. doi:10.1115/98-GT-587.
FREE TO VIEW

An improvement in the efficiency of land based gas turbines is sought by replacing air with steam for the internal cooling of hot gas path components. However, the degradation behavior and mechanisms associated with the interaction of superalloys with steam at high temperatures are unknown. This information is very crucial for the reliability as well as for useful life prediction of the gas turbine components. In this work, oxidation in steam has been studied for three nickel based superalloys commonly used in the gas turbines. These are IN 738, Inconel 617, and CMSX-4. The study was carried out at four different test temperatures in the range of 800° C to 950° C for times up to 1400 hours. The results indicated significantly different oxidation behavior in steam compared to air for all the alloys tested.

Commentary by Dr. Valentin Fuster
1998;():V005T12A023. doi:10.1115/98-GT-589.
FREE TO VIEW

There is an increasing interest in development of manufacturing methods for Dual Property BLISKs, consisting of creep resistant airfoils and fatigue resistant disks bonded together by a durable joint. Optimum heat treatments are, however, very different for creep resistant single crystal CMSX-4 and fatigue resistant poly-crystalline Udimet 720 selected in this study, but fortunately the first ageing treatment for CMSX-4 (1140°C, 2–6h, AC) is similar to the partial solution treatment of U 720 HS2 (1115°C, 4h, OQ).

Based on this, diffusion bonding was performed by HIP at 1120°C and 200 MPa argon pressure for 4h, followed by cooling to 400°C. Subsequently, a shortened Udimet 720 HS2 two-step ageing treatment was adopted by heating to 650°C for 6h followed by cooling to 400°C, heating to 760°C for 2h, and finally cooling to R.T. under remaining HIP pressure.

Plasma etching followed by thin (80 nm) PVD coating with either nickel or titanium were used to clean and protect the polished surfaces before joining. The selection of coatings was governed by the possibility to reduce oxidized nickel by flushing with hydrogen at 330°C during evacuation of the HIP capsules, and by the large solubility of oxygen in titanium.

Hot tensile testing was performed at 750°C on both joined and reference materials subjected to the modified heat treatment. Initially solution treated Udimet 720 and CMSX-4 comprised the reference materials. The testing showed that joints with Ni-PVD coatings were almost as strong as Udimet 720 (although with very limited elongation), while the joints with Ti-PVD coatings were weaker.

Commentary by Dr. Valentin Fuster

Ceramics

1998;():V005T13A001. doi:10.1115/98-GT-030.
FREE TO VIEW

Gas turbine combustor concepts designed to give improved control of NOx emissions require the usage of hot uncooled walls. The main material properties needed in this application include mechanical and chemical stability at temperatures in excess of 1400°C for long times (>10 000 hours). Composites made from single crystal oxide fibre reinforced oxide with a compatible high temperature stable weak oxide interphase are potential candidate materials to meet these requirements. Alumina was chosen as a model material, unidirectional and 2D composites were processed and a suitable weak zirconia interphase was designed. The process was scaled-up to make production of larger panels and components possible. Mechanical testing was carried out at room temperature to characterise the performance of the material in the as produced and thermally aged condition. Room temperature mechanical properties compared well with other current ceramic composites and excellent high temperature stability was demonstrated. The applicability of the composite as a material for uncooled combustor walls is to be further assessed by evaluation in a combustor test rig. Results from computational fluid dynamics and finite element calculations as well as results from combustor rig tests of monolithic and composite ceramic tiles will be presented.

Commentary by Dr. Valentin Fuster
1998;():V005T13A002. doi:10.1115/98-GT-104.
FREE TO VIEW

The Research Institute of Advanced Materials Gas-Generator (AMG), which is a joint effort by the Japan Key Technology Center and 14 firms in Japan, has, since fiscal year 1992, been conducting technological studies on an innovative gas generator that will use 20% less fuel, weigh 50% less, and emit 70% less NOx than the conventional gas generator through the use of advanced materials.

Within this project, there is an R&D program for applying ceramic matrix composite (CMC) liners to the combustor, which is a major component of the gas generator.

In the course of R&D, continuous SiC fiber-reinforced SiC composite (SiCF/SiC) was selected as the most suitable CMC for the combustor liner because of its thermal stability and formability.

An evaluation of the applicability of the SiCF/SiC composite to the combustor liner on the basis of an evaluation of its mechanical properties and stress analysis of a SiCF/SiC combustor liner was carried out, and trial SiCF/SiC combustor liners, the largest of which was 500mm in diameter, were fabricated by the filament winding and PIP (polymer impregnation and pyrolysis) method.

Using a SiCF/SiC liner built to the actual dimensions, a non-cooling combustion test was carried out and even when the gas temperature was raised to 1873K at outlet of the liner, no damage was observed after the test. Through our studies we have confirmed the applicability of the selected SiCF/SiC composite as a combustor liner.

In this paper, we describe the present state of the R&D of a CMC combustor liner.

Commentary by Dr. Valentin Fuster
1998;():V005T13A003. doi:10.1115/98-GT-116.
FREE TO VIEW

At the Institut für Thermische Strömungsrnaschinen, University of Karlsruhe (ITS) a design technology has been introduced to reduce the mechanically and especially the thermally induced stresses in ceramic components. The concept is based on a three-layered construction (outer ceramic shell - heat insulating layer - metallic core) and an optimization of the thicknesses of the single layers, in order to obtain a homogenous temperature distribution in the ceramic structure. The optimization is performed by finite element analyses in combination with failure probability calculations.

This methodology has been applied to increase the reliability of a first stage Sintered Silicon Carbide (SSiC) ceramic nozzle vane of a stationary gas turbine (70MW/1400°C). As a result it was found that the mechanically and thermally induced loads have been reduced considerably and do not exceed 100MPa, thus achieving adequate life based upon failure probability calculations. Even in a trip situation (fuel cutoff), when the highest loads do occur, the calculations demonstrate a significantly reduced failure probability.

The results of the finite element analyses were verified by simulating the typical operating conditions after fuel cutoff in a test rig.

Commentary by Dr. Valentin Fuster
1998;():V005T13A004. doi:10.1115/98-GT-133.
FREE TO VIEW

A design for an impingement cooled and vented silicon nitride turbine vane for potential application to the Pratt & Whitney/Turbo Power and Marine FT8 aeroderivative stationary turbine is presented. The final design reduced the cooling air requirement for this first stage turbine vane by several percent of total core flow, compared to the current cobalt alloy bill of materials airfoil. Careful attention to achieving uniform internal and external temperature distributions, and design of the trailing edge vests, were necessary to ensure a design having a high Probability of Survival. The NASA CARES code was used to determine probabilistic fast fracture reliability of the silicon nitride airfoil.

Commentary by Dr. Valentin Fuster
1998;():V005T13A005. doi:10.1115/98-GT-181.
FREE TO VIEW

A program is being performed under the sponsorship of the United States Department of Energy, Office of Industrial Technologies, to improve the performance of stationary gas turbines in cogeneration through the selective replacement of metallic hot section components with ceramic parts. The program focuses on design, fabrication, and testing of ceramic components, generating a materials properties data base, and applying life prediction and nondestructive evaluation (NDE). The development program is being performed by a team led by Solar Turbines Incorporated, and which includes suppliers of ceramic components, U.S. research laboratories and an industrial cogeneration end user.

The Solar Centaur 50S engine was selected for the development program. The program goals included an increase in the turbine rotor inlet temperature (TRIT) from 1010°C (1850°F) to 1121°C (2050°F), accompanied by increases in thermal efficiency and output power. The performance improvements are attributable to the increase in TRIT and the reduction in cooling air requirements for the ceramic parts. The ceramic liners are also expected to lower the emissions of NOx and CO.

Under the program uncooled ceramic blades and nozzles have been inserted for currently cooled metal components in the first stage of the gas producer turbine. The louvre-cooled metal combustor liners have been replaced with uncooled continuous-fiber reinforced ceramic composite (CFCC) liners. Modifications have been made to the engine hot section to accommodate the ceramic parts.

To-date all first generation designs have been completed. Ceramic components have been fabricated, and are being tested in rigs and in the Centaur 50S engine. Field testing at an industrial co-generation site was started in May, 1997. This paper will provide an update of the development work and details of engine testing of ceramic components under the program.

Commentary by Dr. Valentin Fuster
1998;():V005T13A006. doi:10.1115/98-GT-186.
FREE TO VIEW

The introduction of a ceramic gas turbine component in commercial power generation service will require significant effort. A careful assessment of the power plant performance benefit achievable from the use of ceramic components is necessary to rationalize the priority of this development compared to other alternatives. This paper overviews a study in which the performance benefit from ceramic components was evaluated for an MS9001FA gas turbine in a combined cycle power plant configuration. The study was performed with guidelines of maintaining constant compressor inlet airflow and turbine exit NOx emissions, effectively setting the combustion reaction zone temperature. Cooling flow estimates were calculated to maintain standard design life expectancy of all components.

Monolithic silicon nitride ceramic was considered for application to the transition piece, stage one and two buckets, nozzles and shrouds. Performance benefit was calculated both for ceramic properties at 1093C (2200F) and for the more optimistic 1315C (2400F) oxidatian limit of the ceramic. Hybrid ceramic-metal components were evaluated in the less optimistic case.

Commentary by Dr. Valentin Fuster
1998;():V005T13A007. doi:10.1115/98-GT-336.
FREE TO VIEW

The use of high performance ceramic thermal barrier coatings in stationary gas turbines requires fundamental knowledge of their fatigue behavior under high temperature gradients and thermal cycling. An experimental method based on rapid laser heating complemented with finite-element calculations was developed in order to identify the major damage mechanisms and to obtain a data set for reliability assessment of thermal barrier coatings for temperature and stress fields similar to gas turbine conditions.

The observed failures are strongly related to the pretreatment procedures such as annealing under high temperature gradients and isothermal long-term oxidation. The vertical crack patterns observed close to the top surface of the Zirconia coating are generated at the moment of rapid cooling. These cracks are induced by high biaxial tensile stresses caused by the temperature gradient and the stress reversion after relaxation of compressive stresses at high temperatures. The long-term fatigue behavior is decisively determined by two processes:

(i) The porous Zirconia loses its damage tolerant properties by densification.

(ii) The growth of an oxide layer at the bond coat degrades adhesion and produces localized stress fields at the interface.

Cyclic loads increase the length of existing in-plane cracks and delaminations rather than enlarging their number. Misfit of the crack flanks and wedge effects are the driving forces for continued crack propagation.

These experimental results are discussed in terms of fracture mechanics.

Commentary by Dr. Valentin Fuster
1998;():V005T13A008. doi:10.1115/98-GT-348.
FREE TO VIEW

The paper describes a test facility for small-scale gas turbines, which basically has been designed and assembled at the Institute of Combustion Engines of the Technical University Berlin. The facility exposes ceramic rotor components to the most significant loads that occur during real gas turbine operation in a clearly predefined manner (high circumferential velocities and highest turbine inlet temperatures). The test facility allows the investigation of bladed radial inflow turbine rotors, as well as — in a preceding step — geometrically simplified ceramic or coated metallic rotors. A newly designed, ceramically lined, variable geometry combustion chamber allows turbine inlet temperatures up to 1450°C (2640 F). A fast thermal shock unit (switching time of about 1s), which is integrated into the test facility between the combustion chamber and the turbine scroll, can be used to create, for example, severe transient temperature gradients within the rotor components to simulate gas turbine trip conditions. In order to generate steady state temperature gradients, especially during disk testing, the rotor components can be subjected to an impingement cooling of the rotor back face (uncoated in case of TBC-testing). The test facility is additionally equipped with a non-contact transient temperature measurement system (turbine radiation pyrometry) to determine the test rotor surface temperature distribution during operation. Apart from the possibilities of basic rotor material investigations, the test facility can also be used to automatically generate compressor and turbine performance characteristics maps. The latter might be used to assess the aerodynamic performance of bladed ceramic radial inflow or mixed flow turbine rotors with respect to manufacturing tolerances due to near-net-shape forming processes (e.g., gelcasting or injection molding).

Commentary by Dr. Valentin Fuster
1998;():V005T13A009. doi:10.1115/98-GT-465.
FREE TO VIEW

Due to promising results in literature, SiC-Si3N4 particle composites in the range 0–100% SiC were evaluated. Focusing on high-temperature properties, mainly Y2O3 was used as sintering additive. Consolidation occurred primarily by hot-pressing, sintering tests were performed for comparison. Besides short-term properties like strength, toughness etc., long-term properties like creep and oxidation behavior were determined.

Results as a function of SiC-content and microstructure were discussed with respect to materials’ performance at high temperatures and possibilities of their production on a technical scale.

Commentary by Dr. Valentin Fuster
1998;():V005T13A010. doi:10.1115/98-GT-479.
FREE TO VIEW

The determination of ‘ultra’-fast fracture strengths of five silicon nitride ceramics at elevated temperatures has been made by using constant stress-rate (“dynamic fatigue”) testing with a series of ‘ultra’-fast test rates. The test materials included four monolithic and one SiC whisker-reinforced composite silicon nitrides. Of the five test materials, four silicon nitrides exhibited the elevated-temperature strengths that approached their respective room-temperature strengths at an ‘ultra’-fast test rate of 33 × 104 MPa/s. This implies that slow crack growth responsible for elevated-temperature failure can be eliminated or minimized by using the ‘ultra’-fast test rate. These ongoing experimental results have shed light on laying a theoretical and practical foundation on the concept and definition of elevated-temperature “inert” strength behavior of advanced ceramics.

Commentary by Dr. Valentin Fuster
1998;():V005T13A011. doi:10.1115/98-GT-480.
FREE TO VIEW

The present study is focussed on the oxidation behavior of nonoxide silicon-based ceramics. Various Si3N4 and SiC ceramics were examined after long term oxidation tests (up to 5000 h) at 1500°C in ambient air. The damage mechanisms were discussed on the basis of a comprehensive chemical and microstructural analysis of the materials after the oxidation tests. The diffusion of oxygen into the material and its further reaction in the bulk of the material were found to be the most critical factors during long term oxidation treatment at elevated temperatures. However, the resulting damage in the microstructure of the materials can be significantly reduced by purposeful microstructural engineering. Using Si3N4/SiC and Si3N4/MoSi2 composite materials provides the possibility to improve the high temperature stability.

Commentary by Dr. Valentin Fuster
1998;():V005T13A012. doi:10.1115/98-GT-489.
FREE TO VIEW

High temperature and long duration applications of monolithic ceramics can place their failure mode in the creep rupture regime. A previous model advanced by the authors described a methodology by which the creep rupture life of a loaded component can be predicted. That model was based on the life fraction damage accumulation rule in association with the modified Monkman-Grant creep rupture criterion. However, that model did not take into account the deteriorating state of the material due to creep damage (e.g., cavitation) as time elapsed. In addition, the material creep parameters used in that life prediction methodology, were based on uniaxial creep curves displaying primary and secondary creep behavior, with no tertiary regime. The objective of this paper is to present a creep life prediction methodology based on a modified form of the Kachanov-Rabotnov continuum damage mechanics (CDM) theory. In this theory, the uniaxial creep rate is described in terms of stress, temperature, time, and the current state of material damage. This scalar damage state parameter is basically an abstract measure of the current state of material damage due to creep deformation. The damage rate is assumed to vary with stress, temperature, time, and the current state of damage itself. Multiaxial creep and creep rupture formulations of the CDM approach are presented in this paper. Parameter estimation methodologies based on nonlinear regression analysis are also described for both, isothermal constant stress states and anisothermal variable stress conditions This creep life prediction methodology was preliminarily added to the integrated design code. CARES/Creep (Ceramics Analysis and Reliability Evaluation of Structures/Creep), which is a postprocessor program to commercially available finite element analysis (FEA) packages. Two examples, showing comparisons between experimental and predicted creep lives of ceramic specimens, are used to demonstrate the viability of this methodology and the CARES/Creep program.

Commentary by Dr. Valentin Fuster
1998;():V005T13A013. doi:10.1115/98-GT-527.
FREE TO VIEW

Machining damage (a surface flaw) and porous-region-flaw (a volume flaw) populations limited the flexure strengths of a commercially available silicon nitride at 25°C, while these same flaws, along with inclusions, limited flexure strengths at 850°C. The machining damage and porous region flaws were the primary interest in the present study because they caused failure at both temperatures. Censoring revealed that the two-parameter Weibull strength distributions representing each flaw population changed as a function of stressing rate (i.e., dynamic fatigue) and temperature. A decrease in the Weibull scaling parameter is recognized as an indication of slow crack growth or time-dependent strength reduction in monolithic ceramics. Available life prediction codes used for reliability predictions of structural ceramic components consider the slow crack growth phenomenon. However, changes in the Weibull modulus are infrequently observed or reported, and typically are not accounted for in these life prediction codes. In the present study, changes in both Weibull parameters for the strength distributions provided motivation to the authors to survey what factors (e.g., residual stress, slow crack growth, and changes in failure mechanisms) could provide partial or full explanation of the observed distribution changes in this silicon nitride. Lastly, exercises were performed to examine the effects of strength distribution changes on the failure probability prediction of a diesel exhaust valve. Because the surface area and volume of this valve were substantially larger than those of the tested bend bars, it was found that the valve’s failure probability analysis amplified some slight or inconclusive distribution changes which were not evident from the interpretation of the censored bend bar strength data.

Commentary by Dr. Valentin Fuster
1998;():V005T13A014. doi:10.1115/98-GT-528.
FREE TO VIEW

The goal of the Ceramic Stationary Gas Turbine(CSGT) Development Program, under the sponsorship of the United States Department of Energy (DOE), Office of Industrial Technologies (OIT), is to improve the performance (fuel efficiency, output power, exhaust emissions) of stationary gas turbines in cogeneration through the selective replacement of hot section components with ceramic parts. Phase II of this program includes detailed engine and component design, procurement and testing. This paper will review the design and test of the first stage ceramic nozzle for the Centaur 50S engine. For this test an uncooled monolithic ceramic nozzle made from SN-88 silicon nitride(NGK Insulators Ltd.) was used.

A major challenge in the successful introduction of ceramic parts into a gas turbine is the design of the interface between the ceramic parts and metallic components. The design and attachment of the ceramic nozzle was greatly influenced by these considerations. Metallic components in the stationary structure of the turbine have been added or redesigned to retrofit the ceramic nozzle into the all metallic Centaur 50S engine.

This paper will also discuss special handling and assembly techniques used to install the ceramic nozzle into the engine. Trial assemblies were used in the engine build process, this proved most beneficial in identifying problems and reducing the risk of damage to the ceramic nozzles. Assembly techniques were designed to reduce assembly loads and to eliminate blind assemblies.

Before installing any ceramic nozzles into the engine they were first required to successfully pass both mechanical and thermal proof tests. Details of these proof tests and the final full load engine test will be described in this paper. The engine test was run at a turbine rotor inlet temperature(TRIT) of 1010°C. Total number of engine starts was six, and the total run time was approximately 10 hours.

Commentary by Dr. Valentin Fuster
1998;():V005T13A015. doi:10.1115/98-GT-529.
FREE TO VIEW

The goal of the Ceramic Stationary Gas Turbine (CSGT) Development Program, under the sponsorship of the United States Department of Energy (DOE), Office of Industrial Technologies (OIT), is to improve the performance (fuel efficiency, output power, and exhaust emissions) of stationary gas turbines in cogeneration through the selective replacement of hot section components with ceramic parts. This program, which is headed by Solar Turbines Incorporated and supported by various suppliers, and national research institutes, includes detailed engine and component design, procurement, and field testing.

A major challenge in the successful introduction of ceramic parts into a gas turbine is the design of the interface between the ceramic parts and metallic hardware. A turbine blade, which incorporated a dovetail root, was designed with such considerations. A relatively thin compliant layer between the ceramic-metallic loading surface was considered for equalizing pressure face load distributions.

Five monolithic siliocn nitride ceramic materials were considered: AS800 and GN10, AlliedSignal Ceramic Components; NT164, Norton Advanced Ceramics; SN281 and SN253, Kyocera Industrial Ceramics Corporation. The probability of survival using NASA/CARES for 30,000 hours of engine operation was calculated for each material. The blade frequencies, stresses, and temperatures were predicted. The influence of the dovetail angle was also analyzed to determine the most optimum configuration. Prior to engine installation all blades underwent extensive nondestructive evaluation and spin proof testing. This paper will review the design, life prediction, and testing of the first stage ceramic turbine blade for the Solar Turbines Centaur 5OS engine.

Commentary by Dr. Valentin Fuster
1998;():V005T13A016. doi:10.1115/98-GT-530.
FREE TO VIEW

An update is presented of the activities of the American Society for Testing and Materials (ASTM) Committee C-28 on Advanced Ceramics. Since its inception in 1986, this committee, which has five standard producing subcommittees, has written and published over 32 consensus standards. These standards are concerned with mechanical testing of monolithic and composite ceramics, nondestructive examination, statistical analysis and design, powder characterization, quantitative microscopy, fractography, and terminology. These standards ensure optimum material behavior with physical and mechanical property reproducibility, component reliability, and well-defined methods of data treatment and material analysis for both monolithic and composite materials. Committee C-28 continues to sponsor technical symposia and to cooperate in the development of international standards. An update of recent and current activities as well as possible new areas of standardization work will be presented.

Commentary by Dr. Valentin Fuster
1998;():V005T13A017. doi:10.1115/98-GT-567.
FREE TO VIEW

Development of an updated single filament test standard for tensile strength determination has been undertaken by a joint government-industry task group under the auspices of the Subcommittee C28.07 of the American Society for Testing and Materials (ASTM). This update was initiated in response to a need expressed by people involved with fiber testing. A fiber testing workshop in 1991 attracted a wide range of attendees with concerns regarding the uniformity of testing methods and comparability of resulting data in the fiber literature. With the inception of a C28.07.07 Task Group for Ceramic Fibers, an effort to conduct round robin testing was begun among interested government and industrial test facilities. User surveys and discussion groups were also employed to determine areas of concern with the current standard test method, ASTM D 3379-75. Shortcomings in the current standard were identified and methods of addressing these areas were discussed with experts including fiber manufacturers, testing facilities and composite manufacturers. Among the main concerns was the method and application of fiber diameter measurements.

Commentary by Dr. Valentin Fuster
1998;():V005T13A018. doi:10.1115/98-GT-569.
FREE TO VIEW

This paper presents a computational methodology for life prediction and time-dependent reliability analysis of ceramic structures under combined effects of static and cyclic fatigue. It involves (1) a crack-growth equation representing damage contributions from both static and cyclic fatigue, (2) a multivariate nonlinear regression model for performing parameter estimation from fatigue data generated by small specimens, and (3) the Batdorf model for structural reliability analysis. A linear superposition of crack-growth rates obtained from the Power-law and Walker-law equations was used. The model assumes that the time-dependent and cycle-dependent crack growth formulation exponents are identical, and that loading frequency and amplitude do not vary over time. For the parameter estimation, the regression was performed using nonlinear least squares and a modified Levenberg-Marquardt algorithm. This methodology was implemented into the integrated design code named CARES/Life (Ceramics Analysis and Reliability Evaluation of Structures/Life). A numerical example is presented to illustrate the parameter estimation component of this methodology. The results suggest that the predicted stress-life curves based on the proposed model can correlate better with experimental data when compared with either Power-law or the Walker-law models individually.

Commentary by Dr. Valentin Fuster
1998;():V005T13A019. doi:10.1115/98-GT-595.
FREE TO VIEW

Strength and fatigue lifetime of structural ceramics under multiaxial stress state have been estimated and compared with experimental data. Biaxial strength tests were done by an anticlastic bending test method at room temperature. Biaxial fatigue tests were done by anticlastic bending and also ring-on-ring test method at 1200°C in air. Fracture probability and lifetime were predicted on the basis of a Weibull multiaxial distribution function and subcritical crack growth, using the results of stress analyses by the finite element method. Modified maximum hoop stress theory including an empirical parameter, T, was applied to the equivalent normal stress in the multiaxial distribution function. The empirical parameter T represents a shear stress sensitivity to mixed-mode fracture due to a grain interlocking effect. It has been confirmed that the predicted fracture probability and the fatigue lifetime agrees well with the experimental data if grain interlocking effects are taking into account.

Commentary by Dr. Valentin Fuster

Structures and Dynamics

1998;():V005T14A001. doi:10.1115/98-GT-005.
FREE TO VIEW

The work presents a method for analyzing the dynamic regime of labyrinth liquid seals. By using the traditional simplifying assumptions for the centered seal (sinusoidal, harmonically varying, first order dynamic perturbation), the approach can be addressed as “quasi” 2D. A numerical coordinate transformation capable to treat displacement perturbations is introduced. The first order mathematical model is then deduced following the same steps as in a previously published work (Arghir et Frêne, 1997b). From this standpoint, the present method can be regarded as an extension of the above mentioned approach which was able to deal only with stator-grooved seals. The method is validated by comparisons with Nordmann and Dietzen’s (1988) theoretical results for a seal with grooves on both stator and rotor and with the experimental results of Staubli’s (1993) test case concerning a general seal.

Commentary by Dr. Valentin Fuster
1998;():V005T14A002. doi:10.1115/98-GT-006.
FREE TO VIEW

Rotor vibration attenuation and structural components isolation in jet engines are achieved with squeeze film dampers, many of them supported on long elastic squirrel cages. Integral squeeze film dampers (ISFDs) are comprised of arcuate pads and wire-EDM webs rendering a compact viscoelastic support. An experimental study is conducted to evaluate the effectiveness of ISFDs in attenuating the imbalance response of a massive test rotor. Measurements of the damper structural stiffness and rotor natural frequencies are detailed. Impact tests on the test rotor supported on its dampers reveal the supporting structure to be very flexible, thus requiring the experimental evaluation of an equivalent stiffness for the damper and supports system. System damping coefficients extracted from impact load excitations vary with the lubricant viscosity and include a significant structural damping from the bearing supports. Rotor coast-down tests demonstrate the ISFDs to damp well the rotor response with peak vibration amplitude proportional (linear) to the imbalance. Viscous damping coefficients estimated from the amplitude response at the critical speeds agree reasonably well with predictions from a full-film, finite element model.

Topics: Dampers , Rotors
Commentary by Dr. Valentin Fuster
1998;():V005T14A003. doi:10.1115/98-GT-008.
FREE TO VIEW

Centrifugal compressors are increasingly required to operate at higher pressures, speeds, and fluid density. In these conditions, compressors are susceptible to rotordynamic instabilities. To remedy this situation, labyrinth seals have sometimes been modified by using shunt injection. In shunt injection, the gas is taken from the diffuser or discharge volute and injected into an upstream chamber of the balance-piston labyrinth seal. The injection direction can be radial or against rotation. This study contains the first measured rotordynamic data for labyrinth seals with shunt injection. A comparison has been made between conventional labyrinth seals, labyrinth seal with shunt injection (radial and against rotation), and a honeycomb seal. Labyrinth seals with injection against rotation are better able to control rotordynamic instabilities than labyrinth seals with radial injection; however, the leakage is slightly higher. The leakage comparison for all seals demonstrates that the honeycomb seal has the best flow control. Test data are presented for a top rotor surface velocity of 110 m/sec, a supply pressure of 13.7 bars, and IPr = 0.95 (injection pressure is 1.05 = 1/0.95 times the seal inlet pressure). For these conditions, and considering effective damping, the labyrinth seal with injection against rotation is better than the honeycomb seal when the pressure ratio across the seal PR<0.45. On the other hand, the honeycomb seal is better when PR>0.45. The effectiveness of the shunt-injection against rotation in developing effective damping is reduced with increasing rotor surface velocity.

Commentary by Dr. Valentin Fuster
1998;():V005T14A004. doi:10.1115/98-GT-013.
FREE TO VIEW

A bulk-flow model for calculation of the dynamic force characteristics in a single cavity, multiple-pocket gas damper seal is presented. Flow turbulence is accounted for by using turbulent shear stress parameters and Moody’s friction factors in the circumferential momentum equation. Zeroth order-equations describe the isothermal flow field for a centered seal, and first-order equations govern the perturbed flow for small amplitude rotor lateral motions. Comparisons to limited measurements from a four-pocket gas damper seal show the current model to predict well the mass flow rate and the direct damping coefficient. For a reference two-bladed teeth-on-stator labyrinth seal, the current model predicts similar rotordynamic coefficients when compared to results from a two control-volume bulk-flow model. Force coefficients from a reference single-cavity, four pocket gas damper depend on the rotor speed and pressure drop with magnitudes decreasing as the rotor whirl frequency increases. The multiple-pocket gas damper seal provides substantially more damping than a conventional labyrinth seal of the same dimensions. The damper seal cross-coupled stiffness coefficients are small though sensitive to the inlet circumferential pre-swirl flow.

Commentary by Dr. Valentin Fuster
1998;():V005T14A005. doi:10.1115/98-GT-014.
FREE TO VIEW

Anti-swirl gas dampers have applications in high-temperature turbomachinery. Nozzles comprising the circumference of the damper inject air against the direction of shaft rotation, providing a force tangential to the rotor surface that acts to damp rotor vibration. The present work involves prediction of experiments by using direct damping and cross-coupled stiffness coefficients given in Vance and Handy (1997) that characterize the rotordynamic performance of an anti-swirl damper. Direct stiffness added at the damper location is vital to representing the changes in critical speeds and the onset and cease of backward whirl given in the experiments. This direct stiffness arises due to the release of air axially across the annulus of the damper. With the addition of this significant direct stiffness, experimental results compare well with the present rotordynamic model. Predictions using the experimentally obtained damping coefficients adequately reproduce the reduction in vibration amplitudes at the critical speeds. However, applying the cross-coupled stiffness coefficients in predictions fails to show increases in the speed at which backward whirl begins and does not reproduce the wrecking instability experienced in the tests. A further study investigates the magnitude of the cross-coupled stiffness coefficient necessary to cause instability in the rotor-damper system.

Topics: Dampers
Commentary by Dr. Valentin Fuster
1998;():V005T14A006. doi:10.1115/98-GT-017.
FREE TO VIEW

Annular seals are known to enhance rotordynamic stability margins and minimize vibration response levels in high-speed rotating machinery. Theoretical predictions for the rotordynamic characteristics of annular seals exist but additional experimental data is needed to properly anchor these results. NASA’s Marshall Space Flight Center (MSFC) has developed an annular seal test rig and facility to experimentally characterize axially-fed annular seals. The objective of MSFC’s annular seal test rig is to obtain the rotordynamic coefficients (direct and cross-coupled stiffness, damping, and added mass) for a variety of high Reynolds number annular seals. The MSFC test rig supports centered-seal testing with inlet pressures up to 138 bars (2000 psi) and flow rates of over 946 liters per minute (250 gpm). The rig’s shaft is powered by a 186 kilowatt (250 horsepower) steam turbine capable of rotational speeds of over 20,000 revolutions per minute (rpm). A description of the identification process used to obtain rotordynamic coefficients is given as well as procedures for ensuring quality data. Experimental results for a smooth annular seal with an L/D = 0.5 is presented. Excellent agreement between experimental and theoretical results is obtained.

Commentary by Dr. Valentin Fuster
1998;():V005T14A007. doi:10.1115/98-GT-018.
FREE TO VIEW

The three-dimensional, multi-stage, unsteady, turbomachinery analysis, TURBO, has been extended to predict the aeroelastic and aeroacoustic response behaviors of a blade row operating within a cylindrical annular duct. In particular, a blade vibration capability has been incorporated so that the TURBO analysis can be applied over a solution domain that deforms with a vibratory blade motion. Also, unsteady far-field conditions have been implemented to render the computational inlet and exit boundaries transparent to outgoing unsteady disturbances and to allow for the prescription of incoming aerodynamic excitations. The modified TURBO analysis has been applied to predict unsteady subsonic and transonic flows. The intent is to partially validate this nonlinear analysis for blade flutter applications via numerical results for benchmark unsteady flows, and to demonstrate this analysis for a realistic fan rotor. For these purposes, we have considered unsteady subsonic flows through a 3D version of the 10th Standard Cascade and unsteady transonic flows through the first stage rotor of the NASA Lewis, Rotor 67 fan. Some general correlations between aeromechanical stabilities and fan operating characteristics will be presented.

Commentary by Dr. Valentin Fuster
1998;():V005T14A008. doi:10.1115/98-GT-027.
FREE TO VIEW

Flexible bearing supports may have a great influence in the calculation of forced response and stability of rotor systems. However, this effect is not always included in rotor analyses since an accurate model of the foundation and pedestals may be difficult and costly to obtain. It is common practice to use either a one degree of freedom model or a full modal analysis to represent the bearing supports. While the one degree of freedom model is easy to set up for computer calculations, it often requires experience to determine values for the stiffness, mass and damping of the model that will accurately represent the support under study. This model, however, fails to capture the dynamics of the system for stability analyses when more than one mode of the support structure is in the range of interest. On the other hand, modal representation provides much more information and can be measured experimentally, but requires measurement of the mode shapes. Even though modal representation can include all the dynamics of the system in the frequency range of interest, it provides much more information than is required for calculation of the rotor response and it is more difficult to use in calculation programs. This paper presents a procedure to include the support characteristics using transfer functions. Transfer functions permit modeling of multi-degree of freedom systems while maintaining the size of a one degree of freedom system (2×2 matrix if rotation at the bearing is not considered). Another advantage of transfer functions is that they can be obtained from existing discrete models, from modal information or can be measured directly. The fixed size of the transfer function matrix permits the method to be easily incorporated into rotor dynamic stability and forced response programs. The method is applied to stability calculations of models of typical industrial machines.

Commentary by Dr. Valentin Fuster
1998;():V005T14A009. doi:10.1115/98-GT-042.
FREE TO VIEW

In the present work the unsteady aerodynamic characteristics of harmonically oscillating fan blades are investigated by applying a time-shifted boundary condition at the periodic boundaries. The direct-store method is used to implement the time-shifted boundary condition in a time-marching Euler/Navier-Stokes solver. Inviscid flow calculations for a flat plate helical fan, in a single-blade passage domain, are used to verify the analysis. The results obtained show good correlation with other published results as well as with the same solver using multiple blade passages stacked together. Significant savings in computer time is realized, especially for smaller phase angles.

Commentary by Dr. Valentin Fuster
1998;():V005T14A010. doi:10.1115/98-GT-043.
FREE TO VIEW

A quasi-three-dimensional multigrid Navier-Stokes solver on single and multiple passage domains is presented for solving unsteady flows around oscillating turbine and compressor blades. The conventional “direct store” method is used for applying the phase-shifted periodic boundary condition over a single blade passage. A parallel version of the solver using the Message Passing Interface (MPI) standard is developed for multiple passage computations. In the parallel multiple passage computations, the phase-shifted periodic boundary condition is converted to simple in-phase periodic condition. Euler and Navier-Stokes solutions are obtained for unsteady flows through an oscillating turbine cascade blade row with both the sequential and the parallel code. It is found that the parallel code offers almost linear speedup with multiple CPUs. In addition, significant improvement is achieved in convergence of the computation to a periodic unsteady state in the parallel multiple passage computations due to the use of in-phase periodic boundary conditions as compared to that in the single passage computations with phase-lagged periodic boundary conditions via the “direct store” method. The parallel Navier-Stokes code is also used to calculate the flow through an oscillating compressor cascade. Results are compared with experimental data and computations by other authors.

Commentary by Dr. Valentin Fuster
1998;():V005T14A011. doi:10.1115/98-GT-051.
FREE TO VIEW

Many industrial machines nowadays are sold based on analysis performed on mathematical models of the rotors, bearings, substructures, and other components. The validity of the analysts therefore depends on the accuracy of the models themselves. When the rotor is available, modal testing may be used to validate the model of the rotor by comparing the calculated and measured free-free natural frequencies and mode shapes. This work presents additional tools for the verification of analytical models against experimental data. These tools use models of the rotor constructed from the measured data and the analytical model. A comparison of the first six calculated and measured free-free natural frequencies and mode shapes for a multi-mass flexible rotor is presented. The natural frequencies compare within 1.8%. The calculated and measured mode shapes were used to construct independent reduced order models of the rotor. These models were used to perform forced response and stability analyses. Forced response functions are presented comparing the forced response characteristics obtained from the two models. This provides a comparison between the measured and calculated forced response functions for the same number of modes. For the stability analysis, identical bearing models were added to both reduced order models. The eigenvalues were calculated using both models for a range of bearing stiffness and damping coefficients and were plotted for comparison.

Topics: Rotors
Commentary by Dr. Valentin Fuster
1998;():V005T14A012. doi:10.1115/98-GT-052.
FREE TO VIEW

A frequent cause of turbomachinery blade failure is excessive vibration due to flutter or forced response. One method for dealing with this problem is to increase blade structural damping using either tip or mid-span shroud designs. Unfortunately, most existing aeroelastic analyses deal with a blade alone model which can not be used for system mode analysis. Therefore, judgments based on past experience are used to determine the acceptability of a shrouded blade design.

A new cyclic symmetry analysis has been developed to predict shrouded blade flutter. The method provides a system approach, over and above the standard blade alone approach, for predicting potential aeroelastic problems. Using the blade natural frequencies and mode shapes from a cyclic symmetry finite element model, the unsteady aerodynamic forces of the system mode are calculated. A cyclic symmetry flutter analysis is then performed.

This analysis has been applied to a typical shrouded fan blade to investigate blade flutter. The predicted system mode flutter demonstrated that the blade alone analysis can be non-conservative.

Commentary by Dr. Valentin Fuster
1998;():V005T14A013. doi:10.1115/98-GT-105.
FREE TO VIEW

In this paper, the effect of blade crack on the mode localization of a rotating blade-disk is studied. Pretwisted taper beams are used to simulate blades of a blade-disk. The crack on the blade can be regarded as a local disorder of this periodically coupled blades system. An application of Hamilton’s principle and Galerkin’s method is used to formulate the equations of motion of the mistuned system. Effects of pretwisted angle, rotating speed and crack depth of the blade on the in-plane and off-plane mode localizations of a rotating system are investigated. Numerical results indicate that the increase of rotating speed, pretwisted angle and crack depth could enhance the localization phenomenon significantly.

Topics: Disks , Blades
Commentary by Dr. Valentin Fuster
1998;():V005T14A014. doi:10.1115/98-GT-106.
FREE TO VIEW

This study focuses on the fully reversed fatigue behavior exhibited by a carbon fiber/polyimide resin woven laminate at room and elevated temperatures. Nondestructive video edge view microscopy and destructive sectioning techniques were used to study the microscopic damage mechanisms that evolved. The elastic stiffness was monitored and recorded throughout the fatigue life of the coupon. In addition, residual compressive strength tests were conducted on fatigue coupons with various degrees of damage as quantified by stiffness reduction. Experimental results indicated that the monotonic tensile properties were only minimally influenced by temperature, while the monotonic compressive and fully reversed fatigue properties displayed greater reductions due to the elevated temperature. The stiffness degradation, as a function of cycles, consisted of three stages; a short-lived high degradation period, a constant degradation rate segment covering the majority of the life, and a final stage demonstrating an increasing rate of degradation up to failure. Concerning the residual compressive strength tests at room and elevated temperatures, the elevated temperature coupons appeared much more sensitive to damage. At elevated temperatures, coupons experienced a much larger loss in compressive strength when compared to room temperature coupons with equivalent damage. The fatigue damage accumulation law proposed for the model incorporates a scalar representation for damage, but admits a multiaxial, anisotropic evolutionary law. The model predicts the current damage (as quantified by residual stiffness) and remnant life of a composite that has undergone a known load at temperature. The damage/life model is dependent on the applied multiaxial stress state as well as temperature. Comparisons between the model and data showed good predictive capabilities concerning stiffness degradation and cycles to failure.

Commentary by Dr. Valentin Fuster
1998;():V005T14A015. doi:10.1115/98-GT-109.
FREE TO VIEW

Squeeze film dampers (SFDs) are effective means to reduce vibrations and to suppress instabilities in rotor-bearing systems. However, at operating conditions while traversing critical speeds with large orbital whirl motions, ingestion and entrapment of air into the thin lands of SFDs generates a bubbly mixture (air in lubricant) which is known to reduce the dynamic film pressures and the overall damping capability. This pervasive phenomenon lacks proper physical understanding and sound analytical modeling. An experimental investigation to quantify the forced performance of a SFD operating with a controlled bubbly mixture is detailed. Tests are conducted in a constrained circular orbit SFD to measure the dynamic squeeze film pressures and journal motion at two whirl frequencies (8.33 and 16.67 Hz) as the air content in the mixture increases from 0% to 100%. The analysis of period-averaged film pressures reveals a zone of uniform low pressure of magnitude equal to the discharge pressure, independently of the mixture composition. The uniform pressure zone extends as the mixture void fraction increases. Radial and tangential film forces are estimated from the dynamic pressures at two axial locations of measurement. The tangential (damping) force decreases proportionally with the mixture volume fraction, while a radial hydrostatic force remains nearly invariant. The experimental results quantify effects previously known by qualitative description only, thus providing a benchmark towards the development of sound theoretical models.

Topics: Lubricants , Stress , Dampers
Commentary by Dr. Valentin Fuster
1998;():V005T14A016. doi:10.1115/98-GT-257.
FREE TO VIEW

Correlation between dynamic strain gage measurements and modal analysis results can be adversely affected by gage misplacement and gage misorientation. An optimization algorithm has been developed which allows the modeled strain gage locations and orientations to be varied within specified tolerances. An objective function is defined based on the least squares sum of the differences between experimental and model results. The Kuhn-Tucker conditions are then applied to find the gage locations and orientations which minimize this objective function. The procedure is applied on a one-time basis considering all measured modes of vibration simultaneously. This procedure minimizes instrumentation error which then allows the analyst to modify the model to more accurately represent other factors, including boundary conditions. Flat plate vibratory data was used to demonstrate a significant improvement in correlation between measured data and model predictions.

Commentary by Dr. Valentin Fuster
1998;():V005T14A017. doi:10.1115/98-GT-263.
FREE TO VIEW

A simplified solution method that enables the estimation of stresses and strains in high temperature components under creep conditions is presented. The solution is derived based on strain energy density considerations and is applicable to both uniaxial and multiaxial stress states. In particular, this simplified method is developed for an efficient estimation of the cyclic stress-strain history at critical locations which needed for fatigue analysis of hot sections under creep conditions where conventional finite element creep analysis becomes extremely time consuming. The input data necessary to perform this simplified solution are the stresses and strains obtained from a linear elastic analyses. If the finite element method (FEM) is used for the linear elastic analysis of components, then the simplified solution method can be programmed as a post processor file. The file uses the linear elastic FEM results and generates an approximate time-dependent analysis. Presented results illustrates the accuracy of the method by comparing with finite element creep analysis results for several hot sections under creep conditions. Also, it is shown that the computational time needed to perform this solution is far less than the conventional finite element creep analysis.

Commentary by Dr. Valentin Fuster
1998;():V005T14A018. doi:10.1115/98-GT-297.
FREE TO VIEW

The rotordynamic behavior of an industrial gas turbine rotor train was assessed on site, and the sensitivity to unbalance was quantified. An outline of the measurement procedure is given.

Differential data reduction with test unbalances was undertaken to minimize the influence of measurement uncertainty. A test unbalance was installed for one run and then shifted by 180° for the consecutive run. With differential data, the effective dynamic properties of the rotor - support - system can be estimated more accurately.

A rotordynamic model was used to identify the support system parameters based on measured data. For the analysis, the anisotropic, elliptical vibration orbits were decomposed into two counter-rotating circular orbits, and the support system parameters identified match the originally predicted values well.

The methods of differential data reduction, rotor train mode shape presentation, elliptical orbit decomposition, and the link of measurement to analytical models with parameter definition are described. Examples from on-site measurements are included for illustration.

Commentary by Dr. Valentin Fuster
1998;():V005T14A019. doi:10.1115/98-GT-305.
FREE TO VIEW

This paper reports the synchronous vibration instability problem (a rare phenomenon) experienced in a high pressure steam turbine rotor (19MW) driving synthesis gas compressor train in a large scale petrochemical complex. The turbine had about one year history of showing infrequently high vibration. Rotor vibrations appeared in an intermittent and irregular fashion and the perturbation frequency was the rotor operating speed of 10,135 rpm. The sealing steam system was found responsible for cropping the vibration. At a definite level of seal steam pressure (0.90 to 1.10 bar-gauge), operating speed and load, the rotor radial vibration response was reached at 4.5 mils as compared to the frequently smooth running level of less than 1.0 mil. Subsequently, the major overhauling of the turbine revealed severely worn and, virtually, non-functional high pressure end labyrinth seals. The paper also elaborates the steam turbine rotordynamics behaviors recorded during excessive levels of vibration.

Commentary by Dr. Valentin Fuster
1998;():V005T14A020. doi:10.1115/98-GT-316.
FREE TO VIEW

Rotor power losses in magnetic bearings cannot be accurately calculated at this time because of the complexity of the magnetic field distribution and several other effects. The losses are due to eddy currents, hysteresis, and windage. This paper presents measured results in radial magnetic bearing configurations with 8 pole and 16 pole stators and two laminated rotors. Two different air gaps were tested. The rotor power losses were determined by measuring the rundown speed of the rotor after the rotor was spun up to speeds of approximately 30,000 rpm, DN = 2,670,000 mm-rpm, in atmospheric air. The kinetic energy of the rotor is converted to heat by magnetic and air drag power loss mechanisms during the run down. Given past publications and the opinions of researchers in the field, the results were quite unexpected. The measured power losses were found to be nearly independent of the number of poles in the bearing. Also, the overall measured rotor power loss increased significantly as the magnetic flux density increased and also increased significantly as the air gap thickness decreased.

A method of separating the hysteresis, eddy current and windage losses is presented. Eddy current effects were found to be the most important loss mechanism in the data analysis, for large clearance bearings. Hysteresis and windage effects did not change much from one configuration to the other.

Commentary by Dr. Valentin Fuster
1998;():V005T14A021. doi:10.1115/98-GT-317.
FREE TO VIEW

The identification of parameters that dictate the magnitude of rotor power losses in radial magnetic bearings is very important for many applications. Low loss performance of magnetic bearings in aerospace equipment such as jet engines and flywheel energy storage systems is especially critical. Two basic magnetic bearing designs are employed in industrial practice today: the homopolar design, where the flux paths are of a mixed radial/axial orientation, and the heteropolar design, where the flux paths are primarily radial in nature.

The stator geometry and flux path of a specific bearing can have a significant effect on the rotor losses. This paper describes the detailed measurement of rotor losses for experimentally comparable homopolar and heteropolar designs. The two test bearing configurations are identical except for geometric features that determine the direction of the flux path. Both test bearing designs have the same air gap length, tip clearance ratio, surface area under the poles, and bias flux levels. An experimental test apparatus was used where run down tests were performed on a test rotor with both bearing designs to measure power losses. Numerous test runs where made for each bearing configuration by running multiple levels of flux density. The components of the overall measured power loss, due to hysteresis, eddy currents, and windage, were determined based on theoretical expressions for power loss. It was found that the homopolar bearing had significantly lower power losses than the heteropolar bearing.

Commentary by Dr. Valentin Fuster
1998;():V005T14A022. doi:10.1115/98-GT-318.
FREE TO VIEW

TMF tests were conducted with bare and aluminide coated single crystal nickel-based superalloy specimens. Temperature cycling was between 400°C and 1100°C with a phase shift (135°) which is typical for damaged locations on turbine blades. Stress response is characterized by a constant range and the formation of a tensile mean stress as a result of relaxation in the high temperature part of the cycle which is in compression. Bare specimens showed crack initiation from typical oxide hillocks. Coated specimens showed life reduction with respect to the bare ones caused by brittle cracking of the coating in the low temperature part of the cycle. Isothermal bending tests of coated specimens confirmed the low ductility of the coating at tempeatures below 600°C but quantitative correlation with the TMF test results failed.

Commentary by Dr. Valentin Fuster
1998;():V005T14A023. doi:10.1115/98-GT-319.
FREE TO VIEW

The tensile response of continuous fibre reinforced ceramic matrix composites (CFCCs) is not expected to show the large variation in strength properties commonly observed for monolithic ceramics. Results of recent investigations on a number of 2D reinforced CFCCs have nevertheless revealed a considerable scatter in the initial elastic modulus, in the first matrix cracking stress and in the failure stress. One school of thought considers that the observed variability is caused by experimental factors. Elaborate testing programmes have been set up to clarify the origins of this scatter by investigation of the effects of control mode, loading rate, specimen shape, etc.. Another school explains the scatter by the presence of (axial) residual stresses in the fibres and in the matrix. Although plausible, this hypothesis is difficult to verify because experimental determination of the residual stress state in CFCCs is not straightforward. In addition, with the available methods it is impractical to determine the residual stresses in every test specimen. This approach is indeed required for establishing the relationship between the magnitude of the residual stresses and the experimentally observed scatter.

At IAM a method has been developed and validated which allows to quantify the axial residual stress state in individual CFCC specimens by subjecting them to intermittent unloading-reloading cycles. The method as well as the derived relationship between residual stress state and scatter in mechanical response will be presented.

Commentary by Dr. Valentin Fuster
1998;():V005T14A024. doi:10.1115/98-GT-324.
FREE TO VIEW

Three different coatings were studied in this work: vacuum plasma-sprayed NiCoCrAlYTa, electrolytically deposited NiCoCrAlYTa and Ni-Pt aluminide diffusion coatings. These three coatings were deposited on AM3 single crystal alloy. The tensile properties of coated single crystal test specimens were investigated. Ductile to Brittle Transition Temperatures (DBTT) were determined from tensile tests. All the coatings were examined before and after testing.

All the tested coatings induce a ductile/brittle transition. Strain rate has a great influence on the transition temperature. The comparison between plasma-sprayed deposition and electrodeposition illustrates the strong influence of coating microstructure. In every case, NiCoCrAlYTa coatings were more ductile, and then less detrimental, than aluminide coatings.

Topics: Coatings , Turbines
Commentary by Dr. Valentin Fuster
1998;():V005T14A025. doi:10.1115/98-GT-344.
FREE TO VIEW

Dynamic loading of an inlet guide vane (IGV) in a transonic compressor is characterized by unsteady IGV surface pressures. This pressure data was acquired for two spanwise locations at a 105% speed operating condition, which produces supersonic relative Mach numbers over the majority of the rotor blade span. The back pressure of the compressor was varied to determine the effects from such changes. Strong bow shock interaction was evident in both experimental and computational results. Variations in the back pressure have significant influence on the magnitude and phase of the upstream pressure fluctuations. The largest unsteady surface pressure magnitude, 40 kPa, was obtained for the near stall mass flow condition at 75% span and 95% chord. Radial variation effects caused by the spanwise variation in relative Mach number were measured. Comparisons to a two-dimensional non-linear unsteady blade/vane Navier-Stokes analysis shows good agreement for the 50% span results in terms of IGV unsteady surface pressure. The results of the study indicate that significant non-linear bow shock influences exist on the IGV trailing edge due to the downstream rotor shock system.

Topics: Pressure , Compressors
Commentary by Dr. Valentin Fuster
1998;():V005T14A026. doi:10.1115/98-GT-356.
FREE TO VIEW

An eight noded mixed shell element for the vibration analysis of rotating cambered helicoidal blades is presented. The strain displacement relations of Gol den’veizer are used and Reissner’s functional is adopted in deriving this element. The effect of initial in-plane stresses is also included. Also, expressions for the pressure difference between the pressure and suction profiles of the rotating blade are derived from two dimensional thin airfoil theories of a stage. The forces thus obtained are applied on the blade to determine the dynamic stresses.

Topics: Stress , Blades , Shells
Commentary by Dr. Valentin Fuster
1998;():V005T14A027. doi:10.1115/98-GT-376.
FREE TO VIEW

Recent technological advancements make hybridization of the magnetic and foil bearings both possible and extremely attractive. Operation of the foil/magnetic bearing takes advantage of the strengths of each individual bearing while minimizing each others weaknesses. In this paper one possible hybrid foil and magnetic bearing arrangement is investigated and sample design and operating parameters are presented. One of the weaknesses of the foil bearings, like any hydrodynamic bearing, is that contact between the foil bearing and the shaft occurs at rest or at very low speeds and it has low load carrying capacity at low speeds. For high speed applications, AMBs are, however, vulnerable to rotor-bending or structural resonances that can easily saturate power amplifiers and make the control system unstable. Since the foil bearing is advantageous for high speed operation with a higher load carrying capacity, and the magnetic bearing is so in low speed range, it is a natural evolution to combine them into a hybrid bearing system thus utilizing the advantages of both.

To take full advantage of the foil and magnetic elements comprising a hybrid bearing, it is imperative that the static and dynamic characteristics of each bearing be understood. This paper describes the development of a new analysis technique that was used to evaluate the performance of a class of gas-lubricated journal bearings. Unlike conventional approaches, the solution of the governing hydrodynamic equations dealing with compressible fluid is coupled with the structural resiliency of the bearing surfaces. The distribution of the fluid film thickness and pressures, as well as the shear stresses in a finite-width journal bearing, are computed. Using the Finite Element (FE) method, the membrane effect of an elastic top foil was evaluated and included in the overall analytical procedure. Influence coefficients were generated to address the elasticity effects of combined top foil and elastic foundation on the hydrodynamics of journal bearings, and were used to expedite the numerical solution. The overall program logic proved to be an efficient technique to deal with the complex structural compliance of various foil bearings. Parametric analysis was conducted to establish tabulated data for use in a hybrid foil/magnetic bearing design analysis. A load sharing control algorithm between the foil and magnetic elements is also discussed.

Topics: Bearings
Commentary by Dr. Valentin Fuster
1998;():V005T14A028. doi:10.1115/98-GT-378.
FREE TO VIEW

This paper describes the rotor dynamic stability analysis and the PTC-10 Class 1 test of a three body centrifugal compressor train for high pressure natural gas injection service. This train had a full load full pressure string test on hydrocarbon gasses to a final discharge pressure of 500 BAR (7250 PSIA). Each compressor is of the back to back configuration, and is equipped with tilting pad seats, damper bearings, and a honeycomb labyrinth at the division wall with shunt holes. The driver is a gas turbine.

Commentary by Dr. Valentin Fuster
1998;():V005T14A029. doi:10.1115/98-GT-379.
FREE TO VIEW

This paper deals with the statistics of the response of a mistuned bladed disk assembly subjected to random excitation. Analytical techniques are developed to compute this statistics for two types of random excitation: white noise and narrow band. The validity of the analytical methods has been established by comparison with the results from numerical simulations. The sensitivities of the response to mistuning have been examined as a function of the width of the frequency band of the random excitation, the dominant frequency of the random excitation and the structural coupling between adjacent blades.

Commentary by Dr. Valentin Fuster
1998;():V005T14A030. doi:10.1115/98-GT-387.
FREE TO VIEW

During creep loading metallic substrates impose deformation on deposited ceramic thermal barrier coatings (TBC). Strain accomodation of the TBC is not attained by plastic deformation, but by means of crack initiation, crack opening, crack propagation or sliding of adjacent crack faces. In technical applications a distinction is made between tolerated or desired cracks perpendicular to the surface, and detrimental cracks parallel to the substrate-coating interface. Thus, TBC can respond to creep deformation by segmentation or spallation, the latter being referred to as failure.

The parameters influencing the probability of either segmentation or spallation are temperature, creep rate, magnitude of creep deformation, layer thickness and microstructure of the TBC. It can be stated that spallation failure probability increases with increasing creep rate, creep deformation and layer thickness. The presence of pores between single spraying layers also strongly augments the likelyhood of spallation. No significant influence of temperature on spallation failure probability can be found in the range from 850°C to 1050°C.

Light microscopy and scanning electron microscopy investigations show that the microstructure of the ceramic TBC changes during creep, and that the density of cracks detected on micrographs with low magnification (x50), increases with increasing creep deformation. On the other hand, the density of microcracks visible with high magnification (x500) is constant, or even decreases with increasing creep deformation. These findings are explained by sintering processes enabled by stress relaxation due to formation of macroscopic cracks perpendicular to the surface as a response to creep deformation.

A relationship between microstructural changes and the emission of acoustic signals recorded during creep is presented.

Commentary by Dr. Valentin Fuster
1998;():V005T14A031. doi:10.1115/98-GT-397.
FREE TO VIEW

The influence of journal’s dynamic angular misalignment effects on single span flexible rotor vibrations is analysed in a simple system. It comprises a cylindrical plain bearing with two narrow lands separated by an annular groove and a Jeffcott based rotor model. Analytical expressions are presented for all 32 dynamic coefficients of the complete linear bearing model, together with simulation results that provide guidelines for early design stage identification of rotor systems sensitive to angular misalignment.

Commentary by Dr. Valentin Fuster
1998;():V005T14A032. doi:10.1115/98-GT-405.
FREE TO VIEW

This paper presents test results from the application of Adaptive Vibration Control (AVC) to a magnetic bearing equipped 6-stage hydrogen process compressor which operates at 20.6 MPa (3000 psi). This application represents the first time AVC has been demonstrated on a fully functional industrial machine. The results demonstrate the reduction in vibration levels possible with AVC operating in each of the two vibration control modes: synchronous position reduction which reduces the synchronous displacement of the shaft as measured at the bearing and synchronous current reduction which reduces the synchronous control effort by the bearings, allowing the shaft to rotate about its inertial center. A discussion is also included on future developments planned for AVC.

Commentary by Dr. Valentin Fuster
1998;():V005T14A033. doi:10.1115/98-GT-406.
FREE TO VIEW

Magnetic bearings, unlike traditional mechanical bearings, consist of a series of components that, when mated together, form a stabilized system. The correct design of the magnetic bearing actuators will provide a cost effective device with low power requirements. The necessity of stability, which a magnetic suspension control system requires in both gain and phase margin, demands the actuators have sufficient bandwidth over the operating range. This paper presents designs for both radial and axial actuators which have attained both high bandwidth and low loss in operation. Test data which verifies the actuator models are presented. Applications of this technology for turbomachinery are described. These include fully magnetically levitated rotor systems, tip clearance control systems and noise and vibration control systems.

Commentary by Dr. Valentin Fuster
1998;():V005T14A034. doi:10.1115/98-GT-407.
FREE TO VIEW

A sliding mode feedback algorithm is proposed to control the vibration of a flexible rotor supported by magnetic bearings. It is assumed that the number of states is greater than the number of sensors. A mathematical model of the rotor/magnetic bearing system is presented in terms of partial differential equations. These equations are then discretized into a finite number of ordinary differential equations through Galerkin’s method. The sliding mode control law is designed to be robust to rotor imbalance and transient disturbances. A boundary layer is introduced around each sliding hyperplane to eliminate the chattering phenomenon. The results from numerical simulations are presented which not only corroborate the validity of the proposed controller, but also show the effects of various control parameters as a function of the angular speed of the rotor. In addition, results are presented that indicate how the current required by the magnetic bearings is affected by control parameters and the angular speed of the rotor.

Commentary by Dr. Valentin Fuster
1998;():V005T14A035. doi:10.1115/98-GT-408.
FREE TO VIEW

An analysis of energy absorption in a tensile test of a CAS/SIC ceramic matrix composite is presented. The analysis is based on the principal parameters related to the strain and fracture mechanisms, as are the pullout length and the matrix crack spacing. The micro-mechanical properties of the material, the interfacial debonding, the frictional stress along the interface and the debonding energy are the core of the analysis. Using these parameters, the energy absorbed during the test, and the elastic energy released after fracture are calculated. The results are compared with those obtained in twenty-one tensile tests at room temperature, in which the strain energy density has been calculated and the permanent strain has been measured.

Commentary by Dr. Valentin Fuster
1998;():V005T14A036. doi:10.1115/98-GT-409.
FREE TO VIEW

Testing and analysis of a profiled leading edge groove tilting pad journal bearing developed for light load operation is described. This bearing was designed for a generic, small, high speed steam turbine operating at projected loads of less than 25 psi (172.4 kPa) and journal surface speeds to 400 ft/s (122 m/s). On the second turbine application, a rotor instability was experienced with the oil flowrate reduced to optimize bearing steady state performance. This instability was eliminated by machining a taper on the exit side of the feed groove on each pad. At the reduced flowrate, the profiled groove bearing greatly improved the operating characteristics of the rotor system by reducing vibration amplitudes and stabilizing operation at speed.

This paper is divided into two sections. The first section compares the rotordynamics analysis with test data that shows improved unbalance response and operating stability with the profiled groove bearing. The second section provides original insight of the effect of the profiled geometry on the bearing flow field using computational fluid dynamics models.

Commentary by Dr. Valentin Fuster
1998;():V005T14A037. doi:10.1115/98-GT-410.
FREE TO VIEW

For the purpose of the efficient analysis of a mistuned bladed disk system, a new analysis method which applies the substructure synthesis method and the modal analysis method is proposed. Using the proposed method, the vibrational characteristics of the grouped blades structure are studied. From the results, it is found that the grouped blades structure is very sensitive to the mistuning. It is also found that the mixed grouped blades structure (a bladed disk system consisting of some different types of grouped blades relating to the number of blades contained) has an undesirable effect on the forced response. Moreover, by comparing the vibrational characteristics of the integral shroud blades (ISB) structure with those of the grouped blades structure, it is clarified that the reliability of the ISB structure is superior to other structures also from the viewpoint of the mistuning.

Topics: Disks , Blades
Commentary by Dr. Valentin Fuster
1998;():V005T14A038. doi:10.1115/98-GT-411.
FREE TO VIEW

This paper deals with the dynamics of an accelerating unbalanced Jeffcott rotor-bearing system mounted on damped, flexible supports. The general equations of motion for such a system are presented and discussed. The rotor response was predicted, via numerical integration, for various cases in runup and rundown conditions and presented in graphical form. The effects of acceleration on the rotor peak amplitude and the speed at which the peak occurs is discussed and compared to steady state predictions.

Topics: Bearings , Rotors
Commentary by Dr. Valentin Fuster
1998;():V005T14A039. doi:10.1115/98-GT-412.
FREE TO VIEW

This paper describes the development of an extremely fast method to obtain the unbalance response of multiple mode rotors supported on Squeeze Film Dampers (SFDs). Planar modal analysis theory is used to model the rotor-SFD system. Undamped critical speed analysis is performed to obtain the rotor eigenvalues and eigenvectors. The SFD nonlinear forces are included in the modal force vector. The system differential equations are constructed for the system and are uncoupled using the orthogonal properties of modal vectors. Assuming circular orbit, consistent with planar modes, the differential equations are converted into algebraic ones. A polynomial in speed is obtained through algebraic manipulations. This polynomial represents the steady state behavior of the rotor-SFD system. The full unbalance response is directly obtained by finding the roots of the polynomial for each particular orbit. This method is extremely fast compared to numerical integration and to iterative methods. The developed method is useful in performing parametric studies and optimum design of SFDs. Twenty five orders of magnitude computer lime savings are reported. Part II of the paper presents parametric studies of an aircraft gas turbine fan rotor supported by an SFD.

Topics: Dampers , Rotors
Commentary by Dr. Valentin Fuster
1998;():V005T14A040. doi:10.1115/98-GT-413.
FREE TO VIEW

This paper is a continuation of Part I, where the advantages of the fast algorithm to obtain the nonlinear response of multi-mode rotors supported on Squeeze Film Dampers (SFDs) are exploited. In Part I, the fast algorithm which relies on planar modal analysis and circular orbits to obtain a polynomial in rotor speed, is described. The advantages of the algorithm showing twenty five orders of magnitude computer time savings are discussed. In this paper, the fast algorithm is used to perform parametric studies on an aircraft gas turbine fan rotor (AGTFR). The parametric studies show the possibility of appropriately locating the SFD, to dampen the rotor modes. In addition, parametric studies are also used to determine the effect of the SFD parameters on the AGTFR unbalance response.

Topics: Dampers , Rotors
Commentary by Dr. Valentin Fuster
1998;():V005T14A041. doi:10.1115/98-GT-484.
FREE TO VIEW

This paper presents important improvements and extensions to a computationally efficient reduced order modeling technique for the vibration analysis of mistuned bladed disks. In particular, this work shows how the existing modeling technique is readily extended to turbomachinery rotors with shrouded blades. The modeling technique employs a component mode synthesis approach to systematically generate a Reduced Order Model (ROM) using component modes calculated from a Finite Element Model (FEM) of the rotor. Based on the total number of degrees of freedom, the ROM is typically two or three orders of magnitude smaller than the FEM. This makes it feasible to predict the forced response statistics of mistuned bladed disks using Monte Carlo simulations. In this work, particular attention is devoted to the introduction of mistuning into the ROM of a shrouded assembly. Mistuning is modeled by projecting the mistuned natural frequencies of a single, cantilever blade with free shrouds onto the harmonic modes of the shrouded blade assembly. Thus, the necessary mistuning information may be measured by testing individual blades.

Commentary by Dr. Valentin Fuster
1998;():V005T14A042. doi:10.1115/98-GT-485.
FREE TO VIEW

In this paper, the 3D shroud contact kinematics of a shrouded blade system is studied. The assumed blade motion has three components, namely axial, tangential, and radial components, which result in a three dimensional relative motion across the shroud interface. The resulting relative motion can be decomposed into two components. The first one is on the contact plane and can induce stick-slip friction. The other component is perpendicular to the contact plane and can cause variation of the contact normal load and, in extreme circumstances, separation of the two contacting surfaces. In order to estimate the equivalent stiffness and damping of the shroud contact an approach is proposed. In this approach, the in-plane slip motion is assumed to be elliptical and is decomposed into two linear motions along the principal major and minor axes of the ellipse. A variable normal load friction force model (Yang and Menq, 1996) is then applied separately to each individual linear motion, and the equivalent stiffness and damping of the shroud contact can be approximately estimated. With the estimated stiffness and damping, the developed shroud contact model is applied to the prediction of the resonant response of a shrouded blade system. The effects of two different shroud constraint conditions, namely 2D constraint and 3D constraint, on the resonant response of a shrouded blade system are compared and the results are discussed.

Topics: Resonance , Blades
Commentary by Dr. Valentin Fuster
1998;():V005T14A043. doi:10.1115/98-GT-490.
FREE TO VIEW

This paper presents a new International Standard Configuration to be added to an already existing set of 10 configurations for unsteady flow through vibrating axial-flow turbomachine cascades. This 11th configuration represents a turbine blade geometry with transonic design flow conditions with a normal shock positioned at 75% real chord on the suction side. Out of a set of test cases covering all relevant flow regimes two cases were selected for publication: A subsonic, attached flow case and an off-design transonic case showing a separation bubble at 30% real chord on the suction side are published. The performed tests are shown to be repeatable and suitable for code validations of numerical models predicting flutter in viscous flows.

The validity of the measured data of the two public cases was examined and comparisons with other tests were conducted. Sometimes a large difference in aerodynamic damping was observed on cases with similar flow conditions. This was investigated at three transonic cases with almost identical inlet flow conditions and only small variations in outlet Mach Number. It was found that the differences in the global damping are due to very local changes on the blade surface in the shock region, which obtain a large influence by the integration because of the discrete measuring points. Hence it is recommended not to look at the global damping for code validations but more precisely to the local values. These show a common tendency, which is reproducible with different numerical methods.

This was demonstrated with a potential model, a linear Euler model, a nonlinear Euler model and a Navier-Stokes solver, all applied to predict flutter of each test case with a 2D/Q3D approach. The limitations of inviscid codes to predict flutter in viscous flow regimes is demonstrated, but also their cost advantage in attached flow calculations. The need of viscous code development and validation is pointed out. This should justify and encourage the publication of thoroughly measured test cases with viscous effects.

Commentary by Dr. Valentin Fuster
1998;():V005T14A044. doi:10.1115/98-GT-505.
FREE TO VIEW

This paper will discuss a study of an innovative design for an advanced turbine rotor, that could have a great impact on future engines. The design challenge is to provide a minimum weight turbine rotor system that can withstand beyond state-of-the-art levels of AN2 (turbine annulus area multiplied by speed squared). An AN2 limit has been reached for High Pressure Turbine (HPT) disks configured in conventional (single web) geometry with state-of-the-art nickel alloys. The problem has reached the point where increased AN2 has been declared a “Break-Through” technology. The twin-web disk has the potential to provide this break through. This paper will present the history of this turbine rotor design, analytical results, material/component processing and concept validation results. All work was performed under an Air Force sponsored program entitled Composite Ring Reinforced Turbine (CRRT).

Topics: Disks
Commentary by Dr. Valentin Fuster
1998;():V005T14A045. doi:10.1115/98-GT-513.
FREE TO VIEW

This paper illustrates the theory behind the development of a General form of a numerically accurate and stable software for determining the lateral natural frequencies, mode shapes, stability and unbalance response of multi-shaft rotor bearing systems which may be gear coupled and supported on flexible anisotropic bearings, such as hydrodynamic bearings as well as rigid rolling elements. Gyroscopic effects, shear deformation and hysteretic damping are included in the analysis. To achieve the improved accuracy and numerical convergence to all relevant roots of the characteristic equations the Riccati transfer matrix is utilized rather than the traditional transfer matrix approach. This more accurate approach introduces poles into the characteristic equation. After cancellation of all poles, convergence to all relevant roots is guaranteed. In this study two different methods of handling in-span conditions are presented and it is proven that after elimination of poles the Riccati transfer matrix method and the transfer matrix method have the same characteristic equation. However with even 25% increase in average computation time, the Riccati transfer matrix method remains more accurate due to a different method of calculating the characteristic equation.

Commentary by Dr. Valentin Fuster
1998;():V005T14A046. doi:10.1115/98-GT-514.
FREE TO VIEW

High cycle fatigue (HCF) is a widespread technical problem in aircraft gas turbines, and can lead to the premature failure of plumbing, pumps, accessory drives and airfoils. Roughly 35% of the problems associated with newer production engines are the direct result of HCF. This paper discusses three innovative technologies that can assist in the avoidance or elimination of high cycle fatigue failures in aircraft gas turbines. This work was performed under the sponsorship and direction of the Defense Advanced Research Projects Agency (DARPA) and the United States Air Force Research Laboratory.

Topics: Gas turbines
Commentary by Dr. Valentin Fuster
1998;():V005T14A047. doi:10.1115/98-GT-548.
FREE TO VIEW

In this paper, the concept of constrained mode shapes is employed to predict the resonant response of a frictionally constrained blade system. For a tuned blade system, the constrained mode shapes can be calculated using a finite element model of a single blade along with the cyclic symmetry constraint that simulates a fully stuck friction contact. The resulting constrained mode shapes are often complex and can be used to obtain the constrained receptance of the frictionally constrained blade. It is shown that by examining each mode’s contribution to the receptance at the friction contact point, the importance of each individual modes to the prediction of the resonant response of a frictionally constrained blade can be determined. Furthermore, by comparing the receptances calculated from free mode shapes and those from constrained mode shapes, it is found that in the neighborhood of the fully slipping region, the prediction of resonant response requires fewer number of modes when using free mode shapes compared to using constrained mode shapes. On the other hand, in the neighborhood of the fully stuck region, it requires fewer number of modes if constrained mode shapes are used. Therefore, when high preload at the friction contact is desirable, such as for shrouded blade systems, using the constrained mode shapes for the prediction of resonant response is preferred. Moreover, the concept of hybrid receptance is introduced so as to yield very accurate prediction of the resonant response based on only very few vibration modes.

Commentary by Dr. Valentin Fuster
1998;():V005T14A048. doi:10.1115/98-GT-570.
FREE TO VIEW

The effect of the coupling of the torsion mode blade vibration to the bending mode flutter in transonic fans has been studied by quasi-3D viscous unsteady calculations. The type of flutter in this research is that of a highly loaded blade with a tip relative Mach number just above unity, for which the mechanism is that the instability of the passage shock wave when it unstarts generates the dominant blade exciting aerodynamic work at its foot on the pressure surface of the blade. The dependence of such flutter on the blade vibration mode, i.e. the amplitude ratio and the phase difference of the bending and torsional components has been studied.

The study showed that the blade exciting aerodynamic work reduced when the torsional component is added in-phase (torsional motion noses up during the upward bending motion) to the bending oscillation. The amplitude of the torsional component was shown to have an optimum amplitude.

It was also shown that this tendency would switch when the shock wave is fully detached, and the blade exciting aerodynamic work would increase by adding the torsional component in-phase to the bending oscillation.

Commentary by Dr. Valentin Fuster
1998;():V005T14A049. doi:10.1115/98-GT-571.
FREE TO VIEW

The results of two investigations, conducted on the aerodynamic response of a turbine blade oscillating in a three dimensional bending mode, are presented in this paper.

The first is an experimental and computational study, designed to produce detailed three dimensional test cases for aeroelastic applications and examine the ability of a 3D time-marching Euler method to predict the relevant unsteady aerodynamics. Extensive blade surface unsteady pressure measurements were obtained for a range of reduced frequency, from a test facility with clearly defined boundary conditions, Bell & He (1997). The test data exhibits a significant three dimensional effect, whereby the amplitude of the unsteady pressure response at different spanwise positions is largely insensitive to the local bending amplitude. The inviscid numerical scheme successfully captured this behaviour, and a good qualitative and quantitative agreement with the test data was achieved for the full range of reduced frequency. In addition, the issue of linearity is addressed and both experimental and numerical tests demonstrate a linear behaviour of the unsteady aerodynamics.

The second, an experimental investigation, considers the influence of tip leakage on the unsteady pressure response of an oscillating turbine blade. Results are provided for three tip clearances. The steady flow measurements show marked increases in the size and strength of the tip leakage vortex for the larger tip gaps and deviations in the blade loading towards the tip section. The changes in tip gap also caused distinct trends in the amplitude of the unsteady pressure at 90% span, which were consistent with those observed for steady flow blade loading. It is the authors opinion, that the existence of these trends in unsteady pressure warrants further investigation into the influence of tip leakage upon the local unsteady flow and aerodynamic damping.

Commentary by Dr. Valentin Fuster
1998;():V005T14A050. doi:10.1115/98-GT-572.
FREE TO VIEW

The present paper shows the comparison between steady-state and unsteady results from two versions of a flow solver and measurements in a linear cascade at transonic flow conditions for forced vibration of a single blade.

Experiments have been conducted in a linear test facility with a cascade composed of five turbine blades. The center blade was forced to oscillate in a pure bending mode with a frequency of 160 Hz. Tailboards are fixed in this test facility at the trailing edges of the border blades to create a periodic flow at the outlet of the cascade.

A 2D inviscid flow solver for unsteady flows through vibrating blade rows was adapted to simulate the flow conditions in the linear cascade.

The initial adaptation of the flow solver was the limitation to only one vibrating blade. Comparisons with measurements showed discrepancies which were hypothesized to be due to the presence of the tailboards which can create reflections of pressure waves and shocks.

Therefore, as a second modification of the program, the periodic boundaries downstream of the blades were replaced by solid walls.

Based on these computations it can be concluded that for transonic flows the tailboards have a significant influence on both the steady-state and unsteady flow in the outlet region as well as on the blade in the shock region. In particular, the unsteady data measured in a linear cascade has to be analyzed very carefully to estimate the influence of the tailboards.

Commentary by Dr. Valentin Fuster
1998;():V005T14A051. doi:10.1115/98-GT-573.
FREE TO VIEW

The work described in this paper is part of a comprehensive research effort aimed at eliminating the occurrence of low pressure turbine blade flutter in aircraft engines. The results of fundamental unsteady aerodynamic experiments conducted in an annular cascade are studied in order to improve the overall understanding of the flutter mechanism and to identify the key flutter parameters. In addition to the standard traveling wave tests, several other unique experiments are described. The influence coefficient technique is experimentally verified for this class of blades. The beneficial stabilizing effect of mistuning is also directly demonstrated. Finally, the key design parameters for flutter in low pressure turbine blades are identified. In addition to the experimental effort, correlating analyses utilizing linearized Euler methods demonstrate that these computational techniques are adequate to predict turbine flutter.

Commentary by Dr. Valentin Fuster
1998;():V005T14A052. doi:10.1115/98-GT-575.
FREE TO VIEW

A design approach to avoid flutter of low pressure turbine blades in aircraft engines is described. A linearized Euler analysis, previously validated using experimental data, is used for a series of parameter studies. The influence of mode shape and reduced frequency are investigated. Mode shape is identified as the most important contributor to determining the stability of a blade design. A new stability parameter is introduced to gain additional insight into the key contributors to flutter. This stability parameter is derived from the influence coefficient representation of the cascade, and includes only contributions from the reference blade and its immediate neighbors. This has the effect of retaining the most important contributions to aerodynamic damping while filtering out terms of less significance. This parameter is utilized to develop a stability map, which provides the critical reduced frequency as a function of torsion axis location. Rules for preliminary design and procedures for detailed design analysis are defined.

Commentary by Dr. Valentin Fuster
1998;():V005T14A053. doi:10.1115/98-GT-580.
FREE TO VIEW

The capabilities of an inviscid quasi three-dimensional linearized unstructured flow solver to correctly predict the stall flutter limit, flutter modes and critical inter-blade phase angles on a transonic rotating shroudless fan model where experimental data exist have been investigated. Three operating points were chosen for investigation at 70% and 95% speed. At 70% speed two points were investigated: one close to the torsional flutter boundary (at the intermediate operating line) and one at the flutter boundary. The 95% speed point was at the flexural flutter boundary. Steady state and unsteady calculations were made at several stream sections per operating point. At each stream section unsteady calculations were performed over the entire range of inter-blade phase angles with different mode shapes (real mode, rigid torsion and rigid bending) at different frequencies. Thus the model was “provoked” with “unphysical” mode shapes and frequencies to be compared to the unsteady solution obtained with the mode shapes and frequencies observed from the experiments. Furthermore all unsteady calculations were made with different mesh densities and solutions from different “tuned” and “untuned” steady-state solutions.

The main conclusion of the validation of the inviscid Q3D Euler model on the Fan C Model Rotating Rig is that the model generally predicts flutter, flutter modes and the critical inter-blade phase angles to be close to the experimentally determined ones.

Commentary by Dr. Valentin Fuster
1998;():V005T14A054. doi:10.1115/98-GT-583.
FREE TO VIEW

The focus of the present investigation is on the estimation of the dynamic properties, i.e. masses, stiffnesses, natural frequencies, mode shapes and their statistical distributions, of turbomachine blades to be used in the accurate prediction of the forced response of mistuned bladed disks. As input to this process, it is assumed that the lowest natural frequencies of the blades alone have been experimentally measured, for example in a broach block test.

Since the number of measurements is always less than the number of unknowns, this problem is indeterminate in nature. Two distinct approaches will be investigated to resolve the shortfall of data. The first one relies on the imposition of as many constraints as needed to insure a unique solution to this identification problem. Specifically, the mode shapes and modal masses of the blades are set to their design/tuned counterparts while the modal stiffnesses are varied from blade-to-blade to match the measured natural frequencies. The second approach, based on the maximum likelihood principle, yields estimates of all the structural parameters of the blades through the minimization of a specified “cost function”. The accuracy of these two techniques in predicting the forced response of mistuned bladed disks will be assessed on simple dynamic models of the blades.

Topics: Disks
Commentary by Dr. Valentin Fuster
1998;():V005T14A055. doi:10.1115/98-GT-586.
FREE TO VIEW

Description: Stress analysis of a GT engine fan shaft was performed for IN718 and Ti. A FE package NISA gave stiffness for unit radial and tangential loads. Results obtained with IN718 were compared with an earlier study by SNECMA. An extension of the study for weight reduction was done with Ti material parameters. The design was adapted to simplify a spring bearing mount for the HP (#3) bearing.

A convoluted spring bearing mount suggested by SNECMA was redesigned using the results for the conical fan shaft to give a more gradual stress distribution. Such a structure would be less likely to yield in LCF and catastrophic failure modes. Comparison between computed results and tabulated results for conical shells was attempted.

Results depended on loading and point of application and mounting. For instance, the fan shaft was fixed on the main rotor and a force applied to the rim and vice versa. Variation in the number of nodes and order of approximation also affected the results. Regions of maximum stress were shown in contour plots, while displacement studies gave modes of deflection.

Commentary by Dr. Valentin Fuster

Controls, Diagnostics and Instrumentation

1998;():V005T15A001. doi:10.1115/98-GT-003.
FREE TO VIEW

This paper addresses the motivations for using a distributed control system architecture, technical challenges, typical functions which are off-loaded to remote terminals, sensor/effector interface issues, data bus selection, technology insertion issues, lessons learned, and objectives for future distributed control implementations. Typical design requirements, constraints, environmental conditions, and operational challenges will be described. Examples of various distributed control system implementations will be discussed, including both propulsion control and flight control examples.

Topics: Propulsion , Flight
Commentary by Dr. Valentin Fuster
1998;():V005T15A002. doi:10.1115/98-GT-016.
FREE TO VIEW

A Distributed Control System (DCS) for a turbine engine has been demonstrated and tested, consisting of prototype Electronic Interface Units (EIUs) connected to data and power busses. In the DCS, a central control computer communicated with smart sensors and smart actuators via a 2.5 megabit/sec digital data bus, using the Fieldbus protocol. Power was distributed to the smart devices as 100-kHz 100V peak AC, allowing light, simple power converters at each smart device. All smart sensors, smart actuators and cables were dual-redundant. The smart actuators received position demand from the central control computer, exchanged data between channels to provide local redundancy management, closed the position loop locally, and reported actuator position to the central controller. Smart sensors converted sensed signals to digital values in engineering units, and performed local built-in tests. Testing of the DCS was done in a closed-loop simulation with an engine model. Frequency response of the DCS was almost identical with the conventional system.

Commentary by Dr. Valentin Fuster
1998;():V005T15A003. doi:10.1115/98-GT-044.
FREE TO VIEW

The application of distributed control systems to turbine engine controls offers the potential for major reductions in development time and costs for the engine control and the engine. Once the data bus and power bus are standardized for elements of a distributed control system, the industry will have a group of sensors, actuators, and controllers that could be interchangeable between applications. Software and hardware will still require modification to fit the specific application, however, great strides will have been made toward a “plug and play” capability between sensors, actuators, and controllers all tied together on the same data bus. The main controller in a distributed control system, except for software, would be interchangeable from engine to engine.

This paper describes the design and development of the electronics for a smart actuator and discusses the design considerations which were used to guide the requirements. Requirements unique to turbine engine applications include temperature environments to 30° C, a severe vibration environment, minimum size and weight, and very high reliability.

The electronics developed for the smart actuator were packaged on credit card sized printed wiring board modules. Two of these modules were packaged in a housing approximately 23×3.4×1.1 inches. The electronics operate from 28 volt DC power and communicate with the rest of the control system via the MEL-STD-1553B data bus. Although a hydraulic actuator was chosen as the demonstration vehicle, the electronic module is adaptable to any servo application and can be expanded to read any of the common engine sensors and operate solenoids.

The chosen actuator was intended as a development tool to expose the design problems of distributed systems. Therefore, this first demonstration unit was designed using electronic components rated for 125° C operation. AlliedSignal is currently a member of a consortium of companies under DARPA sponsorship developing a family of SOI (silicon-on-insulator) integrated circuits rated for 200° C operation. Our current 125° C design is compatible with the new devices being developed. A 200° C unit is planned for 1998. Further improvements in the metalization used in the SOI devices will allow reliable long term operation to about 300° C. Devices for this higher temperature range are expected to be available in 1999.

Commentary by Dr. Valentin Fuster
1998;():V005T15A004. doi:10.1115/98-GT-045.
FREE TO VIEW

The increasing use of embedded intelligence to produce smart sensors and actuators offers great potential benefits in adopting a distributed control strategy in aerospace applications. There are many advantages to be gained: fewer and shorter buses, intrinsic partitioning, smaller control box size, increased health monitoring, increased system flexibility and reduced vulnerability to hazardous events. However, this has to be traded off against the problems of greater complexity, processor diversity, accessibility, exposure of electronics to severe environments, power distribution, software production costs and management. In this paper how multi-objective optimisation can be applied to this problem for the application of military gas turbine engines is considered.

Commentary by Dr. Valentin Fuster
1998;():V005T15A005. doi:10.1115/98-GT-061.
FREE TO VIEW

The increasing numbers of commercial decision support tools based on Case-Based Reasoning (CBR) technology reflects the interest in the use of this technology to aid fault diagnosis, not only in gas turbine applications but also in the wider engineering domain. CASSIOPEE from CFM and ARIADNE from British Airways are but two examples of these types of support tool in the civil gas turbine engine arena. They reflect the trend of developing such systems from a historical database of maintenance arisings and necessarily apply to older engine designs. Further characteristics of which include older and less sophisticated aircraft maintenance systems, which provide limited diagnostic information and rely more heavily upon the experience of maintenance personnel to isolate engine faults successfully. It is this diagnostic experience which these CBR tools have attempted to capture and emulate; the case for a CBR-based tool in this instance is well proven.

More recent aircraft design have resulted in a more comprehensive on-board maintenance systems to aid maintenance personnel diagnose airframe and engine faults. Such an approach should rely less heavily upon the experience of the maintenance crew to isolate a fault successfully. The paper reports on the design and development of two CBR-based maintenance decision support tools for the Rolls-Royce Trent 800 and RB211-524 engines. One system is currently undergoing a trial to refine the design of the maintenance case base (the database over which the CBR engine searches to match the current fault symptoms to a historical event). The lessons learned from this field trial and future prospects for such a CBR-based maintenance aid are discussed.

Topics: Design , Gas turbines
Commentary by Dr. Valentin Fuster
1998;():V005T15A006. doi:10.1115/98-GT-098.
FREE TO VIEW

The frequency-domain identification of gas turbine dynamics is discussed. Models are directly estimated from engine data and used to validate linearised thermodynamic models derived from the engine physics. This work is motivated by the problems previously encountered when using time-domain methods. A brief overview of frequency-domain techniques is presented and the design of appropriate multisine test signals is discussed. Practical results are presented for the modelling of the fuel feed to shaft speed dynamics of a twin-spool engine. The gathered data are analysed and the frequency response functions of the engine are estimated. The identification of parametric s-domain models is discussed in detail and a comparison made between the identified models and the linearised thermodynamic models. The influence of engine nonlinearities on the linear models is also examined.

Commentary by Dr. Valentin Fuster
1998;():V005T15A007. doi:10.1115/98-GT-099.
FREE TO VIEW

System identification plays an important role in advanced control systems for jet engines, in which controls are performed adaptively using data from the actual engine and the identified engine. An identification technique for jet engine using the Constant Gain Extended Kalman Filter (CGEKF) is described. The filter is constructed for a two-spool turbofan engine. The CGEKF filter developed here can recognize parameter change in engine components and estimate unmeasurable variables over whole flight conditions. These capabilities are useful for an advanced Full Authority Digital Electric Control (FADEC). Effects of measurement noise and bias, effects of operating point and unpredicted performance change are discussed. Some experimental results using the actual engine are shown to evaluate the effectiveness of CGEKF filter.

Topics: Jet engines
Commentary by Dr. Valentin Fuster
1998;():V005T15A008. doi:10.1115/98-GT-100.
FREE TO VIEW

The feasibility of Model Predictive Control (MPC) applied to a laboratory gas turbine installation is investigated. MPC explicitly incorporates (input- and output-) constraints in its optimizations, which explains the choice for this computationally demanding control strategy. Strong nonlinearities, displayed by the gas turbine installation, cannot always be handled adequately by standard linear MPC. Therefore, we resort to nonlinear methods, based on successive linearization and nonlinear prediction as well as the combination of these. We implement these methods, using a nonlinear model of the installation, and compare them to linear MPC. It is shown that controller performance can be improved, without increasing controller execution-time excessively.

Commentary by Dr. Valentin Fuster
1998;():V005T15A009. doi:10.1115/98-GT-101.
FREE TO VIEW

Condition monitoring of engine gas generators plays an essential role in airline fleet management. Adaptive diagnostic systems are becoming available that interpret measured data, furnish diagnosis of problems, provide a prognosis of engine health for planning purposes, and rank engines for scheduled maintenance. More than four hundred operations worldwide currently use versions of the first or second generation diagnostic tools.

Development of a third generation system is underway which will provide additional system enhancements and combine the functions of the existing tools. Proposed enhancements include the use of artificial intelligence to automate, improve the quality of the analysis, provide timely alerts, and the use of an Internet link for collaboration. One objective of these enhancements is to have the intelligent system do more of the analysis and decision making, while continuing to support the depth of analysis currently available at experienced operations.

This paper presents recent developments in technology and strategies in engine condition monitoring including:

1) application of statistical analysis and artificial neural network filters to improve data quality;

2) neural networks for trend change detection, and classification to diagnose performance change; and

3) expert systems to diagnose, provide alerts and to rank maintenance action recommendations.

Commentary by Dr. Valentin Fuster
1998;():V005T15A010. doi:10.1115/98-GT-129.
FREE TO VIEW

The condition monitoring section at the Brazilian power utility CEMIG is implementing an effective condition-based maintenance strategy that ensures the over 40 power plants spread out over a large area operate with minimal downtime and at a minimal maintenance cost. The condition monitoring system needed to fulfil CEMIG’s needs for the larger plants did not exist, so it was decided to integrate several monitoring systems for this purpose. A computerized, permanently installed vibration monitoring system is planned to be integrated to other systems dedicated to specific periodic machine condition monitoring applications (e.g. air gap monitoring, oil analysis, magnetic flux monitoring, partial discharge analysis). This integrated monitoring approach results in a distributed system with a single system technique for alarm handling, and a user interface and database for analysis, diagnosis and fault correlation. The vibration monitoring system will also be extended for importing process data from the existing distributed supervisory and control system for monitoring calculated performance parameters such as efficiency and head. Testing is also under way for investigating the possibility of more effectively monitoring cavitation without purchasing a separate stand-alone system. Several of the larger plants at CEMIG will eventually be remotely monitored this way, but this paper focuses primarily on the monitoring system, strategy and current operating experience at the Nova Ponte hydroelectric power station. Even before integrating the other monitoring systems, the installed condition monitoring section played a large role in ensuring the plant operates safely, cost effectively and with maximum availability. Although the monitoring system is installed at a hydro-electric power station, some examples are briefly given on how the same integrated monitoring system approach could equally be advantageous in detecting and/or diagnosing certain faults within gas turbines and compressors.

Commentary by Dr. Valentin Fuster
1998;():V005T15A011. doi:10.1115/98-GT-155.
FREE TO VIEW

Fault Detection and Accommodation Logic easily occupies at least fifty percent of the real time embedded software application program in the contemporary Full Authority Digital Electronic Engine Control (FADEC). Indeed, if a layman were to look at the logic distribution in a FADEC he might think the main purpose of the FADEC was to detect failures! This, of course, is not the case. It is true, however, that real world applications must be programmed to deal with the inevitable fact that system failures will occur. The success of the application program then depends not only on how well it controls the engine, but how well it detects, isolates and accommodates failures. In other words, the fault tolerance of an application is just as important as the robustness of the control algorithms.

Over the past two decades many basic Fault Detection and Accommodation (FDA) techniques have been developed and have become commonplace in FADECs. These techniques have been developed in such a way as to maximize an application’s Built In Test (BIT) effectiveness for both “hard”, “soft”, and intermittent failures, while minimizing failure annunciation due to “false alarm” events. These techniques have been implemented in both hardware and software.

This paper will present and discuss these basic FDA techniques for input signals, output signals, processors, memory devices and communication devices. Many of the newer FDA techniques will also be discussed. An overview of the key components of FDA, and the established “norms” for effective FDA will be added for completeness. Since the structure of FDA is dependent upon system architecture, the system under consideration is chosen as the Dual-Channel Full Authority Digital Electronic Engine Control. This is the configuration that is most commonly used in commercial and military engine control applications.

Commentary by Dr. Valentin Fuster
1998;():V005T15A012. doi:10.1115/98-GT-156.
FREE TO VIEW

In increasing the availability and reducing the maintenance costs of gasturbines in critical applications, there has been an increasing trend towards Condition Monitoring, Diagnostic and Decision Support systems (CMD&D). The idea is to reduce the number of routine preventative maintenance procedures, in favor of techniques that continuously monitor the health and diagnose faults as they develop. Early isolation of the root cause of problems, allows the operators to better plan the corrective maintenance actions, taking into account the gasturbine and the other units with which it collaborates.

Clearly, this requires expertise in many fields such as Data Acquisition, Trending, Aerothermodynamic Performance, Vibration, Emissions, Fault Isolation, etc. In addition, it is increasingly becoming expensive to have experts available at all locations, so the solution obviously lies in (Semi-)autonomous systems. Their task would be to monitor the health and diagnose symptoms of machinery, perhaps in collaboration with junior experts locally and senior experts at a remote center.

This report outlines some of the results of a short but broad technical exploration in the questions of why none of the existing packages have achieved widespread acceptance, and how a turbine control manufacturer can play a significant role. A conceptual solution is presented which can meet the challenging functionality, data abstraction, automation and communication requirements by integration of expertise and modules from different groups. However, details on how to turn the concept into a product have been puposely ommitted in order to safeguard proprietory activity.

Commentary by Dr. Valentin Fuster
1998;():V005T15A013. doi:10.1115/98-GT-157.
FREE TO VIEW

This paper presents some problems inherent in the application of a laser two-focus anemometry technique to measurement in high-speed small-scale compressors. Improvements in data processing are described, in relation to reducing the acquisition time. Measurement uncertainties and their possible estimation are discussed. An optimized blade pitch partition during synchronized measurements within rotating blade passages is presented using a prediction of shadow zones. Two examples highlight the resulting benefits both in terms of reduction of the acquisition time and improvement of the azimuthal resolution.

Commentary by Dr. Valentin Fuster
1998;():V005T15A014. doi:10.1115/98-GT-158.
FREE TO VIEW

In order to prevent machine malfunctions and to determine the machine operating state, it is necessary to use correct measurements from actual system inputs and outputs. This requires the use of techniques for the detection and isolation of sensor faults.

In this paper an approach based on analytical redundancy which uses dynamic observers is suggested to solve the sensor fault detection and isolation problem for a single-shaft industrial gas turbine. The proposed technique requires the generation of classical residual functions obtained with different observer configurations. The diagnosis is performed by checking fluctuations of these residuals caused by faults.

Commentary by Dr. Valentin Fuster
1998;():V005T15A015. doi:10.1115/98-GT-168.
FREE TO VIEW

Engine overhaul shops need a reliable analytical methodology to pinpoint the cause(s) of engine test-cell under-performance to aid the overhaul decision-making process. Gas path analysis codes have been somewhat successful, but have not been entirely satisfactory. Previous works [Doel, 1994] have raised the idea that if other information could be integrated with the gas path analysis results, it may be possible to achieve better results.

This paper presents a diagnostic system developed for the CF6 family of engines. The system integrates test cell measurements and the gas path analysis program results with information regarding engine operational history, build-up workscope, and direct physical observations in a Bayesian belief network. The paper lays out the nature of the problem and the system requirements and design.

The system produces a diagnosis while following a cost-effective diagnostic process using value of information calculations. This is illustrated through sample cases.

Commentary by Dr. Valentin Fuster
1998;():V005T15A016. doi:10.1115/98-GT-169.
FREE TO VIEW

In the turbomachinery field, many diagnostic systems utilize databases with symptoms corresponding to the most frequent operation faults. Thanks to Computational Fluid Dynamics (CFD), databases can be created without costly experiments, whereas the use of unstructured grids in combination with parallel processing makes the whole task easy and fast to accomplish. In this paper, a procedure that builds up a database for gas-turbine fault diagnosis is demonstrated. Advanced CFD tools that operate concurrently on multi-processor platforms are used. The so-prepared database contributes to the identification of faults through the analysis of the unsteady pressure signals that correspond to hypothetical sensors located in the inter-blade region. The pressure signals are post-processed in a similar way to the one experimentalists employ for fast-response pressure measurements. Symptoms related to displaced and/or twisted blades in an industrial high-speed compressor cascade, at design and off-design operating conditions, are analyzed.

Commentary by Dr. Valentin Fuster
1998;():V005T15A017. doi:10.1115/98-GT-170.
FREE TO VIEW

One of the main requirements for modern FADEC systems is to implement great computing power with many interfaces and to keep the FADEC hardware effort to a minimum. On the other side the criticality potential of computer failures is considered as ‘hazardous’. The trend in FADEC development is to implement even more complex functions into the control software which consequently increases the authority and therefore the criticality potential of computer failures. In the mid 80’s a double computer system was used to performed a parallel execution of the control software with identical input parameters to output identical results. A difference in any one of these computer results causes the comparator hardware to output a failure indication. This was considered to have a 100% coverage of computer failures. The problem with this system was certainly the relatively large hardware overhead and the limited intelligence of the comparator logic. Some other FADEC systems have implemented only a Watch Dog Timer and Bus Access Supervisory hardware to detect computer malfunctions. With this method the proof for the achievements of the safety requirements have become almost impossible since adequate fault models of the computer components are difficult to establish due to their increasing functional complexity. This paper describes how to develop the safety features for the Computer Design from the Engine Control System Safety Requirements to achieve a full coverage of the potentially critical failure effects with fault tolerant failure recovery functions and a minimum of hardware overhead.

Topics: Safety , Computers
Commentary by Dr. Valentin Fuster
1998;():V005T15A018. doi:10.1115/98-GT-171.
FREE TO VIEW

The interest of the world electricity market in the gas turbine and the gas-steam combined cycles has given a significant stimulus to the development of numerical simulations, useful in the detailed analyses of dynamic behaviour and off-design performance of the gas plant. This paper describes the latest improvements of a physical simulator built to predict the off-design and dynamic behaviour of single shaft, heavy-duty gas turbine plants. The analysis has been applied to the V94.2 and V64.3 models of the Ansaldo Energia production, that feature a compressor equipped with variable inlet guide vanes and an air bleed system cooled turbine. The mathematical model is non linear, based on the lumped parameter approach and described by a set of differential and algebraic equations. In particular, the next version presents a turbine model entirely rebuilt, based on a simplified, one-dimensional row by row calculation method. The equations system and the structure of the new code are here described, and the steady-state results of the dynamic model are compared with the results of an Ansaldo static code with a good correspond in the performance prediction. The model has been developed in the Matlab-Simulink® environment using some FORTRAN subroutines.

Commentary by Dr. Valentin Fuster
1998;():V005T15A019. doi:10.1115/98-GT-172.
FREE TO VIEW

A non linear dynamic model has been developed in order to simulate the dynamic behavior of single-shaft or multi-shaft regenerative gas turbines. The aim of the work is to provide a fast and reliable model for the synthesis of the controllers and the study of critical dynamic situations. Some significant features characterize the proposed model:

- an efficient one-dimensional model is used in order to properly model a stationary counter-flow regenerator;

- a stage by stage model is provided for the air cooled turbine expansion in order to take into account the blade thermal transients.

The mathematical model and the numerical methodology are described in the paper.

A single-shaft regenerative cycle gas turbine is analyzed. In order to design a multivariable controller for this plant, a linearized model is developed from the non-linear model. The gas turbine is described by the transfer functions that relate the input variables (fuel rate and variable inlet guide vanes) to the state variables (shaft speed and turbine output temperature). Accuracy and effects of non-linearities are described.

Commentary by Dr. Valentin Fuster
1998;():V005T15A020. doi:10.1115/98-GT-183.
FREE TO VIEW

This paper describes two dual-laser probe integrated fiber optic systems for measuring blade tip clearance in rotating turbomachinery. The probes are nearly flush with the casing inner lining resulting in minimal flow disturbance. The two probes are closely spaced in a circumferential plane and are slanted at an angle relative to each other so that the blade tip traverse time of the space between the two laser beams varies with the tip radius allowing determination of the tip clearance at the rotor operating conditions. The tip clearance can be obtained for all the blades in a rotor with a single system, provided there are no synchronous vibrations present at a particular operating condition. These probes were installed in two holders; one provided an included angle between the probes of 20 degrees, and the other provided an included angle of 40 degrees. The two configurations were calibrated in a vacuum spin rig facility that is capable of reproducing realistic blade tip speeds.

Commentary by Dr. Valentin Fuster
1998;():V005T15A021. doi:10.1115/98-GT-224.
FREE TO VIEW

The heat flux measurement is one of the essential parameter for the diagnostic of thermal systems. In the high temperature environment there are difficulties in differentiating between the convective and radiation component of heat flux on the heat transfer surface. A new method for heat flux measurement is being developed using a porous sensing element. The gas stream flowing through the porous element is used to measure the heat received by the sensor surface exposed to the hot gas environment and to control whether or not the sensing element receives the convection component of the total heat flux. It is possible to define a critical mass flow rate corresponding to the destruction of the boundary layer over the sensing element. With subcritical mass flow rate the porous sensing element will receive both the convective and radiative heat fluxes. A supercritical mass flow rate will eliminate the convective component of the total heat flux. Two consecutive measurements considering respectively a critical and a sub-critical mass flow rate can be used to determine separately the convection and radiation heat fluxes.

A numerical model of sensor with appropriate boundary condition has been developed in order to perform analysis of possible options in the design of the sensor. The analysis includes: geometry of element, physical parameters of gas and solid and gas flow rate through the porous element.

For the optimal selection of the relevant parameters an experimental set-up was designed, including the sensor element with corresponding cooling and monitoring system and high temperature radiation source. Applying the respective measuring procedure the calibration curve of the sensor was obtained. The linear dependency of the heat flux and respective temperature difference of the gas was verified. The accuracy analysis of the sensor reading has proved high linearity of the calibration curve and accuracy of ± 5%.

Commentary by Dr. Valentin Fuster
1998;():V005T15A022. doi:10.1115/98-GT-234.
FREE TO VIEW

The subject of this paper is the development and first analysis of the performance of a new booster bleed valve control logic to improve the surge margin of the booster compressor in the new BR715 turbofan.

The booster compressor is driven by the LP spool of the two spool engine, comprising an aerodynamically tuned system out of fan root and booster compressor. The flow delivered at the exit of the booster compressor must be accepted by the following HP compressor. This matching is not always possible, since the rates of change of the spool speeds during transients behave differently as do pressure rise and mass flow. Hence the transient handling of the booster compressor is difficult with respect to its surge stability. To aid in this situation it is a common practice to position a booster bleed valve after the booster compressor, which passes bleed air into the bypass duct.

The BR715 jet engine takes advantage of a newly developed logic driving this valve. Monitoring input parameters such as spool and flight speeds as well as altitude, the logic provides the capability of appropriately positioning the valve, thereby maintaining an optimum performance of the engine throughout the whole operating range. Furthermore the logic allows for reaction on thrust reverser deployment or more serious events such as surge and foreign object damage.

Much effort has been put into the development of the newly designed transient valve logic, which in its main part is based on the rate of change of the pressure at the exit of the HP compressor. This logic ensures a sufficient surge margin in case of deceleration or reslams of the engine.

A simulation of the logic was set up, connected to a synthesis program for the two spool engine and used for the development. After this basic testing, the logic was implemented in the BR715 engine, which first ran in April 1997. The actual data from basic handling tests performed with this engine were analysed and checked against simulation results. The paper describes the logic in detail and contains first results. They prove a proper working of the logic and show the strong effect of the booster bleed valve.

Topics: Compressors , Surges
Commentary by Dr. Valentin Fuster
1998;():V005T15A023. doi:10.1115/98-GT-264.
FREE TO VIEW

This paper presents the first attempt to stabilize rotating stall in a single-stage transonic axial flow compressor with inlet distortion using active feedback control. The experiments were conducted at the NASA Lewis Research Center on a single-stage transonic core compressor inlet stage. An annular array of 12 jet-injectors located upstream of the rotor tip was used for forced response testing and to extend the compressor stable operating range. Results for radial distortion are reported in this paper.

First, the effects of radial distortion on the compressor performance and the dynamic behavior were investigated. Control laws were designed using empirical transfer function estimates determined from forced response results. The transfer functions indicated that the compressor dynamics are decoupled with radial inlet distortion, as they are for the case of undistorted inlet flow. Single-input-single-output (SISO) control strategies were therefore used for the radial distortion controller designs.

Steady axisymmetric injection of 4% of the compressor mass flow resulted in a reduction in stalling mass flow of 9.7% relative to the case with inlet distortion and no injection. Use of a robust H controller with unsteady non-axisymmetric injection achieved a further reduction in stalling mass flow of 7.5%, resulting in a total reduction of 17.2%.

Commentary by Dr. Valentin Fuster
1998;():V005T15A024. doi:10.1115/98-GT-265.
FREE TO VIEW

This paper presents the first attempt to stabilize rotating stall in a single-stage transonic axial flow compressor with inlet distortion using active feedback control. The experiments were conducted at the NASA Lewis Research Center on a single-stage transonic core compressor inlet stage. An array of 12 jet injectors located upstream of the compressor was used for forced response testing and feedback stabilization. Results for a circumferential total pressure distortion of about one dynamic head and a 120° extent (DC(60) = 0.61) are reported in this paper. Part I (Spakovszky et al. (1998)) reports results for radial distortion.

Control laws were designed using empirical transfer function estimates determined from forced response results. Distortion introduces coupling between the harmonics of circumferential pressure perturbations, requiring multi-variable identification and control design techniques. The compressor response displayed a strong first spatial harmonic, dominated by the well known incompressible Moore-Greitzer mode.

Steady axisymmetric injection of 4% of the compressor mass flow resulted in a 6.2% reduction of stalling mass flow. Constant gain feedback, using unsteady asymmetric injection, yielded a further range extension of 9%. A more sophisticated robust controller allowed a reduction in stalling mass flow of 10.2% relative to steady injection, yielding a total reduction in stalling mass flow of 16.4%.

Commentary by Dr. Valentin Fuster
1998;():V005T15A025. doi:10.1115/98-GT-308.
FREE TO VIEW

The effectiveness of jet actuation for active modal control of rotating stall is investigated experimentally. The dominant physical effects of injection, such as momentum and mass addition, are elucidated. The results indicate that several of the theoretical assumptions used in past studies of jet injection for rotating stall control must be revised. An updated model of the compression system with jet actuation which allows for the effect of control feedback dynamics to be adequately characterized is developed and verified with forced response measurements. It predicts the right trends of movement of the critical pole. Preliminary active control results are presented, among which is a 5.5% range extension in downstream flow coefficient.

Commentary by Dr. Valentin Fuster
1998;():V005T15A026. doi:10.1115/98-GT-347.
FREE TO VIEW

The various parameters appearing along an engine’s gas path, such as flows, pressures, temperatures, speeds, etc., vary not only with power condition but also with the ambient conditions at the engine’s inlet. Since a change in inlet temperature and/or pressure will contribute to an attendant change in a gas path parameter’s value, it would be difficult to characterize the aero-thermodynamic relationships between gas turbine engine parameters, (even at a constant engine operating point) unless the ambient conditions are somehow accounted for. This is usually accomplished through the use of corrected engine parameters. Although most of these corrections are well known by practitioners in the industry, knowledge of their origin does not appear to be as commonplace. The purpose of this paper is to fill that gap and furnish a summary of the commonly used corrections for the “major” gas path parameters that are used in performance analysis, diagnostics and control design, and to offer a derivation of these corrections. We will suggest both an analytic approach as well as an empirical approach. The latter can be used to establish the correction for parameters not directly addressed in this paper, as well as to fine tune the correction factors when actual engine data is available.

Topics: Gas turbines
Commentary by Dr. Valentin Fuster
1998;():V005T15A027. doi:10.1115/98-GT-361.
FREE TO VIEW

This paper describes an experiment on a GHH BORSIG Type THM 1304-10 Gas Turbine engine to test the effects of variable vane setting on the vibration behaviour of the blades in all 10 stages of the axial compressor. The rotor was fitted with a network of strain-gauges. An analogue telemetry system was arranged using standard hardware and special application software to display in real-time and to log the full range of frequencies and amplitudes for all instrumented blades. The data acquisition system is described together with a presentation of the live display which allowed engineers to interact with measured results to maximise the benefits of the test whilst all strain-gauges were still functional. Tests were arranged to maximise the vibration data collected at all points before gauge mortality was experienced. Prior to the test, blades were vibrated statically to determine shapes of the first four vibration modes. The paper discusses the fixing techniques for the gauges, the modal shape measurement technique and the calibration of the strain-gauges. The telemetry system architecture and multiplexing arrangement are described together with examples of typical test data and the conclusions concerning the effects on blade vibration of different variable inlet guide vanes (IGV) settings.

Commentary by Dr. Valentin Fuster
1998;():V005T15A028. doi:10.1115/98-GT-362.
FREE TO VIEW

Advances in silicon-on-insulator (SOI) integrated circuit technology and the steady development of wider band gap semiconductors like silicon carbide are enabling the practical deployment of high temperature electronics. High temperature civilian and military electronics applications include distributed controls for aircraft, automotive electronics, electric vehicles and instrumentation for geothermal wells, oil well logging and nuclear reactors. While integrated circuits are key to the realization of complete high temperature electronic systems, passive components including resistors, capacitors, magnetics and crystals are also required. This paper will present characterization data obtained from a number of silicon high temperature integrated evaluated over a range of elevated temperatures and aged at a selected high temperature. This paper will also present a representative cross section of high temperature passive component characterization data for device types needed by many applications. Device types represented will include both small signal and power resistors and capacitors. Specific problems encountered with the employment of these devices in harsh environments will be discussed for each family of components. The goal in presenting this information is to demonstrate the viability of a significant number of commercially available silicon integrated circuits and passive components that operate at elevated temperatures as well as to encourage component suppliers to continue to optimize a selection of their product offerings for operation at higher temperatures. In addition, systems designers will be encouraged to view this information with an eye toward the conception and implementation of reliable and affordable high temperature systems.

Commentary by Dr. Valentin Fuster
1998;():V005T15A029. doi:10.1115/98-GT-380.
FREE TO VIEW

A calibration equation is derived linking the non-dimensional entropy generation rate per unit area with the non-dimensional aerodynamic wall shear stress and free stream pressure gradient. It is proposed that the latter quantities, which can be measured from surface gauges, be used to measure the profile entropy generation rate. It is shown that the equation is accurate for a wide range of well-defined laminar profiles. To measure the dimensional entropy generation rate per unit area requires measurement of the thickness of the boundary layer. A general profile equation is given and used to show the range of accuracy of a further simplification to the calibration. For flows with low free stream pressure gradients, the entropy generation rate is very simply related to the wall shear stress, if both are expressed without units. An array of heated thin film sensors is calibrated for the measurement of wall shear stress, thus demonstrating the feasibility of using them to measure profile entropy generation rate.

Commentary by Dr. Valentin Fuster
1998;():V005T15A030. doi:10.1115/98-GT-414.
FREE TO VIEW

Over the past several decades, gas turbine control systems have evolved from hydro-mechanical systems to full authority redundant electronic systems. One advanced technology with potential to revolutionize the way engine system designers build new products is high temperature distributed controls. Distributed control systems put electronics close to control functions and reduce the number of interconnects between central processors and sensors or effectors. In distributed systems, power and data buses take the place of multiple discrete analog wire bundles found in centralized control systems. Distributed modules interconnected with power and data buses control effectors such as hydraulic actuators or solenoid valves and read sensors to measure pressures, temperatures and speeds. With distributed controls, many gas turbine applications will require high temperature electronics ruggedized to survive the demanding environment.

For these new systems, manufacturers must determine how to maximize the use of standard interfaces and electronic components and minimize the use of custom parts. Two particular areas would benefit the aerospace industry include distributed system power supplies and communication data buses since these designs play an important role in system cost, weight, size and reliability. Interface standardization will benefit engine manufacturers by lowering system cost and enabling inter-changeability of distributed engine control components from different suppliers. Careful attention to architectural design details for the power supplies and data buses can lead to systems that meet the needs of end users.

Commentary by Dr. Valentin Fuster
1998;():V005T15A031. doi:10.1115/98-GT-455.
FREE TO VIEW

Moving power generation equipment is a major undertaking under the best of circumstances. This paper documents the project to provide more reliable backup power for the island of Gotland, off the coast of Sweden in the Baltic Sea. All the power for this scenic island is normally supplied by two sub-sea, HVDC cables from the Swedish mainland. In the early 1990’s, the power company started the planning to redeploy some of their peaking generation assets more effectively on Gotland. Providing the island with more generating capacity and faster recovery in the event of a partial or total loss of the DC tie-line was deemed a worthy endeavor.

Preparations to move the two 60 MW AVON Quad-pack gas turbine-generators were initiated late in 1994. In addition to commencing green-field site construction to receive the units, the company initiated a controls retrofit of the 1970-vintage units with new fault-tolerant combustion and generation automatic voltage regulation control systems to the relocated turbines.

The scope and timeline for this complex, fast-track project are described along with specific challenges. Results versus expectations are also documented.

Topics: Emergency power
Commentary by Dr. Valentin Fuster
1998;():V005T15A032. doi:10.1115/98-GT-456.
FREE TO VIEW

This paper presents the main assumptions and targets of the strain and stress states modelling of the steam turbine components. The analysis is undertaken due to the main components of HP and IP turbine sections (inner and outer casings, valve bodies and rotors) working under a significant load. Stress modelling has been divided into two parts: a simplified analysis and the detailed one. The former has been proposed in this way that the on-line calculations can be done. It’s based on special functions built upon multi-variant heat process simulations. The functions mentioned make it possible to describe the maximum component stress from some measured temperatures. This part should meet the requirements due to the stress state modelling during a usual turbine operation.

The detailed analysis comprises the unsteady operating conditions, bearing the stamp of intensive heating or cooling processes. As a result of the advanced calculation methods and full stress models being used the analyses is carried off-line.

Topics: Steam turbines
Commentary by Dr. Valentin Fuster
1998;():V005T15A033. doi:10.1115/98-GT-457.
FREE TO VIEW

The Boeing 777 airplane has been designed with a Thrust Asymmetry Compensation System (TAC). This system reduces flight crew workload during single-engine flight operation by automatically commanding rudder deflection to compensate for different levels of thrust between the left and right engines. The airplane is equipped with two Engine Data Interface Units (EDIU) which compute thrust based on engine parameters. This paper will describe the thrust calculation, the fault detection requirements and the resultant fault monitors, and the processes used to validate the EDIU logic.

Topics: Thrust , Propulsion , Aircraft
Commentary by Dr. Valentin Fuster
1998;():V005T15A034. doi:10.1115/98-GT-458.
FREE TO VIEW

This paper presents an overview of current research efforts aimed at improving turbine-engine structural-instrumentation capabilities. Emphasis is placed on non-intrusive concepts that will be applicable to the advanced engines currently in use, or concepts being developed for initial operational testing shortly after the turn of the century. Technologies to be described include: blade-tip deflection sensors to determine dynamic stress, thermographic phosphors to measure metal temperature, pressure-sensitive paints and air etalons to measure dynamic pressure, and micro-electro-mechanical systems to assess a variety of parameters. Advantages, potential problems and limitations of each system are presented, and an assessment of the applicability of each system to either the Integrated High Performance Turbine Engine Technology or the High Cycle Fatigue initiative is made.

Commentary by Dr. Valentin Fuster
1998;():V005T15A035. doi:10.1115/98-GT-491.
FREE TO VIEW

The Propulsion Instrumentation Working Group (PIWG) was formed to cooperatively address critical propulsion engine development test instrumentation and sensor issues. Members of PIWG include the Air Force Arnold Engineering and Development Center, Air Force Wright Laboratory, AlliedSignal Engines, Allison Engine Company, General Electric Aircraft Engines, NASA Lewis Research Center, and Pratt & Whitney with the Ohio Aerospace Institute serving as the administrator to the working group. This paper describes the benefits of this cooperative effort, defines the mission statement, and describes instrumentation and sensor issues currently being addressed.

Commentary by Dr. Valentin Fuster
1998;():V005T15A036. doi:10.1115/98-GT-499.
FREE TO VIEW

This paper presents measurements of static pressures on the stator vane suction side of a high-speed single stage fan using the technique of pressure sensitive paint (PSP). The paper illustrates development in application of the relatively new experimental technique to the complex environment of internal flows in turbomachines. First, there is a short explanation of the physics of the PSP technique and a discussion of calibration methods for pressure sensitive paint in the turbomachinery environment. A description of the image conversion process follows. The recorded image of the stator vane pressure field is skewed due to the limited optical access and must be converted to the meridional plane projection for comparison with analytical predictions. The experimental results for seven operating conditions along an off-design rotational speed line are shown in a concise form, including performance map points, midspan static tap pressure distributions, and vane suction side pressure fields. Then, a comparison between static tap and pressure sensitive paint data is discussed. Finally, the paper lists shortcomings of the pressure sensitive paint technology and lessons learned in this high-speed fan application.

Topics: Stators
Commentary by Dr. Valentin Fuster