0

Manufacturing Materials and Metallurgy

1997;():V004T12A001. doi:10.1115/97-GT-117.
FREE TO VIEW

Turbine inlet temperatures over the next few years will approach 1650°C (3000°F) at maximum power for the latest large commercial turbo fan engines, resulting in high fuel efficiency and thrust levels approaching 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, design technology for stresses and airflow, single crystal and directionally solidified casting process improvements and the development and use of rhenium (Re) containing high γ′ volume fraction nickel-base superalloys with advanced coatings, including full-airfoil ceramic thermal barrier coatings. Re additions to cast airfoil superalloys not only improve creep and thermo-mechanical fatigue strength but also environmental properties, including coating performance. Re dramatically slows down diffusion in these alloys at high operating temperatures.

A team approach has been used to develop a family of two nickel-base single crystal alloys (CMSX-4® containing 3% Re and CMSX®−10 containing 6% Re) and a directionally solidified, columnar grain nickel-base alloy (CM 186 LC® containing 3% Re) for a variety of turbine engine applications. A range of critical properties of these alloys is reviewed in relation to turbine component engineering performance through engine certification testing and service experience.

Industrial turbines are now commencing to use this aero developed turbine technology in both small and large frame units in addition to aero-derivative industrial engines. These applications are demanding, with high reliability required for turbine airfoils out to 25,000 hours, with perhaps greater than 50% of the time spent at maximum power. Combined cycle efficiencies of large frame industrial engines is scheduled to reach 60% in the U.S. ATS programme. Application experience to a total 1.3 million engine hours and 28,000 hours individual blade set service for CMSX-4 first stage turbine blades is reviewed for a small frame industrial engine.

Commentary by Dr. Valentin Fuster
1997;():V004T12A002. doi:10.1115/97-GT-118.
FREE TO VIEW

A new procedure being developed in British Standards for the assessment of creep-rupture data is described, and evaluated with trial data sets of gas turbine blading materials. The procedure is applied in phases. An important development by statistical experts is a framework for the main assessment phase which uses maximum-likelihood fitting methods for the treatment of unfailed test points and error variance. The framework selects models from a standard suite (together with any other linear models supplied by the assessor) using statistical criteria, but also incorporates metallurgical judgement. The improved representation of the experimental data compared with previous fitting methods, and the associated statistical tests indicate that the new procedure can be used to derive rupture strength values for gas turbine materials with confidence.

Topics: Creep , Gas turbines , Rupture
Commentary by Dr. Valentin Fuster
1997;():V004T12A003. doi:10.1115/97-GT-138.
FREE TO VIEW

Recent work done by Aptech Engineering Services, Inc. (APTECH) has demonstrated that many hot section gas turbine components are not achieving their intended service lives and/or maintenance intervals, falling short by a factor of as much as ten. As summarized in (Makansi, 1996), this observation has been confirmed by numerous industry and literature reports of failures, degradation, reduced availability and increased maintenance for a wide variety of turbines and manufacturers.

This paper focuses on hot corrosion damage as one common mechanism that is frequently the cause of failure to reach expected service intervals. In general, attempts are made to limit hot corrosion damage by specifying impurity limits for fuel oil, combustion air, NOx reduction injection water and compressor washing water. Strict adherence to manufacturers recommended impurity limits is often very difficult to achieve, particularly if the machine is frequently cycled through harsh (but not uncommon) operating transients. A discussion of various manufacturer specifications and their implications will be presented, followed by two case studies which demonstrate that extensive hot corrosion damage can result even when manufacturers specifications for fuel, air and water purity are (apparently) being met. A proposed modification of traditional impurity limit specifications is offered, which would include guidelines for acceptable corrosion rate limits. Practical recommendations for reducing the potential for hot corrosion are also offered.

Topics: Failure
Commentary by Dr. Valentin Fuster
1997;():V004T12A004. doi:10.1115/97-GT-167.
FREE TO VIEW

In this paper the objectives and implementation of the company Materials Database at European Gas Turbines (EGT) are described. In the initial stages, a materials audit identified the existing functional requirements for materials information, the properties required by each function, and the flow of information. The main requirement of the database is to provide materials design data. Related, but separate databases for bibliographic and test data were also created. Technical procedures are necessary to define and control the content, use and maintenance of the Database, with an example being the creation, review and approval of materials property sets. Written procedures were also prepared to describe the flow of data resulting in relationships between a unique material specification, the ‘design data’ and its source, and the applications of the material for EGTs engines.

The database and technical procedures developed by EGT may be used as a model for other manufacturers developing their own computerised systems. It is shown that data may be represented in several ways, and it is argued that the structure of the data evolves during the implementation stage. Current standards and key publications (eg. Newton, 1993, and Buttner and Kröckel, 1989) which define how to build a materials database do not, so far, define the information content for gas turbine materials. With increasing electronic exchange of data amongst turbine manufacturers, and with their material suppliers and external test houses, there is an urgent need for standardisation of exchange methods. It is concluded that new standards for formatting and exchange of material property data will be of general benefit to the gas turbine industry.

Commentary by Dr. Valentin Fuster
1997;():V004T12A005. doi:10.1115/97-GT-168.
FREE TO VIEW

In this study, the microstructure of IN738LC turbine blades ground by two different vendors was examined. Turbine blades from a grinding vendor reporting the occasional appearance of “linear indications” were evaluated. The indications were found to be grain boundary cracks, some over 1 mm in depth, running perpendicular to the grinding direction. The cracked blades showed signs of surface melting which can be regarded as a prerequisite for grain boundary cracking in the grinding of IN738LC. Blades ground by a different vendor, not reporting grinding problems, were also examined and no surface melting was seen. No fundamental difference was seen in the as-cast microstructure between the blades from the two vendors. All the observations in this study thus indicate that the grinding cracks in the blades from the first vendor resulted from improper or uncontrolled grinding conditions.

Commentary by Dr. Valentin Fuster
1997;():V004T12A006. doi:10.1115/97-GT-293.
FREE TO VIEW

Steam has been proposed as an efficient cooling medium for the hot section components in the advanced gas turbines replacing compressed air for more efficient cooling allowing the turbine to operate at higher efficiency. To evaluate the effect of steam on oxidation and corrosion of hot section materials, three superalloys (X-45, Inconel-617 and IN-738LC) were exposed to steam at 840°C. Steam environments used were (a) steam generated from deionized water, (b) steam generated from deionized water with 5ppm each of NaCl and Na2SO4 and (c) steam generated from deionized water with 50 ppm each of NaCl and Na2SO4. The respective exposure times were 3900, 2950, and 1450 hours. IN-738LC showed severe internal oxidation in steam. In contaminated steam the hot corrosion damage was maximum in Inconel-617. X-45 showed less oxidation damage than IN-738LC and less hot corrosion than Inconel-617.

Topics: Gas turbines , Steam
Commentary by Dr. Valentin Fuster
1997;():V004T12A007. doi:10.1115/97-GT-348.
FREE TO VIEW

Gas tungsten arc welding (GTAW) using AC or DC current types has been utilised for the manufacture and repair of turbine engine components for many years. The weld repair of single crystal components such as turbine blade tips is also important. However, welding of a single crystal alloy has its associated problems (which will be discussed later). SMP14 is an advanced single crystal, Ni-base superalloy used for turbine blading. This alloy has enhanced mechanical properties at elevated temperatures when compared to equiaxed, directionally solidified and first generation single crystal superalloys. The objective of this task is to investigate the weldability of SMP14 using the new variable polarity GTAW process. Metallurgical investigations were undertaken to evaluate the microstructure of the welded region. Microporosity was found in the weld and there was no evidence of recrystallization, sub-grains beneath the weld or microcracking/microfissuring. Tensile test evaluations revealed that a high strength weld equivalent to equiaxed MAR-M247, Ni-based superalloy could be achieved. This appears to be the highest weld mechanical strength achieved to date when joining any Ni-base single crystal alloy together. Fractography was also utilised to analyse the fracture surfaces of the tensile test samples. Oxidation tests also revealed that the oxidation resistance of the weld was good and will be suitable for weld tip restoration where the weld needs to have good oxidation resistance.

Commentary by Dr. Valentin Fuster
1997;():V004T12A008. doi:10.1115/97-GT-349.
FREE TO VIEW

To estimate the residual life of heavy duty gas turbine hot section components, we studied the microstructure of aged, stress aged and creep ruptured γ′ strengthened alloy for combustor parts. The coarsening rate of precipitate was evaluated, and the activation energy for coarsening was found to be 327 kJ/mol which is slightly higher than the activation energy of volume diffusion for Al, Ti in a Ni solid solution. According to the experimental results, stress dependency of coarsening is not apparent, and the mean diameter of γ′ for creep ruptured specimens is not sensitive to temperature under the same stress conditions. This phenomenon can be explained with some assumptions. Using these results, methods for estimating metal temperature and creep damage based on microstructural information are performed for stress aged samples. These estimated results were compared with actual values.

Commentary by Dr. Valentin Fuster
1997;():V004T12A009. doi:10.1115/97-GT-350.
FREE TO VIEW

In order to develop the life prediction method under creep-fatigue loading for gas turbine combustion transition piece, creep-fatigue tests were carried out on both as-received and aged Ni-based superalloy Nimonic 263. Crack initiation and propagation behaviors for the smooth specimen were observed. An unique relationship was obtained between life fraction and the maximum surface crack length under triangular wave shape loading tests, except the results for the trapezoidal wave loading tests. The latter results were due to the over estimation of the surface crack length at the crack initiation. These were caused from an oxide film break during straining. In the case of removing the oxide film before the measurement of surface crack, the relationship between life fraction and the maximum surface crack length obtained as unique relationship regardless of triangular and trapezoidal strain wave shapes. Using the life prediction method proposed, which is based on maximum surface crack length, the damage of combustion transition piece materials in service was evaluated.

Commentary by Dr. Valentin Fuster
1997;():V004T12A010. doi:10.1115/97-GT-363.
FREE TO VIEW

The failure mode for thermal barrier coatings in gas turbine engines is spallation at or near the ceramic to metal interface. We propose that the two most important factors leading to this failure are the change in the bond strength and bond stress with cycling. Five methods of measuring stress near the ceramic bond coat interface and four methods of bond strength measurement were investigated. Laser fluorescence and enhanced laboratory x-ray methods have the most potential for stress measurement, while the promising bond strength measurement methods are direct pull testing, chevron notch fracture toughness, and laser induced ultrasonic spallation. The stress and strength as a function of thermal cycles was determined for one set of yttria stabilized zirconia coated single crystal samples. Both a decreasing residual compression stress in the oxide layer that joins the ceramic to the metal and a decreasing bond strength were measured as a function of thermal cycles.

Commentary by Dr. Valentin Fuster
1997;():V004T12A011. doi:10.1115/97-GT-365.
FREE TO VIEW

Industrial gas turbine components are subject, in the course of their operating life, to various kinds of damages, requiring repair processes during periodical overhauling operations. Blades, in particular, suffer from creep, corrosion, wear phenomena.

The majority of blade damage is currently repaired by means of manual TIG welding, with a filler metal which is often different from the blade alloy. This leads to an inferior metallurgical and mechanical condition of the repaired area as compared to the base metal. Besides, the nickel superalloys of the blades are often subject to cracking during welding operations.

A process of laser welding for the repair of the airfoil tip has been introduced and optimized, to improve the characteristics of the repaired component. Powder of the same alloy of the part is used as filler metal, and the process is carried out using a Nd:YAG laser, equipped with a 6–axis CNC motion control. The original blade geometry is rebuilt by multi–layer cladding, then the blade is submitted to machining operations, NDT testing and heat treatment.

The optimizing activity has been performed with the aid of microstructural characterization, chemical composition checking (by EDX microanalysis), hardness and stress rupture testing of the welded specimens.

Commentary by Dr. Valentin Fuster
1997;():V004T12A012. doi:10.1115/97-GT-372.
FREE TO VIEW

This summary describes the repair and refurbishment process development for a set of compressor blades that suffered from heavy pitting corrosion. The objective was to restore the surface of the corroded blades and regain fatigue strength, while maintaining the proper airfoil profile.

Metallurgical evaluation during the development and refurbishment indicated where improvements were needed and later on controlled the progress of the repair. Mechanical tests, such as fatigue and hardness, were performed to ensure that mechanical property criteria were met following the complete refurbishment of the blades.

The resulting repair scheme is generic and can be used on most types of gas turbine compressor blades.

Commentary by Dr. Valentin Fuster
1997;():V004T12A013. doi:10.1115/97-GT-373.
FREE TO VIEW

The life expectancy of gas turbine components is strongly dependant on the service conditions in which they operate. Diverse factors can reduce the life dramatically, or can extend it far beyond what was originally predicted by the Original Equipment Manufacturer. In order to examine the condition of the casing, compressor blading, stationary turbine vanes, and rotating turbine blades, nondestructive inspection has been performed more frequently as part of an overhaul. During inspections, visual examination is usually the only method used. Hot corrosion, oxidation or overaging of the blades may not be obvious during such an inspection. The most accurate method of evaluation involves the removal of the parts from the unit, complete visual and penetrant inspection, metallurgical evaluation, and destructive mechanical testing. Yet, complete disassembly only takes place if obvious problems are found. Microstructural analysis and physical testing in such situations add much to the condition assessment. Although many in-situ techniques have been in place for years, applying them to the evaluation of gas turbine components has been limited. When nondestructive inspection is done regularly, the art of evaluation and life assessment can become more scientific. The risk of failure can be reduced and the component life extended. This paper describes the procedure for preparation and examination.

Commentary by Dr. Valentin Fuster
1997;():V004T12A014. doi:10.1115/97-GT-374.
FREE TO VIEW

A study of the variation in fatigue crack propagation behavior with respect to specimen orientation, aging and loading was done for HAYNES® 242™ Alloy. Fatigue crack growth tests were conducted at 650°C in vacuum. Heat treatment of the material resulted in enhanced fatigue crack growth properties. The influence of orientation on the fatigue crack growth properties was found to be minimal, although in some cases, where the notch was oriented perpendicular to the rolling direction, marginal improvement is observed. R-ratio was found to affect the fatigue crack growth behavior significantly. Fatigue crack growth rates increased when the R-ratio was changed from 0 to 0.5 and then decreased when it was further increased to 0.8. Fracture surfaces were characterized by scanning electron microscopy to explain the crack growth behavior. These results have been studied in light of the crack growth characteristics to obtain an understanding of the material behavior when subjected to fatigue loading.

Commentary by Dr. Valentin Fuster
1997;():V004T12A015. doi:10.1115/97-GT-389.
FREE TO VIEW

The hot gas path section components of land based turbines require materials with superior mechanical properties and good hot corrosion and oxidation resistance. These components are generally coated with either a diffusion coating (aluminide or platinum aluminide) or with an overlay coating (MCrAlY) to provide additional hot corrosion and/or oxidation protection. These coatings degrade due to inward and outward diffusion of elements during service. Outward diffusion of aluminum results in formation of a protective oxide layer on the surface. When the protective oxide spalls, aluminum in the coating diffuses out to reform the oxide layer. Accelerated oxidation and failure of coating occur when the Al content in the coating is insufficient to reform a continuous alumina film. This paper describes development of a coating life prediction model that accounts for both oxidation and oxide spallation under thermal mechanical loading, as well as diffusion of elements that dictate the end of useful life. Cyclic oxidation data for aluminide and platinum aluminide coatings were generated to determine model constants. Applications of this model for predicting cyclic oxidation life of coated materials are demonstrated. Work is underway to develop additional material data and to qualify the model for determining actual blade and vane coating refurbishment intervals.

Topics: Coatings , oxidation
Commentary by Dr. Valentin Fuster
1997;():V004T12A016. doi:10.1115/97-GT-423.
FREE TO VIEW

In service degradation of MCrAlY coated camberline cooled blades resulted after 16,000 hours of operation. The peripheral cooling design was developed to reduce the operating metal temperature, thereby extending the coating life and extending the refurbishment cycle. To evaluate the effectiveness of this design, camberline and peripheral cooled blades were coated with MCrAlY type coating for evaluation. The blades were installed in a 501D5 combustion turbine for rainbow testing in 1991. In this paper, the metallurgical evaluation results of the coated peripheral style blades exposed to 24,000 hours of operation is presented.

Commentary by Dr. Valentin Fuster
1997;():V004T12A017. doi:10.1115/97-GT-424.
FREE TO VIEW

Single crystal (SC) nickel based superalloys have been used in aero engine applications for a long time. The SC casting techniques are not yet widely used in the land based engine components because of their much larger size which makes casting and the subsequent heat treatment more difficult. In large casting, microporosity and elemental segregation are more severe. The γ/γ′ eutectic structure and dendritic arm spacing are also larger due to large casting size. All these defects will affect the heat treatment process and the subsequent service properties, including the resistance to oxidation and hot corrosion. In this work, as-cast and heat treated CMSX-4 specimens from small and large blades were used to study the effects of cast defects and heat treatment on Na2SO4 -induced hot corrosion resistance of the single crystal superalloy. The tests were carried out between 900 to 1000°C, for times ranging than 10 to 600 hours in the presence of a Na2SO4 deposit. The specimens from as-cast large blade underwent catastrophic attack after a short exposure time. But the initiation time for catastrophic attack approximately doubled in the case of wholly solution heat treated specimens. It is thus likely that although the hot corrosion resistance of this single crystal material is not as good as that of IN738, proper heat treatment can improve its hot corrosion resistance.

Commentary by Dr. Valentin Fuster
1997;():V004T12A018. doi:10.1115/97-GT-425.
FREE TO VIEW

Advanced turbines with improved efficiency require materials that can operate at higher temperatures. Availability of these materials would minimize cooling flow requirements and thus, improve the efficiency of a turbine. Advanced processing such as directional solidification (DS), can improve temperature capability of the majority of Ni based superalloys. However, results of earlier work on IN-738 reveal that the DS process does not significantly improve temperature capability of this alloy. A research program was initiated to develop a corrosion resistant Ni-based DS blade material for land-based turbines.

In this program, eight heats with varied Cr, Al, Ti, Ta, and W contents were selected for evaluation. Screening tests performed on these heats in the DS condition include tensile, creep, and corrosion. The results of experimental heats were compared with those of IN-738 in the equiaxed condition. From these results, two chemistries offering approximately 100°F temperature advantage at typical row 1 turbine blade operating stress, were selected for castability and further mechanical property evaluation. Several row 1 solid and cored turbine blades were successfully cast. The blades were evaluated for grain structure and mechanical properties. Tests were also conducted to evaluate the effects of withdrawal rates on properties. These results are summarized in this paper.

Commentary by Dr. Valentin Fuster
1997;():V004T12A019. doi:10.1115/97-GT-426.
FREE TO VIEW

Two different, non-Re containing single crystal (SX) superalloys are defined primarily for industrial turbine application. The alloys, CMSX®-11B and CMSX-®-11C, contain respective chromium levels of about 12.5% and 14.5%. Both materials develop unique and extremely good blends of hot corrosion and oxidation resistance. They exhibit extremely good testability, employ relatively simple solution heat treatments and provide creep strength which is as good or better in comparison to other first generation SX materials such as CMSX-2/3, PWA 1480 and René N4. Moreover, at certain engine-pertinent temperature/stress conditions, and particularly in long-term tests (greater than 1,000 hours duration), the alloys appear to exhibit density corrected rupture strengths which are similar or better than CMSX-4® and other second generation SX casting superalloys.

Commentary by Dr. Valentin Fuster
1997;():V004T12A020. doi:10.1115/97-GT-427.
FREE TO VIEW

The rhenium containing second generation single crystal alloy CMSX-4® was introduced for turbine blading in the Solar® Turbines Incorporated Mars® T-14000 engine in 1990. Based on the initial success with the first 4,000 hour engine test, the alloy was confirmed as bill of material for the stage 1 turbine blades in the T-14000 engine. Field experience has been excellent and there are now about 1,500,000 hours total accumulated engine time with several engines having completed their first overhaul cycle (Typically 30,000 hours). Use of this alloy is now being considered by Solar for other turbine airfoil components.

This update provides the results of metallurgical evaluations on high time blades. A detailed analysis of the substrate and coating, microstructure following the 25,000 hours field exposure was conducted and the results reported herein. A brief discussion of the effect of PtAl coatings on fatigue properties is also included.

Commentary by Dr. Valentin Fuster
1997;():V004T12A021. doi:10.1115/97-GT-428.
FREE TO VIEW

The Department of Energy’s Turbine Airfoil Manufacturing Technology Program was initiated at PCC Airfoils, Inc./General Electric Power Generation Group. The specific program goal is to define manufacturing methods for single crystal (SX) technology to be applied to airfoil components for power generation applications. A number of technical issues are being addressed and these form the task structure of the program and will be discussed including:

• Alloy Melt Practice to Reduce Sulfur Content in Alloys

• Modification/Improvement of SX Casting Process

• Core Materials and Design

• Grain Orientation Control

Melt desulfurization has successfully reduced sulfur to below 1 part/million with accompanying improvements in oxidation resistance compared to undesulfurized material. Casting variables have been identified which can optimize casting quality. Core material evaluations to date indicate the initial core compositions need to be optimized to eliminate core break during processing operations. Grain orientation control addresses the formation of a database to enable decisions to be made concerning the establishment of grain limit defect criteria. Liberalizing the defect criteria will have a positive effect on the producibility and yields associated with the SX castings. The planned testing has been formulated.

Commentary by Dr. Valentin Fuster
1997;():V004T12A022. doi:10.1115/97-GT-429.
FREE TO VIEW

Future advancements in the efficiency and reliability of Industrial Gas Turbines (IGT) will be closely tied to the application of advanced materials, together with increasingly sophisticated turbine hot section designs. An example of this trend is illustrated by the recent design of a first stage blade component for an advanced IGT concept utilizing the third generation single crystal superalloy CMSX-10. It is anticipated that alloy CMSX-10 will permit the use of increased turbine firing temperatures with reduced cooling flows compared to previous recuperated turbine designs, while maintaining acceptable blade durability and life-cycle cost.

This paper discusses some of the design/materials analyses and cost studies performed on the blade, which ultimately led to the consideration of alloy CMSX-10 for the IGT application. The solid modeling and finite element blade design methods which allowed the incorporation of state-of-the-art cooling technology and aerodynamics are described. Alloy CMSX-10 characteristics, particularly mechanical properties and microstructural stability considerations, are discussed. Additionally, the results of a recent casting demonstration in an IGT blade configuration are presented. Finally, future tasks supporting the application of the alloy are outlined, such as coatings development efforts and the DOE/ORNL sponsored Land Based Turbine Casting Initiative; activities sponsored through a cooperative agreement with the United States Department of Energy within the Advanced Turbine System (ATS) Program.

Commentary by Dr. Valentin Fuster
1997;():V004T12A023. doi:10.1115/97-GT-430.
FREE TO VIEW

Howmet is currently conducting a Department of Energy funded Advanced Turbine Systems program, the Land Based Turbine Casting Initiative, to scale single crystal casting technological advances developed for aircraft gas turbine engines up to land based gas turbine sized components. The program encompasses technical activities in the areas of: low sulfur alloys (low sulfur melt processing treatments), casting process development (casting process, mold, core and wax), post-cast process development (mold, core, and gating removal, heat treatment, HIP, and inspection), and casting defect tolerance level establishment (freckles, off-orientation, low angle boundaries, and recrystallized grains).

This presentation will highlight recent progress in the areas of low sulfur alloy, large single crystal casting process development, and 2.5D reconstruction X-ray inspection system development. Howmet has developed a melt treatment to reduce the sulfur content of the nickel-based superalloy ingot it manufactures. Sulfur contents achieved with this technique will be compared to sulfur levels present in currently available ingot. The effect of the reduced sulfur contents on oxidation performance is being investigated and will be reported. The casting process development activities are foundry experiments to determine the effect of different process factors on both grain and dimensional quality. The status of casting activities to scale aeroengine processes up to land based sized components will be presented. The 2.5D reconstruction uses X-ray metrology and multiple 2D X-ray views of the casting to reconstruct the 3D geometry of selected features. The large cross sections and long path lengths inherent in large utility land based components will be particularly challenging.

Topics: Casting , Turbines
Commentary by Dr. Valentin Fuster
1997;():V004T12A024. doi:10.1115/97-GT-486.
FREE TO VIEW

The maintenance period of critical gas turbine components operating in advanced D, E and F technology industrial turbines is often determined by the life of the coating system applied to protect the base alloy substrate. The progressively higher firing temperatures used in all of the advanced engine designs results not only in very high metal surface temperatures but also in very high temperature gradients and concomitant thermal stresses Induced in part by the complex cooling systems. In order to develop optimum component life strategies, it is important to establish the actual operating conditions of each component, and, to define the predominant degradation modes.

Metallurgical life assessment of these advanced component designs has identified several distinct coating/base metal failure mechanisms not generally encountered in earlier generation turbines, in addition to the more usual degradation modes, accelerated by the increases in temperature. A review and examples of these degradation mechanisms encountered on service exposed coating systems currently used by some of today’s major manufactures are presented.

Commentary by Dr. Valentin Fuster
1997;():V004T12A025. doi:10.1115/97-GT-487.
FREE TO VIEW

The desire to improve the performance and efficiency of gas turbine engines has led to higher operating temperatures in the turbine sections of the engine. Present materials and materials under development for hot section turbine blades and vanes are not inherently resistant to hot corrosion, and therefore require protective coatings. In the past two decades this has led to increased use of thermally sprayed MCrAlY coatings, both as stand-alone overlay and as a bond coat for thermal barrier coatings.

This paper reviews the issues involved in thermally sprayed MCrAlY and TB coatings onto hot section blades and vanes of industrial gas turbines. The generation of a specification for coating acceptance and its practical implications are discussed. The issues in applying such coatings will be discussed, along with references to manufacturing issues on the shop floor. The difficulties inherent in applying a line-of-sight coating to complex geometric shapes will be discussed, with particular reference to robotics spraying. The utility of using a design-of-experiment approach to satisfy the user will be reviewed. The testing, evaluation, and performance characteristics of typical coatings are discussed.

Commentary by Dr. Valentin Fuster
1997;():V004T12A026. doi:10.1115/97-GT-531.
FREE TO VIEW

Laboratory testing was conducted on air plasma sprayed (APS) and electron beam-physical vapor deposited (EB-PVD) thermal barrier coatings (TBCs) applied onto nickel alloy specimens. As-coated chemistry, microstructure, and bond strength of the TBC systems were evaluated. Cyclic oxidation tests that simulated industrial gas turbine environments were also conducted on the various thermal barrier coatings. This study evaluated the effects of ceramic and metallic coating compositions and application processes on coatings microstructure and performance. The relative cyclic performance of the TBC systems was determined from the laboratory tests.

Commentary by Dr. Valentin Fuster
1997;():V004T12A027. doi:10.1115/97-GT-532.
FREE TO VIEW

This paper describes examination of the microstructure/composition and mechanical properties (22–950 °C) in over aluminized CoCrAlY coatings of advanced gas turbine blades using scanning Auger microprobe and a small punch (SP) testing method. Aluminized coatings consisted of layered structure divided into four regions; (I) Al enriched and Cr depleted region, (II) Al and Cr graded region, (III) fine grained microstructure with a mixture of Al and Cr enriched phases and (IV) Ni/Co interdifusion zone adjacent to the interface. SP specimens were prepared in order that the specimen surface would be located in the various coating regions. SP tests indicated a strong dependence of the fracture properties on the various coatings regions. Coating zones I and II with high micro hardness showed much easier formation of brittle cracks in a wide temperature range, compared to regions III and IV The coating region III had lower room temperature ductility than the zone IV. However, the ductility in the coating zone III exceeded that in the region IV above 730 °C due to a precipitous ductility increase. The integrity of aluminized coatings while in-service is discussed in light of the variation of the low cycle fatigue life as well as the ductility in the layered structure.

Commentary by Dr. Valentin Fuster

Ceramics

1997;():V004T13A001. doi:10.1115/97-GT-067.
FREE TO VIEW

NGK Spark Plug Co., Ltd. has been developing various silicon nitride materials, and the technology for fabricating components for ceramic gas turbines (CGT) using theses materials. We are supplying silicon nitride material components for the project to develop 300 kW class CGT for co-generation in Japan. EC-152 was developed for components that require high strength at high temperature, such as turbine blades and turbine nozzles.

In order to adapt the increasing of the turbine inlet temperature (TIT) up to 1,350 °C in accordance with the project goals, we developed two silicon nitride materials with further unproved properties: ST-1 and ST-2. ST-1 has a higher strength than EC-152 and is suitable for first stage turbine blades and power turbine blades. ST-2 has higher oxidation resistance than EC-152 and is suitable for power turbine nozzles.

In this paper, we report on the properties of these materials, and present the results of evaluations of these materials when they are actually used for CGT components such as first stage turbine blades and power turbine nozzles.

Commentary by Dr. Valentin Fuster
1997;():V004T13A002. doi:10.1115/97-GT-154.
FREE TO VIEW

Experimental and theoretical work concerning the application of ceramic components in small high temperature gas turbines has been performed for several years. The significance of some non-oxide ceramic materials for gas turbines in particular is based on their excellent high temperature properties. The application of ceramic materials allows an increase of the turbine inlet temperature resulting in higher efficiencies and a reduction of pollution emissions.

The inherent brittleness of monolithic ceramic materials can be virtually reduced by reinforcement with ceramic fibers leading to a quasi-ductile behavior. Unfortunately, some problems arise due to oxidation of these composite materials in the presence of hot gas flow containing oxygen.

At the Motoren- und Turbinen Union, München GmbH, comprehensive investigations including strength, oxidation, and thermal shock tests of several materials that seemed to be appropriate for combustor liner applications were undertaken. As a result, C/C, SiC/SiC, and two C/SiC-composites coated with SiC, as oxidation protection, were chosen for examination in a gas turbine combustion chamber.

To prove the suitability of these materials under real engine conditions, the fiber reinforced flame tubes were installed in a small gas turbine operating under varying conditions. The loading of the flame tubes was characterized by wall temperature measurements.

The materials showed different oxidation behavior when exposed to the hot gas flow. Inspection of the C/SiC-composites revealed debonding of the coatings. The C/C- and the SiC/SiC-materials withstood the tests with a maximum cumulated test duration of 90 hours without damage.

Commentary by Dr. Valentin Fuster
1997;():V004T13A003. doi:10.1115/97-GT-155.
FREE TO VIEW

A U. S. Department of Energy (DOE) program with the Allison Engine Company will demonstrate ceramic vanes in an industrial turbine. First-stage ceramic vanes and their metallic mounts are to be designed, fabricated, and operated in a relatively short-term engine test (up to 50 hr). The ceramic vanes and mounts will then be retrofitted into an existing turbine for operation at a commercial site for an extended duration test (up to 8000 hr). The ceramic vanes and metallic mounts have been designed. Thermal and stress analyses of the vanes have calculated acceptable fast fracture stress levels and probabilities of survival exceeding 99.99% for turbine continuous power and emergency shutdown (thermal shock) conditions. The maximum calculated steady-state stress is 169 MPa (24.5 ksi) at a material temperature of 1182°C (2160°F). Consequently, currently available ceramics appear to provide acceptable fast fracture strengths for use in industrial turbines. Long-term materials tests will evaluate the life times and retained strength of ceramics at stress and temperature levels in the range calculated from the ceramic vane analyses. The results of these tests will support the decision on which vane material will be used in the long duration turbine demonstration. A successful demonstration could provide a basis for incorporating first-stage ceramic vanes into current generation industrial turbines and also the introduction of ceramic airfoils into downstream rows of future high temperature Advanced Turbine System (ATS) engines.

Topics: Ceramics , Turbines
Commentary by Dr. Valentin Fuster
1997;():V004T13A004. doi:10.1115/97-GT-156.
FREE TO VIEW

The design and development testing of a full scale, low emissions, ceramic combustor for a 5500 HP industrial gas turbine are described. The combustor was developed under a joint program conducted by the U.S. DOE and Solar Turbines.

The ceramic combustor is designed to replace the production Centaur 50S SoLoNOx burner which uses lean-premixed combustion to limit NOx and CO to 25 and 50 ppm, respectively. Both the ceramic and production combustors are annular in shape and employ twelve premixing, natural gas fuel injectors. The ceramic combustor design effort involved the integration of two CFCC cylinders (76.2 cm [30 in.] and 35.56 cm [14 in.] diameters) into the combustor primary zone.

The ceramic combustor was evaluated at Solar in full scale test rigs and a test engine. Performance of the combustor was excellent with high combustion efficiency and extremely low NOx and CO emissions. The hot walls of the ceramic combustor played a significant role in reducing CO emissions. This suggests that liner cooling air injected through the metal production liner contributes to CO emissions by reaction quenching at the liner walls. It appears that ceramics can serve to improve combustion efficiency near the combustor lean limit which, in turn, would allow further reductions in NOx emissions.

Approximately 50 hours of operation have been accumulated using the ceramic combustor. No significant deterioration in the CFCC liners has been observed. A 4000 hour field test of the combustion system is planned to begin in 1997 as a durability assessment.

Commentary by Dr. Valentin Fuster
1997;():V004T13A005. doi:10.1115/97-GT-157.
FREE TO VIEW

For years of time there have been conducted the investigations of gas-turbine engine parts made of carbon-carbon and ceramic materials.

This paper presents mainly the results of works done to create engine components of ceramic materials. There are given the investigation results on development of equipment and methods intended for use in determining the characteristics of heat-resistant non-metallic materials under ultra high temperature conditions. The unique tooling is developed to be used for conducting mechanical tests in different conditions (vacuum, protective medium, air) at temperatures up to 2200°C. There are considered three possible fields of application of ceramic materials, that are, turbine (1), combustion chamber and other stator components operating at high temperatures (2), bearings (3). Different ceramic elements are designed and manufactured, their structural strength is investigated in the laboratory faculties and also as part of engine gas generators.

Commentary by Dr. Valentin Fuster
1997;():V004T13A006. doi:10.1115/97-GT-284.
FREE TO VIEW

A three year program to evaluate the feasibility of using monolithic silicon nitride ceramic components in gas turbines was conducted. The use of ceramic materials may enable design of turbine components which operate at higher gas temperatures and/or require less cooling air than their metal counterparts. The feasibility evaluation consisted of three tasks: 1) Expand the material properties database for candidate silicon nitride materials, 2) Demonstrate the ability to predict ceramic reliability and life using a conceptual component model and 3) Evaluate the effect of proof testing on conceptual component reliability. The overall feasibility goal was to determine whether established life and reliability targets could be satisfied for the conceptual ceramic component having properties of an available material. Fast and delayed fracture reliability models were developed and validated via thermal shock and tensile experiments. A creep model was developed using tensile creep data. The effect of oxidation was empirically evaluated using four-point flexure samples exposed to flowing natural gas combustion products. The reliability- and life-limiting failure mechanisms were characterized in terms of temperature, stress and probability of component failure. Conservative limits for design of silicon nitride gas turbine components were established.

Commentary by Dr. Valentin Fuster
1997;():V004T13A007. doi:10.1115/97-GT-317.
FREE TO VIEW

A team led by Solar Turbines Incorporated is conducting a three phase program 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. Preliminary and concept design were completed in Phase I. Detailed design, component fabrication, and rig and engine testing are all being conducted in Phase II. Field engine testing will be performed in Phase III. This review summarizes progress on Phases II and III for the program for 1996.

In 1996 the primary activities involved testing of uncooled first stage silicon nitride blades and SiC/SiC CFCC liners in a Centaur 50S engine modified to accept the ceramic components. Cumulative engine test experience by the end of November, 1996 has reached 52 hrs. The longest operating time on a single engine build at full load is 16 hours. Ceramic parts were also proof tested in rigs prior to engine testing. Preparations are currently underway for a 4,000 hour field test at the enhanced oil recovery site of ARCO Western Energy in Bakersfield, California.

Commentary by Dr. Valentin Fuster
1997;():V004T13A008. doi:10.1115/97-GT-318.
FREE TO VIEW

The fast-fracture failure probability of a prototype NT451 SiAlON diesel exhaust valve was predicted using volume-censored flexure and tensile strength data. A tandem set of life prediction codes developed by AlliedSignal Engines was used to make these predictions in this exercise; these codes also yielded 95% confidence estimates on determined two-parameter Weibull distributions and failure probabilities. The prediction of fast-fracture failure probability was compared to actual NT451 SiAlON valve fast-fracture strengths. All the valves failed from volume-flaws, so volume-censored input strength data were required for the fast-fracture failure probability predictions. In addition to the failure probability predictions, the effects of several independent parameters on strength, and ultimately on the failure probability, were also examined in this exercise: the specimen geometry tested; specimen machining parameters consisting of grinding direction and grinding wheel grit size; and the utilization of a post-machining anneal. Within 95% confidence estimates, the effect of specimen geometry, machining parameters, and a post-machining anneal had no effect on the calculated Weibull Moduli and scaling parameters and the consequential predicted reliability.

Commentary by Dr. Valentin Fuster
1997;():V004T13A009. doi:10.1115/97-GT-319.
FREE TO VIEW

The desirable properties of ceramics at high temperatures have generated interest in their use for structural applications such as in advanced turbine systems. Some of these ceramic components, such as vanes and rotors, are subjected to concurrent tensile and compressive stress fields. Design lives for such systems can exceed 10,000 hours. Such long life requirements necessitate subjecting the components to relatively low stresses. The combination of high temperatures and low stresses typically places failure for monolithic ceramics in the creep regime. The objective of this paper is to present a design methodology for predicting the lifetimes of structural components subjected to concurrent transient tensile and compressive creep stress states. In this methodology, failure generally starts at or near the most highly stressed point and subsequently propagates across the section. The creep rupture life is divided into two stages. The first is called the stage of latent failure. During this stage the damage accumulates until it becomes critical at some point within the component, and failure begins. Damage due to compressive stresses is assumed to be negligible. Subsequently, the second stage, named the propagation of failure, takes place. Component failure occurs at the end of this stage when the total carrying capacity of the structure is expended. This methodology utilizes commercially available finite element packages and takes into account the time varying creep stress distributions (stress relaxation). The creep life of a component is divided into short time steps, during which, the stress distribution is assumed constant. The damage is calculated for each time step based on a modified Monkmon-Grant creep rupture criterion. Failure is assumed to commence when the normalized accumulated damage at a point in the body is equal or greater than unity. For tensile/compressive stress states, rupture is assumed to take place when the damage zone is large enough so that the component is no longer able to sustain load. The corresponding time will be the creep rupture life for that component. Flexural and C-ring data of siliconized silicon carbide KX01 material are used to test the viability of this methodology. The NASA integrated design code CARES/Creep (Ceramics Analysis and Reliability Evaluation of Structures/Creep) which utilizes this damage accumulation model was used for this purpose. It was found that the methodology described in this paper yielded reasonable creep rupture life predictions given the amount of scatter in the data.

Commentary by Dr. Valentin Fuster
1997;():V004T13A010. doi:10.1115/97-GT-320.
FREE TO VIEW

The relationship between strength distribution or ground ceramics and machining condition was examined. The strength of ground silicon carbide ceramics at the failure probability of 1 percent, which is termed 1 % strength, was estimated as a function of the maximum grain depth of cut g, which is used as an index of grinding conditions. It is assumed that the fracture is governed by the competing failure mechanism between pre-existing flaws and machining cracks and that the Weibull modulus of machining cracks was independent of grinding condition. The deviation from estimated value is noticeable in small g region although experimental data of the 1% strength of ground specimen approaches that of polished ones with diminishing g. The observed discrepancy is inferred to be attributed to an increase in stress intensity factor by the interaction between a pre-existing flaw and its adjacent machining crack.

Topics: Machining , Ceramics
Commentary by Dr. Valentin Fuster
1997;():V004T13A011. doi:10.1115/97-GT-321.
FREE TO VIEW

Silicon nitride monolithic structural ceramic components have recently been introduced as production parts in commercial aircraft turbomachinery pump and seal applications to take advantage of their unique thermomechanical properties. Additionally, extensive efforts are in progress to develop, evaluate, and productionize silicon nitride components for commercial aircraft turbomachinery hot sections, industrial power generation turbines, and automotive hybrid vehicle turbogenerators. AlliedSignal Ceramic Components has developed a family of in-situ reinforced silicon nitride materials for these applications and is developing and implementing a suite of component fabrication processes to achieve production-viable manufacturing of complex shaped components, including turbine seals, blades, nozzles, wheels, and combustors. A key focus of the manufacturing process development is the need to achieve low cost fabrication of components in order to meet cost targets required for commercial introduction. Finally, the status and plans for a number of aerospace, industrial, and automotive turbomachinery applications are discussed, including commercial aircraft turbomachinery production components (pump and seal parts) and development components (auxiliary power unit turbine blades and nozzles, and propulsion engine wheels and starter wheels), nozzle, blade, wheel and combustor components for automotive hybrid vehicle turbogenerators, and turbine blades and nozzles for industrial power generation gas turbines.

Commentary by Dr. Valentin Fuster
1997;():V004T13A012. doi:10.1115/97-GT-322.
FREE TO VIEW

The European EUREKA project, EU 209, otherwise known as AGATA (Advanced Gas Turbine for Automobiles), is a programme dedicated to the development of three critical ceramic components — a catalytic combustor, a radial turbine wheel and a static heat exchanger — for a 60 kW turbogenerator in an hybrid electric vehicle. These three components, which are of critical importance to the achievement of low emissions and high efficiency, have been designed and developed and will be manufactured and tested as part of a full scale feasibility study. AGATA is a joint project conducted by eight commercial companies and four research institutes in France and Sweden. Silicon nitride ceramics play an important role both in the development of the catalytic combustor and for the radial turbine wheel. This paper outlines the main results of the AGATA project with special emphasis to the development of HIPed Si3N4 combustor and turbine wheel. AC Cerama has developed a HIPed Si3N4 material designated CSN 101. This material has been selected for the catalytic combustor afterburner as well as for the radial turbine wheel. Mechanical properties of the CSN 101 Si3N4 have been found to be at the level of the best available high temperature Si3N4 materials. A new glass encapsulation technique using an interlayer between the glass and the silicon nitride has been shown to give material with excellent strength and creep resistance with as-HIPed surface finish.

Commentary by Dr. Valentin Fuster
1997;():V004T13A013. doi:10.1115/97-GT-323.
FREE TO VIEW

Alumo-boron-nitride-based heat-resistant, thermostable, shrinkage-free structural ceramic material was developed and manufactured by vacuum-compression impregnation technology. Actually, all the high-temperature path parts for 2.5 MW gas-turbine engine were made of this ceramic material [1]. The results of the granulometric, microstructural, phase and other characterization into the material are presented. New conditions of interaction between the boron nitride and aluminium were identified.

Topics: Ceramics , Boron
Commentary by Dr. Valentin Fuster
1997;():V004T13A014. doi:10.1115/97-GT-332.
FREE TO VIEW

The effect of oxidation on the stress-rupture behavior of fiber bundles was modeled. It is shown that oxidation-induced fiber strength degradation results in the delayed failure of the associated fiber bundle and that the fiber bundle strength decreases with time as t−1/4. It is also shown that the temperature dependence of the bundle loss of strength reflects the thermal dependence of the mechanism controlling the oxidation of the fibers. The effect of gauge length on the fiber bundle strength was also analyzed. Numerical examples are presented for the special case of Nicalon™ fibers.

Topics: Fibers , Stress , oxidation , Rupture
Commentary by Dr. Valentin Fuster
1997;():V004T13A015. doi:10.1115/97-GT-355.
FREE TO VIEW

In order to evaluate the durability of silicon-carbides (SiC) and silicon-nitrides (Si3N4), we studied the oxidation behavior of SiC and Si3N4 in 1500°C combustion gas flow. We found that the exposure to the combustion gas flow resulted in the weight losses of those ceramics due to the partial disappearance of the oxidized surface layer.

We investigated the effects of sintering aids and high speed gas flow as possible factors for the disappearance of the oxide layer. Two kinds of SiC, without sintering aids and sintered with B4C, were used as test specimens. After the exposure to combustion gas flow conditions of 1500°C, 150m/s, 0.18MPa, the weight loss rate and thickness of the oxide layer were quite the same for each specimen of SiC. The existence of sintering aids did not have any effect on the disappearance of the oxide layer. To investigate the effect of gas flow, we set each specimen in a tube made of SiC to protect it from the gas flow. The tube had two holes each acting both as inlet and exhaust vents. Consequently, the oxide layer formed thickly. But at the spots on the specimen facing the holes, the oxide layer was thin. Hollows occurred on the specimen of SiC at these spots. It seems that the existence of gas flow is a very important factor in the disappearance of the oxide layer.

Alumina (Al2O3) and zirconia (ZrO2) as oxide ceramics were exposed to the combustion gas flow. The weight of these also decreased. There is a possibility that the weight loss of ceramics in combustion gas flow is caused by degradation of oxide layer on their surface from erosion and hot corrosion due to some oxide scales coming from the test equipment.

Commentary by Dr. Valentin Fuster
1997;():V004T13A016. doi:10.1115/97-GT-356.
FREE TO VIEW

The given paper submits some results of implementation of the first two steps of the program (Table 1) of development of a low-toxic high-efficiency metal-ceramic GTE of 2.5MW power intended for stationary power application, this R&D works having been carried out at NITI EM (Research-Technological Institute for Power Engineering) since 1990.

Commentary by Dr. Valentin Fuster
1997;():V004T13A017. doi:10.1115/97-GT-390.
FREE TO VIEW

A numerical study is made of the thermal characteristics of semitransparent materials exposed to simultaneous conduction and radiation between concentric cylinders. For extremely high-temperature applications, where radiative transfer plays an important role, ceramic-matrix composites, considered as semitransparent materials, are being explored for potential use in turbines and compressors components, spacecraft structures, engine control systems and nuclear reactors. Through the use of a gray model and the the two flux method, specialized equations are developed that generate a system of nonlinear ordinary differential equations. To facilitate the solution of this system, an iterative strategy is adopted. In order to demonstrate the versatility and accuracy of the proposed methodology, the results of several numerical experiments are presented and compared with benchmark solutions.

Commentary by Dr. Valentin Fuster
1997;():V004T13A018. doi:10.1115/97-GT-413.
FREE TO VIEW

All-oxide ceramic composites as a material with potential for long life-time applications at temperatures in the 1400–1600°C range in combustion environments were studied. The properties of available polycrystalline and single crystal oxide fibres were summarised. The literature on stable weak interfaces in all-oxide composites was reviewed. Composites with single crystal fibres, a polycrystalline matrix of the same material as the fibres, and a compatible high temperature stable weak oxide interphase was suggested to be the most promising approach. Recent progress in an ongoing European project aiming at development, scale-up and property evaluation of all-oxide composites is reported. The composite will be applied to a simple prototype combustor tile and tested in a combustor rig.

Commentary by Dr. Valentin Fuster
1997;():V004T13A019. doi:10.1115/97-GT-534.
FREE TO VIEW

Advanced ceramics such as alumina, silicon carbide and silicon nitride (monolithics and composites) have properties that suggest application in gas turbine engines. However, the production of components from these materials is very different from that typical of superalloys and this has limited the range of applications for ceramics in gas turbines. The manufacturing freedom offered by the recently developed technologies termed “rapid prototyping,” RP, or equivalently, “solid freeform fabrication,” SFF, may enable a much wider range of applications to be served in the future. RP was developed to allow production of form-and-fit models without the need for tooling and has proven to be a key assel in the design of new components as well as for the implementation of design changes to existing ones. Direct SFF using engineering materials to prototype components is undergoing continued development and is expected to provide an enabling technology that promises to change design philosophies for components made from ceramics (and other powder-based materials). In this paper, the opportunities for SFF in gas turbine applications are discussed, a brief state-of-the-art overview of RP and its application to engineering ceramics is provided, and a particular process, CAM-LEM, is highlighted.

Commentary by Dr. Valentin Fuster

Structures and Dynamics

1997;():V004T14A001. doi:10.1115/97-GT-009.
FREE TO VIEW

Honeycomb annular seals are an attractive design alternative due to their superior static and dynamic performance. However, their implementation in industrial practice has been delayed by the following characteristics: a) manufacturing time can be appreciable, and b) they can seriously damage the shaft if rubbing occurs. To minimize these problems, “hole-pattern” gas damper seals, which are formed by simply drilling holes into an annular smooth seal, were manufactured and tested. The hole-pattern damper seal stator can be made of high-strength plastic materials which are less likely to damage a shaft during rubbing. The experimental results presented demonstrate that, compared to a honeycomb seal, a hole-pattern damper seal with 3.18 mm hole diameters and a high percentage of hole surface has achieved: (a) an average of 12 percent reduction in leakage rate, and (b) considerably higher effective damping, especially under high speeds and low inlet pressure ratio conditions.

Topics: Dampers , Leakage
Commentary by Dr. Valentin Fuster
1997;():V004T14A002. doi:10.1115/97-GT-011.
FREE TO VIEW

The cause of the low-frequency vibration (subsynchronous vibration) of a high pressure turbine was investigated by the analytical study and vibration exciting test for the actual machine in operation. From the results, it is found that the low-frequency vibration is caused by the decrease of the rotor system damping at high-loading operating conditions. As a countermeasure, a squeeze-film damper is designed in order to increase the damping of the rotor system. After the verification test of the squeeze-film damper’s capability in the workshop, it was installed on the actual turbine. Vibration exciting tests for the high pressure turbine under the actual operating conditions were carried out. These field tests confirmed that the damping of the rotor system was increased as expected in the design and consequently the low-frequency vibrations disappeared completely under all operating conditions.

Commentary by Dr. Valentin Fuster
1997;():V004T14A003. doi:10.1115/97-GT-012.
FREE TO VIEW

Squeeze film dampers (SFDs) provide vibration attenuation and structural isolation to aircraft gas turbine engines which must be able to tolerate larger imbalances while operating above one or more critical speeds. Rotor-bearing-SFD systems are regarded in theory as highly nonlinear, showing jump-phenomena and even chaotic behavior for sufficiently large levels of rotor imbalance. Yet, few experimental results of practical value have verified the analytical predictions. A test rig for measurement of the dynamic forced response of a three-disk rotor (45 kg) supported on two cylindrical SFDs is described. The major objective is to provide a reliable data base to validate and enhance SFD design practice and to allow a direct comparison with analytical models. The open-ends SFD are supported by four-bar centering structures each with a stiffness of 3.5 MN/m. Measured synchronous responses to 9,000 rpm due to various imbalances show the rotor-SFD system to be well damped with amplification factors between 1.6 and 2.1 while traversing cylindrical and conical modes critical speeds. The rotor amplitudes of motion are found to be proportional to the imbalances for the first mode of vibration, and the damping coefficients extracted compare reasonably well to predictions based on the full-film open-ends SFD. Tight lip (elastomeric) seals contribute greatly to the overall damping of the test rig. Measured dynamic pressures at the squeeze film lands are well above ambient values with no indication of lubricant dynamic cavitation as simple theoretical models dictate. The measurements show absence of non-linear behavior of the rotor-SFD apparatus for the range of imbalances tested.

• The research program is a joint effort funded by the National Science Foundation (NSF) and the TAMU Turbomachinery Research Consortium (TRC).

Topics: Dampers , Rotors
Commentary by Dr. Valentin Fuster
1997;():V004T14A004. doi:10.1115/97-GT-013.
FREE TO VIEW

This paper summarizes the development of a finite element rotordynamic solution used in a closed loop simulation for a magnetic bearing rotor system in a gas turbine engine. A magnetic bearing controlled rotor is analyzed, and the state dynamics matrix [A], the shaft control influence matrix [B], and the sensor matrix [C] are constructed. Bode plots of the state-space transfer function are also constructed and compared to the results of the rotor dynamic model.

Commentary by Dr. Valentin Fuster
1997;():V004T14A005. doi:10.1115/97-GT-017.
FREE TO VIEW

The rotor power losses in magnetic bearings are due to eddy currents, hysteresis, and windage. The influence of air gap magnetic flux density and air gap thickness is not well understood at this time. This paper presents measured results in two magnetic bearing radial configurations with a laminated rotor. The rotor power losses were evaluated by measuring the rundown speed of the rotor, in air, after the rotor was spun up to speeds of approximately 30,000 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. 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. Hysteresis and windage effects did not change much from one configuration to the other. The measured rotor power loss increased significantly as the magnetic flux density increased and also increased significantly as the air gap thickness decreased.

Commentary by Dr. Valentin Fuster
1997;():V004T14A006. doi:10.1115/97-GT-018.
FREE TO VIEW

Force versus coil currents and air gap measurements were obtained for an 8-pole planar radial magnetic actuator constructed from laminated silicon iron. Static force measurements were made for journal eccentricities up to 2/3 of the nominal actuator radial clearance and various coil currents spanning the expected operating range. Three theoretical force models of varying degrees of complexity were developed using magnetic circuit theory and constant magnetic material properties. All three models were used to reduce the experimental data and an optimized expression representing the actuator force as a function of journal position and stator coil currents was found. The resulting optimized calibration model produced a proportionality constant and equivalent iron length significantly different from theoretically determined values, 29% and 130% greater respectively.

A detailed error analysis was conducted to quantify the uncertainty in the bearing calibration relationship such that uncertainty bounds can be applied to the in situ actuator force measurements.

Hysteresis testing was conducted for various journal positions. Hysteresis effects were shown to be approximately 2% of the peak force when comparing the differences between the actuator force as the currents were increasing and the forces generated when the currents were decreasing. The actuator frequency response was also examined. An actuator bandwidth of at least 700 Hz was determined. Above 700 Hz the actuator frequency response could not be distinguished from the test fixture frequency response.

Commentary by Dr. Valentin Fuster
1997;():V004T14A007. doi:10.1115/97-GT-019.
FREE TO VIEW

Friction dampers are often used in turbine design to attenuate blade vibration to acceptable levels so as to prolong blades’ service life. A wedge damper, also called a self-centering blade-to-blade damper, can provide more design flexibility to meet various needs in different operating conditions when compared with conventional platform dampers. However, direct coupling of the two inclined friction interfaces of the wedge damper often leads to very complex contact kinematics. In Part I of this two-part paper, a dual-interface friction force model is proposed to investigate the coupling contact kinematics. The key issue of the model formulation is to derive analytical criteria for the stick-slip transitions that can be used to precisely simulate the complex stick-slip motion and, thus, the induced friction force as well. When considering cyclic loading, the induced periodic friction forces can be obtained to determine the effective stiffness and damping of the interfaces over a cycle of motion. In Part II of this paper, the estimated stiffness and damping are then incorporated with the harmonic balance method to predict the forced response of a blade constrained by wedge dampers.

Commentary by Dr. Valentin Fuster
1997;():V004T14A008. doi:10.1115/97-GT-020.
FREE TO VIEW

In the second part of this paper, the application of the proposed dual-interface model to the prediction of the forced response of a blade constrained by wedge dampers will be presented. When considering cyclic loading, the induced friction forces and contact normal loads are combined so as to determine the effective stiffness and damping of the friction interfaces over a cycle of motion. The harmonic balance method is then used to impose the approximate stiffness and damping of the friction interfaces to a linear structure model of the blade. This approach results in a set of nonlinear algebraic equations that can be solved to yield the forced response of the blade excited by harmonic external forces. The predicted forced response can then be used to optimize a given damper design, namely to determine the dynamic weight at which the maximum reduction of resonant response is obtained. In order to illustrate the capacity of the proposed method and to examine its accuracy, the forced response of a test beam is examined. The prediction is also compared with the results of lab tests to validate the proposed dual-interface friction force model.

Commentary by Dr. Valentin Fuster
1997;():V004T14A009. doi:10.1115/97-GT-021.
FREE TO VIEW

An overset grid method is applied to the solution of single and multi-passage cascade flows with a compressible Navier-Stokes solver. C-type grids around individual blades are overset onto a Cartesian background grid. Overset grids are allowed to move in time relative to each other as prescribed by the oscillatory plunging motion. The overset grid method uses a simple, robust numerical algorithm to localize moving boundary points and to interpolate solution variables across intergrid boundaries. Computational results and comparisons with single/staggered, deforming grid solutions are presented for in- and out-of-phase multi-passage flows through a compressor cascade. Very good agreement is obtained against the deforming grid solutions.

Commentary by Dr. Valentin Fuster
1997;():V004T14A010. doi:10.1115/97-GT-022.
FREE TO VIEW

A probabilistic design code is being developed for high energy disks and rotors used in aircraft gas turbine engines. This code is intended to augment, not replace, the current safe-life approach to the design of these critical components. While the code will ultimately be applicable to a range of disk alloys, initial emphasis has been placed on titanium alloys containing hard alpha defects. The approach involves developing an enhanced defect distribution for hard alpha, obtaining crack initiation data for hard alpha and fatigue crack growth data for three titanium alloys, and integrating this information into a software code that is sufficiently efficient that it can be routinely used for design and life prediction.

Topics: Design , Rotors , Turbines , Aircraft
Commentary by Dr. Valentin Fuster
1997;():V004T14A011. doi:10.1115/97-GT-058.
FREE TO VIEW

The Fast Wavelet Transform (FWT) is a powerful new tool which can be used for vibration analysis and condition monitoring of advanced rotating machinery. The main advantage of wavelet analysis for condition monitoring is that so-called wavelet maps can be produced showing three dimensional plots of amplitude versus frequency and time. This is in contrast to Fast Fourier Transform (FFT) analysis, in which the time domain of the signal is lost. The wavelet maps provide striking visual indications of tiny changes in machine behaviour which cannot be detected in a normal frequency spectrum. This improves the chances of averting catastrophic failures and expands the time window available to take corrective action. Additional advantages of wavelet analysis over FFT analysis include: (1) no requirements for periodicity of the signal, (2) extremely fast computation, (3) the location of patterns in the time domain and (4) an effective detection of high frequency details.

Wavelet analyses of all types are available but user-friendly information is hard to come by and this has a detrimental effect on progress towards practical commercial applications. Thus, the main purpose of this paper is to provide a simple and clear introduction to wavelet analysis and its use in machine condition monitoring. The paper has been written for an audience having some familiarity with spectrum analysis but no prior knowledge of wavelets.

Commentary by Dr. Valentin Fuster
1997;():V004T14A012. doi:10.1115/97-GT-064.
FREE TO VIEW

As the need for specialist decision-making has increased with the volume of data produced by modern monitoring systems, and the current trend towards downsizing and external sourcing (for example, specialist consultant companies) has continued, the demand for computer-based expert systems for automatic machine condition (vibration and process) analysis and diagnosis has intensified.

Various levels of success have been achieved, but most expert systems available today do not reflect the actual reasoning process of a human expert; are inherently obsolete for the continuous learning capability required for dynamic applications; and/or require considerable skills in computer simulation or statistical methods to update the system.

In this paper, new techniques and tools are presented that address the basic elements in the reasoning process of a human expert, and offer solutions to the practical implementation of effective and reliable automatic machine diagnosis. Essential tools for optimum automatic spectrum analysis are first introduced, and then a method presented that allows system results to be automatically qualified and improved upon to reflect actual machine conditions. The paper then introduces neural-network technology as a means of implementing a workable, user-defined knowledge base that can be used to augment the expert system with the user’s own knowledge and experience, and the idiosyncrasies of individual machines.

Commentary by Dr. Valentin Fuster
1997;():V004T14A013. doi:10.1115/97-GT-070.
FREE TO VIEW

The main differences in the calculated behaviour of an unbalanced rotor on two-lobe oil film bearings, in terms of mean statical position, 1 × rev. and 2 × rev. components in steady-state conditions, between the results obtained by means of the linearized and the non linear oil film models are shown.

The errors introduced by the linearized model are not negligible and can affect significantly the accuracy of results of numerical simulations. This may be unacceptable if one uses the measured vibrations in the bearings for identifying the exciting forces on the rotor in a model based diagnostic approach.

Therefore a method is presented which utilises the linear model of the rotor and the non linear oil-film forces in the identification procedure.

Topics: Bearings
Commentary by Dr. Valentin Fuster
1997;():V004T14A014. doi:10.1115/97-GT-075.
FREE TO VIEW

The various design aspects of an industrial gas turbine rotor need special attention to ensure safe and reliable operation of the entire machine. The rotor has to fulfill the following main criteria and tasks: carry the blading, transmit the torque, define the aerodynamic channel geometry, guide the cooling air, provide sealing and bearing surfaces, and exhibit acceptable rotordynamic behavior. Additional criteria linked to the rotor design can be established, i.e., ease of manufacture and assembly, maintainability (where needed), transient blading tip clearances, rotational inertia, balance of axial thrust, and safety against fault conditions (blade loss, short circuit, hot-gas ingestion, etc.) with possibilities for subsequent repair. Industrial practice shows that there are several design approaches possible.

This paper describes the general requirements and how to assess the strength and the dynamic behavior under transient and steady state conditions. Additionally, the individual existing main design principles, i.e., stacked discs with individual centering and tie rod types, and monolithic or welded integral rotor design, are compared. Differences in levels of loading and in the dynamic behavior are outlined. Finally, an assessment of the individual rotor designs under fault conditions is given, and critical areas of the component under such conditions are discussed.

Commentary by Dr. Valentin Fuster
1997;():V004T14A015. doi:10.1115/97-GT-085.
FREE TO VIEW

A mathematical model is developed to analyze the unsteady flow through an harmonically oscillating cascade of airfoils, including separated flow. The model incorporates an inverse integral boundary layer solution with the time–marching Euler analysis NPHASE. An embedded composite grid formulation is incorporated, specifically a deforming C–grid embedded in a Cartesian H–grid, thereby simplifying grid generation. To reduce computational requirements, Fourier series unsteady periodic boundary conditions are implemented. The integral turbulent boundary layer model is closed with steady correlations adopted in a quasi–steady manner. To couple the inviscid and viscous solutions, the viscous effect is modeled in the unsteady Euler solution in a quasi–steady manner by a transpiration boundary condition. An isolated airfoil is used to compare the steady interaction model with experimental data. Then a flat plate cascade is used to verify the unsteady flow solver with linear theory predictions. An experimental unsteady aerodynamics data set of a loaded cascade with separated meanflow executing torsional oscillations compared favorably with the analysis. The code is then utilized to study the effect of flow separation on the unsteady aerodynamics.

Commentary by Dr. Valentin Fuster
1997;():V004T14A016. doi:10.1115/97-GT-086.
FREE TO VIEW

This paper demonstrates reduction of stator unsteady loading due to forced response in a large-scale, low-speed, rotor/stator/rotor axial compressor rig by clocking the downstream rotor. Data from the rotor/stator configuration showed that the stator response due to the upstream vortical disturbance reaches a maximum when the wake impinges against the suction surface immediately downstream of the leading edge. Results from the stator/rotor configuration revealed that the stator response due to the downstream potential disturbance reaches a minimum with a slight time delay after the rotor sweeps pass the stator trailing edge. For the rotor/stator/rotor configuration, with Gap1= 10% chord and Gap2= 30% chord, results showed a 60% reduction in the stator force amplitude by clocking the downstream rotor so that the time occurrence of the maximum force due to the upstream vortical disturbance coincides with that of the minimum force due to the downstream potential disturbance. This is the first time, the authors believe, that beneficial use of flow unsteadiness is definitively demonstrated to reduce the blade unsteady loading.

Topics: Compressors , Rotors , Blades
Commentary by Dr. Valentin Fuster
1997;():V004T14A017. doi:10.1115/97-GT-101.
FREE TO VIEW

A computational approach is described for the rapid and systematic prediction and evaluation of the onset of dynamic stall due to rapid incidence changes or unsteady pitch or plunge motions. The method combines an unsteady, two-dimensional panel code with a two-dimensional boundary layer code. The panel code provides incompressible, inviscid flowfields about arbitrary airfoils undergoing prescribed motions. The boundary layer code computes laminar, transitional and turbulent regimes, with transition onset predicted by Michel’s criterion. Presented results demonstrate that the delay in dynamic stall onset is directly related to the dynamic pressure lag, in agreement with previous Navier-Stokes simulations, but in apparent disagreement with several aspects of the ‘moving wall’ analogy suggested in the past as an explanation for delayed dynamic stall onset.

Commentary by Dr. Valentin Fuster
1997;():V004T14A018. doi:10.1115/97-GT-102.
FREE TO VIEW

During the design process of compressor and turbine blades the investigation of flutter phenomena becomes increasingly important since higher load and better efficiency are desired. As an improvement on the numerical analysis and prediction of unsteady flow through turbomachine cascades with vibrating blades a time accurate Navier Stokes code for S1-stream surfaces SAFES1 is presented within the scope of this paper. To validate the code, numerical results for sub- and transonic test cases of a turbine and a compressor cascade are compared with experimental data. Their good agreement and comparison with Euler calculations show the necessity to take into account viscous effects. To cope with shock waves and areas of separation in laminar or turbulent flow, the fully non linearized Navier Stokes equations are solved using an algebraic turbulence model by Baldwin and Lomax. An approximative upwind flux difference splitting scheme suggested by Roe is implemented. Third order spatial accuracy can be achieved by the MUSCL technique in conjunction with a TVD scheme and a flux limiter by van Albada. By applying either an explicit or an implicit scheme the algorithm can give second order temporal accuracy. The implicit scheme exactly describes the time dependent solution by following a Newton subiteration for every time step.

The blades are discretized in a single passage by a C- or O-type grid. The harmonic motion of the blades is bending or torsion or both simultaneously in a non-rotating or rotating frame of reference. For the chosen mode of oscillation the time dependent axial and circumferential blade forces are determined as well as the resulting moment and damping coefficient. To handle a phaseshift between the motion of the blades a direct store method is used. For the unsteady grid movement a fast grid generation is performed in the core region.

Commentary by Dr. Valentin Fuster
1997;():V004T14A019. doi:10.1115/97-GT-103.
FREE TO VIEW

Axial ribs have been added to gas turbine stationary casings to reduce casing ovalization and allow tighter blade tip clearances to be maintained. Test experience showing blade rub patterns on stator casings due to casing ovalization is presented. Classical solution techniques are presented which demonstrate trends and the relative roundness and attenuation benefits of adding the axial ribs which increase the order of the casing response. The results of finite element studies using 2D solid modeling techniques are compared to the classical solutions. Examples of applications on industrial gas turbine hardware are presented.

Commentary by Dr. Valentin Fuster
1997;():V004T14A020. doi:10.1115/97-GT-104.
FREE TO VIEW

This papers describes the design and development of a touchdown bearing system. The touchdown bearing is one of the important elements in any high speed magnetic bearing system. The touchdown bearing functions to fully off-load or load share with the magnetic bearing and prevent catastrophic system failures. The proper design of this component is a critical step in the design of the total system and was not treated as a stand-alone sub-component. The impact loads which these bearings have to endure is a challenge to the designer. The high temperature environment on this bearing adds additional constraints to the design options which the designer has to work with. A design was produced, which will function under these conditions. The use of silicon nitride balls with sliver coated CPM Rex 20 races will be described. The bearing is mounted on a replaceable journal, which in turn is mounted on the magnetic bearing. Design and test data for use of these bearing materials in a high temperature gas turbine application is included in this paper.

Commentary by Dr. Valentin Fuster
1997;():V004T14A021. doi:10.1115/97-GT-105.
FREE TO VIEW

A complete set of unsteady blade surface pressure measurements is presented for a single turbine blade oscillating in a three dimensional bending mode. Results are provided for five spanwise sections at 10%, 30%, 50%, 70% and 90% of span. Steady blade pressure measurements and five-hole probe traverses at the inlet and exit planes of the test section, are also included.

The test facility operates at low speed and the working section consists of a single turbine blade mounted in a profiled duct. A rigid blade with constant section was used, and a three dimensional bending mode realised by hinging the blade at root and driving the tip section. The low speed and scale of the test facility allowed low oscillation frequencies (5 to 20 Hz) to be employed, in order to match realistic reduced frequencies. This enabled the unsteady blade surface pressure response to be recorded with externally mounted pressure transducers. The validity of this technique is examined. Results from the test facility demonstrate a noticeable three dimensional behaviour of the unsteady flow.

Commentary by Dr. Valentin Fuster
1997;():V004T14A022. doi:10.1115/97-GT-106.
FREE TO VIEW

This study concerns the development of a finite element model to support design improvements in elastomeric seals subject to high temperature and pressure, such as in aircraft engines. Existing finite element codes familiar to the authors do not couple thermal and mechanical fields, nor do they implement thermomechanical contact models suitable for highly deformable materials. Recently, the authors have introduced a thermohyperelastic constitutive model for near-incompressible elastomers. In two subsequent studies, using the constitutive model, a method has been introduced for finite element analysis of coupled thermomechanical response, including boundary contributions due to large deformation and variable contact. A new thermomechanical contact model has also been introduced to accommodate the softness of elastomers. The method has been implemented in a special purpose code which concerns a seal compressed into a well. Several computations are used to validate the code. Simulations of a seal in an idealized geometry indicate rapid pressure increase with increasing compression and temperature.

Commentary by Dr. Valentin Fuster
1997;():V004T14A023. doi:10.1115/97-GT-107.
FREE TO VIEW

Component fatigue life is a major concern in the design of gas turbine Auxiliary Power Units as it directly influences the reliability and life cycle cost of the end product. Accordingly, there is heavy emphasis placed upon designing components which safely maximize their fatigue life. Typical industry practice for managing fatigue has relied on what is commonly referred to as the “safe life approach” in which retirement lives are analytically determined for components and hardware is removed from service before fatigue related failures can occur.

The safe life approach is deterministic in nature since stress analysis results based on minimum geometry, material properties and maximum load are used to set a single life for the component. However, service experience shows that fatigue failures can occur before service life and that actual service lives are distributed over a large range of values as a result of variables not accounted for by deterministic methods.

In order to better achieve the goal of minimizing product life cycle costs while recognizing the variable nature of fatigue lives, Sundstrand Aerospace (SA) has developed a Life Management Plan (LMP) which includes probabilistic methods to augment the company’s standard safe life methodology.

Sundstrand’s LMP builds on the safe life methodology by using statistical distributions along with Monte Carlo simulations to predict initial component cracking rates. These initial predictions are used to guide an inspection program which provides actual cracking data. As this inspection data base grows, the initial simulation is modified to include the inspection data and the predicted failure rates are updated. This provides Sundstrand with a tool to manage failure risk in the field as well as to provide early warning of negative trends.

This paper will discuss how Sundstrand’s LMP was devised and implemented as well as what lessons have been learned and what changes are planned for future incorporation. A case history will be cited to illustrate how the LMP has worked, comparing predictions to actual experience.

Commentary by Dr. Valentin Fuster
1997;():V004T14A024. doi:10.1115/97-GT-108.
FREE TO VIEW

A typical hot structural component within an engine such as composite combustor liner is computationally simulated and probabilistically evaluated in view of the numerous uncertainties associated with the structural, material, and thermo-mechanical load variables (primitive variables) that describe the combustor. The combustor is evaluated for local stresses. Results show that the scatter in the combined stress near the support is significantly dependent upon the uncertainties in the through thickness thermal gradients, the liner material thickness, the coefficient of thermal expansion, and the axial and both the axial and shear moduli.

Commentary by Dr. Valentin Fuster
1997;():V004T14A025. doi:10.1115/97-GT-109.
FREE TO VIEW

The design and fatigue life analysis of many fabricated sheet metal components which operate in a hot gas path environment is made uncertain by the difficulty in predicting the gas dynamic forces to which the component will be subjected. This is particularly true for ducts which are exposed to minimal static loading such as inter-turbine ducts, which have little pressure drop across them but are subjected to highly non-uniform flow from the upstream turbine exit. Finite element analysis and static testing has revealed — that certain modal frequencies respond much more significantly than others and where these coincide with gas dynamic frequencies a resonance is set up and rapid and catastrophic failure will occur.

Through a combination of analysis and engine testing it has been possible to identify the damaging gas dynamics and modal frequencies. This has enabled rules to be developed which assist in the design of such ducts to avoid high cycle fatigue failure.

This paper describes the development of the design process and by means of a typical example shows its application and effect on the design of an inter-turbine duct.

Commentary by Dr. Valentin Fuster
1997;():V004T14A026. doi:10.1115/97-GT-110.
FREE TO VIEW

The reliability of gas turbine driven compressor units is a major factor in the ability of a pipeline system to satisfy delivery or service requirements. However, despite it’s importance, unit reliability is not typically considered in design or operations. At Nova Gas Transmission Ltd (NGTL), a reliability evaluation methodology, referred to as Sustainable Capacity, has been developed to quantify service expectations through the explicit consideration of the effects of gas turbine reliability. The methodology, which has been adapted from one widely used in the electric power industry, is based on the representation of the load and capacity processes with statistical models. Time-series representing simultaneous load and capacity profiles are generated and evaluated to determine reliability expectations. A cost-benefit methodology which recognizes the variation in the value attributed to pipeline reliability by individual customers is also used to assess the economic efficiency of any proposed system modification. Although developed as a system design tool, the Sustainable Capacity methodology can also be used to quantify the value of gas turbine maintenance strategies provided the effect on unit reliability can be quantified.

Commentary by Dr. Valentin Fuster
1997;():V004T14A027. doi:10.1115/97-GT-111.
FREE TO VIEW

Turbomachine design is directed at performance, accomplished with blading operating near to aeromechanics stability limits. Currently, linearized friction damper models are used to predict these stability limits. However, during engine development, these linearized solutions are often found to be inadequate. Thus, for accurate blade row stability predictions, the nonlinear effects of friction dampers must be considered. This paper is directed at investigating the effects of friction damper nonlinearities on the response of a linear system and also the applicability of linear and nonlinear aeromechanics analyses. This is accomplished utilizing both linear and nonlinear aeromechanics models to analyze a nonlinear aeromechanics system with only one nonlinearity — friction damping. The linearized friction damper results are shown to be sufficient for preliminary design and for operating conditions away from stability limits. However, nonlinear friction dampers must be considered for final design or optimization near stability limits.

Topics: Friction , Dampers , Airfoils
Commentary by Dr. Valentin Fuster
1997;():V004T14A028. doi:10.1115/97-GT-112.
FREE TO VIEW

Active magnetic bearings (AMBs) while offering many unique design and operational opportunities for advanced rotor systems, require some form of backup or auxiliary bearing in the event of a component failure or the onset of high transient loads. A zero clearance auxiliary bearing (ZCAB) has recently been conceived and a prototype system tested. The ZCAB presented in this paper uses a series of interconnected rollers to surround a shaft. In the open position, a clearance exists between the ZCAB rollers and the shaft. When the shaft drops on the ZCAB due to either an AMB failure or transient shock, the rollers move circumferentially and radially inward to eliminate the clearance and re-center the shaft. Besides centering the shaft, the law shaft-to-ZCAB traction coefficient and composite support dynamic characteristics eliminate the possibility of backward whirl. This paper presents the design methodology used, results of an analytical design study, including time transient analysis, as well as preliminary feasibility prototype testing under simulated AMB failure and transient shock conditions. The test rotor was supported by a rolling element bearing at one end and an integrated magnetic bearing/ZCAB support system at the other end. Both rotor drop and shock tests were performed with this configuration. Experimental results under simulated AMB failure and transient shock conditions demonstrated successful operation of the ZCAB.

Commentary by Dr. Valentin Fuster
1997;():V004T14A029. doi:10.1115/97-GT-113.
FREE TO VIEW

Equations of motion are presented for a two degree of freedom system representing a rotor suspended with an active magnetic bearing. The rotor is subjected to both a rotating unbalance force and a steady load, which is balanced by differential bias in the vertical magnet pair. A coupling between the coordinate directions resulting from the spatial arrangement of magnets is included in the equations of motion, which renders them nonlinear even though the closed loop control on each axis is linear. Results of a path-following analysis of the equations show that the addition of a steady load has several effects. Compared with previous results for an unloaded system, the response is now asymmetric due to unequal stiffnesses, and the region of unacceptable high amplitude response extends over a larger frequency range. In addition, the frequency ranges over which multiple stable solutions coexist is expanded. New results for the unloaded case are also presented that show domains of attraction to large amplitude stable response due to an applied impulse. There is a strong dependence of the domains of attraction on the timing of the impulse.

Commentary by Dr. Valentin Fuster
1997;():V004T14A030. doi:10.1115/97-GT-114.
FREE TO VIEW

Magnetic bearings, unlike traditional mechanical bearings, consist of a series of components when mated together, form a stabilized system. A series of four papers will summaries the results carried out at Draper in the development of a high temperature magnetic bearing suspension system for a gas turbine application. Part I [1] will document our approach for rotordynamics modeling of the turbine shaft and the development of models for use in our simulation programs. Part II [2] documents the simulation efforts and the control system which resulted from this effort. Parts III and IV [3] document the design and fabrication of the magnetic bearing actuators and the auxiliary touchdown bearings.

This paper, part III, deals with the design of the high temperature magnetic bearing actuators. Two radial and one axial magnetic bearing actuator were designed to meet the requirements for the turbine application. No bias coils are included in these design. The biasing flux is provided by current from the control power amplifiers. All the coils are made from ceramic coated copper wire and are terminated to high temperature connectors designed into the actuators. The new high strength Hiperco 50 HS material was chosen for the rotor lamination material for the radial bearings. A customized heat treatment process for this material in a high vacuum environment was developed to insure the maximum strength was obtained with the maximum magnetic properties. High temperature ceramic coated copper wire and bonding and potting material used for the coil assembly were tested up to 650 degrees C without failures.

Commentary by Dr. Valentin Fuster
1997;():V004T14A031. doi:10.1115/97-GT-115.
FREE TO VIEW

A new friction damper has been designed by Volvo Aero Corporation. It is used in the high pressure turbine stage of a turbojet engine. The objective of this paper was to find the optimal weight of the new damper that minimizes the blade response amplitude for six and nine engine order excitation and to compare the new damper design with that currently used. Another objective was to compare how well simulation results agree with experimental results from spin pit tests. Simulations were made with a damper model that incorporates the possibility of both micro- and macro-slip in the blade-damper contact interface. Turbine blades were modeled using finite element beam elements. Experimental data were provided from spin pit tests with a completely bladed high pressure turbine rotor. Results show that the simulation model can be used to give qualitative results but has to be further developed to incorporate mistuning effects and coupled modes of vibration for the blade. The spin pit test shows that the new damper design is more efficient in reducing resonance stresses than the old design. It was not possible to see if simulations predict the right optimal damper weight by comparing with experimental data because the rotor could not be excited up to the design point.

Commentary by Dr. Valentin Fuster
1997;():V004T14A032. doi:10.1115/97-GT-116.
FREE TO VIEW

During the operation process in many types of fluid flow machines the rotating blades pass through various resonances e.g. during run-up or run-down or other transient conditions. Therefore, for the high cycle fatigue problem of the blades it might be important to consider the transient vibratory response of the blades during these passages through resonance and to get knowledge about the occuring maximum vibratory stresses.

In the paper, approximate formulas are presented which allow the estimation of the maximum transient response of the blades. Thereby, the influence of the change of the natural frequencies due to the increasing or decreasing centrifugal force field during the run-up or run-down, respectively, is taken into consideration. Basically, the approximate formulas are based on a linear change of the natural frequencies versus time and on a linear viscous type of damping. Extensions to account for parabolic changes which are more realistic for centrifugal effects and for non-linear damping models e.g. friction damping or turbulence damping are discussed.

Commentary by Dr. Valentin Fuster
1997;():V004T14A033. doi:10.1115/97-GT-186.
FREE TO VIEW

A computational method is presented for predicting the unsteady aerodynamic response of a vibrating blade row which is part of a multistage turbomachine. Most current unsteady aerodynamic theories model a single blade row isolated in an infinitely long duct. This assumption neglects the potentially important influence of neighboring blade rows. The present ‘coupled mode’ analysis is an elegant and computationally efficient method for modelling neighboring blade row effects. Using this approach, the coupling between blade rows is modelled using a subset of the so-called spinning modes, i.e. pressure, vorticity, and entropy waves which propagate between the blade rows. The blade rows themselves are represented by reflection and transmission coefficients. These coefficients describe how spinning modes interact with, and are scattered by, a given blade row. The coefficients can be calculated using any standard isolated blade row model; here we use a linearized full potential flow model together with rapid distortion theory to account for incident vortical gusts. The isolated blade row reflection and transmission coefficients, inter-row coupling relationships, and appropriate boundary conditions are all assembled into a small sparse linear system of equations which describes the unsteady multistage flow. A number of numerical examples are presented to validate the method and to demonstrate the profound influence of neighboring blade rows on the aerodynamic damping of a cascade of vibrating airfoils.

Commentary by Dr. Valentin Fuster
1997;():V004T14A034. doi:10.1115/97-GT-187.
FREE TO VIEW

This paper is concerned with the dynamic behavior of geared rotor systems supported by squeeze film dampers, wherein coupled bending torsion vibrations occur. Considering the imbalance forces and gravity, it is shown that geared rotors exhibit chaotic behavior due to non linearity of damper forces. The route to chaos in such systems is established. In geared rotor systems, it is shown that torsional excitation can induce lateral vibrations. It is shown that squeeze film dampers can suppress large amplitudes of whirl arising out of torsional excitation.

Topics: Rotors
Commentary by Dr. Valentin Fuster
1997;():V004T14A035. doi:10.1115/97-GT-189.
FREE TO VIEW

Friction dampers are installed underneath the blade platforms to improve the reliability. Because of centrifugal forces the dampers are pressed onto the platforms. Due to dry friction and the relative motion between blades and dampers, energy is dissipated, which results in a reduction of blade vibration amplitudes.

The geometry of the contact is in many cases like a Hertzian line contact. A three-dimensional motion of the blades results in a two-dimensional motion of one contact line of the friction dampers in the contact plane. An experiment with one friction damper between two blades is used to verify the two-dimensional contact model including microslip.

By optimizing the friction dampers masses, the best damping effects are obtained.

Finally, different methods are shown to calculate the envelope of a three-dimensional response of a detuned bladed disk assembly (V84.3-4th-stage turbine blade) with friction dampers.

Commentary by Dr. Valentin Fuster
1997;():V004T14A036. doi:10.1115/97-GT-190.
FREE TO VIEW

Most of the turbomachine manufacturers are using a non-contact measurement system to measure blade vibration by two probes located in the casing radially outside a blade row. All blades of a row can be observed by this system, therefore it may be considered to be an ideal solution for measuring the vibration response of mistuned blades which have quite different vibration levels from blade to blade in resonance conditions.

By a simple calculation model a non-contact blade vibration measurement was simulated. It was found that for mistuned coupled blades the non-contact measurement results show some deviations from the actual vibration amplitudes. Different mistuning conditions have been investigated by a parameter study to estimate the inherent measurement error of the non-contact principle. Measurements at a circumferentially coupled gas turbine blading performed by a non-contact system and by strain gauges did also show deviations in the results obtained by the two measurement methods.

Commentary by Dr. Valentin Fuster
1997;():V004T14A037. doi:10.1115/97-GT-191.
FREE TO VIEW

Precise non-linear aeroelastic modeling of shrouded bladed-disc assemblies is generally beyond present capacities and analyses often assume that the behavior of the coupled system remains linear and retains a cyclic symmetrical property. In this paper, several models of shrouded assemblies, in the particular case of fully slipping interfaces, are examined and compared. Considering the cyclic symmetrical property of the structure, only the model where shroud segments can slip and propagation relations are applied in the direction normal to the interface plane, should be used. A reduced model based on a direct discretisation of the whole assembly is presented and validated. The application is based on a first stage shrouded fan. The influence of varying the interface shroud angle is examined in terms of frequency, mode shape and aeroelastic damping.

Topics: Design , Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A038. doi:10.1115/97-GT-201.
FREE TO VIEW

The governing equations of motion for a rotating flexible blade-rigid disk-flexible shaft system are derived. The bladed disk is attached at one end of an asymmetric shaft with uniformly distributed mass, mass moment of inertia, and stiffness. The shaft is held by two isotropic supports; one at the far end from the bladed disk, modeled by two translational and two rotational springs, and an intermediate support, modeled by two translational springs only. The effect of shaft asymmetry on the dynamics behavior of the rotating bladed disk shaft system is examined over a wide range of rotational speed, and for different combinations of springs’ stiffness, which determines the type of shaft supports. The cantilever, and the simply supported shaft with an over hang can be looked upon as special cases of the described above shaft configuration, since the former is obtained by assigning large stiffness for both translational and rotational springs at the end support, and zero spring stiffness at the intermediate one, whereas the latter is obtained by assigning large stiffness for the translational springs at both supports and zero stiffness for the rotational springs. Stability boundaries are calculated, and presented in terms of shaft asymmetry and rotor speed for given bearing stiffness.

Topics: Stability , Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A039. doi:10.1115/97-GT-229.
FREE TO VIEW

An quasi three-dimensional time-linearized Euler method has been developed to compute unsteady flows around oscillating blades. In the baseline method, unsteady flow is decomposed into a steady flow plus a linear harmonically varying unsteady flow. Both the steady flow equations and the unsteady perturbation equations are solved using a pseudo time-marching method. Based upon this method, a novel nonlinear harmonic Euler method has been developed. Due to the nonlinearity of the aerodynamic governing equations, time-averaging generates extra “unsteady stress” terms. These nonlinear effects are included by a strongly coupled approach between the perturbation equations and the time-averaged equations. Numerical results demonstrate that nonlinear effects are very effectively modelled by the nonlinear harmonic method.

Commentary by Dr. Valentin Fuster
1997;():V004T14A040. doi:10.1115/97-GT-230.
FREE TO VIEW

A transient, nonlinear analysis was developed and used to study the effect of shock machine testing on a gas turbine simulator supported by homopolar, permanent magnet bias magnetic bearings. The magnetic bearing nonlinearities modeled included saturation effects, clearance effects, and integrator and current limits. Free vertical travel of the shock machine anvil table supporting the simulator was also modeled. The magnetic bearing model was coupled to characteristic matrix based models of the rotor and support system and integrated to produce a time simulation of system performance. The results indicate saturation of the magnetic bearing for brief periods following impacts significant enough to exceed design load capacity, followed by recovery to stable operation in less than a second. The analysis was used to evaluate sizing for the magnetic bearing and backup bearing systems and to evaluate the control system strategy.

Commentary by Dr. Valentin Fuster
1997;():V004T14A041. doi:10.1115/97-GT-231.
FREE TO VIEW

State feedback controller designs allow for pole-placement in an effective manner, but reduction of static offset is difficult. On the other hand, classical control methodology allows for the increase of system type and the elimination of static offset. An integrally augmented state feedback controller provides the benefits of standard feedback designs while allowing for the elimination of static offsets (through the increase of system type). Static offset is a particular problem with magnetic bearing supported rotor systems, in that gravitational effects, current biasing, and operational loading tend to exacerbate this problem. In order to assess the effectiveness of this technique, an integrally augmented state feedback controller is developed, implemented, and tested for a magnetic bearing supported rotor system. Results for several selected configurations are presented and compared. Some conclusions and recommendations concerning the effectiveness of integrally augmented state feedback controller designs are presented.

Commentary by Dr. Valentin Fuster
1997;():V004T14A042. doi:10.1115/97-GT-232.
FREE TO VIEW

The flow through labyrinth seals of turbomachinery generates forces which can cause self-excited vibrations of the rotor above the stability limit. The stability limit is reached at a specific rotating speed or power. The continuous growth in of power density and rotating speed necessitates an exact prediction of the stability limit of turbomachinery.

Usually the seal forces are described with dynamic coefficients. A new, easy-to-handle identification procedure uses the stability behavior of a flexible rotor to determine the dynamic coefficients. Systematic measurements with a great number of labyrinth seal geometries lead to reasonable results and demonstrate the accuracy and sensitivity of the procedure. A comparison of the various methods used to minimize the excitation indicates which seal is more stable and will thus improve the dynamic behavior of the rotor.

Commentary by Dr. Valentin Fuster
1997;():V004T14A043. doi:10.1115/97-GT-233.
FREE TO VIEW

The efficiency and the characteristics of unshrouded centrifugal compressors depend on the thermodynamic and aerodynamic design of the stage and especially of the impeller. Both, the efficiency and the characteristics, can strongly be influenced by a variation of the geometry of the blade structure.

The design of the leading edge depends on the given inlet flow angle distribution and the outlet part of the blade is fixed by the desired outlet angle distribution. Due to the manufacturing process, the selected material, the solidity and the blade vibration, the restrictions and limitations have to be taken into consideration during the aerodynamic optimization process of the blade geometry.

By means of Finite Element Code computations of the solidity and the vibration behaviour the possibilities, the restrictions and the limits of a modern aerodynamic design especially in the inlet and outlet part of the impeller blades were investigated.

The results of this work could demonstrate the importance of integrating the Finite Element Code computation in the design process — not only to see the limits, but also to try and realize new ideas of design optimization.

Commentary by Dr. Valentin Fuster
1997;():V004T14A044. doi:10.1115/97-GT-234.
FREE TO VIEW

A method is presented for reducing the lateral response of an imbalanced rotor accelerating or decelerating through its first lateral bending critical speed by using a variable acceleration rate. A lumped parameter model along with a numerical integration scheme is used to simulate the response of a simply supported, single disk rotor during fast acceleration and deceleration through critical speed. The results indicate that the maximum response and/or the total vibrational energy of a rotor passing through the critical speed can be reduced significantly by using a variable acceleration schedule. That is, reducing the acceleration rate after the nominal critical speed is passed. These predictions were verified experimentally for a single disk rotor.

Topics: Rotors , Vibration
Commentary by Dr. Valentin Fuster
1997;():V004T14A045. doi:10.1115/97-GT-235.
FREE TO VIEW

The bending mode flutter of a modern transonic fan has been studied using a quasi-3D viscous unsteady CFD code. The type of flutter in this research is that of a highly loaded blade with a tip relative Mach number just above unity, commonly referred to as transonic stall flutter. This type of flutter is often encountered in modern wide chord fans without a part span shroud.

The CFD simulation uses an upwinding scheme with Roe’s 3rd-order flux differencing, and Johnson and King’s turbulence model with the later modification due to Johnson and Coakley. A dynamic transition point model is developed using the en method and Schubauer and Klebanoff’s experimental data.

The calculations of the flow in this fan reveal that the source of the flutter of IHI transonic fan is an oscillation of the passage shock, rather than a stall. As the blade loading increases, the passage shock moves forward. Just before the passage shock unstarts, the stability of the passage shock decreases, and a small blade vibration causes the shock to oscillate with a large amplitude between unstarted and started positions. The dominant component of the blade excitation force is due to the foot of the oscillating passage shock on the blade pressure surface.

Commentary by Dr. Valentin Fuster
1997;():V004T14A046. doi:10.1115/97-GT-236.
FREE TO VIEW

A thermomechanical fatigue (TMF) cycle, intended to represent aero-engine blade working conditions, was selected for carrying out tests on uncoated and aluminide coated SRR99 samples until test piece failure. Optical images of the surface of test pieces were collected during testing to monitor surface crack initiation and accompanying transformations. Using these images, surface changes were quantified as a function of time. Post test, each sample was taken through an incremental polishing procedure to allow damage in the tested material to be studied as a function of depth, using optical microscope based quantitative metallography and scanning electron microscopy (SEM). The relationship between the observed surface changes and the damage built-up subsurface was examined. Differences in damage density on the surface and subsurface planes between coated and uncoated samples could accommodate the observed life reductions.

Commentary by Dr. Valentin Fuster
1997;():V004T14A047. doi:10.1115/97-GT-237.
FREE TO VIEW

The near-tip fields of small Case B cracks in power-law hardening materials are investigated under generalized plane-strain and general yielding conditions by finite element analyses. The results for two different crack orientations are examined and compared. The results indicate that the plastic deformation patterns near the tips of the cracks of two different orientations are remarkably similar in terms of the global coordinates. The results of the J integral from the finite element analyses are used to correlate to a fatigue crack growth criterion for Case B cracks. The trends of constant ΔJ contours on the Γ-plane for two cracks of different orientations are virtually the same. Further, the trends are compared reasonably well with those of the experimental results of constant fatigue life and constant fatigue crack growth rate.

Commentary by Dr. Valentin Fuster
1997;():V004T14A048. doi:10.1115/97-GT-238.
FREE TO VIEW

Stress ratios are traditionally used to infer maximum stresses in blades from strain measurements. This method may not be applicable to all the modes of modern low aspect ratio (LAR) blades since LAR blades often have high frequency “tip” modes which are so closely spaced that slight changes in structural properties can cause significant mode shape changes.

In this paper, it is first shown that the actual tip modes of a LAR blade can be well approximated as linear combinations of the “nominal” modes. The stress field of the blade can then be estimated by calculating the modal content from multiple strain measurements. However, since placement and gage inaccuracies can introduce significant errors in the calculation, it is necessary to find the optimal gage placement which minimizes the error in estimated stresses. An error estimate using “norms” and an efficient optimization method are developed for this purpose.

Commentary by Dr. Valentin Fuster
1997;():V004T14A049. doi:10.1115/97-GT-249.
FREE TO VIEW

The effect of a single-edge fatigue crack on the torsional vibration of shafts is investigated. A generalized variational principle is used to formulate the equation of motion and associated boundary conditions for the free vibration of a nonrotating shaft with a fatigue crack of arbitrary size and location. The fatigue crack is introduced in the form of a single-edge crack. The stress and strain of the cracked shaft are determined by introducing a crack function and a displacement function into the shaft’s compatibility relations. The crack function is designed to have the maximum value at the cracked section and decay exponentially away from the crack along the shaft’s longitudinal direction. A displacement function is constructed to modify the in-plane displacement and its slope near the single-edge crack. The natural response of the free-free shaft is calculated through a Galerkin procedure. The results indicated a clear change in the natural frequencies of the cracked nonrotating shaft.

Commentary by Dr. Valentin Fuster
1997;():V004T14A050. doi:10.1115/97-GT-250.
FREE TO VIEW

Turbine blades and vanes operate in a hostile environment which leads to deterioration of these components overtime. This paper describes detailed calculations to predict the vibratory response of a high-pressure turbine blade due to the excitation produced by a single distressed upstream vane in a modern turbofan engine. The approach includes detailed computational fluid dynamics (CFD) analysis of the steady flow field produced by the distressed vane, Fourier decomposition of the flow variables to determine the harmonic content, unsteady CFD analysis to determine the resulting vibratory response of the blade, and crack propagation analysis to determine blade life. Predictions of vibratory stress and threshold crack size are summarized as functions of vane distress level. The results, which indicate that this type of vane distress can indeed be a significant excitation source for the blades, are shown to be in good agreement with engine experience. The method provides, for the first time, a quantitative approach to setting limits for acceptable levels of vane distress in the field.

Topics: Turbine blades
Commentary by Dr. Valentin Fuster
1997;():V004T14A051. doi:10.1115/97-GT-275.
FREE TO VIEW

Matrix cracking in ceramic matrix composites with fine grained fibers at high temperatures will be governed by fiber creep, as relaxation of the fibers eliminates crack tip shielding. Using a time dependent bridging law which describes the effect of creeping fibers bridging a crack in an elastic matrix, crack growth initiation and history have been modeled. For a stationary crack, crack tip stress intensity factors as a function of time are presented to predict incubation times before subcritical crack growth. Two crack growth studies are reviewed: a constant velocity approximation for small-scale bridging, and a complete velocity history analysis which can be used to predict crack length as a function of time. The predictions are summarized and discussed in terms of identifying various regimes of crack growth initiation, subcritical growth, and catastrophic matrix cracking.

Commentary by Dr. Valentin Fuster
1997;():V004T14A052. doi:10.1115/97-GT-299.
FREE TO VIEW

A procedure for the vibration analysis of rotating cambered helicoidal blades using curved shell finite element based on Mindlin theory is given. The strain displacement relations of Gol den’veizer are used in the development of a thick doubly curved helicoidal shell element including initial in-plane stress effect. The method is applied to rotating cambered pretwisted blades.

Commentary by Dr. Valentin Fuster
1997;():V004T14A053. doi:10.1115/97-GT-359.
FREE TO VIEW

A high cycle fatigue failure of a low pressure turbine blade was investigated. Strain gauge tests of a running engine indicated a high dynamic response of the blade at the nozzle passing frequency. This could be attributed to the excitation of a bladed disc mode of vibration. A Finite Element analysis of the low pressure turbine blades and discs, together with bench testing of the complete structure, confirmed the existence of a high frequency 2nd Nodal Diameter mode of vibration. The levels of dynamic strain determined through strain gauge tests were found to be sufficient enough to explain the failure at the given location.

Having understood the problem, the situation was resolved through the use of Finite Element analysis with a short term modification to the original blade aerofoil to prevent the mode from being excited.

An aero/mechanical re-design of both the low pressure turbine rotor and the stator was undertaken to resolve the problem by both returning the blade to avoid high frequency excitation, and also by reducing the forcing effect of the nozzle passing frequency. The new design has been validated through strain gauge tests and endurance tests. A further improvement in performance was also obtained.

Commentary by Dr. Valentin Fuster
1997;():V004T14A054. doi:10.1115/97-GT-391.
FREE TO VIEW

In thermal barrier coatings (TBC) residual stresses develop during cool down from processing temperature due to the thermal expansion mismatch between the different layers (substrate, bond coat, and TBC). These residual stresses can initiate micro cracks at the bond coat/TBC interface and can lead to debonding at the bond coat/TBC interface. The effect of voids or crack like flaws at the interface can be responsible for initiating debonding and accelerate the oxidation process. Effect of oxide layer growth between bond coat and ceramic layer (TBC) can be modeled as volume increase. In this work we represent this change in volume as an induced pressure across the interface. Mixed-mode fracture analysis of a thin circular delamination in an-axisymmetrically multi layer circular plate is developed. Geometrical nonlinearity is included in the analysis, since we have a large deflection case. The elastic deformation problem of a circular plate subjected to a clamped boundary condition at the edge of the delamination, an out of plane pressure load, and a compressive stress due to thermal mismatch between different layers, was solved numerically using a Rayleigh-Ritz method. The strain energy release rate was evaluated by means of the path-independent M-integral. The numerical results of this problem based on the energy method were verified using finite element method. Both methods correlate well in predicting the energy release rate for Mode I and Mode II, deflection, and postbuckling solutions. The energy release rates G, for both Mode I and Mode II using virtual crack extension method were evaluated. The specimen was cooled down from processing temperature of 1000 °C to 0 °C. The variation of the properties as a function of temperature was used for analysis. It was found that the use of temperature dependent properties in contrast to constant properties provides significantly different values of J-integral and G.

Commentary by Dr. Valentin Fuster
1997;():V004T14A055. doi:10.1115/97-GT-400.
FREE TO VIEW

Nowadays, the tilting-pad journal bearings are submitted to more and more severe operating conditions. The aim of this work is to study the thermal and mechanical behavior of the bearing during the transient period from an initial steady-state to a final steady-state (periodic). In order to study the behavior of this kind of bearing under dynamic loading (Fdyn) due to a blade loss, a nonlinear analysis, including local thermal effects, realistic boundary conditions and bearing solid deformations (TEHD analysis) is realized. After a comparison between theoretical results obtained with four models (ISO, ADI, THD and TEHD) and experimental data under steady-state operating conditions (static load Ws), the evolution of the main characteristics for three different cases of the dynamic load (Fdyn/Ws<1, Fdyn/Ws=1 and Fdyn/Ws>1) is discussed. The influence of the transient period on the minimum film thickness, the maximum pressure, the maximum temperature and the shaft orbit is presented. The final steady-state is obtained a long time after the appearance of a dynamic load.

Commentary by Dr. Valentin Fuster
1997;():V004T14A056. doi:10.1115/97-GT-404.
FREE TO VIEW

In this paper, a novel approach to determine reliable estimates of the moments of the steady state resonant response of a randomly mistuned bladed disk is presented and the use of these moments to accurately predict the corresponding distribution of the amplitude of blade vibration is described. The estimation of the moments of the response is accomplished first by relying on a “joint cumulant closure” strategy that expresses higher order moments in terms of lower order ones. A simple modeling of the error terms of these approximations is also suggested that allows the determination of an improved, or accelerated, estimate of the required moments. The evaluation of the distribution of the amplitude of blade response is then accomplished by matching the moments computed by the cumulant closure with those derived from a three-parameter model recently derived. A first order approximation of the moments obtained for a simple structural model of a bladed disk yields a new parameter that can be used as a measure of the localization of the forced response. Then, numerical results demonstrate that the method provides extremely accurate estimates of the moments for all levels of structural coupling which in turn lead to a description of the amplitude of blade response that closely matches simulation results. Finally, a comparison with existing perturbation techniques clearly shows the increased accuracy obtained with the proposed joint cumulant closure formulation.

Topics: Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A057. doi:10.1115/97-GT-410.
FREE TO VIEW

An investigation of thermal barrier coatings on a metal substrate was conducted when the assembly was subjected to both thermal beating and mechanical edge loads generated by interference in adjoining expansion gaps. Both finite element and closed form solution models were developed and compared. The results of the analyses predict that the application of both thermal and mechanical edge loads on the edge of ceramic/metal composites can produce severe local edge spallation in the ceramic coating when an inadequate expansion gap is provided.

Commentary by Dr. Valentin Fuster
1997;():V004T14A058. doi:10.1115/97-GT-444.
FREE TO VIEW

Axial compressors have inherently unsteady flow fields because of relative motion between rotor and statnr airfnils. This relative motion leads to viscous and inviscid (potential) interactions between blade rows. As the number of stages increases in a turbomachine, the buildup of convected wakes can lead in progressively more complex wake/wake and wake/airfnil interactions. Variations in the relative circumferential positions of stators or rotors can change these interactions, leading to different unsteady forcing functions on airfoils and different compressor efficiencies. The current study uses an unsteady, two-dimensional thin-layer Navier-Stokes zonal approach to investigate the unsteady aerodynamics of stator clocking in a low-speed 2 ½-stage compressor. Relative motion between rotors and stators is made possible by the use of systems of patched and overlaid grids. Results include surface pressures instantaneous forces and efficiencies for a 2 ½-stage compressor configuration.

Commentary by Dr. Valentin Fuster
1997;():V004T14A059. doi:10.1115/97-GT-445.
FREE TO VIEW

A simple analytical model is derived for the prediction of time dependent deformation and damage response of metal matrix composites under fiber direction loading. The model can be used in conjunction with a number of viscoplastic constitutive models to describe the matrix material behavior. Damage in the form of progressive fiber fractures is incorporated using a mechanistic approach. The criterion for fiber fractures can be based on statistical information on fiber strength. When used in conjunction with a prescribed failure condition for a composite, the model provides a means for predicting composite life under general thermomechanical load conditions. Based on comparison of results with detailed finite element analyses and with laboratory test data, it appears that the simple model provides reasonably accurate predictions.

Commentary by Dr. Valentin Fuster
1997;():V004T14A060. doi:10.1115/97-GT-446.
FREE TO VIEW

The hot stage components of a large industrial gas turbine engine operate under extremely arduous conditions and as a result have a finite life. In particular, the first stage blades and vanes are expensive to replace and there are strong financial incentives to maximise their useful life and to refurbish them, where possible, to extend that life. In order to achieve this life extension, it is essential to have a detailed knowledge of the conditions under which the blades operate and of their responses to those conditions.

Detailed thermal and stress analyses have been carried out on the first stage blades and vanes of a large gas turbine, operating in a combined cycle power generation plant. The boundary conditions for the analysis were taken from an aerothermal analysis of the engine. From these, the finite element method was used to compute temperature distributions and then stress distributions in both the blades and vanes. Consideration was given to the effect of thermal transients and also to the effect of creep in allowing some redistribution of the thermal stresses.

The resulting stress and temperature distributions were used to estimate damage accumulation in the material due to fatigue, creep and, where appropriate, oxidation. These damage estimates have then been applied to predict crack initiation. Estimates have also been made of likely crack growth behaviour. Damage predictions have been compared with observations made on blades removed from service. The damage accumulation has implications for the operation, maintenance and blade replacement strategies for the engine.

Topics: Gas turbines , Blades
Commentary by Dr. Valentin Fuster
1997;():V004T14A061. doi:10.1115/97-GT-447.
FREE TO VIEW

Satisfactory means for evaluating the vibration behaviour of gas turbine installations where it is impractical to obtain a satisfactory model of the support structure and where it has natural frequencies in the operating range are not yet available. This paper investigates the feasibility of identifying the modal parameters of a flexibly supported rigid casing to predict the unbalance response of the contained rotor. Assumed known are the location of the principal axes of inertia of the casing. Numerical experiments demonstrate that the identification procedure works well even with a measurement accuracy of only two significant digits. Thus, the procedure is expected to be practically feasible, though no sensitivity studies have yet been carried out to determine the limitations imposed by the assumed location of the principal axes.

Commentary by Dr. Valentin Fuster
1997;():V004T14A062. doi:10.1115/97-GT-448.
FREE TO VIEW

This paper suggests that the subsynchronous “instability” observed in many high speed, high performance turbomachines while operating in the supercritical speed range may in some cases be a stable form of lightly damped vibrations. They could be excited by low frequency process forces due to unsteady flow conditions. The non-linearity in the mass, stiffness or damping of the system may have provided a coupling or frequency transformation between the excitation forces and the subsynchronous vibrations. Depending on the kind of non-linear characteristics the critical speeds as defined for a linear system may become regions of “instability”. The degree of non-linearity of the bearing-seal-rotor system has an influence on the sensitivity of the machine to subsynchronous vibrations. Some forcing mechanisms are presented, including non-identical rotor blades, inlet flow distortion and rotating stall. The effect on response of mode shape, internal shaft rotatory damping and frequency dependence of bearing damping at subsynchronous frequencies are discussed. It is recommended that the unsteady fluid dynamic forces, together with the effects of non-linear dynamic characteristics be further investigated to provide more experimental evidence for this hypothesis.

Commentary by Dr. Valentin Fuster
1997;():V004T14A063. doi:10.1115/97-GT-449.
FREE TO VIEW

Measurements on a large rotor dynamic test rig in the Turbomachinery Laboratory with a concrete and steel foundation have shown some critical speeds which can only be predicted if the foundation is included in the model. Mobility transfer function measurements on the foundation were made to obtain the required parameters for including the foundation in the rotordynamic computer model. These measurements eliminate the necessity for building a large finite element model of the foundation structure and allow the use of fast transfer matrix codes for modeling rotor bearing systems which have strong foundation participation. A method first described by Nicholas was used to get a set of equivalent bearing coefficients from a combination of the foundation parameters with the true bearing coefficients. Measurements made by the present authors show that the modal damping of the foundation can be obtained easily from the mobility curves, and that the improved critical speed predictions can be obtained in some cases by using only one foundation mode to calculate the equivalent bearing coefficients.

Topics: Rotors
Commentary by Dr. Valentin Fuster
1997;():V004T14A064. doi:10.1115/97-GT-482.
FREE TO VIEW

Initiated in 1994, a transonic fan in blisk configuration is being developed at MTU with the intention to replace a conventionally bladed titanium fan rotor stage. A major challenge in the structural development process of this blisk stage was to find a blade geometry with a sufficient bird strike resistance. This paper gives an overview of the combined analytical and experimental design process towards the final blisk blade geometry capable of resisting a worst-case 1 lb low speed bird strike at a typical aircraft take-off operating situation without blade loss.

Starting with a first blade geometry a rig ingestion test in 1995 revealed an insufficient bird strike potential of this original blisk blade standard. The intensive analysis of the test result also showed that the analytical model established at MTU had to be adapted with respect to some decisive aspects of the analytical approach. This led to a revised method of meshing the blade leading edge, to a revised failure criterion of the material model and to a revised concept in modelling the bird slicing effect. This was done on the basis of several ‘static’ shooting tests on cantilever steel specimens and on single blade segments of the original blisk. The intention of these tests was to verify the analytical model with respect to quantitatively reliable predictions of local large strain (up to 60% plastic strain) at high rates of strain.

With these revised analytical methods the blisk geometry was redesigned (increased blade hub cross sections) and with a rig ingestion test of a 1 lb pigeon at take-off rotational speed the required capability of the final blisk geometry was verified in 1996.

Commentary by Dr. Valentin Fuster
1997;():V004T14A065. doi:10.1115/97-GT-498.
FREE TO VIEW

A swirl damper intended for high temperature turbomachinery applications was experimentally investigated for its damping characteristics. The results were disappointing in that only a small amount of damping was obtained. Instead, an instability was produced at a speed which exhibited backward whirl. The rotor rig that was used exhibited backward whirl over a speed range between two critical speeds. This backward whirl could be prematurely induced by the use of the swirl gas damper. At high damper pressures, the rotor system exhibited an instability. The exact nature of this instability is not known, however, possible causes are presented. The observations discussed in this paper are also important to those who design swirl brakes or anti-swirl guide vanes for labyrinth seals. The principles involved are similar in that the flow into a labyrinth seal is turned to swirl anti-rotationally as it enters the seal, same as the swirl gas damper.

Commentary by Dr. Valentin Fuster
1997;():V004T14A066. doi:10.1115/97-GT-499.
FREE TO VIEW

A numerical model for the calculation of resonance stationary response of mistuned bladed disc is presented. The bladed disc model includes all important effects on a rotating system of the real geometry. The excitation forces were calculated by a code on the basis of two-dimensional compressible flow (to M < 0.8) for thin airfoil blades. The calculations presented in this paper show that centrifugal stress, and the values of excitation forces, play an important role in considering the influence of mistuning on the response level.

Topics: Resonance , Vibration , Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A067. doi:10.1115/97-GT-500.
FREE TO VIEW

This paper presents a Dynamic Finite Element (DFE) formulation, based on the Dynamic Stiffness Matrix (DSM) approach, for vibrational analysis of spinning beams. The constituent members are considered to be linearly tapered as well as centrifugally stiffened. A non-dimensional formulation is considered, and the frequency dependent trigonometric shape functions are used to find a single frequency dependent element matrix (called DSM) which has both mass and stiffness properties. An adapted bisection method based on a Sturm sequence root counting technique, is used to find the first four out-of-plane flexural natural frequencies of a cantilevered linearly tapered (in height) beam for different non-dimensional rotating speeds. The results have been compared to those found by finite elements method using Hermite beam elements. Much better convergency rates are found by this method when comparing to conventional finite element methods.

Commentary by Dr. Valentin Fuster
1997;():V004T14A068. doi:10.1115/97-GT-501.
FREE TO VIEW

The results of an experimental investigation of the effects of random blade mistuning on the free dynamic response of bladed disks are reported. Two experimental specimens are considered: a nominally periodic twelve-bladed disk with equal blade lengths, and the corresponding mistuned bladed disk, which features slightly different, random blade lengths. In the experiment, both the spatially extended modes of the tuned system and the localized modes of the mistuned system are identified. Particular emphasis is placed on the transition to localized mode shapes as the modal density in various frequency regions increases. Excellent qualitative and quantitative agreement is obtained between experimental measurements and results obtained by finite element analysis. Experimental results are additionally used to validate a component mode-based, reduced-order modeling technique for bladed disks. This work reports the first systematic experiment carried out to demonstrate the occurrence of vibration localization in bladed disks.

Topics: Vibration , Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A069. doi:10.1115/97-GT-502.
FREE TO VIEW

The results of an experimental investigation on the effects of random blade mistuning on the forced dynamic response of bladed disks are reported. Two experimental specimens are considered: a nominally periodic twelve-bladed disk with equal blade lengths, and the corresponding mistuned bladed disk, which features slightly different blades of random lengths. Both specimens are subject to traveling-wave excitations delivered by piezo-electric actuators. The primary aim of the experiment is to demonstrate the occurrence of an increase in forced response blade amplitudes due to mistuning, and to verify analytical predictions about the magnitude of these increases. In particular, the impact of localized mode shapes, engine order excitation, and disk structural coupling on the sensitivity of forced response amplitudes to blade mistuning is reported. This work reports one of the first systematic experiments carried out to demonstrate and quantify the effect of mistuning on the forced response of bladed disks.

Topics: Vibration , Disks
Commentary by Dr. Valentin Fuster
1997;():V004T14A070. doi:10.1115/97-GT-503.
FREE TO VIEW

No literature is currently available which has evaluated finite element power loss models for magnetic bearings and compared the results to experimental results. In this paper a finite element model of the magnetic and electric fields in magnetic bearings, including the motion of the magnetic material in the rotor, is developed. It evaluates the two dimensional magnetic vector potential, magnetic flux density, electric field, eddy current, and power losses in an example magnetic bearing configuration. Results were obtained for both a solid rotor and a laminated rotor. For a solid rotor, both the magnetic flux density and eddy current plots at high rotational speeds are concentrated at the outer edge of the rotor. The ratio of calculated solid to laminated losses is found to be in the range of measured results by other authors. An effective axial conductivity was employed to model a laminated rotor and compared to experimental loss measurements. The correlation between measured and calculated results is quite good for a range of rotor speeds, magnetic flux density, and air gap thickness.

Commentary by Dr. Valentin Fuster

Controls, Diagnostics and Instrumentation

1997;():V004T15A001. doi:10.1115/97-GT-001.
FREE TO VIEW

Nondestructive characterization of non-serviced high-temperature coatings can be considered one of the important factors to achieve a higher level of structural integrity of advanced gas turbines.

The present paper describes an innovative eddy current technique especially developed for measuring the thickness of metallic (MCrAlY) coatings applied by vacuum plasma spray on Ni-base superalloys. Conventional eddy current techniques, well established for quality control of coating thickness, are not applicable in this case because of the low difference of electrical conductivities of conting and base materials, which is a consequence of their quite similar physical and chemical properties.

The new technique employs fast frequency scanning of the electromagnetic field in the range 100 kHz – 10 MHz, corresponding to probing depths from 1 mm to 0.1 mm. Dedicated hardware has been developed, featuring high sensitivity, stability and harmonic rejection.

Analysis of the measured data (i.e. probe impedance vs frequency), in order to estimate the relevant diagnostic parameters (coating thickness, coating and base metal electrical conductivities), is carried out on the basis of a theoretical model of interaction between a plane electromagnetic wave and test piece.

The results of tests performed on a non-serviced first stage blade are reported and compared with reference destructive data. Reliability, accuracy and practical applicability of the method meets the requirements for in-shop quality control.

Commentary by Dr. Valentin Fuster
1997;():V004T15A002. doi:10.1115/97-GT-006.
FREE TO VIEW

This paper discusses the problem of measuring unsteady pressure in a high temperature environment using standard transducers. Commercially available cooling adapters for these transducers use water as the cooling medium to provide thermal protection. This arrangement is suitable only for some test bed applications and not suitable for integration into inflight active control systems.

An assessment of the cooling effectiveness of a commercial water cooled adapter using air as the cooling medium is presented using an experimentally validated finite element heat transfer model. The assessment indicates survival of an air-cooled transducer, itself rated to 235°C, at source flow temperatures up to 800°C.

Commentary by Dr. Valentin Fuster
1997;():V004T15A003. doi:10.1115/97-GT-016.
FREE TO VIEW

With the advent of electronically alterable memories in electronic gas turbine engine control systems, there is now the opportunity for updating software in the field. Field loading provides a means to economically correct problems or introduce enhancements to system operation through the electronic control. In this paper we describe the characteristics of high integrity reprogramming systems used to update engine controls in-the-field.

Pratt & Whitney Aircraft supports two methods for in-service reprogramming of Electronic Engine Controls (EECs), These two methods are PC Laptop based loaders and ARINC loaders. This discussion will focus on the capabilities provided to support in-the-field reprogramming of engine controls. The flexibility, integrity, and the benefits of field reprogramming provided by these software loading systems will be explained. These reprogramming systems provide a PC based application and ARINC based systems for either on-wing reprogramming or on-board reprogramming directly from a flight deck device to the EEC.

The PC Loader reprogramming utilities allow field personnel to reprogram engine control application software and/or constants and configuration information using a suitably equipped IBM PC or compatible computer. These utilities are intended to be operated per Service Bulletin authorization only. They require a PC compatible computer (presumably a laptop model) with 2 UART interface cards, an interfacing cable, and the new software to be loaded. The rigor and manner of the integrity checks to ensure proper loading of the control is essential to an acceptable loading system.

There are two types of ARINC-based loaders: on-wing loaders and on-board loaders. Both types enable the operator to upload application, trim, and/or configuration software to the engine control. Additionally the ARINC 615 device allows operators to download fault and configuration data from the control. Each type of loader uses a specially formatted file to control the sequence of operations involved in a data loading session. The on-wing loader utilizes a specially designed portable data loader which connects directly to the EEC via dedicated cabling through the control’s ARINC connectors. This type of data loader contains software which communicates via an ARINC 615 protocol to a peer software entity running on the EEC. The on-board loader uses the aircraft’s central maintenance computer system to communicate with the EEC over the aircraft’s ARINC 629 data bus. It also operates using a peer-to-peer communication protocol with the EEC. The ARINC 629 loader requires no extra equipment or cabling, nor does it require the EEC to be accessible for attachment of cables.

Commentary by Dr. Valentin Fuster
1997;():V004T15A004. doi:10.1115/97-GT-027.
FREE TO VIEW

The diagnosis of gas turbine sensor faults requires models of the system to calculate estimates of the measured output system variables.

The model set-up phase is of great importance since the reliability of the diagnostic tool depends on the model accuracy.

In the paper two different methodologies of I/O linear model set-up are analyzed and compared to find the more simple and general one.

The first methodology consists in obtaining the I/O linear models by directly linearizing the physical laws (system modeling).

The second one uses statistical methods (system identification) to calculate model parameters from time series data measured on the machine. The models used are of the ARX (Auto Regressive with eXternal input) type. The number of models and the measured variables correlated by each of them have been determined in order to obtain unambiguous fault signatures for each sensor.

The system identification method proves to be more suitable to the system modeling because of its greater simplicity in the fault diagnosis application.

Commentary by Dr. Valentin Fuster
1997;():V004T15A005. doi:10.1115/97-GT-028.
FREE TO VIEW

A physics based one-dimensional generic non linear dynamic model for a gas generator engine which is expected to enable future model based system oriented activities such as system control design, engine dynamic analysts, variable geometry optimization, air bleed extraction optimization, active surge and/or rotating stall avoidance control is presented. The model can be easily partitioned to obtain a one-dimensional compressor dynamic model which can simulate onset and development of surge and demonstrate existence of rotating stall. The notion of surge domain is introduced and it is shown how the dynamic model can quantitatively predict it. Other model predictions such as surge frequency and air bleed extractions are compared with experimental results.

Commentary by Dr. Valentin Fuster
1997;():V004T15A006. doi:10.1115/97-GT-029.
FREE TO VIEW

A practical magnetic-bearing control system has been designed based upon modeling and simulation of the dynamics of a jet engine turbine shaft and bearing system. Simulations include models for flexible rotor dynamics, magnetic actuators, auxiliary touchdown bearings, ordinary and extraordinary external loads, and disturbances from rotor imbalance, stator vibration, and noise. The shaft model includes a motor-generator which acts as an uncontrolled negative stiffness.

The control system is decentralized, running independently for each of the five physical axes of control (1 axial, 4 radial). The fundamental algorithm is classical PID: proportional for broadband stiffness, integrator (with anti-windup) for high load-carrying capacity, and derivative to dampen disturbances. Additional phase lead is provided via a first-order pole-zero pair. The vibration due to rotor imbalance is eliminated by an autobalancing algorithm. Compensation for magnetic actuator non-linearity and varying rotor-stator gap is provided by feedback of sensed magnetic flux, using sensor coils built into the actuator. The control design can be readily implemented using a commercial Digital Signal Processing system. The magnetic bearing actuators will be driven with commercial power amplifiers via customized front-end electronics.

Based upon simulations, the design goal has been achieved of keeping the shaft within two mils of its desired location at the magnetic bearings, under all normal loads. Under extreme external loads, the capacity of the magnetic bearings will be exceeded and touchdown will occur upon backup mechanical bearings. Simulation shows that the control design handles this critical event, which determines the force slew rate required from the actuators.

Commentary by Dr. Valentin Fuster
1997;():V004T15A007. doi:10.1115/97-GT-030.
FREE TO VIEW

Integrated life, vibration and performance monitoring/diagnostics capable of detecting and classifying developing engine faults is critical to reducing engine operating and maintenance costs while optimizing the life of “hot section” engine components (Troudet and Merrill, 1990). Advanced fault pattern recognition and classification techniques utilizing complex finite-element and empirical models of structural and performance related engine areas can now be accessed in a real-time monitoring environment (Dietz et al., 1989). Integration and implementation of these proven technologies presents a great opportunity to significantly enhance current engine health diagnostic capabilities and safely extend engine component life (Ali and Crawford, 1988).

Commentary by Dr. Valentin Fuster