ASME Conference Presenter Attendance Policy and Archival Proceedings

2015;():V07AT00A001. doi:10.1115/GT2015-NS7A.

This online compilation of papers from the ASME Turbo Expo 2015: Turbine Technical Conference and Exposition (GT2015) represents the archival version of the Conference Proceedings. According to ASME’s conference presenter attendance policy, if a paper is not presented at the Conference, the paper will not be published in the official archival Proceedings, which are registered with the Library of Congress and are submitted for abstracting and indexing. The paper also will not be published in The ASME Digital Collection and may not be cited as a published paper.

Commentary by Dr. Valentin Fuster

Emerging Methods in Design and Engineering

2015;():V07AT27A001. doi:10.1115/GT2015-42279.

This paper describes a theoretical approach to shift individual natural frequencies of centrifugal compressor impeller blades. The approach applies sizing optimization of blade’s geometry using a gradient-based optimization method. Calculation of gradients is carried out by the finite-difference method. A new centrifugal compressor blade profile generator incorporating a blade parametrization procedure is developed. The blade’s geometry is parametrized using intuitive geometric parameters. Five design parameters related to the length of the sectional profile generator line, profile thicknesses and rotation angles at hub and shroud are defined for each of the blade sectional profiles. In addition, two global design parameters are defined to control rigid rotation of the blade hub and shroud sections in circumferential direction. Four nonlinear optimization problems containing multiple frequency constraints and constraints on the static equivalent stresses are considered. The optimization aims are either shifting a particular natural frequency of a blade or minimization of blade’s mass. For instance, one of the considered optimization problems is to decrease the 1st natural frequency of an impeller blade by 5%, while the 2nd and the 3rd natural frequencies must be simultaneously increased by 5%. The analysis is applied to the centrifugal compressor of a small-size turboprop engine. A three-dimensional finite element model of the impeller blade is developed in ANSYS Mechanical software package to perform static and modal analyses. The results of the optimization show that the code can meet defined objectives and constraints with reasonable accuracy. A detailed comparison of optimized profiles with the baseline geometry is provided.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A002. doi:10.1115/GT2015-42365.

To achieve reverse objectives in engine design, advanced modelling and analysis methods are among the key research technologies. In the presented work, a robust design optimization of a first stage high pressure turbine blade has been carried out. This blade derives from a current production of a Rolls-Royce aero engine.

The motivation of this work is to show that the methodology of robust design optimization can be applied to high pressure turbine blades. A fully automated workflow, which encapsulated the integral blade design and analysis process, has been used. The main workflow objective is a representative life value of the external surface of the blade. In addition, the workflow enables the engineering uses to consider sub objectives like mass, efficiency and life at critical locations of the blade. These can also be taken into account in the multi-objective robust design optimization. This research also focuses on the use of surrogate models, with attention to the delivery of a physically correct result. For this purpose, the validation of the applied methods has a huge significance and a toolbox was created to generate and evaluate the quality of the surrogate models.

In the present case sixteen geometry parameters were considered. In order to show that this methodology is not limited to geometry variation, parameters for material specification and for boundary conditions were varied in addition. The surrogate model was trained by the workflow generated DoE-data and could be used for different kinds of optimization.

As a conclusion, it has been demonstrated that the methodology can be used for the engineering design process of turbine blades, while delivering physically correct results. The different techniques for surrogate modelling were examined and compared. With the help of these surrogate models, an optimization of life, mass and efficiency with 22.5 million evaluations was possible. Finally, an overview of the methodology for the case of a real world turbine blade could be given, and an improved blade in the sense of multi-objective robust design was found.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A003. doi:10.1115/GT2015-42661.

The design and development process of an aero engine is a complex and time-consuming task that involves many disciplines and company departments with different objectives and requirements. Along the preliminary design phase, multiple concepts are assessed in order to select a competitive technology. The engine design process, which was traditionally subdivided into modular component tasks, is nowadays considered as a multi-disciplinary workflow. Having recognized the need for developing advanced turbine preliminary design tools, this work focuses on enhancing the integration of turbine design disciplines, improving the accuracy of models and speeding the time to generate models.

The proposed process facilitates an automated turbine Secondary Air System (SAS) and turbine discs concept definition. Furthermore, the process of CAD models and flow network models generation is accelerated via automation of the engineering workflow. This is accomplished through a novel Java based data model, where the design of turbine discs and SAS features is captured in a programmable framework.

In the application section, the preliminary design definition of a reference HP turbine subsystem is replicated using the newly developed common design environment. The automated workflow is then used to generate the corresponding CAD models, recognize the subsystem flow network, and generate the 1D flow network model. The results are then compared to the experimentally validated model of a reference engine. As conclusion, the automated workflow offers a quick and parametric model generation process, while providing a good level of fidelity for the preliminary design phase.

Topics: Design , Turbines
Commentary by Dr. Valentin Fuster
2015;():V07AT27A004. doi:10.1115/GT2015-42818.

Bird strikes have been a concern to aviation safety in both civil and military aircrafts. The external surfaces of an aircraft which include wing leading edges are susceptible to bird-strikes. Recently topology optimization is used to realize an aircraft wing concept design using Aluminum in (2009) [1] and optimize its weight. To make it lighter further, a composite wing was derived in (2010) [2]. Here, Fibre Metal Laminates (FMLs) with layers of aluminium alloy and high strength glass fibre have been used. The impact analysis is performed using a SPH model for the bird. Based on the modal analysis, the impact time is determined to capture the modes with high participation factor. The response obtained in time domain is converted to frequency domain and it is shown that the response has predominantly these modal components.

For the given FMLs wing configuration the stress levels obtained are well within the orthotropic yield limits of the structure. All the layers of the composite structure are found to be intact.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A005. doi:10.1115/GT2015-43150.

High cycle fatigue due to mode localization caused by geometric and material mistuning is one of the leading failure risks of integrally bladed rotors (IBRs). Due to the computational analysis cost of full wheel models, IBR mistuned response amplifications are often modeled with reduced order models (ROMs). However, many developed ROMs are based on nominal mode assumptions that do not consider mode shape variations that have been shown to impact predicted mistuned response. Geometrically mistuned finite element models (FEMs) do account for mode shape variations but are notoriously difficult to construct and analyze. Recent advancements in optical scanning have enabled the rapid acquisition of highly accurate dense point clouds representative of manufactured hardware. Previous research pioneered a novel method to automatically and robustly construct an FEM directly from tessellated scan data, this research adds new mesh quality verification algorithms and experimentally validates this algorithm using results from traveling wave excitation. Sensitivity to mesh and point cloud density are also assessed to determine a best practice for creation of the as manufactured mistuned rotor model.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A006. doi:10.1115/GT2015-43179.

An Alstom tool is described for the automated and simultaneous design optimisation of 2 and 4-hook T-root grooving of multiple steam turbine rotor stages in order to minimise the peak stress. The finite element axisymmetric thermal-stress calculation is performed with Abaqus in a few hours on modest hardware. The tool embeds Python scripting to facilitate the rotor groove model definition and meshing within Abaqus/CAE, with emphasis placed on minimising the effort for the initial setup. Rotor groove shapes are described with B-splines, maintained and modified within the in-house tool. Their shape is progressively refined as directed by a hybrid evolutionary-gradient based optimisation engine in order to achieve the minimum stress objective. In the region of highest stress, the groove boundary shape adjusts as the optimisation proceeds to conform to the local contours of stress. Application to a low pressure steam turbine rotor demonstrates comparable or lower stresses with this tool compared to those from manual expert optimisation. The method can be readily extended to other geometric entities on the rotor described with B-spline curves, e.g. cavities, seals.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A007. doi:10.1115/GT2015-43192.

Blade tip timing (BTT) is a commonly used non-intrusive stress measurement system to estimate the operational stresses within an engine’s rotors without the costly installation of strain gauges that can add additional stiffness to the rotor. BTT systems are now standard on many engine tests and ensure safe operations by avoiding running near maximum rotor stress limits. Since these systems measure blade time of arrival (TOA), processes are applied to first convert this data to displacement and then to stress. This effort focuses on the conversion of displacement to stress where the the classic approach utilizes nominal geometry obtained from an “as-designed” nominal model and creates computes the mode shapes using finite element analysis (FEA). The predicted mode shapes of the cyclic analysis reveal the relationship between maximal blade stress and tip displacement for a given nominally designed rotor. However, manufactured rotors deviate from nominal design due to inherent variability in the machining procedures. It is now possible through high fidelity optical geometry collection systems to obtain more accurate BTT limits using measured IBR geometry from as-manufactured rotors. It will be shown that due to the high variability of blade-to-blade geometry obtained from an optically scanned rotor that the BTT limits can vary significantly between blades. A method is also developed that allows comparisons between cyclic sector and full rotor FEA. This research suggests to optimize BTT probe placement not only to measure the maximum expected deflection given blade tip mode shapes, but also to account to for blade to blade geometric variation.

Topics: Modeling , Blades , Geometry
Commentary by Dr. Valentin Fuster
2015;():V07AT27A008. doi:10.1115/GT2015-43837.

Diffusion-bonded titanium hollow warren structures have been successfully used on aircraft engine components, such as fan blade, etc., for which failure behavior under impact load is one of the major considerations during design. Studies show that the welding seam de-bonding is the major failure mechanisms under the impact loads. To effectively simulate the de-bonding, phenomenological effective bond strength needs to be properly defined for both the girder region and the edges. In this paper, the bond strength at the hollow region is calibrated through quasi-static tensile tests, in which the specimens are properly designed to represent the cross-section of the hollow region of a typical warren structure. Then, the strength of the edge region is estimated through an inverse method based on the analysis-test correlation of a hollow panel impact test. Such multi-factor coupled seam failure criteria can provide reasonable accordance with test result in the simulation of impact failure of the hollow structure. This means the model with such failure criteria incorporated can provide a useful reference to the aircraft engine hollow warren structure components design.

Commentary by Dr. Valentin Fuster
2015;():V07AT27A009. doi:10.1115/GT2015-43952.

Modified Modal Domain Analysis (MMDA) is a method to generate an accurate reduced order model (ROM) of a bladed disk with geometric mistuning. An algorithm based on MMDA ROM and a state observer is developed to estimate forcing functions for synchronous (including integer multiples) conditions from the dynamic responses obtained at few nodal locations of blades. The method is tested on a simple spring-mass model, finite element model (FEM) of a geometrically mistuned academic rotor and FEM of a bladed rotor of an industrial scale transonic research compressor. The accuracy of the forcing function estimation algorithm is examined by varying the order of reduced-order model and the number of vibration output signals.

Topics: Rotors
Commentary by Dr. Valentin Fuster
2015;():V07AT27A010. doi:10.1115/GT2015-44140.

Grid convergence in finite element analysis, despite a wide variety of tools available to date, remains an elusive and challenging task. Due to the complex and time consuming process of remeshing and solving the finite element model (FEM), convergence studies can be part of the most arduous portion of the modeling process and can even be impossible with FEMs unassociated with CAD. Existing a posteriori methods, such as relative error in the energy norm, provide a near arbitrary indication of the model convergence for eigenfrequencies. This paper proposes a new approach to evaluate the harmonic convergence of an existing model without conducting a convergence study. Strain energy superconvergence takes advantage of superconvergence points within a FEM and accurately recovers the strain energy within the model using polyharmonic splines, thus providing a more accurate estimate of the system’s eigenfrequencies without modification of the FEM. Accurate eigenfrequencies are critical for designing for airfoil resonance avoidance and mistuned rotor response prediction. Traditional error estimation strategies fail to capture harmonic convergence as effectively as SES, potentially leading to a less accurate airfoil resonance and rotor mistuning prediction.

Commentary by Dr. Valentin Fuster

Fatigue, Fracture and Life Prediction

2015;():V07AT28A001. doi:10.1115/GT2015-42156.

A critical plane approach in combination with principal component analysis (PCA) for determining dominant damage factors (DDFs) was developed for single crystal nickel superalloys at elevated temperature. Maximum resolved shear stress (RSS), maximum slip rate and other 2 mesoscopic parameters on the critical plane, defined as the preferential slip plane, were selected as damage parameters. Correlation analysis results indicated that there were strong correlations (i.e. multicollinearity) among the selected parameters. To address this issue, PCA was performed to eliminate the effect of multicollinearity and the DDFs were determined as well. Based on the DDFs a life model was proposed and then validated by the fatigue experimental results. Most of the experimental lives are within the factor three of the predicted ones. The life model has a relatively simple form with reliable constants which facilitates the application in industry design.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A002. doi:10.1115/GT2015-42681.

In this study the thermomechanical fatigue (TMF) crack initiation of the single-crystal nickel-base superalloy MD2 is investigated and evaluated. A series of experiments are performed of smooth specimens loaded in the nominal [001] and [011] crystal orientations, subjected to both in-phase and out-of-phase TMF loading conditions. Considering the inherent internal structure of crystallographic slip planes in single-crystals, a number of critical-plane approaches are evaluated to enable a good description of the TMF crack initiation. These are evaluated using finite element simulations and a post-process, in which crystallographic entities are extracted and compared to the experimental TMF life. A good correlation is achieved for two of the critical-plane approaches. These are able to predict the TMF crack initiation taking into account the elastic and plastic anisotropy, the tension/compression asymmetry and the creep relaxation present in the material.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A003. doi:10.1115/GT2015-42701.

Testing and simulation of aero engine spectra with dwell times are reported in this paper. The modelling concept used is built on LEFM and provides a history dependent evolution description of dwell damage and its interaction with cyclic load. The simulations have been carried out for three spectra, 1) cyclic loads, 2) combined sustained load and cyclic loads and 3) slow load ramps and cyclic loads, all for surface cracks at 550°C for Inconel 718. All simulations show reasonable good agreement with experimental results. Prediction of multiple tests of several batches is also provided to show statistical scatter.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A004. doi:10.1115/GT2015-42709.

In this paper scatter in crack growth for dwell time loadings in combination with overloads has been investigated. Multiple tests were performed for surface cracks at 550°C in the commonly used high temperature material Inconel 718. The test specimens originate from two different batches which also provides for a discussion of how material properties affect the dwell time damage and overload impact. In combination with these tests an investigation of the microstructure was also carried out, which shows how it influences the growth rate. The results from this study show that, in order to take overloads into consideration when analysing spectrum loadings containing dwell times, one needs a substantial amount of material data available as the scatter seen from one batch to the other is of significant proportions.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A005. doi:10.1115/GT2015-42772.

The majority of fatigue crack growth tests worldwide makes use of the compact tension (CT) specimens that are not necessarily representative of cracks developing under service conditions in highly stressed components of a jet engine. Over the years other geometries have been designed to facilitate a study of relatively small, semi- or quarter-elliptically shaped surface flaws subjected to high tensile and compressive stresses. Despite an extensive use by the aerospace community, practical aspects of testing and data analysis relevant to the complex surface- or corner-flawed geometries are not regulated by a commonly accepted set of rules.

Two types of test specimens — CT and surface-crack tension (SCT) — were machined from a forged and heat treated Inconel 718. For both geometries the crack orientation and propagation direction with respect to the original forging were the same.

The CT specimens were tested in accordance with the ASTM Standard E647 as well as using an alternative compression pre-cracking procedure. After correct application of the compression pre-cracking to the CT geometry both approaches had yielded reasonably consistent results. At high ΔK values both studied geometries also produced similar results, however, as the ΔK values decreased, a trend towards slower crack growth rates in the SCT specimens became evident.

In order to address a so-called small-crack effect, several SCT specimens received much smaller crack starter notches produced by the focused ion beam (FIB) technology.

The results of the present study highlight the importance of the appropriate material properties for accurate and reliable service life prediction of the critical aerospace propulsion hardware with particular emphasis on the influence of specimen/crack geometry and test method on the fatigue crack propagation behavior of jet engine alloys.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A006. doi:10.1115/GT2015-42820.

Thermo-mechanical fatigue (TMF) crack growth modelling has been conducted on Inconel 718 with dwell time at maximum load. A history dependent damage model taking dwell damage into account, developed under isothermal conditions, has been extended for TMF conditions. Parameter determination for the model is carried out on isothermal load controlled tests at 550–650°C for surface cracks, which later have been used to extrapolate parameters used for TMF crack growth. Further, validation of the developed model is conducted on a notched specimen subjected to strain control at 50–550°C. Satisfying results are gained within reasonable scatter level compared for test and simulated number of cycles to failure.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A007. doi:10.1115/GT2015-42865.

The paper will describe the main outcomes of a risk assessment performed by General Electric O&G (GE O&G) in 2013 on a large number of gas turbine rotors that had exceeded their design end of life (EOL).

The assessment involves a large number of medium frames (e.g. MS3002J, MS5002C) plus small industrial gas turbines (e.g. MS1002, PGT10).

The design end of life is the lifespan inside which the risk of uncontained failure is expected to be improbable and this means the failure has a probability of occurrence of 0.12 total events or less during the life of five hundred gas turbines operating 8000 hours per year for 30 years [1].

GE gas turbine rotors are designed to operate with no need of inspection and maintenance till their design EOL (alias serviceable life).

The gas turbines end users may exceed EOL, provided the rotors have been subjected to a specific investigation that includes teardown and a full inspection.

The disregard of this recommendation may have catastrophic consequences on the gas turbines and on surrounding equipment and personnel.

Such recommendation was provided to customers the first time in 2005 through GETIL 1576 [2], the recommendation was renewed in 2013 with a new communication NIC 13.16 [3].

The risk assessment has essentially focused on rotating parts, because rotors (i.e. shaft, wheel, spacer) are the highest gas turbine energy components, that even if designed with high safety margins, they are not immune from time (e.g. creep) and/or cyclic dependent damage (e.g. fatigue).

The assessment has been carried out by following MIL-STD-882C [1] guidelines and in accordance to the principles defined by ISO 12100-2010 [4].

The failure risks have been estimated with the help of specific tools by using field data and life calculations performed with physics-based models.

The paper describes the methodology used to support the risk assessment and this means the “rotor life management methodology” that was developed by GE more than ten years ago, and was subsequently improved by including statistical assessments.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A008. doi:10.1115/GT2015-43071.

The stringent structural requirements posed on aircraft engines, especially the high pressure turbine blades, result from the diversity of the extreme operational conditions they are subjected to. The accurate life assessment of the blades under these conditions therefore demands accurate analytical tools and techniques, and also an elaborate understanding of the operational conditions. Given the drive to reduce cost related to experimental testing, numerical approaches are often adopted to aid in the initial design stages. With recent advancement in numerical modelling, the simultaneous integration of the various numerical codes of fluid flow and structural analysis (otherwise known as fluid-structure interaction) is projected to provide reliable input into fatigue life prediction programs. This study adopts the numerical method of fluid-structure interaction to investigate the fatigue properties of the Aachen turbine test case. A load-time history obtained for the high stress monitor position is superimposed on that from a quasi-static FE solution, and used as input into a fatigue estimation tool. The low cycle fatigue (LCF) is estimated using the Basquin-Coffin-Manson correlation with corrections for mean stress and multi-axial fatigue effects. An FFT analysis of the fluctuating aerodynamic loads show signals with significant high frequency content. There is noticeable increased energy signal at the rotor inlet as compared to stator inlet. The stator inlet signals, however, are characterized by multiple resonances of frequency with lower energy content. By avoiding the resonances, the fatigue analysis predicts a safe design with a safety factor level of 3 for the rotor.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A009. doi:10.1115/GT2015-43191.

Many Alstom heavy-duty gas turbines with a silo combustor are in service and moreover undergoing upgrades for performance augmentation, lifetime extension, and emission reduction. The combustor liners, which are exposed to high gas temperatures, may require design tuning for these upgrades, and therefore reliable simulation of their behavior is of utmost importance.

This paper focuses on a case study of transient behavior of the Hot Gas Casing, a transition liner between compressor, turbine and silo combustor. Following a three-dimensional thermal assessment based on computational fluid dynamics, detailed structural analysis is done to identify the drivers behind different types of Hot Gas Casing deformation, observed after upgraded combustor introduction.

Thermal Barrier Coating application is proposed to reduce Hot Gas Casing bending. The solution was confirmed analytically and successfully introduced in the field. Good correlation of field findings and finite element prediction was found. It is shown that only a combined effect of thermal deformation, cyclic loading and creep may explain the observed behavior of the part.

A detailed design sensitivity study is performed, comparing different approaches to deformation reduction: application of ribs, corrugations and wall thickness increase. The best solution in terms of manufacturability and its impact on part deformation is chosen. It is found that wall corrugation does not provide the desired effect due to the nature of part deformation.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A010. doi:10.1115/GT2015-43303.

The failure of vital components is not uncommon in the gas turbine industry. In the event excessive degradation occurs within a component, e.g. extensive cracking in a turbine blade or vane, solutions exist to either repair or replace defective parts. Such parts are readily accessible and mostly exchangeable in the field, limiting the amount of outage time and assessment required for defective parts. When more critical components exhibit extreme wear or cracking, e.g. a crack in a rotor disk, repairs typically necessitate a complete rotor destack and refurbishment or have the potential to require the replacement of individual disks. In extreme cases, defects found in rotor disks can be known to retire an entire compressor or turbine rotor. The OEM solution of replacing disks puts a substantial cost on the customer, thus providing an incentive for characterization and advanced analyses to determine the residual life in critical rotating components.

Considered an advanced analysis, linear elastic fracture mechanics (LEFM) provides the theory and fundamental structure to conduct crack growth analyses in components that exhibit nominally elastic behavior. Successful implementation of LEFM requires extensive characterization of the material, engine operating boundary conditions, and high fidelity finite element models. Upon the detection of a flaw, whether an internal or external indication, the results from finite element analyses can be used to derive the crack tip stress field and subsequent crack tip driving parameters. These parameters are then utilized in a comprehensive crack propagation model, calibrated to temperature- and load-dependent material data, to determine the number of cycles to unstable propagation. As a result, the remaining life of a component with a given indication is readily obtained, enabling our engineering team to provide a thorough life assessment of critical rotating components.

An overview of the linear elastic fracture mechanics crack growth analyses conducted is presented, with a special emphasis on compressor and turbine disks.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A011. doi:10.1115/GT2015-43312.

Utilising a comprehensive design tool suite and in-house knowledge, the root cause analysis of High Cycle Fatigue (HCF) failures on the first stage compressor blade in a gas turbine (GT) engine is presented in this paper.

Based on the experience gained from the root cause analysis more reliable lifetime prediction for compressor blade design is possible. Tip timing measurements of blades in service have been evaluated to obtain valuable information about the real vibratory stress levels of the first stage compressor blades, to validate design methods, and to identify the failure mechanism. Metallurgical evaluation and testing of the failed parts were also used as part of the investigation. During the failure investigations the impacts of important damage mechanisms, not identified in the original design process, were determined. These damage mechanisms included, water droplet erosion of blades due to impingement on highly stressed regions of airfoil, mistuning within manufactured blade sets, and a very strong transient aerodynamic loading.

Considering the damage mechanisms, improved design methodologies have been used to design robust compressor blade sets. These have been validated in successful engine tests. Subsequent continued successful operation of new blade designs has been recorded in engines during an extended validation period.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A012. doi:10.1115/GT2015-43333.

Industrial and aeroderivative gas turbines when used in CHP and CCPP applications typically experience an increased exhaust back pressure due to pressure losses from the downstream balance-of-plant systems. This increased back pressure on the power turbine results not only in decreased thermodynamic performance but also changes power turbine secondary flow characteristics thus impacting lives of rotating and stationary components of the power turbine.

This Paper discusses the Impact to Fatigue and Creep life of free power turbine disks subjected to high back pressure applications using Siemens Energy approach.

Steady State and Transient stress fields have been calculated using finite element method. New Lifing Correlation [1] Criteria has been used to estimate Predicted Safe Cyclic Life (PSCL) of the disks. Walker Strain Initiation model [1] is utilized to predict cycles to crack initiation and a fracture mechanics based approach is used to estimate propagation life. Hyperbolic Tangent Model [2] has been used to estimate creep damage of the disks.

Steady state and transient temperature fields in the disks are highly dependent on the secondary air flows and cavity dynamics thus directly impacting the Predicted Safe Cyclic Life and Overall Creep Damage. A System-level power turbine secondary flow analyses was carried out with and without high back pressure. In addition, numerical simulations were performed to understand the cavity flow dynamics. These results have been used to perform a sensitivity study on disk temperature distribution and understand the impact of various back pressure levels on turbine disk lives.

The Steady Sate and Transient Thermal predictions were validated using full-scale engine test and have been found to correlate well with the test results. The Life Prediction Study shows that the impact on PSCL and Overall Creep damage for high back pressure applications meets the product design standards.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A013. doi:10.1115/GT2015-43335.

Industrial and aeroderivative gas turbines when used in CHP and CCPP applications typically experience an increased exhaust back pressure due to losses from the downstream balance-of-plant systems. Also, gas turbines for mechanical drive application have a wide operating envelope which leads to a fluctuating back pressure that varies with change in exhaust flows. This increased back pressure on the power turbine results in increased exhaust gas temperatures and aerodynamic loading that can influence the mechanical integrity and life of Power Turbine Exhaust System.

This Paper discusses the Impact to Fatigue and Creep life of free power turbine exhaust system subjected to high back pressure applications using Siemens Energy approach.

Steady state and transient temperature fields were predicted using finite element method. These predictions were validated using full-scale engine test and are found to correlate well with the test results.

Full Scale strain gauge survey of the exhaust hood was undertaken at ambient conditions at various pressure levels to validate the structural boundary conditions of lifing models. Strain Predictions were found in good agreement with measured strain gauge data.

Steady State and Transient stress fields have been estimated using validated structural and thermal finite element models. Walker Strain Initiation model [1] is utilized to predict Low Cycle Fatigue Life and Larson Miller Parameter Creep Model has been used to estimate creep damage to the exhaust system.

The Life Prediction Study shows that the exhaust system design for high back pressure applications meets the product design standards.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A014. doi:10.1115/GT2015-43380.

With increasing use of renewable energy sources, an industrial gas turbine is often a competitive solution to balance the power grid. However, life robustness approaches for gas turbine components operating under increasingly cyclic conditions, is a challenging task. Ductile superalloys, as Haynes 230, are often used in stationary gas turbine hot parts such as combustors. The main load for such components is due to non-homogeneous thermal expansion within or between parts. As the material is ductile there is considerable redistribution of stresses and strains due to inelastic deformations during the crack initiation phase. Therefore, the subsequent crack growth occurs through a material with significant residual stresses and strains. In this work, fatigue crack propagation experiments, including the initiation phase, have been performed on a single edge notched specimen under strain controlled conditions. The test results are compared to fracture mechanics analyses using the linear ΔK and the nonlinear ΔJ approaches, and an attempt to quantify the difference in terms of a life prediction is made. For the tested notched geometry, material and strain ranges, the difference in the results using ΔKeff or ΔJeff are larger than the scatter seen when fitting the model to the experimental data. The largest differences can be found for short crack lengths, when the cyclic plastic work is the largest. The ΔJ approach clearly shows better agreement with the experimental results in this regime.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A015. doi:10.1115/GT2015-43479.

Thermal Mechanical Fatigue (TMF) is emerging to be a major damage accumulation mode in the first-stage turbine blades of the F class industrial turbines. In this work, the effect of engine operating temperatures on the TMF life of the first-stage W501F turbine blade has been studied. The turbine inlet temperature (TIT) for the base load engine operation is predicted using thermodynamics based engine analysis. A computational fluid dynamics (CFD) based cascade analysis of the gas path was performed to predict the metal temperatures and temperature gradients. A higher average TIT with a sharper gradient compared to the TIT profile for the base load was also assumed to numerically represent a more severe engine operating test case. A material physics based TMF life prediction approach was adopted to identify the fracture critical locations (FCL). The cyclic life to crack nucleation under different engine operating conditions was also predicted. The predicted FCL and life matched with field experience. A FCL was also predicted inside the cooling channels of the blade, suggesting the need to develop suitable inspection techniques.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A016. doi:10.1115/GT2015-43755.

Modern gas turbine bolts experience severe operational conditions due to high temperatures and elevated axial stresses, generated by the tightening couple applied during the turbine assembly. In such conditions the relaxation of the initial stress due to viscous phenomena has to be taken into account in order to guarantee the proper operation of the turbine.

Relaxation modelling can either be based on strain controlled relaxation tests or load controlled creep tests. Both solutions present difficulties: relaxation tests entail critical experimental issues, whereas creep tests may not be significant for the given strain controlled operational condition of a gas turbine bolt. Some of these problems will be described in the paper and solutions will be provided.

The performances of several models for stress relaxation quantification will be compared, highlighting advantages and disadvantages of each approach. In particular, great emphasis will be given to those aspects which are relevant for bolt design or tightening load calculation. For instance, some important requirements are: firstly, the possibility to implement the given model easily in finite element calculations; secondly, the possibility to accurately calculate the relaxation in the second life of a serviced bolt after re-tightening; lastly, the possibility to reduce as much as possible the time required for the experimental tests.

In order to evaluate the coefficients of the different models considered in the study, creep tests were performed at 450°C and 475°C with applied stresses producing a strain ε = 1% in a time range of 1000–10000h and stress relaxation tests were performed at the same temperatures with initial strain in the range of 0.2%. After some stress relaxation, the specimens were reloaded at the initial stress several times in order to simulate the aforesaid service conditions of bolts.

In the paper it will be shown how a valid model, capable of predicting the stress relaxation with acceptable accuracy, can be fed either by creep or relaxation tests, provided that the experimental tests and the related data elaboration are conducted with the proper methodology. This scenario provides the engineer responsible for material model creation with a remarkable flexibility, essential to fulfill the requirements of modern GT design, in terms of accuracy, promptness of data collection and possibility of FEM implementation.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A017. doi:10.1115/GT2015-44128.

Fatigue life and reliability are the critical problems for long blades design due to complicated stress state, wet steam and aggressive environment. In this report, the effects of stress ratio, surface properties, steam, and sodium-chloride (NaCl) aqueous environments on the fatigue strength and fracture mechanisms of Ti-6Al-4V alloy have been investigated. Results indicate that residual compressive stress decreases and vanishes finally with increasing stress ratio. Compared to fatigue crack originating from surface for annealed specimens, the fatigue crack initiation sites are located in the interior of the specimen due to the effect of residual stress when low stress ratios are present. Fatigue experiments have been performed in saturated steam with low oxygen content at 100°C and NaCl aqueous at 80 °C. Results indicate that, for 0.1 stress ratio loading conditions, steam environment demonstrates the most serious effect on the endurance limit with 12.3% reduction of fatigue strength. NaCl aqueous leads to the 9.6% drop in fatigue strength corresponding to 107 cycles of design life. For all corrosion environments, cracks originated from the surface and no corrosion pits were observed.

Commentary by Dr. Valentin Fuster
2015;():V07AT28A018. doi:10.1115/GT2015-44132.

High-pressure low-pressure (HLP) combined rotors have been gradually used in advanced power plants. In the present work, tensile, impact, fatigue and creep experiments were conducted to comprehensively investigate mechanical properties of a newly developed combined rotor steel 25Cr2NiMo1V based on microstructure influence. Fatigue crack growth (FCG) rates were obtained by both constant amplitude method and the load-shedding technique. A new method based on cyclic plastic zone size being equalling to grain size was introduced to defferentiate corresponding FCG data in the Paris regime and the near-threshold regime. Results show that 25Cr2NiMo1V steel has sufficient lower temperature strength and toughness in LP zone, and good high temperature creep properties in HP zone. Fracture Appearance Transition Temperature (FATT) at center core of LP is lower than 3°C with a tensile strength of 850 MPa, and creep rupture strength of HP for 105 h is respected to reach 165 MPa at 566°C. By comparison with other combined rotor materials in literature, the properties of 25Cr2NiMo1V steel enable it particularly suitable to HLP rotor material for advanced combined cycle power plants.

Commentary by Dr. Valentin Fuster

Probabilistic Methods

2015;():V07AT29A001. doi:10.1115/GT2015-42392.

Calibration and uncertainty quantification for gas turbine (GT) performance models is a key activity for GT manufacturers. The adjustment between the numerical model and measured GT data is obtained with a calibration technique. Since both, the calibration parameters and the measurement data are uncertain the calibration process is intrinsically stochastic. Traditional approaches for calibration of a numerical GT model are deterministic. Therefore, quantification of the remaining uncertainty of the calibrated GT model is not clearly derived. However, there is the business need to provide the probability of the GT performance predictions at tested or untested conditions. Furthermore, a GT performance prediction might be required for a new GT model when no test data for this model are available yet. In this case, quantification of the uncertainty of the baseline GT, upon which the new development is based on, and propagation of the design uncertainty for the new GT is required for risk assessment and decision making reasons. By using as a benchmark a GT model, the calibration problem is discussed and several possible model calibration methodologies are presented. Uncertainty quantification based on both a conventional least squares method and a Bayesian approach will be presented and discussed. For the general nonlinear model a fully Bayesian approach is conducted, and the posterior of the calibration problem is computed based on a Markov Chain Monte Carlo simulation using a Metropolis-Hastings sampling scheme. When considering the calibration parameters dependent on operating conditions, a novel formulation of the GT calibration problem is presented in terms of a Gaussian process regression problem.

Commentary by Dr. Valentin Fuster
2015;():V07AT29A002. doi:10.1115/GT2015-42505.

Modern large frame F, G, & J class gas turbine flow path component design requires the complex integration of multiple design disciplines that traditionally reside with distinct specialists. A traditional design system for an actively cooled gas turbine blade includes aerodynamicists, heat transfer engineers, structural engineers and a failure or lifetime prediction engineers passing information through a manual process with a small number of iterations between the disciplines. Design or manufacturing engineers can also be involved ensuring manufacturability and policing best practices in a predominantly deterministic design system. Over the past few decades robust design or probabilistic design philosophies along with cluster computing advancements have accelerated the release of commercial software that allows for the automation of multiple analytical evaluations at off design points. These software codes allow for process automation of several independent codes executed multiple times at various conditions for automated design of experiment (DOE), and reliability analysis using Monte Carlo or other advanced approximate probabilistic methods across the entire design system.

In this paper, the authors are presenting a novel approach of using a commercially available process integration tool to fully integrate a series of other commercially available tools for a root cause failure analysis of a F class turbine blade rather than a new make design. The analysis incorporated a computational fluid dynamics model (CFD) to define inlet temperatures and pressure profiles, a fully conjugate thermal analysis interacting with a finite element (FEA) solver linked to a proprietary creep lifetime prediction model. A DOE was executed to define the meaningful parameters and help rule out potential causes of failure such as loss of coating or compromised cooling system as a contributing factor of the failure which greatly reduced the amount of time and money needed for the investigation. A probabilistic failure analysis was then executed and surrogate models created for quick probabilistic assessment for different operating conditions. This allowed for validation against fleet history to explain the single engine failure not previously observed. Setting up the non-deterministic models initially allowed for rapid redesign in less than 1 month time with confidence that the true root cause was identified and mitigated. It further allowed for feedback and calibrations to the traditional design system methodology.

Commentary by Dr. Valentin Fuster
2015;():V07AT29A003. doi:10.1115/GT2015-43110.

The competitive ability of jet engine maintenance companies depends mainly on turn around time and overhaul costs. Both airline and maintenance companies need the best possible accuracy regarding the prediction of emerging costs and time of engine maintenance process to secure their operation. Estimating the deterioration status of engine modules prior to disassembling is one of the greatest challenges for the maintenance process. In a pilot project a Bayesian belief network (BBN) has been developed to determine the deterioration condition of the General Electric CF6-80C2 first stage high pressure turbine (HPT) nozzle guide vane (NGV). The aim of this paper is to extend the used BBN techniques to the HPT first and second stage rotor blades and the second stage vanes. Thereby, its objective is to prove the successful application of the developed method for constructing a BBN for component hardware forecast.

The BBN is composed of following parameters: component repair history, region, on-wing cycles, airfoil material, thrust rating, engine wing position and customer segment. Performing statistical data analysis and combining these parameters with expert knowledge result in component specific BBNs. These nets provide a moderate forecast accuracy of 59 percent for the first stage rotor blades, 65 percent for the second stage rotor blades and promising 89 percent for the second stage NGVs.

The paper concludes that a BBN has very good qualities to forecast the hardware condition of HPT components impressively shown by virtue of the nozzles. Therefore, it is worth to transfer the developed method to other modules in order to accurately predict the degradation of the components in an unconventional way.

Commentary by Dr. Valentin Fuster
2015;():V07AT29A004. doi:10.1115/GT2015-43693.

Methods for efficient variance based global sensitivity analysis of complex high-dimensional problems are presented and compared. Variance decomposition methods rank inputs according to Sobol indices which can be computationally expensive to evaluate. Main and interaction effect Sobol indices can be computed efficiently in the Kennedy & O’Hagan framework with Gaussian Processes (GPs). These methods use the High Dimensional Model Representation (HDMR) concept for variance decomposition which presents a unique model representation when inputs are uncorrelated. However, when the inputs are correlated, multiple model representations may be possible leading to ambiguous sensitivity ranking with Sobol indices. In this work we present the effect of input correlation on sensitivity analysis and discuss the methods presented by Li & Rabitz in the context of Kennedy & O’ Hagan framework with GPs. Results are demonstrated on simulated and real problems for correlated and uncorrelated inputs and demonstrate the utility of variance decomposition methods for sensitivity analysis.

Commentary by Dr. Valentin Fuster


2015;():V07AT30A001. doi:10.1115/GT2015-42556.

Process fluid lubrication of rotating machinery offers advantage of compactness and efficiency while dispensing with complicated oil lubricant supply systems. Prior work in a dedicated test rig demonstrated the performance of water lubricated radial and thrust bearings into high speed and high load conditions. The application related to a high performance rocket engine turbo pump. The test rig was revamped to operate with gas bearings in a program aiming to measure the performance of gas thrust bearings. The gas bearings for lateral support of the rotor are of hybrid type (hydrostatic/hydrodynamic) with flexure pivots and multiple ports for inlet gas pressurization. The paper details the design of the flexure pivot bearings and predictions of the lateral rotordynamics of the rotor supported on the hybrid gas bearings. Troubleshooting operation of the test rotor supported on the novel gas bearings followed with preliminary runs with the bearings supplied with air at 7.9 bar, then 6.5 bar and at 5.1 bar, and shaft speeds to 25 krpm (surface speed=50 m/s). The data recorded showed a very lightly damped system with a critical speed at ∼6 krpm, and susceptible to excite sub synchronous whirl motions when operating above the first critical speed. Ignoring the initial warnings, the operator persisted in operating the rotor to a high speed of 28 krpm while lowering the air supply pressure to 5.1 bar into the bearings. Suddenly, the shaft experienced large amplitude sub synchronous whirl motions, contacted the bearings, and produced a catastrophic failure. The incident produced much damage including a broken coupling, a twisted rotor, sheared covers, and welded pads into the bearing casing. Post-mortem analysis shows the failure is due to a sub synchronous whirl instability of the first rigid body rotor-bearing mode also exacerbated by the rotor approaching second natural frequency of the rotor-bearing system. The rotordynamics model includes the rotor rigidly connected to a long quill shaft and coupling produces results in agreement with the last vibration data set acquired prior to the incident. The experience demonstrates the need for following proper operating procedures while also paying attention to early evidence that could have prevented the mishap.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A002. doi:10.1115/GT2015-42729.

Misalignment is a usual phenomenon in rotating machines. The rotor centerlines are not collinear at the couplings and the rotors operate in incorrect axial positions in a multi-span rotor. The effects of misalignment of flexible rotor system are summarized as the variation of joint stiffness and additional misalignment excitation force based on the dynamic model established.

The variation of joints stiffness is difficult to describe, meanwhile the misalignment excitation and rotor unbalance changes with different assembly and operating conditions. The distributions of these parameters which have significant effect on rotor dynamics are unknown, but the intervals of uncertain parameters are usually easier to get. An interval analysis method based on Taylor expansion and direct integration, which solves the dynamic response of rotor system under complex excitations including misalignment and multi unbalance with different frequencies and excitation points is presented. The differential equation of rotor system is formulated by combination of the matrixes of an actual rotor system finite element model and interval excitation vectors. The responses of a single spool and two spools with misalignment and unbalance are calculated by the interval analysis method. The results indicate that the method is effective and reflects some dynamic influence of misalignment and unbalance on rotor system. Second harmonic frequency appears, and rotor orbit is irregular. The response reflects the uncertain interval distribution characteristics, and the frequency components on different locations of the rotor have different characteristics.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A003. doi:10.1115/GT2015-42816.

The aim of this paper is to provide some basis for the design and assembly of a rod-fastened rotor with Hirth coupling. The rod-fastened rotor is comprised of a series of discs clamped together by a central tie rod or several tie rods on the pitch circle diameter. The key difference between a rod-fastened rotor and an integrated one is the existence of contact interfaces. The contact status of contact interface in the rod-fastened rotor is the key concern for accurate rotor dynamic analysis. Therefore, the method of accurately describing the slippage status and contact status is presented in this paper. The approach of eliminating the slippage and making the radial contact pressure distribution more uniform is also presented. According to the characteristics of Hirth coupling, one model of a turbine end rotor with Hirth coupling of a heavy duty gas turbine was built. The three-dimensional finite element contact method and non-linear behaviors such as friction were also taken into account. The effect of pre-tightening forces, centrifugal forces and overhung rim lengths on the radial slippage including initial radial slippage usi and dynamic radial slippage usd of contact interface was determined. A dimensionless coefficient cr was also defined to describe the radial contact pressure distribution of contact interface which was influenced by the values of pre-tightening forces, centrifugal forces and wheel rim lengths respectively. The results of Hirth coupling indicate that the initial radial slippage increases with the pre-tightening forces, and for a fixed pre-tightening force, usi decreased with the increase of overhung rim length. In addition, there is an optimum rim length to eliminate the dynamic radial slippage usd produced by the change of the centrifugal force. Through the analysis of contact pressure distribution, we know that the reasonable design of the load relief trough processed in the overhung rim makes the contact pressure distribution more uniform. Finally, the effect of temperature load on the radial slippage and contact pressure distribution was investigated.

Topics: Gas turbines , Rotors
Commentary by Dr. Valentin Fuster
2015;():V07AT30A004. doi:10.1115/GT2015-42862.

Turbochargers are commonly used in reciprocating compressors and internal combustion engines to improve overall efficiency, thereby reducing fuel requirements. In reciprocating compressor applications, turbochargers typically operate in the range of 15,000 rpm to 30,000 rpm. These turbomachines operate at higher rotational speed in automotive applications, often exceeding 100,000 rpm. These high speeds result in bearings that are often highly non-linear, with large limit cycle shaft orbits and high subsynchronous vibration levels. These large orbits also result in much higher power losses than would be observed for the same bearing with low vibration levels. These devices are often used in automotive applications, where there are significant cost pressures, ruling out more expensive bearing options such as tilting pad bearings. There is a need for bearing designs that are effective at stabilizing turbochargers and are also low cost.

In this paper, a theoretical turbocharger for a reciprocating compressor application is considered. The initial design incorporated plain axial groove bearings. Several replacement bearing options were considered, including pressure dam bearings and tilting pad bearings to improve rotor stability. The pressure dam bearing is used to impose a load on the shaft, making it run off center. This feature reduces subsynchronous whirl instability and also reduces power loss. The tilting pad bearing eliminates self-excited forces from the bearings but has increased power loss when compared to fixed-pad bearings. Starved lubrication of the tilting pad bearing reduced the power loss but also reduced the stability margin. The application considered for this paper is a larger turbocharger with a rotational speed of 25,000 rpm, and is unstable with conventional bearings.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A005. doi:10.1115/GT2015-42956.

For rotors supported by active magnetic bearings (AMBs), clearance bearings are commonly used to provide backup support under loss of AMB functionality. Test data from real machines shows that rotor vibration during touchdown on backup bearings may involve steady forward whirling at a sub-synchronous frequency. This excitation is believed to be due to friction forces transmitted between the rotor and a bearing end-face under axial load. This paper proposes a new analytical approach to model and predict such friction-driven forward whirl behaviors. A set of constraint equations are derived that relate a circular whirl motion of arbitrary orbital speed to the frequency response functions of the rotor-housing structure. This model is coupled with an evaluation of Coulomb friction associated with slip between the rotor and the supporting end-face of a thrust bearing. The resulting equations can be used to compute a set of possible whirl motions via a root-finding procedure. A case study is undertaken for a 140 kW energy storage flywheel. Model-based predictions are compared with measured data from spin-down tests and show a good level of agreement. The study confirms the role of friction-related forces in driving forward-whirl response behaviors. It also highlights the key role of housing and machine support characteristics in response behavior. This influence is shown to be complex and not open to simple physical interpretation. Therefore, the proposed analytical method is seen as a useful tool to investigate this influence while avoiding the need for time consuming numerical simulations.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A006. doi:10.1115/GT2015-42977.

With the ever increasing of the centrifugal compressor capability, such as large scale LNG and CO2 re-injection, the stability margin evaluation and instability control is crucial both for the centrifugal compressor OEM and end users. An industrial scale experimental test rig with two 5-pads tilting pad journal bearings and two active magnetic bearings was introduced. Tilting pad bearings were used to support the rotor, and active magnetic bearings were used to exert static load and sub-synchronous sine load to excite the rotor. The logarithmic decrement identification method of rotor-bearing system was presented. Here, weighted instrument variable and directional frequency response function were combined to improve anti-noise performance and separate forward/ backward whirl vibration. Thermoelastohydrodynamic bearing model was used to investigate the effect of hot clearance on the stability prediction of centrifugal compressor. Also, the contribution of bearing specific load on the stability was studied numerically. Also, experimental work was carried out to investigate the damping ratio of rotor under different rotating speed and bearing specific loads. The numerical results correlate with the test results very well. The results of the investigation indicate that: (i) Hot clearance is very important for centrifugal compressor rotordynamic instability prediction; (ii) The effect of pad mechanical deformation are smaller than that of thermal deformation; (iii) With the increase of bearing specific load, the logarithmic decrement is decreasing.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A007. doi:10.1115/GT2015-43096.

Squeeze Film Dampers (SFDs) are effective to ameliorate shaft vibration amplitudes and to suppress instabilities in rotor-bearing systems. Compact aero jet engines implement ultra-short length SFDs (L/D ≤ 0.2) to satisfy stringent weight and space demands with low parts count. This paper describes a test campaign to identify the dynamic forced response of an open ends SFD (L=25.4 mm, D=125.7 mm), single film land and oil fed through three holes (120° apart), operating with similar conditions as in an aircraft engine. Two journals make for two SFD films with clearances cA=0.129 mm and cB=0.254 mm (small and large). The total oil wetted length equals Ltot=36.8 mm that includes deep end grooves, width and depth = 2.5 × 3.8 mm, for installation of end seals. In the current experiments, the end seals are not in place. A hydraulic static loader pulls the bearing cartridge (BC) to a preset static eccentricity (eS) and two electromagnetic shakers excite the BC with single frequency loads to create circular orbits, centered and off-centered, over a prescribed frequency range ω=10–100Hz. The whirl amplitudes range from r=0.05cA–0.6cA and r=0.15cB–0.75cB while the static eccentricity increases to eS=0.5cA and eS=0.75cB, respectively. Comparisons of force coefficients between the two identical dampers with differing clearances show that the small clearance damper (cA) provides ∼4 times more damping and ∼1.8 times the inertia coefficients than the damper with large clearance (cB). The test results demonstrate damping scales with ∼1/c3 and inertia with ∼1/c, as theory also shows. Analysis of the measured film land pressures evidence that the deep end grooves contribute to the generation of dynamic pressures enhancing the dynamic forced response of the test SFDs. A thin film flow model with an effective groove depth delivers predictions that closely match the test damping and inertia coefficients. Other predictions, based on the short length bearing model, use an effective length Leff ∼1.17L to deliver damping coefficients 15% larger than the experimental results; however, inertia coefficients are ½ of the identified magnitudes. The experiments and analysis complement earlier experimental work conducted with centrally grooved SFDs.

Topics: Dampers
Commentary by Dr. Valentin Fuster
2015;():V07AT30A008. doi:10.1115/GT2015-43334.

Rotor system of modern aero-engine together with the case mounted on the wing represents a uniform system that is supposed to be considered as a rotor-bearings-foundation structure. Nowadays rotating machinery such as modern aircraft engines usually designed, marketed and sold for the most part based on analytical and numerical predictions. In such a way methods to incorporate the foundation effect in rotordynamic calculations are very important. For investigation purposes a rotor-foundation test rig, which can simulate the aero-engine’s typical operating condition such as wing vibration and hard landing, was built to study the influence of foundation behavior on the dynamic characteristics of rotor system. To predict natural frequencies for the full system simplified models based on FEM approach were created. Moreover, simple numerical model was created to study influence of foundation kinematic excitation on behavior of rotor disk orbit. Furthermore simulation results were compared with experimental to understand influence of main parameters which define foundation vibration on rotor deviation from normal operation condition. Obtained numerical and experimental results can help to understand principles of rotating structure-foundation interaction when the later one subjected to excitation and could be further used for improving of more complicated models for design and enhancement of aircraft engines.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A009. doi:10.1115/GT2015-43337.

The dynamic analysis of the multi-shaft turbine rotor equipped with a spur gear pair for the various gear parameters is studied. Main components of the multi-shaft turbine rotor system include the outer shaft, the inner shaft, the impeller shaft, the oil shaft and the ball bearings. The global assumed mode method (GAMM) is applied to model the rotor motion and the system equation of motion is formulated using Lagrange’s approach. The dynamic behavior of the geared multi-shaft turbine rotor system includes the natural frequency, mode shape and unbalanced response. Numerical results show that large vibration amplitude is observed in steady state at self-excited rotating speed adjacent to the natural frequency. There is no influence of the various pressure angle, modulus, and modification coefficients on unbalance response. Contrary to above cases, the variation of the system unbalance response is dominated by the tooth types rather than the other gear parameters.

Topics: Gears , Rotors , Turbines
Commentary by Dr. Valentin Fuster
2015;():V07AT30A010. doi:10.1115/GT2015-43405.

It was surprisingly reported that a generator rotor could not be balanced to an acceptable vibration level by weights at two balance planes at the drive end (DE) and the non-drive end (NDE) fan rings. Both real measured vibration data and rotordynamic calculations indicate that the rotor at rated speed of 3600 rpm appears to run just above the 2nd critical speed (couple or conical mode). However, couple weights (same weights placed at both DE and NDE with 180-degree-out-of-phase) have little effect on 1X vibration response. A third balance plane had to be utilized to effectively reduce vibration. This paper uses measured data and rotordynamic modelling to explain these findings. It is found that the 4th mode could affect synchronous vibration response at rated speed significantly besides the 3rd mode. The two fan ring balance planes appear to be near the nodal points of the 4th mode, which explains ineffectiveness of the couple weights to vibration response at rated speed in the field. Measured data from real machines including influence vectors are presented from third balance planes such as the coupling and the exciter ends, besides the fan ring wheels. The 3rd and 4th rotordynamic modes are also given along with unbalance response studies.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A011. doi:10.1115/GT2015-43434.

This paper presents an evaluation of various rotordynamic parameters at commercial vehicle turbochargers, which are operated supercritically in full-floating hydrodynamic journal bearing systems. The evaluation is conducted by using an experimental approach to determine the performance of the rotor-bearing-system in a real-life assembly at a hot gas test bench. This takes support stiffness, external heating and the excitation by seals, thrust bearings and gas forces into account, while Engine-specific excitation is not present. The system’s ability to carry additional unbalance load at different oil support pressures without the occurrence of mixed friction throughout a complete run-up is assessed. By executing this assessment for multiple assemblies with different bearings, rotors and oil types, the influence of main design and boundary parameters on the effective journal bearing performance of turbochargers is quantified.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A012. doi:10.1115/GT2015-43481.

The ability to accurately predict the response of rotating machinery to external forces and to assess system-level stability for different modes is crucial from a reliability and preventive maintenance perspective. Geared systems, in particular, contain many complexities, which may lead to instability and even chaotic vibration behavior. No methods for determining the effects of the dynamic meshing forces on the vibrations of complete shaft/bearing systems have been proposed in the literature. Several time-transient and steady-state models for analyzing gear forces and deflections have been proposed, but they focus primarily on the dynamics of the gearbox itself and neglect vibration transmission through the remainder of the drive-train. Models that do incorporate other components of the drive-train propose simplified lumped-parameter models for the shafts and bearings. Recent models have used the finite element method to couple the lateral, torsional, and axial degrees-of-freedom of geared shaft systems to the forces and moments exchanged between the gears via stiffness matrices. Other models in literature capture the backlash non-linearity and the state-varying mesh stiffness and observe the time-transient response of the gearbox and simplified shaft/bearing structure. A finite element formulation of complete geared systems, which couples the axial, lateral, and torsional degrees-of-freedom, is developed in which the shaft is modeled with Timoshenko beam elements and captures the forces and moments due to gyroscopic effects, and rotational accelerations due to start-up. It includes the capability of modeling non-linear contact loss due to backlash clearance and parametric excitations resulting from the state-varying mesh stiffness and solves the time-transient state equations for the displacements and velocities of the shafts using the direct Runge-Kutta method.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A013. doi:10.1115/GT2015-43534.

The calculated dynamic response of an excited rotating system can be significantly affected by uncertainties in its mechanical properties, such as mass, stiffness, geometrical imperfections, or loadings. For this reason, it is essential to understand and quantify the influence of uncertain parameters on the predicted rotor response.

This paper aims to optimize the propagation of random input uncertainties for a rotordynamic problem and assess their influence on the dynamic behaviour of an unbalanced rotor. The Harmonic balance method (HBM) and a non-intrusive Polynomial Chaos Expansion (PCE) are used to evaluate the stochastic response of a finite element rotor. The proposed stochastic approach is based on a numerical quadrature calculation of integrals for finding the coefficients of the PCE.

The method is initially applied to evaluate the stochastic response of a linear rotodynamic system, leading to the original concept of stochastic Campbell diagram and further extended to nonlinear rotordynamic problems, using the Asymptotic Numerical Method (ANM).

Commentary by Dr. Valentin Fuster
2015;():V07AT30A014. doi:10.1115/GT2015-43547.

Gas-expanded lubricants (GELs) have the potential to increase bearing energy efficiency, long-term reliability, and provide for a degree of control over the rotordynamics of high-speed rotating machines. Previous work has shown that these tunable mixtures of synthetic oil and dissolved carbon dioxide could be used to maximize the stability margin of a machine during startup by controlling bearing stiffness and damping. This allows the user to then modify the fluid properties after reaching a steady operating speed to minimize bearing power loss and reduce operating temperatures. However, it is unknown how a typical machine would respond to rapid changes in bearing stiffness and damping due to changes in the fluid properties once the machine has completed startup. In this work, the time-transient behavior of a high-speed compressor was evaluated numerically to examine the effects of rapidly changing bearing dynamics on rotordynamic performance. Two cases were evaluated for an 8-stage centrifugal compressor: an assessment under stable operating conditions as well as a study of the instability threshold. These case studies presented two contrasting sets of transient operating conditions to evaluate, the first being critical to the viability of using GELs in high-speed rotating machinery. The fluid transitions studied for machine performance were between that of a polyol ester synthetic lubricant and a GEL with a 20% carbon dioxide content. The performance simulations were carried out using a steady-state thermoelastohydrodynamic (TEHD) bearing model, which provided bearing stiffness and damping coefficients as inputs to a time-transient rotordynamic model using Timoshenko beam finite elements. The displacements and velocities of each node were solved for using a fourth order Runge-Kutta method and provided information on the response of the rotating machine due to rapid changes in bearing stiffness and damping coefficients. These changes were assumed to be rapid due to 1) the short lubricant residence times calculated for the bearings, and 2) rapid mixing due to high shear rates in the machine bearings causing sudden changes in the fluid properties. This operating condition was also considered to be a worst-case scenario as an abrupt change in the bearing dynamics would likely solicit a more extreme rotordynamic response than a more gradual change, making this analysis quite important. The results of this study provide critical insight into the nature of operating a rotating machine and controlling its behavior using gas-expanded lubricants, which will be vital to the implementation of this technology.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A015. doi:10.1115/GT2015-43803.

A state-of-the-art, rotary-disc atomizer driven by a permanent-magnet electric motor and supported by active magnetic bearings (AMB) was designed, fabricated, and tested as part of a spray-dryer system within a pharmaceutical-processing plant. The atomizing process imposed several challenges on the AMBs, including large, highly-dynamic rotor imbalances and large, quasi-periodic external radial impulses. Several design changes were systematically implemented to mitigate the effects of large rotor imbalances. A novel impulse detection and recovery system was introduced to alleviate the effects of external impulses. These changes, which have steadily improved the operability and reliability of the machine, are described here along with field test data.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A016. doi:10.1115/GT2015-44115.

In turbomachines, the transfer of energy between the rotor and the fluid does not — in theory — result in lateral forces on the rotor. In positive displacement machines, on the other hand, the transfer of energy between the moving and stationary components usually results in unbalanced pressure fields and forces. In [1] the authors developed a model to predict the dynamic forces in twin screw pumps, showing that the helical screw shape generates hydraulic forces that oscillate at multiples of running speed.

The work presented here attempts to validate the model in [1] using a clear-casing twin screw pump. The pump runs in both single and multiphase conditions with exit pressure up to 300 KPa and a flow rate 0.6 liter per second. The pump was instrumented with dynamic pressure probes across the axial length of the screw in two perpendicular directions to validate the dynamic model. Two proximity probes measured the dynamic rotor displacement at the outlet to validate the rotordynamics model and the hydrodynamic cyclic forces predicted in [1].

The predictions were found in good agreement with the measurements. The amplitude of the dynamic pressure measurements in two perpendicular plans supported the main assumptions of the model (constant pressure inside the chambers and linear pressure drop across the screw lands). The predicted rotor orbits at the pump outlet in the middle of the rotor matched the experimental orbits closely. The spectrum of the response showed harmonics of the running speed as predicted by the model.

The pump rotor’s calculated critical speed was at 24.8 krpm, roughly 14 times the rotor’s running speed of 1750 rpm. The measured and observed excitation frequencies extended out to nine times running speed, still well below the 1st critical speed. However, for longer twin-screw pumps running at higher speed, the coincidence of a higher-harmonic excitation frequency with the lightly damped 1st critical speed should be considered.

Commentary by Dr. Valentin Fuster
2015;():V07AT30A017. doi:10.1115/GT2015-44129.

The paper presents an analysis methodology in order to investigate coupling between rotating shaft and bladed disks under anisotropic supporting structures. Solid models of the bladed disks are first developed as reference models in the rotating frame in order to take into account the non-linear geometrical stiffening under centrifugal loading. A correlation procedure is then used in order to elaborate equivalent axisymmetric models based on Fourier multi-harmonic finite elements. These equivalent models are then transferred in the inertial frame and coupled with the anisotropic supporting structures. The analyses in the inertial frame show coupling between the flexible shaft and the equivalent bladed disks. Veering phenomena occur when the rotation increases due to bladed disk flexibility. Mixed backward and forward modes due to the anisotropic supporting devices are also investigated.

Commentary by Dr. Valentin Fuster

Bearing and Seal Dynamics

2015;():V07AT31A001. doi:10.1115/GT2015-42190.

This paper proposes and studies the non-parametric system identification of a foil-air bearing (FAB) and its application to the frequency-domain nonlinear analysis of a foil-air bearing rotor system. This research is motivated by two advantages: (i) it removes computational limitations by replacing the air film and foil structure state equations by a displacement/force relationship; (ii) if the identification is based on empirical data, it can capture complications that cannot be easily modelled. A numerical model of the FAB is identified using a recurrent neural network (RNN). The training data sets are taken from the simultaneous time domain solution of the air film, foil and rotor equations. The RNN FAB model identified at a single speed is then validated over a range of speeds in two ways: (i) by subjecting it to several sets of input-output data that are different from those used in training; (ii) by using it in the harmonic balance (HB) solution process for the unbalance response of a rotor-bearing system. In either case, the test results using the identified model show good agreement with the exact results obtained using the air film and foil equations, demonstrating the great potential of this method, in the absence of self-excitation effects.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A002. doi:10.1115/GT2015-42289.

Static and dynamic load tests were performed on a three-pad, rocker-pivot, tilting-pad journal bearing (TPJB) with three interchangeable pad configurations, each with measurably different pad flexibilities. Measured dynamic-stiffness data for the bearing were readily fitted by a frequency-independent, constant-coefficient [K][C][M] model.

The test bearing had a 101 .74 mm diameter with L/D = 0.6 Tests were conducted over the speed range of 6 to 12 krpm, with unit loads varying from .172 to 1.724 MPa. An ISO VG 46 lubricant was used as the test fluid.

Pad flexibility was characterized as the change in the pad’s bending stiffness or the change in pad thickness. A finite-element model was created to predict the structural bending stiffness of each pad configuration, showing a significant pad-flexibility increase as pad thickness decreased.

To examine the effect of pad flexibility on the rotordynamic coefficients, the measured results were compared across pad configurations and showed that the pad-flexibility increase reduced the direct damping coefficients by 12–20%. As pad flexibility increased, the direct stiffness coefficients could increase or decrease, depending on the unit load. They varied from an increase of 12% at low unit loads to a decrease of 3% at high unit loads.

Results show that the pad’s structural bending stiffness or flexibility is important when predicting the bearing’s dynamic performance. Damping is consistently over-predicted when neglecting pad flexibility.

A non-dimensional pad flexibility parameter αflex was developed. It related the average deflection across the pad surface to the pad’s arc length and was to relate the pad flexibility of multiple bearings of different sizes.

A bearing code was used to predict the percent change in direct damping coefficients for rigid-pad/flexible-pivot and flexible-pad/flexible-pivot models for a surface speed of 54 m/s and a unit load of 783 kPa for the three pad configurations tested here plus five additional tested bearings from the literature. For the minimum-pad-thickness configuration tested here, the code predicted a 20% drop in predicted Cxx (off-load-axis direct damping) when comparing a model that included pad flexibility with a model that neglected pad flexibility. In terms of αflex, the two thinnest pad configurations tested here are quite flexible compared to both TPJB’s pads used in industry and previously-tested TPJB pads.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A003. doi:10.1115/GT2015-42331.

Static and thermal characteristics (measured and predicted) are presented for a 4-pad, spherical-seat, TPJB with 0.5 pivot offset, 0.6 L/D, 101.6 mm nominal diameter, and 0.3 preload in the load-between-pivots orientation. One bearing is tested four separate times in the following four different lubrication configurations: (1) flooded single-orifice (SO) at the bearing shell, (2) evacuated leading edge groove (LEG), (3) evacuated spray-bar blocker (SBB), and (4) evacuated spray-bar (SB). The LEG, SBB, and SB are all considered methods of “directed lubrication”. These methods rely on lubrication injected directly to the pad/rotor interface. The same set of pads is used for every test to maintain clearance and preload; each method of lubrication is added as an assembly to the bearing. Test conditions include surface speeds and unit loads up to 85 m/s and 2.9 MPa respectively. Static data includes rotor-bearing eccentricities, and attitude angles. Thermal data include measured temperatures from sixteen bearing thermocouples. Twelve of the bearing thermocouples are embedded in the babbitt layer of the pads while the remaining four are oriented at the leading and trailing edge of the loaded pads exposed to the lubricant. Bearing thermocouples provide a circumferential and axial temperature gradient. The pivot stiffness (pad and pivot in series) is measured and incorporated into predictions.

Topics: Lubrication
Commentary by Dr. Valentin Fuster
2015;():V07AT31A004. doi:10.1115/GT2015-42332.

This paper studies the effect of brush seal segmentation on the seal performance characteristics. A brush-labyrinth sealing configuration arranged of one brush seal downstream and two labyrinth fins upstream is studied experimentally and theoretically. The studied brush seal is of welded design installed with zero cold radial clearance. The brush seal front and back rings as well as the bristle pack are segmented radially in a single plane using the electrical discharge machining technique. The segmentation procedure results in loss of bristles at the site of the cuts altering the leakage flow structure in the seal and its performance characteristics. Two test rigs are used to obtain leakage, as well as rotordynamic stiffness and damping coefficients of the seal at different pressure ratios. The CFD-based model is used to predict the seal performance and to study in detail local changes in the flow field due to the segmentation. A back-to-back comparison of the performance of non-segmented and segmented brush seals, as well as baseline labyrinth seal is provided. The obtained results demonstrate that the segmentation in general negatively affects the performance of the studied brush-labyrinth sealing configuration. However, the segmented brush seal shows increased direct damping coefficients.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A005. doi:10.1115/GT2015-42336.

Measured and predicted rotordynamic characteristics are presented for a 4-pad, spherical-seat, TPJB with 0.5 pivot offset, 0.6 L/D, 101.6 mm nominal diameter, and 0.3 preload in the load-between-pivots orientation. One bearing is tested four separate times in the following four different lubrication configurations: (1) flooded single-orifice (SO) at the bearing shell, (2) evacuated leading edge groove (LEG), (3) evacuated spray-bar blocker (SBB), and (4) evacuated spray-bar (SB). The same set of pads is used for every test to maintain clearance and preload; each method of lubrication is added as an assembly to the bearing. Test conditions include surface speeds and unit loads up to 85 m/s and 2.9 MPa respectively. Dynamic data includes four sets (one set for each bearing configuration) of direct and cross-coupled rotordynamic coefficients derived from measurements and fit to a frequency independent KCM model. The pivot stiffness (pad and pivot in series) is measured and incorporated into predictions.

Topics: Lubrication
Commentary by Dr. Valentin Fuster
2015;():V07AT31A006. doi:10.1115/GT2015-42461.

In high speed gearbox systems, the lubrication is generally provided using nozzles to create small oil jets that feed oil into the meshing zone. It is essential that the gear teeth are properly lubricated and that enough oil gets into the tooth spaces to permit sufficient cooling and prevent gearbox failure. A good understanding of the oil behaviour inside the gearbox is therefore desirable, to minimize lubrication losses and reduce the oil volume involved, and ensure gearbox reliability.

In order to reach these objectives, a comprehensive numerical study of a single oil jet impinging radially on a single spur gear teeth has been carried out using the Volume of Fluid (VOF) method. The aims of this study are to evaluate the resistant torque produced by the oil jet lubrication, and to develop a physical understanding of the losses deriving from the oil-gear interaction, studying the droplets and ligaments formation produced by the breaking up of the jet as well as the formation of an oil film on the surface of the teeth.

URANS calculations have been performed with the commercial code ANSYS FLUENT and an adaptive mesh approach has been developed as a way of significantly reducing the simulation costs. This method allows an automatic mesh refinement and/or coarsening at the air-oil interface based on the volume of fluid gradient, increasing the accuracy of the predictions of oil break-up as well as minimizing numerical diffusion of the interface. A global sensitivity analysis of adopted models has been carried out and a numerical set-up has been defined. Finally several simulations varying the oil injection angle have been performed, in order to evaluate how this parameter affects the resistant torque and the lubrication performances.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A007. doi:10.1115/GT2015-42472.

The present work is focused on the pneumatic hammer instability in an aerostatic bearing with shallow recesses and orifices of four different diameters. Operating conditions were zero rotation speed, zero load and different supply pressures. The diameters of the tested orifices were large compared to the usual practice and correspond to a combined inherent and orifice restriction. The theoretical analysis was based on the CFD evaluation of the ratio between the recess and the feeding pressure and on the “bulk flow” calculation of the rotordynamic coefficients of the aerostatic bearing. Calculations showed an increase of the direct stiffness with decreasing the orifice diameter and increasing the supply pressure and, on the other hand, a decrease toward negative values of the direct damping with decreasing the orifice diameter. These negative values of the direct damping coefficient indicate pneumatic hammer instabilities. In parallel, experiments were performed on a floating bearing test rig. Direct stiffness and damping coefficients were identified from multiple frequency excitations applied by a single shaker. Experiments were performed only for the three largest orifices and confirmed the decrease of the direct damping with the orifice diameter and the supply pressure. The identification of the rotordynamic coefficients was not possible for the smallest available orifice because the aerostatic bearing showed self-sustained vibrations for all feeding pressures. These self-sustained vibrations are considered the signature of the pneumatic hammer instability. The paper demonstrates that aerostatic bearings with shallow recesses and free of pneumatic hammer instabilities can be designed by adopting orifice restrictors of large size diameter.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A008. doi:10.1115/GT2015-42483.

Tilting-pad journal bearings (TPJBs) support highly loaded high speed rotors with high requirements on rotordynamic behavior. Typical applications therefore are integrally geared compressors, where the gear force of the high speed pinions (HSPs) is at least one magnitude higher than the gravity force.

A continuously rising demand for increasing the overall efficiency of integrally geared compressors leads to a necessity to expand TPJBs operation limits. Allowable limits of actual bearings often result in limitations for thermodynamical compressor design. In addition, bearings generate 20% to 40% of the mechanical power losses in integrally geared compressors. Improvements in the bearing design have to be performed in order to meet the challenges of rising loads and speeds.

This paper presents an optimization of TPJBs for integrally geared compressors to meet the further demand of higher operation limits.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A009. doi:10.1115/GT2015-42577.

A modified Reynolds equation for bump type gas foil thrust bearing was established with consideration of the gas rarefaction coefficient. Under rarefied gas lubrication, the Knudsen number which was affected by the film thickness and pressure was introduced to the Reynolds equation. The coupled modified Reynolds and lubricating film thickness equations were solved using Newton-Raphson Iterative Method and Finite Difference Method. By calculating the load capacity for increasing rotor speeds, the lift-off speed under certain static load was obtained. Parametric studies for a series of structural parameters and assembled clearances were carried out for bearing optimization design. The results indicate that with gas rarefaction effect, the axial load capacity would be decreased, and the lift-off speed would be improved. The rarefied gas has a more remarkable impact under a lower rotating speed and a smaller foil compliance coefficient. When the assembled clearance of the thrust bearing rotor system lies in a small value, the lift-off speed increases dramatically as the assembled clearance decreases further. Therefore, the axial clearance should be controlled carefully in assembling the foil thrust bearing. It’s worth noting that the linear uniform bump foil stiffness model is not exact for large foil compliance ∼0.5, especially for lift-off speed analysis, due to ignoring the interaction between bumps and bending stiffness of the foil.

Topics: Thrust bearings
Commentary by Dr. Valentin Fuster
2015;():V07AT31A010. doi:10.1115/GT2015-42735.

This paper presents a novel method for identifying the dynamic parameters of a gas bearing, whose force coefficients are strong functions of frequency. The method is based on the analysis of the phase diagram with the model assuming a mass-damper-spring system with time-dependent force coefficients. Usually, it is necessary a controlled mechanism to find the transfer function, this condition limits the application of the method. On the other hand, estimation of the damping and stiffness parameters under real loading is very cumbersome and requires a special care on identifying the excitation forces. One of the main difficulties is the isolation of noise and those vibration signals with an unidentified source. In this work, the excitation force was taken from the unbalance loading of a rotor test. Therefore, there is no need for a special test rig. The dynamic parameters can be estimated analyzing data from the actual rotor mounted on the gas bearings. Identifying the parameters that cause gas bearing instabilities is a big challenge. The gas properties are very sensitive to temperature and pressure changes, and, as a consequence the bearing rotor-dynamic coefficients change drastically and the rotor behaves chaotically, which means that the dynamic parameters are nonlinear. In this research a methodology based on the phase diagram construction to identify nonlinear instabilities of gas bearings is presented. The results show the method capability to estimate the dynamic coefficients by the analysis of the energy variation.

Among nonparametric methods, the phase diagram or phase space is in use to identify nonlinearities in dynamic systems. The identification is conducted through the analysis of the energy variations. The energy variations can be represented as a three dimensional function E(x,v,t). In this way the phase diagram can be related to the frequency and the dynamic parameters of the system. According to Taken’s theorem, a dynamic system can be obtained by reconstructing the phase diagram. Then, using this method, the damping and stiffness coefficients are estimated.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A011. doi:10.1115/GT2015-42776.

Tilting pad journal bearings (TPJBs) supporting high performance turbomachinery rotors have undergone steady design improvements to satisfy ever stringent operating conditions that include large specific loads due to smaller footprints, and high surface speeds that promote flow turbulence and thus larger drag power losses. Simultaneously, predictive models continuously evolve to include minute details on bearing geometry, pads and pivots’ configurations, oil delivery systems, etc. In general, predicted TPJB rotordynamic force coefficients correlate well with experimental data for operation with small to moderately large unit loads (1.7 MPa). Experiments also demonstrate bearing dynamic stiffnesses are frequency dependent, best fitted with a stiffness-mass like model whereas damping coefficients are adequately represented as of viscous type. However, for operation with large specific loads (> 1.7 MPa), poor correlation of predictions to measured force coefficients is common. Recently, an experimental effort [1] produced test data for three TPJB sets, each having three pads of unequal thickness, to quantify the effect of pad flexibility on the bearings’ force coefficients, in particular damping, over a range of load and rotational speed conditions. This paper introduces a fluid film flow model accounting for both pivot and pad flexibility to predict the bearing journal eccentricity, drag power loss, lubricant temperature rise and force coefficients of typical TPJBs. A finite element pad structural model including the Babbitt layer is coupled to the thin film flow model to determine the mechanical deformation of the pad surface. Predictions correlate favorably with test data, also demonstrating that pad flexibility produces a reduction of up to 34% in damping for the bearing with the thinnest pads relative to that with the thickest pads. A parametric study follows to quantify the influence of pad thickness on the rotordynamic force coefficients of a sample TPJB with three pads of increasing preload, rp = 0, 0.25 (baseline) and 0.5. The bearing pads are either rigid or flexible by varying their thickness. For design considerations, dimensionless static and dynamic characteristics of the bearings are presented versus the Sommerfeld number (S). Pad flexibility shows a more pronounced effect on the journal eccentricity and the force coefficients of a TPJB with null pad preload than for the bearings with larger pad preloads (0.25 and 0.5), in particular for operation with a small load or at a high surface speed (S>0.8).

Commentary by Dr. Valentin Fuster
2015;():V07AT31A012. doi:10.1115/GT2015-42806.

The 2D Reynolds equation is the traditional method used for solving hydrodynamic lubrication problems, whereas the full 3D Navier-Stokes (NS) equation has been a new and attractive approach in recent years. Unfortunately, the conventional cavitation models which were successful in Reynolds equation have encountered difficulties when they are implemented in NS equation, thus the proper modeling of cavitation has been a fundamental and important issue for the application of NS equation in bearing problems. A novel cavitation model was derived by the authors, which is based on the mechanism of the gaseous cavitation in submerged bearings. In this paper, this cavitation model is implemented in both 2D Reynolds and 3D NS equations. Different governing equations with various cavitation models are then used to analyze two typical oil-film bearings, i.e. a plain journal bearing and a pocketed thrust washer, so as to illustrate the advantage of the presented gaseous cavitation model. It is found that the gaseous model shows accurate prediction of the bearing overall performance as well as the cavitation region for different bearings, no matter which governing equation it is applied with. Especially, when used with the NS equation, the gaseous cavitation model has distinct advantages in accuracy and robustness compared with the commonly used Half-Sommerfeld model and the Rayleigh-Plesset model. Thus it is concluded that this model provides a universal and efficient approach for modeling cavitation in both 2D Reynolds and 3D NS equations, and it will help to prompt the future application of NS equation in more complex bearing problems.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A013. doi:10.1115/GT2015-42923.

Pocket damper seals are used as replacements for labyrinth seals in high-pressure centrifugal compressors at the balance piston location or center seal location to enhance rotordynamic stability. A concern exists that this enhanced stability will be lost at high positive inlet preswirl.

Numerical results of frequency-dependent rotordynamic force coefficients and leakage flow rates were presented and compared for a fully-partitioned pocket damper seal (FPDS) and a labyrinth seal at high positive and negative inlet preswirl, using a proposed transient CFD method based on the multi-frequency elliptical orbit whirling model. The negative preswirl indicates a fluid swirl in a direction opposite to rotor rotation at seal inlet. Both seals have identical diameter and sealing clearance. The full 3D concentric CFD model and mesh were built for the labyrinth seal and FPDS, respectively. The accuracy and availability of the present transient CFD numerical method were demonstrated with the experiment data of frequency-dependent rotordynamic coefficients of the labyrinth seal and FPDS at zero and high positive preswirl conditions. The numerical boundary conditions include two high positive preswirl, two high negative preswirl and a zero preswirl. Numerical results show that the effect of inlet preswirl on the direct force coefficients is weak, but the effect on the cross-coupling stiffness and effective damping is dramatic. Both two seals possess negative effective damping at lower excitation frequencies due to positive preswirl, and the crossover frequency of effective damping term increases with increasing positive preswirl. Negative preswirl produces negative cross-coupling stiffness and positive effective damping over the whole excitation frequency range. Increasing negative preswirl is a stabilizing factor for annular gas seals which results in a significant increase in the effective damping and a decrease in the crossover frequency. It is desirable to reduce the inlet preswirl to zero or even negative through applications of negative-swirl brakes and negative injection devices.

Topics: Dampers
Commentary by Dr. Valentin Fuster
2015;():V07AT31A014. doi:10.1115/GT2015-43095.

The most recent development in centrifugal compressor technology is towards wet gas operating conditions. This means the centrifugal compressor has to manage a liquid phase which is varying between 0 to 3% Liquid Volume Fraction (LVF) according to the most widely agreed definition. The centrifugal compressor operation is challenged by the liquid presence with respect to all the main aspects (e.g. thermodynamics, material selection, thrust load) and especially from a rotordynamic viewpoint. The main test results of a centrifugal compressor tested in a special wet gas loop [1] show that wet gas compression (without an upstream separation) is a viable technology. In wet gas conditions the rotordynamic behavior could be impacted by the liquid presence both from a critical speed viewpoint and stability wise. Moreover the major rotordynamic results from the previous mentioned test campaign [2] show that both vibrations when crossing the rotor first critical speed and stability (tested through a magnetic exciter) are not critically affected by the liquid phase. Additionally it was found that the liquid may affect the vibration behavior by partially flooding the internal annular seals and causing a sort of forced excitation phenomenon.

In order to better understand the wet gas test outcomes, the authors performed an extensive CFD analysis simulating all the different types of balance piston annular seals used (namely a Tooth on Stator Labyrinth Seal and a Pocket Damper Seal). They were simulated in both steady state and transient conditions and finally compared in terms of liquid management capability.

CFD simulation after a proper tuning (especially in terms of LVF level) showed interesting results which are mostly consistent with the experimental outcome. The results also provide a physical explanation of the behavior of both seals, which was observed during testing.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A015. doi:10.1115/GT2015-43161.

In recent years, small-size aerodynamic bearings for turbomachines such as blowers and compressors have attracted considerable attention for increasing rotational speed. These kinds of bearings require excellent stability at high speeds and durability in a high-temperature environment. Foil bearings are one of the most suitable candidates that can satisfy these requirements but their structure is very complicated, and it is difficult to control their manufacturing accuracy. It is well known that flexibly supported herringbone-grooved aerodynamic journal bearings have excellent stability at high speeds and they are relatively easy to manufacture compared with foil bearings. Moreover, their dynamic characteristics can be easily solved numerically. In this paper, a flexibly supported herringbone-grooved aerodynamic journal bearing using straight spring wires made of stainless steel is proposed to provide a simple and reliable support system for a bearing bush. Six straight spring wires were assembled into a hexagonal shape into which the bearing bush was inserted. The threshold speed of instability of the proposed aerodynamic bearing was investigated numerically and experimentally. For this investigation, the nonlinear orbit method was adopted in numerical calculations. This investigation found that straight spring wires could steadily support the bearing bush and provide a simple and reliable support system for the bearing bush and that a 6-mm-diameter rigid rotor with a mass of 4.8 g supported by the proposed aerodynamic journal bearings could stably rotate at speeds of more than 0.7 million rpm.

Topics: Wire , Bearings , Rotors , Springs
Commentary by Dr. Valentin Fuster
2015;():V07AT31A016. doi:10.1115/GT2015-43242.

Rotordynamic and leakage data are presented for a see-through tooth-on-rotor (TOR) labyrinth seal with comparisons to a see-through tooth-on-stator (TOS) labyrinth seal. Measurements for both seals are also compared to predictions from XLLaby. Both seals have identical diameter and can be considered as relatively long labyrinth seals. The TOR seal has a length-to-diameter ratio of 0.62, whereas the TOS seal is longer and has a length-to-diameter ratio of 0.75. Both seals also differ by number of teeth, tooth height, and tooth cavity length. TOR labyrinth tests were carried out at an inlet pressure of 70 bar-a (1,015 psia), pressure ratios of 0.4, 0.5, and 0.6, rotor speeds up to 20,200 rpm, a radial clearance of 0.1 mm (4 mils), and three preswirl ratios. For comparison, TOS labyrinth tests were run at identical conditions as the TOR tests but for only one positive preswirl ratio.

TOR labyrinth measurements show a pronounced dependence of rotordynamic coefficients on rotor speed, especially when compared to prior documented TOS labyrinth seal tests run at a radial clearance of 0.2 mm (8mils). The TOR labyrinth cross-coupled stiffness is higher in magnitude and increases at a higher rate than that of the TOS labyrinth across all test speeds. However, the TOR labyrinth effective damping was determined to be greater due to higher measurements of direct damping. Measured leakage rates for the TOR labyrinth were approximately 5–10% less than the TOS labyrinth. XLLaby underpredicted the rotordynamic coefficients for both seals. However, as with measurements, it predicted the TOR labyrinth to have higher effective damping than the TOS labyrinth.

Topics: Rotors , Stators , Leakage
Commentary by Dr. Valentin Fuster
2015;():V07AT31A017. doi:10.1115/GT2015-43332.

Floating ring annular seals represent one of the solutions for controlling leakage in high speed rotating machinery. They are generally made of a carbon ring mounted in a steel ring for preserving their integrity. Low leakage is ensured by the small clearance of the annular space between the carbon ring and the rotor. Under normal operating conditions, the ring must be able to “float” on the rotor in order to accommodate its vibration. Impacts between the carbon ring and the rotor may occur when the annular seal is locked up against the stator and the amplitude of rotor vibrations are larger than the radial clearance. This situation is prohibited because it rapidly leads to the destruction of the carbon ring.

The present work presents experimental results obtained for floating ring annular seals of 38 mm, tandem mounted in a buffer seal arrangement. The rotation speed was comprised between 50 Hz and 350 Hz and maximum pressure drop was 7 bar. For these operating conditions the floating ring follows the rotor vibrations without any impacts.

Comparisons were made with a theoretical model based on the equations of motion of the floating ring driven by mass inertia forces, hydrostatic forces in the (main) annular seal and by friction forces on its radial face (also named the “nose” of the seal). The friction coefficient on the nose of the floating ring was estimated from Greenwood and Williamson’s model for mixed lubrication. The present analysis validates the theoretical model used for predicting the dynamic response of the floating ring for a given rotor motion.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A018. doi:10.1115/GT2015-43386.

A fast and efficient method for evaluating bearing coefficients of the fixed geometry bearings is presented. In a typical industrial application, where the accuracy of the solution is desired, this paper presents a method whose accuracy is verified to be good by the benchmark study. Reynolds equation is solved to obtain non-dimensionalised static and rotor-dynamic characteristics for a pre-defined bearing pad geometry. The solution in the form of non-dimensional functions is obtained for a 2 dimensional space representing all possible journal loci for any load vector orientation. Laminar flow is considered in the analysis, although the method of analysis can be extended to Turbulent flow regime. The analysis method is most efficient for isoviscous boundary condition. A pad assembly method for the fixed pad journal bearings is presented. Any fixed pad bearing geometry including multi-pad bearings, preload with any load vector orientation can be evaluated using this method. In this paper, demonstrating cases for a four-pad bearing are presented.

Topics: Bearings , Geometry
Commentary by Dr. Valentin Fuster
2015;():V07AT31A019. doi:10.1115/GT2015-43460.

A theoretical algorithm for the analysis of bidirectional interaction of combined journal and thrust bearings is presented. While many theoretical and experimental investigations on the operating behavior of single journal and thrust bearings can be found only few results for combined bearings are available. However, combined bearings interact by exchanging lubricant and heat which can affect significant changes of boundary conditions compared to a single bearing application. Therefore, a novel procedure is developed to combine two separate codes for journal and thrust bearings in order to iteratively determine the coupling boundary conditions due to the special design of the entire bearing unit. The degree of interaction strongly depends on the type of lubrication. In a first step predictions are verified by measurement data for a combined bearing with a fixed-pad offset-halves journal bearing and a directed lubricated tilting-pad thrust bearing. Experiments were conducted on a high speed test rig up to sliding speeds of 107 m/s at the mean radius of the thrust bearing. As expected the interaction of the two oil films is comparably low in the investigated speed and load range for this bearing design because of the active lubrication of both bearings and the low hydraulic resistance of the thrust bearing. In order to theoretically investigate interaction of thrust and journal bearings in more details a combined bearing with fixed-pad thrust parts lubricated exclusively by the side flow of the journal bearing is studied. A variation of modeling level, pocket design of the journal part, thrust load and rotating frequency provides the following results: (i) hydraulic and energetic interaction have to be modelled in details, (ii) the axial flow resistance of the pockets strongly influences flow rates and the pressure level at the interfaces (iii) the level of interface pressure rises with increasing thrust loads and decreasing rotor speed, (iv) the axial bearing clearance is rather of minor importance for the investigated bearing. Finally, improvements in order to predict operating conditions more precisely are comprehensively discussed.

Topics: Thrust bearings
Commentary by Dr. Valentin Fuster
2015;():V07AT31A020. doi:10.1115/GT2015-43487.

High performance turbomachinery demands high shaft speeds, increased rotor flexibility, tighter clearances in the flow passages, advanced materials, and increased tolerance to imbalances. Operation at high speeds induces severe dynamic loading with large amplitude shaft motions at the bearing supports. Typical rotordynamic models rely on linearized force coefficients (stiffness K, damping C, and inertia M) to model the reaction forces from fluid film bearings and seal elements. These true linear force coefficients are derived from infinitesimally small amplitude motions about an equilibrium position. Often, however, a rotor-bearing system does not reach an equilibrium position and displaces with motions amounting to a sizable portion of the film clearance; the most notable example being a squeeze film damper (SFD). Clearly, linearized force coefficients cannot be used in situations exceeding its basic formulation. Conversely, the current speed of computing permits to evaluate fluid film element reaction forces in real time for ready numerical integration of the transient response of complex rotor-bearing systems. This approach albeit fast does not help to gauge the importance of individual effects on system response. Presently, an orbit analysis method estimates force coefficients from numerical simulations of specified journal motions and predicted fluid film reaction forces. For identical operating conditions in static eccentricity and whirl amplitude and frequency as those in measurements, the computational physics model calculates instantaneous damper reaction forces during one full period of motion and performs a Fourier analysis to characterize the fundamental components of the dynamic forces. The procedure is repeated over a range of frequencies to accumulate sets of forces and displacements building mechanical transfer functions from which force coefficients are identified. These coefficients thus represent best fits to the motion over a frequency range and dissipate the same mechanical energy as the nonlinear mechanical element. More accurate than the true linearized coefficients, force coefficients from the orbit analysis correlate best with SFD test data, in particular for large amplitude motions, statically off-centered. The comparisons also reveal the fallacy in representing nonlinear systems as simple K-C-M models impervious to the kinematics of motion.

Topics: Bearings , Fluid films
Commentary by Dr. Valentin Fuster
2015;():V07AT31A021. doi:10.1115/GT2015-43519.

Recent developments in the aeronautic domain focus on the optimization of the lubrication oil system for civil aircraft gas turbine engines, in order to reduce air and oil consumptions. Specifically, over the last few decades, as brush seals have shown tremendous leakage performance in sealing secondary flows compared to classic labyrinth seal, an increasingly popular idea is to extend their utilization to bearing chambers applications.

In the frame of the European FP7 E-Break project, the Aero-Thermo-Mechanics department of ULB collaborates with French aircraft engine manufacturer SNECMA in order to investigate experimentally the brush seal behaviour in an environment simulating the bearing chamber working conditions. The aim is first to deepen the brush seal behaviour knowledge by identifying the most influential geometric parameters acting on the leakage performance on both sides of the seal (oil and air), and on its wear, and by evaluating the friction torque and the dissipated heat.

The paper will first highlight the effect of the brush seals geometric parameters on the air consumption and the torque friction. Results highlight a trade-off to be made between these two performance levels. Also, relations have been developed to predict the performance of a carbon brush seal with non-canted bristles. The bristle free length and the axial density must carefully be chosen first to dictate the brush seal porosity. The distance between the backing plate and the front plate acts as a secondary parameter to adjust the bristle pack stiffness, and it is proposed to mount such a carbon brush seal with a reduced interference to limit the effect of the brush seal wear on the air consumption. Finally, the carbon brush seals performance was compared with the latest ones, with promising results being shown to expect carbon brush seals to be employed at a higher scale in bearing chambers in the future.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A022. doi:10.1115/GT2015-43650.

Compliant gas foil bearings are composed of two layers of thin metallic foil and a thin film of gas to support the journal. The bottom foil creates an elastic structure which supports the top foil. This support structure can take a variety of shapes that range from a series of bumps around the circumference to a series of overlapping leaves. The top foil and the rotation of the rotor create a wedge of air that supports the rotor. The complaint foil structure deforms in response to the pressure developed within the gas film. These bearings have several advantages over conventional fluid film bearings. These advantages include reduced weight due to the elimination of the oil system, stable operation at higher speeds and temperatures, low power loss at high speeds and long life with little maintenance. Some disadvantages of gas foil bearings are low load capacities at low speed and modest stiffness and damping values. Due to these properties, compliant gas foil bearings are commonly used in specialized applications such as compressors for aircraft pressurization, engines for turboshaft propulsion, air cycle machines, turboexpanders, and small micro turbines. The ability to predict the behavior of these bearings and design them to meet the needs of the application is invaluable to the design process. This behavior can include things such as bearing stiffness, damping, and load capacity. Currently most foil bearing analysis tools involve some sort of coupling between hydrodynamics and structural analyses. These analysis tools can often have convergence issues and can require the use of empirically derived characteristics. This paper reviews the current status of the compliant gas foil bearings research, focusing mainly on the journal bump-type gas foil bearings and the development of the analysis tools for these bearings. This paper contributes to the field by making recommendations of the future developments of the analytical tools of journal bump-type gas foil bearings.

Topics: Foil bearings
Commentary by Dr. Valentin Fuster
2015;():V07AT31A023. doi:10.1115/GT2015-43734.

High speed rotors supported on bump-type foil bearings (BFBs) often suffer from large sub synchronous whirl motions. Mechanically preloading BFBs through shimming is a common, low cost practice that shows improvements in rotordynamic stability. However, there is absence of empirical information related to the force coefficients (structural and rotordynamic) of shimmed BFBs. This paper details a concerted study towards assessing the effect of shimming on a first generation BFB (L=38.1 mm, D =36.5 mm). Three metal shims, 120° apart, are glued to the inner surface of the bearing cartridge and facing the underside of the bump foil strip. The shim sets are of identical thickness, either 30 μm or 50 μm. Static load tests show that shimming produces nonlinear static load vs. deflection curves leading to a larger structural stiffness than for the bearing without shims. Torque measurements during shaft acceleration also demonstrate a shimmed BFB has a larger friction coefficient. For a static load of 14.3 kPa, dynamic loads with a frequency sweep from 250 Hz to 450 Hz are exerted on the BFB, without and with shims, to estimate its rotordynamic force coefficients while operating at ∼50 krpm (833 Hz). Similar measurements are conducted without shaft rotation. Results are presented for the original BFB (without shims) and the two shimmed BFB configurations. The direct stiffnesses of the BFB, shimmed or not, increase with excitation frequency thus evidencing a mild hardening effect. The BFB stiffness and damping coefficients decrease slightly for operation with rotor speed as opposed to the coefficients when the shaft is stationary. For frequencies above 300 Hz, the direct damping coefficients of the BFB with 50 μm thick shims are ∼ 30% larger than the coefficients of the original bearing. The bearing structural loss factor, a measure of its ability to dissipate mechanical energy, is derived from the direct stiffness and damping coefficients. The BFB with 50 μm thick shims has a 25% larger loss factor — average from test data collected at 300 Hz to 400 Hz — than the original BFB. Further measurements of rotor motions while the shaft accelerates to ∼50 krpm demonstrate the shimmed BFB (thickest shim set) effectively removes sub synchronous whirl motions amplitudes that were conspicuous when operating with the original bearing.

Topics: Foil bearings
Commentary by Dr. Valentin Fuster
2015;():V07AT31A024. doi:10.1115/GT2015-43815.

Numerous papers have investigated the behavior of dry-friction whip and whirl; most of them consider contact between a rotor and stator at a single location. For rotors running on multiple magnetic bearings, air bearings, or bushings, equipment failure may result in rub at more than one location. For these cases, it is important to have an analytical model that characterizes possible regions of two-point contact dry-friction whip and whirl.

The current work presents a general model to predict possible whirl regions for multi-contact dry-friction whip and whirl, allowing for an arbitrary phase between contact locations. In theory this method can be applied to more than two contact locations; however, a two-point contact example case is developed and compared to results from an experimental test rig developed to demonstrate multi-contact dry-friction whip and whirl in the current work.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A025. doi:10.1115/GT2015-43826.

Water lubricated bearings used in nuclear coolant pumps and sub-sea applications exhibit large lubricant inertia forces in the magnitude order of viscous forces. To model these bearings the traditional Reynolds equation is not adequate. An extended Reynolds equation is developed in this study which takes into account the turbulence and inertia effects: both convective and temporal. The most complete form of temporal inertia which applies to the turbulent regime as well, is developed that consists of primary and secondary temporal inertia terms. The convective inertia model follows Constantinescu’s approach [1,2]. The turbulence model is also Constantinescu’s which is tuned with a CFD work. The dynamic coefficients including the lubricant added mass coefficients of a full cylindrical fixed geometry water bearing are obtained. It is observed that the convective inertia increases the load capacity and stability of the bearing. Significant lubricant added mass coefficients comparable to the shaft mass are calculated, which exhibit destabilizing effects in general.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A026. doi:10.1115/GT2015-43959.

The aim of this study was to evaluate the structural performance of multilayer gas foil bearings (GFBs). Three test GFBs and a dummy rotor with a diameter of 28 mm were manufactured for application to a series of structural tests. The test bearings were designed with an identical axial length of 22.5 mm, a radial clearance of 0.170 mm, and a mechanical preload of 0.050 mm at a 50% offset within the (machined) lobe bearing housings. All the bearings had multilayer (doublelayer) thin foils supported by an elastic bump strip layer, i.e., a configuration with a top foil, shim foil, and bump strip layer. Note that the shim foil is installed between the top foil and bump strip layer to enhance the dry-friction damping effect. The first bearing (Type A) had a single top foil, single shim foil, and single bump strip layer. All had an arc length of 360°. The second version (Type B) had a single top foil with an arc length of 360°, three shim foils with an arc length of 120°, and three bump strip layers with an arc length of 120°. The third version (Type C) had three top foils, three shim foils, and three bump strip layers. All had an arc length of 120°. The top foil is spot-welded to a key that is inserted into an axial key slot in the bearing housing. The shim foil and bump strip layer are inserted into an axial foil slot in the bearing housing. A series of static load-deflection tests were conducted on the test GFBs floating on the fixed, non-rotating test rotor. The measured results for the bearing deflection and structural stiffness were found to be in very good agreement with the model predictions for a GFB with a mechanical preload. In general, the test results were found to exhibit a similar radial sway space (or assembly radial clearance) and structural stiffness for all three test GFBs. Small local hysteresis loops appeared as the magnitudes of the load increased, thus determining the local structural stiffness and structural loss factor versus the displacement. The local stiffness was found to increase while the loss factor decreased as the magnitude of the displacement increased. The estimated structural loss factor can be as large as 0.9 within the radial sway space under low-load conditions.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A027. doi:10.1115/GT2015-43982.

The closed loop Brayton cycle with super critical CO2 (S-CO2) as an operating fluid is an attractive alternative to conventional power cycles due to very high power density. Foil gas bearings using CO2 is the most promising for small S-CO2 turbomachinery but there are many problems to address; large power loss due to high flow turbulence, lack of design/analysis tool due to non-ideal gas behavior, and lack of load capacity when they are used for large systems.

This paper presents high level design/analysis tool involving three-dimensional thermo-hydrodynamic analyses of radial foil bearings considering real gas effect and flow turbulence inside the film. Simulations are performed for radial foil bearing with 34.9mm in diameter lubricated with CO2 and N2 under various ambient conditions up to above 40 bar gauge pressure. The simulation results using the turbulence model still under-predict the measured data in open literature. However, the error between the prediction and measurements decreases as either speed or ambient pressure increases. In addition, general behavior of substantial increase in power loss with ambient pressure agrees with the measured data. The simulation results indicate the importance of detailed THD analysis of the foil bearings for prediction of power loss under severe turbulent condition.

A conceptual layout of rotor system for 10MWe S-CO2 loop is also presented along with realistic rotor weight and bearing load. A hybrid foil bearings with diameter of 102mm is suggested for gas generator rotor, and its power losses and minimum film thicknesses at various operating conditions are presented.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A028. doi:10.1115/GT2015-43999.

Improvement of the load capacity of gas foil thrust bearings (GFTBs) is important to broadening their application in oil-free microturbomachinery (<250 kW) with high power density. Although GFTBs have the significant advantage of low friction without the use of lubrication systems compared to oil film thrust bearings, their inherently low load capacity has limited their application. The aim of the present study was to develop a design guideline for increasing the load capacity of GFTBs. The Reynolds equation for an isothermal isoviscous ideal gas was used to calculate the gas film pressure. To predict the ultimate load capacity of the GFTB, the pressure was averaged in the radial direction of the gas flow field used to deflect the foil structure. The load capacity, film pressure profile, and film thickness profile were predicted for a GFTB with an outer radius of 55 mm, inner radius of 30 mm, and eight foils each of arc length 45°. The predictions showed that the load capacity of the GFTB increased with increasing rotor speed and decreasing minimum film thickness, and was always lower than the analytically determined limit value for infinite rotor speed (obtained by simple algebraic equations). A parametric study in which the ramp extent (or inclined angle) was increased from 5° to 40°, and the ramp height from 0 to 0.320 mm, revealed that the GFTB had an optimal ramp extent of ∼22.5° and ramp height of ∼0.030 mm for maximum load capacity. Interestingly, the optimal values were also valid for a rigid-surface bearing. The predicted load capacities for a ramp extent of ∼22.5° and increasing ramp height from 0.030 to 0.320 mm were compared with experimental data obtained from a previous work. The predictions for a ramp height of 0.155 mm were in good agreement with the experimental data for all three test GFTBs with outer radii of 45, 50, and 55 mm, respectively. In addition, this paper shows that the predicted drag torque increases linearly with increasing rotor speed and decreasing minimum film thickness, and nonlinearly with decreasing ramp height. The drag torque significantly increased only for ramp heights below the optimal value. The predictions imply that the optimal ramp height improves the load capacity of the GFTB with little change in the drag torque.

Commentary by Dr. Valentin Fuster
2015;():V07AT31A029. doi:10.1115/GT2015-44059.

Oil whip is a self-excited vibration in a hydrodynamic bearing which occurs when the rotation speed is above approximately twice the first natural frequency. Because of this, the oil whip phenomenon limits the operational speed of a rotor system on hydrodynamic bearings. Below the oil whip threshold, the related phenomenon of oil whirl can cause large vibrations at frequencies below half the rotation speed.

A method is presented for stabilizing oil whip and oil whirl in a hydrodynamic bearing with an active magnetic bearing (AMB). The AMB controller is designed with μ-synthesis model-based robust control utilizing the Bently-Muszynska fluid film bearing model, which predicts the unstable phenomena. Therefore, the resulting AMB controller stabilizes the natural instability in the hydrodynamic bearing. Rotor speed is taken into account by use of a parametric uncertainty such that the method is robust to changes in running speed.

The proposed method is demonstrated on an experimental hydrodynamic bearing test rig. Details of the test rig and implementation of the AMB controller design are presented. Waterfall plots for the controlled and uncontrolled system are presented which demonstrate the improved stability limit.

Commentary by Dr. Valentin Fuster

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In