18 th Blade Mechanics Seminar: Abstracts

Transcription

1 18 th Blade Mechanics Seminar: Abstracts ZHAW Zurich University of Applied Sciences IMES Institute of Mechanical Systems in Winterthur, Switzerland Wednesday, September 11 th, 2013 Eulachpassage, TN E0.58, Technikumstrasse 71, 8401 Winterthur Hans Mårtensson GKN Aerospace Sweden AB, Sweden Title key note presentation: Insight into the Design of Turbomachinery Blades for Flutter Avoidance The presentation starts from basic concepts on mechanisms that cause or control flutter phenomena. It is described how the aerodynamics of a vibrating blade can cause an unstable situation due to an interaction controlled by the structural mode shapes. A few computational examples illustrate the particular problems in computing the aerodynamic damping. The particular challenges of computing flutter limits are explained in part by the aerodynamic damping of the least damped mode being the result of a small difference between large numbers. The difficulty in assigning design margins is highlighted in systems with low structural damping by the use of the S-curve. But, the numerical flutter assessment must be related to large uncertainties of aerodynamic damping. Inherent material damping is usually negligible and much of the structural damping in blade systems therefore comes from friction on contact surfaces. Often the characterization poses a problem as many physical effects must be accounted for. A challenge is posed by the basic nonlinearity of a friction damper as a limit cycle tends to result. Current practices to formulating design criteria for flutter are discussed first from a point of view of comparing results with other machines in a best of class approach. The strength of the aeroelastic interaction is then described by way of a characteristic damping, where performance goals and structural properties can be assessed. Utilizing aeroelastic models allows for adding flutter margin by introducing an intentional mistuning. An example of this is given in the case of a transonic blade, where positive flutter margins could be established by mistuning.

2 Zdenek Kubin Doosan Skoda Power, Czech Republic Damping Identification for Various Materials & Types of Turbine Blading Regarding steam turbine blade vibrations, damping of blade as well as bladed disc mode shapes is one of the most important parameters in terms of steam turbine operation. A value of the parameter depends on properties of the material used for manufacturing of the blades, the discs and other construction elements such as blade roots, shrouds, tiebosses (snubbers) and dampers. The necessity to know the amount of total damping is discussed in terms of coupled mode frequency and unstalled flutter. This article deals with a comparison of material damping measured on special beams and mode shapes damping for particular blade couplings. The whole identification procedure of the damping together with its specifics is also presented. First, an identification technique for material damping ratio is introduced and its results are given for different materials. The material damping ratio is assessed as material strain dependent and comparison to friction damping is discussed. Subsequently, the damping ratio of mode shapes of bladed disc under rotation is identified taking into account two alternatives. The alternatives differ in such a way that blades have been free standing for the first time and then coupled with friction dampers. Outcomes presented in the article illustrate a good agreement between damping ratio of bladed disc mode shapes with free standing blades and the material used for manufacturing the disc. On the other hand, damping ratio of bladed disc mode shapes with friction dampers is significantly different and strongly dependent on blade vibration amplitudes as well as nodal diameters of bladed disc mode shapes. Achim Zanker École Polytechnique Fédérale de Lausanne, Switzerland Experimental Aeroelastic Analysis of Grouped Turbine Blades The presentation focuses on controlled vibration measurements performed to analyze the aerodynamic damping behavior of grouped blades (cluster). The measurements were carried out in the non rotating annular test facility at EPF Lausanne. In specific two clusters are selected, the 2

3 first cluster simulates two blades welded in pair with a vibration direction perpendicular to the blade chord, the second cluster simulates a four blade vane segment with a vibration direction in torsion mode. The results of the cluster measurements are compared to reference single blade measurements with similar vibration directions. The focus of the comparison is to identify possible similarities between the single blade and cluster blade aerodynamic damping results. Artem Seidenberg Alstom, Switzerland Application of an Aeroelasticity Code Reproducing Dynamic Stress Level Determined by Engine Measurements A Systematic Approach Resonant vibration of gas turbine blades can implicate critical vibrational stresses from the HCF point of view. Even for a resonance-free design for nominal engine speed, an excitation of eigenmodes may occur e.g. during transients. Assessing this state of resonance can be realized applying an aeroelasticity prediction tool. This tool has been applied to assess a compressor vane for which strain gauge measurement data are available. Main focus was on studying the effects by moving from an ideal to a real geometry including deviations due to manufacturing. Further different analysis methods like full annulus and fast frequency response have been applied. Moustapha Mbaye Turbomeca, France Non-Synchronous Vibration Phenomena on an Experimental Compressor Stage Non-synchronous vibration (NSV) is the interaction of an aerodynamic instability with turbomachinery blade vibrations. This phenomenon generally occurs away from a stalled condition, at non-integral multiples of the shaft rotational frequencies and occurs at one dominant frequency. This presentation shows a non-synchronous vibration phenomena experienced on an experimental compressor stage. The aerodynamic instability is shows and its interaction in specific conditions with the blade vibrations is also described. 3

4 Damian Vogt University of Stuttgart, Institute of Thermal Turbomachinery Sensitivity of Aerodynamic Damping Values Determined from Unsteady Blade Surface Pressures Measurements The continuous trend towards increased power densities of turbomachines while providing a wide operating range presents a demanding challenge to the aeromechanical design. Among the range of possible aeromechanical problems, flutter is considered as a very serious one as it can lead to disintegration of components in a very short period of time. Flutter denotes a self-excited and selfsustained vibration phenomenon, that occurs when the aerodynamic damping becomes negative and outbalances positive mechanical damping. The design for flutter-free turbomachines involves nowadays greatly the use of sophisticated numerical prediction tools. Such tools solve the unsteady flow field in vibrating blade rows and consequently allow to determine the aerodynamic damping from an energy consideration. In order to validate flutter prediction tools, component tests are commonly performed, in which the unsteady blade surface pressure is measured in an oscillating blade row and condensed to an aerodynamic damping value. However, the accuracy of aerodynamic damping values that are determined experimentally in such manner depend heavily on the accuracy and spatial resolution of the measured unsteady blade surface pressure. The proposed publication discusses aspects or error propagation as well as flow field effects on the sensitivity of experimentally determined aerodynamic damping values. 4

5 Bernhard Klein ABB Turbo Systems, Switzerland Tip Timing Measurements on Turbochargers - State of the Art and Systems, Future Developments ABB started Tip Timing measurements on gas and steam turbines First publications were 1980 (Roth, BBC). First tests on Turbochargers were in Tip Timing and strain gauge measurements typically are done in parallel during blade vibration development tests for ABB s turbochargers. Highlights: Resolutions down to 2 micrometer and good correlation between Tip Timing and strain gauge can be achieved. Challenges: High modes, total sensor count, multiple sensor planes and sensor position are some of the future challenges to achieve Tip Timing only. Raphael Friedli Zurich University of Applied Sciences, Switzerland Tip Timing Analysis in Consideration of Axial Rotor Movement Tip timing is a well-established noncontact measurement technique to analyze turbine blade vibrations and other properties of a running turbo machine. In order to analyze vibration induced stresses in the blades the exact position of the measurement location has to be known. This can be achieved by the usage of multiple axial measurement planes which also guarantees a proper detection of all types of vibration modes. But this implies a large number of measurement probes if conventional fitting techniques are used, since the probes to identify the order of vibration have to lie in one plane in sufficient numbers. In the present work a general method has been developed to determine blade vibrations and the underlying mode shapes with a minimum count of measurement probes. Further an analysis has been made to show the reliability for an unambiguous determination of single and superposed vibration modes. 5

6 Igor Putchkov Alstom Power, Russia Pressure pulsation technique for blade oscillation measurement and resonance diagnostics Pressure pulsation method (PPM) was developed in Central Institute of Aviation Motors (CIAM), Russia, for monitoring of compressor blades resonances, rotating stall and flutter phenomena during engine commissioning. PP method is based on gas or air pressure pulsation measurement, that are recorded and then, the after-test high-frequency spectral analysis is performed. Doppler frequency, frequency of the blade motion, first down/up frequencies, k-th down/up frequencies and corresponding amplitudes are the diagnostic parameters. The usage of PP method for turbine blades would give the advantage of a cheaper technique compared to the strain gages, capacitive (tip timing) and optical methods. Compared to the capacitive method, the PP technique is less sensitive to tip clearance and it is easier to apply than the optical experimental approach. PP method has been verified in the gas turbine engine. The measured blades dynamics based on the ordinary strain gauging data and PPM results are compared to each other in detail. The results of testing evaluated with the after-test high-frequency spectral analysis in CIAM are presented. Authors: A. Arkhipov, Alstom Power, Moscow, Russia I. Putchkov, Alstom Power, Moscow, Russia / Lead Author 6

7 Letian Wang GE Global Research, Niskayuna, USA Fast Aeromechanics Analysis of Turbine Blades Based on Shell Elements For turbine and compressor blades, the frequency margins and avoidance check is usually not performed in the early concept design stage due to the complexity of 3D modeling and time consuming cyclic symmetric modal analysis. It may bring difficulties in later preliminary and detail design if the blade frequencies are far off from the desired range. A fast finite element-based tool has been developed to estimate the natural frequency ranges and trends of cooled turbine bladed disk assemblies during conceptual design. The tool uses shell elements and models the airfoil, shank and disk to achieve more than 1000 times reduction in computation time and allow for multiple conceptual design iterations. The results obtained are shown to be within 5% of the frequencies obtained from the full-fidelity finite element model. The tool takes as input the airfoil external geometry file, certain geometrical parameters for the disk, shank and airfoil internal cavities, and parameters defining the excitations in the system. It then builds up a shell element-based model, and performs an appropriate modal analysis at the relevant harmonic indices, including frequency margin computation for critical modes. The tool also provides options for visualizing mode shapes and generating Campbell diagrams. Since many features are not finalized in the conceptual design stage, sensitivity studies can be performed to show the effect of critical features. Effects of internal and external geometrical changes on the aeromechanics performance of the system may be efficiently down selected using this tool in conceptual design stage. Natural frequency trends of specific modes as a result of small geometric changes can be studied and used to avoid any crossings at operating conditions. The effects of other parameters such as material properties, temperature, operating conditions (speed etc.) can be assessed with the tool. 7

8 Armin Schmid ABB Turbo Systems, Switzerland Reduced Order Model Using Classical Craig-Bampton Approach: Application to Mistuned Compressor Wheels of Turbochargers Due to manufacturing tolerances mistuning of turbocharger compressor wheels refers mainly to small variations of geometrical properties between each sector. The model of reduced order presented is based on classical component mode synthesis technique (Craig-Bampton) applied to each cyclic symmetric sector of the wheel. To goal of the model is to easily capture different geometric blade-mistuning patterns within numerical statistical analysis. The approach that was chosen here is advantageous, since the analyses can be continuously realized in the commercial FE-Solver ABAQUS. Starting point for the reduction is always a certain mistuned sector. The formulation used is not based on an updating technique of a reduced order tuned model, in contrast to the more common methods. The eigensolutions and the frequency response obtained with this solution in ABAQUS are promising also considering higher mode families. Letian Wang GE Global Research, Niskayuna, USA Optimal Design of Turbine Blade Dampers Using a Hybrid Computational / Experimental Approach Friction dampers (under-platform, shroud contact, etc) are regularly employed in turbomachinery to reduce turbine blade vibration levels to acceptable limits. However, optimizing the dampers for a specific application is typically hard due to lack of adequate theoretical understanding of the underlying physics and complexity involved in conducting reliable & useful experiments. Typically, a rotating rig (wheel-box test) with a few instrumented blades and assembled dampers is used to experimentally evaluate damping. These tests are expensive and cannot be employed during early design stage, making it very difficult to run damper trade-off studies during this stage. In this paper, the authors present Reduced Order Model (ROM) approaches to simulate nonlinear behavior with friction damping for turbine blades. A practical approach to optimally design underplatform and tip shroud dampers, using a novel hybrid scheme that leverages design-of- 8

9 experiments (DOE) based computational optimization combined with bench-top jugular experiments to identify, validate and down select dampers before expensive wheel-box test. The paper will demonstrate the approach using a practical industrial turbine damper optimization application and will highlight how this approach can be used to improve the first-time-yield of damper designs during early engine design tollgates. 9