A Procedure to Predict Influence of Acceleration and Damping of Blades Passing Through Critical Speeds on Fatigue Life

J.S. Rao Chief Science Officer, Altair Engineering Mercury 2B Block, 5th Floor, Prestige Tech Park, Sarjapur Marathahalli Outer Ring Road, Bangalore, Karnataka, 560103, India js.rao@altair.com

S.Suresh Altair Engineering Mercury 2B Block, 5th Floor, Prestige Tech Park, Sarjapur Marathahalli Outer Ring Road, Bangalore, Karnataka, 560103, India suresh. rao@altaircom

Rejin Ratnakar Altair Engineering Mercury 2B Block, 5th Floor, Prestige Tech Park, Sarjapur Marathahalli Outer Ring Road, Bangalore, Karnataka, 560103, India rejjnratnakar@altair.com R. Narayan Altair Engineering Mercury 2B Block, 5th Floor, Prestige Tech Park, Sarjapur Marathahalli Outer Ring Road, Bangalore, Karnataka, 560103, India narayanrangarajan@altair.com

www.altairproductdesign.com copyright Altair Engineering, Inc. 2011

www.altairproductdesign.com

Abstract
Turbine Blades suffer fatigue damage when they cross over a critical speed during start up and shut down conditions. The stress response is usually determined from quasi-steady analysis through resonance with an assumed damping. This response above fatigue limit can be divided into several steps to reach the peak value at critical speed and then fall after passing the critical. For a given acceleration of the rotor, one can then determine the number of cycles at each of these stress levels and assess cumulative damage for one crossing. In this paper, the effect of acceleration and damping on the magnitude of peak stress and where it occurs in the vicinity of critical speed is included in determining the damage suffered by a blade while passing through the critical speed.

1.0 Introduction
Fatigue is essentially due to alternating stresses; Wohler (I) conducted the earliest tests to define S-N diagrams. The influence of the mean stress component is presented in Goodman diagrams (2). Bagci (3) proposed a fourth order relation which is shown to give good results for ductile materials. S-N diagrams and mean stress diagrams are combined to define the fatigue failure surface under elastic conditions, see (4). The fatigue failure surface defined thus is applicable when the stress field is entirely in the elastic region. Ludwik (5) and Hollomon (6) defined the stress-strain relation in the plastic region, = Kn where K is the strength coefficient and n is strength exponent For globally elastic and locally plastic structures, Neuber (7) hypothesis is used for relating the nominal and local cyclic stresses and strains, e.g., (8, 9). For fatigue loading, using fatigue stress concentration factor Kf local stress and strain ranges are related to nominal stress range

Socie et al (10) suggest that Kf can be taken to be the theoretical stress concentration factor. Based on Basquin (11) for endurance and Manson (12) and Coffin (13) works, we have a life relation for crack initiation given by

Palmgren (14) and Miner (15) defined a linear cumulative damage law. The nonlinear theories proposed include Marco and Starkey (16), Corten and Dolan (17), Haibach (18), Marin (19), Henry (19), Gatts (20) and Manson (21). Rao et al (22) made a comparison of lives from different nonlinear approaches and pointed out the relative merits of the different approaches. Turbine blades suffer fatigue damage when they cross over a critical speed during start up and shut down conditions. The stress response is usually determined from quasi-steady analysis, i.e. determining response at different operating speeds through resonance with an

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

assumed damping. This response above fatigue limit can be divided into several steps to reach the peak value at critical speed and then fall after passing the critical. For a given acceleration of the rotor, one can then determine the number of cycles at each of these stress levels and assess cumulative damage for one crossing. This process is adopted by Rao, Narayan and Ranjith (23) to determine the damage suffered by a blade during a cross-over of the critical speed. Damping has been identified long ago as a key parameter in blade design. Rowett (24) conducted tests on elastic hysteresis in steel in 1914. Effects of friction and loose mounting were studied by Hansen et al. (25). Sinha and Griffin (26) studied analytically the effects of static friction on the forced response of frictionally damped turbine blades. Rao et al (27) developed a test rig to determine a nonlinear damping model that depends on strain amplitude at a reference point in a given mode at a given speed. An analytical process to obtain such a damping model was developed by Rao et al (28). Using this damping model, damage suffered by a blade during the crossover of a critical speed is determined This was a breakthrough in a true simulation of Iifing without depending on assumed stimulus values and damping values. In the works above, the effect of acceleration and damping on the peak stress and speed where it occurs in the vicinity of critical speed are ignored. The influence of nonlinear damping on the life with angular acceleration is first demonstrated by Rao and vyas (29) using energy methods for defined blade geometries; the influence of stress raisers in the blade root area was not included. This necessitates adoption of finite element methods from suitable commercial codes to achieve the required results. 1.1 Overview of the Process A simplified outline of what is involved in the process is first given. The blade or bladed-disk is first prepared as a CAD model and a FE model is prepared for analysis. The steady loads usually can be centrifugal loads and mean gas loads. The steady gas loads can be negligible, particularly in low pressure stages and may be neglected. The steady state stress field is then obtained using where possible cyclic symmetry conditions and centrifugal loading. In gas turbines, thermo mechanical analysis may be conducted to determine any thermal stresses and include them in the steady stress field where required. We then need the alternating stress field to be able to determine fatigue life. Usually there will be always some unsteady force on the blades due to load changes, speed changes etc at speeds not necessarily in the vicinity of resonance. This alternating stress field is very small and has very little influence on fatigue. The significant effect on life arises from alternating stress field when the blade crosses a resonance. Therefore, a modal analysis is made to determine the natural frequencies and plot a Campbell diagram to identify the critical speeds where resonance occurs. At these speeds, the alternating stress field is determined and damage is calculated.
Copyright Altair Engineering, Inc., 2011 3

www.altairproductdesign.com

The first step is to find the resonant stress at each of the critical speeds. We need two additional inputs for this, the magnitude of alternating force field and damping to perform a linear analysis. The alternating force field is determined from a CFD analysis. Damping poses considerable problems in this analysis. Usually it is simplified as equivalent viscous damping without recognizing the influence of strain in the blade (which makes it nonlinear and this is one of the aspects of this paper). The usual process is then to make a forced vibration analysis using the force field with excitation at the critical speed under question and an assumed damping value. This analysis also poses some difficulty since the exact natural frequency to a high accuracy of decimal places may be required at low values of damping and one may not pick up the correct value at resonance. An alternate route to deteInline the resonant stress is to treat the alternating load as steady load and obtain equivalent steady state stress field and then use quality factor 1/2 to get the correct peak stress. This can also pose problems since the pressure field may change phase from node to node in the vane. Therefore a forced vibration analysis may be resorted with an almost zero excitation frequency, say 0.0001 rad/s to obtain the equivalent steady stress field and then use the quality factor. The procedure adopted here to take into effect of acceleration and damping (for present only material damping is included) on the damage suffered by a blade during the crossover of a critical speed is as follows: 1. Treating the alternating loads on the blade as steady, the equivalent static response is first determined, see (30). 2. Lazan's law (31) is used to determine the material damping characteristic as a function of strain amplitude at a reference point of the blade at the given operating speed and mode of vibration 3. A nonlinear approach, see (29) is adopted to determine the correct peak stress at resonance using an iterative approach to match the damping value and strain at the reference point. 4. While the blade accelerates through the critical speed, the peak occurs slightly to the right of classical resonance and also the peak value of response may be lower than the classical resonant stress. While performing transient analysis with the influence of acceleration included, it is initially assumed that the damping obtained in step 3 above without including acceleration effect is valid 5. Then a transient analysis of the bladed-disk with the damping thus determined in step 3 is obtained for a given acceleration through the critical speed.

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

6. This response is filtered to capture the model component of interest and determine the dynamic stress as a function of instantaneous speed. 7. In step 6 with peak response having decreased with the given acceleration, the damping is also changed An iterative process is adopted until the damping value and the response have converged. 8. The resulting frequency response is then used to assess damage suffered by the blade while passing through the critical speed. The Blade (now named Altair TurboManager) code reported in (23 and 28) is adopted and generalized to develop a simulation tool to include the acceleration effects as described above and determine life at design stage. Nomenclature

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

2.0 Lifing of Turbomachine Components Turbomachinery blades pass through several resonances during start up and shut down operations that cause cumulative damage leading to failures. In some cases, the blades operate very close to resonance at operating speed. It is important to estimate life of these blades at design stage so as to reduce number of prototype tests. The lifing calculations involve the following: 1. 2. 3. 4. 5. Mean stress field determination Natural frequencies, Mode shapes Campbell diagram Nonlinear modal damping Alternating pressure field from stage flow interference Peak stress and speed at which it occurs for a given acceleration of the rotor through critical speed 6. Damage fraction through critical speed crossing and 7. Cumulative damage during start-up and shut-down. The Blade code is extended to consider the acceleration and damping effects as discussed below. 2.1 Blade Model For simplicity and saving computation time, only the blade of the bladed-disk model (32) is adopted for the analysis as shown in Fig. 1. There are 60 pre-twisted 290 mm long blades on the disk. The blade root bottom radius is 248 mm. The operating speed is 8500 rpm. There are 70 nozzles. Using mapped meshing options a solid element mesh with 8 nodes is generated, by capturing all the critical regions with finer mesh. Mesh around the singularities at root fillet regions where peak stresses occur are captured with 2 to 3 layers of elements with element size as low as 0.235 rom. Total number of elements in the model are 31090 with 24649 nodes.

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

Figure 1: Blade Model Showing the FE Mesh 2.2 Mean Stress Field Blades are subjected to steady stress fields due to gas loads, thermal loads, and centrifugal loads under normal conditions of operation. The gas loads are determined from CFD analysis of the gas path, which have a steady part and an unsteady part at nozzle passing frequencies. Because of the airflow in the compressor flow path or the hot gas path in turbine blades, they are subjected to thermal loading during the transient period of start-up and shut-down, so they will form a mean load at a given steady operation or an overall cyclic load for each start-up and shut-down operation. The blades are also subjected to mean loads due to centrifugal loading which could be substantial in low pressure compressor and turbine blades that will push the structure into globally elastic and locally plastic conditions. Cyclic symmetry can be utilized in assessing these stress fields. The steady stress field determination is well established and can be directly imported by the user at the start of life estimation in to Blade (23, 28). Of particular interest is the result at the stress raiser location. Usually, the centrifugal loads lead to local plastic conditions in the root; an elasto-plastic analysis can give a near true picture in the stress raiser location. However, this is not necessary since true stress and strain conditions are assessed by Neuber hypothesis as given in equations (1) and (2) based on the surrounding elastic field. Alternatively, a user can perform the analysis directly using codes of his choice.

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

2.3 Natural Frequencies and Mode Shapes The natural frequencies and mode shapes are determined by using any well established code. Fig. 2 shows the Campbell diagram upto 1000 RPM speed. The intersection of II bending mode (261.5 Hz) with I NPF excitation gives one critical speed 291 RPM. The transient effect study on life is illustrated for crossing this critical speed.

Figure 2: Campbell Diagram 2.4 Nonlinear Material Damping Damping is the key element in assessing life of the blade. Usually a simple viscous damping model is used with equivalent damping determined from a test. Such a model is inadequate since material damping is highly dependent of the state of stress in the blade. Nonlinear damping model was quantified through experiments by Rao et al (27); the equivalent viscous damping is expressed as a function of strain amplitude at a reference point in a given mode of vibration at a given speed of rotation. An analytical procedure using Lazan's hysteresis law (31) to determine such a nonlinear model is briefly explained below. The total damping energy D0 (Nm) is given by

where v the volume. The Loss factor n is

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

where Wo is the total strain energy (Nm). Equivalent Viscous Damping is then given by

where C is the equivalent viscous damping (N-s/m), n is the natural frequency (rad/s) and K is the modal stiffness (N/m). For increased strain amplitudes, the orthonormal reference strain amplitudes, stress and strain energy are multiplied by a factor F to obtain the equivalent viscous damping C e at various strain amplitudes as given in (6).

For the critical speed under consideration, 291 rpm, the material damping obtained is shown in Fig. 3.

Copyright Altair Engineering, Inc., 2011

www.altairproductdesign.com

Figure 3: 3Material Damping in Mode II at 291 RPM It can be seen that even under low speed of operation at 291 rpm, the damping value more than doubled from 0.00027 to 0.00057 with strain increased from 1 x 10-5 to 6 X 10-5 in the same mode. 2.5 Excitation Force Atransient CFD analysis for the flow path interference in the stage can be carried out in a suitable code and the dynamic loads (or nodal pressures) can be calculated. The pressure field used in this analysis is expressed by

The pressures are applied on both pressure and suction surfaces in four areas along the span as shown in Fig. 4 and Table 1as given in equation (7).

Copyright Altair Engineering, Inc., 2011

10

www.altairproductdesign.com

Table 1: Alternating Pressure Magnitudes Po 2.6 Resonant Stress under Quasi-Steady Analysis Treating the alternating load as static, the equivalent steady stress distribution is first obtained. Since the centrifugal loads should be included while determining this stress distribution, it is convenient to perform a forced vibration analysis without damping at an excitation frequency close to zero value, say 0.01 rad/s. The resonant stress is then obtained by multiplying the steady stress field produced from treating the alternating loads as steady field with quality factor = 1/2. An iterative approach is used to determine the correct value of; and the resonant stress. 2.7 Effect of Acceleration

Copyright Altair Engineering, Inc., 2011

11

www.altairproductdesign.com

Five different acceleration values are chosen, as given in Table 2. Peak stress values and the frequency where it occurs are also given in this Table.

Figure 5: Quasi-Steady Response at Critical Speed

Table 2: Effect of Acceleration on Peak Stress and Critical Speed It can be seen that the maximum stress decreases with high accelerations taking the blade through critical speed quickly without allowing it to build to conventional resonant conditions and also that the peak stress occurs at a speed slightly beyond the conventional critical speed on the Campbell diagram.

Copyright Altair Engineering, Inc., 2011

12

www.altairproductdesign.com

2.8 Effect of Acceleration and Damping Next the influence of damping for four values of = 0.0001, 0.005, 0.01 and 0.04 is considered and the transient response is obtained. The peak stress value obtained as a function of angular acceleration for no damping and four other dampings is given in Fig. 6. The peak stress value decreases rapidly for no damping and very low damping of 0.0001. With damping the peak stress value decreases for a given acceleration, however the acceleration has little influence when damping is high in the system. Fig. 7 gives the critical speed (the speed at which peak response occurs) as a function of acceleration for no damping and four other damping values. With acceleration the critical speed increases nearly 2\1% in the range of acceleration considered from 500 to 2000 rad/s2 for all damping values. Also for a given acceleration, the critical speed is high with no or very little damping.

Figure 6: Stress vs. Acceleration

Copyright Altair Engineering, Inc., 2011

13

www.altairproductdesign.com

Figure 7: Critical Speed vs. Acceleration

Table 3 gives the summarized results. The peak stress from no damping decreased gradually from 168.65 MPa to 166.61, 99.67,67.71 and 20.76 MPa as damping is increased to 0.0001, 0.005, 0.01 and 0.04 respectively for 500 rad/sec2 acceleration. The critical speed, speed at which peak stress occurred also shifted gradually from 26311 Hz to 268.16, 266.94, 26573 and 262.09 Hz respectively. Similar trends are observed for higher acceleration values.

Copyright Altair Engineering, Inc., 2011

14

www.altairproductdesign.com

Table 3: Influence of Damping and Acceleration We note from Table 3 that for a given acceleration, say 500 rad/s 2 the maximum stress has not fallen by half from 166.61 to 83.305 MPa, when damping is doubled from 0.0001 to 0.5. Even though damping has doubled, the transient conditions with blade accelerating at 500 rad/s2 did not allow a full reduction of stress by half as we would have obtained from quasisteady analysis. This influence is predominant at higher angular accelerations. Therefore, the stress reduction is more influenced by acceleration rather than damping. We also note that the peak response occurs at the farthest speed beyond the classical critical speed with low damping values and approaches critical speed with very high damping; also the speed at which peak response occurs increases with acceleration. 2.9 Estimation of Peak Transient Response and Damping The determination of resonant response at critical speed under quasi-steady conditions of operation is presented in section 2.6. With acceleration of the blade, the correct value of damping and peak stress are determined as follows. From quasi-steady analysis, the damping ratio at resonance is first determined; in this case it is 0.0011. With this damping, the strain amplitude under 1000 rad/s2 acceleration is obtained as 8.1 x 10-5. The damping ratio correspondingly falls to 0.000625 as given in Fig. 3. For this damping, the peak value in transient response with the same acceleration 1000 rad/s 2 is determined. Since damping has decreased, the peak value goes up and for this new peak, the damping value changes etc. Iteration is done until there is no change in the peak value and damping value for this acceleration. Table 4 gives the iterations. It can be seen that convergency is obtained quickly.

Copyright Altair Engineering, Inc., 2011

15

www.altairproductdesign.com

Table 4: Iterations for Correct Damping in Transient Conditions Therefore the correct damping for 1000 rad/s2 is 0.000634. With this damping, the peak Stress at the stress raiser point is obtained and given by 155.2 MPa. Compared with quasi-steady response in Fig. 5 the effect of acceleration 1000 rad/s 2 is 10 reduce damping from 0.0011 to 0.000634. Under quasi-steady conditions such a reduction of damping would have increased the stress; instead the 1000 rad/s 2 acceleration did not allow the stress build with this reduction in damping. Instead this acceleration has predominantly influenced the peak stress and decreased it considerably from 275 MPa to 155.2 MPa. The stress response a! stress raiser point for this condition is obtained as shown in Fig. 8.

Figure 8: Transient Response of Stress at Stress Raiser

Copyright Altair Engineering, Inc., 2011

16

www.altairproductdesign.com

3.0 Cumulative Damage Through Resonance The Blade code developed earlier is adopted here with modifications to account for the acceleration and considering the time domain response instead of frequency domain for quasi-steady response results. The mean stress and alternating stresses are determined as described in section 2. 3.1 Material Definition The fatigue material properties required for the analysis are defined as shown in Figs 9 and 10. 3.2 Damage through Resonance Since the response is in time domain as shown in Fig. 8, it is easier to count cycle by cycle rather than frequency domain approach used earlier. Each cycle is counted with its stress level and fraction of damage for this cycle is determined. Counting all the cycles above the fatigue limit, the cumulative damage is obtained by simply adding these individual damage values. The M.arco-S1arkey nonlinear damage rule (16) can be applied for decreasing stress levels 10 obtain nonlinear damage fraction.

Figure 9: Fatigue material definition

Copyright Altair Engineering, Inc., 2011

17

www.altairproductdesign.com

Figure 10: Fatigue modification definition

Fig 11 gives the damage for four different values of damping ratio, 0.0, 0.0001, 0.005 and 0.001 respectively. The damage decreases with increase in angular acceleration and damping. Both S-N and E-N approaches are applicable, here for illustration S-N method is used.

Figure 11: Damage vs. Acceleration

3.3 Damage under Transient Stress Response Once the damping under transient conditions through resonance is identified, e.g., = 0.000634, with the response given in Fig. 8, cumulative damage in crossing this critical speed can be estimated following section 3.2, which is 3.31x10-7.

Copyright Altair Engineering, Inc., 2011

18

www.altairproductdesign.com

4.0 Concluding Remarks A code is developed to determine the life of turbomachine blades using the HyperWorks platform to help a designer speed up the design process. The special features included in this version of the code are: 1. Determining transient response under accelerating conditions of the blade. 2. Determining correct damping ratio for a blade accelerating through a critical speed 3. Accurate determination of resonant stress from the nonlinear damping model for a given acceleration 4. Determination of cumulative damage during stress raising and dropping through a resonance by linear as well as nonlinear theories 5. Adopt SN or EN method as appropriate in high or low cycle fatigue calculations 6. Determine life in terms of start-up and shut-down operations or operation at a continuous resonant conditions A code like this can help the designer in complete simulation and reduction of number of prototype tests and thus considerably decrease the design cycle time. In particular, it will help the designer in choosing possible acceleration of the rotor to achieve the required life targets.

Acknowledgements The authors acknowledge encouraging support from Altair India in this effort.

References 1. Wohler, A., Versuche tiber die Festigkeit der Eisenbahnwagen. Achsen, Zeitschrift jiir Bauwesen, 1860 2. Heywood, R. B., Designing Against Fatigue, Chapman and Hall, 1962 3. Bagci, C, Fatigue Design of Machine Elements Using Bagci Line Defining the Fatigue Failure Surface Line (Mean Stress Diagram), Mechanism andMachine Theory, v. 16, 1981, p. 339 4. Rao, 1. S., Turbomachine Blade Vibration, John Wiley, 1991 and New Age International., 1997 5. Ludwik, P, Elemenm der Technologischen Mechanik, Springer Publication, 1909 6. Hollomon, J. H., Tensile Deformation, Trans. American Inst. Mining andMetallurgicaIEngrs., vol. 162, 1945, p. 268 7. Neuber, H., Theory of Stress Concentration for Shear Strained Prismatical Bodies with Arbitrary Nonlinear Stress Strain Law, Trans. ASME, J Applied Mechanics, vol. 8, 1961, p.544

Copyright Altair Engineering, Inc., 2011

19

www.altairproductdesign.com

8. Topper, T. H., Wetzel, R. M and Morrow, Jo Dean, Neuber's Rule Applied to Fatigue of Notched Specimens, J of Materials, vol. 4, No.1, March 1969, p. 200 9. Impellizzeri 1. F., Cumulative Damage Analysis in Structural Fatigue, Effects of Environment and Complex Load History on Fatigue Life, STP-462, ASTM, 1970, p. 40 10. Socie, D. F., Dowling, N. E. and Kurath, P, Fatigue Life Estimation of Notched Members, Proc. 15th Symp. On Fracture Mechanics, STP 833, 1984, p. 284 11. Basquin, O. H., The Experimental Law of Endurance Tests, Proc. ASTM, vol. 10,1910, p. 625 12. Manson, S. S., Behavior of Materials under Constant Thermal Stress, Heat Transfer Symposium, University of Michigan, Engng. Research Institute, 1953, p. 9 13. Coffin, 1. F. Jr., A Study of the effects of Cyclic Thermal Stresses on A Ductile Material, Trans ASME, vol. 76, 1954,p.931 14. Palmgren, A., Die Lebensdauer von Kugellagem, ZVDI, v. 68, 1924, p. 339 15. Miner, M. A., Cumulative Damage in Fatigue, Trans ASME, J. AppliedMechanics, vol. 67, 1945, P A159 16. Marco, S. M. and Starkey, W 1., A Concept of Fatigue Damage, Trans ASME, v. 76, 1954, p. 627 17. Corten, H. T. and Dolan, T. J, Cumulative Fatigue Damage, Proc. Inti. Con! on Fatigue of Metals, ASME and IME, 1956, p. 235 18. Haibach, E., Betriebsfestigkeit Verfahren und Daten zur Bauteilberechnung, VDI.Verlag GmbH, Dusseldorf, 1989 19. Henry, D. 1., Theory of Fatigue Damage Accumulation in Steel, Trans. ASME, v. 77, 1955, p. 913 20.Gatts, R R., Application of a Cumulative Damage Concept to Fatigue, Trans. ASME, v. 83, Series D, No.4, 1961, p. 529 21.Manson, S. S., Frecke, J. C. and Ensign, C. R., Application of a Double Linear Damage Rule to Cumulative Fatigue, Fatigue Crack Propagation, STP.415, American Society for Testing and Materials, Philadelphia, 1967, p 384 22 Rao, J. S, Pathak, A. and Chawla, A., Blade Life A Comparison by Cumulative Damage Theories, J. Of Engng for Gas Turbines and Power, vol. 123, No.4, 2001, p. 886 23 Rao, 1. S, Narayan, R. and Ranjith M. C, Lifing of Turbomachinery Blades - A Process Driven Approach, GT2008.50231, ASME Turbo Expo 2008, June 9.13, 2008 Berlin, Germany
Copyright Altair Engineering, Inc., 2011 20

www.altairproductdesign.com

24. Rowett, F. E., Elastic Hysteresis in Steel, Proe. Roy. Soc., vol. 89, 1914 25. Hansen M. P et al., A Method of Evaluating Loose Blade Mounting as a means of suppressing Turbine and Compressor Blade Vibration, Proc. SESA, vol. 10, 1953, p.103 26 Sinha, A. and Griffin, J., Effects of Static Friction on the Forced Response of Frictionally Damped Turbine Blades, J. Engng. Power and Gas Turbines, ASME, v. 106, 1984, p. 65 27. Rao, 1. S., 'Yas, N. S. and Gupta, K., Blade Damping Measurement in A Spin Rig with Nozzle Passing Excitation Simulated by Electromagnets, Shock & Vib. Bull, 56, Pt 2, 1986, p.109 28 Rao, 1. S, Narayan, R. Ranjith M C. and Rejin, R., Blade Lifing with Material and Friction Damping, The Future of Gas Turbine Technology, 4th International Conference, 15-16 October 2008, Brussels, Belgium 29. Rao, J S. and 'Yas, N. S., Determination of Blade Stresses under Const3nt Speed and Transient Conditions with Nonlinear Damping, J of Engng for Gas Turbines and Power, Trans ASME, vol. 118, No.2, 1996, p. 424 30. Rao, J S., Peraiah, K Ch and Uday Kumar Singh, Estimation of Dynamic Stresses in Last Stage Steam Turbine Blades under Reverse Flow Conditions, Proc. VETOMAC - IV, Hyderabad, 17-19 Dec 2007 31. Lazan, B. J, Damping of Materials and Members in Structural Mechanics, Pergammon Press, New York, 1968. 32. Rao, J S. and Suresh, S., Blade Root Shape Optimization, The Future of Gas Turbine Technology, 3rd International Conference, 11-12 October 2006, Brussels

Copyright Altair Engineering, Inc., 2011

21