Exploring How Shroud Constraint Can Affect Vibratory Response in Turbomachinery

1993 ◽  
Author(s):  
M.-T. Yang ◽  
J. H. Griffin

Unusual resonant responses were observed during spin pit tests of shrouded blades. The unusual behavior consisted of blades that exhibited strong response over a broad range of frequencies. The frequencies of peak response were considerably lower than predicted and were also lower than those observed in other tests of nominally identical wheels. In addition, the tracking plots of blade amplitudes versus frequency were truncated rather than displaying the usual sharp peaks seen in other tests. The unusual response is potentially dangerous since high vibratory response could be excited over a broad range of operating speeds. The blades were designed to be free standing with gaps of approximately 10 milli-inches between neighboring shrouds. It is hypothesized that shroud contact at the blade tips could cause the unusual vibratory response. A simple model of a blade with shroud contact is developed in order to determine if shroud contact could cause blades to vibrate in the observed manner. The model is unusual in that it seeks to explore how contact could cause the blade to respond at lower (rather than higher) frequencies since contact typically increases the constraints on a structure and raises its natural frequencies.

1995 ◽  
Vol 117 (1) ◽  
pp. 198-206
Author(s):  
M.-T. Yang ◽  
J. H. Griffin

Unusual resonant responses were observed during spin pit tests of shrouded blades. The unusual behavior consisted of blades that exhibited strong response over a broad range of frequencies. The frequencies of peak resonance were considerably lower than predicted and were also lower than those observed in other tests of nominally identical wheels. In addition, the tracking plots of blade amplitudes versus frequency were truncated rather than displaying the usual sharp peaks seen in other tests. The unusual response is potentially dangerous since high vibratory response could be excited over a broad range of operating speeds. The blades were designed to be free standing with gaps of approximately 0.25 mm between neighboring shrouds. It is hypothesized that shroud contact at the blade tips could cause the unusual vibratory response. A simple model of a blade with shroud contact is developed in order to determine if shroud contact could cause blades to vibrate in the observed manner. The model is unusual in that it seeks to explore how contact could cause the blade to respond at lower (rather than higher) frequencies since contact typically increases the constraints on a structure and raises its natural frequencies.


Author(s):  
J. M. Allen ◽  
L. B. Erickson

A NASTRAN finite element analysis of a free standing gas turbine blade is presented. The analysis entails calculation of the first four natural frequencies, mode shapes, and relative vibratory stresses, as well as deflections and stresses due to centrifugal loading. The stiffening effect of the centrifugal force field was accounted for by using NASTRAN’s differential stiffness option. Natural frequencies measured in a rotating test correlated well with computed results. Areas of maximum vibratory stress (fundamental mode) coincided with the three zones of crack initiation observed in a metallographic examination of a fatigue failure. Airfoil stress distributions were found to be significantly different from that predicted by generalized beam theory, especially near the airfoil-platform junction.


Author(s):  
D. S. Whitehead ◽  
D. H. Evans

An analysis is presented to predict flutter in a wheel of turbine blades which are connected together into a number of identical groups. The natural frequencies and mode shapes of a group are assumed to be known. The unsteady aerodynamic coefficients for free-standing blades are assumed to be known from an unsteady aerodynamic program, and FINSUP is used here. The work fed into the vibration by the aerodynamic forces is then calculated. This is illustrated by two examples of low pressure steam turbine blade rows GR-1 and GR-2. On GR-1 the three modes considered are all found to be stable, but on GR-2 the lowest frequency mode shows some instability. Tying the blades together in groups is found to be stabilizing. Blade response, measured by a Blade Vibration Monitor at two different installations, is shown for a range of operating conditions. The measured responses indicate the GR-1 blade is stable whereas the GR-2 blade shows, at the lowest frequency, high response that is dependent on turbine operating conditions.


2003 ◽  
Vol 68 (6) ◽  
Author(s):  
S. T. Wang ◽  
X. F. Han ◽  
Z. Q. Liu ◽  
A. Cady ◽  
M. D. Radcliffe ◽  
...  

Author(s):  
Hans D. Giesecke

Large turbine generators have torsional modes of vibration that can be excited from the electrical grid by torques applied through the generator. The most significant of these torques has a frequency at twice the grid frequency and is due to the negative sequence current in the generator caused by operation at unbalanced load or during grid transients. When the twisting modes of the low pressure turbine rotors combine with the vibratory modes of the last few stages of blade rows, and the frequency of the combined torsional mode is close to the frequency of the exciting torques, significant vibratory response of the shaft and blades can occur. The accumulated fatigue damage caused by such vibration over time can result in failure of the blades. Since this low damped torsional vibration can not be seen on any of the plant instrumentation, it can result in the loss of low pressure blades with little or no warning. To ensure that the turbine generator is not susceptible to damage from the torsional vibratory response of these modes, it is necessary to confirm that the torsional frequencies are sufficiently removed from the frequency of the exciting torques when the turbine generator is operating. For a large turbine generator, the torsional modes of concern are often between the 15th to 25th mode of vibration. Analysis techniques may not be able to determine the frequency of these modes within the accuracy required to ensure that they are not excited. The only reliable way to determine the natural frequencies of such modes with sufficient accuracy is to measure them directly while the turbine generator is operating. On-line monitoring is often the preferred approach for such measurements since it does not impact the operation of the plant and it determines the torsional natural frequencies at the plant operating conditions. Torsional natural frequencies tend to change as a function of turbine generator speed while the turbine generator is off-line and as a function of power while the turbine generator is on-line. On-line monitoring uses sensitive instrumentation and time averaging techniques to determine the torsional natural frequencies of a turbine generator from random vibration of the shaft while the turbine generator is operating. Identifying the torsional mode that is associated with each measured frequency requires the combination of a good analytic model of the turbine generator and an understanding of how the torsional frequencies react to specific changes in operating parameters. The analytic and measurement techniques that have been developed through experience and implemented during numerous on-line measurements are described in this paper. These techniques have also been used to determine blade stress response levels to torsional excitation in order to evaluate the susceptibility of a specific turbine generator to damage from torsional vibration. In this regard, there is some evidence that the torsional response of the turbine generator can be amplified by the steam flow through the blade path. Finally, these techniques can be used to evaluate any specific transient that occurs during operation of the plant with respect to its impact on fatigue usage of the turbine blades and shaft. If necessary, modifications can be designed to shift the torsional natural frequencies away from the problem torques once the complete response of the turbine generator to torsional excitation is understood.


1997 ◽  
Vol 499 ◽  
Author(s):  
Lars Stixrude ◽  
R. M. Wentzcovitch ◽  
C. Da Silva ◽  
B. Kiefer

ABSTRACTWe discuss the behavior of two minerals of the same composition, Mg2SiO4 but different structures: forsterite and ringwoodite. Ab initio plane-wave pseudopotential results are discussed in terms of the full elastic constant tensors of these phases and their elastic anisotropy. The structures of the two minerals, based on pseudo-hep and pseudo-fee packing of oxygens, respectively, show very different behavior at high pressure. While the elastic anisotropy of olivine depends weakly on pressure between 0 and 25 GPa, the anisotropy of ringwoodite decreases with pressure initially, vanishing at 17 GPa before increasing again at higher pressure. This unusual behavior is understood in terms of a change of sign of the combination of elastic constants c2+2c44-c11, and a resulting interchange of fast and slow directions of acoustic wave propagation. To gain insight into the origin of their elastic behavior, the ab initio results are compared with a simple model based on the O sub-lattice.


2017 ◽  
Vol 139 (10) ◽  
Author(s):  
Tomokazu Tachikawa ◽  
Daisuke Iba ◽  
Nobuaki Kurita ◽  
Morimasa Nakamura ◽  
Ichiro Moriwaki

In order to improve the accuracy of skived gears by means of reducing vibrations that are often observed during the cutting process, a simple model for calculating the cutting forces of skiving process is presented and also its effectiveness was discussed. The model is characterized by simple geometrical calculations, and the cutting forces were assumed as sum of vectors that represent the penetration of cutting edges. In the model, multiple cutting edges that are simultaneously meshing with the workpiece were considered. Distinguished oscillating frequencies of the calculated cutting forces and the natural frequencies of the clamped workpiece and of the cutter were carefully analyzed in order to predict the cutter rotation speed that was most likely to reduce undesired vibrations. Processing experiments conducted at several cutter rotation speeds showed that the predicted cutter rotation speed which could significantly reduce undesired vibrations was very effective and enabled the quality of a skived gear to be improved. Consequently, the proposed calculation model was enough effective and useful for operation conditions such as cutter rotation speed being determined.


1983 ◽  
Vol 20 (03) ◽  
pp. 257-271
Author(s):  
Michael G. Parsons

The effect of propeller coupling on the torsional and longitudinal vibration of marine propulsion shafting is studied. Recent research on the nature and computation of propeller added mass and damping is reviewed. It is now possible to estimate the inertia coupling and velocity coupling characteristics which couple the torsional and longitudinal vibration of a marine propeller. Regression equations suitable for estimating the torsional and longitudinal added mass and damping of 4-, 5-, 6-, and 7-bladed Wageningen B-Series propellers are presented in the Appendix. The torsional and longitudinal modeling of a typical marine propulsion plant is reviewed. The numerical techniques and computer programs used in solving the free vibration problem for natural frequencies and mode shapes and the forced vibration problem for vibratory response are introduced. Results for a realistic numerical example are presented to compare the natural frequencies, mode shapes, and vibratory response which are obtained when the propeller coupling is neglected with those obtained with the propeller coupling included. In general, the natural frequencies are shown to change by less than 2 percent. The modes shapes can show significant change. Vibratory response can be either increased or decreased. Some cases show that neglecting the propeller coupling can result in more than a 50 percent underprediction of the vibratory response. The coupling effects are greatest when a torsional natural frequency obtained with the propeller coupling neglected and a longitudinal natural frequency obtained with the propeller coupling neglected converge.


2019 ◽  
pp. 0309524X1988243
Author(s):  
Erik Möllerström ◽  
Fredric Ottermo

A simple model for accounting for tower mass when estimating the first-mode eigenfrequency of a semi-guy-wired tower has been derived. This extends previous work where an analytical model of the semi-guy-wired tower of a 200-kW vertical-axis wind turbine was developed. The model was primarily used to estimate the eigenfrequencies as a result of adding guy wires to a free-standing tower (thus creating a semi-guy-wired setup). However, a weakness with the model was that the tower mass was accounted for in a rough way that essentially ignored the guy wires, which gave a larger-than-necessary error. In this work, an effective top mass, that takes into account the tower mass and the constraints from the guy wires, is derived to achieve a higher accuracy when estimating the first-mode eigenfrequency. This, together with the earlier models, gives a more complete method to estimate the eigenfrequencies for a semi-guy-wired wind turbine.


Sign in / Sign up

Export Citation Format

Share Document