Particle Image Velocimetry Analysis on the Effect of Stator Loading on Transonic Blade-Row Interactions

2012 ◽  
Vol 134 (6) ◽  
Author(s):  
Scott B. Reynolds ◽  
Steven E. Gorrell ◽  
Jordi Estevadeordal

Experiments were performed to investigate interactions between a loaded stator and transonic rotor. The blade row interaction (BRI) rig was used to simulate an embedded transonic fan stage with realistic geometry (thin trailing edge), which produces a wake through diffusion. Details of the unsteady flow field between the stator and rotor were obtained using PIV. Flow-visualization images and PIV data that facilitate analysis of vortex shedding, wake motion, and wake-shock-interaction phenomena are presented. Stator wake and rotor-bow-shock interactions were analyzed for three stator/rotor axial spacings and two stator loadings. Specific shed vortices and wake topological features were isolated for each configuration. The data analysis focuses on measuring the vortex size, strength, and location as it forms on the stator trailing edge and propagates downstream into the rotor passage. It was observed that vortex shedding is synchronized to the passing of a rotor bow shock. Results show that the circulation of a vortex increased by 19% to 23% from far to close spacing due to the increased strength of the rotor bow shock impacting the stator trailing edge. Reduction in stator loading decreased shed vortex circulation for the same stator/rotor axial spacing by 20% to 25%. Pitchwise radius of vortices also decreased by 13% to 19% from far to close spacing. Such changes in vortex size and strength should be accounted for to predict the effect of unsteady blade-row interactions on transonic compressor performance.

Author(s):  
Scott B. Reynolds ◽  
Steven E. Gorrell ◽  
Jordi Estevadeordal

Experiments have been performed to investigate interactions between a loaded stator and transonic rotor. The Blade Row Interaction (BRI) rig is used to simulate an embedded transonic fan stage with realistic geometry (thin trailing edge) which produces a wake through diffusion. Details of the unsteady flow field between the stator and rotor were obtained using PIV. Flow-visualization images and PIV data that facilitate analysis of vortex shedding, wake motion, and wake-shock-interaction phenomena are presented. Stator wake and rotor-bow-shock interactions are analyzed for three stator/rotor axial spacings, and two stator loadings. Specific shed vortices and wake topological features are isolated for each configuration. The data analysis focuses on measuring the vortex size, strength, and location as it forms on the stator trailing edge and propagates downstream into the rotor passage. It was observed that vortex shedding is synchronized to the passing of a rotor bow shock. Results show that the circulation of a vortex increased by 19 to 23% from far to close spacing due to the increased strength of the rotor bow shock impacting the stator trailing edge. Reduction in stator loading decreased shed vortex circulation for the same stator/rotor axial spacing by 20 to 25%. Pitchwise radius of vortices also decreased by 13 to 19% from far to close spacing. Such changes in vortex size and strength should be accounted for to predict the effect of unsteady blade-row interactions on transonic compressor performance.


Author(s):  
Steven E. Gorrell ◽  
David Car ◽  
Steven L. Puterbaugh ◽  
Jordi Estevadeordal ◽  
Theodore H. Okiishi

The effects of varying axial gap on the unsteady flow field between the stator and rotor of a transonic compressor stage are important because they can result in significant changes in stage mass flow rate, pressure rise and efficiency. Some of these effects are analyzed with measurements using Digital Particle Image Velocimetry (DPIV) and with time-accurate simulations using the 3D unsteady Navier-Stokes CFD solver TURBO. Generally there is excellent agreement between the measurements and simulations, instilling confidence in both. Strong vortices of the wake can break up the rotor bow shock and contribute to loss. At close spacing vortices are shed from the trailing edge of the upstream stationary blade row in response to the unsteady, discontinuous pressure field generated by the downstream rotor bow shock. Shed vortices increase in size and strength and generate more loss as spacing decreases, a consequence of the effective increase in rotor bow shock strength at the stationary blade row trailing edge. A relationship for the change in shed vorticity as a function of rotor bow shock strength is presented that predicts the difference between close and far spacing TURBO simulations.


Author(s):  
Steven E. Gorrell ◽  
Theodore H. Okiishi ◽  
William W. Copenhaver

A previously unidentified loss producing mechanism resulting from the interaction of a transonic rotor blade-row with an upstream stator blade-row is described. This additional loss occurs only when the two blade rows are spaced closer together axially. Time-accurate simulations of the flow and high-response static pressure measurements acquired on the stator blade surface reveal important aspects of the fluid dynamics of the production of this additional loss. At close spacing the rotor bow shock is chopped by the stator trailing edge. The chopped bow shock becomes a pressure wave on the upper surface of the stator that is nearly normal to the flow and that propagates upstream. In the reference frame relative to this pressure wave, the flow is supersonic and thus a moving shock wave that produces an entropy rise and loss is experienced. The effect of this outcome of blade-row interaction is to lower the efficiency, pressure ratio, and mass flow rate observed as blade-row axial spacing is reduced from far to close. The magnitude of loss production is affected by the strength of the bow shock and how much it turns as it interacts with the trailing edge of the stator. At far spacing the rotor bow shock degenerates into a bow wave before it interacts with the stator trailing edge and no significant pressure wave forms on the stator upper surface. For this condition, no additional loss is produced.


2003 ◽  
Vol 125 (2) ◽  
pp. 336-345 ◽  
Author(s):  
Steven E. Gorrell ◽  
Theodore H. Okiishi ◽  
William W. Copenhaver

A previously unidentified loss producing mechanism resulting from the interaction of a transonic rotor blade row with an upstream stator blade row is described. This additional loss occurs only when the two blade rows are spaced closer together axially. Time-accurate simulations of the flow and high-response static pressure measurements acquired on the stator blade surface reveal important aspects of the fluid dynamics of the production of this additional loss. At close spacing the rotor bow shock is chopped by the stator trailing edge. The chopped bow shock becomes a pressure wave on the upper surface of the stator that is nearly normal to the flow and that propagates upstream. In the reference frame relative to this pressure wave, the flow is supersonic and thus a moving shock wave that produces an entropy rise and loss is experienced. The effect of this outcome of blade-row interaction is to lower the efficiency, pressure ratio, and mass flow rate observed as blade-row axial spacing is reduced from far to close. The magnitude of loss production is affected by the strength of the bow shock and how much it turns as it interacts with the trailing edge of the stator. At far spacing the rotor bow shock degenerates into a bow wave before it interacts with the stator trailing edge and no significant pressure wave forms on the stator upper surface. For this condition, no additional loss is produced.


1999 ◽  
Vol 121 (3) ◽  
pp. 428-435 ◽  
Author(s):  
D. L. Sondak ◽  
D. J. Dorney

Vortex shedding in a turbomachine blade row is affected by the passing of blades in the adjacent downstream blade row, but these effects have not been examined in the literature. A series of flow simulations has been performed to study vortex shedding in a turbine stage, and to quantify the blade interaction effects on the unsteady pressure response. The numerical issues of spatial order of accuracy and the use of Newton subiterations were investigated first. Second-order spatial accuracy was shown to be inadequate to model the shedding frequency response and time-averaged base pressure accurately. For the small time step employed for temporal accuracy, Newton iterations were shown to be unnecessary. The effects of the adjacent blade row were examined by comparing the shedding frequency response for the stage simulations to the response for isolated cascades. The vane shedding was shown to occur exactly on a series of harmonics of the blade passing frequency for the stage case, compared to a single predominant frequency for the isolated cascade. Losses were also examined in the wake region. It was shown that close to the trailing edge, losses were mainly due to wake mixing. Farther downstream of the trailing edge, losses were predominantly due to the trailing edge shock wave.


Author(s):  
Douglas L. Sondak ◽  
Daniel J. Dorney

Vortex shedding in a turbomachine blade row is affected by the passing of blades in the adjacent downstream blade row, but these effects have not been examined in the literature. A series of flow simulations has been performed to study vortex shedding in a turbine stage, and to quantify the blade interaction effects on the unsteady pressure response. The numerical issues of spatial order of accuracy and the use of Newton subiterations were first investigated. Second order spatial accuracy was shown to be inadequate to accurately model the shedding frequency response and time-averaged base pressure. For the small time step employed for temporal accuracy, Newton iterations were shown to be unnecessary. The effects of the adjacent blade row were examined by comparing the shedding frequency response for the stage simulations to the response for isolated cascades. The vane shedding was shown to occur exactly on a series of harmonics of the blade passing frequency for the stage case, compared to a single predominant frequency for the isolated cascade. Losses were also examined in the wake region. It was shown that close to the trailing edge, losses were mainly due to wake mixing. Farther downstream of the trailing edge, losses were predominantly due to the trailing edge shock wave.


Author(s):  
Kenneth P. Clark ◽  
Steven E. Gorrell

Multiple high-fidelity time-accurate computational fluid dynamics simulations were performed to investigate the effects of upstream stator loading and rotor shock strength on vortex shedding characteristics in a single stage transonic compressor. Three loadings on the upstream stator row of decreased, nominal, and increased were studied. The time-accurate URANS code TURBO was used to generate periodic, quarter annulus simulations of the Blade Row Interaction compressor rig. It was observed that vortex shedding was synchronized to the passing of a rotor bow shock. Results show that vortex size and strength increase with stator loading. “Normal” and “large” shock-induced vortices formed on the stator trailing edge immediately after the shock passing, but the “large” vortices were strengthened at the trailing edge due to a low velocity region at the suction surface. This low velocity region was generated upstream on the suction surface from a shock-induced thickening of the boundary layer or separation bubble. The circulation of the “large” vortices was greater than the “normal” vortices by a factor of 1.7, 1.9 and 2.1 for decreased, nominal and increased deswirler loadings. At decreased loading only 24% of the measured vortices were considered “large” while at nominal loading 58% were “large”. An understanding of the unsteady interactions associated with blade loading and rotor shock strength in transonic stages will help compressor designers account for unsteady flow physics at design and off-design operating conditions.


2005 ◽  
Vol 128 (4) ◽  
pp. 616-626 ◽  
Author(s):  
Steven E. Gorrell ◽  
David Car ◽  
Steven L. Puterbaugh ◽  
Jordi Estevadeordal ◽  
Theodore H. Okiishi

The effects of varying axial gap on the unsteady flow field between the stator and rotor of a transonic compressor stage are important because they can result in significant changes in stage mass flow rate, pressure rise, and efficiency. Some of these effects are analyzed with measurements using digital particle image velocimetry (DPIV) and with time-accurate simulations using the 3D unsteady Navier-Stokes computational fluid dynamics solver TURBO. Generally there is excellent agreement between the measurements and simulations, instilling confidence in both. Strong vortices of the wake can break up the rotor bow shock and contribute to loss. At close spacing vortices are shed from the trailing edge of the upstream stationary blade row in response to the unsteady, discontinuous pressure field generated by the downstream rotor bow shock. Shed vortices increase in size and strength and generate more loss as spacing decreases, a consequence of the effective increase in rotor bow shock strength at the stationary blade row trailing edge. A relationship for the change in shed vorticity as a function of rotor bow shock strength is presented that predicts the difference between close and far spacing TURBO simulations.


2015 ◽  
Vol 137 (12) ◽  
Author(s):  
Kenneth P. Clark ◽  
Steven E. Gorrell

Multiple high-fidelity time-accurate computational fluid dynamics simulations were performed to investigate the effects of upstream stator loading and rotor shock strength on vortex shedding characteristics in a single-stage transonic compressor. Three loadings on the upstream stator row of decreased, nominal, and increased loading in conjunction with three axial spacings of close, mid, and far were studied for this analysis. The time-accurate urans code turbo was used to generate periodic, quarter annulus simulations of the blade row interaction (BRI) compressor rig. It was observed that vortex shedding was synchronized to the passing of a rotor bow shock. Results show that vortex strength increases linearly with stator loading and rotor bow shock strength. “Normal” and “large” shock-induced vortices formed on the stator trailing edge (TE) immediately after the shock passing, but the large vortices were strengthened at the TE due to a low-velocity region on the suction surface. This low-velocity region was generated upstream on the suction surface from a shock-induced thickening of the boundary layer or separation bubble. The circulation of the large vortices was greater than the normal vortices by a factor of 1.7, 1.83, and 2.04 for decreased, nominal, and increased deswirler loadings. At decreased loading, only 24% of the measured vortices were considered large, while at nominal loading 58% were large. A model was developed to predict shock-induced vortex circulation from a known rotor bow shock strength, stator diffusion factor, and near-wake parameters. The model predicts the average vortex circulation very well, with 5% difference between predicted and measured values. An understanding of the unsteady interactions associated with blade loading and rotor shock strength in transonic stages will help compressor designers account for unsteady flow physics at design and off-design operating conditions.


AIAA Journal ◽  
2001 ◽  
Vol 39 ◽  
pp. 787-793
Author(s):  
Wei Ning ◽  
Li He

Sign in / Sign up

Export Citation Format

Share Document