Effects of Solid Particle Ingestion Through an HP Turbine

Author(s):  
Adel Ghenaiet

Modern gas turbines operate in severe dusty environments, and because of such harsh operating conditions, their blades experience significant degradation in service. This paper presents a numerical study of particle dynamics and erosion in an hp axial turbine stage. The flow field is solved separately from the solid phase and constitutes the necessary data in the particle trajectories simulations using a Lagrangian tracking model based on the finite element method. Several parameters consider a statistical description such as particle size, shape and rebound, in addition to the turbulence effect. A semi empirical erosion correlation is used to estimate erosion contours and blades deteriorations, knowing the locations and conditions of impacts. The trajectory and erosion results show high erosion rates over the pressure side of NGV near trailing edge, in addition to extreme erosion observed toward the root corner, due to high number of particles impacting with high velocities. On the suction side, erosion is mainly over a narrow strip from leading edge. Erosion in the rotor blade is shown along the leading edge and spreading over the fore of the blade suction side, owing to a flux of particles entering at high velocities and incidence. On the pressure side, regions of dense erosion are observed near the leading edge and trailing edge as well as the tip corner. Critical erosion spots seen over NGV and rotor blade are signs of a premature failure.

Author(s):  
Dieter E. Bohn ◽  
Karsten A. Kusterer

A leading edge cooling configuration is investigated numerically by application of a 3-D conjugate fluid flow and heat transfer solver, CHT-Flow. The code has been developed at the Institute of Steam and Gas Turbines, Aachen University of Technology. It works on the basis of an implicit finite volume method combined with a multi-block technique. The cooling configuration is an axial turbine blade cascade with leading edge ejection through two rows of cooling holes. The rows are located in the vicinity of the stagnation line, one row is on the suction side, the other row is on the pressure side. The cooling holes have a radial ejection angle of 45°. This configuration has been investigated experimentally by other authors and the results have been documented as a test case for numerical calculations of ejection flow phenomena. The numerical domain includes the internal cooling fluid supply, the radially inclined holes and the complete external flow field of the turbine vane in a high resolution grid. Periodic boundary conditions have been used in the radial direction. Thus, end wall effects have been excluded. The numerical investigations focus on the aerothermal mixing process in the cooling jets and the impact on the temperature distribution on the blade surface. The radial ejection angles lead to a fully three dimensional and asymmetric jet flow field. Within a secondary flow analysis it can be shown that complex vortex systems are formed in the ejection holes and in the cooling fluid jets. The secondary flow fields include asymmetric kidney vortex systems with one dominating vortex on the back side of the jets. The numerical and experimental data show a good agreement concerning the vortex development. The phenomena on the suction side and the pressure side are principally the same. It can be found that the jets are barely touching the blade surface as the dominating vortex transports hot gas under the jets. Thus, the cooling efficiency is reduced.


Author(s):  
Adel Ghenaiet

This paper presents a numerical study of particle laden gas flow through a two-stage hp axial turbine, by means of an in-house code based on the Lagrangian tracking model and the finite element method. As fly-ash solid particles trajectories and locations of impacts are predicted, the local erosion rates and the deteriorations of blades are assessed. The computed trajectories provide a detailed description of particles behaviors and reveal that particle impacts on the aft of vane pressure side usually lead to significant variations in the directions of particles to the next rotor blade, and subsequently particles impact the suction side. The plots of equivalent erosion rates indicate the vanes and blades locations which suffer more erosion. The first vane pressure surface is impacted more than any other component, but higher rates are seen at the top corner from trailing edge. The critical regions of erosion wear in the first rotor are observed over the top of blade leading edge extending along the tip as well as a rounding of the top corner from trailing edge. In the second vane, the regions of higher erosion are revealed over the last third of leading edge and the top corner extending along tip. The erosion in the second rotor is over a large area of suction side till the tip corner. The predicted areas of extreme erosion, also shown by the deteriorated profiles, are indicators for anticipated vanes and blades failures.


Author(s):  
Seung Chul Back ◽  
Garth V. Hobson ◽  
Seung Jin Song ◽  
Knox T. Millsaps

An experimental investigation has been conducted to characterize the influence of surface roughness location and Reynolds number on compressor cascade performance. Flow field surveys have been conducted in a low-speed, linear compressor cascade. Pressure, velocity, and flow angles have been measured via a 5-hole probe, pitot probe, and pressure taps on the blades. In addition to the entirely smooth and entirely rough blade cases, blades with roughness covering the leading edge; pressure side; and 5%, 20%, 35%, 50%, and 100% of suction side from the leading edge have been studied. All of the tests have been done for Reynolds number ranging from 300,000 to 640,000.Cascade performance (i.e. blade loading, loss, and deviation) is more sensitive to roughness on the suction side than pressure side. Roughness near the trailing edge of suction side increases loss more than that near the leading edge. When the suction side roughness is located closer to the trailing edge, the deviation and loss increase more rapidly with Reynolds number. For a given roughness location, there exists a Reynolds number at which loss begins to visibly increase. Finally, increasing the area of rough suction surface from the leading edge reduces the Reynolds number at which the loss coefficient begins to increase.


Author(s):  
S. Naik ◽  
J. Krueckels ◽  
M. Gritsch ◽  
M. Schnieder

This paper investigates the aerodynamic and film cooling effectiveness characteristics of a first stage turbine high lift guide vane and its corresponding downstream blade. The vane and blade geometrical profiles and operating conditions are representative of that normally found in a heavy-duty gas turbine. Both the vane and the blade airfoils consist of multi-row film cooling holes located at various axial positions along the airfoil chord. The film cooling holes are geometrically three-dimensional in shape and depending on the location on the airfoil; they can be either symmetrically fan shaped or non-symmetrically fan shaped. Additionally the film cooling holes can be either compounded or in-line with the external flow direction. Numerical studies and experimental investigations in a linear cascade have been conducted at vane and blade exit isentropic Mach number of 0.8. The influence of the coolant flow ejected from the film cooling holes has been investigated for both the vane and the blade profiles. For the nozzle guide vane, the measured film cooling effectiveness compared well with the predictions, especially on the pressure side. The suction side film cooling effectiveness, which consisted of two pre-throat film rows, proved very effective up-to the suction side trailing edge. For the blade, there was a reasonable comparison between the measured and predicted film cooling effectiveness. Again the blade pre-throat fan shaped cooling holes proved very effective up-to the suction side trailing edge. For the vane, the impact of varying the blowing ratios showed a strong variation in the film cooling effectiveness on the pressure side. However, on the blade, the effect of varying the blowing ratio had a greater impact on the suction side film effectiveness compared to the pressure side.


Author(s):  
Stefan Wolff ◽  
Leonhard Fottner ◽  
Sabine Ardey

In order to close the gap in knowledge concerning the influence of periodic unsteady inflow conditions on the mixing process of film cooling jets, time resolved flow velocities and turbulent fluctuations in the injection zone of a linear large scale high pressure turbine cascade with leading edge film cooling were measured by means of the 3D hot-wire anemometry. The periodic impinging wakes are generated by a wake generator consisting of moving bars upstream of the cascade inlet plane. Near the stagnation point the blades are equipped with one row of film cooling holes on the suction side and one row on the pressure side. Mach number and Reynolds number are set to values typically found in modern gas turbines. At a density ratio of unity the blowing ratio is set to M = 0.7 and the Strouhal number is set to Sr = 0.31. The general flow structures which were determined by the steady state measurements — i.e. normal jet in cross flow behavior with the kidney shape vortex on the suction side and a second counter rotating pair of vortices underneath the kidney vortex on the pressure side — have been detected by the unsteady measurements as well. The large recirculation zone behind the pressure side injection hole, caused by the strong adverse pressure gradient and the lift off of the coolant jet, is not suppressed by the passing wakes but rather reinforced promoting potential hot spots in this area. The suction side coolant jet has almost disappeared when the wake leaves the suction side region. It undergoes a recovery process until the next wake hits again on the suction side.


Author(s):  
Sabine Ardey ◽  
Leonhard Fottner

To increase the understanding of the aerodynamic processes dominating the flow field of turbine bladings with leading edge film cooling, isothermal investigations were carried out on a large scale high pressure turbine cascade. Near the stagnation point the blades are equipped with one row of film cooling holes on the suction side and one on the pressure side. Blowing ratio, turbulence intensity, Mach number, and Reynolds number are set to values typically found in modern gas turbines. Experimental data of the cascade flow were obtained by pneumatic probes and static pressure tappings. The flow field was visualized by Schlieren and oil flow techniques. For detailed investigations near the blowing holes the Laser Transit Velocimetry and the three dimensional Hot Wire Anemometry were used. The flow field measurements in the near hole region of the suction side show the typical kidney shaped vortex pair. A local suction peak on the pressure side causes a large recirculation area behind the holes on the pressure side and induces separation bubbles in between the pressure side holes. This leads to the generation of two pairs of vortices: The kidney-vortex is located on top of a second vortex pair and a trough flow that fills up the deficit of the recirculation. Thus the film cooling air is detached from the pressure side surface. In addition to the mean flow vectors Reynolds stress components are a good means to judge the propagation of the jet. In spite of the complex flow pattern occurring on each single jet, the surveyed loss-increase due to the leading edge blowing can be predicted by the mixing layer model.


Author(s):  
B. D. J. Schreiner ◽  
M. Wilson ◽  
Y. S. Li ◽  
C. M. Sangan

Abstract Endwall contouring is used to increase the aerodynamic efficiency of both compressor and turbine stages in industrial gas turbines and aeroengines. The complex interaction between the secondary air-leakage, used to cool the turbine disc, and the mainstream gas path, leads to an unsteady flow field that is challenging to compute. Current endwall designs have shown sensitivity to the introduction of secondary air, with stage efficiency improvements being reduced, or in the limit, eliminated altogether. A computational study of an engine-representative turbine stage was conducted using an unsteady RANS solver. Previously published computations of the baseline axisymmetric endwall were validated with experimental data from a geometrically similar test rig. Understanding from this prior study was used to inform the design process for contoured endwalls, namely through the identification of three key geometric features: the leading-edge feature; the suction-side trough; the pressure-side trough. The baseline axisymmetric endwall showed periodic unsteadiness, large secondary flow features and an egress plume which dominated the aerodynamics of the stage. The implementation of a suction-side trough (i.e. making the endwall non-axisymmetric) reduced the magnitude of the unsteadiness by controlling the path of the egress plume. The trough also reduced the span-wise migration of the egress plume through the passage and provided modest control over pitchwise position. In corroboration with the findings of other authors, the introduction of a leading-edge feature was also used to reduce the leading edge horseshoe vortex,. The pressure-side trough enabled the prominence of the leading-edge feature to be enhanced, however it increased the span-wise migration of the egress plume. Insight generated from computations of the three distinct geometric features resulted in an improved endwall concept; the improved endwall demonstrated a 0.4% net efficiency gain for the stage relative to the cylindrical baseline.


2008 ◽  
Author(s):  
H. Wu ◽  
F. Soranna ◽  
T. Michael ◽  
J. Katz ◽  
S. Jessup

Recent upgrades to the turbomachinery facility at JHU enable measurements of performance, as well as flow structure, turbulence and cavitation within a water-jet pump. The rotor, stator and pump casing in this optically index-matched facility are made of acrylic that has the same optical index of refraction as the working fluid, a concentrated solution of NaI in water. The essentially “invisible” blades allow unobstructed view and access to optical flow measurement techniques. Initial tests in water focus on observations on occurrence of cavitation in the vicinity of the narrow tip-gap. For the present design and operating conditions, near the leading edge, cavitation in the tip corner of the pressure side causes accumulation of bubbles along the pressure side that extends to mid blade. As rollup of a tip vortex starts, these bubbles cross the tip gap to the suction side, and become primary nuclei for cavitation inception within the tip leakage vortex (TLV). Bursting of this tip vortex as it migrates towards the pressure side of the neighboring blade generates a cloud of bubbles along the aft section of the passage. As the flow in the tip gap increases upstream of the trailing edge, cavitation also develops within the gap, along the pressure side corner.


Author(s):  
Björn Laumert ◽  
Hans Mårtensson ◽  
Torsten H. Fransson

This paper presents the results from three-dimensional (3D) steady and unsteady Navier-Stokes computations, performed on the transonic VKI BRITE EURAM test turbine stage. The work aimed at giving deeper insight in the aerodynamics of the turbine stage. The analysis has been carried out with the nominal stator trailing edge ejection slot geometry and cooling flow ejection. Additionally a simplified rounded stator trailing edge was employed. The results from the unsteady computations were compared with measured pressure perturbation traces at 22 locations around the rotor blade at midspan. Computations with both the ejection slot and the rounded stator trailing edge geometry were in good agreement with the measurements on the pressure side and half chord of the rotor blade’s suction side. Measurements and computations showed less good agreement downstream a weak shock on the suction side of the rotor blade. The measured pressure double peak in the rotor blade leading edge region is only observed in the computations with the ejection slot geometry.


Author(s):  
Prasert Prapamonthon ◽  
Huazhao Xu ◽  
Jianhua Wang ◽  
Ge Li

The thermal efficiency of gas turbine engines increases with turbine inlet temperature (TIT) directly. However, the TIT is limited by the allowable temperature of current blade materials. Film cooling technique is an effective method to maintain turbine vane working smoothly under high TIT conditions. The adiabatic film effectiveness has been widely employed to understand film cooling mechanism. Therefore, the prediction of the adiabatic effectiveness of gas turbine engines under real operating conditions is essential. The showerhead film cooled turbine vane reported by L. P. Timko (NASA CR-168289) is adopted in the present study. There are two rows of film holes on the leading edge, three rows on the pressure side, and two rows on the suction side. All holes are cylindrical, which are placed at an angle of 45 degrees to the vane surface in the span-wise direction. This numerical investigation discusses the influences of free stream turbulence intensity on the adiabatic film effectiveness in the vane leading edge region and its vicinity. Five two-equation turbulence models based on Reynolds Averaged Navier-Stokes (RANS) are employed to predict the adiabatic film effectiveness under real operating conditions at a blowing ratio (BR) of 1.41 and three free stream turbulence intensities (Tu=3.3, 10, and 20%). The adiabatic film effectiveness on the vane surface at 8, 52.5, and 89% span in an x/C range between −0.4 and 0.4 is presented. Obviously, the numerical results predicted by all five models show that on the suction side, the increasing free stream turbulence intensity can reduce film effectiveness except at 8% span. On the pressure side, the RNG k-ε, Realizable k-ε and SST k-ω models predict the same trend of the adiabatic film effectiveness, especially the RNG k-ε and SST k-ω models. Those three models predict that the locally adiabatic film effectiveness (especially near film holes) can be improved when turbulence intensity increases. However, at a span of 89% within the x/C range between −0.4 and −0.2, all k-ε models and SST k-ω model predict that the increase of turbulence intensity can reduce the adiabatic film effectiveness. In addition, the film effectiveness contours show a significant variation of film effectiveness predicted by the five turbulence models on the leading edge when turbulence intensity increases. For the near-pressure side, all models except the Standard k-ω model predict that the high turbulence intensity can reduce the film spreading from film holes dramatically.


Sign in / Sign up

Export Citation Format

Share Document