Comparison of a Conventional Thermal Analysis of a Turbine Cascade to a Full Conjugate Heat Transfer Computation

Author(s):  
Christoph Starke ◽  
Erik Janke ◽  
Toma´sˇ Hofer ◽  
Davide Lengani

Recent development in commercial CFD codes offers possibilities to include the solid body in order to perform conjugate heat transfer computations for complex geometries. The current paper aims to analyse the differences between a conjugate heat transfer computation and conventional uncoupled approaches where a heat transfer coefficient is first derived from a flow solution and then taken as boundary condition for a thermal conduction analysis of the solid part. Whereas the thermal analyses are done with a Rolls-Royce in-house finite element code, the CFD as well as the conjugate heat transfer computation are done using the new version 8 of the commercial code Fine Turbo from Numeca International. The analysed geometry is a turbine cascade that was tested by VKI in Brussels within the European FP6 project AITEB 2. First, the paper presents the aerodynamic results. The pure flow solutions are validated against pressure measurements of the cascade test. Then, the heat transfer from flow computations with wall temperature boundary conditions is compared to the measured heat transfer. Once validated, the heat transfer coefficients are used as boundary condition for three uncoupled thermal analyses of the blade to predict its surface temperatures in a steady state. The results are then compared to a conjugate heat transfer method. Therefore, a mesh of the solid blade was added to the validated flow computation. The paper will present and compare the results of conventional uncoupled thermal analyses with different strategies for the wall boundary condition to results of a conjugate heat transfer computation. As it turns out, the global results are similar but especially the over-tip region with its complex geometry and flow structure and where effective cooling is crucial shows remarkable differences because the conjugate heat transfer solution predicts lower blade tip temperatures. This will be explained by the missing coupling between the fluid and the solid domain.

Author(s):  
Tom Heuer ◽  
Bertold Engels ◽  
Patrick Wollscheid

One of the most challenging tasks in designing a turbocharger is to guarantee a sufficient lifetime. Turbine housings are critical parts due to their very complex geometry and consequently complicated temperature and stress distributions. Therefore, high thermal loads as well as thermo-mechanical fatigue have to be considered. Calculating the thermal stress distribution in the turbine housing, steady state and transient, can indicate the regions of crack initiation. From this information selective design improvements can be deduced to increase the component lifetime. But the quality of the stress analysis is strongly dependent on a reliable temperature distribution. Taking into account the interdependency of heat transfer between solid walls and fluid, conjugate heat transfer (CHT) calculations can provide temperature data of high accuracy. Since a transient CHT-calculation is still beyond state of the art, a new approach has been developed. Two steady state CHT-calculations serve to determine heat transfer coefficients at engine brake and full load. Beginning with the engine brake temperature distribution, it is assumed that the gas temperature and the mass flow change immediately. Therefore heat transfer coefficients at full load serve as a boundary condition for a subsequent transient solid body calculation simulating the acceleration process. For the deceleration process the full load temperature field is combined with the engine brake heat transfer coefficients. Monitor points give information about the steepest temperature gradients in the material. At discrete time points a steady state stress analysis has to be performed to detect the regions of highest loads. This subsequent step is essential because in a complex geometry like in a spiral housing with a divider and regionally different wall thicknesses, the stress maxima are not necessarily located at the same places as the temperature peaks. For the two steady state CHT-calculations the turbine wheel has been included in order to consider a realistic flow field. Compared to a transient calculation the degree of abstraction is as low as possible because the assumed frozen rotor boundary condition takes into account centrifugal and coriolis forces. This paper demonstrates the calculation procedure considering a twin-entry turbine housing with an integrated manifold designed for a truck application. The computational results are in excellent agreement with thermal shock test data. A second loop with an improved design proves the success of the method.


Author(s):  
Reby Roy ◽  
B. V. S. S. S. Prasad ◽  
S. Srinivasa Murthy

The conjugate heat transfer in a stationary cylindrical cavity with a rotating disk and fluid through-flow is analysed at various rotational speeds ranging from 10000 to 50000 rpm by using a finite volume commercial code. The numerical model and code are validated for a problem, which involves rotation and fluid through-flow. A reduction of the thermal boundary layer thickness and increase in the heat transfer coefficients are observed with increase in the rotational speed. Marked differences are noticed between the Nusselt numbers obtained from the conjugate and constant temperature analyses.


Author(s):  
Kuo-San Ho ◽  
Christopher Urwiller ◽  
S. Murthy Konan ◽  
Jong S. Liu ◽  
Bruno Aguilar

This paper explores the conjugate heat transfer (CHT) numerical simulation approach to calculate the metal temperature for a cooled gas turbine blade. ANSYS CFX14.0 code was selected as the computational fluid dynamic (CFD) tool to perform the CHT simulation. The two-equation SST turbulence model with automatic wall treatment was employed. A full engine test with Silicon Carbide (SiC) chip measurements was performed and used to validate the CHT results. Metal temperatures calculated with the CHT model were compared to engine test data. The results demonstrated good agreement between predicted and measured airfoil metal temperatures. The blade cooling flow prediction was matched to the flow network analysis. This paper describes a process to calculate convection heat transfer coefficients (HTC) for cooling passages and airfoil surfaces using CHT results. This process was made possible because local wall heat flux and fluid temperatures were known. This approach assisted in calibrating an in-house conduction thermal model for steady state thermal analyses.


2001 ◽  
Vol 123 (4) ◽  
pp. 704-708 ◽  
Author(s):  
A. A. Ameri

Experimental and computational studies have been performed to investigate the detailed distribution of convective heat transfer coefficients on the first-stage blade tip surface for a geometry typical of large power generation turbines (>100 MW). In a previous work the numerical heat transfer results for a sharp edge blade tip and a radiused blade tip were presented. More recently several other tip treatments have been considered for which the tip heat transfer has been measured and documented. This paper is concerned with the numerical prediction of the tip surface heat transfer for radiused blade tip equipped with mean-camberline strip (or “squealer” as it is often called). The heat transfer results are compared with the experimental results and discussed. The effectiveness of the mean-camberline strip in reducing the tip leakage and the tip heat transfer as compared to a radiused edge tip and sharp edge tip was studied. The calculations show that the sharp edge tip works best (among the cases considered) in reducing the tip leakage flow and the tip heat transfer.


Author(s):  
Jun Su Park ◽  
Namgeon Yun ◽  
Hokyu Moon ◽  
Kyung Min Kim ◽  
Sin-Ho Kang ◽  
...  

This paper presents thermal analyses of the cooling system of a transition piece, which is one of the primary hot components in a gas turbine engine. The thermal analyses include heat transfer distributions induced by heat and fluid flow, temperature, and thermal stresses. The purpose of this study is to provide basic thermal and structural information on transition piece, to facilitate their maintenance and repair. The study is carried out primarily by numerical methods, using the commercial software, Fluent and ANSYS. First, the combustion field in a combustion liner with nine fuel nozzles is analyzed to determine the inlet conditions of a transition piece. Using the results of this analysis, pressure distributions inside a transition piece are calculated. The outside of the transition piece in a dump diffuser system is also analyzed. Information on the pressure differences is then used to obtain data on cooling channel flow (one of the methods for cooling a transition piece). The cooling channels have exit holes that function as film-cooling holes. Thermal and flow analyses are carried out on the inside of a film-cooled transition piece. The results are used to investigate the adjacent temperatures and wall heat transfer coefficients inside the transition piece. Overall temperature and thermal stress distributions of the transition piece are obtained. These results will provide a direction to improve thermal design of transition piece.


Author(s):  
Long-gang Liu ◽  
Chun-wei Gu ◽  
Xiao-dong Ren

Convective cooling channels are applied in a two-dimensional compressor vane to use the intercooling method to improve the efficiency of Brayton cycle and reduce the temperature of the vane. In this paper, we analyze the effect of coolant to the aerodynamic performance and heat transfer performance of the main stream and the vane. For the case of a two-dimensional compressor vane NACA65-(12A2I8b)10, the vane which has five convective cooling channels has been numerically simulated in different test conditions by discontinuous Galerkin (DG) method. The coolant is supercritical carbon dioxide whose pressure is 10MPa. Conjugate heat transfer method has been used in this paper. The numerical simulation result is similar to the experiment data and has been compared with the result of the vane without cooling channels to prove the effect of cooling channels. Cooling channels have large effect on the distribution of temperature and heat transfer coefficient. In addition, the relationship between Nu and Re on the fluid-solid interface has been analyzed and a suitable empirical equation has been obtained. This work analyzes the effect of intercooling system in the compressor and give several advice on future engineering applications in aero engines and gas turbines.


Author(s):  
X. Liu ◽  
W. Rodi

A detailed experimental study has been conducted on the wake-induced unsteady flow and heat transfer in a linear turbine cascade. The unsteady wakes with passing frequencies in the range zero to 240 Hz were generated by moving cylinders on a squirrel cage device. The velocity fields in the blade-to-blade flow and in the boundary layers were measured with hot-wire anemometers, the surface pressures with a pressure transducer and the heat transfer coefficients with a glue-on hot film. The results were obtained in ensemble-averaged form so that periodic unsteady processes can be studied. Of particular interest was the transition of the boundary layer. The boundary layer remained laminar on the pressure side in all cases and in the case without wakes also on the suction side. On the latter, the wakes generated by the moving cylinders caused transition, and the beginning of transition moves forward as the cylinder-passing frequency increases. Unlike in the flat-plate study of Liu and Rodi (1991a) the instantaneous boundary layer state does not respond to the passing wakes and therefore does not vary with time. The heat transfer increases under increasing cylinder-passing frequency even in the regions with laminar boundary layers due to the increased background turbulence.


Author(s):  
Andreas Jeromin ◽  
Christian Eichler ◽  
Berthold Noll ◽  
Manfred Aigner

Numerical predictions of conjugate heat transfer on an effusion cooled flat plate were performed and compared to detailed experimental data. The commercial package CFX® is used as flow solver. The effusion holes in the referenced experiment had an inclination angle of 17 degrees and were distributed in a staggered array of 7 rows. The geometry and boundary conditions in the experiments were derived from modern gas turbine combustors. The computational domain contains a plenum chamber for coolant supply, a solid wall and the main flow duct. Conjugate heat transfer conditions are applied in order to couple the heat fluxes between the fluid region and the solid wall. The fluid domain contains 2.4 million nodes, the solid domain 300,000 nodes. Turbulence modeling is provided by the SST turbulence model which allows the resolution of the laminar sublayer without wall functions. The numerical predictions of velocity and temperature distributions at certain locations show significant differences to the experimental data in velocity and temperature profiles. It is assumed that this behavior is due to inappropriate modeling of turbulence especially in the effusion hole. Nonetheless, the numerically predicted heat transfer coefficients are in good agreement with the experimental data at low blowing ratios.


Author(s):  
Vikrant Saxena ◽  
Hasan Nasir ◽  
Srinath V. Ekkad

A comprehensive investigation of the effect of various tip sealing geometries is presented on the blade tip leakage flow and associated heat transfer of a scaled up HPT turbine blade in a low-speed wind tunnel facility. The linear cascade is made of four blades with the two corner blades acting as guides. The tip section of a HPT first stage rotor blade is used to fabricate the 2-D blade. The wind tunnel accommodates an 116° turn for the blade cascade. The mainstream Reynolds number based on the axial chord length at cascade exit is 4.83 × 105. The upstream wake effect is simulated with a spoked wheel wake generator placed upstream of the cascade. A turbulence grid placed even farther upstream generates the required free-stream turbulence of 4.8%. The center blade has a tip clearance gap of 1.5625% with respect to the blade span. Static pressure measurements are obtained on the blade surface and the shroud. The effect of crosswise trip strips to reduce leakage flow and associated heat transfer is investigated with strips placed along the leakage flow direction, against the leakage flow and along the chord. Cylindrical pin fins and pitch variation of strips over the tip surface are also investigated. Detailed heat transfer measurements are obtained using a steady state HSI-based liquid crystal technique. The effect of periodic unsteady wake effect is also investigated by varying the wake Strouhal number from 0. to 0.2, and to 0.4. Results show that the trip strips placed against the leakage flow produce the lowest heat transfer on the tips compared to all the other cases with a reduction between 10–15% compared to the plain tip. Results also show that the pitch of the strips has a small effect on the overall reduction. Cylindrical pins fins and strips along the leakage flow direction do not decrease the heat transfer coefficients and in some cases enhance the heat transfer coefficients by as much as 20%.


Author(s):  
Taher Schobeiri ◽  
Eric McFarland ◽  
Frederick Yeh

In this report the results of aerodynamic and heat transfer experimental investigations performed in a high Reynolds number turbine cascade test facility are analyzed. The experimental facility simulates the high Reynolds number flow conditions similar to those encountered in the space shuttle main engine. In order to determine the influence of Reynolds number on aerodynamic and thermal behavior of the blades, heat transfer coefficients were measured at various Reynolds numbers using liquid crystal temperature measurement technique. Potential flow calculation methods were used to predict the cascade pressure distributions. Boundary layer and heat transfer calculation methods were used with these pressure distributions to verify the experimental results.


Sign in / Sign up

Export Citation Format

Share Document