High Resolution Heat Transfer Measurements on the Stator Endwall of an Axial Turbine

Author(s):  
Benoit Laveau ◽  
Reza S. Abhari ◽  
Michael E. Crawford ◽  
Ewald Lutum

In order to continue increasing the efficiency of gas turbines, an important effort is made on the thermal management of the turbine stage. In particular understanding and accurately estimating the thermal loads in a vane passage is of primary interest to engine designers looking to optimize the cooling requirements and ensure the integrity of the components. This paper focuses on the measurement of endwall heat transfer in a vane passage with a 3D airfoil shape and cylindrical endwalls. It also presents a comparison with predictions performed using an in-house developed RANS solver featuring a specific treatment of the numerical smoothing using a flow adaptive scheme. The measurements have been performed in a steady state axial turbine facility on a novel platform developed for heat transfer measurements and integrated to the nozzle guide vane row of the turbine. A quasi-isothermal boundary condition is used to obtain both the heat transfer coefficient and the adiabatic wall temperature within a single measurement day. The surface temperature is measured using infrared thermography through small view ports. The infrared camera is mounted on a robot-arm with six degrees of freedom to provide high resolution surface temperature and a full coverage of the vane passage. The paper presents results from experiments with two different flow conditions obtained by varying the mass flow through the turbine: measurements at the design point (ReCax = 7,2.105) and at a reduced mass flow rate (ReCax = 5,2.105). The heat transfer quantities, namely the heat transfer coefficient and the adiabatic wall temperature, are derived from measurements at 14 different isothermal temperatures. The experimental data are supplemented with numerical predictions that are deduced from a set of adiabatic and diabatic simulations. In addition, the predicted flow field in the passage is used to highlight the link between the heat transfer patterns measured and the vortical structures present in the passage.

2014 ◽  
Vol 137 (4) ◽  
Author(s):  
Benoit Laveau ◽  
Reza S. Abhari ◽  
Michael E. Crawford ◽  
Ewald Lutum

In order to continue increasing the efficiency of gas turbines, an important effort is made on the thermal management of the turbine stage. In particular, understanding and accurately estimating the thermal loads in a vane passage is of primary interest to engine designers looking to optimize the cooling requirements and ensure the integrity of the components. This paper focuses on the measurement of endwall heat transfer in a vane passage with a three-dimensional (3D) airfoil shape and cylindrical endwalls. It also presents a comparison with predictions performed using an in-house developed Reynolds-Averaged Navier–Stokes (RANS) solver featuring a specific treatment of the numerical smoothing using a flow adaptive scheme. The measurements have been performed in a steady state axial turbine facility on a novel platform developed for heat transfer measurements and integrated to the nozzle guide vane (NGV) row of the turbine. A quasi-isothermal boundary condition is used to obtain both the heat transfer coefficient and the adiabatic wall temperature within a single measurement day. The surface temperature is measured using infrared thermography through small view ports. The infrared camera is mounted on a robot arm with six degrees of freedom to provide high resolution surface temperature and a full coverage of the vane passage. The paper presents results from experiments with two different flow conditions obtained by varying the mass flow through the turbine: measurements at the design point (ReCax=7.2×105) and at a reduced mass flow rate (ReCax=5.2×105). The heat transfer quantities, namely the heat transfer coefficient and the adiabatic wall temperature, are derived from measurements at 14 different isothermal temperatures. The experimental data are supplemented with numerical predictions that are deduced from a set of adiabatic and diabatic simulations. In addition, the predicted flow field in the passage is used to highlight the link between the heat transfer patterns measured and the vortical structures present in the passage.


1960 ◽  
Vol 11 (3) ◽  
pp. 269-284
Author(s):  
J. S. Przemieniecki

SummaryA set of design charts is presented for the calculation of transient temperature and thermal stress distributions in thermally thick plates subjected to aerodynamic heating.The method is particularly useful for determining temperatures and thermal stresses in plates with an arbitrary variation of the heat transfer coefficient and the adiabatic wall temperature of the boundary layer. The present method is based on repetitive applications of the exact analytical solution to a unit triangular variation of the adiabatic wall temperature and a constant heat transfer coefficient. The actual variation of the adiabatic wall temperature is represented as a series of straight lines while the heat transfer coefficient is approximated by a step function. The temperature distribution through the plate is separated into linear and “self-equilibrating” temperature distributions to facilitate thermal stress calculations; these distributions can be obtained directly from the design charts presented in this paper.The general principle of this semi-numerical method is also applied to thermally thin plates subjected to arbitrary heating conditions.


Author(s):  
X. C. Li ◽  
J. Zhou ◽  
K. Aung

One of the most fundamental concepts in heat transfer is the convective heat transfer coefficient, which is closely related with the flow Reynolds number, flow geometry and the thermal conditions on the heat transfer surface. To define the heat transfer coefficient, a reference temperature is needed besides the surface temperature and heat flux. The reference temperature can be chosen differently, such as the fluid bulk mean temperature (for internal flows) and the temperature at the far field (for external flows). For complicated flows, the adiabatic wall temperature, defined as the wall temperature when the surface heat flux is zero, is commonly adopted as the reference temperature. Other options can also be applied to complicated flows. This paper analyzed some of the potential selections of the reference temperature for different flow settings, including film cooling, jet impingement with cross flows and a mixing flow in a straight duct with or without internal heat source. Both laminar and turbulent flows are considered with different boundary conditions. Dramatic changes of heat transfer coefficient are observed with different reference temperatures. In some special conditions the heat transfer coefficient becomes negative, which means the heat flux has a different direction with the driving temperature difference defined. An innovative method is proposed to calculate the heat transfer coefficient of complicated flows with constant surface temperature.


2003 ◽  
Vol 125 (3) ◽  
pp. 538-546 ◽  
Author(s):  
Andrew C. Chambers ◽  
David R. H. Gillespie ◽  
Peter T. Ireland ◽  
Geoffrey M. Dailey

Transient liquid crystal techniques are widely used for experimental heat transfer measurements. In many instances it is necessary to model the heat transfer resulting from the temperature difference between a mixture of two gas streams and a solid surface. To nondimensionally characterize the heat transfer from scale models it is necessary to know both the heat transfer coefficient and adiabatic wall temperature of the model. Traditional techniques rely on deducing both parameters from a single test. This is a poorly conditioned problem. A novel strategy is proposed in which both parameters are deduced from a well-conditioned three-test strategy. The heat transfer coefficient is first calculated in a single test; the contribution from each driving gas stream is then deduced using additional tests. Analytical techniques are developed to deal with variations in the temperature profile and transient start time of each flow. The technique is applied to the analysis of the heat transfer within a low aspect ratio impingement channel with initial cross flow.


2004 ◽  
Vol 126 (4) ◽  
pp. 597-603 ◽  
Author(s):  
Srinath V. Ekkad ◽  
Shichuan Ou ◽  
Richard B. Rivir

In film cooling situations, there is a need to determine both local adiabatic wall temperature and heat transfer coefficient to fully assess the local heat flux into the surface. Typical film cooling situations are termed three temperature problems where the complex interaction between the jets and mainstream dictates the surface temperature. The coolant temperature is much cooler than the mainstream resulting in a mixed temperature in the film region downstream of injection. An infrared thermography technique using a transient surface temperature acquisition is described which determines both the heat transfer coefficient and film effectiveness (nondimensional adiabatic wall temperature) from a single test. Hot mainstream and cooler air injected through discrete holes are imposed suddenly on an ambient temperature surface and the wall temperature response is captured using infrared thermography. The wall temperature and the known mainstream and coolant temperatures are used to determine the two unknowns (the heat transfer coefficient and film effectiveness) at every point on the test surface. The advantage of this technique over existing techniques is the ability to obtain the information using a single transient test. Transient liquid crystal techniques have been one of the standard techniques for determining h and η for turbine film cooling for several years. Liquid crystal techniques do not account for nonuniform initial model temperatures while the transient IR technique measures the entire initial model distribution. The transient liquid crystal technique is very sensitive to the angle of illumination and view while the IR technique is not. The IR technique is more robust in being able to take measurements over a wider temperature range which improves the accuracy of h and η. The IR requires less intensive calibration than liquid crystal techniques. Results are presented for film cooling downstream of a single hole on a turbine blade leading edge model.


Author(s):  
Andrew C. Chambers ◽  
David R. H. Gillespie ◽  
Peter T. Ireland ◽  
Geoffrey M. Dailey

Transient liquid crystal techniques are widely used for experimental heat transfer measurements. In many instances it is necessary to model the heat transfer resulting from the temperature difference between a mixture of two gas streams and a solid surface. To non-dimensionally characterise the heat transfer from scale models it is necessary to know both the heat transfer coefficient and adiabatic wall temperature of the model. Traditional techniques rely on deducing both parameters from a single test. This is a poorly conditioned problem. A novel strategy is proposed in which both parameters are deduced from a well conditioned three test strategy. The heat transfer coefficient is first calculated in a single test; the contribution from each driving gas stream is then deduced using additional tests. Analytical techniques are developed to deal with variations in the temperature profile and transient start time of each flow. The technique is applied to the analysis of the heat transfer within a low aspect ratio impingement channel with initial cross flow.


Author(s):  
Kenneth W. Van Treuren ◽  
Zuolan Wang ◽  
Peter T. Ireland ◽  
Terry V. Jones ◽  
S. T. Kohler

Recent work, Van Treuren et al. (1993), has shown the transient method of measuring heat transfer under an array of impinging jets allows the determination of local values of adiabatic wall temperature and heat transfer coefficient over the complete surface of the target plate. Using this technique, an inline array of impinging jets has been tested over a range of average jet Reynolds numbers (10,000–40,000) and for three channel height to jet hole diameter ratios (1, 2, and 4). The array is confined on three sides and spent flow is allowed to exit in one direction. Local values are averaged and compared with previously published data in related geometries. The current data for a staggered array is compared to those from an inline array with the same hole diameter and pitch for an average jet Reynolds number of 10,000 and channel height to diameter ratio of one. A comparison is made between intensity and hue techniques for measuring stagnation point and local distributions of heat transfer. The influence of the temperature of the impingement plate through which the coolant gas flows on the target plate heat transfer has been quantified.


1983 ◽  
Vol 105 (4) ◽  
pp. 878-883 ◽  
Author(s):  
A. Haji-Sheikh ◽  
M. Mashena ◽  
M. J. Haji-Sheikh

An analytical method for the numerical calculation of the heat transfer coefficient in arbitrarily shaped ducts with constant wall temperature at the boundary is presented. The flow is considered to be laminar and fully developed, both thermally and hydrodynamically. The method presented herein makes use of Galerkin-type functions for computation of the Nusselt number. This method is applied to circular pipes and ducts with rectangular, isosceles triangular, and right triangular cross sections. A three-term or even a two-term solution yields accurate solutions for circular ducts. The situation is similar for right triangular ducts with two equal sides. However, for narrower ducts, a larger number of terms must be used.


Author(s):  
Robert J. Moffat

In all electronics cooling situations, and many other practical situations, the surface temperature varies in the stream-wise direction. In these cases, defining the heat transfer coefficient using the adiabatic temperature of the surface instead of the mixed mean temperature of the coolant result in significant benefits. The resulting definition is hadiabatic. The theoretical and practical bases for hadiabatic are presented. Examples of its use in electronics cooling are described to show the operational advantages this approach offers. Turbulence strongly affects heat transfer. A simple, turbulence-based correlation is presented that yields an estimate of the heat transfer coefficient good enough for preliminary design estimates and often as accurate as can be relied upon from CFD calculations using present codes.


Sign in / Sign up

Export Citation Format

Share Document