Experimental and Theoretical Investigations of Heat Transfer in Closed Gas-Filled Rotating Annuli

1995 ◽  
Vol 117 (1) ◽  
pp. 175-183 ◽  
Author(s):  
D. Bohn ◽  
E. Deuker ◽  
R. Emunds ◽  
V. Gorzelitz

The prediction of the temperature distribution in a gas turbine rotor containing closed, gas-filled cavities, for example in between two disks, has to account for the heat transfer conditions encountered inside these cavities. In an entirely closed annulus, forced convection is not present, but a strong natural convection flow exists, induced by a nonuniform density distribution in the centrifugal force field. Experimental investigations have been made to analyze the convective heat transfer in closed, gas-filled annuli rotating around their horizontal axes. The experimental setup is designed to establish a pure centripetal heat flux inside these annular cavities (hot outer, and cold inner cylindrical wall, thermally insulated side walls). The experimental investigations have been carried out for several geometries varying the Rayleigh number in a range usually encountered in cavities of turbine rotors (107 < Ra < 1012). The convective heat flux induced for Ra =1012 was found to be a hundred times larger compared to the only conductive heat flux. By inserting radial walls the annulus is divided into 45 deg sections and the heat transfer increases considerably. A computer program to simulate flow and heat transfer in closed rotating cavities has been developed and tested successfully for annuli with isothermal side walls with different temperatures giving an axial heat flux. For the centripetal heat flux configuration, three-dimensional steady-state calculations of the sectored annulus were found to be consistent with the experimental results. Nevertheless, analysis of unsteady calculations show that the flow can become unstable. This is analogous to the Be´nard problem in the gravitational field.

Author(s):  
D. Bohn ◽  
E. Deuker ◽  
R. Emunds ◽  
V. Gorzelitz

The prediction of the temperature distribution in a gas turbine rotor containing closed, gas-filled cavities, for example in between two discs, has to account for the heat transfer conditions encountered inside these cavities. In an entirely closed annulus forced convection is not present, but a strong natural convection flow exists, induced by a non-uniform density distribution in the centrifugal force field. Experimental investigations have been made to analyze the convective heat transfer in closed, gas-filled annuli rotating around their horizontal axis. The experimental set-up is designed to establish a pure centripedal heat flux inside these annular cavities (hot outer, and cold inner cylindrical wall, thermally insulated side walls). The experimental investigations have been carried out for several geometries varying the Rayleigh number in a range usually encountered in cavities of turbine rotors (1007 < Ra < 1012). The convective heat flux induced for Ra = 1012 was found to be a hundred times larger compared to the only conductive heat flux. By inserting radial walls the annulus is divided into 45° sections and the heat transfer increases considerably. A computer programme to simulate flow and heat transfer in closed rotating cavities has been developed and tested successfully for annuli with isotherm side walls with different temperatures giving an axial heat flux. For the centripedal heat flux configuration, three-dimensional steady state calculations of the sectored annulus were found to be consistent with the experimental results. Nevertheless, analysis of unsteady calculations show that the flow can become unstable. This is analogous to the Bénard problem in the gravitational field.


Author(s):  
Patrick H. Oosthuizen ◽  
Jane T. Paul

Two-dimensional natural convective heat transfer from vertical plates has been extensively studied. However, when the width of the plate is relatively small compared to its height, the heat transfer rate can be greater than that predicted by these two-dimensional flow results. Because situations that can be approximately modelled as narrow vertical plates occur in a number of practical situations, there exists a need to be able to predict heat transfer rates from such narrow plates. Attention has here been given to a plate with a uniform surface heat flux. The magnitude of the edge effects will, in general, depend on the boundary conditions existing near the edge of the plate. To examine this effect, two situations have been considered. In one, the heated plate is imbedded in a large plane adiabatic surface, the surfaces of the heated plane and the adiabatic surface being in the same plane while in the second there are plane adiabatic surfaces above and below the heated plate but the edge of the plate is directly exposed to the surrounding fluid. The flow has been assumed to be steady and laminar and it has been assumed that the fluid properties are constant except for the density change with temperature which gives rise to the buoyancy forces, this having been treated by using the Boussinesq approach. It has also been assumed that the flow is symmetrical about the vertical centre-plane of the plate. The solution has been obtained by numerically solving the full three-dimensional form of the governing equations, these equations being written in terms of dimensionless variables. Results have only been obtained for a Prandtl number of 0.7. A wide range of the other governing parameters have been considered for both edge situations and the conditions under which three dimensional flow effects can be neglected have been deduced.


Author(s):  
Patrick H. Oosthuizen

Most studies of convective heat transfer in window-blind systems assume that the flow over the window-blind arrangement is two-dimensional. In some cases, however, three-dimensional flow effects can become important. The present study was undertaken to determine how significant such effects can be for the particular case of a window covered by a simple plane blind. Only convective heat transfer has been considered. The situation considered is only an approximate model of the real window-blind situation. The window is represented by a rectangular vertical isothermal wall section embedded in a large vertical adiabatic plane wall surface and exposed to a large surrounding "room" in which the temperature is lower than the window temperature. The plane blind is represented by a thin vertical wall having the same size as the "window" which offers no resistance to heat transfer across it and in which conductive heat transfer is negligible. The gaps between the blind and the window at the sides and at the top of the window-blind system are assumed to be open. The flow has been assumed to be laminar and it has been assumed that the fluid properties are constant except for the density change with temperature which gives rise to the buoyancy forces. The solution has been obtained by numerically solving the three-dimensional governing equations written in dimensionless form. The effects of the dimensionless governing variables on the window Nusselt number have been numerically examined.


2016 ◽  
Vol 138 (11) ◽  
Author(s):  
Nathan J. Greiner ◽  
Marc D. Polanka ◽  
James L. Rutledge ◽  
Andrew T. Shewhart

Convective heat transfer from a fluid to a surface is an approximately linear function of driving temperature if the properties within the boundary layer are approximately constant. However, in environments with large driving temperatures like those seen in the hot sections of gas turbine engines, significant property variations exist within the boundary layer. In addition, radiative heat transfer can be a significant contributor to the total heat transfer in a high-temperature environment such that it can not be neglected. As a result, heat transfer to the surface becomes a nonlinear function of driving temperature and the conventional linear heat flux assumption cannot be employed to characterize the convective heat transfer. The present study experimentally examines the nonlinearity of convective heat flux on a zero-pressure-gradient flat plate with large freestream to wall-temperature differences. In addition, the need to account for the radiative component of the overall heat transfer is highlighted. Finally, a method to account for the effects of both variable properties and radiation simultaneously is proposed and demonstrated. Overall, the proposed technique provides the means to quantify the independent contributions of radiative and variable property convective heat transfer to the total conductive heat transfer to or from a surface in a single experiment.


1996 ◽  
Vol 118 (1) ◽  
pp. 11-19 ◽  
Author(s):  
D. Bohn ◽  
R. Emunds ◽  
V. Gorzelitz ◽  
U. Kru¨ger

Increasing the thermal efficiency by higher turbine inlet temperatures is one of the most important aims in the area of gas turbine development. Because of the high temperatures, the turbine vanes and blades have to be cooled, and also knowledge of the mechanically and thermally stressed parts in the hottest zones of the rotor is of great interest. The prediction of the temperature distribution in a gas turbine rotor containing closed, gas-filled cavities, for example, in between two disks, has to account for the heat transfer conditions encountered in these cavities. In an entirely closed annulus, forced convection is not present, but a strong natural convection flow exists, induced by a nonuniform density distribution in the centrifugal force field. In Bohn et al. (1994), experimental and numerical investigations on rotating cavities with pure centripetal heat flux had been carried out. The present paper deals with investigations on a pure axially directed heat flux. An experimental setup was designed to realize a wide range of Ra numbers (2·108< Ra < 5·1010) usually encountered in cavities of gas turbine rotors. Parallel to the experiments, numerical calculations have been conducted. The numerical results are compared with the experimental data. The numerical scheme is also used to account for the influence of Re on heat transfer without changing Ra. This influence could not be pointed out by experiments, because a variation of the Re–Ra characteristic of the employed annuli was not possible. It was found that the numerical and experimental data are in quite good agreement, with exception of high Ra, where the numerical scheme predicts higher heat transfer than the experiments show. One reason may be that in the experiments the inner and outer cylindrical walls were not really adiabatic, an assumption used in the numerical procedure. Moreover, the assumption of a two-dimensional flow pattern may become invalid for high Ra. The influence of three-dimensional effects was studied with the three-dimensional version of the numerical code. In contrast to the radial directed heat transfer, it was found that Nu is much smaller and depends strongly on Re, whereas the radial heat transfer is only weakly influenced by Re.


2015 ◽  
Vol 138 (1) ◽  
Author(s):  
Sergio Lavagnoli ◽  
Cis De Maesschalck ◽  
Guillermo Paniagua

Turbine rotor tips and casings are vulnerable to mechanical failures due to the extreme thermal loads they undergo during engine service. In addition to the heat flux variations during the engine transient operation, periodic unsteadiness occurs at every rotor passage, with amplitude dependent on the tip gap. The development of appropriate predictive tools and cooling schemes requires the precise understanding of the heat transfer mechanisms. The present paper analyses the nature of the overtip flow in transonic turbine rotors running at tight clearances and explores a methodology to determine the relevant flow parameters that model the heat transfer. Steady-state three-dimensional Reynolds-averaged Navier–Stokes (RANS) calculations were performed to simulate engine-like conditions considering two rotor tip gaps, 0.1% and 1%, of the blade span. At tight tip clearance, the adiabatic wall temperature is no longer independent of the solid thermal boundary conditions. The adiabatic wall temperature predicted with the linear Newton's cooling law was observed to rise to unphysical levels in certain regions within the rotor tip gap, resulting in unreliable convective heat transfer coefficients (HTCs). This paper investigates different approaches to estimate the relevant flow parameters that drive the heat transfer. A novel four-coefficient nonlinear cooling law is proposed to model the effects of temperature-dependent gas properties and of the heat transfer history. The four-parameter correlation provided reliable estimates of the convective heat transfer descriptors for the 1% tip clearance case, but failed to model the tip heat transfer of the 0.1% tip gap rotor. The present study allows experimentalists to retrieve information on the gap flow temperature and convective HTC based on the use of wall heat flux measurements. The use of nonlinear cooling laws is sought to improve the evaluation of the rotor heat transfer data while enhancing the understanding of tight-clearance overtip flows.


2020 ◽  
Vol 0 (0) ◽  
Author(s):  
Yijie Zhang ◽  
Juhong Jia

AbstractThis article aimed to study the characteristics and mechanisms of 3D heat transfer through clothing involving the air gap. A three-dimensional finite volume method is used to obtain the coupled conductive, convective, and radiative heat transfer in a body-air-cloth microclimate system. The flow contours and characteristics of temperature, heat flux, and velocity have been obtained. The reason for the high flux and temperature regions was analyzed. Computational results show that the coupled effect of the air gap and the airflow between the skin and garment strongly influences the temperature and heat flux distribution. There are several high-temperature regions on the clothing and high heat flux regions on the body skin because the conductive heat flux can cross through the narrow air gap and reach the cloth surface easily. The high-speed cooling airflow brings about high forced convective heat flux, which will result in the temperature increase on the upper cloth surface. The radiative heat flux has a strong correlation with the temperature gradient between the body and clothing. But its proportion in the total heat flux is relatively small.


Sign in / Sign up

Export Citation Format

Share Document