THERMAL MODELLING OF HEAT TRANSFER IN A ROTATING DISC DUE TO A SUPERPOSED FLOW OF COOLING AIR

Equipment ◽  
2006 ◽  
Author(s):  
D. Zhang ◽  
S. Varah ◽  
G. D. Lock ◽  
M. Wilson
Author(s):  
G. H. Dibelius ◽  
M. Heinen

The local heat transfer from a plane rotating disc enclosed in a casing has been studied experimentally. The disc of 800 mm diameter can be run up to 2000 min−1 at axial distances between disc and casing varied between 5 and 55 mm. Centrifugal or alternatively centripetal flow of cooling air at rates up to ṁ = 1 kg/s can be applied, both with or without an inlet swirl. With the disc rotating in a closed casing (ṁ = 0 kg/s) the influence of the characteristic dimensionless groups on the local heat transfer has been investigated. At a fixed radius, a variation of the local Reynolds Number by either speed or density results in corresponding changes of the heat transfer. However, with a variation of the radius different heat transfer-Re relations are found. In fact, the temperature distribution in the gas caused by the heat flux results in an additional influence of free convection, to be expressed by a Grashof Number. This is confirmed by a comparison of the experimental results with calculations based on Reynolds Analogy and measured friction coefficients. The discrepancies found can be explained only, if in addition to the limitations of the analogy, the influence of free convection is taken into account. Additional results of ongoing experiments concerning the influence of the geometry of the cavity between disc and casing, of the coolant flow rate and of the swirl are presented.


Author(s):  
Abdul A. Jaafar ◽  
Fariborz Motallebi ◽  
Michael Wilson ◽  
J. Michael Owen

In this paper, new experimental results are presented for the flow in a co-rotating disc system with a rotating inner cylinder and a stationary stepped outer casing. The configuration is based on a turbine disc-cooling system used in a gas turbine engine. One of the rotating discs can be heated, and cooling air is introduced through discrete holes angled inward at the periphery of this disc. The cooling air leaves the system through axial clearances between the discs and the outer casing. Some features of computed flows, and both measured and computed heat transfer, were reported previously for this system. New velocity measurements, obtained using Laser Doppler Anemometry, are compared with results from axisymmetric, steady, turbulent flow computations obtained using a low-Reynolds-number k-ε turbulence model. The measurements and computations show that the tangential component of velocity is invariant with axial location in much of the cavity, and the data suggest that Rankine (combined free and forced) vortex flow occurs. The computations fail to reproduce this behaviour, and there are differences between measured and computed details of secondary flow recirculations. Possible reasons for these discrepancies, and their importance for the prediction of associated heat transfer, are discussed.


Author(s):  
Robert Pilbrow ◽  
Hasan Karabay ◽  
Michael Wilson ◽  
J. Michael Owen

In most gas turbines, blade-cooling air is supplied from stationary pre-swirl nozzles that swirl the air in the direction of rotation of the turbine disc. In the “cover-plate” system, the pre-swirl nozzles are located radially inward of the blade-cooling holes in the disc, and the swirling air flows radially outwards in the cavity between the disc and a cover-plate attached to it. In this combined computational and experimental paper, an axisymmetric elliptic solver, incorporating the Launder-Sharma and the Morse low-Reynolds-number k-ε turbulence models, is used to compute the flow and heat transfer. The computed Nusselt numbers for the heated “turbine disc” are compared with measured values obtained from a rotating-disc rig. Comparisons are presented, for a wide range of coolant flow rates, for rotational Reynolds numbers in the range 0.5 × 106 to 1.5 × 106, and for 0.9 < βp < 3.1, where βp is the pre-swirl ratio (or ratio of the tangential component of velocity of the cooling air at inlet to the system to that of the disc). Agreement between the computed and measured Nusselt numbers is reasonably good, particularly at the larger Reynolds numbers. A simplified numerical simulation is also conducted to show the effect of the swirl ratio and the other flow parameters on the flow and heat transfer in the cover-plate system.


Author(s):  
Hasan Karabay ◽  
Robert Pilbrow ◽  
Michael Wilson ◽  
J. Michael Owen

This paper summarises and extends recent theoretical, computational and experimental research into the fluid mechanics, thermodynamics and heat transfer characteristics of the so-called cover-plate pre-swirl system. Experiments were carried out in a purpose-built rotating-disc rig, and the Reynolds-averaged Navier-Stokes equations were solved using 2D (axisymmetric) and 3D computational codes, both of which incorporated low-Reynolds-number k-ε turbulence models. The free-vortex flow, which occurs inside the rotating cavity between the disc and cover-plate, is controlled principally by the pre-swirl ratio, βp: this is the ratio of the tangential velocity of the air leaving the nozzles to that of the rotating disc. Computed values of the tangential velocity are in good agreement with measurements, and computed distributions of pressure are in close agreement with those predicted by a one-dimensional theoretical model. It is shown theoretically and computationally that there is a critical pre-swirl ratio, βp,crit, for which the frictional moment on the rotating discs is zero, and there is an optimal pre-swirl ratio, βp,opt, where the average Nusselt number is a minimum. Computations show that, for βp < βp,opt, the temperature of the blade-cooling air decreases as βp increases; for βp > βp,opt, whether the temperature of the cooling air increases or decreases as βp increases depends on the flow conditions and on the temperature difference between the disc and the air. Owing to the three-dimensional flow and heat transfer near the blade-cooling holes, and to unquantifiable uncertainties in the experimental measurements, there were significant differences between the computed and measured temperatures of the blade-cooling air. In the main, the 3D computations produced smaller differences than the 2D computations.


2000 ◽  
Vol 122 (3) ◽  
pp. 442-450 ◽  
Author(s):  
Hasan Karabay ◽  
Robert Pilbrow ◽  
Michael Wilson ◽  
J. Michael Owen

This paper summarizes and extends recent theoretical, computational, and experimental research into the fluid mechanics, thermodynamics, and heat transfer characteristics of the so-called cover-plate pre-swirl system. Experiments were carried out in a purpose-built rotating-disc rig, and the Reynolds-averaged Navier-Stokes equations were solved using two-dimensional (axisymmetric) and three-dimensional computational codes, both of which incorporated low-Reynolds-number k-ε turbulence models. The free-vortex flow, which occurs inside the rotating cavity between the disc and cover-plate, is controlled principally by the pre-swirl ratio, βp: this is the ratio of the tangential velocity of the air leaving the nozzles to that of the rotating disc. Computed values of the tangential velocity are in good agreement with measurements, and computed distributions of pressure are in close agreement with those predicted by a one-dimensional theoretical model. It is shown theoretically and computationally that there is a critical pre-swirl ratio, βp,crit, for which the frictional moment on the rotating discs is zero, and there is an optimal pre-swirl ratio, βp,opt, where the average Nusselt number is a minimum. Computations show that, for βp<βp,opt, the temperature of the blade-cooling air decreases as βp increases; for βp>βp,opt, whether the temperature of the cooling air increases or decreases as βp increases depends on the flow conditions and on the temperature difference between the disc and the air. Owing to the three-dimensional flow and heat transfer near the blade-cooling holes, and to unquantifiable uncertainties in the experimental measurements, there were significant differences between the computed and measured temperatures of the blade-cooling air. In the main, the three-dimensional computations produced smaller differences than the two-dimensional computations. [S0742-4795(00)01902-5]


Author(s):  
Iraj Mirzaee ◽  
Paul Quinn ◽  
Michael Wilson ◽  
J. Michael Owen

In the system considered here, corotating “turbine” discs are cooled by air supplied at the periphery of the system. The system comprises two corotating discs, connected by a rotating cylindrical hub and shrouded by a stepped, stationary cylindrical outer casing. Cooling air enters the system through holes in the periphery of one disc, and leaves through the clearances between the outer casing and the discs. The paper describes a combined computational and experimental study of the heat transfer in the above system. In the experiments, one rotating disc is heated, the hub and outer casing are insulated, and the other disc is quasi-adiabatic. Thermocouples and fluxmeters attached to the heated disc enable the Nusselt numbers, Nu, to be determined for a wide range of rotational speeds and coolant flow rates. Computations are carried out using an axisymmetric elliptic solver incorporating the Launder-Sharma low-Reynolds-number k-ε turbulence model. The flow structure is shown to be complex and depends strongly on the so-called turbulent flow parameter, λT, which incorporates both rotational speed and flow rate. For a given value of λT, the computations show that Nu increases as Reϕ, the rotational Reynolds number, increases. Despite the complexity of the flow, the agreement between the computed and measured Nusselt numbers is reasonably good.


2017 ◽  
Vol 48 (3) ◽  
pp. 263-281 ◽  
Author(s):  
Meiqi Wang ◽  
Bin Liu ◽  
Yan Wen ◽  
Haoran Liu

Author(s):  
Gaowen Liu ◽  
Zhao Lei ◽  
Aqiang Lin ◽  
Qing Feng ◽  
Yan Chen

The pre-swirl system is of great importance for temperature drop and cooling air supply. This study aims to investigate the influencing mechanism of heat transfer, nonuniform thermodynamic characteristics, and cooling air supply sensitivity in a pre-swirl system by the application of the flow control method of the pre-swirl nozzle. A novel test rig was proposed to actively control the supplied cooling air mass flow rate by three adjustable pre-swirl nozzles. Then, the transient problem of the pre-swirl system was numerically conducted by comparison with 60°, 120°, and 180° rotating disk cavity cases, which were verified with the experiment results. Results show that the partial nozzle closure will aggravate the fluctuation of air supply mass flow rate and temperature. When three parts of nozzles are closed evenly at 120° in the circumferential direction, the maximum value of the nonuniformity coefficient of air supply mass flow rate changes to 3.1% and that of temperature changes to 0.25%. When six parts of nozzles are closed evenly at 60° in the circumferential direction, the maximum nonuniformity coefficient of air supply mass flow rate changes to 1.4% and that of temperature changes to 0.20%. However, different partial nozzle closure modes have little effect on the average air supply parameters. Closing 14.3% of the nozzle area will reduce the air supply mass flow rate by 9.9% and the average air supply temperature by about 1 K.


Sign in / Sign up

Export Citation Format

Share Document