Characterization of Turbulent Heat Transfer in Ribbed Pipe Flow

2016 ◽  
Vol 138 (4) ◽  
Author(s):  
Changwoo Kang ◽  
Kyung-Soo Yang

In the current investigation, we performed large eddy simulation (LES) of turbulent heat transfer in circular ribbed-pipe flow in order to study the effects of periodically mounted square ribs on heat transfer characteristics. The ribs were implemented on a cylindrical coordinate system by using an immersed boundary method, and dynamic subgrid-scale models were used to model Reynolds stresses and turbulent heat flux terms. A constant and uniform wall heat flux was imposed on all the solid boundaries. The Reynolds number (Re) based on the bulk velocity and pipe diameter is 24,000, and Prandtl number is fixed at Pr = 0.71. The blockage ratio (BR) based on the pipe diameter and rib height is fixed with 0.0625, while the pitch ratio based on the rib interval and rib height is varied with 2, 4, 6, 8, 10, and 18. Since the pitch ratio is the key parameter that can change flow topology, we focus on its effects on the characteristics of turbulent heat transfer. Mean flow and temperature fields are presented in the form of streamlines and contours. How the surface roughness, manifested by the wall-mounted ribs, affects the mean streamwise-velocity profile was investigated by comparing the roughness function. Local heat transfer distributions between two neighboring ribs were obtained for the pitch ratios under consideration. The flow structures related to heat transfer enhancement were identified. Friction factors and mean heat transfer enhancement factors were calculated from the mean flow and temperature fields, respectively. Furthermore, the friction and heat-transfer correlations currently available in the literature for turbulent pipe flow with surface roughness were revisited and evaluated with the LES data. A simple Nusselt number correlation is also proposed for turbulent heat transfer in ribbed pipe flow.

2015 ◽  
Vol 138 (1) ◽  
Author(s):  
Changwoo Kang ◽  
Kyung-Soo Yang

In the present investigation, turbulent heat transfer in fully developed curved-pipe flow has been studied by using large eddy simulation (LES). We consider a fully developed turbulent curved-pipe flow with axially uniform wall heat flux. The friction Reynolds number under consideration is Reτ  = 1000 based on the mean friction velocity and the pipe radius, and the Prandtl number (Pr) is 0.71. To investigate the effects of wall curvature on turbulent flow and heat transfer, we varied the nondimensionalized curvature (δ) from 0.01 to 0.1. Dynamic subgrid-scale models for turbulent subgrid-scale stresses and heat fluxes were employed to close the governing equations. To elucidate the secondary flow structures due to the pipe curvature and their effect on the heat transfer, the mean quantities and various turbulence statistics of the flow and temperature fields are presented, and compared with those of the straight-pipe flow. The friction factor and the mean Nusselt number computed in the present study are in good agreement with the experimental results currently available in the literature. We also present turbulence intensities, skewness and flatness factors of temperature fluctuations, and cross-correlations of velocity and temperature fluctuations. In addition, we report the results of an octant analysis to clarify the correlation between near-wall turbulence structures and temperature fluctuation in the vicinity of the pipe wall. Based on our results, we attempt to clarify the effects of the pipe curvature on turbulent heat transfer. Our LES provides researchers and engineers with useful data to understand the heat-transfer mechanisms in turbulent curved-pipe flow, which has numerous applications in engineering.


1957 ◽  
Vol 35 (4) ◽  
pp. 410-434
Author(s):  
A. W. Marris

A vorticity transfer analogy theory of turbulent heat transfer is developed first for the case of fully developed turbulent flow under zero transverse pressure and temperature gradients such as that in the annulus between concentric cylinders rotating with different angular velocities or in a "free vortex". The mean flow is assumed to be two-dimensional. The theory, which requires that the turbulence be statistically isotropic, yields a temperature distribution in agreement with experiment except in narrow regions immediately adjacent to the boundaries. An argument is given to show that the boundary layer thickness should be of the order of the reciprocal of the square root of the mean velocity, these boundaries are introduced, and Nusselt moduli are defined and their dependence on Reynolds and Prandtl numbers is investigated.The temperature distributions for the case of non-zero transverse temperature and pressure gradients, i.e. for the case of flow in a curved channel in which the fluid does not flow back into itself, are then obtained and the applicability of the simpler equations for zero transverse gradients to this case is investigated.


2017 ◽  
Vol 139 (7) ◽  
Author(s):  
Changwoo Kang ◽  
Kyung-Soo Yang

The present study aims at explaining why heat transfer is enhanced in turbulent ribbed-pipe flow, based on our previous large eddy simulation (LES) database (Kang and Yang, 2016, “Characterization of Turbulent Heat Transfer in Ribbed Pipe Flow,” ASME J. Heat Transfer, 138(4), p. 041901) obtained for Re = 24,000, Pr = 0.71, pitch ratio (PR) = 2, 4, 6, 8, 10, and 18, and blockage ratio (BR) = 0.0625. Here, the bulk velocity and the pipe diameter were used as the velocity and length scales, respectively. The ribs were implemented in the cylindrical coordinate system by means of an immersed boundary method. In particular, we focus on the cases of PR ≥ 4 for which heat transfer turns out to be significantly enhanced. Instantaneous flow fields reveal that the vortices shed from the ribs are entrained into the main recirculating region behind the ribs, inducing velocity fluctuations in the vicinity of the pipe wall. In order to identify the turbulence structures responsible for heat transfer enhancement in turbulent ribbed-pipe flow, various correlations among the fluctuations of temperature and velocity components have been computed and analyzed. The cross-correlation coefficient and joint probability density distributions of velocity and temperature fluctuations, obtained for PR = 10, confirm that temperature fluctuation is highly correlated with velocity-component fluctuation, but which component depends upon the axial location of interest between two neighboring ribs. Furthermore, it was found via the octant analysis performed for the same PR that at the axial point of the maximum heat transfer rate, O3 (cold wallward interaction) and O5 (hot outward interaction) events most contribute to turbulent heat flux and most frequently occur.


2006 ◽  
Vol 129 (4) ◽  
pp. 425-433 ◽  
Author(s):  
B. A. Younis ◽  
B. Weigand ◽  
S. Spring

Fourier’s law, which forms the basis of most engineering prediction methods for the turbulent heat fluxes, is known to fail badly in capturing the effects of streamline curvature on the rate of heat transfer in turbulent shear flows. In this paper, an alternative model, which is both algebraic and explicit in the turbulent heat fluxes and which has been formulated from tensor-representation theory, is presented, and its applicability is extended by incorporating the effects of a wall on the turbulent heat transfer processes in its vicinity. The model’s equations for flows with curvature in the plane of the mean shear are derived and calculations are performed for a heated turbulent boundary layer, which develops over a flat plate before encountering a short region of high convex curvature. The results show that the new model accurately predicts the significant reduction in the wall heat transfer rates wrought by the stabilizing-curvature effects, in sharp contrast to the conventional model predictions, which are shown to seriously underestimate the same effects. Comparisons are also made with results from a complete heat-flux transport model, which involves the solution of differential transport equations for each component of the heat-flux tensor. Downstream of the bend, where the perturbed boundary layer recovers on a flat wall, the comparisons show that the algebraic model yields indistinguishable predictions from those obtained with the differential model in regions where the mean-strain field is in rapid evolution and the turbulence processes are far removed from local equilibrium.


1977 ◽  
Vol 99 (1) ◽  
pp. 12-19 ◽  
Author(s):  
T. M. Kuzay ◽  
C. J. Scott

Experimental investigations of turbulent heat transfer are made in a large-gap annulus with both rotating and nonrotating inner cylinder. The vertical annular channel has an electrically heated outer wall; the inner wall is thermally and electrically insulated. The axial air flow is allowed to develop before rotation and heating are imparted. The resulting temperature fields are investigated using thermocouple probes located near the channel exit. The wall heat flux, wall axial temperature development, and radial temperature profiles are measured. For each axial Reynolds number, three heat flux rates are used. Excellent correlation is established between rotational and nonrotational Nusselt number. The proper correlation parameter is a physical quantity characterizing the flow helix. This parameter is the inverse, of the ratio of axial travel of the flow helix in terms of hydraulic diameter, per half revolution of the spinning wall.


2013 ◽  
Vol 65 (3) ◽  
Author(s):  
Chiranth Srinivasan ◽  
Dimitrios V. Papavassiliou

This work serves a two-fold purpose of briefly reviewing the currently existing literature on the scaling of thermal turbulent fields and, in addition, proposing a new scaling framework and testing its applicability. An extensive set of turbulent scalar transport data for turbulent flow in infinitely long channels is obtained using a Lagrangian scalar tracking approach combined with direct numerical simulation of turbulent flow. Two cases of Poiseuille channel flow, with friction Reynolds numbers 150 and 300, and different types of fluids with Prandtl number ranging from 0.7 to 50,000 are studied. Based on analysis of this database, it is argued that the value and the location of the maximum normal turbulent heat flux are important scaling parameters in turbulent heat transfer. Implementing such scaling on the mean temperature profile for different fluids and Reynolds number cases shows a collapse of the mean temperature profiles onto a single universal profile in the near wall region of the channel. In addition, the profiles of normal turbulent heat flux and the root mean square of the temperature fluctuations appear to collapse on one profile, respectively. The maximum normal turbulent heat flux is thus established as a turbulence thermal scaling parameter for both mean and fluctuating temperature statistics.


Sign in / Sign up

Export Citation Format

Share Document