You are here

Heat Transfer, Friction, and Turbulent Analysis on Single Ribbed-Wall Square Channel

Download pdf | Full Screen View

Date Issued:
2017
Abstract/Description:
An experimental investigation of heat transfer and friction behavior for a fully developed flow in a non-rotating square channel was conducted under a wide range of Reynolds numbers from 6,000 to 180,000. The rig used in this study was a single ribbed wall variant of Ahmed et al.'s [ 1 ] rig from which results of this rig were compared. Ahmed et al.'s rig was a replica of Han et al.'s square channel [ 2 ] used to validate their work, and expand the Reynolds number range for both heat transfer and friction data. The test section was 22 hydraulic diameters (Dh) long, and made of four aluminum plates. One rib roughened bottom wall, and three smooth walls bounded the flow. Glued brass ribs oriented at 45(&)deg; to the flow direction, with a ratio of rib height to channel hydraulic diameter (e/Dh) and a ratio of pitch to rib height (p/e) of 0.063 and 10, respectively, lined the bottom wall. A 20Dh long acrylic channel with a continuation of the test section's interior was attached at the inlet of the test section to confirm the fully developed flow. Heat transfer tests were conducted in a Reynolds number range of 20,000 to 150,000. During these tests, the four walls were held under isothermal conditions. Wall-averaged, and module-averaged Nusselt values were calculated from the log-mean temperature differences between the plate surface temperature and calculated, by energy balance, fluid bulk temperature. Streamwise Nusselt values become constant at an x/Dh of 8 within the tested Reynolds number range. Wall averaged Nusselt values were determined after x/Dh=8, and scaled by the Dittus-Boelter correlation, Nuo, for smooth ducts to yield a Nusselt augmentation value (Nu/Nuo). Non-heated friction tests were conducted from a Reynolds number range of 6,000 to 180,000. Pressure drop along the channel was recorded, and channel-averaged Darcy-Weisbach friction factor was calculated within the range of Reynolds number tested. Scaling the friction factor by the smooth-wall Blasius correlation, fo, gave the friction augmentation (f/fo). The thermal performance, a modified ratio of the Nusselt and friction augmentation used by Han et al. [ 2 ], was then calculated to evaluate the bottom-line performance of the rig. It was found that the Nusselt augmentation approached a constant value of 1.4 after a Reynolds number of 60,000 while friction augmentation continued to increase in a linear fashion past that point. This caused the overall thermal performance to decline as Reynolds number increased up to a certain point. Further studies were conducted in an all acrylic, non-heated variant of the rig to study the fluid flow in the streamwise direction on, and between two ribs in the fully developed region of the channel. Single-wire hot-wire anemometry characterized velocity magnitude profiles with great detail, as well as turbulence intensity for Reynolds numbers ranging from 5,000 to 50,000. As the Reynolds number increased the reattachment point between two ribs remained about stationary while the turbulence intensity receded to the trailing surface of the upstream rib, and dissipated as it traveled. At low Reynolds numbers, between 5,000 and 10,000, the velocity and turbulence intensity streamwise profiles seemed to form two distinct flow regions, indicating that the flow over the upstream rib never completely attached between the two ribs. Integral length-scales were also derived from the autocorrelation function using the most turbulent signal acquired at each Reynolds number. It was found that there is a linear trend between Reynolds number and the integral length-scale at the most turbulent points in the flow. For example, at Re=50,000 the most the length scale found just past the first rib was on the order of two times the height of the rib. Rivir et al. [ 30 ] found in a similar case that at Re = 45,000, it was 1.5 times the rib height. Several factors could influence the value of this integral length-scale, but the fact that their scale is on the order of what was obtained in this case gives some level of confidence in the value.
Title: Heat Transfer, Friction, and Turbulent Analysis on Single Ribbed-Wall Square Channel.
17 views
8 downloads
Name(s): Vergos, Christopher, Author
Kapat, Jayanta, Committee Chair
Vasu Sumathi, Subith, Committee CoChair
Ahmed, Kareem, Committee CoChair
University of Central Florida, Degree Grantor
Type of Resource: text
Date Issued: 2017
Publisher: University of Central Florida
Language(s): English
Abstract/Description: An experimental investigation of heat transfer and friction behavior for a fully developed flow in a non-rotating square channel was conducted under a wide range of Reynolds numbers from 6,000 to 180,000. The rig used in this study was a single ribbed wall variant of Ahmed et al.'s [ 1 ] rig from which results of this rig were compared. Ahmed et al.'s rig was a replica of Han et al.'s square channel [ 2 ] used to validate their work, and expand the Reynolds number range for both heat transfer and friction data. The test section was 22 hydraulic diameters (Dh) long, and made of four aluminum plates. One rib roughened bottom wall, and three smooth walls bounded the flow. Glued brass ribs oriented at 45(&)deg; to the flow direction, with a ratio of rib height to channel hydraulic diameter (e/Dh) and a ratio of pitch to rib height (p/e) of 0.063 and 10, respectively, lined the bottom wall. A 20Dh long acrylic channel with a continuation of the test section's interior was attached at the inlet of the test section to confirm the fully developed flow. Heat transfer tests were conducted in a Reynolds number range of 20,000 to 150,000. During these tests, the four walls were held under isothermal conditions. Wall-averaged, and module-averaged Nusselt values were calculated from the log-mean temperature differences between the plate surface temperature and calculated, by energy balance, fluid bulk temperature. Streamwise Nusselt values become constant at an x/Dh of 8 within the tested Reynolds number range. Wall averaged Nusselt values were determined after x/Dh=8, and scaled by the Dittus-Boelter correlation, Nuo, for smooth ducts to yield a Nusselt augmentation value (Nu/Nuo). Non-heated friction tests were conducted from a Reynolds number range of 6,000 to 180,000. Pressure drop along the channel was recorded, and channel-averaged Darcy-Weisbach friction factor was calculated within the range of Reynolds number tested. Scaling the friction factor by the smooth-wall Blasius correlation, fo, gave the friction augmentation (f/fo). The thermal performance, a modified ratio of the Nusselt and friction augmentation used by Han et al. [ 2 ], was then calculated to evaluate the bottom-line performance of the rig. It was found that the Nusselt augmentation approached a constant value of 1.4 after a Reynolds number of 60,000 while friction augmentation continued to increase in a linear fashion past that point. This caused the overall thermal performance to decline as Reynolds number increased up to a certain point. Further studies were conducted in an all acrylic, non-heated variant of the rig to study the fluid flow in the streamwise direction on, and between two ribs in the fully developed region of the channel. Single-wire hot-wire anemometry characterized velocity magnitude profiles with great detail, as well as turbulence intensity for Reynolds numbers ranging from 5,000 to 50,000. As the Reynolds number increased the reattachment point between two ribs remained about stationary while the turbulence intensity receded to the trailing surface of the upstream rib, and dissipated as it traveled. At low Reynolds numbers, between 5,000 and 10,000, the velocity and turbulence intensity streamwise profiles seemed to form two distinct flow regions, indicating that the flow over the upstream rib never completely attached between the two ribs. Integral length-scales were also derived from the autocorrelation function using the most turbulent signal acquired at each Reynolds number. It was found that there is a linear trend between Reynolds number and the integral length-scale at the most turbulent points in the flow. For example, at Re=50,000 the most the length scale found just past the first rib was on the order of two times the height of the rib. Rivir et al. [ 30 ] found in a similar case that at Re = 45,000, it was 1.5 times the rib height. Several factors could influence the value of this integral length-scale, but the fact that their scale is on the order of what was obtained in this case gives some level of confidence in the value.
Identifier: CFE0007138 (IID), ucf:52318 (fedora)
Note(s): 2017-08-01
M.S.M.E.
Engineering and Computer Science, Mechanical and Aerospace Engineering
Masters
This record was generated from author submitted information.
Subject(s): Heat Transfer -- Friction Factor -- Length-Scales -- Hotwire -- Ribbed-Channel -- Internal Duct Cooling
Persistent Link to This Record: http://purl.flvc.org/ucf/fd/CFE0007138
Restrictions on Access: public 2018-02-15
Host Institution: UCF

In Collections