Shrink Oven Analysis

Transcript

ABSTRACT KIRAN, THIRUMARAN. Optimization of Air Flow and Heat Transfer in a Convective Heat Shrink Oven. (Under the direction of Dr. Stephen D. Terry.) The optimization of the flow and heat transfer characteristics inside a convective heat shrink oven poses a complex flow field problem. The plenum blower wheel used to recirculate the air creates a flow field that is highly non-uniform along the length of the oven. This causes uneven heating of the bulk air flow resulting in reduced quality of wrapped product. In the initial part of this thesis the different flow regions that directly affect the performance of the oven are throughly studied by experimental data collection. A computer model of the baseline oven is built with an implicit model of a blower using CFX [ANSYS] and the flow is visualized. The simulation is verified with the experimental results and is used as a tool to analyze design modifications. The experimental data together with computer simulations give us a good understanding of the nature of the flow and its coupling with the heat transfer characteristics inside the oven. This knowledge is used to develop design modifications and retrofits to create a uniform flow and temperature along the length of the oven. Various modifications to the design are developed and their effects on the flow field are predicted by the computer model. The most viable of these options are taken to the experimental stage where they are a built and tested on the oven. Separate design modifications have been built and tested to tackle the various aspects of the oven performance. In this process a CFD model was developed that could give insight into the flow dynamics of the oven and can be used as a tool. In the final part of the thesis a new oven design with cross-flow fan at its core is put forward and its various advantages over a centrifugal blower are discussed. © Copyright 2014 by Thirumaran Kiran All Rights Reserved Optimization of Air Flow and Heat Transfer in a Convective Heat Shrink Oven by Thirumaran Kiran A thesis submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the Degree of Master of Science Mechancial Engineering Raleigh, North Carolina 2014 APPROVED BY: Dr. Alexei V. Saveliev Dr. Herbert M. Eckerlin Dr. Stephen D. Terry Chair of Advisory Committee DEDICATION To all the people and places in my life. ii BIOGRAPHY Kiran Thirumaran was born and raised in Chennai,India. He was a curios kid and became a fan of the scientific method at a very young age. This continued throughout his teens which made him take up engineering. He graduated in 2012 with his Bachelor of Engineering in Aeronautical Engineering from Anna University. In order to further his education and get a global exposure he decided to attend graduate school in Mechanical Engineering at North Carolina State University. There he learned a great deal in his work with the Industrial Assessment Center and was able to obtain invaluable practical knowledge. iii ACKNOWLEDGEMENTS I would first like to thank Mr.Adam Duncan, Design Engineer at Axoncorp for providing me the freedom and necessary assistance to carry out this project. I would also like to thank Prithwish Kundu, doctoral candidate at North Carolina State University for his assistance with computational tools and all my colleague at the Industrial Accessement Center for their continued support throughout the course of the thesis. Finally, and most importantly, I would like to thank Dr. Stephen Terry for giving me the opportunity to work at the Industrial Assessment Center and on this project. The practical knowledge he has imparted has given me the confidence to take on any engineering task. iv TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii CHAPTER 1 Design . . . . . . . . . . . . . . . 1.1 Basic Design . . . . . . . . . . . . . . . 1.2 Tunnel Components . . . . . . . . . 1.2.1 Outer Shell . . . . . . . . . . . 1.2.2 Upper Blower Wheel . . . . 1.2.3 Panel . . . . . . . . . . . . . . . 1.2.4 Lower Blower Wheel . . . . 1.2.5 Inner Shell . . . . . . . . . . . 1.2.6 Heating Element . . . . . . . 1.3 Product Testing . . . . . . . . . . . . . 1.3.1 Dimensional Correctness 1.3.2 Alignment . . . . . . . . . . . 1.3.3 Fan Operation . . . . . . . . 1.3.4 Heating Coils . . . . . . . . . 1.4 Oven Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 3 4 4 4 6 10 13 15 15 15 15 15 16 CHAPTER 2 Shrink Oven Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Analysis Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Experimental Tools . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Computational Tools . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Flow Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Quantitative Analysis of the Jet Stream. . . . . . . . . . . . 2.2.2 Qualitative Analysis of the Flow in the Inner Shell . . . . 2.2.3 Quantitative Analysis at Fan Inlet . . . . . . . . . . . . . . . 2.2.4 Qualitative Analysis of the Flow in the Tunnel Section . 2.3 Heat Transfer Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Heat Transfer Relations . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Heat Transfer Coefficient - Experimental Result . . . . . 2.4 Design Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 18 18 21 24 24 33 40 42 46 47 50 52 CHAPTER 3 Design Modifications . . . . . . . . . . . . . . 3.1 Turbulator . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Inner Shell Re-Design . . . . . . . . . . . . . . . . . . . 3.2.1 Effect on flow Characteristics . . . . . . . 3.2.2 Effect on Heat Transfer Characteristics . 3.3 Plates with Slits . . . . . . . . . . . . . . . . . . . . . . . 3.4 Plenum Redesign . . . . . . . . . . . . . . . . . . . . . . 3.5 Cross-flow Fans . . . . . . . . . . . . . . . . . . . . . . . 3.5.1 Fan Sizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 55 59 60 62 64 65 70 70 v . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 3.5.2 Scope of the Tangential Oven . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 APPENDIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Appendix A Additional Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 vi LIST OF TABLES Table 2.1 Table 2.2 Mesh Properties of Various Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Velocity at the Outlet of each cross section of the Blower Blade . . . . . . . . . . . . . . 36 vii LIST OF FIGURES Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 1.5 Figure 1.6 Figure 1.7 Isometric ,Longitudinal and Transverse section of the Heat Shrink Oven . . . . . . . 2 The Front View of the Heat Shrink Oven with . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Panel with slots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Inlet and Outlet Velocity Vectors for a Backward Curved Centrifugal Blower Blade 7 Inlet and Outlet Velocity triangles for a Backward Curved Centrifugal Blower Blade 8 Inlet and Outlet Velocity triangles for a Backward Curved Centrifugal Blower Blade 10 Heating Coils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Figure 2.7 Figure 2.8 Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 AccuSense F900 Sensors [F90] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermocouple - Type T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 20 Amp split-core AC current sensor - CTV-A [Ct] . . . . . . . . . . . . . . . . . . . . . H22 Energy Logger and FlexSmart Analog Module [Ene] . . . . . . . . . . . . . . . . . . . Temperature distribution along the length of the slots (Left Panel and Right Panel) Velocity distribution at the slots - Left Panel and Right Panel . . . . . . . . . . . . . . . Dependence of Temperature with Mean Velocity at the slots of the Left panel . . . Turbulence fluctuation with time at the start of the front slot (Near Tunnel Entrance) Turbulence fluctuation with time at 2/3r d the length of the front slot . . . . . . . . . Turbulence fluctuation with time at 1/3r d the length of the rear slot . . . . . . . . . . Turbulence fluctuation with time at the end of the rear slot (Tunnel Exit) . . . . . . Dependence of Mean Velocity with turbulence strength along the slots of the Left panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dependence of Temperature with turbulence strength along the slots of the Left panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inner Shell with given boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . Streamlines along the XZ plane at various heights . . . . . . . . . . . . . . . . . . . . . . . Comparison of Computational and Experimental results. . . . . . . . . . . . . . . . . . Flow properties at the Plenum Opening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation Setup of the Tunnel Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flow of Cold ambient air through the Tunnel . . . . . . . . . . . . . . . . . . . . . . . . . . . Tempertaure Distribution at various cross-sections along the length of the tunnel Power Drawn by the heating coil with the oven set at 140 F . . . . . . . . . . . . . . . . . 18 19 20 21 25 26 27 29 29 31 31 Effect of Rib Turbulator on the flow field in the Inner Shell. . . . . . . . . . . . . . . . . Rib Turbulator - Configuration B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effects of Rib turbulators on the temperature of the jet stream . . . . . . . . . . . . . . Redesigned Inner Shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of gradually reducing hydraulic diameter on the flow field in the Inner Shell. Temperature and Turbulence Intensity distribution along the length of the front slot of the Left Panel for the Redesigned Inner Shell . . . . . . . . . . . . . . . . . . . . . . Various Slit Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of the Mean velocity at the slots of the left panel for various slit configuration with the baseline values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 57 58 60 61 Figure 2.13 Figure 2.14 Figure 2.15 Figure 2.16 Figure 2.17 Figure 2.18 Figure 2.19 Figure 2.20 Figure 2.21 Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 viii 34 35 37 39 40 41 44 45 46 51 63 65 66 Figure 3.9 Various Plenum Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.10 Flow of Cold Ambient Air through the Redesigned Tunnel . . . . . . . . . . . . . . . . Figure 3.11 Effect of the new plenum design on the temperature distribution in the tunnel section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.12 Cross-Flow or Tangential Fan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 . 68 Figure A.1 Figure A.2 Figure A.3 Figure A.4 Figure A.5 . . . . . Dependence of Temperature with Mean Velocity at the slots of the Right panel Turbulence fluctuation with time at 1/2 of the length of the front slot . . . . . . . Turbulence fluctuation with time at end of the front slot . . . . . . . . . . . . . . . . Turbulence fluctuation with time at start of the rear slot . . . . . . . . . . . . . . . . Turbulence fluctuation with time at 1/2 of the rear slot . . . . . . . . . . . . . . . . . ix . . . . . . 69 . 71 79 80 80 81 81 CHAPTER 1 Design A Heat Shrink Tunnel is a convective heat oven that uses impinging streams of hot air to shrink-wrap label films around bottles. Axon Corp, a pioneer in manufacturing convection based heat shrink ovens make a product that has been in the market for over 20 years. The product has undergone continuous improvements throughout its lifetime in the market and is time-tested and successful. The company always on the lookout for further optimization opportunities have brought NCState on-board to help with Research and Development. This project is an attempt at that very objective and concentrates on the study of the fluid and thermal characteristics of the oven. The flow properties as well as its interaction with the heating elements are first studied and then matched with the requirements of the oven. The goal is to develop a design for a heat shrink tunnel that is optimized to its functions and provide Axon with a competitive edge in the market. 1.1 Basic Design Axon’s heat shrink tunnels are designed to work with both tamper evident bands and sleeve labels. The basic design consist two concentric shells one inside the other, the air flow through these shells 1 1.1. BASIC DESIGN CHAPTER 1. Design are independent and are driven by a two separate blower wheel that are mounted on the same shaft. The inner shell houses the heating coils that are placed symmetric on either of its side columns. A tunnel section that runs throughout the length of the oven defines the working area through which the product to be shrink wrapped is sent by means of a conveyor, the direction of which defines the entrance and exit of the tunnel. The inner shell supplies the hot air needed for the shrinking process as impinging jet streams discharges into the tunnel section through slots in the panels mounted onto the inner shell. Figure 1.1 Isometric ,Longitudinal and Transverse section of the Heat Shrink Oven 2 1.2. TUNNEL COMPONENTS CHAPTER 1. Design Figure 1.2 The Front View of the Heat Shrink Oven with 1.2 Tunnel Components The tunnel components whose design defines the characteristics of the flow and the heat transfer in the system include, 1. Outer Shell. 2. Inner Shell. 3. Panel. 4. Upper Blower Wheel. 5. Lower Blower Wheel. 6. Heating Coils. 7. Plenum 3 1.2. TUNNEL COMPONENTS CHAPTER 1. Design The Inner shell, lower wheel blower and the heating coils are discussed in detail as they are responsible for the flow in the tunnel section. 1.2.1 Outer Shell The Outer Shell houses the upper blower wheel that cools the electronics on top of it. The outer shell is 53.34 cm in height and 44.45 cm wide and could be 91.4 cm or 61 cm long depending upon the model. The only differences between the two models of the oven is the length of the tunnel section while all the other dimensions and the components remain the same. 1.2.2 Upper Blower Wheel The Upper Blower Wheel draws air at ambient condition from the room through the openings in the electronic housing. The openings are covered with filters that prevent dust from affecting the electronics. The air flows over the electronics, into the fan and is dispersed radially into the columns of the outer shell. The air flows through the columns further cooling the inner shell before being let out through various slots on the front, back and along the sides. The flow through the outer shell never directly affects the flow in the inner shell, Though it indirectly affects the flow by removing energy from it in the form of heat. Thus the flow through the outer shell can be seen to serve two purposes, 1. Remove heat from the electronics. 2. Prevent over heating of the oven. 1.2.3 Panel The Panels are fit onto the Inner Shell on either sides and provide slots for the impinging jet streams. The slot design shown in Figure 1.3 is the most commonly used design and would be the one used for majority of the analysis. There can be two or more pairs of slots in each panel based on the size and shape of the product to be shrink wrapped but most slot designs have an inclined slot that is modified according to the product to be wrapped and horizontal slots at the bottom which are present in most panel design. The most common slot design used in the 91.4 cm model of the tunnel has horizontal slots that run for 80.6 cm while the inclined panels start 20.32 cm from the entrance and run for 62.9 cm. The slots are 1 cm in thickness and thus the slots have a cross sectional area of 143.5c m 2 on each panel. The 4 1.2. TUNNEL COMPONENTS CHAPTER 1. Design mean velocity of the jet stream as measured at the slot is 4.41 m/s (Will be discussed in detail in Chapter 2). Figure 1.3 Panel with slots ˙ ) through the inner shell is calculated to be The mass flow rate (m ˙ = ρAV m (1.1) Where, ρ - Density in k g /m 3 A - Cross sectional area in m 2 V - Velocity of the flow in m /s For the preliminary calculations of the flow the heater is switched off and only the air flow through the various components is studied.Thus the air flow would be at room conditions throughout the oven and the density of air is constant at 1.204 k g /m 3 .Thus the Mass Flow Rate is given by, ˙ m ‹ m  kg × 143.5 × 10−4 (m 2 ) × 4.41 m3 s kg = 0.076 s = 1.204  This is the mass flow rate through all the slots in one panel. 5 1.2. TUNNEL COMPONENTS 1.2.4 CHAPTER 1. Design Lower Blower Wheel Even though the lower and upper blower wheels are driven by the same motor they are independent in terms of the fluid flow they handle. The lower blower wheel draws in air through the openings in the plenum which are at the roof of the tunnel section and moves the air at an angle to the sides of the inner shell. A plate on top of the wheel prevents the air to seep into the upper blower. Using the mass flow rate calculated in the previous section, the dimensions of the fan and the properties on the suction side of the fan, the flow properties at the outlet of the fan is calculate using velocity triangles. From the Conservation of Mass we have, ρAV F a n = 2 [ρAV Sl o t s ] Thus the flow through the fan is 0.152 kg/s. The velocity of the air at the inlet side of the fan blades Vf 1 can be calculated thus, Vf 1 = m F˙a n ρ f a n A F a n I nl e t (1.2) The fan is 20.57 cm in diameter at the inlet side and 10.79 cm in breadth giving a cross sectional area of 697 c m 2 . The air through the slots is mixed with some air at atmospheric conditions and recirculated to the inner shell through the fan. Thus the air temperature at the inlet side would be below the temperature of the jet streams depending on the amount of outside air during the actual operation of the oven. For this calculation however, the heater is turned off and thus the fan inlet is at room conditions. Substituting the values in Eq. 1.2 we get a velocity at the inlet side of the fan blades to be 1.8 m/s. 1.2.4.1 Velocity Triangles The velocity vector at the inlet side of the fan blade is radial at any point along its circumference while the velocity at the outlet is displaced angularly depending on the blade angle and the angular velocity of the blower wheel. Thus we need the velocity triangles to find the flow properties at the outlet side of the blower wheels. The Figure 1.4 gives the Inlet and Outlet Velocity Vectors for a typical Backward Curved Centrifugal Blower Blade [DG]. The V1 (V1 = Vf 1 ) and V2 refer to the absolute velocity at the inlet and outlet of 6 1.2. TUNNEL COMPONENTS CHAPTER 1. Design Figure 1.4 Inlet and Outlet Velocity Vectors for a Backward Curved Centrifugal Blower Blade the blades, this is the velocity observed from a stationery reference plane. U1 and U2 are the angular velocity of the blade at its entrance and the exit. The Vr 1 and Vr 2 are the relative velocities which refer to the velocity that the blade sees while rotating with an angular velocity. β2 is angle of the rotor blade at the outlet while α is the angle between the absolute velocity and the plane of rotation. The magnitude of radial component of the outlet velocity can be calculated by writing the continuity equation between the inlet and outlet side of the blower wheel or between the outlet side of the blower wheel and the slots[BRMy ]. The outlet diameter of the fan is 23.18 cm. The density remains the same across the fan as the pressure drop is small and there is minimal temperature rise or drop. Further the breadth of the blower wheel is the same on either side. Thus using known data the velocity at the outlet of the fan blades Vf 2 can be calculated from the continuity equation as Vf 2 = Vf 1 D1 D2 1.8 m /s × 20.57 m = 23.18 m = 1.6 m /s . 7 1.2. TUNNEL COMPONENTS CHAPTER 1. Design Vf decreases from inlet to the outlet as the tip of the fan has a greater rotational velocity compared to the hub. Thus a lot more of the radial component of velocity is converted to whirl or angular component. The blower wheels run at the maximum speed of 1475 RPM. Thus the angular velocity is calculated by, U1 U1 Similarly, U2 rev r a d 1 m i n 0.2057 2π m mi n r e v 60 s e c 2 = 15.87m/s = 1475 = 17.89 m /s “U" gives the angular velocity of the blades and “Vf 2 " gives the magnitude of the radial component of the output velocity. Now the magnitude of the whirl velocity “Vw 1 " needs to be calculated to define the outlet flow. The whirl velocity and the radial velocity are the resolved components of the actual outlet velocity “V2 ". For this calculation the inlet and outlet velocity triangles are formed using the velocity vectors. Figure 1.5 Inlet and Outlet Velocity triangles for a Backward Curved Centrifugal Blower Blade The Figure 1.5 represents the inlet flow conditions. The value of “β1 " is calulated from the radial and angular velocity as, β1 = a r c t a n  Vf 1 ‹ U1 Knowing “β1 " the relative velocity for the inlet flow can be calculated from the absolute and angular velocity. 8 1.2. TUNNEL COMPONENTS CHAPTER 1. Design Vr 1 = U1 c o s β1 Solving the above equations we get the relative velocity at the inlet to be 15.75 m/s. The outlet triangle is a little different compared to the inlet as the rotation of the fan and the blade angle add a whirl velocity to the flow. But similar calculation can be made for the outlet flow as well knowing the blade angle and using the outlet velocity triangles. The blades in the lower blower wheel are symmetrically curved backwards and kept at an angle such that a tangent drawn to the blade at the inlet is perpendicular to the direction of rotation while it is almost parallel to the direction of rotation at the outlet of the blower wheel. The tangent drawn to any point on the blade represents the direction of the flow relative to the blade at that cross section. Thus the angle the tangent forms with the direction of rotation of the blade gives the blade angle. Thus from the discussion the blade angle is 90◦ at the inlet and 0◦ at the outlet and is between these two values at any point in between. A more in depth analytical model of the fan is developed in Chapter 2 when the fan is to be simulated. For the preliminary calculations let us take an arbitrary blade angle, say 10◦ to proceed with the calculation. This represents a point in the blade that is closer to the outlet and can be assumed to be rotating at 17.89 m/s which is rotational velocity at the outlet. cot β2 = U2 − Vw 2 Vf 2 In the above equation the whirl component of the velocity Vw 2 is the only unknown and is calculated as 8.82 m/s.The radial and whirl component define the outlet flow of the lower wheel blower.The magnitude of the absolute velocity and its direction can be calculated as tan α2 V2 2 = Vf 2 /Vw 2 = Vf22 + Vw22 An absolute velocity of 8.93 m/s is calculated at the outlet of the lower wheel which is which is at an angle of 10.28 ◦ to the axis of rotation. This represents the flow properties at the outlet for an arbitrary cross section of the blade where the blade angle is 10◦ . 9 1.2. TUNNEL COMPONENTS 1.2.5 CHAPTER 1. Design Inner Shell The Inner shell is 43.2cm in height, 31.75cm wide and 91.4cm in length, a 19.05cm x 30.48cm tunnel section runs right through the center of the shell throughout its length. The tunnel section splits the shell into two columns on either sides. The panels are fixed to the inner shell and together they form the flow area that houses the heating coils and the lower blower wheel.The shell opens to the tunnel section through the slots in the panel. In-order to be consistent the standard inclined outlet panel is considered for all the calculation. Figure 1.6 Inlet and Outlet Velocity triangles for a Backward Curved Centrifugal Blower Blade In this preliminary analysis of the inner shell we try to quantify the approximate pressure drop in the shell. This is the pressure against which the lower blower wheel is operating against. If the pressure drop is increased by adding resistance, the flow across the fan is reduced. Knowing the size of the slots, the velocity of flow through the column and slots the major and minor losses for the inner shell can be calculated. To find the pressure drop in the system a streamline is drawn from the fan outlet to the slots. The pressure drop can be written from the Bernoulli equation along the stream line.  ‹ ∆P V 2 F l X = + k ρg 2g Dh X ∆P = hf + hmi no r ρg Where, 10 1.2. TUNNEL COMPONENTS CHAPTER 1. Design ∆P - Pressure Drop hmi no r - Minor head loss in m h f - head loss due to friction m . Lets consider a cross sectional plane near the mid section of the tunnel length where the flow can be considered to be one dimensional for the purpose of calculating the losses.The streamline is drawn along this cross section. Head loss due to minor losses is calculated from the equation: hmi no r = kV 2 2g Where, k - Minor Loss Co-efficient V 2 - Velocity of the flow in m /s g - Acceleration due to gravity in m /s 2 The flow exits the inner shell through either the inclined slots or the horizontal slots that have a similar minor head loss associated with them.Note that the air stream exits through either one of the slots and hence the minor loss at the outlet is not to be doubled for the system. The minor loss coefficient for such an outlet is 2.8 [Min] and the velocity through it 4.41 m/s, the mean velocity at the slots. This gives a minor head loss of 2.78m due to the slots. Another minor loss is associated with the 90◦ bend in the inner shell. The radial velocity Vf 2 at the outlet side of the fan discussed in the previous section is used to calculate this loss. This is purely a theoretical value calculated using the continuity equation. In reality the velocity coming out of the centrifugal blower wheel has a radial and whirl component making the flow two dimensional. Analytical calculation of the minor losses in two dimensional flow is difficult and since the most significantly loss in this system is due to the slots this effect can be safely ignored. The minor loss coefficient for a sharp 90◦ bend is 1. Thus a velocity of 1.6 m/s through this bend would give a minor loss head of 0.13m. For the same stream line of the flow the major losses are considered. The major losses are due to the friction in the duct. Friction to the flow is given by the walls of the column where the flow 11 1.2. TUNNEL COMPONENTS CHAPTER 1. Design velocity is calculated knowing the mass flow through the column to be 0.076 k g /s and the cross section area to be 580.64 c m 2 to be 1.08 m/s The head loss associated with friction is calculated using the Darcy-Weisbach equation given by, h f = fD L V2 Dh 2g Where, fd - Darcy friction coefficient. L - Length of Duct. Dh - Equivalent Hydraulic Diameter - 4(Area)/Perimeter V 2 - Velocity of the flow in m /s g - Acceleration due to gravity in m /s 2 The hydraulic diameter of the duct is calculated to be 0.119m while its length is 0.3048m, the height of the column. The Darcy friction coefficient is calculated from the Colebrook equation given by,   ε/Dh 1 2.51 + p . = −2.0l o g p 3.7 fd R e fd Where, ε - Surface Roughness m . R e - Reynolds Number - V .Dh /ν The flow velocity through the column is 1.08m/s and the kinematic viscosity is 15.11 × 10−6 m 2 /s (ν for air at 20◦ C).This gives a Reynolds number of 8510 corresponding to that of a turbulent flow. (The flow is laminar up to Re = 2300 for flow through a duct). Thus knowing the Reynolds number and taking the surface roughness for aluminum to be 0.001 × 10−3 m the friction factor is calculated to be 0.0323 from the Colebrook formula. Substituting the known values into the Darcy-Weisbach equation we get the head loss due to friction 12 1.2. TUNNEL COMPONENTS CHAPTER 1. Design to be 0.005m.Thus we see how the major head loss and in fact even the minor loss due to the bend is negligible compared to the minor loss due to the slot. This minor loss at the slot is purely a function of the velocity through the slots. This velocity was got by averaging the values got from experiment. Thus we see that not including the complexities of the flow inside the inner shell will not cause much change to the total pressure drop. The head loss due to the minor losses is 2.91 m while that due to friction is 0.005 m. The pressure drop in the inner shell is calculated from the Bernoulli equation. 2.91 m of air flow head loss corresponds to 34 Pa. This is the pressure the fan must operate against and is commonly used in association with the fan curves to find the optimum range of operation for a fan. Unfortunately the fan curves for a separate blower wheel is not available and thus such a analysis was not possible. 1.2.6 Heating Element Tubular heaters are used to provide the necessary temperature rise to the jet steam. They are housed at the lower half of the inner shell on both sides. The heater consist of a helical resistance coil, made out of an alloy of 80% Nickel and 20% chromium, which is stretched and centered in a magnesium oxide metal sheath. Figure 1.7 Heating Coils 13 1.2. TUNNEL COMPONENTS CHAPTER 1. Design The element is bent as shown in the Figure 1.7 to fit into the inner shell of the oven. With a length of 48.2 cm and a bend radius of 2 cm the total heated length is 211 cm.The element is 1.2cm in diameter, uses 3kW and operates at 480V.Knowing the dimensions and the power used, the watt density of the heater element can be calculated. Watt/Density( i nWc h 2 ) is defined as the heating element power divided by the actively heated surface area of the element. Element Wattage π(Element Diameter) x (Element Heated Length) 3000W = π(0.475)(4(19) + 3(0.787π)) W = 24 i nc h2 W = 3.6 c m2 Watt Density(q˙ ) = Watt density governs the element sheath and internal resistance wire temperature and is a useful measure for predicting relative heating element temperature when comparing different alternatives. This has to be matched with the velocity of the fluid flow to prevent overheating of the heating element. The operating temperature of the element is an important consideration for heat transfer efficiency and life. The velocity of air has to be matched with the watt density of the element to avoid element overheating. Let us proceed to calculate the capacity of the tubular heating coils.The amount of kW of heat needed to be provided to an air stream at a particular velocity to raise its temperature by δT is given by Q r e q ui r e d (k W ) = 1.08 × C F M × ∆T 3413 B T U /k W h (1.3) Where, CFM - Cubic feet per minute. ∆ T - Temperature Rise (F) Knowing the kW we can calculate the temperature rise the heating coil is expected to give for the calculated flow of air. The mean velocity of the air flow over the heating coil can be calculated by using the continuity equation between the slots and the cross section of the inner shell where the 14 1.3. PRODUCT TESTING CHAPTER 1. Design coils are attached. This gives a velocity value of 1.08 m/s across the heating coil. Air at 1.08 m/s passes through the inner shell area which is 6.35 cm width and 91.4 cm in length. This gives a air flow of 226 m 3 /h r or 133 CFM. Thus the temperature rise is calculated from Eq. 1.3 to be 69 F. 1.3 Product Testing Before the establishment of the baseline performance it is good practice to test the system for its proper functioning. Each component of the oven was tested for its proper functioning before experiments where carried out. The following steps were undertaken towards testing the product. 1.3.1 Dimensional Correctness The oven was measured to check if it was in accordance to the design parameters. The length, breath and height of the tunnel section, the outer cover and the panels where measured and found to be as discussed in the tunnel design in Chapter 1. 1.3.2 Alignment Improper alignment is a factor that could potential cause a huge error in the experimental data if not corrected. This is true especially in the heat shrink oven which is highly sensitive to the flow properties. The two centrifugal fans are responsible for the flow through oven, they were adjusted to be perpendicular to the floor. The heating element is mounted in the center of the main body and is shorter in length than the main tunnel body. The coils had to be symmetrically placed on each column of the inner shell so that air would pass over the same cross section of the heating coils yielding symmetric temperature on both sides. 1.3.3 Fan Operation The oven design dictates the fan rotates in the counter clockwise direction at a constant speed of 1450 rpm. The rotational direction fan was checked. Further the clearance of the fan was made as small as possible to avoid spillage of air from the inner to the outer shell or vice versa. 1.3.4 Heating Coils The Heating coils where first visually inspected for observable damage. This was followed by a “Continuity Test” to test for proper functioning of the heating elements and to make sure they operating properly. A multi-meter was used to get the resistance across the heating coil element.The 15 1.4. OVEN REQUIREMENTS CHAPTER 1. Design heating coils used operate at 480V and though they are rated for 3000W ,measurements showed that they draw 3500W when operated. Further The resistance measured across the heating element read 65 Ω which agrees with the value calculated by Ohms Law. V =IR , P =V I Combining these two equations we have, R V2 P 4802 = 3500 = 65.8 Ω = Both the coils had the same resistance of approximately 65 Ω which was in accordance to what was calculated. 1.4 Oven Requirements In order to recognize the critical parameters of the oven, the requirements for the proper shrinking of the wraps is to be established. The requirement can be summarized as, “The oven should be designed to produce impinging jet stream at a certain temperature that could be directed at a certain velocity around the film label symmetrically for its proper shrinking.” The major parameters for which the oven is to be designed for are set by this requirement. These values vary with the product to be shrink wrapped and hence the oven should have the flexibility to develop air jets for a wide temperature and velocity range. These parameters are listed below. 1. Temperature of the Jet Stream. 2. Mean Velocity of Jet Stream. 3. Turbulence of the Jet Stream. An additional parameter that enhances the shrinking process and directly affects the efficiency of the oven is also studied along with the major parameters. • Temperature along the Tunnel Section. 16 CHAPTER 2 Shrink Oven Analysis Understanding the flow physics is of prime importance in solving any fluid dynamics problem. With this in mind, much time has been spent to get a clear picture of the flow before any design modification was considered. In order to design a better model of the oven the existing model had to be scrutinized. A three phase 480V unit was setup at the “Structures Lab” at North Carolina State University to study its flow and heat transfer characteristics. The oven was analyzed through extensive data collection and computer simulations. These go hand in hand while trying to understand the behavior of flow in the oven and its interaction which the heating element. To develop air jets as dictated by the requirements in Chapter 1, two characteristics of the oven needs to be studied. 1. The Flow Characteristics 2. The Heat Transfer Characteristics Before moving forward to discuss the characteristics of the oven lets look at the tools and methods used to study the oven. 17 2.1. ANALYSIS TOOLS 2.1 CHAPTER 2. Shrink Oven Analysis Analysis Tools Various methods where employed to examine the oven and understand its flow and heat transfer characteristic. Experiments where carried out to record data of the most significant parameters to establish a baseline performance, computer simulations where used to study qualitatively the flow field in the inner shell and the tunnel section while analytical models where used to recognize the various parameters that affect the heat transfer characteristics. 2.1.1 Experimental Tools A data acquisition system was setup using AccuSense F900 Air Velocity Sensor, Type-T and Type-K thermocouples and 20A current transducers to measure the various parameters of the oven. The Type T - Thermocouples was interfaced with Onset U12-014 data logger while the transducer and the air velocity sensor were connected to the H22 Energy logger. The loggers records the data over a period of time which could be readout using the HOBOware data logger software. 2.1.1.1 Anemomenter-AccuSense F900 Sensors Figure 2.1 AccuSense F900 Sensors [F90] 18 2.1. ANALYSIS TOOLS CHAPTER 2. Shrink Oven Analysis The AccuSense F900 Sensors uses thermal anemometer technology, which measures heat loss from a heated device. Two thermistors on the sensor head detect airflow and ambient temperature simultaneously. The airflow reading is then compensated for temperature within the specified calibration range, linearized, and made available at the output. Each sensor is individually calibrated, which makes the F900 fully interchangeable. The F900 has a operating range of 50 to 140 F and has an accuracy of 0.05 m/s up to 10 m/s. 2.1.1.2 Thermocouples Figure 2.2 Thermocouple - Type T A thermocouple is a temperature sensor that works based on “The Seebeck Effect” . It consists of two dissimilar metals, joined together at one end. When the junction of the two metals is heated or cooled a voltage is produced that can be correlated back to the temperature. The Type T thermocouple has copper and constantan [an alloy of copper and nickel ] as its dissimilar metals. Type T thermocouples have a range in the −250 to 350◦ C range [Dow10]. Type T thermocouples have a sensitivity of about 43 µV /◦ C thus can measure temperature changes as small as 1◦ C. 19 2.1. ANALYSIS TOOLS CHAPTER 2. Shrink Oven Analysis The Type K thermocouple has chromel and alumel as its dissimilar metals.It has a wider measuremnt range and is accurate from −200 to 1250◦ C and has a sensitivity of about 41 µV /◦ C. Type K was used to measure the temperature on the walls of the heating coil for heat transfer coefficient calculations. 2.1.1.3 Current Transducer Figure 2.3 2 - 20 Amp split-core AC current sensor - CTV-A [Ct] Current transducer is a type of sensor that measures the magnetic flux of a power conductor to sense drive motor currents for machinery and process equipment and transmits an analog milliamp or voltage signal to control systems. ONSET 2 - 20 Amp split-core AC current sensor - CTV-A was used analone with a U12 data logger to measure the power used by individual heating coils. The heating coils are rated for 3KW and operate at 480V ,this gives a current range that is well suited to be measured accurately by a 20A CT. 2.1.1.4 H22 Energy Logger and FlexSmart Analog Module The HOBO Energy Logger is a modular, reconfigurable data logging system. The H22 gives the flexibility for expansion up to three FlexSmart Modules and six Smart Sensors. During the initial 20 2.1. ANALYSIS TOOLS CHAPTER 2. Shrink Oven Analysis testing phase of the oven Onset Thermocouples that plug in straight into the smart sensor input port of the H22 logger where used. Along with the data acquisition software this was used to monitor simultaneously upto six temperature readings and three velocity readings at different location in the tunnel. Figure 2.4 H22 Energy Logger and FlexSmart Analog Module [Ene] The FlexSmart Analog module is a DC signal-conditioning module for the Onset H22-001 HOBO Energy Logger.The module features input protection and signal filtering, as well as delta-sigma A/D conversion and factory calibration. Sensors/transducers are connected to the module via a seven-pin Phoenix-style detachable screw terminal connector. Once the sensors/transducers are connected, the module can then be configured using HOBOware Pro software. 2.1.2 Computational Tools Computer simulations were used as an aid in visualizing the flow and provide a base to test design modifications. A CFD model was developed to study qualitatively the flow field in the inner shell and the tunnel section of the oven.A Reynolds Averaged Navier Stokes (RANS) solver was chosen as 21 2.1. ANALYSIS TOOLS CHAPTER 2. Shrink Oven Analysis it was predicted to be sufficient to capture the quality of the flow and provide useful solutions. The ANSYS workbench provides a common interface for all components of the CFD analysis. 2.1.2.1 Designing Tool ANSYS DesignModeler software allows the import of CAD models from other modeling software while also providing tools for construction of geometry from the ground up. The ANSYS DesignModeler product is fully parametric in the sense it can, with parametric meshing and parametric solver use the same geometry for multiple design variations. Most of the models where made using the in-built tools in the DesignModeler for ease of connecting them with the other tools of the workbench. The models where made to actual size to avoid scaling during the analysis. 2.1.2.2 Meshing Meshing (also called grid generation) is the process of splitting flow domains into sub domains which are primarily composed of triangles or quadrilaterals for 2D geometry or tetrahedral or hexahedral in 3D geometry. Governing equations are discretized and solved in every single sub domain. The sub domains are called cells or elements. Combined, they are collectively called mesh. Mesh generation is one of the most critical aspects of engineering simulation. Too many cells may result in long solver runs, and too few may lead to inaccurate results. The model was meshed using ANSYS Meshing which provides a means to balance these requirements by automated algorithms. The most common meshing techniques used include the “Curvature” advanced size function which was used for the default domain and “element size" function for domains that need a finer mesh based on the requirements of the simulation. Some of the domains that were fine meshed using the sizing function include the fan, slots and the turbulators in the redesign stage. Table 2.1 Mesh Properties of Various Designs Geometry # of Nodes # of Elements Avg. Skewness Inner Shell 1,069,997 5,948,649 0.217 Plates with slits 1,003,180 5,563,693 0.217 Tunnel Section 79,891 408,244 0.244 Plenum Redesign 160,409 861,452 0.222 The properties of the mesh is given in the table for various geometries that have been simulated 22 2.1. ANALYSIS TOOLS CHAPTER 2. Shrink Oven Analysis in this project. Best practice is to keep the skewness to be less than 0.85 and thus an average mesh skewness of about 0.22 indicate the mesh to be of high quality. A grid dependency study was carried out to obtain grid independent solutions. In a grid dependency study, a family of grids is created by refining the previous grid and the solution is calculated until the solution stops changing. 2.1.2.3 Solver The ANSYS workbench provides two flow solvers, CFX and Fluent.The CFX is a Reynolds-Averaged Navier-Stokes (RANS) based solver while Fluent has the capability to run Large Eddy Simulations (LES) in addition to RANS. Both models use the steady-state Navier-Stokes equations to predict the velocity and the pressure fields in the laminar regime. They assume that the velocity field does not vary with time, and get an accurate prediction of the flow behavior. However as the Reynolds number increases, the flow field exhibits small eddies, and the timescales of the oscillations become so short that it is computationally not feasible to solve the Navier-Stokes equations. In this flow regime, the solver either uses a Reynolds-Averaged Navier-Stokes (RANS) formulation, which is based on time-averaged equations of motion for fluid flow while the Large Eddy Simulation uses filtering to separate the flow fluid, all flow scales larger than the filter size specified will be exactly calculated and the scales smaller than filter size will be modeled. Thus the turbulence model determines how the flow field is solved at high Reynolds Number.Thus deciding the best model is a key issue in fluid flow simulations. In order to get a accurate solution it is important to choose a solver best suited for the flow under study. The flow in the inner shell, for which most of the simulation is done, is similar to a turbulent flow in a duct where most of the fluid meanders at high speed through the core of the channel while bumping into and rebounding from the walls. These collisions lead to the formation of slower eddy-filled layers of fluid along both walls, these layers serve as “lubricant” for the relative motion between the fast core fluid and the stationary walls. These type of flows have been studied using a SST turbulence model [Men03] which is a RANS-based turbulence model. CFX has an inbuilt option to choose SST model to solve for the turbulence. The shear stress transport (SST) formulation [Men94] solves for k; the turbulent kinetic energy, and epsilon (k−ε) in the free stream and solves for ω the specific rate of dissipation of kinetic energy and k near the walls(k−ω). It does not use wall functions and tends to be most accurate when solving the flow near the wall.The use of a k−ω formulation in the inner parts of the boundary layer makes the model accurate near the walls and the switch to a k−ε behavior in the free-stream avoids the 23 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis common k−ω problem that the model is too sensitive to the inlet free-stream turbulence properties. The flow is very uneven along the length of the inner shell due to the whirl velocity of the fan. Though the flow is chaotic it doesn’t change with time as could be seen in an uniform results got at the slots with time. Thus a steady state simulation was good enough to analyze the flow. 2.2 Flow Characteristics The flow of hot air throughout the oven is studied in this section. How this air is heated would be dealt in the heat transfer characteristics section. The flow characteristics in the outer shell is neglected as in does not affect the flow that is responsible for the shrinking of the wraps. The following where identified to be the most important flows in the oven and were studied in detail. 1. Flow Characteristics in the Inner Shell • Quantitative Analysis of the Jet Stream. • Qualitative Analysis of the flow in the Inner Shell. 2. Flow Characteristics in the Tunnel Section. • Quantitative Analysis at the fan inlet. • Qualitative Analysis of the flow in the Tunnel Section. 2.2.1 Quantitative Analysis of the Jet Stream. The impinging jet stream is the flow, out of the inner shell and is directly responsible for the film label to shrink around the bottles. Three important parameters of this jet stream is studied, these parameters form the baseline which would be improved upon in the redesign stage. They are, 1. Temperature of the Jet Stream. 2. Mean Velocity of Jet Stream. 3. Turbulence of the Jet Stream. 2.2.1.1 Temperature of the Jet Stream The temperature of the air jet at the slots is one of the parameters for which the tunnel is to be designed for. This temperature defines the temperature profile throughout the length of the tunnel section. The main objective of the project is to come up with a uniform temperature profile across 24 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis the length of the tunnel section and having a symmetric temperature on either sides of the panels is of paramount importance in achieving this. Thus establishing a baseline temperature distribution at the slots was one of the most foremost things to be done. 240 220 Fahrenheit 200 180 160 Left Panel 140 Right Panel 120 Slot Length Figure 2.5 Temperature distribution along the length of the slots (Left Panel and Right Panel) The plot in Figure 2.5 shows huge difference in temperature between the slots on the left and right panels along the length of tunnel. This is not the temperature distribution you would expect given the tunnel is symmetric in design along the lengthwise axis and similar heating coils are placed on either sides of the panel. Further these heating coils where checked for proper functioning and alignment during the product testing phase. A temperature difference along the symmetric axis 25 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis hints a variation of airflow across the coils. Correcting this temperature gradient would significantly improve the efficiency of the tunnel and improve the quality of the shrink wrapping process. 2.2.1.2 Mean Velocity of Jet Stream The velocity profile across the slots defines the second most important parameter for which the tunnel is to be designed for. A minimum velocity is needed for the proper shrinking of the film. The F900 anemometer was used to find the velocity profile across the slots on either sides. The readings were taken over a period of sixty seconds and averaged to get the mean velocity at a given point. 5 4.5 4 m/s 3.5 3 2.5 Left Panel Right Panel 2 1.5 1 Slot Length Figure 2.6 Velocity distribution at the slots - Left Panel and Right Panel 26 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis It is seen from the plot in Figure 2.6 that the velocity distribution along the length is not uniform as well. This is in-line with our theory that the temperature gradient across the length of the tunnel is caused by a variation in the air flow across the coils. The heating coils are placed next to the slots along its length inside the inner shell. Hence the flow properties at any given point in the slot can be taken to be the flow property at the element cross section next to it. To visulazie this dependence between the temperature and mean velocity of the jet stream they are plotted together for the slots on the left panel in Figure 2.7. 5 240 4.5 220 4 Fahrenheit 3 m/s 3.5 200 180 2.5 160 2 Temperature 140 Mean Velocity 1.5 120 1 Slot Length - Left Panel Figure 2.7 Dependence of Temperature with Mean Velocity at the slots of the Left panel It can be seen from Figure 2.7 that the velocity and the temperature are inversely related for 27 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis most of the regions of flow. Wherever there is a drop in the mean velocity there is a corresponding increase in the temperature and vice versa. This is true in the case of the right panel as well (refer to Appendix). For a specified cross section lower mean velocity implies lower volume of air that the coil needs to heat up with the same power, this explains the inverse variation of temperature with mean flow velocity. 2.2.1.3 Turbulence of the Jet Stream Directionality is a important functionality of the jet stream that renders the tunnel to be flexible to shrink wraps products of various configuration. Turbulence defines the extent to which the air jet could be controlled. Thus the turbulence should be as low as possible to get a high quality jet stream. Also the turbulence could potentially affect temperature of the jet stream at the slots which is the major parameters of the oven design. Thus establishing a turbulence baseline is significant to understand the performance of the oven. Turbulent eddies create fluctuations in velocity. If the flow were steady and laminar then u = u¯ for all time (t), where the over-bar denotes a time average. For turbulent flow, however, the velocity record includes both a mean and a turbulent component. We decompose the flow as follows. u (t ) = u¯ + u 0 t Where, u¯ - Mean Velocity. u 0 - Turbulence fluctuation with time. This is commonly called a Reynolds decomposition [Mul06]. Measurements of velocity were recorded at various points along the four inclined slots for a period of sixty seconds to get an idea of the turbulence levels. These readings are plotted for the same points on the left and right panel slots to get an idea of the turbulence levels of the mixing stream. The turbulence level at a few locations (marked in the slot representation below each plot) are shown to study its behavior. The Turbulence fluctuation with time at a point near the tunnel entrance is shown in Figure 2.8. It can be seen from the plot that the turbulence level at this point on either jet streams is fairly low. with very low turbulence in the jet stream from the slot in the right panel.The left jet stream is a little more turbulent when compared to the jet stream from the slots in the right panel. This might explain its lower mean velocity and a higher temperature as seen in Figure 2.6 and Figure 2.5. In Figure 2.9 we see that the jet from the left panel is highly turbulent a little past the mid point of the front slot. Again referring to the plots in Figure 2.6 and Figure 2.5 we find that the mean velocity at this point is very low while the temperature at this point is at its peak for the given temperature setting. This 28 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 Left Panel Right Panel 2 1.5 1 Time(s) Probe Figure 2.8 Turbulence fluctuation with time at the start of the front slot (Near Tunnel Entrance) 5 4.5 4 Velocity (m/s) 3.5 3 2.5 2 Left Panel Right Panel 1.5 1 Time(s) Probe Figure 2.9 Turbulence fluctuation with time at 2/3r d the length of the front slot 29 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis phenomenon could be explained by establish the relationship between turbulence level and the heat transfer rate of convection. Turbulent flow promotes better mixing of the fluid compared to laminar flow and thus have better heat transfer characteristics. The effect of turbulence on the heat transfer can be explained in terms of heat transfer relations, h = f (N u) Nu = f (R e , P r ) Where, h - Convective heat transfer coefficient of the fluid. Nu - Nusselt number Re - Reynolds number Pr - Prandtl number We know that the Reynolds number is large for a highly turbulent flow. Therefore Higher Turbulence =⇒ Higher Re =⇒ Higher Nu =⇒ Better Heat transfer Now lets look at the turbulence levels in the rear slots. The probe was placed at 1/3rd the length of the rear slot which represents a location which is equidistant from the mid point of the tunnel and opposite in direction as the location of the probe for the previous experiment whose characteristics where discussed in Figure 2.9. From the values plotted in Figure 2.10 it is seen that jet stream leaving the slots on right panel have a higher turbulence at the rear end of the tunnel.They could be matched with the mean velocity and turbulence values as was done in previous cases. The left jet stream has a very low turbulence at this location. The switch in turbulence from the left side to the right from the front end to the rear end might be due to the effect of the rotation of the fan affecting the flow. This effect will be discussed in the next section when the quality of the flow is studied. The plot in Figure 2.11 shows that the jet stream on the left and right side have almost the same mean velocity and turbulence levels near the tunnel exit.This case further confirms the theory that the flow characteristics, namely mean velocity and turbulence level, are responsible for the temperature 30 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 2 Left Panel Right Panel 1.5 1 Time(s) Probe Figure 2.10 Turbulence fluctuation with time at 1/3r d the length of the rear slot 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 Left Panel 2 Right Panel 1.5 1 Time(s) Probe Figure 2.11 Turbulence fluctuation with time at the end of the rear slot (Tunnel Exit) 31 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis gradient along the length of the slots as at this point near the exit of the tunnel the temperature of the two stream are almost the same as seem in the Figure 2.5. Also this is the point in the tunnel farthest from the fan and have the least fan effect. The front slots do not start at the entrance of the tunnel and hence have more fan effect than the end of the rear explaining there higher turbulence. The turbulence of the two stream that mix inside the tunnel section is different at every point(expect at the entrance and exit ends) and flips its turbulence intensity at around the mid point of the tunnel. The four examples where taken to explain this effect. The velocity fluctuations for other locations of the slots are given in the Appendix. 2.2.1.4 Effect of Turbulence Strength on Flow Parameters of the Jet Stream For ease of comparison and to better visualize the effects of this turbulence on the temperature of the jet stream, the characteristics of the turbulence is calculated at each point of consideration. The turbulent motion is approximately random and we need to use statistical concepts to characterize them. In theory the velocity record is continuous and the mean can be evaluated through integration. However, in practice the measured velocity records are a series of discrete points, u(i). Below, an over-bar is used to denote a time average over the time interval (t), which is the 60 seconds over which the readings where taken in our case. Some of the main properties of a turbulence include [Nepll] Mean Velocity u¯ = t Z u (t ) dt continuous record 0 = Turbulence Fluctuation N 1 X ui N i =1 discrete,equi-spaced points u 0t = u (t ) + u¯ continuous record u 0i = u (i ) + u¯ discrete,equi-spaced points 32 (2.1) (2.2) 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis = Æ u¯ 0 (t )2 continuous record v u N u1 X = t (u 0 )2 discrete,equi-spaced points N i =1 i p Turbulence Intensity = u r m s u Turbulence Strength ur m s (2.3) (2.4) The subscript “rms" stands for root-mean-square. The root mean square of the velocity defined in Eq.2.4 can be extended for velocities in other directions, v(t) and w(t), as well. In this analysis we are concerned only with the velocity along x-direction as the jet out of the slot is linear. To get an idea of the turbulence at any other region of the tunnel where all three velocity components exist, the analysis of the velocity components in all three directions is required.A larger u r m s indicates a higher level turbulence. The relationship between the turbulence level in the air stream and its mean velocity can be established by calculating the turbulence strength given by the statistical method explained earlier. The turbulence strength was calculated by taking an time average for the 60 sec period of measurement at discreet points along the slot length using the relation below. ur m s v u X 60 t1 (u 0 )2 = 60 i =1 i discrete,equi-spaced points (2.5) The results are plotted versus the temperature along the length of the slots. It is seen from the Figure 2.12 that wherever there is a spike in the turbulence strength there is a mean velocity drop. This velocity drop as seen in the previous section causes the temperature to go up. Thus in order to get a uniform temperature distribution along the length of the slot the reason for turbulence creation has to be found and corrected. 2.2.2 Qualitative Analysis of the Flow in the Inner Shell The analysis of the experimental data has given us an idea of the relationship between the various flow parameters of the jet stream. Though we where able to relate the turbulence levels and the temperature of the jet stream, the reason behind the turbulence is not clearly understood. To recognize the cause behind the turbulence creation we need a proper understanding of the quality of the flow field in the inner shell. Computer simulations aid us in visualizing the flow and provide a base to test design modifications. 33 CHAPTER 2. Shrink Oven Analysis 0.6 6 0.5 5 0.4 4 Mean Velocity Turbulence Strength 0.3 3 0.2 2 0.1 1 0 0 Mean Velocity (m/s) Turbu;ence Intensity (m/s) 2.2. FLOW CHARACTERISTICS Slots Length Figure 2.12 Dependence of Mean Velocity with turbulence strength along the slots of the Left panel 34 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 0.6 240 Temperature Turbulence Strength 0.5 220 0.4 0.3 m/s Fahrenheit 200 180 0.2 160 0.1 140 120 0 Slot Length - Left Panel Figure 2.13 Dependence of Temperature with turbulence strength along the slots of the Left panel 35 2.2. FLOW CHARACTERISTICS 2.2.2.1 CHAPTER 2. Shrink Oven Analysis Simulation Setup - Inner Shell The Design module and the CFD meshing tool available in ANSYS workbench was used to model and mesh the geometry of the flow field in the inner shell. For the simulation of the flow in the inner shell the blower provides the input boundary condition and thus it had to be modeled. The blower wheel was implicitly modeled in CFX without actually having to include any of its geometry. Analytically calculated data for the flow at the fan outlet was used as the boundary condition representing the blower. The flow velocity values where calculated and loaded into the CFX-Solver as inlet velocity in cylindrical co-ordinates system. This allowed for the accounting of the whirl velocity at the outlet of the fan. The radial and whirl velocity at each cross section of the blade was calculated. The values were then averaged for the entire blade and was used as the inlet boundary condition for the simulation.This calculation is shown in the Table 2.2. Note that the rotational velocity and the blade angle change as one moves from the inlet side to the outlet side of the blade. The rotational velocity is a function of the diameter at that cross section while the blade angle is taken to change with a constant decrement as it is symmetric. The calculation is the same as shown in chapter one. Table 2.2 Velocity at the Outlet of each cross section of the Blower Blade β2 D2 U2 Vf 2 c o t (β ) Vw 2 s i n(β ) Vr 2 V2 α2 6 0.2318 17.89 1.59 9.5144 2.78 0.1045 15.20 3.20 29.74 14 0.2294 17.71 1.61 4.0108 11.24 0.2419 6.67 11.35 8.17 22 0.2271 17.53 1.63 2.4751 13.49 0.3746 4.35 13.59 6.89 30 0.2247 17.34 1.65 1.7321 14.49 0.5000 3.30 14.58 6.49 38 0.2223 17.16 1.67 1.2800 15.03 0.6157 2.71 15.12 6.32 46 0.2199 16.98 1.68 0.9657 15.35 0.7193 2.34 15.44 6.26 54 0.2176 16.79 1.70 0.7266 15.56 0.8090 2.10 15.65 6.24 62 0.2152 16.61 1.72 0.5317 15.70 0.8829 1.95 15.79 6.26 70 0.2128 16.43 1.74 0.3640 15.80 0.9397 1.85 15.89 6.29 78 0.2105 16.25 1.76 0.2126 15.87 0.9781 1.80 15.97 6.33 86 0.2081 16.06 1.78 0.0699 15.94 0.9976 1.78 16.04 6.37 36 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis Figure 2.14 Inner Shell with given boundary conditions The average values of the radial velocity was calculated to be 1.69 m/s while that of the whirl velocity was found to be 14.14 m/s. These values serve as the inlet boundary conditions for the simulation while the outlet is defined by a pressure boundary condition at the slots which is set at zero relative pressure corresponding to the the atmospheric conditions. The solver is set to “Steady State" analysis with no heat transfer or thermal radiations. Air at 25 0 C is chosen to be the fluid in the domain from the materials library. The Shear Stress transport model ,discussed earlier , is set to solve for turbulence and a non-buoyant buoyancy model is chosen as the simulation we are concern with is that of a fan driven flow. Advanced options dealing with mesh 37 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis deformation and domain motion are set to “none" as they are not necessary for our calculations and the reference pressure is set to 1 bar. 2.2.2.2 Simulation Result - Inner Shell The solver is run until the results converge. CFD-Post, a post processing interface is used to analyze the output of the simulation. Studying the surface streamlines in various planes it is seen that the large vortex are formed at the front left and the rear right end of the inner shell. This is found to be due to the nature of the fan flow and its interaction with the walls of the inner shell. The rotation of the fan creates a whirl velocity much greater than its radial velocity, this implies that more air leaves the fan tangentially than radially. The direction of rotation of the fan along with the geometry determines the region which would receive this higher tangential air flow. These regions have a steady flow of air through them while the areas that are void of the supply air become regions of low pressure with chaotic flows associated with them. The discharge of the fan can be visualized from the streamlines in the left most model in Figure 2.15. This shows the surface streamlines along a XZ plane at the very top of the inner shell where the fan is housed. The color of the streamlines represent the velocity in x-direction. Considering the lower half of the figure it can be seen that, due to the whirl velocity and the geometrical constrains, most of the streamlines migrate towards the lower right. This is the section on the right at the entrance of the tunnel and it has a steady stream of air passing through it. The flow through this section is congruent to the left section at the exit of the tunnel where all the air leaving the other side of the fan in the tangential direction flows through. This nature of the flow can be matched with the experimental results which showed that the slots in these regions had a high mean velocity and low turbulence. It can also be seen from the other two models in Figure 2.15 how the flow is in regions that don’t get sufficient air (bottom left and top right). A low pressure is created in these regions due to insufficient air supplied to it making the tangential air hitting the front and rear panels and the steady air supplied to the top left and bottom right regions to rush towards towards them creating a chaotic flow leading to the formation of large vortices. A vortex is created wherever the two opposites streams meet. The picture at the right shows the flow streamlines along a plane that is at the slot openings while the picture in the middle is of the streamlines along the XZ plane a little above it. 38 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis Figure 2.15 Streamlines along the XZ plane at various heights The accuracy of the computer model is checked by comparing the velocity at the slots with their corresponding experimental values. While the turbulence fluctuation at the slots could not be captured in the model as that would require a transient simulation. The non uniform mean velocity profile caused due to the inherit turbulence in the inner shell is captured in the model as shown in Figure 2.16. The profiles match almost perfectly while the magnitude of the velocity got through computer simulation is lower and could be explained by the fact that density variations where not accounted for in the model. The “Air at 25 0 C" fluid was used for the simulation while the experiment was carried at a set value of 60 0 C (140 0 F) . The higher temperature implies a lower density and thus a higher velocity for the same mass flow rate by continuity. 39 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 5 4.5 Mean Flow Velocity - m/s 4 3.5 3 2.5 2 Left Panel - Experimental Left Panel - Computational 1.5 1 Slot Length Figure 2.16 Comparison of Computational and Experimental results. 2.2.3 Quantitative Analysis at Fan Inlet The blower wheel creates a negative pressure at the inlet that causes air to be drawn into the fan from the tunnel section through the plenum box. This suction affects the flow in tunnel section and it is useful to establish a baseline for this flow velocity.The plenum box has a circular opening of diameter of 19cm through which the air is drawn in. The velocity at different points at the opening section of the plenum box was measured using the velocity sensor.The turbulence at the points where calculated using data measured over a period of 60 seconds as for the jet streams.The turbulence strength and the mean velocity are plotted along the radius of the opening in Figure 2.17. The flow at the suction side of the fan is studied to modify the flow in the tunnel section. Scope for 40 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis improving the flow exists by the redesign of the plenum box.Hence the values of the flow where calculated at the entrance of the plenum.The velocities and temperature at the slots on the other hand are definite parameters that need to be met by the oven and thus they where measured without the plenum in place to negate its effects and study the effect of the fan and inner shell separately. It is seen that the velocity is maximum at the center of the circular passage, this is expected given Velocity 3 2.5 Mean Velocity (m/s) 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Radius of Opening (inch) Figure 2.17 Flow properties at the Plenum Opening that the centrifugal fan creates a pressure difference at the suction side in bands that decrease with the radius.The turbulence at the tip of the passage is possible due to the flows interaction with the plenum box creating eddies. The turbulence for this flow is not a significant factor and the main parameter to be measured out of this experiment is the mean velocity across the plenum 41 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis passage.This was calculated to be approximately 2.12 m/s. This value can also be calculated from the continuity equation between the fan inlet and the slot as they form the inlet and outlet of the inner shell. The average temperature at the slots is 144 0 F and the average temperature measured at the fan inlet is 158 0 F, their corresponding densities are used in the equation below. (ρAv )Sl o t s vF a n I nl e t = (ρAv )F a n I nl e t (ρAv )Sl o t s = Aρ P a n e l O p e ni ng 0.996 × 0.01435 × 4.41 0.0332 × 1.029 = 1.845m /s . = Thus the Velocity at fan inlet has been measured by experiment and has been calculated analytically. • Experimental Value = 2.12 m/s • Analytical Value = 1.85 m/s Thus the error calculated for the velocity at the fan inlet calculated between the experiment and analytical model is 12%. This parameter is helpful to verify the computer simulation that will be used to study the nature of the flow in the tunnel section. 2.2.4 Qualitative Analysis of the Flow in the Tunnel Section The nature of the flow in the Tunnel Section is a direct indication of the performance of the tunnel. Getting uniform temperature and velocity profiles along the entire tunnel in all three directions is one of the criteria of a good design and was the target set for the project. A thorough understanding of the flow field in the tunnel section and the factors affecting it is important towards designing a oven that has a uniform flow properties in its tunnel section. A uniform temperature and velocity distribution would improve the quality of the wrapping and reduce the total energy usage by allowing the oven to be operated at a lower temperature for the same operation. Three flows interact inside the tunnel section and their interaction define the nature of flow in tunnel. They are the impingement jet stream of air that enter the tunnel from the slots on both sides, the air being drawn into the plenum by the fan and the outside makeup air that enters the tunnel section through the blinds. The properties of these flows can be studied by the measurement of the following parameters. 1. Flow Characteristics of the Jet Stream. 42 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 2. Flow Properties at Fan Inlet 3. Flow Properties of the Outside makeup air. The flow Characteristics of the jet stream and the flow properties at the fan inlet have already been discussed in much detail. Along with this the flow property of the make up air that enters the tunnel from outside define the flow in the tunnel section. To study qualitatively the interaction of these flows the tunnel section is modeled on CFX. 2.2.4.1 Simulation Setup - Tunnel Section The design module and the CFD meshing tool available were used to model and mesh the geometry of the flow field in the inner shell. The boundary conditions used for the simulations where calculated from the experimental values of the flow rate. While the slots served as the “Outlet Boundary Condition" for the inner shell simulations it is the “Inlet Boundary Condition" for the simulation of the tunnel section. While it is possible to use the output flow at the slots got in the previous simulation as the input to simulate the tunnel section, it was not used due to fact that the flow was uneven and we wanted to model the flow in an ideal tunnel section. This was done so that the effects of the plenum design could be studied without the turbulence effects of the flow that needs to be rectified. The total pressure of the flow had to calculated to define the inlet boundary condition 1 P0 = P + ρv 2 2 (2.6) Where, P - Static pressure, P0 - Total pressure which is constant along any streamline ρ - Density. v - Flow Velocity. The “Total Pressure" was calculated from Eq. 2.6 using the mean velocity got through experiments (4.41 m/s) and was given as the inlet boundary condition. The flow velocity at the plenum inlet calculated through analytical methods and verified by experiment in the previous section was used as the outlet boundary conditions for the simulation. This is considered a “Robust" boundary condition for CFX analysis where the static pressure at the outlet and the mass flow rate at the inlet are part of the solution. The entrance and exit of the tunnel section where modeled as openings 43 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis with air free to move in or out based on the flow conditions inside the tunnel section. An ideal tunnel section where the velocity and temperature profiles of the mixing jet streams are uniform is symmetric along the longitude axis. This symmetry is made use of to bring down the computational demand. Only the right half of the tunnel was modeled and the center plane was given a “Symmetric Plane" boundary condition. Figure 2.18 Simulation Setup of the Tunnel Section 2.2.4.2 Simulation Result - Tunnel Section The solver is run until the results converge and the flow is studied in CFD-Post. It is seen from 3D streamlines drawn from the “Opening" boundary condition to the “Outlet” in Figure 2.19 that there is a significant outside air that is drawn into the tunnel section. The path this air flow takes is dependent on the velocity and direction of the jet stream. The hot air jet gets discharged through the slots and the streams from opposite sides collide. This 44 2.2. FLOW CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis make them to change direction and form a vortex as shown for various cross section along the tunnel length in Figure 2.20. A low pressure region is created in areas enclosed by this vortex through which the outside air transverse through the length of the tunnel section. The cold outside air is responsible for the cold spots in the tunnel section and it could be observed that these spots, seen as blue regions in Figure 2.20, are created almost throughout the length of the tunnel. If the cold makeup air is not allowed to transverse into the center of the tunnel section, where it is eventually drawn into the tunnel, the area of cold spots in the tunnel can be reduced. Figure 2.19 Flow of Cold ambient air through the Tunnel 45 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis Figure 2.20 Tempertaure Distribution at various cross-sections along the length of the tunnel 2.3 Heat Transfer Characteristics In order to optimize the heat transfer between the tubular heating elements and the air flowing past it we need to understand the heat transfer mechanism in the oven. While computational simulation 46 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis and experimental results where used to analyze the flow characteristics in the oven, analytical tools along with experiment results are used to analyses the heat transfer characteristics. 2.3.1 Heat Transfer Relations In this section we try to establish the relationship between the various parameters that affect the heat transfer characteristics. The aim of this analytical study is to find out the controllable parameters that affect the temperature of the bulk air flow. Once these are identified we can look into changing the design to make better the heat transfer in the oven. All three heat transfer mechanisms are present in the oven design. The helical resistance coil is heated by resistance to the electric current and it conducts the heat to the metal sheath. The sheath transfers its heat to the air stream by convection and to the walls of the inner shell through radiation. The outer walls of the inner shell are cooled by the air forced over it by the upper blower wheel through convection while the inner walls have there heat transfered to the tunnel section. Thus all the heat produced is either dissipated through convection or radiation. Qp r o d u c e d = Q c o n v e c t i v e + QR a d i a t i v e (2.7) Radiation heats up the walls of the inner shell while heat transfered by convection heats up the bulk air stream. Though the heated walls transfer some of this heat back to the air stream the rest of it is lost to the environment. While it may seem common sense that in an optimized design most of the heat produced is carried away by convection and the radiation component is kept at a minimum, it may not prove to be the best way to heat by the bulk air because of the coupled nature of the two modes of heat transfer. Let us look at the convective and radiative heat transfer equations to understand this better. Q c o n v˙e c t i o n = h A (Tw − Tb ) Where, q˙ - Q˙ /A - Heat flux - Heat transfer per unit time per unit area. Q˙ - Heat transferred per unit time by convection. A - Heat transfer area of the surface Tw - Temperature of the element. Tb - Bulk Temperature of the air stream. 47 (2.8) 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis h - Convective heat transfer coefficient QR = εσ(Tw − Ts )4 (2.9) Where, QR - Heat transferred per unit time by radiation. ε - Emissivity σ - Stefan-Boltzmann constant = 5.67108 W m 2 K 4 . Tw - Temperature of the element. Ts - Temperature of the side walls. We see that the element temperature Tw couples the two modes of heat transfer. The side walls of the inner shell are constantly being cooled by the air flow through the outer shell and heat dissipation to the environment. Thus the surface temperature of the side walls Ts can be assumed to be independent of the Radiative heat transfer. QR ∝ Tw 4 (2.10) We see that the radiation component of heat transfer is proportional to the wall temperature raised to its fourth power. Thus to have minimum heat dissipated by radiation we need to keep the element temperature low. But this would also mean a low temperature of the bulk air flow whose relation can be deduced from Eq. 2.8 =⇒ Tb = Tw − q˙ h. (2.11) We see that higher the element temperature higher is the bulk temperature. Thus to get a higher air temperature at the slots we need to keep the element hot and deal with the unavoidable radiation. In fact it is seen from experiments that the heating coils are heated to a very high temperature and glow red almost throughout the operation of the oven. This implies that the amount of airflow across the coils is not enough to remove the heat produced by convection alone and that significant radiation is present. The element temperature Tw and the heat flux q˙ are functions of the power supplied to it, the only controllable parameter independent of the power supplied in Eq. 2.11 is the convective heat transfer coefficient. This unlike the coefficients of other modes of heat transfer is not a function of the material alone but also depends on the flow properties and by changing the design of the flow 48 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis area the flow field could be controlled. While increasing the convective heat transfer coefficient should increase the bulk temperature of the flow by equation Eq. 2.11 it is also to be noted that increasing the convective heat transfer coefficient also decreases the element temperature. Thus it becomes a trade off between the parameters and it is possible to know the results of a heat transfer enhancement method only through experimentation. Further the value of “h" is very difficult to analytical measure for a complex flow such as in the inner shell thus making experimentation to find which parameter affects the heat transfer characteristics the most is the only viable option. It is useful however to know the flow parameters that affect the convective heat transfer coefficient. For this, let us look at the convective heat transfer mechanism closer to known these parameters. 2.3.1.1 Mechanism of Convective Heat Transfer The general concept of convection is that near the heating element wall there is a thin layer in which heat is transferred basically by conduction. Outside of this region is high mixing. The thickness of the layer is not a fluid property and it depends on velocity (Reynolds number) and structure of the wall surface. Generally this thickness δ is not known and it is customary to calculate the heat transfer using k f l ui d /δ . This quantity is what is termed as the convective heat transfer coefficient. The heat transfer coefficient “h" depends upon flow properties and is calculated from established empirical relations for Nusselt number. In heat transfer at a boundary within a fluid, the Nusselt number (Nu) is the ratio of convective to conductive heat transfer across the boundary. For forced convection in turbulent pipe flow the Nusselt number is given by the Dittus-Boelter equation. It is a explicit function for calculating the Nusselt number and is tailored for smooth pipes. For a flow in which the fluid is heated it is given by 4/5 N u D = 0.023R eD P r 0.4 Most functions of Nusselt number are written for fully developed flows where the boundary layer extends to the centerline. Inside the inner shell where the air comes in contact with the heating coils, the flow is not fully developed. The manner in which the Nusselt decays from inlet to fully developed conditions depends on the nature of thermal and velocity boundary layer development in the entry region, as well as the surface thermal condition. While a averaged Nusselt number is calculated for a laminar flow these effects of entry and surface thermal conditions are less pronounced for turbulent flow and can be neglected. Also since the local convection coefficient varies around the periphery 49 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis of a tube, approaching zero at its corners, the Dittus-Boelter correlation may be used as a first approximation with the hydraulic diameter, irrespective of the surface thermal condition.[Internal Flow Heat Transfer Correlations Chapter 8]. The various parameters that affect this the heat transfer coefficient has been discussed in the previous chapter, lets proceed to find by experiment the overall heat transfer coefficient for the baseline model of the oven at an arbitrary operating condition. 2.3.2 Heat Transfer Coefficient - Experimental Result In an effort to find the heat transfer coefficient an experiment was setup to record in 1 second intervals the current drawn by the heating coil. The Heating coil is rated for 3 kW but it does not operate at maximum power throughout. A temperature controller that monitors a thermocouple and controls the power suppled to maintain the set temperature. A Type-T thermocouple was used to read the temperature of the flow at various places of the inner shell while a Type-K thermocouple was used to get the wall temperature of the heating elements as it was expected to be higher than the operating range of a Type-T. An electric heater is 100% efficient and all the power that goes into the element is converted to heat. Thus the current drawn by the element averaged over a period of time will give the heat transfer rate for this period. The oven was set to a “Set Value” of 140 0 F (60 0 C) and the reading of the temperature and current where recorded when the oven was maintaining this set value. The current power drawn by the element is shown in Figure 2.21 Heat transfer coefficient is the proportionality coefficient between the heat flux and the thermodynamic driving force for the flow of heat (i.e. the temperature difference, ∆T ). Thus the overall heat transfer coefficient can be written as, U= Q ∆T q˙ - Q˙ /A - Heat flux - Heat transfer per unit time per unit area. Q˙ - Heat transferred per unit time. A - Heat transfer area of the surface. Tw - Temperature of the Element(F). Tb - Bulk Temperature of the air stream. 50 (2.12) 2.3. HEAT TRANSFER CHARACTERISTICS CHAPTER 2. Shrink Oven Analysis 5000 4000 POWER (W) 3000 2000 1000 0 -1000 18:05:20 18:05:24 18:05:28 18:05:33 18:05:37 TIME PERIOD 18:05:41 18:05:46 18:05:50 Figure 2.21 Power Drawn by the heating coil with the oven set at 140 F 51 18:05:54 2.4. DESIGN DEFECTS CHAPTER 2. Shrink Oven Analysis U - Overall heat transfer coefficient. Averaging the values in Figure 2.21 we get the power drawn to be 1723 W. This is the heat produced by the coil to maintain the temperature of the oven at the set value of 140 0 F. The average temperature of the air stream before the heating coil (inlet) is found to be 161 0 F. This value is an average of the temperatures recorded at various points in the inner shell before the heating coil, while the set value of temperature for the oven is maintained by a thermometer at a specific location. This may explain why the set value at which the oven is supposed to be maintained is at a lower temperature than the average value. This is used only as a reference and its inaccuracy does not affect the performance of the oven. The temperature past the heating coil is given mean temperature at the slots. This is calculated from the data plotted in Figure 2.5 as 177F. Now the average bult temperature of air stream is given by, Tb Ti nl e t + To u t l e t 2 = 169. = The wall temperature during this period was found to be 500F on an average giving an overall heat transfer coefficient of 70.87W /m 2 F from Eq. 2.12. 2.4 Design Defects The analysis of the Shrink Oven has helped us to identify the major defects of the design from the standpoint of its flow characteristics. These defects are considered to be the setbacks to achieving a uniform temperature through the tunnel length. These shortcomings in the design along with their causes is listed below. 1. Non uniform flow properties along the length of the inner shell. • Direction of the fan’s outflow. • Interaction of the outflow with the walls of the inner shell. 2. Uneven Temperature distribution of the mixing Jet Streams. • The Non-Uniform flow across the heating coils along its length resulting in different heat transfer rates. 3. Cold Spots in the Tunnel Section 52 2.4. DESIGN DEFECTS CHAPTER 2. Shrink Oven Analysis • Uneven Temperature Distribution of the mixing Jet Streams. • Addition of cold outside air. The ideas generated and the retrofits built to overcome these defects are discussed in Chapter-3. 53 CHAPTER 3 Design Modifications The knowledge of the flow and thermal characteristics has given us the understanding necessary to make modifications to the design to get the desired results. Our main priority is to make the temperature profile more uniform throughout the length of the tunnel. From the analysis in the previous chapter it was understood that the difference in flow field turbulence levels at the inner shell is responsible for the difference in temperature and thus if we create uniform turbulence throughout the inner shell the temperatures could be matched. Various retrofits have been experimented to achieve this and will be discussed in this chapter. These include, 1. Turbulators. 2. Inner Shell Redesign. 3. Plates with Slits. 4. Cross-flow Fans It is to be noted that the experiments on the retrofits where done on the 61 cm model due to ease of manufacturing smaller retrofits. While the flow pattern remain the same, the value of the outlet temperature and the turbulence level may vary and hence the baseline shown in this chapter will vary with the initial data analyzed to study the performance. 54 3.1. TURBULATOR 3.1 CHAPTER 3. Design Modifications Turbulator Various turbulators are experimented with to create turbulence in regions that corresponds to lower temperature. It is known that turbulators in uniform flows generate secondary flow motions, increase the degree of flow turbulence and change the mean velocity fields in velocity boundary layers, which are responsible for the heat transfer augmentation [Tog13]. While extensive study has been done for uniforms flows in air duct only few could be related to the complex flow motions associated with the shrink oven. A lot of research has been done on various configurations of the turbulators to enhance heat transfer [Ala14]. Most of these are done for a uniform flow and may not work for the unique flow in the inner shell. Thus based on the review of the currently available literature a few configurations for creating a turbulence was experimented with. The first design considered was a basic rectangular rib. The position of the rib was to be decided based on the region to be made turbulent. If we divide the inner shell into four sections the flows at the front right and the rear left have steady flow. A 20 cm long rectangular rib made of aluminum was placed 6 cm above the heating coils on the front right section of the inner shell. Before experimentation the effect of the ribs are visualized using computer simulation.The results of the simulation are shown in Figure 3.1 The simulation on the left is of the baseline model without a rib. The flow velocity along vertical direction is shown in the YZ plane. The flow from the fan hits the side walls and flows along it as shown by the blue region along the side walls. It is seen that right next to it is the region in green where the air is almost stagnant or moving up after hitting the lower plate. This creates vortices’s in this region but as we have seen from experiment it is not good enough to give a high turbulence. Now lets us examine the flow with a rib as shown in the simulation on the right. The YZ plane is near the entrance of the tunnel and thus the column to the right represents the region where the flow is stable. The turbulator is kept in this column, 6cm above the region where the heating coils would be attached. The rib creates a low pressure region below it which cause vortex to form in this region. This can be seen from the red region in the simulation result which represents a back flow at this region. This flow pattern would mean a better mixing of the flow and more contact with the hot tube of the heating coils and thus a higher temperature of the jet stream is expected with the the turbulator. The single rib model represents Configuration A and was the very first turbulator to be tested. For the next setup, configuration B, an addition rib is added 9cm above this and on the opposite 55 3.1. TURBULATOR CHAPTER 3. Design Modifications Figure 3.1 Effect of Rib Turbulator on the flow field in the Inner Shell. 56 3.1. TURBULATOR CHAPTER 3. Design Modifications Figure 3.2 Rib Turbulator - Configuration B 57 3.1. TURBULATOR CHAPTER 3. Design Modifications wall. This rib is 30 cm long and extends beyond the first turbulator, the configuration was setup this way to see the effects of the second turbulator in combination with the first turbulator and by itself and can be seen in Figure 3.2. The experiments where carried out and the temperature readings at the front slots are taken and compared with the baseline. It is to be noted the baseline values don’t correspond with the values discussed in the chapter 2 as all the retrofits where tested on a 61cm tunnel while the analysis in Chapter 2 are done on the longer tunnel, the differences between the models have already been discussed. The values are plotted in Figure 3.3 210 200 190 Temperature (F) 180 170 160 150 140 Right Panel Baseline Left Panel Baseline Right panel with Config A Right Panel with Config B 130 120 Front Slot Figure 3.3 Effects of Rib turbulators on the temperature of the jet stream 58 3.2. INNER SHELL RE-DESIGN CHAPTER 3. Design Modifications The top red line represents the temperature of the left panel slot and is the target temperature to be reached by the jet stream out of the slot in the right panel. The blue line at the bottom is the baseline temperature of the slot in the right panel. The two lines in between are the temperatures of jet stream out of the right panel slots with the given turbulator configuration. It is seen that the turbulence creation enhances the heat transfer that is seen as a rise in temperature of the jet stream. Also it could be noted that the temperature correspond to the length of the slots. The green line which represents the single rib “configuration A” drops mid way through the length of the slot , this is because the slot runs only to this length while the blue line that represents “configuration B” is constantly above the baseline throughout as the ribs run throughout the length of the slot. From the above results we see that the turbulators prove to be very useful tool to increase the temperature where it is required. While this proves to be a viable option it has a few shortcomings associated with it. • It is difficult to design turbulators that would work for complex slot designs. • By increasing the turbulence we compromise on the directionality of the jet stream. Thus in the other retrofits considered an uniform temperature distribution is sorted by decreasing the turbulence in regions of high turbulence instead of increasing the turbulence for steady flows. 3.2 Inner Shell Re-Design The inner shell was redesigned to prevent the formation of large vortices in the flow field and reduce the turbulence for the whole of the inner shell by gradually decreasing the flow cross section. It is theorized that by reducing the flow cross section the flows would be restricted from transversing across the length of the inner shell. This would break down the vortex and flatten out temperature peaks that are caused by the high turbulence. The design is shown in Figure 3.4 59 3.2. INNER SHELL RE-DESIGN CHAPTER 3. Design Modifications Figure 3.4 Redesigned Inner Shell 3.2.1 Effect on flow Characteristics ˙ = ρAV m Re = v DH ν The inner shell is redesigned such that the Reynolds number and effectively the turbulence of the flow is brought down by decreasing the hydraulic diameter of the duct. But in order to satisfy the continuity relations the velocity must increase with decrease in diameter to keep the mass flow rate constant. In the inner shell however decreasing the diameter also increases the pressure drop in the duct. This implies that the fan has to operate against a higher pressure and thus the mass flow rate is decreased effectively decreasing the mean velocity through the duct [Ste].Thus it is theorized that the Reynolds number would be reduced if the diameter reduces gradually. In fact it is noted during 60 3.2. INNER SHELL RE-DESIGN CHAPTER 3. Design Modifications the testing of this design that the mean velocity dropped from 2.98 to 2.65 m/s (Values got for 61 cm oven) in the front slot of the left panel where the readings where taken.Thus the overall Reynolds number is lowered by this redesign. A CFX simulation is done to visualize the flow in the new design of the inner shell. Figure 3.5 Effect of gradually reducing hydraulic diameter on the flow field in the Inner Shell. It is seen from Figure 3.5 that the size of the vortex formed in the baseline has been greatly reduced in the redesign. On the left column the blue regions represents the turbulent flow that is defected from front panel. This flow causes the formation of vortices as it transverses along the length of the tunnel until they are stopped by the steady flow from the other end. It is seen from the simulation that in the redesign this flow traverses a lesser length before it collides with steady 61 3.2. INNER SHELL RE-DESIGN CHAPTER 3. Design Modifications upstream flow while it happens almost near the region of the fan in the baseline. 3.2.2 Effect on Heat Transfer Characteristics To look at how the change in flow would affect the heat transfer characteristics lets recall the DittusBoelter equation from Chapter-2. 4/5 N u D = 0.023R eD P r 0.4 It is seen that the Nusselt number varies with the Reynolds number raised to the power of 4/5. Thus with the decrease in Reynolds number, the Nusselt number and effectively the heat transfer coefficient would go down.Also recalling from Chapter-2 the equation for the temperature of the bulk flow. =⇒ Tb = Tw − q˙ h In the redesign of the inner shell we intend to decrease the value of Tb by reducing the Reynolds number of the flow which in effect reduces the heat transfer coefficient by lowering the Nusselt Number [PSG08]. The interaction of the various parameters has already been discussed and with an increase in the static pressure of the system which reduces the mass flow rate through the inner shell the complexities involved is increased. We need to depend on experimental testing to see how the reduced turbulence would affect the heat transfer characteristics. The reading are taken at the front slots of the left panel, one of the regions where the turbulence is predominant.In Figure 3.6 The blue and the green give the baseline turbulence intensity and temperature respectively. The red line represents the turbulence intensity of the redesigned inner shell. It can be seen that the turbulence levels have gone down which in turn has reduced the temperature of the impinging jet streams given by the purple line. But the drop in temperature is not as significant as expected and the temperatures of the jet stream from the slots on the right panel are lower that this. The aim of the redesign was to close the gap between the temperature profiles of the mixing jet stream. Though a perfect match of the temperature profiles was not expected a bigger drop would have made it possible to fine tune the design to make the temperatures even. Given there is still a big temperature difference between the two mixing jet streams it is concluded that it is not possible to achieve uniform temperature distribution along the slot length by this method. Though the tem- 62 3.2. INNER SHELL RE-DESIGN CHAPTER 3. Design Modifications 0.6 210 200 0.5 0.4 180 170 0.3 160 0.2 Temperature (F) Turbulence Intensity (m/s) 190 150 140 Turbulence - Baseline 0.1 Turbulence - Redesigned Temperature - Baseline 130 Temperature - Redesigned 0 120 Front Slot - Left Panel Figure 3.6 Temperature and Turbulence Intensity distribution along the length of the front slot of the Left Panel for the Redesigned Inner Shell 63 3.3. PLATES WITH SLITS CHAPTER 3. Design Modifications perature difference could not be corrected this redesign could still be used when high directionality is needed as it significantly reduces turbulence intensity. During the design phase of the retrofit an compromise had to be achieved between designing a shell that would be most optimized for the flow and the economical feasibility of the design. While a aerodynamically optimum design could be made with curved edges and the required length to prevent the formation of the vortex, the dimensional and manufacturing constrains prevent such a design. 3.3 Plates with Slits This retrofit is another attempt to make the flow across the heating coil and along the length of the slots uniform. The plates are placed in the inner shell such that they split the inner shell into two zones. The high pressure zone housing the blower is separated by the plates from the low pressure zone which houses the heating coils. It is predicted that this will isolate the vortex formed in the inner shell to the top blower zone creating a more uniform flow over the heating coils making the temperature and the mean velocity at the slots uniform along its length. Air flows between the two zones through the slits in the plates and it is to be noted that in order to create high pressure in the top blower zone the slit area should be lesser than the area of the slots on the panel. If this area is greater, there is no pressure differences created and the plates will only act as obstacles in the flow creating turbulence. A couple of slit design models were made and the flow through them where simulated, some of these designs are shown in Figure 3.7. The computer model of the baseline effectively predicted the mean flow distribution along the length of the slot and the slits were evaluated based on this criteria. It was found that the triple slot design with three slots of width 0.5 cm each running 90 cm along the length created the most uniform mean velocity distribution. The plate with a single slit created a high velocity steam that ricochet of the base plate creating vortices, this made the mean velocity distribution non-uniform. The results for the single slit configuration is largely chaotic and is plotted for different width (1 cm and 0.5 cm) in the plot in the Appendix. The double slits where found to create an uniform flow but weren’t as effective as the three slit plate. The mean velocity profiles at the slots with the double and triple slit plates are compared with that of the experimental results and are plotted in Figure 3.8. It is seen that the mean velocity distribution is most uniform for the three slit model. The computer model has been shown to predict the results of the baseline and other design modifications with a reasonable levels of accuracy. The only parameter in this simulation that may cause the simulation to fail is the static pressure that the fan can operate against. This value is not 64 3.4. PLENUM REDESIGN CHAPTER 3. Design Modifications known as the fan curves are not available. The backward curved blowers are made to operate at high pressure and it can be assumed that the fan will not fail or drastically reduce its flow with the introduction of the plates. Due to this possibility the true effects of the retrofit can be found only through experiment but this analysis gives a good starting point for the experiments. Figure 3.7 Various Slit Configurations 3.4 Plenum Redesign The air in the tunnel section is drawn into the blower through a single circular opening below it. Even in designs that has a plenum to house the third heating coil, the air is still drawn through a same sized circular hole in the plenum box. This means that all the outside air drawn into the tunnel needs to enter the blower through this opening. It was seen from the analysis of the tunnel section in Chapter-2 that cold spots are formed in the tunnel section due to this cold ambient air as it flows from the tunnel entrance/exit to the blower inlet. The plenum was redesigned such that air is drawn evenly throughout the length of the tunnel section and not just through the opening below the blower wheel. This creates a more uniform temperature distribution in the tunnel section and would make the tunnel more efficient by reducing the cold spots. A couple of plenum design models were made and the flow through them where simulated, some of these designs are shown in Figure 3.9. 65 3.4. PLENUM REDESIGN CHAPTER 3. Design Modifications 5 4.5 Mean Flow Velocity - m/s 4 3.5 3 2.5 Double Slit - Computational Baseline - Computational 2 Triple Slit - Computational 1.5 1 Slot Length - Left Panel Figure 3.8 Comparison of the Mean velocity at the slots of the left panel for various slit configuration with the baseline values 66 3.4. PLENUM REDESIGN CHAPTER 3. Design Modifications Figure 3.9 Various Plenum Designs The design that was most found to give the most uniform temperature distribution was “Plenum Design 4" in Figure 3.9. The rectangular plenum slots draw in air from the whole width of the tunnel section. The varying cross section of the plenum slots, with the smallest slot near the center and the largest at the ends helps create a pressure distribution such that uniform air is drawn in along the length. The flow of the cold ambient air through the tunnel is shown by the 3D streamlines in Figure 3.10. It is seen that the cold air entering into the tunnel is drawn in through the outer plenum slots and thus do not traverse through the tunnel section. This helps in reducing the cold spots of the tunnel and gives a more uniform temperature distribution. 67 3.4. PLENUM REDESIGN CHAPTER 3. Design Modifications The temperature profiles for various cross sections along the length of the tunnel section with the redesigned plenum box is compared with that of the baseline tunnel configuration in Figure 3.11. It is seen that while the profile at the ends of the tunnel are fairly similar, the tunnel with a redesigned plenum has a better temperature distribution as we approach the center. The size of the cold spot decreases at a faster rate than with the baseline tunnel and it gradually vanishes near the center of the tunnel. Figure 3.10 Flow of Cold Ambient Air through the Redesigned Tunnel 68 3.4. PLENUM REDESIGN CHAPTER 3. Design Modifications Figure 3.11 Effect of the new plenum design on the temperature distribution in the tunnel section 69 3.5. CROSS-FLOW FANS 3.5 CHAPTER 3. Design Modifications Cross-flow Fans The design modifications discussed above where efficient in optimizing the current model of the convective heat oven. Exploring new ideas to put forward the best design for the tunnel led to the consideration of cross flow fans to replace the centrifugal fans. This section discusses the concept development and the consideration involved in developing a new design that would incorporate the cross flow fan. Most on the problems discussed in Chapter-2 is due to the unsymmetrical nature of the flow of air caused by the centrifugal fan. Though these effects where mitigated by the design modification developed they were not eliminated. The only way to totally avoid the problem caused by turbulence is to create a new flow by replacing the centrifugal fan with a new setup for air circulation. A cross-flow or tangential fan as shown in Figure 3.12 is usually long in relation to the diameter, so the flow approximately remains two-dimensional away from the ends. Unlike the centrifugal fan, the main flow moves transversely across the impeller, passing the blading twice. The flow within a cross-flow fan may be broken up into three distinct regions: a vortex region near the fan discharge, called an eccentric vortex, the through-flow region, and a paddling region directly opposite. Both the vortex and paddling regions are dissipative, and as a result, only a portion of the impeller imparts usable work on the flow. The cross-flow fans are compact, quiet and can provide high pressure coefficient. While a more detailed analysis is required to fully develop a heat shrink oven with tangential fans, a few of the concepts of the design that might prove constructive to consider are discussed below. The various retrofits developed for the ovens for flow control and heat transfer enhancement could be avoided by developing the new product to encompass all these operation variations of the product in the design. 3.5.1 Fan Sizing A new tunnel is built around the cross-flow fan by using the data available for the requirements of the shrinking tunnel. This allows for tailoring the oven design towards meeting these requirements. The fan was sized such that the required velocity at the panel opening of 4.41 m/s is met. The mass flow rate that needs to be achieved by the fan in order to get this velocity at the opening is given by 70 3.5. CROSS-FLOW FANS CHAPTER 3. Design Modifications Figure 3.12 Cross-Flow or Tangential Fan 71 3.5. CROSS-FLOW FANS CHAPTER 3. Design Modifications Continuity Equation. (ρAv )Sl o t s =⇒ vF a n D i s c h a r g e = (ρAv )F a n D i s c h a r g e = (ρAv )Sl o t s (ρA)F a n D i s c h a r g e (3.1) (3.2) The density of air is not a constant and it varies with temperature. Air can be safely considered to be ideal for our case of high temperature and low pressure (pressure change in cross-flow fan is low) as intermolecular forces become important only at lower temperature or higher pressure. Thus the Temperature vs Density relation can be derived from the Ideal gas equations. The fan is to be sized for its peak load. From Eq. 3.2 it is seen that the velocity required at discharge is inversely proportional to the density at that point. Thus the fan would need to discharge more air when, ρF a n D i s c h a r g e < ρSl o t s This situation never occurs in our model as the heating coil is placed in between, thus for our model the peak load on the fan would be when the air is flowing from the fan discharge to the panel opening with very small or no temperature rise (Heating coils go off when the temperature throughout is at the set value). Thus for peak load calculation the density of air is considered constant ,thus from Eq. 3.2 we have vF a n D i s c h a r g e = (Av )Sl o t s (A)F a n D i s c h a r g e For the consideration of fan sizing the size of the model is kept unchanged and thus the areas enclosed by the panels is the same as for the centrifugal fan model.For a cross flow fan the cross section of flow is the same as its outlet discharge area (6.35 x 91.44 cm) , the outlet of the cross flow fan would be discharges this cross sectional area, Thus A F a n D i s c h a r g e = 0.058m 2 and as calculated in Chapter-1 A Sl o t s = 0.01435m 2 vF a n D i s c ha r g e = 0.01435 × 4.41/0.058 = 1.09m /s 72 3.5. CROSS-FLOW FANS CHAPTER 3. Design Modifications The volume of air that need to be discharged by the cross-flow fan is given by, Vf a n = vF a n d i s c h a r g e × C SF l o w = 1.09 × 0.058 = 0.0632m 3 /s A fan that provides this volume of air is to be chosen to provide the necessary airflow through the oven as was provided by the centrifugal fan.It is to be noted during the selection of the fan that a fan with this volume flow might come in different sizes and the one which would ideally fit into the inner shell is to be chosen. 3.5.2 Scope of the Tangential Oven An oven designed around a tangential fan has reduced turbulence and a more uniform mean velocity distribution directionality. This new flow field provides for scope of the design that could be revisited while taking this concept to the design stage. These are discussed below. 1. Finned heating coils could be considered. Though the finned elements have a lower element temperature they enhance the heat transfer by increasing the cross sectional area exposed to the flow when compared to tubular heaters. While the complexities involved in the heat transfer optimization has been discussed in the thesis it would be interesting to study which heating coil would be a best fit for the flow through a cross-flow fan. 2. Without a twin blower wheel setup the need for the outer shell powered by a separate centrifugal fan becomes debatable. While the electronics could be housed separately the outer panels should be “Safe to touch" for operators. The flow in the outer shell and the cooling it provides has been largely been neglected in our analysis as it doesn’t directly affect the flow in the tunnel section. Further study is required to make proper modification to the outer shell to better fit the cross flow design. There is a huge scope of design improvement if its made possible to bring down the outer temperature to an acceptable value by suitable insulation without a need for air flow. 3. The new design concept gives an additional functionality to the oven. The air flow on each side is independently controllable as they are provided by separate fans. The proper use of this functionality to shrink labels onto complex contours is another area that should be analyzed. 73 3.6. CONCLUSION CHAPTER 3. Design Modifications 4. Now that the fan has been sized its optimum position inside the inner shell is to be determined. Recirculation of hot air is a significant functionality of the oven given that it operates at high temperatures. Thus the fans have to posited such that optimal amount of hot air is recirculated. This was one of the primary reasons for a double tangential fan design as its difficult to recirculate the air evenly by running the oven with a single tangential fan. 3.6 Conclusion In an effort to develop a more preferable flow field for the convective heat shrink oven the nature of the flow and its properties where scrutinized using analytical models and computer simulations. This made it possible to identify the unique problems associated with the fluid and thermal characteristics of the oven that hindered the optimized operation of its various components. Design modifications ideas were developed to correct the flow field in the inner shell and the tunnel section of the oven. These ideas were validated using the computer model and tested by experimentation. While it was possible to bring down the non uniformity in the flow characteristics by making the design modifications it was not possible to totally eradicate the problems associated with the oven using the current setup to recirculate air. Thus in the final section of this thesis a conceptual model of the oven that uses cross-flow fans to circulate the air is suggested. The inherent problems of the oven are theorized to be overcome by this new setup because of the uniform nature of the outlet flow in a cross flow oven. While the design modifications where developed and tested on a heat shrink oven, the concept behind them are universal and could be applied to any thermal setup that has similar needs. Hot air jets are required for various applications such as the soldering of electronic components to printed circuit boards and to bake food products. All of them use similar setup to create hot zones inside a control space and the design modifications can be used in these ovens to get a more preferable flow field. 74 REFERENCES [Ct] 2 - 20 Amp split-core AC current sensor - CTV-A - User Manual. Onset Computer Corporation. [F90] AccuSense F900 Air Velocity and Air Temperature Sensor - User Manual. DegreeC. [Ala14] Alam, T. et al. “Heat and flow characteristics of air heater ducts provided with turbula˚ review”. Renewable and Sustainable Energy Reviews 31.C (2014), pp. 289–304. torsUA [BRMy ] Bruce R. Munson Alric P. Rothmayer, T. H. O. W. W. H. Fundamentals of Fluid Mechanics. Wiley, (May 15, 2012). [Dow10] Dowell, D. “A Critical Look at Type T Thermocouples in Low-Temperature Measurement Applications”. English. International Journal of Thermophysics 31.8-9 (2010), pp. 1527– 1532. [DG] Dr. G.Biswas Dr. S.Sarkar, D. S. S. Module 7 - Fans and Blowers. Indian Institute of Technology Kanpur. URL: http://nptel.ac.in/courses/Webcourse-contents/IITKANPUR/machine/ui/Course_home-9.htm. [Ene] HOBO Energy Logger - Data Logger & Modules - User Guide. Onset Computer Corporation. [Men94] Menter, F. R. “Two-equation eddy-viscosity turbulence models for engineering applications”. AIAA Journal 32.8 (1994), pp. 1598–1605. [Men03] Menter, F. R. et al. “Ten Years of Industrial Experience with the SST Turbulence Model”. Ed. by Hanjalic, K. et al. Begell House, Inc., 2003. [Min] Minor loss coefficient diagrams for air ductwork, bends, expansions, inlets and outlets - SI units. URL: http://www.engineeringtoolbox.com/air-duct-minor-lossdiagram-d_332.html. [Mul06] Muller, P. The Equations of Oceanic Motions. Cambridge University Press, 2006. [Nepll] Nepf, H. Transport Processes in the Environment - Velocity profiles and turbulence. Massachusetts Institute of Technology. Fall 2008. URL: http://ocw.mit.edu/courses/ civil-and-environmental-engineering/1-061-transport-processes-inthe-environment-fall-2008. [PSG08] Piyush Sabharwall, V. U. & Gunnerson, F. “Effect Of Mass Flow Rate On The Convective Heat Transfer Coefficient: Analysis For Constant Velocity And Constant Area Case”. Nuclear technology A. 2009, vol. 166, nˇr 2, pp. 197-200 [4 pages] [bibl. : 6 ref.] (2008). 75 REFERENCES REFERENCES [Ste] Stevens, M. AMCA Publication 201 - Fans and Systems. Air Movement and Control Association. [Tog13] Togun, H. et al. “A CFD study of turbulent heat transfer and fluid flow through the channel with semicircle rib”. Clean Energy and Technology (CEAT), 2013 IEEE Conference on. 2013, pp. 312–316. 76 APPENDIX 77 APPENDIX A ADDITIONAL PLOTS 78 APPENDIX A. ADDITIONAL PLOTS 250 6 5.5 230 5 210 4.5 3.5 m/s Fahrenheit 4 190 170 3 2.5 150 2 Temperature 130 Mean Velocity 1.5 110 1 Slot Length - Right Panel Figure A.1 Dependence of Temperature with Mean Velocity at the slots of the Right panel 79 APPENDIX A. ADDITIONAL PLOTS 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 Left Panel 2 Right Panel 1.5 1 Time(s) Probe Figure A.2 Turbulence fluctuation with time at 1/2 of the length of the front slot 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 2 Left Panel Right Panel 1.5 1 Time(s) Probe Figure A.3 Turbulence fluctuation with time at end of the front slot 80 APPENDIX A. ADDITIONAL PLOTS 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 Left Panel 2 Right Panel 1.5 1 Time(s) Probe Figure A.4 Turbulence fluctuation with time at start of the rear slot 5.5 5 4.5 Velocity (m/s) 4 3.5 3 2.5 Left Panel 2 Right Panel 1.5 1 Time(s) Probe Figure A.5 Turbulence fluctuation with time at 1/2 of the rear slot 81