Thermal Analysis of a Hot Gas Duct for a High Temperature Gas Cooled Nuclear Reactor by Viram Pandya An Engineering Project Submitted to the Graduate Faculty of Rensselaer Polytechnic Institute in Partial Fulfillment of the Requirements for the degree of MASTER OF ENGINEERING IN MECHANICAL ENGINEERING Approved: _________________________________________________ Ernesto Gutierrez-Miravete, Engineering Project Adviser Rensselaer Polytechnic Institute Hartford, Connecticut December 2010 © Copyright 2010 by Viram Pandya All Rights Reserved ii CONTENTS LIST OF TABLES ............................................................................................................. v LIST OF FIGURES .......................................................................................................... vi LIST OF SYMBOLS ....................................................................................................... vii ACKNOWLEDGMENT .................................................................................................. ix ABSTRACT ...................................................................................................................... x 1. Introduction.................................................................................................................. 1 1.1 Evolution of Nuclear Power ............................................................................... 1 1.2 Early Helium Cooled Reactors .......................................................................... 2 1.3 General Atomics Design .................................................................................... 3 2. Methodology ................................................................................................................ 5 2.1 Assumptions ....................................................................................................... 5 2.2 Critical Radius of Insulation .............................................................................. 5 2.3 Mathematical Model .......................................................................................... 6 2.3.1 Steady State Heat Conduction................................................................ 6 2.3.2 Governing Equations .............................................................................. 6 2.4 Thermal Resistances........................................................................................... 8 2.5 Design Properties ............................................................................................... 9 2.6 Heat Exchanger Theory.................................................................................... 12 2.7 Convective Heat Transfer ................................................................................ 13 2.8 2.7.1 Entry Length......................................................................................... 13 2.7.2 Turbulence............................................................................................ 14 Model Description ............................................................................................ 16 2.8.1 Simplified Model ................................................................................. 16 2.8.2 Real Model ........................................................................................... 17 3. Results........................................................................................................................ 19 3.1 Insulation Thermal Conductivity ..................................................................... 19 iii 3.2 Hot and Cold Stream Exit Temperatures ......................................................... 19 3.3 COMSOL Modeling ........................................................................................ 19 4. Conclusion ................................................................................................................. 26 5. References.................................................................................................................. 27 Appendix A – Thermal Resistances Calculation ............................................................. 29 Appendix B – Heat Exchanger Calculation ..................................................................... 31 Appendix C – Entry-Length Calculation ......................................................................... 32 iv LIST OF TABLES Table 1: HGD Physical Properties ................................................................................... 10 Table 2: HGD Fluid Properties ........................................................................................ 10 Table 3: Helium Properties at 1223 K and 763 K (both at 3 MPa) ................................. 11 Table 4: Alloy 800H Properties ....................................................................................... 11 Table 5: Ceramic Insulation Properties ........................................................................... 12 Table 6: Outlet Temperature Summary ........................................................................... 13 Table 7: COMSOL Simplified Model Dimensions ......................................................... 17 Table 8: Hot and Cold Stream Outlet Temperatures ....................................................... 21 v LIST OF FIGURES Figure 1: Nuclear Power Evolution Timeline Showing Generation Breakdown [18]....... 2 Figure 2: General Atomics Gas-Turbine Modular Helium Reactor (GT-MHR) [16] ....... 4 Figure 3: Cross section of the HGD [6] ............................................................................. 4 Figure 4: Variation of heat transfer rate with radius [17] .................................................. 6 Figure 5: Thermal Resistance Model with Multiple Layers [3] ........................................ 9 Figure 6: Graphical Representation of Time-Averaged and Fluctuating Parameter [1] . 15 Figure 7: COMSOL Simplified Model Illustration ......................................................... 16 Figure 8: Magnified View of Simplified Model Mesh .................................................... 17 Figure 9: COMSOL Real Model Illustration ................................................................... 18 Figure 10: Magnified View of Real Model Mesh ........................................................... 18 Figure 11: Plot of Hot Stream Outlet Temperature vs. Cold Stream Inlet Velocity ....... 20 Figure 12: Real Model – Temperature Through All Layers at HGD Midpoint .............. 22 Figure 13: Simple Laminar Model Close-up at HGD Midpoint...................................... 22 Figure 14: Real Laminar Model Close-up at HGD Midpoint.......................................... 23 Figure 15: Simple Turbulent Model – Temperature Along Separator Centerline ........... 24 Figure 16: Real Turbulent Model – Temperature Along Insulation Centerline .............. 24 Figure 17: HGD Breakdown for Thermal Resistance Setup ........................................... 29 vi LIST OF SYMBOLS k = thermal conductivity (W/m·K) h, hhot, hcold = convection heat transfer coefficient (W/m2·K) rcr = critical radius of insulation (m) r = radius (m) qr = heat flux in radial direction (W) A = area normal to direction of heat transfer (m2) Rt,cond = thermal resistance for conduction (K/W) Rt,conv = thermal resistance for convection (K/W) L = length of cylinder (m) P = Pressure (Pa) T = Temperature (K) T0 = Absolute Temperature = 273 K Cp = specific heat, constant pressure (J/kg*K) C, Chot, Ccold = heat capacity rate (W/K) Cr = heat capacity ratio U = overall heat transfer coefficient (W/m2*K) NTU = number of transfer units ε = effectiveness q, qmax, Qmax = heat transfer rate (W) ṁ = mass flow rate (kg/s) T = temperature (K) Re = Reynolds number µ = dynamic viscosity (kg/m*s) ν = kinematic viscosity (m2/s) α = thermal expansion coefficient (/ºC) ρ = density (kg/m3) θ = angular position (radians) V = velocity (m/s) Dh = hydraulic diameter (m) Pr = Prandtl number vii LIST OF SYMBOLS (continued) Nu = Nusselt number xh = hydrodynamic entry length (m) xt = thermal entry length (m) u = x-component of velocity v = y-component of velocity ū = time-averaged x-component of velocity u’ = fluctuating x-component of velocity F = body force (N) e = internal energy (J) i = enthalpy (J) τ = shear stress (Pa) x = axial direction (direction going through the HGD) vr = velocity in the radial direction (m/s) vθ = velocity in the θ direction (m/s) viii ACKNOWLEDGMENT I would like to thank my family and friends for their support and encouragement throughout the semester. I would also like to thank Professor Gutierrez and the entire RPI staff for their assistance and persistence to continue challenging myself and to stay on track. ix ABSTRACT Current designs of high temperature helium cooled nuclear reactors call for two different pressure vessels: one which contains the nuclear core to produce heat and one to convert the heat to electricity or some other useful form. Connecting these two vessels is a coaxial tube, commonly called the hot gas duct. The inner tube of the hot gas duct transports hot helium from the core to the power conversion vessel while the outer tube takes helium cooled from the various heat exchangers back to the reactor. This project describes a model developed to carry out a thermal analysis on this duct using both advanced heat transfer concepts and finite element computer software to model the hot gas duct of a currently operating experimental reactor. It examines the thermal insulation necessary to minimize heat losses and the various layers within the duct to compensate for large temperature gradients and thermal expansion. In addition, the thermal conductivity of the insulation is calculated along with exit temperatures of both the hot and cold streams. x 1. Introduction Although dormant for over 30 years, reactor design and development has been ramping up recently in anticipation of the nuclear renaissance: the revival of the nuclear power industry to meet today’s demand for green energy. Many people are hesitant about nuclear power due to the radioactive waste that is produced and required to be stored underground. However, many of these people agree that supplying the world’s energy demand without carbon emissions is impossible with renewable energy sources alone. Recent advances in nuclear reactor design may have the answer to a cleaner and safer commercial nuclear power industry. 1.1 Evolution of Nuclear Power The timeline of commercial nuclear power is broken up into numerous generations as shown in Figure 1. The earliest research reactors and prototypes consisted of Generation I in the 50’s and 60’s followed by Generation II reactors built up to the end of the 90’s. Much of the nuclear reactors we are familiar with today such as the Pressurized Water Reactor (PWR) and the Boiling Water Reactor (BWR) are Generation II. Generation III and III+ have already been designed and many are currently being built or have been approved to be built in countries such as Japan, France and China. These reactors incorporate advances in nuclear technology such as passive safety systems, more fuel burn up resulting in less waste at shutdown, and increased thermal efficiency. All these improvements result in a reactor with a longer lifetime and smaller waste profile. Even as Generation III reactors are being built, Generation IV reactors are underway in different phases from research and development to conceptual design, incorporating advances such as the ability to use nuclear waste as fuel, greater energy extraction from a given amount of fuel, and improved safety systems. Advances in the construction of commercial plants are also being incorporated such as the use of steel plate construction and modularization techniques. A popular Generation IV reactor design is the High Temperature Gas-cooled Reactor (HTGR) which is graphite moderated and helium cooled. One of the biggest advantages of the HTGR is the ability to obtain core outlet temperatures as high as 1 1000°C. This high temperature justifies uses other than electricity production such as process heat for hydrogen production, desalination and oil refining. Additionally, the reactor is designed using only passive safety systems. Even in the event of a loss of coolant casualty, which in today’s plants is a severe problem causing all sorts of safety systems to trip online, the HTGR would be able to remove heat from the core simply using natural convection and conduction. Figure 1: Nuclear Power Evolution Timeline Showing Generation Breakdown [18] 1.2 Early Helium Cooled Reactors The first helium cooled nuclear reactor to produce electricity was built at the Peach Bottom Atomic Power Station in Pennsylvania and operated from 1967 to 1974 [19]. Two helium reactors were built a few years prior in England and Germany and operated as experimental reactors. The Peach Bottom unit was very successful, producing a maximum of 40 MW to the grid at a thermal efficiency of approximately 39% [11]. In 1976 the Fort St. Vrain Generating Station came online in northern Colorado. This was also a high temperature helium cooled reactor which operated until 1989. Although generally successful near the end of its commercial life, this design was plagued by problems initially. The biggest problem originated in the complex steam turbine helium circulators. The water lubricated bearing design of the circulators allowed 2 water to enter the inert environment and create devastating corrosion issues [20]. However, both the Peach Bottom and Fort St. Vrain reactors proved that a helium cooled, graphite moderated nuclear reactor was not only viable, but inherently safe as well. Although they were generally successful, the pressurized water reactor was chosen over its helium cooled counterpart mainly due to Captain Hyman Rickover’s decision to use the pressurized water reactor as a prototype for naval submarines, claiming they were simpler and closer to maturity. The helium cooled reactor was put on the shelf until its recent emergence to lay the groundwork for the next generation of reactors. 1.3 General Atomics Design A General Atomics proposed design for an HTGR is shown below in Figure 2. The single tube connecting the two vessels is the hot gas duct (HGD), which is a coaxial pipe. A typical cross section of the HGD is shown in Figure 3. Hot helium from the core runs through the liner tube while cooler helium from the power production vessel runs through the pressure tube. The elevated temperatures and moderate pressures up to about 4 MPa that the HGD experiences also create significant stresses and thermal gradients. In addition, insulation must be added between the hot and cold tubes to minimize heat losses. Flow characteristics must also be determined to account for convective effects. Insulation is also used to ensure the pressure tube is kept below its maximum allowable temperature. The proper materials must be purchased or engineered, as in the case of the Japan Atomic Energy Agency (JAEA) to build the High Temperature Test Reactor (HTTR). 3 Hot gas duct Figure 2: General Atomics Gas-Turbine Modular Helium Reactor (GT-MHR) [16] Figure 3: Cross section of the HGD [6] 4 2. Methodology 2.1 Assumptions Since a thermal analysis of a coaxial tube with turbulent flow conditions is complex, numerous assumptions were made to help simplify the problem. Assume: Both hot and cold helium streams are smooth pipes with a negligible friction factor. When performing hand calculations, the case of constant heat rate per unit of tube length was used. This is in contrast to the constant surface temperature case. There is no natural convection. Steady state conditions for both hot and cold streams. All material and fluid properties are constant. 2.2 Critical Radius of Insulation For radial systems there exists an optimal insulation thickness. This can be proven by the fact that there are competing effects when increasing insulation. The first effect is that the conduction resistance increases when adding insulation. The competing effect is that as insulation is added, the total surface area of the system increases, causing the convection resistance to decrease. Thus there must be an optimal insulation thickness that minimizes heat loss. This is called the critical radius of insulation and is given by the following equation: 𝑟𝑐𝑟 = Where 𝑘 ℎ (1) k = thermal conductivity (W/m·K) h = convection heat transfer coefficient (W/m2·K) Above this critical radius, the heat flux decreases with increasing thickness of insulation while below it, the heat flux increases up to the critical radius as Figure 4 below illustrates. 5 Figure 4: Variation of heat transfer rate with radius [17] 2.3 Mathematical Model 2.3.1 Steady State Heat Conduction The heat equation for steady-state conditions with no heat generation in a cylinder is: 1𝑑 𝑑𝑇 (𝑘𝑟 ) = 0 𝑟 𝑑𝑟 𝑑𝑟 (2) The appropriate form of Fourier’s law for a cylindrical surface is: 𝑞𝑟 = −𝑘𝐴 Where 𝑑𝑇 𝑑𝑟 (3) r = radius (m) qr = heat flux in radial direction (W) A = area normal to direction of heat transfer (m2) 2.3.2 Governing Equations The continuity, momentum and energy equations of a compressible fluid in vector form are shown below: 6 Continuity: 𝐷𝜌 + 𝜌(𝛻 ∙ 𝑉) = 0 𝐷𝑡 (4) Momentum: 𝐷𝑉 = −∇𝑃 + ∇2 𝑉 + 𝜌𝐹 𝐷𝑡 (5) 𝐷 𝑉2 (𝑒 + ) + ∇ ∙ 𝑃𝑉 + ∇ ∙ (𝜏 ∙ 𝑉) − ∇ ∙ 𝑘∇𝑇 = 𝜌(𝑉 ∙ 𝐹) 𝐷𝑡 2 (6) 𝜌 Energy: 𝜌 Where V is the fluid velocity, F is a body force, e is the internal energy and τ is the shear stress. Applying the assumption that there is steady state flow cancels the time derivative terms. Other reasonable assumptions which can be made include no body forces, no heat generation or additional mechanical work, and no diffusion. Using cylindrical coordinates and keeping only velocity in the axial direction since it is the only significant component (the derivatives of vr and vθ with respect to r are very small) simplifies the equations to those shown below. Note that further simplification of Equation 9 will lead to the heat equation shown in Equation 2. Continuity: 𝜕𝑢 1 𝜕 (𝑟𝑣) = 0 + 𝜕𝑥 𝑟 𝜕𝑟 (7) Momentum: 𝜕𝑢 𝜕𝑢 𝑑𝑃 1 𝜕 𝜕𝑢 + 𝜌𝑣 =− + (𝑟𝜇 ) 𝜕𝑥 𝜕𝑟 𝑑𝑥 𝑟 𝜕𝑟 𝜕𝑟 (8) 𝜕𝑖 𝜕𝑖 1 𝜕 𝜕𝑇 𝜕𝑢 𝑑𝑃 + 𝜌𝑣 − (𝑟𝑘 ) − 𝜇( )2 − 𝑢 =0 𝜕𝑥 𝜕𝑟 𝑟 𝜕𝑟 𝜕𝑟 𝜕𝑟 𝑑𝑥 (9) 𝜌𝑢 Energy: 𝜌𝑢 Where u and v are the fluid velocities in the x and y axes respectively, i is the enthalpy, µ is the dynamic viscosity and P is the pressure. 7 2.4 Thermal Resistances For the case of one dimensional heat transfer with no internal energy generation and constant properties, thermal resistances can be used to model the conduction of heat, similar to electrical resistances for electrical charge. The resistances can be added in series in the case of multiple layers as shown in Figure 5. Refer to Appendix A for the thermal resistance calculation performed to determine the thermal conductivity of the insulation in the HGD. The thermal resistances for conduction and convection in a cylindrical wall are as follows: 𝑅𝑡,𝑐𝑜𝑛𝑑 = 𝑟 ln(𝑟2 ) 1 2𝜋𝐿𝑘 (10) 1 ℎ𝐴 (11) 𝑅𝑡,𝑐𝑜𝑛𝑣 = Where Rt,cond = thermal resistance for conduction (K/W) Rt,conv = thermal resistance for convection (K/W) L = length of cylinder (m) 8 Figure 5: Thermal Resistance Model with Multiple Layers [3] 2.5 Design Properties The physical dimensions of the hot gas duct were extracted from [6]. These are the dimensions of the planned very high temperature gas cooled reactor (VHTR) being researched at the Korea Atomic Energy Research Institute (KAERI) since data on the General Atomics design could not be obtained. A study in [6] sizes the inner diameter of the HGD pressure vessel according to three situations where the flow velocity, flow rate, or dynamic pressure of the hot helium is the same as the cold. We use the case of identical flow rates. Also, although the length of the HGD is not specifically given, we use information about the distance between the reactor and power production vessels to make an educated assumption on its length. Tables 1 and 2 show the design parameter values used in this study. 9 Table 1: HGD Physical Properties Inner radius (m) Outer radius (m) Hot Helium Area (m2) Thickness (m) 0.386 0.468 Liner Tube 0.386 0.393 0.007 0.017 Insulation 0.393 0.513 0.120 0.342 Inner Tube Annulus (Cold Helium) 0.513 0.523 0.010 0.033 0.523 0.626 0.103 0.372 Pressure Tube 0.626 *Length of HGD = 8.0 m 0.691 0.065 0.269 Table 2: HGD Fluid Properties Hot Helium Temp (K) 1223 Cold Helium Temp (K) 763 Pressure (both) (MPa) 3 Flow rate (both) (kg/s) Flow Velocity Hot Helium (m/s) Flow Velocity Cold Helium (m/s) 84 65 82 Helium properties vary significantly with temperature and pressure. The following temperature and pressure dependent equations from [10] are calculated at 1223 K and 763 K and 3 MPa in Table 3. The specific heats at constant pressure and constant volume do not vary significantly. −1 𝜌 (𝐷𝑒𝑛𝑠𝑖𝑡𝑦) = 0.17623 ∗ 𝑃 𝑇⁄ 𝑇0 [1 + 0.53 ∗ 10−3 ∗ 𝑃 1.2 ] (𝑇⁄𝑇 ) 0 𝑘𝑔 𝑚3 𝑇 0.7 𝑘𝑔 𝜇 (𝐷𝑦𝑛𝑎𝑚𝑖𝑐 𝑉𝑖𝑠𝑐𝑜𝑠𝑖𝑡𝑦) = 1.865 ∗ 10−5 ∗ ( ) 𝑇0 𝑚∗𝑠 (12) (13) 𝑘 (𝑇ℎ𝑒𝑟𝑚𝑎𝑙 𝐶𝑜𝑛𝑑. ) −4 ∗𝑃) = 0.144(1 + 2.7 ∗ 10 −4 𝑇 0.71∗(1−2∗10 ∗ 𝑃) ( ) 𝑇0 10 𝑊 𝑚∗𝐾 (14) Where P = Pressure (bar) T = Temperature (K) T0 = Absolute Temperature = 273 K Table 3: Helium Properties at 1223 K and 763 K (both at 3 MPa) at 1223 K at 763 K 3 1.18 1.88 Dynamic Viscosity (kg/m*s) 5.33E-05 3.83E-05 5195 5195 0.418 0.300 Density (kg/m ) Specific Heat, const pressure (J/kg*K) Thermal conductivity (W/m*K) The tubes of the HGD are made of Alloy 800H- a high nickel content (32%) alloy that is highly resistant to oxidation and corrosion at elevated temperatures. The required properties of Alloy 800H from [13] are listed in Table 4. The average thermal conductivity value is the arithmetic average between the values at 1223 K and 763 K. Table 4: Alloy 800H Properties Density (kg/m3) 8030 Specific Heat, const pressure (J/kg*K) Thermal conductivity (W/m*K) 500 Coeff. Thermal expansion (m/ºC) 30.8 at 1223 K 24.55 Average 18.3 at 763 K 1.78E-05 The insulation within the HGD is made of a fibrous ceramic insulation material mainly consisting of silicon dioxide (SiO2) and alumina (Al2O3). The fibers are wrapped in a stainless steel net to prevent it from entering the helium stream [9]. Commercial ceramic insulation products consisting of alumina fibers were found from [14] and its properties are listed in Table 5 below. 11 Table 5: Ceramic Insulation Properties Density (kg/m3) Specific Heat, const pressure (J/kg*K) Thermal conductivity (W/m*K) 192 1130 0.150 at 811 K 0.235 at 1033 K 0.320 at 1255 K 2.6 Heat Exchanger Theory Although they have opposite goals, the HGD and a counter flow coaxial heat exchanger are almost identical. Since the HGD has insulation between its two streams, its purpose is to hold in heat rather than exchange it. Because of this similarity, heat exchanger theory can be used to analytically determine the exit temperatures of the hot and cold streams by employing the effectiveness-NTU method. First, the heat capacity rate of each stream must be calculated and the stream with the lower value is considered Cmin. 𝐶 = 𝑚̇ ∗ 𝐶𝑝 (15) 𝐶𝑚𝑖𝑛 𝐶𝑚𝑎𝑥 (16) 𝐶𝑟 = We then calculate the overall heat transfer coefficient, U, to determine the number of transfer units, NTU. Table 11.3 from [2] tabulates effectiveness equations in terms of NTU. Once the effectiveness of the heat exchanger is known, the actual heat transfer rate can be determined along with exit temperatures using Equations 20 through 22. 𝑈= 1 1 1 + ℎ𝐻𝑜𝑡 ℎ𝐶𝑜𝑙𝑑 𝑁𝑇𝑈 = 𝜀= 𝑈𝐴 𝐶𝑚𝑖𝑛 𝑁𝑇𝑈 1 + 𝑁𝑇𝑈 12 (17) (18) (19) 𝑞 = 𝜀 ∗ 𝑞𝑚𝑎𝑥 = 𝜖 ∗ 𝐶𝑚𝑖𝑛 ∗ (𝑇𝐻𝑜𝑡 − 𝑇𝐶𝑜𝑙𝑑 ) 𝑇𝐻𝑜𝑡,𝑜𝑢𝑡𝑙𝑒𝑡 = 𝑇𝐻𝑜𝑡,𝑖𝑛𝑙𝑒𝑡 − (20) 𝑞 𝑚̇ ∗ 𝐶𝑝 (21) 𝑞 𝑚̇ ∗ 𝐶𝑝 (22) 𝑇𝐶𝑜𝑙𝑑,𝑜𝑢𝑡𝑙𝑒𝑡 = 𝑇𝐶𝑜𝑙𝑑,𝑖𝑛𝑙𝑒𝑡 + See Table 6 for a summary of results and Appendix B for detailed calculations. Table 6: Outlet Temperature Summary Cmin (W/K) U (W/m2*K) NTU ε qmax (W) Hot stream outlet temp (K) Cold stream outlet temp (K) 436380 720 0.032 0.031 2.00E+08 1209 777 2.7 Convective Heat Transfer The sheer volume and velocity of helium flowing through the HGD is a substantial amount which follows certain velocity and thermal profiles. Since both the hot and cold flows are turbulent, there is significant “mixing” due to random motion which tends to increase the heat transfer rate. In addition, the turbulence of the streams creates a boundary layer with the surface of the pipes where viscous effects cannot be neglected. 2.7.1 Entry Length The development of the boundary layer is caused by the no-slip boundary condition which says the velocity of the fluid at the pipe wall must be zero. In this boundary layer, the velocity profile varies with both the x and y axes where the x-axis is the axial coordinate along the pipe length and the y-axis is the coordinate from one inner pipe surface to the other. Viscous effects must also be taken into account within the 13 boundary layer. The boundary layer grows in thickness until it has completely filled the pipe. The point along the x-axis where this occurs is called the hydrodynamic entry length. Beyond this, the flow is considered to be fully developed and the velocity profile varies with y only. The same situation occurs with the temperature of the fluid flowing through the pipe. The temperature of the fluid very close to the pipe must be equal to the surface temperature of the pipe. In this case a thermal boundary layer develops until it reaches the thermal entry length. The hydrodynamic and thermal entry lengths are different for laminar and turbulent flow except for flows where the Prandtl number is equal to one. In this case the hydrodynamic boundary layer develops at the same rate as the thermal boundary layer. The equations for turbulent flow are given below as they will represent the HGD. 𝑥ℎ = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ (23) 𝑥𝑡 = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ ∗ 𝑃𝑟 (24) From Appendix C it is shown that the only stream which reaches any type of fully developed flow is the cold stream. This stream becomes thermally fully developed after travelling 5.7 meters through the pipe. 2.7.2 Turbulence The transition to turbulence occurs when laminar flow becomes unstable due to minor disturbances. When the disturbance, such as the effects of a rough pipe, occurs, the velocity changes and the inertia forces associated with this velocity change create instability by magnifying the disturbance [15]. The turbulent flow undergoes random fluid particle fluctuations and contains unsteady velocity components in all three axes. Due to this complexity, most turbulent flow problems can only be solved using software. Due to the random nature of turbulence, fluids undergoing turbulent flow involve fluctuating parameters, such as velocity. The parameters must be broken down into a mean value and a fluctuating value as shown in Figure 6. Thus, the value of the parameter, u, at some time t is equal to: 𝑢(𝑡) = 𝑢̅ + 𝑢′ 14 (25) This is called the Reynolds average of the parameter where ū is the time-averaged (mean) value and u’ is the fluctuating value. The Reynolds average of u(t) is defined as: 𝑡0 +𝑡 1 𝑢̅(𝑡) = ̅̅̅̅̅̅̅̅ 𝑢̅ + 𝑢′ = lim ∫ 𝑡→∞ 𝑡 𝑡 0 1 𝑡0 +𝑡 (𝑢̅ + 𝑢′ )𝑑𝑡 = 𝑢̅ + lim ∫ 𝑢′𝑑𝑡 = 𝑢̅ 𝑡→∞ 𝑡 𝑡 0 Where t0 to t is a length of time that is used to integrate out the turbulent fluctuations in the averaging process. Reynolds averaging can also be applied to the Navier-Stokes and energy equations as shown below. Reynolds –averaged Navier-Stokes equation: 𝑢̅ 𝜕𝑢̅ 𝜕𝑢̅ 1 𝜕𝑃̅ 𝜕 𝜕𝑢̅ + 𝑣̅ =− + (𝜈 − ̅̅̅̅̅̅ 𝑢′ 𝑣 ′ ) 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜕𝑦 𝜕𝑦 (26) Reynolds-averaged energy equation: 𝑢̅ 𝜕𝑇̅ 𝜕𝑇̅ 𝜕 𝜕𝑇̅ + 𝑣̅ = (𝛼 − ̅̅̅̅̅̅ 𝑣 ′𝑇 ′) 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦 (27) In the above equations, ̅̅̅̅̅̅ 𝑢′ 𝑣 ′ is a turbulent shear stress and ̅̅̅̅̅̅ 𝑣 ′ 𝑇 ′ is a turbulent heat flux in the direction normal to the main flow. Figure 6: Graphical Representation of Time-Averaged and Fluctuating Parameter [1] 15 2.8 Model Description The model of the HGD was built using the Finite Element Model (FEM) in COMSOL Multiphysics. Two different orientations were used when modeling the HGD: simplified and real. The simplified model underwent both laminar and turbulent flow analyses while the real model only underwent turbulent analysis, and in each case the thermal boundary condition at the end of the pressure tube was an inward heat flux of 0 W/m2. 2.8.1 Simplified Model In the simplified model, the hot and cold streams were approximately equal in area and were separated by a very thin piece of Alloy 800H, herein called the separator. Another piece of Alloy 800H surrounded the entire model from the outside, acting as the pressure tube. The length of this model was 4 meters compared to the actual length of 8. The simplified model was used to establish the correct laminar and turbulent conditions and served as a baseline to compare the real model. Figure 7 and Table 7 illustrate the simplified model and its dimensions. The mesh used in COMSOL for the simplified model consists of 5504 elements and is shown in Figure 8. Cold stream Axis of symmetry Pressure tube Hot stream Separator Figure 7: COMSOL Simplified Model Illustration 16 Table 7: COMSOL Simplified Model Dimensions Thickness (m) 0.386 0.010 0.358 Component Hot stream Separator Cold stream Pressure tube 0.060 Figure 8: Magnified View of Simplified Model Mesh 2.8.2 Real Model The real model is a replica of the actual HGD with identical dimensions to that of Table 1. All the tubes were added with their proper radius as well as the layer of insulation. Properties used for the insulation were that of alumina ceramic fiber taken from Table 5. This is the closest material with published properties since the data for the insulation used on the HGD is unavailable. This model, as shown in Figure 9, only underwent turbulent analysis since this is the realistic condition of the HGD. The mesh for the real model consists of 2300 elements and is shown in Figure 10. It has significantly fewer elements compared to the simplified model due to the fact that the real model has more layers, many of which are very thin, as well as the increased length of the HGD. 17 Axis of symmetry Cold stream Hot stream Pressure tube Liner tube Insulation Inner tube Figure 9: COMSOL Real Model Illustration Figure 10: Magnified View of Real Model Mesh 18 3. Results 3.1 Insulation Thermal Conductivity As calculated in Appendix A, the thermal conductivity of the insulation is 1.95 W/m*K using the method of thermal resistances. Reference [9] calculates the thermal conductivity to be 0.47 W/m*ºC through thermal tests. This difference results in a large error percentage. The most likely reason for this large error is because [9] calculates the thermal conductivity through thermal tests that model the HGD of the HTR-10 instead of the KAERI design. HTR-10 is an experimental helium cooled reactor based in China whose layout and design parameters are different. Nonetheless, the aluminum oxide fiber insulation is identical. 3.2 Hot and Cold Stream Exit Temperatures The extremely low effectiveness calculated in Appendix B makes sense since the presence of insulation will significantly reduce the amount of heat that will be lost by the hot stream, making for an ineffective heat exchanger. Through the 8 meter length of the HGD, the hot helium decreases 14 K which is gained by the cold helium. In reality, the cold helium temperature rise will not be as much since the insulation will contain most of the heat. 3.3 COMSOL Modeling Figure 11 below shows the variation of the hot stream outlet temperature as the cold stream inlet velocity is varied between 30 m/s and 300 m/s. In all cases, the inlet temperatures for the hot and cold stream were 1223 K and 763 K respectively. The hot stream inlet velocity was kept constant at 65 m/s. To determine the outlet temperatures, the Boundary Integration function was used after solving the problem. This resulted in a value of temperature integrated over the area, so the hot stream outlet area of 0.468 m2 was divided out to arrive at the temperature. 19 Hot Stream Outlet Temperature (K) 1224 1222 1220 Simplified Turbulent 1218 Simplified Laminar 1216 Real Turbulent 1214 1212 0 50 100 150 200 250 300 Cold Stream Inlet Velocity (m/s) Figure 11: Plot of Hot Stream Outlet Temperature vs. Cold Stream Inlet Velocity The hot stream outlet temperature is nearly independent of the cold stream inlet velocity for the real model. The thick insulation and large mass flow rates of both streams can account for this. The simplified models exhibit similar behavior although the laminar model does not lose as much energy. This result reinforces the fact that turbulence causes efficient mixing and greater heat transfer. In addition, the outlet temperatures of the hot and cold streams were determined for all models and listed in Table 8 along with the calculated value obtained in Appendix B. Again, COMSOL’s Boundary Integration tool was used to determine the temperature integrated over the area. The outlet temperatures were determined for the normal operating conditions shown in Table 2. 20 Table 8: Hot and Cold Stream Outlet Temperatures Model Simple Laminar Simple Turbulent Real Turbulent Calculated Outlet Temperature (K) Cold Hot Stream Stream 770.24 1214.88 769.81 763.07 777 1218.09 1222.95 1209 The real model exhibits the smallest temperature change in both streams whereas the calculated value undergoes the largest. This result makes logical sense since the real model is highly insulated and the calculated values included many assumptions about the characteristics of the flow in order to use a tabulated Nusselt number. Figure 12 was plotted to illustrate the temperature change over the cross-section of the HGD. The data is taken at the longitudinal midpoint of the HGD at 4 meters and begins at the center of the hot stream, going radially outward to the edge of the pressure tube. A close-up image of the separator in the simplified model and the insulation in the real model is shown in Figures 13 and 14 respectively. 21 Figure 12: Real Model – Temperature Through All Layers at HGD Midpoint Figure 13: Simple Laminar Model Close-up at HGD Midpoint 22 Figure 14: Real Laminar Model Close-up at HGD Midpoint The linear temperature gradient between the hot and cold streams shows the benefit of insulation as it creates a gradual change in temperature, preventing large thermal gradients from forming and causing unnecessary stresses. The images reinforce this point since the simple model has an uneven temperature distribution in the hot stream along the separator. This could cause uneven thermal expansion within the HGD and cause fractures over time. The real model, on the other hand, displays an even temperature reduction from hot to cold streams. Since a large focus of the thermal analysis of the HGD relies on the insulation, plots of the temperature vs. HGD length along the centerline of the insulation or separator are provided in Figures 15 and 16 for the turbulent case. Both plots were taken with the HGD under normal operating conditions. 23 Figure 15: Simple Turbulent Model – Temperature Along Separator Centerline Figure 16: Real Turbulent Model – Temperature Along Insulation Centerline 24 Again one can see why the presence of insulation is beneficial as its temperature varies linearly from inlet to outlet with a temperature difference of about 6 K. Conversely, the separator in the simplified model varies by about 400 K. This shows that the insulation stays at about the same temperature through the HGD while the separator undergoes large fluctuations, reinforcing the value provided by insulation. 25 4. Conclusion The thermal analysis of the HGD was successful in accomplishing the given tasks: calculating the thermal conductivity of the insulation, calculating the exit temperatures and analyzing a finite element model of the HGD. However, it highlights the difficulty in obtaining accurate results when performing hand calculations. Calculation of the thermal conductivity of the insulation and the outlet temperatures of the hot and cold streams was accomplished using thermal resistances and heat exchanger theory respectively. Unfortunately, the proprietary nature of materials used and their properties forced the use of data from two different reactors which skewed results of the conductivity. Additionally, many assumptions had to be made in determining the outlet temperatures which led to greater heat transfer than would normally occur. The COMSOL finite element models proved useful in examining the thermal nature of the HGD. The insulation is superior in preventing heat transfer between the two streams as well as minimizing thermal gradients. It also allows for a uniform thermal expansion throughout the HGD which must be taken into account when constructing the reactor as this section will expand and cannot be constrained on both ends. 26 5. References 1. Munson, Bruce R., Donald F. Young and Theodore H. Okiishi. Fundamentals of Fluid Mechanics. Hoboken: John Wiley & Sons, Inc, 2006. 2. Incropera, Frank P., David P. Dewitt, Theodore L. Bergman and Adrienne S. Lavine. Fundamentals of Heat and Mass Transfer. Hoboken: John Wiley & Sons, Inc, 2007. 3. Bird, R. Byron, Warren E. Stewart and Edwin N. Lightfoot. Transport Phenomena. Hoboken: John Wiley & Sons, Inc, 2002. 4. Cengel, Yunus A., and Michael A. Boles. Thermodynamics: An Engineering Approach. New York: McGraw-Hill, 2006. 5. Hishida, Makoto, Kazuhiko Kunitomi, Ikuo Ioka, et al. “Thermal Performance Test of the Hot Gas Ducts of Hendel.” Nuclear Engineering and Design 83 (1984): 91-103. 6. Song, Kee-nam and Yong-wan Kim. “Preliminary Design Analysis of Hot Gas Ducts for the Nuclear Hydrogen System.” Journal of Engineering for Gas Turbines and Power 131 (Jan 2009). 7. Bröckerhoff, P., J. Singh, H. Schmitt, J. Knaul, et al. “Status of Design and Testing of Hot Gas Ducts.” Nuclear Engineering and Design 78 (1984): 215-221. 8. Inagaki, Yoshiyuki, Kazuhiko Kunitomi, Masatoshi Futakawa, Ioka Ikuo and Yoshiyuki Kaji. “R&D on high temperature components.” Nuclear Engineering and Design 233 (2004): 211-223. 9. Huang, Z.Y., Z.M. Zhang, M.S. Yao and S.Y. He. “Design and experiment of hot gas duct for the HTR-10.” Nuclear Engineering and Design 218 (2002): 137145. 10. Petersen, Helge. “The Properties of Helium.” Danish Atomic Energy Commission (1970). 11. Kingrey, K.I. “Fuel Summary for Peach Bottom Unit 1 High-Temperature GasCooled Reactor Cores 1 and 2.” Idaho National Engineering and Environmental Laboratory (April 2003): 5-9. 12. http://www.cdeep.iitb.ac.in/nptel/Mechanical/Heat%20and%20Mass%20Transfe r/Conduction/Module%202/main/2.6.4.html 13. Sandmeyer Steel Company. http://www.sandmeyersteel.com/A800-A800HA800AT.html 14. MatWeb. “Thermal Ceramics Kaowool Blanket.” http://www.matweb.com/search/datasheet.aspx?MatGUID=cb830e74bc69422aa5 60a7b57494955a 15. Kays, William, Michael Crawford and Bernhard Weigand. Convective Heat and Mass Transfer. New York: McGraw-Hill, 2005. 16. General Atomics website. http://gt-mhr.ga.com/description.php 17. Steady State Conduction Module. http://www.cdeep.iitb.ac.in/nptel/Mechanical/Heat%20and%20Mass%20Transfe r/Conduction/Module%202/main/2.6.4.html 18. U.S. Department of Energy. Generation IV Nuclear Energy Systems. http://www.ne.doe.gov/geniv/neGenIV1.html 27 19. Wikipedia contributors. "Peach Bottom Nuclear Generating Station." Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 18 Nov. 2010. Web. 4 Dec. 2010. 20. Wikipedia contributors. "Fort St. Vrain Generating Station." Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 13 Jul. 2010. Web. 4 Dec. 2010. 28 Appendix A – Thermal Resistances Calculation The thermal resistance method was used to determine the thermal conductivity of the insulation and compare it to the value given in [9]. Figure 17 below illustrates the problem setup. Figure 17: HGD Breakdown for Thermal Resistance Setup 𝑞𝑟𝑎𝑑𝑖𝑎𝑙 = 𝑅𝑇𝑜𝑡𝑎𝑙 𝑇∞,𝐻 − 𝑇∞,𝐶 𝑅𝑇𝑜𝑡𝑎𝑙 𝑟 𝑟 𝑟 𝑙𝑛 𝑟2 𝑙𝑛 𝑟3 𝑙𝑛 𝑟4 1 1 1 2 3 = + + + + 2𝜋𝑟1 𝐿ℎ𝐻𝑜𝑡 2𝜋𝑘𝐴 𝐿 2𝜋𝑘𝑖 𝐿 2𝜋𝑘𝐴 𝐿 2𝜋𝑟5 𝐿ℎ𝐶𝑜𝑙𝑑 (28) (29) Where kA is the thermal conductivity of Alloy 800H and ki is the thermal conductivity of the insulation. In order to determine the convection coefficients (h1 and h5) of the hot and cold helium, the Reynolds number is first calculated to determine if the flow is laminar, turbulent, or in transition. 𝑅𝑒𝐻𝑜𝑡 𝑅𝑒𝐶𝑜𝑙𝑑 𝑘𝑔 𝑚 𝜌𝑉𝐷ℎ 1.18 𝑚3 ∗ 65 𝑠 ∗ 2 ∗ 0.386 𝑚 = = = 1.11 ∗ 106 𝑘𝑔 𝜇 −5 5.33 ∗ 10 𝑚 ∗ 𝑠 𝑘𝑔 𝑚 𝜌𝑉𝐷ℎ 1.88 𝑚3 ∗ 82 𝑠 ∗ (2 ∗ 0.626 − 2 ∗ 0.523)𝑚 = = = 8.29 ∗ 105 𝑘𝑔 𝜇 3.83 ∗ 10−5 𝑚 ∗ 𝑠 29 Therefore, both flows are turbulent. The Prandtl number is also calculated to determine how the velocity and thermal diffusivity are developing. 𝐽 −5 𝑘𝑔 𝜇𝐶𝑝 5.33 ∗ 10 𝑚 ∗ 𝑠 ∗ 5195 𝑘𝑔 ∗ 𝐾 𝑃𝑟𝐻𝑜𝑡 = = = 0.662 𝑊 𝑘 0.418 𝑚 ∗ 𝐾 𝐽 −5 𝑘𝑔 𝜇𝐶𝑝 3.83 ∗ 10 𝑚 ∗ 𝑠 ∗ 5195 𝑘𝑔 ∗ 𝐾 𝑃𝑟𝐶𝑜𝑙𝑑 = = = 0.663 𝑊 𝑘 0.300 𝑚 ∗ 𝐾 Since the Prandtl numbers are approximately equal to 1, the heat and momentum are diffused through the fluid at the same rates and certain simplifications can be made. The Nusselt number for the hot and cold streams is calculated from the Dittus-Boelter equation where n=0.4 for heating and n=0.3 for cooling. 𝑁𝑢𝐻𝑜𝑡 = 0.023 ∗ 𝑅𝑒 4/5 ∗ 𝑃𝑟 𝑛 = 0.023 ∗ (1.11 ∗ 106 )4/5 ∗ (0.662)0.3 = 1394 𝑁𝑢𝐶𝑜𝑙𝑑 = 0.023 ∗ 𝑅𝑒 4/5 ∗ 𝑃𝑟 𝑛 = 0.023 ∗ (8.29 ∗ 105 )4/5 ∗ (0.663)0.4 = 1060 ℎ𝐻𝑜𝑡 ℎ𝐶𝑜𝑙𝑑 𝑊 𝑁𝑢𝐻𝑜𝑡 ∗ 𝑘 1394 ∗ 0.418 𝑚 ∗ 𝐾 𝑊 = = = 755 2 𝐷ℎ 2 ∗ 0.386 𝑚 𝑚 ∗𝐾 𝑊 1060 ∗ 0.300 𝑚 ∗ 𝐾 𝑁𝑢𝐶𝑜𝑙𝑑 ∗ 𝑘 𝑊 = = = 1543 2 𝐷ℎ (2 ∗ 0.626 − 2 ∗ 0.523) 𝑚 𝑚 ∗𝐾 𝑞𝑟𝑎𝑑𝑖𝑎𝑙 = ℎ𝐻𝑜𝑡 ∗ 𝐴 ∗ (𝑇𝐻𝑜𝑡 − 𝑇𝐶𝑜𝑙𝑑 ) = 755 𝑊 ∗ 0.468 𝑚2 ∗ (1223 − 763)𝐾 ∗𝐾 𝑚2 = 162536 𝑊 All the constants are now known except for the thermal conductivity of the insulation (ki). 𝑞𝑟𝑎𝑑𝑖𝑎𝑙 = 𝑇∞,𝐻 − 𝑇∞,𝐶 𝑟2 𝑟3 𝑟4 𝑙𝑛 𝑙𝑛 𝑙𝑛 1 𝑟1 𝑟2 𝑟3 1 + + + + 2𝜋𝑟1 𝐿ℎ𝐻𝑜𝑡 2𝜋𝑘𝐴 𝐿 2𝜋𝑘𝑖 𝐿 2𝜋𝑘𝐴 𝐿 2𝜋𝑟5 𝐿ℎ𝐶𝑜𝑙𝑑 𝑘𝑖 = 1.95 30 𝑊 𝑚∗𝐾 Appendix B – Heat Exchanger Calculation 𝐶𝐻𝑜𝑡 = 𝐶𝐶𝑜𝑙𝑑 = 𝑚 ∗̇ 𝐶𝑝 = 84 𝑘𝑔 𝐽 𝑊 ∗ 5195 = 436380 𝑠 𝑘𝑔 ∗ 𝐾 𝐾 We have calculated hhot and hcold in Appendix A and use these values to calculate U: 𝑈= 𝑞𝑚𝑎𝑥 1 1 1 = 720 𝑊 𝑚2 ∗ 𝐾 1 + 𝑊 𝑊 ℎ𝐻𝑜𝑡 ℎ𝐶𝑜𝑙𝑑 755 2 1543 2 𝑚 ∗𝐾 𝑚 ∗𝐾 𝑊 720 2 ∗ 𝜋 ∗ 2 ∗ 0.386 𝑚 ∗ 8 𝑚 𝑈𝐴 𝑚 ∗𝐾 𝑁𝑇𝑈 = = = 0.032 𝑊 𝐶𝑚𝑖𝑛 436380 𝐾 𝑁𝑇𝑈 0.032 𝜀= = = 0.031 1 + 𝑁𝑇𝑈 1 + 0.032 𝑊 = 𝐶𝑚𝑖𝑛 (𝑇𝐻𝑜𝑡,𝑖𝑛𝑙𝑒𝑡 − 𝑇𝐶𝑜𝑙𝑑,𝑖𝑛𝑙𝑒𝑡 ) = 436380 ∗ (1223 − 763) 𝐾 = 2.00 ∗ 108 𝑊 𝐾 + 1 1 = 𝑞𝑎𝑐𝑡𝑢𝑎𝑙 = 𝜖 ∗ 𝑞𝑚𝑎𝑥 = 0.031 ∗ 2.00 ∗ 108 𝑊 = 6.23 ∗ 106 𝑊 𝑇𝐻𝑜𝑡,𝑜𝑢𝑡𝑙𝑒𝑡 = 𝑇𝐻𝑜𝑡,𝑖𝑛𝑙𝑒𝑡 − 𝑞 6.23 ∗ 106 𝑊 = 1223 𝐾 − = 1209 𝐾 𝑊 𝑚̇ ∗ 𝐶𝑝 436380 𝐾 𝑇𝐶𝑜𝑙𝑑,𝑜𝑢𝑡𝑙𝑒𝑡 = 𝑇𝐶𝑜𝑙𝑑,𝑖𝑛𝑙𝑒𝑡 + 𝑞 6.23 ∗ 106 𝑊 = 763 + = 777 𝐾 𝑊 𝑚̇ ∗ 𝐶𝑝 436380 𝐾 31 Appendix C – Entry-Length Calculation From Appendix A we know the values of the Reynolds number, Prandtl number and hydraulic diameter for the hot and cold streams. Hydrodynamic Entry-Length: 𝑥ℎ,ℎ𝑜𝑡 = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ = 4.4 ∗ (1.11 ∗ 106 )1/6 ∗ 2 ∗ 0.386 𝑚 = 34.6 𝑚 𝑥ℎ,𝑐𝑜𝑙𝑑 = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ = 4.4 ∗ (8.29 ∗ 105 )1/6 ∗ 2 ∗ (0.626 − 0.523) 𝑚 = 8.5 𝑚 Thermal Entry-Length: 𝑥𝑡,ℎ𝑜𝑡 = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ ∗ 𝑃𝑟 = 4.4 ∗ (1.11 ∗ 106 )1/6 ∗ 2 ∗ 0.386 𝑚 ∗ 0.662 = 22.9 𝑚 𝑥𝑡,𝑐𝑜𝑙𝑑 = 4.4 ∗ (𝑅𝑒)1/6 ∗ 𝐷ℎ ∗ 𝑃𝑟 = 4.4 ∗ (8.29 ∗ 105 )1/6 ∗ 2 ∗ (0.626 − 0.523)𝑚 ∗ 0.663 = 5.7 𝑚 32