| General informatio : | / c)1980. a)x, 172 P. : b)Illu. ; c)30 Cm. + e)computer file. a)Turbines./Hetherington, H.J, / Supervisor./University of Dundee, / Faculty of Engineering, -Applied Science department./a)Includes bibliographical references./a)Hetherington, H.J, e)Supervisor./a)University of Dundee, b)Faculty of Engineering, b)Applied Science department./A model is presented for solving the system of turbulent forced heat convection in a horizontal square duct with fully developed velocity distribution and developing temperatures. Air is the only fluid considered. Similar algebraic expressions for turbulent viscosity and conductivity were assigned in each of three layers into which the duct cross-section was divided This implies considerable similarity between the velocity and temperature fields. Two sets of thermal boundary condition~ were imposed on the duct walls. Firstly, the ,two vertical duct sides are isothermal and the horizontal ones adiabatic, while secondly, all sides are isothermal. The numerical solution was carried out by the finite element method based on the Galerkin proFess. Triangular elements were used in solving the momentum equation, and tetrahedral elements in the case of the energy equation. The computer program was constructed to enable calculation for any duct length without increasing the computer capacity needed, and with adequate accuracy. Another program obtained the system parameters at any cross-section, by means of the Simpson method. An experimental rig was constructed in accordance with the ’above system, providing a velocity development length, and two isothermal and two adiabatic walls in the heating section. Wall shear velocity, axial velocity, and temperature were measured at the duct exit. These data guided the choice’ of values of all the model assignable constants. W The solution was determined for constant,~briefly, for variable fluid properties. Two alternative techniques for the/a)A model is presented for solving the system of turbulent forced heat convection in a horizontal square duct with fully developed velocity distribution and developing temperatures. Air is the only fluid considered. Similar algebraic expressions for turbulent viscosity and conductivity were assigned in each of three layers into which the duct cross-section was divided This implies considerable similarity between the velocity and temperature fields. Two sets of thermal boundary condition~ were imposed on the duct walls. Firstly, the ,two vertical duct sides are isothermal and the horizontal ones adiabatic, while secondly, all sides are isothermal. The numerical solution was carried out by the finite element method based on the Galerkin proFess. Triangular elements were used in solving the momentum equation, and tetrahedral elements in the case of the energy equation. The computer program was constructed to enable calculation for any duct length without increasing the computer capacity needed, and with adequate accuracy. Another program obtained the system parameters at any cross-section, by means of the Simpson method. An experimental rig was constructed in accordance with the ’above system, providing a velocity development length, and two isothermal and two adiabatic walls in the heating section. Wall shear velocity, axial velocity, and temperature were measured at the duct exit. These data guided the choice’ of values of all the model assignable constants. W The solution was determined for constant,~briefly, for variable fluid properties. Two alternative techniques for the/ Thesis(PH.D.) - University of Dundee. Faculty of engineering. Applied Science department. a)Thesis(PH.D.) - University of Dundee. Faculty of engineering. Applied Science department. a)Text in english, abstracts in arabic and ebglish./ |