Prediction and measurement of heat transfer in a switched reluctance generator

This research is an experimental investigation of the convective film coefficients on the unique pole faces of a high speed rotor operating within a generator. The specific geometry studied is that of a switched reluctance generator under development by the United States Air Force for use on jet aircraft as an integrated power unit (IPU) to provide auxiliary and emergency power to the aircraft's electrical system. The generator consists of a four pole rotor, nominally 4.30 inches in diameter, housed within a stator assembly. The clearance gap between the rotor pole tip and the stator wall is nominally 0.020 inches. The rotor is designed to rotate at 55,000 rpm.; Rotor cooling is achieved by convection to air flowing axially through the rotor cavities. The convective film coefficient for each unique pole face, as well as an overall film coefficient for the four pole faces are inferred from the rate at which insulated test rates cool when subjected to various rates of rotor rotational speed and axial airflows.; The high velocity of the rotor tips in close proximity to the stator wall results in aerodynamic heating of the air in the gap region. Convective film coefficients associated with high speed flow must be referenced to the adiabatic wall temperature to account for the effects of viscous heating. An estimate of the adiabatic wall temperature in the rotor gap region is developed analytically starting with the fundamental case of high-speed, wall-driven flow. These analytical results for wall-driven flow are verified experimentally using a smooth poled rotor. The analytically predicted film coefficients are within 10% of the experimental results.; Experiments are conducted for rotor speeds ranging from 5,000 to 30,000 rpm and axial airflows ranging from 0.036 to 0.073 kg/s (0.08 to 0.16 lbm/s). The convective film coefficients for each pole face as well as average values representing heat loss from the combined faces are presented in terms of Stanton and Nusselt numbers and correlated to two dimensionless parameters: Rotor Reynolds number and the ratio of axial velocity in the rotor cavities to the rotor pole tip speed. The correlations predict the experimental estimates of film coefficient within 15%.