Measurement of heat transfer in a switched reluctance generator

Abstract

INTRODUCTION This paper presents an analytical and experimental investigation of the convective heat transfer between a cylinder rotating at high speed within a stationary outer cylinder with a small clearance gap. The investigation is the basis of a larger study to determine the convective film coefficients on the unique faces of a four-pole high-speed rotor under development for use within a switched reluctance electric generator. Analytical models for the velocity and temperature distributions of the air in the annular gap are developed. Applying the concepts of high-speed external boundary layer flow, an adiabatic wall temperature is defined that accounts for the effects of viscous heat generation within the gap region. Using these models, convective film coefficients are developed based on the adiabatic wall temperature. The analytical results are compared to experimental data from which the film coefficients were inferred. It is found that the values of Stanton number determined from experimental data are in excellent agreement with values predicted by the theoretical model when the level of shear stress is increased in a manner suggested by experimental windage data. Copyright © 1998 The American Institute of Aeronautics and Astronautics, Inc. All rights reserved. As part of the United States Air Force's More-Electric Aircraft initiative, a gearless, oilless integrated power unit (IPU) is under development for use on jet aircraft. The IPU combines the functions of a conventional auxiliary power unit (APU) and emergency power unit (EPU). Its function is to provide power to the 270 Vdc electrical bus of the aircraft and is being designed to provide a continuous power output of 125 kw. The IPU is turbine powered and comprised of a switched reluctance generator mounted on a common shaft between the compressor discharge and turbine and will operate at speeds up to 55,000 rpm. The generator employs a four-pole rotor with an overall pole-tip to pole-tip diameter of 109.2 mm (4.30 in.). The design clearance between the pole tips and surrounding stator wall is nominally 0.508 1.016 mm (0.020 0.040 in.). At 55,000 rpm, the pole tip speed is 314 m/s and viscous heat generation will clearly be present in the gap region. The generator is air cooled by the discharge-from the turbine compressor. Effective cooling of the rotor and stator is one of the many interrelationships involved in the optimal design of the integrated system. Estimating the convective film coefficient on the unique rotor pole faces is 1 American Institute of Aeronautics and Astronautics Copyright© 1998, American Institute of Aeronautics and Astronautics, Inc. essential for the overall thermal modeling of the generator. The University of Dayton has fabricated experimental test apparatus and developed analysis tools to test the actual rotor geometries at speeds ranging up to 30,000 rpm. Five test rotors of the same cross-section as the proposed rotors for the EPU were manufactured. Each test rotor was uniquely insulated to channel most heat transfer through the rotor pole face of interest. As part of the testing, a smooth cylindrical rotor, referred to as Rotor 6, was fabricated to provide experimental data to be used in validating the analysis of experimental test data. Analysis of the data for Rotors 1 through 5 is in progress; however, analysis of the test data for Rotor 6 has been completed. The following is a presentation of the theoretical model of the velocity and temperature fields in the annular gap between an inner rotating cylinder and outer stationary cylinder and a comparison of the theoretical convective film coefficient to the experimental data for Rotor 6. ANALYTICAL MODEL OF CONVECTTVE FILM COEFFICIENT FOR NONLAMINAR. WALL DRIVEN FLOW WITH VISCOUS HEATING The convective film coefficient was estimated by analyzing the airflow in the annular gap between the rotor and stator surfaces. The gap height is much smaller than the rotor radius so that this flow may be considered as wall driven flow between two flat plates as shown in Figure 1. In Figure 1 the rotor has been illustrated as the lower, stationary surface with heat flux qR transferred from the rotor to the air. The upper surface is shown as the stator surface with heat flux qs — qR + e transferred u(y) Upper surface moves at speed V, u(h) = V u(y), velocity distribution Thickness of viscous sublayer, Stationary lower surface, u(0) = 0