Characteristic measurement system for automotive class switched reluctance machines

Abstract

Flux-linkage characteristics and torque characteristics, which are regarded as fingerprints of switched reluctance machines (SRM), are indispensable fundamental data required in modeling the machine behavior for both simulation and control purposes. In contrast to other types of electrical machines, e.g. induction machines and synchronous machines, SRMs are basically characterized by a strongly nonlinear behavior due to typical operation in the magnetic saturation region. Hence, no analytical function can be applied to describe characteristics of SRMs precisely. An accurate SRM model inevitably needs the complete machine characteristics in form of data-intensive look-up tables to represent the machine behavior. Generally, the SRM characteristics can be calculated by finite element (FE) simulation. However, real SRMs can be different from the simulation model due to end-winding effects, which are normally not considered in the FE-simulation model. Hence, the experimental measurement is preferred. Measuring and preparing the SRM characteristics is a complicated and time-consuming task, if done manually. The time-expense as well as possible human errors in determining the SRM characteristics can be minimized, if the measurement procedures are automated. This paper presents an automated characteristic measurement system for SRM. The measurement system was designed for automotive class SRMs, e.g. starter-generator, hybrid or main propulsion motor. The applied measurement methods, the system design and construction are described in the paper. Furthermore, constraints and discussions concerning the measurement accuracy are presented