Evaluation and Modeling of Rotor Position Sensor Characteristics for Electric Traction Motors 2016-01-1065
Vehicles driven by electric or hybrid technologies have the advantage that a high torque potential can be used from the start, hence the initial vehicle acceleration is higher compared to conventional propulsion concepts . The speed-torque characteristic of electric machines is nearly ideal for the use in automotive applications and electrical machines can be controlled with a high efficiency. The aim of the present work is the examination of different sensor technologies, which are used in such automotive applications to measure the rotor position of electric motors. The project includes the assessment and evaluation of different sensor technologies, e.g. resolver, eddy current sensors and sensors based on magneto-resistive effects. The quality of the sensor angular measurement depends on different parameters, for example misalignment in planar direction, longitudinal direction, tilt angle, temperature, rotational speed and supply voltage. For evaluation of all these influencing factors, a specific test bench with a maximum speed of 24000 rpm, a high precision automated positioning system of the device under test, a variable voltage supply and a temperature range from -40 °C up to +160 °C was built up. Target was to study various sensor concepts in view of requirements for automotive applications. All measurements are compared with a high-precision reference position sensor system for evaluating the angular error of the device under test. In this paper, a methodology based on Design of Experiments is introduced to evaluate the sensor error characteristics as a function of the mentioned sensitive parameters. Different experiment designs are examined, with the target to reduce the overall number of test runs and to generate as much information by keeping the number of experiments as low as possible. The results of the experiments are used to compute mathematical models of the sensor error, for example quadratic response surface models.