Steel Gear Fault Diagnostics Using Non-Contact Magnetic Rotational Position Sensors

Abstract

This thesis reports on the investigation of a cost effective method of diagnosing gear faults using non-contact magnetic rotational position sensors. Like all new methods, the proposed sensor needed to be compared directly to an industry standard technique for measuring shaft rotations. The sensor was tested for its ability to record shaft rotational speed under a variety of experimental conditions. The sensor provided optimal measurements when directly compared to an industry standard technique of completing the same task. Using the fully calibrated sensor, experiments were conducted, which involved measuring the presence of faults in a 1:1 gearbox comprised of steel spur gears. Measuring the dynamic transmission error over the course of one full gear rotation allowed for identification of various gear fault types and severity levels. A large dynamic transmission error is the result of the output shaft lagging the input shaft, which ultimately represents the presence of a gear fault. The dynamic transmission error, over the course of one shaft rotation, was also investigated for the presence of a smooth engagement profile. A smooth engagement profile is the result of the sensor accurately detecting each individual tooth entering and exiting the gear mesh. The sensor demonstrated a great deal of promise for the application of a cost effective method for condition monitoring and fault diagnostics in a gearbox containing steel gears and loaded under realistic conditions.