Content area
Abstract
The work in this dissertation is focused on ultrasonic inspection of bearing components. Roller bearings sustain the heavy loads railcars can often encounter in the rail industry and consist of an inner and outer ring with rollers transferring the load between the two. Rollers in smaller bearings are typically spherical with the contours of the rings conforming closely to this spherical geometry. Tapered roller bearings were designed to increase this load capacity by increasing the contact area between the rollers and raceways by using rollers of cylindrical shape. This design allows for the support of heavier railcars but introduces a higher likelihood of encountering a defect below the contact surfaces due to increased contact area between rollers and rings.
In this work, metrics for ultrasonic quantification of inclusion content are described and applied to railcar bearings. An analysis program is then developed to quantify inclusion content in a consistent manner regardless of the part geometry such that meaningful comparisons between different steels can be made. Ultrasonic scanning methods and reference parts for equipment setup and calibration are also described.
In a similar manner, an alternative mode of defect detection is employed to detect inclusions in this critical region utilizing surface waves which propagate only near the surface of the part. The analysis code developed also characterizes these defects in terms of similar metrics for comparison between steels. Inspection routines and novel reference parts for verification of experimental setup and parameters are developed.
Predictions of the ultrasonic attenuation of waves in the presence of several types of defects are presented as a result of the analysis and theoretical development. The work presents predictions of detection limits in terms of both defect size and type, as well as the incorporation of steel characteristics such as the average grain size.
Bulk inspection of defects is modeled considering backscattered signals as a function of defect size and type, as well as the grain size of the steel used. Predictions of backscattered amplitude are given for defects with sizes and types that are of concern for fatigue considerations. Previous studies utilize an approach that models grain noise and defect backscatter independently. This work presents a unified approach to this problem that has not been performed previously.
