Mechanical & Materials Engineering, Department of

 

First Advisor

Joseph A. Turner

Committee Members

Kamlakar Rajurkar, Keegan Moore

Date of this Version

8-2024

Document Type

Thesis

Citation

A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science

Major: Mechanical Engineering and Applied Mechanics

Under the supervision of Professor Joseph A. Turner

Lincoln, Nebraska, August 2024

Comments

Copyright 2024, Trevor K. Adelung. Used by permission

Abstract

While in service, railroad bearings must withstand large loads to ensure smooth operation. These large loads can lead to failure due to rolling contact fatigue (RCF), for which small cracks initiate just beneath the surface of the components. As a preventative measure, bearings are manufactured with a hardened surface layer called the effective case depth (ECD). Current methods in industry use destructive hardness testing to quantify the ECD. While accurate, this method is costly and time consuming. Ultrasonic nondestructive testing would provide a more efficient and cost-effective method to determine ECD. In this thesis, diffuse ultrasonic shear backscatter was used to characterize the ECD in bearing inner rings (cones). The spatial variance was measured and an exponentially modified Gaussian (EMG) was fit to the curve. The EMG is characterized by four parameters, in which the peak arrival time and peak amplitude were shown to have linear trends with ECD. Predictions were made using these trends based on a linear fit. Both a classification and regression machine learning model were implemented to predict case depth. It was determined that the best predictor of ECD was the peak arrival time. Both machine models also made accurate predictions. This study demonstrates that ultrasonic nondestructive testing is viable as a method to predict changes in the ECD, allowing a quicker, less expensive alternative for measuring ECD in railroad bearing components. Further research could optimize this method for efficient use in industry applications.

Advisor: Joseph A. Turner

Share

COinS