Date of Award

8-2017

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

Committee Member

Dr. Gregory Mocko, Committee Chair

Committee Member

Dr. Laine Mears

Committee Member

Dr. Rodrigo Martinez-Duarte

Abstract

Gear grinding is a process used to improve the surface finish of machined gears to increase their lifespan and decrease noise during their operation. Large scale gear grinding produces finished gears at a competitive cost but tool wear plays an important factor in the final quality. The objective of this research is to identify how process parameters during the gear grinding process vary and determine if they can predict the noise associated with gears in final assembly. Specifically, this research records the vibrations on the grinding wheel and decomposes them using a Fast Fourier Transform (FFT). The vibration patterns at the grinding wheel mesh frequency are studied using two design variables that characterize the tool, a) grinding wheel diameter (d) and b) location along the grinding wheel width (y). These variables correspond to geometrical positions on the tool over its lifetime. This was followed by measuring parts machined at sections of the grinding wheel (varying y values) that recorded the highest and lowest vibrations to evaluate if the vibrations influenced the surface finish of the gears. Finally the gears are installed in gearboxes and tested for noise made due to running gears to evaluate if there was a difference in noise based on the gear geometries and the machining location on the tool Analyzing vibration data for 2868 parts machined using a full tool, the results of an ANOVA and two sample t-tests showed a statistical difference between the vibrations recorded at different sections of the grinding wheel. Vibrations at y4 are higher than the vibrations at y34 by 3.035 mg while vibrations at y4 are higher than the vibrations at y3 by 2.12 mg. Analyzing the geometrical data for 313 gears over four y locations, the results show that the surface roughness of left gear profiles machined at y4 is greater than left gear profiles machined at y34 by 0.458 microns. The roughness of left gear profiles machined at y4 is greater than the left gear profiles machined at y3 by 0.167 microns. Additionally, the roughness of right profiles machined at y4 were lesser than those machined at y34 by 0.175 microns. Finally, 294 gears were tested in gearboxes and the statistical results show that gears machined at y4 were louder than gears machined at y34 by 1.088 dB while there was no statistical difference in noise made by gears machined at y4 and y3. The future scope of this work will be to perform similar studies on different processes and determine if limits can be set to identify when rougher parts are machined and removed from serial production. This may also be achieved by taking samples from production failures and use them as a knowledge base to determine if quality can be determined by on-line monitoring systems.

Share

COinS