Date of Award

Winter 2009

Project Type

Dissertation

Program or Major

Mechanical Engineering

Degree Name

Doctor of Philosophy

First Advisor

Barry K Fussell

Abstract

The main goal of this dissertation is to pave the way towards a milling simulation software platform that enables academic modeling research to be easily incorporated into industrial practice. The most significant contribution is creation of numeric structures that enable utilization of any milling cutting force model with any type of milling cutting tool in a computationally efficient manner. Efficiency and accuracy of the force and surface finish modeling are the main focus of this study.

Force modeling is an important part of milling research since it is directly connected to tool health and workpiece quality. A number of force models are investigated for feedrate selection, assuming that calibration is limited to a spindle motor power sensor. Although more restricted, motor power sensors are cheaper and more practical alternative to table dynamometers for force model calibration purposes. Calibration of the force models with a motor power sensor is derived and their feasibility and accuracy is evaluated by a number of experimental cuts. It is shown that each force model performs much better than the simple Material Removal Rate model that is predominantly used in industry. Also, advantages of the different models under different cutting conditions are discussed.

Significant problems to industrial use of milling models include the excessive computational time required by the algorithms and the difficulty of easily incorporating a large number of cutting tool types. Various numeric structures, which can be used as a Software Development Kit (SDK), are developed to address these problems. These structures allow utilization of any force model with any cutting tool in a computationally efficient manner. Also, the structures make the milling related programming much simpler and flexible.

A surface modeling program is created using the structures and evaluated through a number of experiments. This program calculates cutting forces, tool vibrations and the resulting peripheral surface. Despite the complexity of the concepts in this program, it is less than 140 lines, and performed well when tested with two force models and three different cutting tools. Force predictions, surface roughness, and surface tolerance were shown to be reasonably accurate under most cutting conditions.

Share

COinS