Determination of Parametric Uncertainties for Regression-Based Modeling of Turning Operations

Published

May 30, 2006

Author(s)

Robert W. Ivester, Laurent Deshayes, Michael L. McGlauflin, Y Marcon

Abstract

Uncertainty associated with model-based predictions of machining performance plays an important role in the application of machining modeling. This paper presents a technique for model development and characterization of the uncertainty in the resulting model parameters. This technique provides a pragmatic basis for mathematical optimization of machining with an allowance for uncertainty in model-based predictions. The workpiece material, American Iron Steel Institute (AISI) 1045 steel, was machined using inserts with four different chip breaker geometries. A regression-based model of process performance based on the Coupled Work Tool (CWT) methodology is calibrated along with the determination of uncertainties associated with key model parameters. Parametric specification of modeling uncertainty enables robust optimization techniques to allow for variable process performance due to uncontrolled factors in machining operations.

Proceedings Title

2006 Transactions of the North American Manufacturing Research Conference

Volume

34

Conference Dates

May 24-26, 2006

Conference Location

Boston, USA

Conference Title

Same

Pub Type

Conferences

Keywords

machining, Modeling uncertainty, robust optimization, smart machining systems

Citation

Ivester, R. , Deshayes, L. , McGlauflin, M. and Marcon, Y. (2006), Determination of Parametric Uncertainties for Regression-Based Modeling of Turning Operations, 2006 Transactions of the North American Manufacturing Research Conference, Boston, USA (Accessed October 30, 2025)