White Rose University Consortium logo
University of Leeds logo University of Sheffield logo York University logo

Mathematical modelling and in-process monitoring techniques for cutting tools.

Oraby, Samy El-Sayed (1989) Mathematical modelling and in-process monitoring techniques for cutting tools. PhD thesis, University of Sheffield.

[img]
Preview
Text (DX190329.pdf)
DX190329.pdf

Download (43Mb)

Abstract

The need is expressed for mathematical models which describe the cutting tool-workpiece interaction and for accurate on-line monitoring of tool-state. These are essential requirements for the achievement of unmanned and computerized machining processes. Techniques are used to design the experiments which substantially reduce the number of tests while providing all the essential information for statistical analysis and for the development of mathematical models. The testing conditions are chosen to reasonably conform with the practical requirements. Multi-coated carbide tool inserts [Sandvik GC435] are used to cut an alloy steel [EN 19] under normal cutting conditions and for a wide range of operating parameters. An accurate and sensitive three-component dynamometer was designed, manufactured and used to measure the tool forces through a BBC microcomputer. Continuous records of the tool vibration have been collected in two different co-ordinate directions simultaneously together with measurements of tool wear and cutting forces. Linear and non-linear regression techniques are used to develop mathematical models for the experimentally measured responses: cutting forces, tool vibration, and tool wear. Special attention is devoted to the identification of the most appropriate models. Each model being capable of representing the tool state throughout its working lifetime. Tool life wear-based models are developed to relate the expected tool lifetime to the operating parameters: speed, feed, and depth of cut. A robust regression analysis technique, used in conjunction with iteratively re-weighting least-squares, has been found to improve the accuracy of the models, and to stabilize its computed residuals through the elimination of the effect of influential observations having high experimental error. Response surface methodology RSM has been used to signify the non-linear nature of tool life response. The force variation has been shown to correlate strongly with the wear progress so that it can be used for accurate in-process determination of tool wear and for monitoring tool state. It has been shown that the variation in the ratio between force components correlates with wear and is independent of the effect of other machining parameters; this enables the approach to be used for a wider range of materials and more extensive operational domain. Study of the power spectral analysis of the tool vibration indicates that among the tool's vibration modes, the first fundamental natural frequency of vibration in the feed direction exhibits a consistent correlation with wear-progress. The Vibration amplitude decreases with the increase of the wear level until it reaches a limit after which it tends to reverse its characteristic. The time at which the characteristic changes is found to closely correspond with the practical end of the tool lifetime. Based on this fact, an in-process approach is investigated to determine the tool life on-line. Also, a model has been developed for tool wear estimation based on a combination of vibration and force; and, very good agreement has been obtained with the experimental data. The validity of the models; their feasibility; and, their industrial significance are confirmed for adaptive control AC systems, and for machinability data base systems MDBS.

Item Type: Thesis (PhD)
Keywords: Computerised machining
Academic Units: The University of Sheffield > Faculty of Engineering (Sheffield) > Mechanical Engineering (Sheffield)
Depositing User: EThOS Import Sheffield
Date Deposited: 25 Oct 2012 14:55
Last Modified: 08 Aug 2013 08:47
URI: http://etheses.whiterose.ac.uk/id/eprint/1814

Actions (repository staff only: login required)