Machining is the most important part of the manufacturing process. Machining deals with the process of removing material from a work piece in the form of chips. Machining is necessary where tight tolerances on dimensions and finishes are required. The common feature is the use of a cutting tool to form a chip that is removed from the work part, called Swarf. Every tool is subjected to wear in machining. The wear of the tool is gradual and reaches certain limit of life which is identified when the tool no longer produce the parts to required quality. There are various types of wear a single point cutting tool may be subjected to in turning. Of these, flank wear on the tool significantly affects surface roughness. The other types of tool wears are generally avoided by proper selection of tool material and cutting conditions. On-line surface roughness measurements gained significant importance in manufacturing systems to provide accurate machining. The acoustic emission (AE) analysis is one of the most promising techniques for on-line surface roughness monitoring. The AE signals are very sensitive to changes in cutting process conditions. The gradual flank wear of the tool in turning causes changes in AE signal parameters. In the present work, investigations are carried for turning operation on mild steel material using high speed steel (HSS) tool. The AE signals are measured by highly sensitive piezoelectric element; the on-line signals are suitably amplified using a high gain pre-amplifier. The amplified signals are recorded on to a computer and then analyzed using MATLAB. A program is developed to measure AE signal parameters like ring down count (RDC), signal rise time (SRT) and root mean square (RMS) voltage. The surface roughness is measured by roller ended linear variable probe, fitted and moved along with tool turret on a CNC lathe machine. The linear movements of probe are converted in the form of continuous signals and are displayed on-line in the computer. This paper proposes to monitor tool wear and surface roughness by acoustic emissions using belief networks. The feature vectors of AE analysis and machining time were used to train the network. The overall success rate of detecting tool wear and surface roughness is high with low error.
Key words: Acoustic emission, surface roughness, turning, belief network.
Copyright © 2019 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0