The cutting tool is the only element in a machine tool that requires frequent changes due to failure. Drill bit wear can cause catastrophic failure that can result in considerable damage to the work piece and the machine tool. Hence, there is an imperative need to keep a watch on the condition of the cutting tools during the machining process. Over the years, a wide variety of on-line or off-line techniques have been investigated for monitoring abnormal cutting tools. A variety of signals such as tool-tip temperature, forces, power, thrust, torque, vibrations, shock pulse, Acoustic Emission (AE) etc., have been used for monitoring tool failure by on-line technique. The detection and monitoring of AE is commonly used to predict tool failure. Present work involves estimation tool flank wear in drilling based on AE parameters viz., RMS, energy, signal strength, count and frequency by empirical methods of analysis like Multiple Regression Analysis and Group Method of Data Handling (GMDH). The experimental work consisted of drilling S.G Cast iron using high-speed steel drill bit and measuring AE parameters from the workpiece using AE measuring system for different cutting conditions. Machining was stopped at regular intervals of time and tool flank wear was measured by Toolmakers microscope. The experimental data were subjected to simpler methods of analysis to obtain a clear insight of the signals involved. The study of AE-time plots showed a similarity with three phases of tool wear, which implies that the measured AE parameters can be related to tool wear. Multiple Regression Analysis and Drilling is a major material removal process in manufacturing. Infact, the drills have been used widely in industry since the industrial revolution. It was estimated that 40% of all the metal removal operations in the aerospace industry is by drilling. Similar to the other cutting tools, after a certain limit, drill wear can cause catastrophic failure that can result in considerable damage to the work piece even to the machine tool [1]. GMDH methods were successful in estimating flank wear based on measured AE parameters. By Multiple Regression Analysis better estimation was obtained at lower cutting conditions. Three criterion functions of GMDH viz., Regularity, Unbiased and Combined were used for estimation with 50%, 62.5% and 75% of data in the training set. Estimation was done upto Level-4. The results of GMDH estimation showed that regularity criterion functions correlates well for the set of input variables compared with unbiased and combined criteria and least error of estimation was found when 75% of data was used in the training set. The optimum level of estimation increased with the increase in the percentage of data in the training set. Comparison of the performance of Multiple Regression Analysis and GMDH indicated that estimation by regularity criterion of GMDH had an edge over Multiple Regression Analysis.

This content is only available via PDF.
You do not currently have access to this content.