DEVELOPMENT OF AN ON-LINE TOOL WEAR MONITORING SYSTEM USING DYNAMIC CUTTING FORCE

Abstract

This project examines the possibility of using the dynamic cutting force to indirectly monitor the tool wear and predict the tool failure on a lathe turning operation. A personal computer can be used for on-line in-process tool wear monitoring by tracking the dynamic force signals of the cutting tool. Findings show that there is a good correlation between the dynamic tangential force component and the flank wear of the tool. A consistent trend is observed when the dynamic force is presented in a frequency spectrum. A characteristic peak frequency relating to the natural frequency of the tool overhang is observed in the frequency spectrum. This peak frequency is found to remain substantially constant throughout the life of the tool. The amplitude of the dynamic force increases steadily as the tool wear progresses and decreases on approaching tool failure. It is deduced that the dynamic force increases as the contact area between the tool and the workpiece increases. On approaching tool failure, the thermal weakening of the tool causes the dynamic force to decrease. A program using Turbo C language is developed for a personal computer to track down the dynamic force signals and to predict tool failure. The computer is equipped with a fast fourier transform (FFT) card to perform real time signal processing and continuous data acquisition. The FFT card has a converter that converts the analogue signals to digital signals. The data acquisition and processing cycle time of the program is about 22 seconds for the 10 Mhz 286 computer. Two criteria were suggested for the prediction of tool failure : (i) when the percentage drop in the dynamic tangential force from its maximum value has exceeded a pre-determined threshold value, (ii) when the gradient of the curve of the dynamic tangential force is at a negative slope. The threshold value for the percentage drop is determined by statistical analysis based the on t-distribution method, the confidence level is taken at 99%. Experimental work was conducted on various types of workpieces and tool inserts. The results show that the consistent trend in the dynamic tangential force can be obtained and that the two criteria could be satisfactorily applied. Similar pattern can also be obtained when using a different normal rake angle of the tool. Studies have shown that the threshold values for the percentage drop in the dynamic tangential force do not vary radically under different cutting speed. However, the threshold values of the percentage drop in the dynamic tangential force when using the titanium coated tool insert is about twice the amount compared to using the uncoated tool insert. This is due to the higher wear resistance of the coated tool insert. The use of the personal computer to monitor in-process tool wear by tracking the dynamic force is found to be feasible. Further improvement in the reliability of the on-line sensor can be achieved by using a faster performance computer.