DefinePK

Abstract

Tool failure is a major and undesirable occurrence affecting the overall operating cost and time as the machining needs to be done once again to fix the mistake. Therefore, this paper introduced an efficient and inexpensive way to overcome the problem by developing tool wear monitoring system using Macro-Fibre Composite (MFC) sensor via alternative statistical signal analysis method, namely Integrated Kurtosis-based Algorithm for Z-notch filter (I-kazTM). A piece of MFC sensor amplified by a power module was mounted on a tool holder in the turning machine to capture vibration signal data using data-logger while cutting the workpiece. The operation ran continuously until criteria of 0.3 mm tool wear achieved with the help of a microscope for wear measurement. The machining was set at 250 and 300 m/min of cutting speeds, while the feed and depth of cut were kept constant at 0.25 mm/rev and 0.12 mm respectively. The raw data were then extracted and observed in time and frequency domain before statistically analysed as soon as the experiment finished. The reliability of I-kazTM method was made to the test by performing correlation with the wear progression data using regression analysis to derive the best equation model and comparing it with one of the global statistical features, namely root means square (rms). The final result indicated that the measured tool wear directly proportional to I-kaz coefficient, where the increment of wear progression increasing the I-kaz coefficient value. It came with the best fit of quadratic polynomial regression models, producing acceptable correlation of determination, R2 of 0.83 and 0.93 while rms having lower values of 0.65 and 0.83. The outcome of the result also showed that the proposed study of using I-kazTM to analyse the vibration signal from MFC sensor was much more reliable than the rms feature. It can be used to monitor tool wear efficiently with 1.8 to 15.9 % of error using I-kazTM while the latter showed a higher percentage of error from 3.4 to 30.1 which nearly as twice as higher.