DEVELOPMENT OF TOOL CONDITION MONITORING SYSTEM FOR FLANK WEAR BY MACHINE LEARNING IN TURNING PROCESS - 16/01/2023 23:41 EST

A thesis about creating a TCMs which will compare the experimental data and simulation data for machining process, turning process using CNC lathe machine. The parameter for the experimental data will be based on 3 main categories, cutting speed, feed rate and depth of cut. there will be three set for each one that will turn into 27 small classes of parameter results from combining those 3 main categories. A vibration analyzer will be used to collect the vibration of each classes. for the classification of the data, there will be 2, which is flank wear<0.1mm and 0.1<flank wear>0.3mm. meaning, the vibration data will only be collected when the flank wear of the cutting tool falls into the classification. the data obtained basically will be acceleration vs Hz, which will be extract using extraction features. the extraction features of the time domain will be maximum, mean, root mean square, variance, standard deviation, skewness, kurtosis, peak-to-peak and crest factor. from all this a type of machine learning method, Support vector machine will be used to compare the data from the experimental data and simulation data. for the validation of result, absolute percentage error(APE) and mean absolute percentage error (MAPE) will be used while for the accuracy, R-squared will be used which can be obtained from RSS and TSS. Therefore, a system need to be develop using only MATLAB that meet all the details above.