Machine learning algorithm-based prediction of machined surface quality in end milling operation

Machine learning (ML) has become an important tool for the development of Industry 4.0. It assists the machining processes by monitoring and maintaining the conditions. Support vector machine (SVM) is one such algorithm of ML used to train and classify the data. The present work uses the SVM for predicting the surface roughness in the end milling of the low-carbon steel. The experiments were performed at nine different combinations of process parameters. Moreover, to monitor the cutting process online, the current drawn is measured using a current sensor. In this regard, a correlation between the current drawn and variation in surface roughness is reported. The average value of the surface roughness was predicted using the SVM at each combination. The results show that the SVM estimates the surface roughness with an approximate error of 0.4 %-10%. On the other hand, the surface roughness variation does not fit well with the current signals due to the variation in tool wear.