Tool Life Prediction in CNC Machine Using Machine Learning and Neural Network Based On Machine Vision

Abstract

Tool Conditioning Monitoring (TCM) is the one of the main part in the field of the machining. In various ways tool conditioning monitoring can be happen like manual way, Indirect way, and Direct way. By these ways of the TCM can find the Remaining useful tool life and tool wear. So, till this time in many industries there is use of the manual way of the tool life by using the microscope and calculation of the wear area or by using machining time guessing of the tool life. So due to that sometimes if there is mistake in the machining time tool can be break before that so it can cause the accident of the error in machined part. In this era many researchers have focused on the Indirect way due to it is very low-cost system, indirect way uses various effective factors like force, voltage etc. so depend on the sensor collected data it predicts the life if tool. But it has limitation of reduced error up to 7-10% based on literature. So, there is use of the Direct way which collects the data from the image of worn tool and based on that data it predicts the life of tool and it reduce the life prediction error up to 5-4%. And if there is use of the effective neural network it reduces the error more. So, looking forward the 4th revolution era of industry, an attempt will make in this project to develop the tool life prediction system by using the machine vision system and Neural network. Various literature are studied to identify the various methods to collect the data from image by image processing techniques and use that data in neural network to predict the tool life and tool wear. There is use of the GUI to conduct this image processing, collection of data and neural network prediction. By using python programming language developed the GUI for the Image processing and using the Binary image extraction technique to collect the data from the image. There is use of the Sigmoid function ANN and ReLU function ANN for prediction of the tool life after collect the data from the experiment. In the experiment there is setup of Industrial camera system and illumination system in CNC to collect the image at prescribed time gapes.