Some Studies on Tool Condition Monitoring Systems Using Computer Vision and artifical Intelligence Techniques

Abstract

Product quality in machined parts is governed by many factors, one of which is the state of wear of the tool, which is one of the most critical factors. Tool wear and life are significant factors that influence part quality. Knowing the condition of the tool wear makes it possible to optimize the tool life and simultaneously maintain the surface quality. Industries rely on historical data to evaluate the tool's useful life. One of the essential requirements for intelligent manufacturing is the reliable predictions of the tool life during machining. It is crucial to monitor the condition of the cutting tool to achieve cost-effective and high-quality machining. Tool conditioning monitoring (TCM) is essential to determining the remaining useful tool life to ensure uninterrupted machining and intelligent manufacturing. Tool wear is traditionally assessed offline using a microscope, which is time-consuming. The condition of the tool can also be done by direct and indirect tool wear measurement and prediction techniques. The tool wear measured using the indirect approach uses parameters that impact tool life. There are methods of online wear measurement proposed in the literature, like correlating some physical parameters to the wear state of the tool. As the processes are indirect, they do not provide exact values of the tool wear but aid in classifying the wear into different states, from mild to severe. Direct methods such as the computer vision system are fast and reliable. In indirect methods, the data acquired from the sensors results in some ambiguity, such as noise, reliability, and complexity, among others. However, in the direct methods, the data is available in images, resulting in significantly less chance of ambiguity with the proper data acquisition system. The direct methods, which provide higher accuracy than indirect methods, involve collecting images of worn tools at different stages of the machining process to predict the tool's life. This work is focused on developing direct tool wear measurements by applying computer vision techniques. It has a negligible interruption in production and helps automate the task of monitoring and replacing tool wear. This study proposes an online tool wear measuring algorithm using edge detection and segmentation techniques. A complementary metal-oxide semiconductor (CMOS) sensor camera captures the wear zone images. The tool’s wear value is extracted by establishing wear boundaries through computer vision, threshold segmentation, edge detection, and morphological operation. The machining tests are performed on a computer numerical control (CNC )lathe machine. The tool wear measured by the proposed technique is compared with the measurements obtained by an optical microscope. The results demonstrated the high detection accuracy of the proposed approach, enabling online tool wear monitoring during turning. The present study proposes a tool wear and life prediction system to examine the progressive tool wear utilizing the artificial neural network (ANN). Experiments were performed on AISI 4140 steel material under dry-cutting conditions with carbide inserts. The cutting speed, feed, depth of cut, and white pixel counts of the tool wear zone are considered input parameters for the proposed model, and the flank wear, along with the remaining tool life, is predicted as the output. The images of the worn tool were captured using an industrial camera during the turning operation at regular intervals. The ANN training set predicts the remaining useful tool life, especially ANN's sigmoid function and Rectified linear unit (ReLU) activation function. The sigmoid function showed an accuracy of 86.5%, and the ReLU function resulted in 93.3% accuracy in predicting tool life. The proposed model's maximum and minimum root mean square error (RMSE) is 1.437 and 0.871 minutes, respectively. The outcomes showcased the ability of computer vision and ANN modelling as the potential approach for developing a low cost industrial TCM system that can measure tool wear and predict tool life in turning operations. Experimental investigation to predict tool life by applying backpropagation algorithms in neural networks is carried out to automate flank wear assessment, predict flank wear, and estimate tool life. Measuring and monitoring tool wear is automated using computer vision techniques. The machining characteristics investigated are speed, feed, and depth of cut. Industrial camera with telecentric lens has been used for online image-based tool wear monitoring. Different back propagation neural network algorithms, such as the Levenberg - Marquardt algorithm, Bayesian regularization, and scaled conjugate gradient, are used to predict the tool life. Performance evaluation of the algorithms is done to find the most accurate algorithm for the tool wear prediction system. This work proposes a computer vision-based tool wear monitoring and tool life estimation system using machine learning methods. Gradient-boosted trees and support vector machines (SVM) techniques are used to estimate tool life. The experimental investigation on the CNC machine is conducted to study the proposed tool wear monitoring system. Experiments are conducted using workpiece material made of alloy steel and coated cutting inserts, and flank wear was monitored. An imaging system consisting of an industrial camera, lens, and LED ring light is mounted on the machine to capture tool wear zone images. Images are then processed by algorithms developed in MATLAB ®. A microscope is also used to measure the wear of the flank surface manually. Validation tests are carried out to determine the accuracy of proposed models, and it is observed that the accuracy of prediction of boosted three and SVM is good with maximum error of 5.89% and 7.56%, respectively. The study's outcome established that the developed system can monitor the tool wear with good accuracy and can be adopted in industries to optimize the utilization of tool inserts. In summary, this study has contributed to a better understanding of implementing condition monitoring systems for cutting tools. Direct computer vision systems are implemented for tool wear measurement and accurately predicting tool life, which is one of the essential elements of automated and intelligent machining processes. It aids in achieving the goal of producing quality production with reduced production costs.