Development of Flank Wear and Surface Roughness Prognosis System in Lathe Machine Based on an Affordable Monitoring System

This chapter delves into the critical aspects of tool wear and surface roughness in lathe machining processes, emphasizing the significance of flank wear as a key indicator of tool condition and workpiece quality. It presents an innovative monitoring system designed to predict flank wear and classify surface roughness using machine learning algorithms. The system employs a piezoelectric disc sensor to collect data on tool deflection, which is then processed by an affordable microcontroller device and transmitted wirelessly to a personal computer. The core of the system lies in its monitoring application, developed on an open-platform Windows application using C# language programming. This application implements a multilayer perceptron (MLP) model to predict flank wear based on cutting speed, accumulated removed material, and cutting time. The predicted flank wear values are then used to classify surface roughness using a k-nearest neighbors (kNN) algorithm. The experimental setup involves various cutting speeds and constant feed rates and cutting depths, with tool wear and surface roughness measurements validated through microscopic observations and surface roughness instruments. The results demonstrate the system's high accuracy in predicting flank wear and classifying surface roughness, making it a valuable tool for enhancing machining processes and ensuring high-quality production. The chapter concludes with a discussion on the system's affordability and potential benefits for industries aiming to elevate their production quality through real-time tool condition monitoring.