Optimising Turning Operation in Precision Manufacturing Using Fused IoT Devices and Machine Learning Approach.
Isaac O. Olalere, Oludolapo Akanni Olanrewaju
Abstract
The two principal components of a turning operation on the machine tool are the cutting tool and workpiece. While the cutting tool condition alongside the machine tool condition, determines the surface finish of a machined part, the workpiece material influences the cutting tool condition. This paper proposes an intelligent precision manufacturing system that employs a machine learning algorithm on the data captured by fused IoT devices for intelligent decision making and condition identification. The data captured during the turning operation includes the thermal image, the vibration condition during the cutting operation and the intermittent wear parameters of the workpiece and cutting tool. Convolutional Neural Network (CNN) algorithm is used to classify, and correlate captured thermal images, while the vibration parameters and wear parameters are trained using Artificial Neural Network (ANN) for classifying the tool and workpiece wear conditions. Measured parameters of the finished part are correlated with the tool condition for optimizing the tool life. Classes indicating the tool and workpiece conditions are established from the trained network of measured data. The significance of the proposed method is that the tool wear condition is measured against the product finish requirement, hence, optimizing the operation by optimally utilizing the tool based on product requirement while reducing downtime due to intermittent condition monitoring.