Analysis and Prediction of Temperature Using an Artificial Neural Network Model for Milling Glass Fiber Reinforced Polymer Composites

Abstract

Milling parts made from glass fiber-reinforced polymer (GFRP) composite materials are recommended to achieve the geometric shapes and dimensional tolerances required for large parts manufactured using the spray lay-up technique. The quality of the surfaces machined by milling is significantly influenced by the temperature generated in the cutting zone. This study aims to develop an Artificial Neural Network (ANN) model to predict the temperature generated when milling GFRP. The ANN model for temperature prediction was created using a virtual instrument developed in the graphical programming language LabVIEW. Predicting temperature is crucial because excessive heat during milling can lead to several issues, such as tool wear and thermal degradation in the polymer matrix. The temperature in the tool–workpiece contact surface during the milling process was measured using a thermography technique with a ThermaCAM SC 640 camera (provided by FLIR Systems AB, Danderyd, Sweden), and the data were analyzed using the ThermaCAM Researcher Professional 2.8 SR-2 software. Experimental research shows that the cutting speed has a much more significant effect on the temperature in the cutting zone compared to axial depth of cut and feed speed. The maximum temperature of 85.19 °C was measured in the tool–workpiece contact zone during machining at a cutting speed of 75.39 m/min, a feed rate of 250 mm/min, and an axial depth of cut of 12 mm. This temperature rise occurred due to the larger contact area and heightened friction resulting from the abrasive characteristics of the reinforcement material.