Methodology for Identification of Quality of clean-Cut surface for IS2062HR sheet metal blanking using Random Forest

Abstract

In the instance of sheet metal blanking, poor cut-surface quality of a blanked surface might cause fit difficulties in the assembly. Due to the uneven surface, cracks may occur, resulting in a loss of surface quality and dimensional precision. The quality of clean-cut surface is investigated using four parameters: punch penetration, shear angle, fracture angle, and burr height are considered for the present study. The depth of punch penetration is gradually raised to identify where the crack starts. Shear angle, fracture angle, and punch penetration following punching, as well as burr height, are considered input factors. On the power press, the uni-punch tool is utilized as a cutting tool, and the processing material is IS2062HR sheet metal. The objective of this work is to use the surface roughness value to classify cut-surface quality into three categories for decision making on fit for assembly operation. To forecast the quality of cut surface, a classification model is created using the Random Forst Classifier method of the Machine Learning approach. The Gini and Entropy index method revealed that the model is 93 percent accurate.