Quality evaluation and predictive analysis of drilled holes in jute/ palm/polyester hybrid bio-composites using CMM and ANN techniques

Abstract

In this study, the evaluation of 75 holes drilled in a hybrid bio-composite jute/palm/polyester plate and controlled by a coordinate measuring machine (CMM) is essential to ensure the quality, dimensional precision, and geometric conformity of the plate. This rigorous process is necessary to meet industrial standards for circularity and cylindricity, which are essential criteria for high-performance applications. Additionally, the integration of artificial neural network (ANN) techniques has revolutionized this approach by enabling precise predictions of key parameters such as delamination, circularity, and cylindricity. In this study, the ANN was trained with 52 samples (70%), while 8 samples (10%) were used for validation and 15 others (20%) for testing at different stages. The results show the influence of feed rate on the delamination factor (Fd) (R2 = 0.98), circularity error (R2 = 0.99), and cylindricity error (R2 = 0.98). This predictive approach significantly improves the reliability and efficiency of the evaluation process.