Experimental and Statistical Evaluation of Drilling induced Damages in Glass Fiber Reinforced Polymer Composites - Taguchi integrated Supervised Machine Learning Approach

Glass reinforced polymer (GFRP) composites are gaining on its usage in various sectors. The drilling of GFRP composite is an inevitable machining operation. The anisotropy feature of the polymer composite makes it little difficult-to-machine material. The drilling of GFRP composite is accompanied by delamination damage. Moreover, the quality of drill, characterized by the hole's surface roughness is also an important response variable to consider. The effect of Feed and drilling speed has been always focused by several researchers, to minimize the damages caused during drilling of GFRP composites but very few have considered the drill tool geometry as an affecting parameter. The present study, thus, investigates the effect of drill tool geometry along with drilling speed and feed on the delamination damage and surface roughness of the drilled hole. The study indicates that drill geometry has the highest significance on the damages considered in the study and contributes more than 75% towards the variance. The supervised learning approach, in terms of linear regression is used in the present work to determine the predicting models for the obtained data. The mathematical models developed using the machine learning approach possess high degree of fitness with all the three R² values being more than 90%.