Closed-loop control of surface preparation for metallizing fiber-reinforced polymer composites

Abstract

This study introduces a novel approach to enhance the surface properties of fiber-reinforced polymer composites through thermal spray coatings, utilizing a metal mesh as an anchor to improve coating adhesion. A critical step in this process is achieving optimal exposure of the metal mesh by sandblasting prior to coating. To address this challenge, we propose a closed-loop control system designed to inspect and blast parts effectively. Our method leverages top-view microscope images as inputs, employing a convolutional neural network (CNN) to correlate these images with the corresponding exposure levels of the metal mesh, measured via a destructive method. Upon training, the CNN model accurately estimates the exposure level solely from the top-view images, facilitating real-time feedback to guide subsequent sandblasting operations. Unlike traditional manual inspection methods, which demand expertise and experience, our automated approach streamlines the inspection process using a cost-effective portable digital microscope. Experimental findings validate the efficacy of our method in successfully discerning surface preparation status with an accuracy rate of 95% and demonstrate its practical utility in closed-loop control. Our study not only offers a robust methodology for quantifying surface preparation data but also presents a significant advancement in automating the inspection process. Moreover, the broader implications of our approach extend to various manufacturing sectors, where defect detection and closed-loop control are crucial for optimizing production efficiency and product quality.