Wettability-Driven void suppression and enhanced mechanical performance in Si₃N₄-Reinforced bamboo Fiber epoxy composites using COCOSO and ANN modeling.

The incorporation of ceramic particles as fillers and natural fibers as reinforcements in polymer composites has gained significant attention in recent years. However, defects such as particle agglomeration, poor wettability, non-homogeneous distribution, and void formation often arise due to unoptimized process parameters. This study aims to enhance wettability between fillers, reinforcement, and the matrix while improving uniform dispersion by optimizing filler percentage, stirring time, and stirring speed. To achieve this, Silicon Nitride (Si₃N₄) fillers were dispersed in bamboo fiber-reinforced epoxy composites using an L₂₇ orthogonal array, with process parameters varied across Si₃N₄ filler content (3%, 6%, and 9%), stirring time (6, 12, and 18 min), and stirring speed (250, 350, and 450 rpm). The response parameters such as void formation, contact angle, and tensile strength were analyzed, and the Combined Compromise Solution (COCOSO) method, integrated with an Artificial Neural Network (ANN), was used to determine the optimal parameter settings and compare experimental results with predictions. The findings reveal that using 6% Si₃N₄, a stirring time of 12 min, and a stirring speed of 350 rpm improves wettability by 25.1% while reducing void formation by 79.3%. These results highlight the importance of optimizing wettability through precise control of filler content and processing parameters to achieve high-performance composites with minimal defects and enhanced mechanical properties.