An Integrated Method for CFRP Drilling Quality Prediction and Multi-Objective Optimization

Carbon fiber-reinforced polymer (CFRP) is extensively applied in aerospace and rail transportation industries, and the quality of CFRP joint holes is crucial for ensuring joint performance and service reliability. However, CFRP drilling involves complex interactions among drilling parameters, tool geometry, and multiple quality indicators, making it difficult to accurately predict drilling quality and identify optimal process parameters. In this study, drilling experiments using twist drills and dagger drills were conducted to analyze the relationships among drilling parameters, drilling physical indicators, and drilling quality indicators. Compared to the twist drill, the dagger drill maintained lower thrust force and cutting temperature, and reduced the delamination factor, burr factor, and drilling wall roughness by 20.6%, 95.5%, and 81.1% on average, respectively. To consider the effect of different drilling indicators on the quality of CFRP drilling holes, a comprehensive fuzzy evaluation prediction model FCE-NN for drilling quality was proposed. The average prediction accuracy reached 91.3% and the drilling indicators FCE were output. The NSGA-II algorithm was employed, and an entropy-weighted TOPSIS method integrated with fuzzy comprehensive evaluation (FCE) was used to rank the Pareto-optimal solutions, thereby achieving multi-objective optimization of the thrust factor, delamination factor, and machining efficiency.