Feng Guangshuo, Liu Bo
School of Mechanical Engineering, University of Science and Technology Beijing, Beijing 100083, China.
Beijing Key Laboratory of Lightweight Metal Forming, Beijing 100083, China.
Materials (Basel). 2024 Jul 18;17(14):3545. doi: 10.3390/ma17143545.
Continuous fiber-reinforced composites are increasingly used in industry for their superior specific modulus and strength. The curing process-induced deformation (PID) has been a critical problem during manufacturing, which always exhibits dispersed values even if the curing process curve and structural parameters remain consistent. This work conducted probability prediction of PID for V-shape composite structures based on the FEM method and data mining. A sequential coupling thermal-chemical-mechanical coupling FE model is established in ABAQUS. The prediction accuracy of the included angle between two sides is verified by the experimental results. Material parameter uncertainties are considered for V-shape structures with different radii and thicknesses. Based on the dataset from the FE model, a decision tree is established and trained to analyze the sensitivity and to predict the probability distribution of PID. The results show that PID increases with the coefficients of thermal expansion in the in-plane perpendicular fiber direction and out-of-plane normal direction. The data-mining method is accurate enough for the PID prediction, and its efficiency provides an additional calculation option in engineering applications.
连续纤维增强复合材料因其优异的比模量和强度而在工业中得到越来越广泛的应用。固化过程引起的变形（PID）一直是制造过程中的关键问题，即使固化过程曲线和结构参数保持一致，其值也总是呈现离散状态。这项工作基于有限元方法和数据挖掘对V形复合结构的PID进行概率预测。在ABAQUS中建立了顺序耦合热-化学-机械耦合有限元模型。通过实验结果验证了两侧夹角的预测精度。考虑了不同半径和厚度的V形结构的材料参数不确定性。基于有限元模型的数据集，建立并训练了决策树，以分析灵敏度并预测PID的概率分布。结果表明，PID随面内垂直纤维方向和面外法线方向的热膨胀系数增加而增大。数据挖掘方法对PID预测足够准确，其效率为工程应用提供了额外的计算选项。
