Probability Prediction of Curing Process-Induced Deformation for V-Shape Composite Structures Based on FEM Method and...

Area of Science:

• Materials Science
• Mechanical Engineering
• Composite Manufacturing

Background:

• Continuous fiber-reinforced composites offer high specific modulus and strength, making them valuable in industry.
• Process-induced deformation (PID) is a critical manufacturing challenge, leading to unpredictable part variations.
• Accurate prediction of PID is essential for quality control and performance optimization.

Purpose of the Study:

• To develop a probabilistic prediction method for PID in V-shape composite structures.
• To investigate the influence of material parameter uncertainties on PID.
• To validate the prediction accuracy using experimental data.

Main Methods:

• A sequential coupling thermal-chemical-mechanical finite element (FE) model was developed in ABAQUS.
• Material parameter uncertainties were incorporated for V-shape structures with varying radii and thicknesses.
• A decision tree model was trained on FE data for PID sensitivity analysis and probability distribution prediction.

Main Results:

• PID was found to increase with coefficients of thermal expansion in specific material directions.
• The prediction accuracy of the included angle was validated against experimental results.
• The data-mining approach demonstrated high accuracy and efficiency for PID prediction.

Conclusions:

• The developed FE and data-mining approach accurately predicts PID in V-shape composite structures.
• This method offers an efficient computational option for engineering applications.
• Understanding PID drivers is crucial for enhancing composite manufacturing processes.