Direct Ink Writing State Prediction Based on Integrated Machine Learning Model Combining Materials Properties and Printing Parameters

ABSTRACT BaTiO 3 /Polydimethylsiloxane (BaTiO 3 /PDMS) polymer‐based ceramic composites are favored for their exceptional piezoelectric and ferroelectric properties. Although direct ink writing 3D printing provides a viable approach for fabricating geometrically complex composite structures, the printability of such composites fundamentally relies on synergistic optimization of ink rheological behavior and deposition parameters. This study proposes a machine learning‐driven framework that establishes an integrated printing state prediction model combining material properties and process parameters. The rheological properties of composite materials with varying matrix ratios and ceramic mass fractions are studied, and the printing state is analyzed under different process parameters. A particle swarm optimization‐enhanced support vector machine (PSO‐SVM) algorithm was developed using 669 groups of experiments for training and testing, achieving 92.48% prediction accuracy in printability assessment through intelligent parameter optimization of penalty factors and kernel function widths in the testing set. The validity of the integrated material‐printing parameters machine learning model was further demonstrated through the successful fabrication of representative structures, demonstrating its robust predictive performance.