Machine learning-aided jet lag prediction for high-accuracy fabrication of melt electrowriting fibrous scaffolds

Melt electrowriting (MEW) is an emerging additive manufacturing technology capable of depositing continuous micro- and submicron-fibers for the fabrication of high-resolution fibrous scaffolds. In this process, jet lag is a distinctive phenomenon that critically affects both process stability and printing accuracy, yet its real-time measurement remains challenging, especially during nonlinear printing. This study developed a machine-learning-assisted framework to predict jet lag length in MEW from five key process parameters: voltage, collector speed, material temperature, nozzle-to-collector distance, and air pressure. Multiple linear regression and random forest were comparatively evaluated. Both models achieved effective prediction, while random forest showed higher predictive accuracy. This improvement is attributed to its greater flexibility in representing potential nonlinear dependencies within the current dataset. The predictive performance of random forest was further validated through the printing of high-accuracy nonlinear fiber patterns. These findings demonstrate the feasibility of data-driven jet lag prediction and provide a practical framework for improving printing accuracy in MEW.