Optimizing System Reliability in Additive Manufacturing Using Physics-Informed Machine Learning

Fused filament fabrication (FFF), an additive manufacturing process, is emerging technology with issues in the uncertainty of mechanical properties and quality printed parts. The consideration all main interaction effects when changing print parameters not efficiently feasible, due to existing stochastic dependencies. To address this issue, a machine learning method developed increase reliability by optimizing input predicting system responses. A structure artificial neural networks (ANN) proposed that predicts response based on observations similar systems. In way, significant for reliable can be determined. ANN part physics-informed pretrained domain knowledge (DK) require fewer full training. This includes theoretical idealized systems measured data. New predictions made without retraining but using further from predicted system. Therefore, are available real time, which precondition use industrial environments. Finally, application bed adhesion FFF discussed evaluated.