Figure 2
The workflow diagram is arranged from left to right using outlined circles connected by right-pointing arrows. The first circle on the left contains the text “Experiment” with the word “Foundation” centered below it. A right-pointing arrow leads to a larger overlapping double circle labeled “F E M” at the top, containing two smaller inner circles. The left inner circle contains the text “Ply-level model,” and the right inner circle contains the text “Validation.” A right-pointing arrow leads to a larger circle labeled “Machine Learning” at the top. Inside this circle, there are two ovals on the left, and between the ovals, there is the text “Neural network” at the center. Above this text, there is a curved arrow pointing from left to right. Below the oval, a curved arrow labeled “Training” loops from right to left and then upward toward the left side of the oval. To the right of the oval is a smaller circle labeled “Impact response prediction.” A final right-pointing arrow leads to italic text reading “Fast predictive surrogate for impact-resistance evaluation.”Scheme of the proposed hybrid experiment-FE modeling-machine learning methodology for impact response prediction of TPU composites