Reproducibility Dataset for "Defect-Aware Thermoforming of Continuous-Fiber Thermoplastic Composites: A Physics-Guided Bayesian Decision Layer for Resource-Aware Process-Window Discovery"

This dataset supports the manuscript “Defect-Aware Thermoforming of Continuous-Fiber Thermoplastic Composites: A Physics-Guided Bayesian Decision Layer for Resource-Aware Process-Window Discovery.” The primary dataset contains a 4,000-design physics-guided emulation benchmark comprising 1,000 Latin-hypercube candidate designs at each of four forming times (1.6, 3.2, 4.8, and 6.4 s). The emulation responses are not direct extractions of the archived FE VTK fields.