Predicting the evolution of microstructure and field quantities under varying processing and loading conditions is a central challenge in computational materials science and metal additive manufacturing (AM). While deep learning (DL) methods offer ultra-fast prediction capabilities post-training, existing models often struggle with poor spatial and temporal extrapolation, high parameter burdens, and an inability to effectively integrate diverse conditioning parameters alongside high-dimensional input fields. To address these bottlenecks, we propose a novel conditional generative adversarial network (CPGAN), which is designed to seamlessly ingest both initial fields and governing condition parameters. The CPGAN framework offers three distinct advantages: (1) it accurately maps the combined effects of initial states and process conditions onto evolved fields; (2) it demonstrates robust extrapolation capabilities across diverse spatial and temporal scales, including the unique ability to natively generate high-resolution rectangular domains; and (3) it achieves superior predictive accuracy and training stability compared to standard convolutional baselines by effectively suppressing spurious artifacts. We validate CPGAN’s performance against rigorous physics-based ground truths across three representative engineering applications: porosity evolution in selective laser sintering (SLS), spatial distribution of 2D von Mises stress fields in solid structures, and the spatiotemporal evolution of grain growth. The results confirm that CPGAN is a highly adaptable and efficient surrogate model, capable of simulating continuous structural and morphological evolutions even when driven by highly non-uniform spatial or temporal kinetics.

Fused deposition modeling (FDM) is one of the most widely deployed additive manufacturing methods, and the mechanical performance of FDM-printed parts is governed by a small set of strongly coupled process parameters. Silk PLA, a PLA-based filament engineered to deliver a high-gloss finish with improved mechanical performance, has received far less attention than commodity PLA, and its parameter–property relationships remain incompletely characterized. In this work, the effects of three FDM printing parameters: (i) layer height (0.10, 0.15, and 0.20 mm), (ii) extrusion temperature (200, 210, and 220 °C), and (iii) print speed (100, 120, and 140 mm/s) on the tensile characteristic of Silk PLA were investigated through a full-factorial design consisting of 27 parameter combinations and 135 ASTM D638 Type-I specimens. Tensile tests were performed on an MTS E42 universal testing frame, while Digital Image Correlation (DIC) was employed to obtain full-field longitudinal and transverse strain distributions and to identify the onset and location of necking. Analysis of variance (ANOVA) was used to assess the statistical significance, while an interpretable machine learning (ML) pipeline combining extreme gradient boosting (XGBoost), Shapley additive explanations (SHAP) values, and partial dependence plots (PDPs) was employed to quantify the relative influence of each parameter and elucidate its effect on the tensile response. Both analyses identified extrusion temperature as the dominant factor governing ultimate tensile strength and Young’s modulus. Partial dependence analysis further revealed that strength gains saturate above 210 °C and are maximized at an intermediate print speed of 120 mm/s, providing actionable guidance for process optimization. The highest tensile strength, 40.68 MPa, was achieved at a layer height of 0.20 mm, an extrusion temperature of 220 °C, and a print speed of 120 mm/s. DIC measurements showed that thinner layers (0.10 mm) produced higher breaking strains and necking that initiated at the gauge-section edges, whereas thicker layers (0.20 mm) shifted necking toward the mid-gauge. Together, the experimental, statistical, and ML results provide a consistent, mechanistically interpretable framework for optimizing FDM process parameters for Silk PLA functional parts.