Explainable CatBoost–TabPFN Hybrid for Mode-Resolved Interlaminar Fracture Toughness Prediction in Fiber-Reinforced Polymer Composites

Fiber-reinforced polymer (FRP) composites offer high specific strength, high stiffness, and good corrosion resistance, which makes them suitable for aerospace, automotive, marine, and wind-energy structures. Their interlaminar fracture behavior is harder to pin down because delamination testing is slow and costly and the response shifts with loading mode. This study presents an explainable hybrid machine learning framework for interlaminar fracture toughness under Mode-I, Mode-II, and Mixed-Mode loading. A weighted ensemble pairs Categorical Boosting (CatBoost), a gradient-boosting method, with the Tabular Prior-Data Fitted Network (TabPFN), a pretrained transformer built for tabular data. A Tree-structured Parzen Estimator (TPE) tunes the CatBoost hyperparameters and the ensemble weight against specimen-grouped cross-validation. The work draws on the public D-STANDART experimental dataset and treats DCB, ELS, and MMB as three separate prediction problems for GIC, GIIC, and GC rather than pooling them into one model. The analyzed D-STANDART specimens comprise pultruded carbon-fiber-reinforced polymer (CFRP) material based on Zoltek PX35 carbon fiber. The data are partitioned at the specimen level, so no specimen contributes records to more than one subset, and feature selection uses the training data alone. SHapley Additive exPlanations (SHAP) then measure how each input shifts the predicted toughness in each mode. The framework reached an R2 of 0.9280 for DCB, 0.9870 for ELS, and 0.9627 for MMB, with RMSE values of 5.1759, 1.5041, and 2.8809 and MAE values of 1.6502, 0.5372, and 0.7000, respectively. Bringing a tabular foundation model into delamination analysis, and keeping the three loading modes apart, lets the framework report mode-specific feature behavior instead of one averaged picture.