Smart Reverse Engineering and Predictive Process Optimization of Titanium Intermetallic Composites For Turbine Blades using Artificial Intelligence and RSM

Lightweight, high-temperature-capable turbine blades are essential for increasing both efficiency and reliability in gas turbine engines. The purpose of this research is to test the thermal behavior, mechanical characteristics, and corrosion resistance of Ti-Al-Nb hybrid alloys for use in turbine blades. A reverse-engineering strategy was used to identify critical thermal and stress points and determine performance requirements through 3D scans, Computer-Aided Design (CAD) drawings, and finite element analysis (FEA). The various compositions of Titanium-Aluminum-Niobium (Ti-Al-Nb) samples were cured in separate heat-treatment processes with different aluminum and niobium contents and cooling rates. Process optimization used Response Surface Methodology (RSM), while a Physics-Informed Neural Network (PINN) with BiLSTM, SHAP, and Zeroth-Order Optimization (PNBM-SHAP-ZOA) was created to predict material properties. Optimum conditions were determined to be 26.3% Al, 11.9% Nb, 850 ℃ curing temperature, and 9.93 ℃/s cooling rate, resulting in a tensile strength of 1098.39 MPa, fracture toughness of 24.13 MPa√m, hardness of 380.52 HV, and weight gain due to oxidation of 0.16 mg/cm 2 . The proposed prediction framework yielded lower prediction errors than those achieved with ANN-GA, CNN-PSO, or traditional RSM, thus proving to be an effective model for designing Ti-Al-Nb turbine blade materials with exceptional performance.