Loading Path Design for Grain Refinement in Hot Forming Process of Ti-6Al-2Zr-1Mo-1V Alloy

Grain refinement is crucial for enhancing the microstructure and mechanical properties of materials. This study employed a new loading method for grain refinement in Ti-6Al-2Zr-1Mo-1V alloy. A particle swarm optimization-backpropagation neural network model was developed to precisely capture the complex rheological behavior of the alloy. The model exhibited outstanding correlation (from 0.99633 to 0.99998). A detailed hot processing map was constructed using experimental and model data. By combining microstructure characteristics, a microstructure deformation mechanism map was developed to identify and isolate the stable dynamic recrystallization fine-grain regions. However, these regions remain chaotic and require further discretization design. By employing finite element simulation, the optimal strain rates of discretized elements were determined, and then, the 3D response surface was fitted. The loading path was ultimately verified through isothermal compression experiments and microstructure analysis. Results indicate that grain sizes under the optimal loading paths are smaller than before, strongly confirming the effectiveness of loading path. Specifically, at 1173 K, grain size under the optimized loading path was 10.45 mu m, and at 1223 K, it was 12.15 mu m.