An intelligent system for predicting HPDC process variables in interactive environment

The selection of optimal parameters in high pressure die casting process (HPDC) has been long recognized as a complex nonlinear problem due to the involvement of a large number of interconnected process variables, each influencing the flow behavior of molten metal inside the die cavity and thus part quality and productivity. In the present work a physical model called Neural Network based Casting Process model (NN-CastPro) has been developed for real time estimation of optimal HPDC process parameters. By submitting a set of four process parameters (having major impact on productivity and part quality) namely, (i) inlet melt temperature, (ii) mold initial temperature, (iii) inlet first phase velocity and (iv) inlet second phase velocity, as input to the NN-CastPro, values for filling time, solidification time and porosity can be obtained simultaneously. The proposed artificial neural network (ANN) model was trained using data generated by ProCast (an FEM-based flow simulation software). The obtained prediction accuracy and enhanced functional capabilities of NN-CastPro show its improved performance over other models available in the literature.