Improved metamodels for the optimization of high-pressure die casting process

The increasing need of improving quality in high-pressure die casting (HPDC) has been boosting the development of models to predict the final outcome and to optimize the process. It is hence promising the use of simplified behavioural models, denoted metamodels, which are abstract descriptions of HPDC. These models represent the relation between some influential process parameters and the final quality, with algebraic explicit equations obtained through least-square regression. The recent researches carried out by the Authors have identified these influential parameters, which represent the physical phenomena responsible for the casting quality. In contrast with the literature, these parameters employ physical quantities that are not instantaneous values of directly measured variables, but rather exploit signal processing and have an integral nature. These results solve the crux in the synthesis of metamodels, i.e. the definition of the correct independent parameters, which is here solved by transforming those adopted for the design of experiment (DOE), through nonlinear relations based on physical considerations. Besides showing some experimental results, the paper outlines a novel paradigm to HPDC optimization. The most influential parameters related to the plunger motion are, in turn, represented through analytical models. Then, once that these parameters are properly chosen, process control can be performed through the injection machine controller and process can be optimized by selecting in advance the best parameters.