Predicting the ultimate grain size of aluminum sheets undergone constrained groove pressing

AbstractThe grain size of constrained groove pressed aluminum has been predicted through the genetic programming approach. “Sheet thickness,” “elongation,” “yield strength,” “ultimate tensile strength,” “total strain,” and “hardness,” along with “primary grain size” of the ultrafine-grained sheets were utilized as input parameters to obtain the ultimate grain size. A total number of 73 available data in the literature were gathered and randomly divided into 60 and 13 sets for algorithm training and testing, respectively. Among the presented models, the one with best performance utilized parameters of total strain, ultimate tensile strength, and primary grain size with 40 chromosomes, 10 head sizes, 2 genes, addition as linking function, and 6 algebraic functions. Being in the excellent-fitness region, minimum errors in both training and testing phases, lower level of complexity, and minimum residual summation of predicted and examined grain sizes were all considered together as criteria for selecting the best model performance. Utilization of both primary grain size and total strain declared that they were inevitable governing parameters while modeling. It was concluded that the predicted results using 28 proposed models and experimented data had good conformity, while all proposed models had reliable applicability for predicting the ultrafine grain sizes. Graphical Abstractᅟ