Small sample multi-objective evolutionary optimization method for components of high-toughness nodular cast iron

Abstract

The invention provides a small sample multi-objective evolutionary optimization method for high-toughness nodular cast iron components, and belongs to the technical field of material component optimization design, the method comprises the following steps: collecting nodular cast iron components in literatures and experimental data of corresponding performance, and carrying out standardization processing; a nodular cast iron component and performance prediction model is constructed based on small sample machine learning, prediction accuracy is improved by combining feature engineering and a data enhancement strategy, and a two-channel prediction model is obtained after training; constructing a virtual performance evaluation environment, and setting a reasonable component range, carbon equivalent and other physical constraint conditions; carrying out optimization search under physical constraint conditions by utilizing a reverse optimization algorithm to obtain a Pareto optimal nodular cast iron component combination meeting tensile strength and elongation targets, and forming a self-learning closed loop; the method solves the problems that in existing high-toughness nodular cast iron component optimization, sample data quality is uneven, multi-target needs cannot be balanced, and a complete process is lacked.

Images