Search

The determination of the relationships between rolling parameters and mechanical properties by high-precision deep learning is important for predicting and controlling the yield strength (YS) of hot rolled steels, however, for most steels, their YS-labeled samples are insufficient for the modeling. In this study, a YS-labeled small dataset and an unlabeled big dataset were firstly collected and then reconstructed the rolling parameters to be microstructures by the physical metallurgical principles. Before being used to label the unlabeled dataset with calculated YS, the strengthening mechanism-based compositions-microstructures-property linkage was optimized by combining the labeled dataset and the particle swarm optimization (PSO) algorithm. To precisely predict YS, the deep neural network (DNN) was initially pre-trained by the big dataset labeled by calculated YS, and then fine-trained by the small dataset labeled by measured YS. Based on it, the effects of compositions, microstructures, and rolling parameters on YS were analyzed, and the results proved to be in good agreement with experimental observations, which presents the solution to implement deep learning by using physical metallurgical principles with a small dataset.

Determining the exact relationships between rolling parameters and mechanical properties is important to predict and even control the mechanical properties of hot rolled Ti micro-alloyed steel products. In this paper, two data dimension reduction strategies, respectively guided by physical metallurgical (PM) principles and data-driven (DD) strategy were proposed to preprocess the high dimensional steel composition and rolling parameters, and a comprehensive set of machine learning (ML) models was developed to predict yield strength (YS) and elongation (EL) combined with the reconstructed inputs. The artificial neural network (ANN) with the inputs including dislocation density, fractions, and sizes of ferrite and TiC precipitate transferred by PM principle outperformed that with the inputs extracted low variance feature elimination (LVFE) and autoencoder (AE) algorithms. Based on the proposed model, the effects of alloying elements, rolling temperatures, and microstructures on mechanical properties were comprehensively analyzed. Also, the variations of microstructures and mechanical properties over the strip length were discussed, which were consistent with the measured results. [Display omitted] • PM-guided dimension reduction was proposed to transfer rolling parameters to microstructures. • The dimension reduction guided by PM performed better than that guided by DD methods with LVFE and AE. • The predicted effects of process and microstructures on YS and EL were consistent with experiments. • The predicted variations of YS and EL over strip length were consistent with measured results. [ABSTRACT FROM AUTHOR]