The solidification and micro- and macro-segregation behaviors of as-cast INCONEL 718 (IN718) alloy at different temperatures under a slow cooling rate (5 °C/min) were investigated in this study. The results indicate that the solid-liquid interface grows into reticulation of hexagons during solidification. The variation trend of the solid fraction and transition rate of the solid phase with solidification time can be well characterized by the Boltzmann and Gaussian distribution, respectively. The order of segregation degree of negative segregation elements is: Fe > Cr > Al. Nb is the most principal positive segregation element, which is abundant in the long-term unsolidified remaining liquid. At the terminal stage of solidification, the increasing tendencies of the Nb and Mo contents in the liquid and the residual liquid density with decreasing temperature reverse due to the formation of the Laves phase. The freckles are most likely to form in the early stages of solidification, at which the liquid fraction is between 0.3 and 0.2, and the temperature range is about 1320 °C to 1310 °C. The information produced is expected to characterize the solidification and segregation behaviors of IN718 alloy when cooled at a slow rate characteristic of larger ingots typical of those required for industrial gas turbines and aircraft engines.

In this paper, the precipitation thermodynamics and growth kinetics of TiN inclusions in GCr15 bearing steel during solidification were calculated in more detail. A more reasonable formula for calculating the segregation of the solute elements was adopted and the stability diagram of TiN precipitation considering solidification segregation was given. By solving equations, the change of the solute element content before and after TiN inclusion precipitation was calculated, and the results were substituted into the kinetic formula of the inclusion growth, which made the kinetic calculation more accurate. Results showed that the most effective way to reduce the precipitation of TiN is to increase the cooling rate and decrease the contents of Ti and N in steel. The effect of Ti content on the size of TiN inclusions is greater than that of N content.

The behavior of MnS, a significant non-metallic inclusion in steel with low sulfur content, is poorly understood, in particular, how the MnS inclusion becomes a favorable secondary phase. To clarify how the sulfur content in DH36 structural steel affects the behaviors of MnS, and the steel properties, samples containing 20–380 ppm sulfur were melted, and the ingots were rolled in a semi-industrial single-roll mill. According to experiments using optical microscopy (OM), scanning electron microscopy (SEM), an automatic inclusion analysis system, and transmission electron microscopy (TEM), the morphology of MnS inclusions changed from thread-shaped to spherical- or spindle-shaped with decreasing sulfur content, and their size and number also decreased. The steel with 20 ppm S contained much more nanoscale MnS, whose pinning effect created a fine grain size in the steel. The reason is that the temperatures of maximum nucleation rate and fastest precipitation of MnS in this sample (900 °C and 920 °C, respectively) are much lower than the soaking temperature. Compared to samples with 70–380 ppm sulfur, the sample with 20 ppm sulfur content demonstrated the highest yield strength, tensile strength, and impact energy. Finally, the industrially produced DH36 samples were analyzed, and the steel with the lower S content had a higher stability and yield strength, in agreement with our laboratory results.

The secondary-phase particles in metals, particularly those composed of non-metallic materials, are often detrimental to the mechanical properties of metals; thus, it is crucial to control inclusion formation and growth. One of the challenges is determining the three-dimensional morphology and inner structures of such inclusions. In this study, a non-aqueous solution electrolytic method and a room-temperature organic technique were developed based on the principle of electrochemistry to determine the three-dimensional morphologies and inner structures of non-metallic inclusions in Al-killed steel, Si-killed steel, and ductile cast iron. The inclusions were first extracted without any damage to the inclusions, and then the collected inclusions were wrapped and cut through Cu ion deposition. The results revealed that the inclusions in Al-killed steel had an irregular morphology, that those in the Si-killed steel were mainly spherical, and that almost all the spheroidal graphite in the ductile cast iron featured a uniform globular morphology. In addition, nucleation was not observed in the inner structures of the inclusions in the Al-killed steel, while some dendritic or rod-like MnS phase precipitates appeared on the silicate inclusion surfaces, and some silicate-rich phases were detected in their inner matrix. For spheroidal graphite, rare-earth oxides (one particle or more) were observed as nuclei in the center of almost every graphite particle. The formation and evolution of inclusions in these types of metals can be better understood by means of the two developed methods.

The highly stochastic nature of rainfall presents significant challenges for the accurate prediction of its time series. To enhance the prediction performance of non-stationary rainfall time series, this study proposes a hybrid deep learning forecasting framework—VMD-IWMA-BiLSTM—that integrates Variational Mode Decomposition (VMD), Improved Whale Migration Algorithm (IWMA), and Bidirectional Long Short-Term Memory network (BiLSTM). Firstly, VMD is employed to decompose the original rainfall series into multiple modes, extracting Intrinsic Mode Functions (IMFs) with more stable frequency characteristics. Secondly, IWMA is utilized to globally optimize multiple hyperparameters of the BiLSTM model, enhancing its ability to capture complex nonlinear relationships and long-term dependencies. Finally, experimental validation is conducted using daily rainfall data from 2020 to 2024 at the Xinzheng National Meteorological Observatory. The results demonstrate that the proposed framework outperforms traditional models such as LSTM, ARIMA, SVM, and LSSVM in terms of prediction accuracy. This research provides new insights and effective technical pathways for improving rainfall time series prediction accuracy and addressing the challenges posed by high randomness.