The nozzle clogging behavior of Ti-bearing IF steel was studied by metallographic analysis, scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), and X-ray diffraction (XRD). According to the experimental results, nozzle clogging primarily appears three layers. There are a lot of large-sized iron particles in the inner layer and mainly slag phase in the middle and outer layers. The principal clog constituents of the inner layer are loose alumina cluster inclusions and granular shaped alumina inclusions, containing iron particles. The clog constituents of the middle layer are mainly dendrite alumina inclusions. The primary phases existing in nozzle clogging are FeO·TiO 2 and FeO·Al 2 O 3 besides α-Al 2 O 3 and α-Fe. The FeO·TiO 2 phases among the deposits adhere the deposits together firmly enough to lead to the inferior castability of Ti-bearing ultra low carbon steel compared with that of Ti-free low carbon Al-killed steel.

Before deoxidation, the main inclusions were Fe-Mn-O inclusions in Ti-bearing ultra low-carbon steel melt. After 3 min Al addition, the inclusions changed to be granular and spherical Al 2 O 3 . Al 2 O 3 cluster, composed of granular and spherical particles with diameter 1-2 mm, formed in 7 to 10 min after Al addition into the melt.

No abstract

No abstract

Combined model in which a tundish is divided into plug flow, well‐mixed flow, and dead region has been widely used for the analysis of residence time distribution curve in an actual tundish. However, the calculation of dead volume in conventional or Sahai's method does not follow its definition in the model assumption, which probably results in incorrect conclusions for flow field optimization of tundish. In this study, two methods for dead volume calculation are discussed, the well‐mixed volume is further divided into equivalent, active, and dead well‐mixed volume. Therefore, the method for dead volume calculation is revised by eliminating the contribution of dead well‐mixed flow. Water modeling experiments of single and multi‐strand tundish are conducted. The new calculation method is applied to analyze the residence time distribution curves and calculate the dead volume fraction in comparison with the other two methods. The results show after the dead well‐mixed flow is eliminated, the dead volume fraction calculated by the new method is smaller than those calculated by conventional and Sahai's method. According to the flow visualization experiments, the dead volume fraction calculated by the new method in this study is more rational than those calculated by the other two methods.