**1. Introduction**

A blast furnace is a reactor with approximately 5000 m<sup>3</sup> of volume to produce pig iron from ore particles. Iron ore (sinter, lump, and pellet) and coke particles are alternately stacked at the topmost layer of the blast furnace, and the hot gas is blown from tuyeres at the bottom of the furnace. Iron ore particles are reduced during the descent, and numerous physical changes and chemical reactions occur between each phase over this period. Thus, it is an extremely complicated system, resulting in the possibility of unfavorable phenomena or serious problems occurring. To avoid these problems, controlling and stabilizing the gas flow in the furnace is of utmost importance because the gas plays a key role for the reduction and the heat source. Thus, keeping the gas flow in the steady state leads to an efficient and low RAR (reducing agent ratio) operations. Therefore, controlling a void fraction in the stacked layer, i.e., the burden distribution at the top of the blast furnace, is the most effective operation for stabilizing the gas flow. Much research has been experimentally conducted and some mathematical models have been proposed to estimate and control the burden distribution [1–5]. These models can give useful information in the daily operation. However, many kinds of particles are usually mixed in ore charging to help the reduction; therefore, it is necessary to analyze the individual

solid particles' behavior in the blast furnace for an in-depth analysis of several phenomena that were previously mentioned. The discrete element method (DEM) [6] is one of the most reliable simulation methods for analysis of the solid particle behavior, and an approach using the computational simulation based on DEM is extremely useful to grasp the phenomena found in the charging process of the blast furnace. Some studies on the modeling of solid flow in a blast furnace have been previously reported, for example, the raceway [7,8], solid flow in the blast furnace [9–11], gas–solid flow [12], hopper flow [13,14], and particle trajectory from the rotating chute [15], and authors have developed burden distribution simulators using DEM [16–20]. Validations of the particle trajectory discharged from the rotating chute were studied, and they showed good correlations [19,20]. This simulator still remains a key issue to validate the simulated burden distribution with experimental results in detail.

In this paper, charging tests were performed using a 1/3-scale experimental burden distribution simulator of the blast furnace to validate the simulated results, and the particle size segregation and ore to coke mass ratio (O/C) were investigated. Furthermore, the particle flow during charging into the experimental blast furnace was modeled using DEM, and the results were compared with the experimental ones.
