2. Methodology

and mass transfer reveals the complexity of a blast furnace and the difficulties to comprehend the process. Since measurements in a blast furnace are very difficult to carry out due to size and limited accessibility, mathematical modelling based on computational fluid dynamics, reaction kinetics and transport phenomena is an applicable tool to investigate into the complex

Multiphase flow phenomena consisting of a fluid, e.g. gas or liquid and a solid phase, could be classified into main categories: On a macroscopic level, all phases are considered as a continuum that has resulted in the well-known two-fluid model [1]. It is a preferred approach due to computational efficiency and convenience for a variety of engineering applications such as process engineering. However, inherent to the averaging concept in the continuum approach, major features of the particulate phase, e.g. material properties, size distribution or shape of individual particles, are lost. Therefore, this gap of information on the particle level usually is

Alternatively, the particles are treated as discrete entities, while the fluid phase in the void space between the particles is still considered as a continuum and, therefore, is referred to as the Combined Continuum and Discrete Model (CCDM) [2, 3]. It bears the valuable advantage that constitutive relations are avoided and the particulate phase is resolved on a fundamental level. An analysis of results related to particles offers a finer resolution than a continuum method and, therefore, leads to a deeper understanding of particle processes which is confirmed by Zhu et al. [4, 5] in a review for the applicability of the CCDM approach. It has developed significantly during the last two decades and represents the dynamics of the particles by the discrete element method (DEM), while the remaining continuum phases are described by differential conversation equations. Hence, CCDM is established as a powerful tool to describe the complex interaction between particles and a fluid as stated by Yu and Xu [6], Feng and Yu [7] and Deen et al. [8]. Although CCDM is the method of choice to reveal underlying physics for challenging multiphase flow applications as reviewed by Zhu et al. [4, 5], current methodologies should be developed for nonspherical particle shapes to accom-

modate efficiently engineering needs and to quantify results for process modelling.

late phase in a gas-fluidised bed as shown by Feng and Yu [13] and Kafuia et al. [14].

At an early stage, only simple multiphase configurations have been investigated [9, 2] that were extended to more and more complex engineering applications such as conveyor belt, cyclone and fluidised bed as demonstrated by Chu and Yu [10]. Chu et al. [11] described the complex flow of water, air, magnetite and coal particles of various sizes in a dense medium cyclone (DMC). Similarly, Zhou et al. [12] were able to model rich/lean combustion of pulverised coal in a complex burner with CCDM. Both predictions agreed remarkedly well with experimental data. CCDM is also well suited to describe the chaotic motion of a particu-

A similar development has been seen for modelling of blast furnaces. Computational fluid dynamics (CFD) as a tool for continuous flow modelling has been applied with success to a large extent as reviewed by Chattopahyay et al. [15, 16]. However, experimental observations [17–21] show that a pure continuous approach to the blast furnace is inaccurate. Therefore, the flow of the solid phase consisting of particles is to be modelled by a discrete approach as suggested by Dong et al. [22] and reviewed by Yu [23] for several engineering applications. Simsek et al. [24] predicted grate ring systems by the CCDM method but obtained only

compensated for by additional closure or constitutive relations.

characteristics of a blast furnace.

126 Iron Ores and Iron Oxide Materials
