Study of gas-fine flow in a packed bed with an application to the ironmaking blast furnace
The iron and steelmaking industry uses coal as the primary reducing agent. The carbon that is the major component of coal is finally released into the environment as carbon dioxide. One of the effective ways to reduce the coke consumption and thus the reduction in greenhouse gases is to introduce coal or other carbonaceous materials through the tuyere. However, this can be done only up to some extent as injection of these materials reduces the bed permeability which in turn affects the operation and productivity of the blast furnace (BF). Previous studies have shown that at a higher pulverised coal injection (PCI) rate and depending on the operating conditions of the furnace, some amount of coal remains unburnt and consequently the ashes and coal particles, in the form of powders, may be entrained in the gas stream or be deposited into the lower zones of the BF. However, the physics of the pulverised coal within the furnace is still not well understood, especially in presence of raceway and tuyere protrusion. Raceway is very important in iron-making BF as its size and shape determines the aerodynamics of BF and thus the heat and mass transfer. Therefore, the study of fine flow in a packed bed is a necessary precursor to understand the above-mentioned processes/phenomena. For the validation purpose, two-dimensional experimental studies were performed at room temperature on the various packed bed such as a rectangular bed with and without a cohesive zone and a cylindrical bed without a cohesive zone. Both the packing and fine particles were made of glass material. Steady state of experiment with fines was achieved based on the recently developed accurate mass balance method which dictates that fines into the system must be equal to fines out of the system. Most research neglect vital characteristics of BF, such as lateral inflow, presence of tuyere protrusion, the raceway’s shape and size, and cohesive zone leading to a significant incongruence with the BF being modelled. The present work incorporates these features into a 2D numerical study of a gas–fines–solid system. The mathematical modelling considers the gas and fines as a Eulerian–Eulerian system with the constant voidage model for the solid phase representing the packing particles. In presence of a cohesive zone, the cohesive blocks were assigned zero porosity. Well-established theoretical relations and correlations are used to determine the inter-phase forces and fines accumulation regions. Particular emphasis is placed on the accurate representation of the raceway formed at the tuyere exit, and three approaches are considered, viz., its absence, a correlation-based prediction, and an iso-stress-based model. The sensitivity analysis is done with a marked interest in the raceway shape and size and accumulation profiles of the fines which are important parameters in the overall flow characterization. Additionally, structural parameters such as cohesive zone configuration, cohesive block porosity, and tuyere protrusion are also varied, and their effect on the raceway and static holdup profiles are characterised. The sensitivity analysis shows that the raceway shape and size play a vital role in the flow of gas-fines and influence the accumulation of fines. While most of the parameters varied have a noticeable effect, certain factors such as fines feed rate and size exhibit negligible consequences. The model predicts the static holdup profile through a correlation based on experimental data. Developed model is able to predict the pressure loss reasonably well against experimental and published data for both clean gas and gas–fine flow in a packed bed with a lateral and bottom injection of gas and fine. Similarly, the model is able to predict correctly the raceway shape & size, pressure profiles in both vertical and horizontal directions as well as fines static holdup both in the presence and absence of cohesive blocks. All the results indicate an excellent agreement with the experimental data and lend credibility to the model and solver.