Abstract
Realizing stable and highly efficient operation of blast furnace under low reducing agent rate condition is one of the key issues to contribute problems of energy, resources and global warming. Under low reducing agent rate operation, chemical and thermal driving force in the blast furnace is weakened, and it is considered that the capability of recovering the process from the fluctuation is deteriorated. Therefore the probing the state inside the blast furnace by utilizing the information from the various sensors is important. This study focused on the relation between pressure distribution on furnace wall and packing structure inside the furnace. The blast furnace is a packed bed reactor with upward gas flow, thus the pressure generally decreases from the tuyere to the top. Partially inverse pressure distribution is sometimes observed as an instable phenomenon of the furnace. This study proposed a mechanism to produce such pressure distribution with simple structural change in packed bed. When a vacant space is formed in the packed bed, the gas flow concentrates to this space and the gas velocity increases. At the downstream end of the space, the gas velocity decreases and generates dynamic pressure. This increase of the pressure forms the inverse pressure distribution. In this study, this mechanism was confirmed through experimental and numerical approaches.
Introduction
In a standard blast furnace—a massive packed-bed reactor—gas is blown in from the bottom (tuyeres) and travels upward through layers of iron ore and coke. Under normal conditions, physics dictates that pressure should steadily decrease as the gas moves toward the top. However, operators often observe "inverse pressure distributions" where sensors show higher pressure at an upper point than at a lower one. This phenomenon is a critical red flag for operational instability, such as "hanging" (where the burden stops descending) or "channeling" (where gas bypasses the burden).
Methodology
The researchers investigated this anomaly through three primary methods:
- Theoretical Modeling: Application of the Ergun equation to model flow resistance and development of a simplified "parallel flow path" model.
- Cold Model Experiments: A rectangular transparent box filled with glass beads (5 mm and 1 mm diameters) to simulate the packed bed. Wire mesh cages were used to create "void spaces" or Highly Permeable Zones (HPZs) of varying sizes.
- Numerical Analysis: 3D CFD (Computational Fluid Dynamics) simulations using the SIMPLER method to visualize gas velocity vectors and isobars that are otherwise invisible in physical experiments.
The Acceleration-Deceleration Mechanism
The core insight of this study is the "nozzle effect" created by internal structural defects. A void space acts as a path of least resistance, causing gas from the surrounding packed bed to be "sucked" into the void, which leads to a massive increase in local gas velocity (acceleration).
As the high-velocity gas hits the dense packed bed at the top of the void, it is forced to decelerate rapidly. According to Bernoulli’s principle, this loss of kinetic energy is converted into dynamic pressure. This localized pressure spike at the downstream end of the void can be large enough to manifest as an inverse pressure distribution on the furnace wall.
Key Results and Findings
- Validation: Numerical simulations closely matched experimental results, confirming that the acceleration-deceleration mechanism is physically sound.
- Geometry Sensitivity: The inversion is most pronounced when the void is longitudinally long and perpendicularly narrow. A narrower void forces higher gas velocities, leading to a larger pressure recovery at the exit.
- Flow Concentration: Simulations showed that even a void taking up only 25% of the cross-sectional area could attract up to 94% of the total gas flow, illustrating why small structural defects can wildly distort wall sensor readings.
- Cohesive Layer Influence: When the void is surrounded by a low-permeability layer (simulating the "cohesive zone"), the gas concentration effect is amplified, making the pressure inversion even more distinct.
Conclusion
This study demonstrates that inverse pressure distributions are not mere sensor errors but are physical indicators of internal void formations or permeability changes. By analyzing the specific shape and magnitude of these inversions, operators can potentially "reverse-engineer" the internal state of the furnace, allowing for more precise real-time adjustments to maintain stability and efficiency.
Summary updated