CFD investigation of industrial cyclone preheaters

Cyclone preheaters are important components in cement and stone wool production plants. They are used to preheat the raw material by recovering the thermal energy of the high-temperature flue gas produced in a kiln or a melting reactor. Despite their deceptive simplicity, the complicated multiphase flow field within cyclone preheaters makes their design and optimization challenging. Conventionally, the design of cyclones is based on empirical approaches developed for simplified, generic, and small-scale cyclones which are not reliable for industrial-scale cases. Therefore, Computational Fluid Dynamics (CFD) has become, in recent years, an essential tool in the analysis, design, and evaluation of cyclones by providing a detailed picture and understanding of the flow field inside these devices.
The present study aimed to develop a well-validated and comprehensive CFD model for simulation of highly loaded large-scale cyclone preheaters. The model should be capable of predicting important performance parameters such as pressure drop, separation efficiency, heat transfer rate, as well as flow pattern and erosion pattern. For this purpose, an Eulerian–Lagrangian multiphase model, i.e., Dense Discrete Phase Model (DDPM) coupled with a robust turbulence model was developed.
The developed model was first validated against experimental measurements of pressure drop in an isothermal highly loaded lab-scale cyclone (diameter of 0.2 m) operating with coarse particles (diameter of 2 mm). The results indicated that the model is promising for the simulation of this complex case. The model was further implemented to simulate an iso-thermal, highly loaded pilot-scale (1.6 m in diameter) cement cyclone operating with relatively fine particles, i.e., Sauter mean diameter of 5 μm. However, preliminary simulations showed that the model highly underpredicted separation efficiency and overpredicted pressure drop. It turned out that this was because the agglomeration phenomenon, which plays a vital role in the case of fine particles, was missing. Therefore, an agglomeration model based on a stochastic Lagrangian inter-particle collision model was then implemented into the CFD model. The new model significantly improved predictions in terms of both pressure drop and separation efficiency.
The performance of the developed model may be influenced by sub-models (e.g., turbulence models, drag models, etc.), model parameters (e.g., particle-particle restitution coefficient and particle-wall rebound coefficient), and numerical parameters (e.g., parcel/grain size, time step size and order of discretization), to different extents. An extensive investigation was performed to suggest a set of sub-models, model parameters, and numerical parameters providing the best prediction of the hydrodynamics of large-scale highly loaded cyclones. The investigation showed that the turbulence model and particle-particle restitution coefficient have the strongest influence on the performance of the developed model.
In the next step, heat transfer modeling was added to the developed model, and an industrial-scale (1.6 m in diameter), non-isothermal, highly loaded cyclone preheater with relatively coarse particles, i.e., Sauter mean diameter of 375 μm, was studied. The studied cyclone operates at Integrate Melting Furnace (IMF) system of a stone wool production plant. Plant operating parameters, i.e., pressure drop and global heat transfer rate and the data from a measurement campaign, i.e., local measurements of gas velocity profiles, using LDA, and gas temperature profiles, using thermocouple and FTIR were used for further validation of the CFD model.
In all these cases, reasonable agreements between the measurements and simulation predictions were obtained. In addition, the developed model was accurate enough to capture the major observed trends caused by the changes in the operating parameters; for example, the improvement in separation efficiency of cyclones due to an increase in particle load is well captured by the developed model.
Lastly, to conduct erosion investigation in highly loaded cyclones, the well-known erosion model of Oka et al. was implemented into the developed CFD model. After successfully testing the model on a lab-scale cyclone, in which erosion was measured experimentally, the model was used to investigate erosion in the industrial cyclone preheater. The predicted erosion pattern and locations with severe erosion were qualitatively matched with observations.
Furthermore, throughout this study, parametric studies on operating conditions including gas and solid flow rates, particle size distribution, and gas inlet temperature were performed and the impacts on pressure drop, separation efficiency, heat transfer rate, and erosion, considered as key performance parameters of cyclone preheaters, were evaluated.
In essence, a CFD model has been developed that is computationally affordable and provides validated predictions of key performance parameters of cyclone preheaters, such as pressure drop, separation efficiency, and heat transfer rate. Additionally, by properly representing critical present physical phenomena, including turbulence, gas-particle interactions, particle-particle interactions, agglomeration, etc., the developed model provides valid predictions of flow patterns, temperature profiles and erosion patterns. Therefore, the developed model can probably also be used for other industrial applications to troubleshoot, redesign, and optimize industrial cyclone preheaters.