CFD Modelling of Melting Cyclones for Stone Wool Production

Stone wool is a fibrous insulation material providing thermal and acoustic insulation as well as fire-resistant properties. Thermal insulation is a key component in reducing the energy consumption, and the related CO₂ emissions, due to the heating of buildings and the demand for insulation materials is expected to continue to rise.

Stone wool is produced by melting rocks and spinning the melt into fibers. Melting the rocks is a very energy-intensive process that has traditionally been done in cupola furnaces. A cupola is a type of shaft furnace related to the blast furnaces used in iron smelting. Similar to the blast furnaces, the cupolas are fired using coke.

The ROCKWOOL Group has developed a new cyclone-based melting process known as the Integrated Melting Furnace (IMF). The IMF system integrates a set of preheater cyclones with a melting cyclone.

In the melting cyclone, the stone materials is blown into the chamber through tangential burner ports along with pulverized coal, preheated combustion air and oxygen. In the strongly swirling flow in the cyclone, molten and partially molten stone are slung to the walls where a melt layer is formed. Due to gravity, the melt flows down the walls and accumulates in the bottom. The use of the melting cyclone gives a more flexible and energy-efficient process compared to the cupola furnaces and the change of fuel from coke to coal and gas further reduces the CO₂ emissions.

The purpose of this work is to develop a modelling approach for the melting cyclone using Computational Fluid Dynamics (CFD). The goal of the final CFD model is to improve the understanding of the process and to help develop the melting cyclone further.

Many different physical phenomena take place within the melting cyclone such as multiphase flow, combustion, phase change and radiation. The developed modelling approach should be able to capture all the dominant physics in the melting cyclone.

For each physical phenomena, the expected behavior in the melting cyclone was compared to previously published models. A set of models was chosen based on their compatibility, their applicability and their computational expense.

The selected models were integrated and modified as needed. The developed models included a modified Eddy Dissipation-model for use in oxygen-enriched combustion, a model for tracking the melting of the raw materials, and coupling of DPM and VOF multi-phase models.

The models were applied to a generic melting cyclone model. The results showed that the models work according to the concept but they are limited in their application due to excessive computational time needed to reach steady state conditions.