Modeling of evaporation processes in glass melting furnaces

The majority of glass furnaces worldwide, apply fossil fuel combustion to transfer heat directly by radiation from the combustion processes to the melting batch and glass melt. During these high temperature melting processes, some glass components, such as: sodium, potassium, boron and lead species will evaporate from the glass melt. There are three main motives to study evaporation processes in glass furnaces. In the first place evaporation of volatile components from the glass melt surface is one of the main causes of particulate and heavy metal emissions of industrial glass furnaces. Secondly, incongruent evaporation of glass melt components might cause depletion of volatile glass compounds at the surface layer of the melt. This process may be a source for glass failures and inhomogeneities in the glass product. Thirdly, volatilization of glass melt components may lead to the formation of aggressive vapors, such as alkali or lead vapors, reacting with the superstructure silica based refractory materials or refractories in the flue gas system. The evaporation rates depend on the process conditions in the glass furnace, like: • Glass melt composition (at the surface); • Temperature of the glass melt surface; • Composition of the atmosphere. Especially the water vapor (air-fuel versus oxy-fuel) and the carbon monoxide contents just above the melt are important; • Exposure time of a glass melt volume to the combustion atmosphere and • Local gas velocities and turbulence intensity just above the glass melt surface. For different glass types, the interaction between furnace atmosphere and the glass melt surface as well as the different evaporation reactions were summarized and discussed qualitatively, by many investigators. However, a universal and accurate mathematical evaporation model to predict evaporation rates of volatile glass components in industrial glass melt furnaces as well as laboratory glass melt furnaces quantitatively, was hardly available so far. Objective The main goal of this study is the development of a general applicable evaporation model, for different types of glass furnaces and different glass compositions. This model should be based on validated mass transfer relations and simulation of the thermodynamic properties of multi-component silicate melts. With this evaporation model the evaporation rates of volatile species from different types of glass melts can be predicted. The model is meant to investigate quantitatively the impact of different process parameters on the evaporation rates of different volatile glass species and depletion of these glass components in the surface layer of the melt. The investigations in this study primarily focus on evaporation from multicomponent silicate glass melts used for the industrial production of glass articles. Approach In this study an evaporation model has been developed and validated by laboratory evaporation tests. For the evaporation model developed here, 5 different steps were required: I. Identification of the main volatile species and dominant evaporation reactions at prevalent conditions for the investigated glass melts. II. Modeling of the mass transport of volatile glass components in the glass melt. The second diffusion law of Fick in combination with known interdiffusion coefficients (from experiments, models or literature) of volatile glass compounds, to describe the diffusion mass transport in a static melt. Additional to this, Computational Fluid Dynamics (CFD) are applied to describe both, the convective and diffusive transport of volatile species from the bulk to the surface of the melt. III. Modeling of mass transport in the gas phase, describing the transport of volatile species from the glass melt surface into the gaseous atmosphere for laminar and turbulent gas flows. IV. Thermodynamic Associated Species Model (ASM) to calculate the chemical activities of volatile glass components in the melt and at the surface of the melt. V. Calculation of the evaporation rates. From the evaporation reaction, its chemical equilibrium constant and the chemical activities of the volatile glass components at the surface of the melt, the saturation pressure of the gaseous reaction products can be determined. The local evaporation rates of individual species, are calculated from the local mass transfer relations or local Nernst boundary layer thickness in the gas phase, the local saturation vapor pressures of the volatile species and the vapor pressures of these species in the bulk gas flow. The local Nernst boundary layer thickness of the Modeling of evaporation processes in glass melting furnaces gas phase above the melt can be determined from the velocity profiles in the gas phase derived from CFD modeling, even for turbulent flows in combustion chambers. Laboratory-scale transpiration evaporation experiments have been used to: A) validate the evaporation modeling results and B) to study evaporation kinetics for sodium-silicate melts, multi-component alkali-lime-silicate melts and alkali-free borosilicate melts. In the transpiration set-up the furnace atmosphere composition, temperature level and gas velocity are controlled. Evaporation rates were measured for sodium, potassium, boron, chloride and sulfur species released from different well defined glass types, with known compositions. As will be shown later on in this summary, a procedure has been developed to derive chemical activities from the evaporation rates, measured during transpiration experiments. The validation of the evaporation model exists of a comparison between the experimentally and from thermodynamic modeling (ASM) derived chemical activities of volatile glass components. Mass transport relations for the gas phase Mass transfer relations and procedures were derived to describe the mass transport of volatile glass components or their volatile reaction products from liquids or melts into the gaseous atmosphere. These mass transfer relations and procedures are developed for the applied laboratory transpiration experiments as well as for industrial glass furnaces. CFD modeling appears to be a useful tool to predict the mass transport of volatile species into a carrier gas for a complex geometrical configuration of a transpiration test set-up. Such CFD models are applied to describe and to understand the fluid dynamics in the gas phase and distribution of volatile species in this phase. Water is used as model liquid to investigate these mass transfer processes in transpiration evaporation tests. Results of the CFD-modeling for water evaporation tests at room temperature have been validated by transpiration experiments. Excellent agreement was found between model results and experiments, as the differences between the experimental measured evaporation rates and the CFD modeling results are less than 2 %. The results of CFD modeling and the results of simple water transpiration evaporation experiments at room temperature are used to obtain relatively simple mass transport relations for a fixed geometry of the transpiration test set-up. It has been shown in this study that these Sherwood relation are applicable for other evaporating liquids and temperatures as well when using the same equipment.