Batch melting model

Batch melting is an integral phenomenon to be considered in the simulation of a glass furnace and it is important to account for radiative and convective heating on top by the combustion gases and convective heat transfer from the glass melt below the batch. The batch model accounts for heat conduction within the batch and the batch velocity is not assumed constant within the tank, but rather it is explicitly calculated. A species transport equation is solved to account for the batch melting, the rate of which could be either thermally or kinetically controlled.

The energy equation includes an additional sink term resulting from the endothermic conversion of batch to glass melt. This is a very general 3D model which can be applied to a wide variety of furnaces (end-fired, side-fired etc.) without any ad-hoc modifications. The presence of cullet in the batch can also be taken into account through an additional heat sink term.

Coupled Simulation of Combustion Space and Refiner

The coupling between the combustion space and the refiner is successfully modeled using two separate submodels, one for the combustion space and the other for the refiner. The combustion space is first calculated with a convective boundary condition. The temperature profile on the bottom boundary is exported and applied to the finer glass surface. The resulting heat flux on the glass surface is then applied back to the bottom boundary of the combustion space, which completes an iteration of coupling. The temperature of selected locations in the refiner is monitored for each iteration. The calculation stops when the temperature difference between iterations is within 10K.

Temperature contours through the plane of the burners

Temperature contours on the wall and pathlines of particles injected from the inlet.

Modeling Front End Systems

It is critical for a glass delivery system (front end system) that bridges the melter and forming to reinsure that glass is conditioned to the requirements of the forming operations while maintaining the highest quality. An improperly designed front end system can cause a number of problems, including poor glass quality and inhomogeneity in the glass thermal profiles. CFD has become an important tool for glass manufacturers to guide and optimize such system designs.

Recently, Dr. Christopher Jian of Owens Corning utilized FLUENT to help in the decision making processes in engineering, operations and business. The temperatures of the front end calculated using FLUENT were in excellent agreement with the measured field data. This established confidence in FLUENT for predictive purposes. FLUENT was then used to optimize the design, and the glass manufacturer was able to achieve significant cost savings.

Vertical temperature profile in a glass delivery system. FLUENT simulation helped to optimize the design and save operating costs. Plot courtesy of Owens Corning.

Electric Boosting

Electrical boosting is used to improve the circulation characteristics of the glass melt and augment the heat transfer from the combustion space to the molten glass. The effect of electric boosting in a glass furnace is simulated with CFD by solving the real and imaginary parts of the electric potential and coupling the resulting Joulean heat dissipation as a source term to the energy equation of the glass melt. The user can choose the location and configuration of the electrodes and has the flexibility to match the total power input to the glass tank by either adjusting the applied voltage to the electrodes or updating the electric current field.

Contours of the real voltage on the walls of the glass tank and the electrodes and the pathlines of particles injected from the inlet.

Single Glass Fiber Drawing

In this process, molten glass is drawn through a nozzle by gravity and attenuated by tension from a winder to form fiberglass. Large free surface deformation and heat transfer are major factors in this process. The ratio of attenuated fiber cross-sectional area to initial cross-sectional area is defined as the draw ratio. The high draw-down ratio (1:10,000 and higher) makes it critical to understand the attenuation process as it influences fiber quality and productivity. Fluent software can be used to simulate advanced phenomena such as the wetting of the busher of fiber, the radius evolution as a function of external thermal effects, and the influence of the air surrounding the fibers.

A single glass fiber drawing was simulated by Dr. Bruno Purnode of Owens Corning using Fluent's polymer processing and glass forming CFD software POLYFLOW. POLYFLOW has unique free surface remeshing algorithms that allow it to handle large draw ratios and to calculate the fiber surface deformation while accounting for the highly non-linear viscosity relationship, surface tension, and special heat transfer laws. Dr. Purnode investigated several flow rates and process conditions of the drawing. The fiber diameter in the axial (draw) direction was compared with the experimental data in literature. The POLYFLOW results, at a draw ratio of 1:19,024, show excellent agreement with the experimental data.

POLYFLOW results, at a draw ratio of 1:19,024, show excellent agreement with the experimental data

Optical Glass Fiber

The quality of optical glass fibers is greatly influenced by heating stage of the fiber drawing process. In this stage, the glass preform is heated radiantly by the cylindrical muffle furnace. A neck-down region is formed that is characterized by the stretching of the glass. A numerical study of the transport phenomena in this drawing process is carried out using FIDAP flow modeling software. A conjugate problem involving the glass and the purge gas regions is solved. The neck-down shape is calculated using the free-surface modeling capability of FIDAP. The flow field, temperature field, and neck-down shape profile are generated based on continuity, momentum, energy and surface force balance. The heat exchange analysis includes coupled convection, conduction and radiation, taking into account the non-gray behavior of the glass surface and the semi transparent behavior of glass.

The neck-down profile obtained, velocity vectors in the purge gas region, speed contours in glass, temperature contours in the glass and the gas region.

Glass Fiber Spinning

The interaction between multiple fibers and air flows within the quench box of a glass fiber spinning process is a very difficult phenomenon to analyze from a theoretical perspective. In the process, continuous strands of fibers are drawn by extruding molten polymers through spinnerette nozzles. The fibers are formed by the balance of the tension from a winder and gravity. Fiber attenuation and interaction with quenching air flows characterizes this process. Quenching air flow pattern and fiber temperature are significantly altered by the entrainment of air due to the drag induced by fiber motion. To aid in the understanding of these complex interactions, Fluent Inc. has developed a fiber module add-on to its FLUENT software.

Steady state melt fiber spinning with a cross stream quenching air flow is successfully simulated using the fiber module. The diameter, temperature and velocity of fibers as functions of distance from the nozzles are calculated together with the flow field in the spinnerette chamber.

Fiber trajectories colored by temperature. The air flow vectors in the quench box.

Float Glass Forming

The process of float glass forming transforms a molten glass pool into a solid ribbon. The molten glass is conveyed on a molten tin bath and is stretched laterally by rolls to enhance ribbon width while it simultaneously cools and solidifies. The placement of overhead heating and cooling units is crucial to produce high optical quality (i.e., very uniform thickness).

Temperature distribution along the glass ribbon. At the beginning of the ribbon (left), the glass is exiting the furnace across a thin slit at a given temperature. At the end of the ribbon (right), after floating on the tin bath and being stretch both in the longitudinal and lateral directions by the numerous rolls located on the border of the ribbon, the thin, flat glass is cooled down and exits the tin bath at a much lower temperature with a small but uniform glass thickness.

The numerical simulation shows the evolution of the temperature, velocity and thickness of the glass ribbon while it is floating on the tin bath. Using this additional information, the glass designer can adjust the operating condition in order to further improve the production process, creating a glass sheet of higher quality.

Float glass forming was successfully modeled by Dr. Rajiv Tiwary and Dr. C.K. Edge of PPG Industries, Inc. using FIDAP. The simulation results provided insight into the float glass forming process, including width and thickness profiles, longitudinal and lateral stresses and residence times. Excellent agreement between the CFD results and plant measurements was obtained. The model was subsequently used for several parametric design studies.

Glass Blowing

The glass blowing process is commonly used each year to produce millions of bottles for water, milk, juice, soda, beer and wine, among others. This process is far more complex than the apparent commonality of the goods it produces. Thermal effects may induce some weaknesses in the bottles and possible risk of breakage. Proper operating conditions ensure a reduction in the weight of each container. This reduction in weight has dramatic economic and ecologic impacts, since reducing the quantity of glass used per bottle means less energy is needed to produce a single bottle and to transport it.

CFD offers a transient model of the different stage of the glass molding process. Each phenomenon can be analyzed in a very slow motion to better understand the process.

Fluent software has been used to successfully improve and optimize the glass blowing process. Using Fluent CFD software, you can better understand subtleties such as the combination of free surface deformation under the combined effect of gravity and the blowing air, the cooling process due to the action of the surrounding air and the contact with the mold, and the mechanical and thermal detection of the contact between the deforming glass and the mold. The better understanding gained from using Fluent software quickly leads to tremendous savings in manufacturing cost and time while ensuring a high quality final product.

Click image to view animation
Stretching and blowing of a glass sample to blow an oil bottle

Glass Pressing

The glass pressing process is extensively used to produce a large number of products. Among others are TV panels, tableware (dishes, asher, drinking glass, etc.) or preform for future blow molding application (press and blow process). Although most people may think that little technology is needed to produce something as common as a drinking glass or a dish, keeping the product quality high and the production costs down are challenges that require advanced technologies to accomplish. The market constantly demands thinner container walls, lighter objects and better textures making glass production far more complex than the apparent commonality of the goods it produces. Too fast of a plunger motion can lead to excessive residual stress or even product breakage during production, thermal effects may induce some weaknesses in the product and possible risk of later breakage. However, proper operating conditions may ensure an item's weight reduction. This competitive advantage can make a substantial difference between two glass companies.
Fluent software has been used successfully to improve the glass pressing process for many years. The combination of free surface deformation under the combined effect of the motion of different solid devices (molds, plunger(s), etc.), gravity, the cooling process (mainly due to the heat exchange with the mold and the plunger), the mechanical and thermal detection of the contact between the deforming glass and the moving and still molds despite the thin slit remaining between all this parts, create a better understanding of the subtleties of the glass pressing process. This better understanding quickly leads to tremendous savings in manufacturing cost and time while ensuring a high quality final product. Glass pressing is probably the application where the deformations are the largest. In Figure 1, the process begins with a spherical sample of glass that becomes a typical dish. Here, the initial grid can not support such an extreme deformation. Therefore, specific advanced techniques such as the adaptive meshing method have been implemented in Fluent's software to ensure both a high robustness of the simulation and an easy to learn problem set-up. During the deformation process, the solver automatically refines the mesh where this is necessary. Refinement may appear where very large deformations occur or where the geometrical details of the mold require them. Intelligent algorithms automatically refine the grid only where and when this is necessary during the run both in 2D and in 3D. These techniques are especially important when small geometrical details are involved in the final product. Fluent's software lets you refine the mesh to close these details while keeping larger elements elsewhere in order to reduce to global computational time.		Figure 1 View animation Glass pressing induces very large deformation and complex thermal contact that can be simulated with Fluent's CFD software Figure 2 View animation 2D simulation of a glass vessel pressing, small geometrical details are also considered during the pressing. Figure 3 View animation 3D glass pressing simulation are routinely run for relatively complex shapes

Gob Forming and Falling

Heating Automotive Windshields During Forming

In automotive windshield production, glass pieces held in position with a steel frame ride through the furnace on rollers while being heated by electric coils from the top and bottom. Spectral radiation through the semi-transparent glass medium, coupled with convection and conduction, heats the glass. Dr. Philip Burnside of PPG Industries, Inc. used the newly implemented Discrete Ordinates Model (DOM) in FLUENT 5 to simulate the heating of glass within a furnace. DOM includes the ability to handle semi-transparency and spectral radiation.

Steady and transient 2-D and 3-D calculations were performed to simulate a stationary piece of glass on a frame and a continuous sheet of glass moving through a furnace. The temperature calculations from FLUENT 5 were validated experimentally for different sets of process conditions and were within a few degrees of the measured values for both the stationary and continuous sheet.

Temperature contours of a continuous glass sheet in a furnace with two heating zones. FLUENT 5 predictions were within a few degrees of measured values.