SCHOTT solutions no. 2/2014 > Research & Development
Real laboratory experiments using the melting crucible make an important contribution to developing simulation models. The parameters and values that are obtained are then transferred over to the actual dimensions of the melting tank. Photo: SCHOTT/A. Sell
Images of reality
SCHOTT uses modeling and simulation to further develop its processes and products in a strategic manner – by experimenting on the computer.
”Make everything as simple as possible, but not simpler!” The model of a glass melting tank for calculating the density of bubbles could also be based on this maxim from Albert Einstein. The 3-D image on a screen shows a bright spectrum of colors inside the tank. Reddish colors mean there are still a lot of bubbles in the glass melt. The bluer the areas, the fewer bubbles it contains. Pretty simple, right? Dr. Christoph Berndhäuser, SCHOTT developer and expert on simulation, grins for a moment before noting: ”This model and its capabilities that we created in-house for evaluating glass quality is really the product of expertise developed over 20 years.”
Today SCHOTT performs mathematical simulations and builds models to optimize and develop nearly all of its technological processes and products – in both its Business Units and in research. And for good reason; simulation often saves the company from having to perform costly trials and allows for tests to be performed on a model to gain insights into a specific system or make forecasts. But that’s not all; the latest computing procedures in combination with increased computer capabilities enable extensive parameter variations that could hardly have been efficiently accomplished using conventional testing. This provides technology companies with a decisive lead in the race to meet higher demands for quality and keep up with ever shorter innovation cycles. This is why SCHOTT relies on simulation and modeling along the entire process chain for manufacturing glass; from melting, fining and hot forming to annealing, ceramization, 3-D shaping and even product certification. Important interrelationships are also identified using actual production data, thanks to the further development and automation of data analysis instruments. And the company already has its sights set on exciting topics for future simulations, including material modeling of glasses, glass-ceramics and plastics (see below ”Multiscale modeling”).
Crystal clear: the glass sample to the right contains considerably fewer bubbles than the one on the left. These types of laboratory results ultimately form the basis for simulation programs and quality models. Photo: SCHOTT/A. Sell
A key topic in the coming years will be to continuously improve advanced models for evaluating the quality of glass by collecting even more know-how in the areas of fining chemistry for melting tanks and the thermodynamics of glass. ”Our customers expect us to be able to accurately assess the quality for our production processes and products. For materials such as our glasses and glassceramics, this means determining the number, size, and composition of the bubbles per kilogram that they still contain,” Dr. Berndhäuser explains and points to his colorful quality model on the screen. To forecast the quality in real tanks, the temperature distributions and flow pattern inside the tank are calculated using special CFD (Computational Fluid Dynamics) programs and are then assessed for the number of bubbles based on quality models developed in-house by the company. Laboratory trials inside a melting crucible make an important contribution to developing these models. They show the bubble size distribution depending on the refining agent used, its concentration, temperature and hold time for every type of glass analyzed. These parameters and values flow into the quality model and are transferred over to the actual size of the tank in the simulation.
Multiscale modeling: looking into the heart of glass
How can certain mechanical or chemical properties of glass be selectively improved? The goal here is to look even more deeply ”into the heart of glass,” just as our company founder Otto Schott once did. Multiscale modeling offers promising approaches for accomplishing this. The starting point is to realize that the many different processes during which material properties form take place on broad spatial and temporal scales. Values such as temperature or stresses can be modeled in the meter to millimeter region on the macroscale. Microstructural, atomistic or even electron-based modeling extends all the way to below the nanometer level, however. These calculations still reach their limits. It is research’s goal to gradually step down to the various levels in order to ultimately tie them into a consistent multiscale model and thus be able to describe and predict material characteristics in a more comprehensive manner. SCHOTT is already on the way to achieving this goal, working on topological models that link together the different glass structures, such as chain, plane and scaffold-shaped structures with specific glass properties. <
So-called tracers are now being used to describe how bubbles behave inside a tank. 100,000 of these simulated markers pass through the various modeled flow pattern and temperature zones on different paths, depending on the furnace design and setup. Statistics then provide information on the size and distribution of the bubbles during and at the end of the melting and refining process. ”The next level of simulation will also include the bubbles’ reactions to the refining process because the physical and chemical conditions change during the process, depending on the bubble load,” Dr. Berndhäuser explains.
These types of questions and insights also mark the starting point for optimized approaches to constructing and setting melting tanks, especially switching over to more environmentally-friendly refining chemicals that don’t contain heavy metals. Research at SCHOTT is currently conducting an in-depth project on this topic. One thing is already clear: advances in the area of simulations definitely will not make it any easier to realize Einstein’s maxim on simplicity.
Process Simulation and Optimization
Process Simulation and Optimization