Ansys Polyflow provides several capabilities that are useful for glass furnace simulations:
a library of finite elements
specific interpolation for the temperature variable
Molten glass has a low conductivity and rather high viscosity. This means that the energy equation requires a finer representation than the momentum equation, since advection-dominated problems are difficult to solve. Polyflow allows, for example, to combine linear velocity and 2×2×2 temperature in 3D.
automatic evolution schemes on all parameters in order to find the solution of highly nonlinear problems
For natural convection simulations of low-conductivity materials such as glass furnaces, a common practice is to start the simulation at a high value of the conductivity and to decrease it toward the solution of the problem. Typically, it is necessary to perform 10 evolution steps to lower the conductivity to its physical value.
Specific models have also been implemented in Ansys Polyflow to improve glass furnace simulations. These are an electrical heating model, a Rosseland correction, and a macroscopic model to account for the influence of air bubblers on the flow field. They are described in the following sections.