Abstract-Improvement of process efficiency and product quality is available through implementation of more complex control algorithms and more accurate process models. It is especially critical for the glass industry production chain since glass production is a complex processes with high energy usage. The accuracy and robustness of advanced control algorithms is strictly dependent on the quality of the underlying mathematical model of the production process. This paper presents the formalisation and an empirical investigation of the hypothesis that a simplified, Finite Element Method (FEM) -based model can capture the closed-loop process dynamics over longer time scales and is suitable for real-time applications. The paper demonstrates how a tradeoff between model complexity and simulation time can be found.