TY - BOOK

T1 - Modelling and implicit solution of industrial furnaces using a multiple 1D finite volume approach

AU - Rath, Andreas

N1 - embargoed until 12-09-2027

PY - 2022

Y1 - 2022

N2 - The optimization of production processes is a key figure to stay competitive in today¿s manufacturing industry. Thus, the research on new possibilities to analyze existing and new manufacturing processes is ongoing a continuously yields new methods to analyze and improve production processes. A state-of-the-art method is to conduct experiments and measurements on representative pilot plants or the real-world production line. However, this approach tends to be expensive because of the construction of pilot plants and/ or the loss of production during the measurement operations. As a result of the high cost, and due the significant improvement of computational power, simulation approaches represent a good alternative apart from experimental setups. Simulations provide a vast selection between different mathematical models that can be adapted easily to describe complicated physical processes. Unfortunately, the more complex and larger a simulation setup gets the more the simulation time increases and require unjustifiable long computation times even on modern computers. Therefore, this work concentrates on this disadvantage and introduces a simulation method that allows to conduct real- time analysis of industrial scale furnaces. Compared to conventional simulation procedures, the proposed method discretizes the simulation geometry in different one- dimensional zones instead of a three- dimensional zone. Additionally, the flow of heat is limited to the respective one- dimensional zones and to the respective zone boundaries where heat can be exchanged with other cell zones. Thereby, the overall number of equations reduces significantly which reduces the required computational power. The temporal discretization is based on an implicit approach which enables an advantage in regard of the applicable time step. However, the disadvantage of the model is the coarse discretization of the simulation domain. Because of this, only averaged and no detailed thermo- technical parameters can be estimated. The aim of this thesis is to develop two fast computing simulation models for different types of furnaces, based on the mentioned theoretical concept. The selected furnace type of the furnaces to develop is on the one hand a tunnel furnace and on the other hand a roller hearth furnace. Both models are subsequently evaluated by provided data from industrial partners. Thereby the yielded results are discussed thoroughly and the respective model¿s potential for the future use as an additional tool for process parameter estimation is discussed.

AB - The optimization of production processes is a key figure to stay competitive in today¿s manufacturing industry. Thus, the research on new possibilities to analyze existing and new manufacturing processes is ongoing a continuously yields new methods to analyze and improve production processes. A state-of-the-art method is to conduct experiments and measurements on representative pilot plants or the real-world production line. However, this approach tends to be expensive because of the construction of pilot plants and/ or the loss of production during the measurement operations. As a result of the high cost, and due the significant improvement of computational power, simulation approaches represent a good alternative apart from experimental setups. Simulations provide a vast selection between different mathematical models that can be adapted easily to describe complicated physical processes. Unfortunately, the more complex and larger a simulation setup gets the more the simulation time increases and require unjustifiable long computation times even on modern computers. Therefore, this work concentrates on this disadvantage and introduces a simulation method that allows to conduct real- time analysis of industrial scale furnaces. Compared to conventional simulation procedures, the proposed method discretizes the simulation geometry in different one- dimensional zones instead of a three- dimensional zone. Additionally, the flow of heat is limited to the respective one- dimensional zones and to the respective zone boundaries where heat can be exchanged with other cell zones. Thereby, the overall number of equations reduces significantly which reduces the required computational power. The temporal discretization is based on an implicit approach which enables an advantage in regard of the applicable time step. However, the disadvantage of the model is the coarse discretization of the simulation domain. Because of this, only averaged and no detailed thermo- technical parameters can be estimated. The aim of this thesis is to develop two fast computing simulation models for different types of furnaces, based on the mentioned theoretical concept. The selected furnace type of the furnaces to develop is on the one hand a tunnel furnace and on the other hand a roller hearth furnace. Both models are subsequently evaluated by provided data from industrial partners. Thereby the yielded results are discussed thoroughly and the respective model¿s potential for the future use as an additional tool for process parameter estimation is discussed.

KW - Simulation

KW - Wärmeübertragung

KW - Prozessoptimierung

KW - Simulation

KW - Heat Transfer

KW - Process Optimization

M3 - Doctoral Thesis

ER -