ANN based grey box modeling of thermodynamics and its dynamic optimization
- Process Systems Engineering
- Masterthesis / Bachelorthesis
- Focus/Key Topic:
- starting immediately
Besides general mass and energy balances dynamic simulation and optimization of chemical processes often includes complex depiction of thermodynamic properties of occurring components. However, the use of thermodynamic models in equation-based solvers for dynamic simulation and optimization can lead to various challenges. Dependencies of thermodynamic state variables for fluid mixtures are often described by cubic equations of state (EOS). The cubic equation has up to three roots for volume V depending on states (temperature T and pressure p). Consequently, the intended root has to be selected from those, which is only partly possible in equation-based development environments for simulation and optimization. In order to still incorporate advantages of EOS, artificial neural networks (ANN) can be applied.
In this work, the use of ANN for grey box modeling of thermodynamics in context with dynamic optimization is to be investigated.
What we expect from you:
- Motivation to work on thermodynamic problems and its implementation in simulation software
- General interest in simulative work
- Basic knowledge of programming are beneficial: Matlab, C++, Modelica
What we can offer you:
- A motivated team in which you can always find a contact person
- An exciting and current research topic
- Extension of your programming skills
If we were able to convince you with this topic, please send us your CV and grade report via email. In case of questions, do not hesitate to call us.