Optimisation-friendly modelling of thermodynamic properties of media

Abstract: The use of dynamic modelling and simulation of thermal power systems has grown recent years and as the performance of computers have increased optimisation of such systems has become possible. When performing such optimisations, there are high numerical requirements on the model used in the system. Due to the nonlinear nature of thermodynamic properties of working fluids, extra consideration has to be made when modelling these. In this thesis such a media model has been implemented through polynomial approximation. It has been shown that such an implementation can be used in optimisation algorithms yielding reliable results. A benchmarking optimisation test case has been constructed in order to test the implementation’s accuracy and performance. The physical models were implemented in Modelica used for dynamic optimisation using the JModelica.org platform. Furthermore it has also been benchmarked against existing media models designed for simulation in complex thermal power systems in order to verify the accuracy of the new model of thermodynamic properties of media.