Simulation model structure numerically robust to changes in magnitude and combination of input and output variables

Mathematical models of refrigeration systems are often based on a coupling of component models forming a “closed loop” type of system model. In these models the coupling structure of the component models represents the actual flow path of refrigerant in the system. Very often numerical instabilities prevent the practical use of such a system model for more than one input/output combination and for other magnitudes of refrigerating capacities.A higher numerical robustness of system models can be achieved by making a model for the refrigeration cycle the core of the system model and by using variables with narrow definition intervals for the exchange of information between the cycle model and the component models.The advantages of the cycle-oriented method are illustrated by an example showing the refrigeration cycle similarities between two very different refrigeration systems.