Non-Azeotropic Mixtures as Working Fluids in Heat Pumps: A theoretical and experimental study
Due to the fact that CFC (Chloro-Fluoro-Carbon) fluids will be forbidden as working fluids in heat pumps for environmental reasons, there is an urgent need for alternatives. As one alternative type of fluid, non-azeotropic mixtures have been suggested. This thesis deals with such mixtures. It includes mixtures with HCFC components, in spite of the recently imposed international restrictions for these fluids too, since this work started well before those restrictions and since good data exist for such mixtures.
The main aim of the thesis work has been to improve our knowledge about mixture behaviour in heat pump systems, with emphasis on thermodynamic differences between mixtures and pure fluids. A comprehensive computer program for simulation of heat pump systems with mixtures has been developed and utilized in several theoretical investigations. Furthermore, an experimental project in a full-scale plant with a non-azeotropic mixture has been carried out.
A study of different methods for comparison between mixtures and pure fluids in terms of calculated performance of a heat pump has been done. It showed that traditional methods for pure fluids can easily yield misleading results when used for mixtures. Reliable results are achieved if the comparisons are made rigorously, with the same heat exchanger area, and transport properties are included properly.
By means of the computer program, non-azeotropic mixture alternatives to CFC12 and CFC114 have been investigated. It was found that mixtures with moderate internal glides (less than approximately 10 K) in many cases can compete with pure fluids in terms of COP, even at small external temperature glides.
Mixture behaviour in two-stage cycles has been investigated in detail theoretically. Thermodynamic differences between a two-stage cycle with flash vessel and one with economizer heat exchanger, both charged with the same non-azeotropic mixture, are explained. It is quantified how the differences influence the heat pump performance. For the heat exchanger type, mixtures gain more in COP when going from a one-stage to a two-stage cycle than a pure fluid does while, for cycles with a flash vessel, this benefit does not exist.
By means of a sensitivity analysis the required accuracy levels of key factors for making performance predictions within a given uncertainty range have been identified. In order to enhance the prediction accuracy it is of greatest interest to increase our knowledge about the heat transfer coefficient at evaporation and the isentropic efficiency of the compressor. No major differences between mixtures and pure fluids were found regarding sensitivities to the investigated factors involved in estimations of heat pump performance.
An experimental comparison between two 1 MW heat pumps was carried out. One was charged with a non-azeotropic mixture of HCFC22-HCFC142b and the other with pure CFC12. Variations in capacity and COP were determined. In spite of unfavourable external conditions for the mixture, comparable COPs and capacities for the two heat pumps were found at equal external conditions. Comparisons were made between experimental and theoretically expected results for the three main parts, the evaporator, condenser and compressor.