Testing and Modeling Scroll Compressors with a view to Integrating them as Expanders into a Rankine Cycle

The aim of this work is to study the integration of a scroll expander into a heat recovery Rankine cycle, whose heat source consists in exhaust gas from an internal combustion engine. An experimental study is carried out on three different scroll compressors adapted to be run in expander mode. The purpose of this experimental study is to measure the performances of these expanders, to judge for the feasibility of a scroll expander on the basis of an existing compressor, and to identify the parameters of a model of expander for the three machines tested. These three machines are characterized by different internal built-in volume ratios and displacements. One of the expanders is oil-free, whereas the other two require lubrication. In terms of temperatures and pressures, the expanders are tested in conditions close to those of the Rankine cycle under consideration. The working fluid is water. One of the expanders is also fed with a binary mixture of water and propylene glycol, since this mixture could combine lubricating and anti-freezing properties. The latter property is necessary for an automotive application. Performances of the expanders are evaluated in terms of measured isentropic effectiveness and delivered mechanical power. In some conditions, isentropic effectiveness reaches values higher than 50%. For two of the expanders, this effectiveness is limited by an important internal leakage due to problems encountered with tip seals. In the range of achieved supply and exhaust pressures, it is shown that the internal built-in volume ratio of the expander is often too small. The under-expansion of steam also limits the isentropic effectiveness. Following this experimental study, the integration of the expander into the heat recovery Rankine cycle is also studied with the help of a global model of the cycle. Performances of the Rankine cycle are evaluated and expressed as a function of the expander characteristics. Finally, an optimum geometry of the expander is determined. Two other working fluids are also considered: a binary mixture of propylene glycol and water and n-pentane. Performances of the cycles working with these two fluids are evaluated and compared to those obtained with water. N-pentane presents the best performances in terms of internal power developed by the expander and power recovered from exhaust gas. The aqueous solution of propylene glycol enters the expander under a two-phase binary mixture. The liquid phase could ensure partly the lubrication, but previous tribology tests have shown that lubricating oil is still necessary