Contribution to the Characterization of Piston Expanders for Their Use in Small-scale Power Production Systems

Piston expanders are well suited for small scale power generation (<50kW) using heat engines such as Ericsson engine or Organic Rankine Cycle (ORC) system. Both systems offer the possibility to use renewable energy sources and are increasingly considered in light of the economic and environmental context. This thesis aims at contributing to the knowledge and the development of piston expanders by presenting modeling, simulation and experimental investigations.
Two types of piston engine models are proposed. The first type (denoted as “detailed”) describes the shaft angle (or time) evolution of the state of the fluid in the control volume limited by the cylinder wall and the piston. The second type of model (denoted as “semi-empirical”) is based on theoretical indicator diagram. A calibration process is proposed and applied with the measurement collected on two different expanders.
The detailed model is first used to simulate an Ericsson engine. The simulations allow for the optimization of the pressure ratio, the determination of the ratio between the swept volumes of the expander and the compressor, the computation of the valve timing and the evaluation of the potential performances.
The main modifications made to turn an available Internal Combustion Engine (ICE) into an Ericsson engine are described. This prototype is then tested but did not run as expected. The identified reasons are a too high leakage rate, a too large cut-off ratio and a too low expander supply temperature. These three features led to insufficient pressure. Nevertheless, a compressor is used to rise the pressure up to 7 bar in order to test the expander part. This expander shows a maximal isentropic efficiency of 59% at 600 RPM with a pressure ratio of 5.
A swash-plate expander integrated into an ORC using R245fa is also investigated in a wide range of operating conditions. Performance maps in terms of pressure ratio and rotational speed are given. Maximal reached isentropic efficiency is 54% at 2500 RPM for a pressure ratio of 10.5.
Experimental results obtained on both expanders are analyzed with the same methodology to assess the importance of each source of losses. This methodology is based on the disaggregation of the isentropic efficiency into different factors. Each of these factors represents one source of losses. In-cylinder pressure measurements are performed in order to measure mechanical losses. Leakage effects are simulated with the detailed model.
Finally, an Ericsson engine-based m-CHP unit and an ORC-based m-CHP unit using the swash-plate expander are simulated with the validated semi-empirical model. Ericsson engine system shows a 6.5% electrical efficiency while the ORC system shows en efficiency of 8.2%. For both systems, possible improvements are simulated. It is shown, among other things, that the benefit of an increase of the built-in volume is limited by the decrease of the mechanical efficiency.