Experimental and numerical analyses of a 5 kWe oil-free open-drive scroll expander for small-scale organic Rankine cycle (ORC) applications.

Highlights • The performance of an oil-free scroll expander is presented. • A total of 75 steady-state points with R245fa have been collected. • The friction losses were quantified to be up to 20% of total power output. • Semi-empirical and Artificial Neural Network models have been compared in a systematic way. • Experimental data and simulation model source-codes have been provided. Abstract Organic Rankine cycles (ORCs) are thermodynamic power cycles designed to generate work from a wide range of heat source conditions. In particular, low-grade waste heat recovery (WHR) (<150 ° C) can be effectively exploited with such systems. The efficiency of an ORC is highly dependent on its expander performance. In the low power output range (<10 kW), scroll expanders are cost-effective. Although a number of researchers have investigated the use of scroll compressors as expanders, very little work has been carried out in modeling and investigating the performance of oil-free expanders. In this work, an experimental evaluation of a newly designed open-drive oil-free scroll expander was performed. The expander had a nominal capacity of 5 kW, built-in volume ratio of 3.5, and was integrated into an ORC test-rig with R245fa as working fluid. The experimental data consisted of 75 data points that were used to map the oil-free operation of the scroll expander over five expander rotational speeds (from 800 rpm to 3000 rpm). Two heat sources inlet temperature, i.e. 85 ° C and 110 ° C, were investigated. The scroll expander achieved a maximum overall isentropic efficiency of 0.58 for the temperature source of 110 ° C, for the imposed specific volume ratio of 6.12 at rotational speed of 1600 rpm. For the same heat source, the maximum expander power output was 3.75 kW for an imposed specific volume ratio of 6.55 and rotational speed of 2500 rpm. Besides the experimental work, the performance of the expander was characterized by means of a semi-empirical model to break-down the different loss terms. A well known model available in the literature was extended to account for the major frictional losses in a scroll machine, i.e. bearings, tip-seals and other sources of friction. Additionally, an Artificial Neural Network (ANN) modeling approach was also proposed to achieve higher accuracy in mapping expander performance for use in system simulation. The experimental data and model source codes are provided as supplementary materials. [ABSTRACT FROM AUTHOR]