Twin screw expanders are widely used for power generation in small scale Organic Rankine Cycles (ORC). In order to increase the efficiency and maximize the generated power, an optimized design and appropriate operating conditions should be used. The use of Computational Fluid Dynamics (CFD) allows the analysis of the flow inside these machines which is impossible to investigate experimentally but which is influencing the performance of the expander. Some of the challenges when performing CFD analysis in these machines are the complexity of the rotors' motion and the properties of the refrigerant. In this paper a 3D CFD analysis of a twin screw expander is presented. The 3D blockstructured grid for the twin screw expander is constructed from the solution of Laplace problems in two-dimensional sections on an unstructured grid of the same geometry. During the transient calculations, grid nodes are moved while keeping the mesh topology. The properties of the refrigerant R245fa have been evaluated using the ideal gas Equation of State (EoS), the Aungier Redlich-Kwong EoS and the CoolProp. 3D CFD analysis of the screw expander showed that the difference in power output between the ideal gas EoS and the Aungier Redlich-Kwong EoS is 8% and between Aungier Redlich-Kwong EoS and CoolProp is negligible for operating conditions of interest. To investigate the performance of the expander, different pressure ratios and rotational speeds were studied for two different designs of the twin screw expander. The flow analysis inside the clearances that are forming leakage paths gives more insight in the performance of the expander. It is concluded that the biggest pressure drop is caused by a throttling loss at the inlet port and therefore an optimized design of the inlet port is necessary.

A twin screw expander is a positive displacement machine used in various applications of waste heat recovery. The performance of this machine is influenced by internal leakages, gas pulsations formed in the inlet pipe and the properties of the refrigerant. In this paper a multi-chamber mathematical model of a twin screw expander is presented to predict its performance. From the mass and energy conservation laws, differential equations are derived which are then solved together with the appropriate Equation of State (EoS) in the instantaneous control volumes. In order to calculate the mass flow rates through leakage paths more accurately, flow coefficients used in the converging nozzle model were derived from 3D Computational Fluid Dynamic (CFD) calculation. Due to high gas pulsation levels at the inlet port, a coupling with a 3D CFD inlet pipe model is introduced in order to better predict throttling losses. The maximal deviation between predictions by the developed model and 3D CFD calculations of the complete machine is around 5% for the mass flow rate and the power output.

No abstract

Small scale Organic

With the increasing importance of minimizing primary energy usage and complying with emission restrictions, a significant interest has been developed towards waste heat recovery from industrial processes. A large portion of this energy is available at low temperatures (350K–400K) but it can be relatively efficiently converted into mechanical power using an Organic Rankine Cycle (ORC). Twin screw expanders can be used as an alternative to turbines with their cheap production costs and well proven efficiencies. In this paper, 3D CFD simulations of a twin screw expander using R245fa as the working fluid are performed. Since the fluid properties show big deviations when using the ideal gas equation of state (EoS), the flow problem has been evaluated using different real gas models. Thermodynamic parameters for the ideal gas EoS, the cubic Aungier Redlich-Kwong EoS and the CoolProp fluid database (open source) were compared in a preliminary study. After that, the models have been included through user-defined functions (UDFs) in ANSYS Fluent and were tested on 3D CFD calculations of a twin screw expander and a simplified expansion model. Several performance indicators such as mass flow rates, pressure-volume diagrams and power output are used to compare different fluid models for R245fa. From the results of this study, it can be concluded that the ideal gas EoS shows big deviations going closer to the saturation vapor line and the deviation in power comparing to the Aungier Redlich-Kwong EoS is around 8%. Conversely, the Aungier Redlich-Kwong EoS and the CoolProp database present very similar results for this case.