Preliminary Design and Optimization of ORC Centripetal Turbines

During the last decade, Organic Rankine Cycle (ORC) turbogenerators have become very at- tractive for the conversion of low-temperature thermal energy sources in the small-to-medium power range. Complex gasdynamic phenomena and strong real-gas effects in the thermody- namic behavior of the working fluid usually characterize the thermo-fluid-dynamics of ORC turboexpanders. The use of Computational Fluid Dynamics (CFD) codes coupled with ac- curate thermo-physical property models is crucial to correctly model the flow expansion and therefore to achieve high-efficiency ORC turboexpanders. The design of ORC turbines is particularly challenging for small power output machines (up to a few hundreds of kWe); in these applications compactness is crucial and single-stage turbines represent a typical solution. As a result, fully supersonic flow conditions are typically adopted in these machines, and dedicated design techniques must be applied to avoid the onset of strong shocks in the stator-rotor gap [1]. In the present work an automated procedure to design single-stage centripetal turbines for ORC systems is presented. The design methodology involves a two-step procedure coupled to a global optimization strategy in order to define the optimal degree of reaction, size and the flow path of the machine. First, the mean-line code zTurbo [2] and a Genetic Algorithm (GA) are used to perform a preliminary design of the machine. The mean flow surfaces of the selected configurations are then optimized by means of the CFD-based throughflow solver TzFlow [3] and a metamodel-assisted GA. This latter design strategy was successfully applied to the design of an axial compressor in [4]. Both numerical codes are coupled with the most accurate equations of state for the thermophysical description of the fluid. The overall procedure is applied to design two turbines with a power output of about 50 and 250 kW. The results are extensively discusse