Robust and Fast Meanline Methodology for Radial Inflow Turbines with Non-ideal Gases for System Level Optimisation of Organic Rankine Cycl

Abstract

Waste heat recovery aims to increase the overall system efficiency and thus, reduce emissions. One technology commonly considered is the Organic Rankine Cycle (ORC) which uses a refrigerant as the working fluid. For an efficient operation of the ORC, the different components, e.g., the turbine, need to be optimised considering the entire system. This paper investigates the capability of meanline modelling for nozzled turbines to model the non-ideal gas effects of different refrigerants. For optimal implementation at a system level, it is important to have sufficient accuracy and low computational cost. The meanline model was compared with CFD for three different working fluids. A 3D parametric model was generated, and a numerical experiment was performed using Latin Hypercube Sampling. Seven meanline design parameters were investigated over a broad design space to show the influence of each respective one on turbine performance. Subsequently, the meanline model was calibrated against a subset of geometries and compared to the entire data sample. The average error stayed below 5%, which is sufficiently accurate for preliminary-stage calculations.