Deep Learning Based Prediction of axial turbine performance for Clean Organic Rankine Cycle system driven by low-temperature heat source

Abstract

The big data performances of Organic Rankine Cycle (ORC) system is widely used in avariety of industrial applications. In some processes, the axial flow turbine at the low temperature heatsource is crucial. The goal of this work is to increase the high efficiency of a tiny axial turbine poweredby an Organic Rankine Cycle (ORC) utilizing numerical studies and deep learning method. To providethe best performance for the Organic Rankine Cycle (ORC) application, various turbine stages aresuggested. Three-dimensional RANS computations are recommended for five different rotational speedsand four mass flow rates ranging from 0.2 to 0.5 kg/s with an inlet temperature of 365 K in order todescribe the hydrodynamic and thermodynamic capabilities and get a big data. The analysis show that thetwo-stage turbine with a fixed rotational speed delivers the highest power output value of 10000W, hasthe highest turbine efficiency of 86%.In a three-stage turbine arrangement operating at steady state, theefficiency of the turbine and its output are 89% and 12000 W, respectively. The maximum values forturbine efficiency and power production are 88% and 12000W, respectively, as a result of the transientcomputational procedure. These results demonstrate the potential for exchanging low temperature heatsources using micro-three-stage axial turbines in Organic Rankine Cycle (ORC) systems .