Publication
The organic Rankine cycle (ORC) system is considered a promising technology to exploit thermodynamic potential of waste heat. The weight and size of standard pumps can penalize the benefits of installing an ORC system on vehicles. Small-scale ORC applications would greatly benefit from more compact and efficient pumps. This paper presents the results of numerical and experimental analyses of small-scale turbopumps for ORC systems. A parameterized design tool is developed, allowing the rapid generation of numerous turbopump geometries and their fluid domains. The design tool creates a dataset of numerous turbopumps with different geometrical parameters. The turbopumps are investigated with CFD analysis, and the accomplished results are analyzed to characterize the influence of tip clearance and splitter blades on the performance (slip factor and head rise) of small-scale ORC turbopumps. The numerical results are employed to infer dimensionless maps (specific speed-specific diameter) and 1D models, which enable the capturing of the influence of down-scaling in the early design process of such machines. In the next step, two turbopumps are designed using the new design tool and then tested experimentally to validate the numerical procedure. The good agreement between experimental performance characteristics and developed models validates the computational results and reduced-order models. Following the experimental validation, the performance of an ORC system designed for the waste heat recovery of truck engines is estimated using the performance characteristics of the experimental turbopumps instead of a commercial multi-stage centrifugal pump. The developed turbopumps are one order of magnitude more compact compared to commercial systems. Further, the comparison suggests that the designed turbopumps improve the targeted ORC’s thermal efficiency by 0.51% and reduce its back-work ratio by 42% and 67%. Keywords: small-scale turbopump, organic Rankine cycle, computational fluid dynamics, experimental investigation, pre-design diagrams, 1D models