[home] [Personal Program] [Help]

---

Session: Session 6: Turbine design I
Session starts: Monday 12 October, 16:20
Presentation starts: 17:00
Room: 1A Europe

---

Choon Seng Wong (Department of Mechanical Engineering, University of Canterbury, New Zealand)
Susan Krumdieck (Department of Mechanical Engineering, University of Canterbury, Private Bag 4800, Christchurch, 8041, New Zealand)

Abstract:
Organic Rankine Cycle (ORC) could be used to generate power from low temperature heat sources or improve overall cycle efficiency in waste heat applications with minimal environmental pollution. The design and development of an ORC turbine, however, is a complex and costly engineering problem. The common refrigerants for an ORC application exhibit non-ideal gas behaviour and some unfavourable characteristics, such as flammability and toxicity. These characteristics further increase the complexity of the design and laboratory testing process of a turbine. Similitude, or similarity concept, is an essential concept in turbomachinery to allow the designer to scale a turbine design to different sizes or different working fluids without repeating the whole design and development process. Similarity concept allows the testing of a turbo-machine in a simple air test bench instead of a full scale ORC test bench. The concept can be further applied to adapt an existing gas turbine as an ORC turbine using different working fluids. This paper aims to scale an industrial gas turbine to different working fluids, other than the fluid the turbine was originally designed for. Three different approaches using the similarity concept were applied on the turbine performance data using compressed air to generate the performance curve for two refrigerants, namely R134a and R245fa. The scaled performance curves derived from the air performance data were compared to the performance map generated using 3D computational fluid dynamics (CFD) analysis tools for R134a and R245fa. The three approaches were compared in term of the accuracy of the performance estimation, and the most feasible approach was selected. The result shows that complete similarity cannot be achieved using two turbo-machines with different working fluids, even at the best efficiency point for particular expansion ratio. Constant Δh0s/a012 is imposed to achieve similarity, but the volumetric ratio is varying using different working fluids due to the variation of sound speed. The differences in the fluid properties and the expansion ratio lead to the deviation in turbine performance parameters, velocity diagram, turbine’s exit swirl angle, and entropy generation. The use of Δh0s/a012 further limits the application of the gas turbine for refrigerants with heavier molecular weight to a pressure ratio less than the designed pressure ratio using air. The specific speed at the best efficiency point with different expansion ratio was shifted to a higher value if higher expansion ratio was imposed. A correction chart for R245fa was attempted to estimate the turbine’s performance at higher expansion ratio as a function of volumetric ratio.