Modeling and characterization of non-ideal compressible flows in unconventional turbines

Abstract

The vast majority of energy conversion systems currently makes use of fossil fuels, whose combustion generates harmful greenhouse gases. Transitioning to renewable energy sources is thus paramount to limiting the environmental impact of human activities on the climate. In this regard, the harvesting of wasted thermal energy constitutes a promising strategy to increase the efficiency of industrial processes and mobile engines. For instance, technologies such as organic Rankine cycle (ORC) systems enable the energy discarded during the conversion processes to the atmosphere to be repurposed and generate CO2-neutral electricity or additional mechanical work.

The efficiency of such systems is subordinate to that of each of the components, among which, is the turbine. Designing more efficient ORC turbines inherently leads to a higher thermodynamic cycle efficiency. However, these turbines operate with complex organic compounds, and part of the expansion process often occurs in the dense vapor state, where the thermodynamic properties exhibit significant deviations from the variations predicted by the ideal gas law. As a consequence, available guidelines for the design of turbomachinery operating with air or steam cannot be used, as they would lead to incorrect sizing and wrong performance estimations. The development of generalized guidelines for turbine design is possible only through a thorough investigation of the internal non-ideal compressible flow inside the vane passage, and by accurately discerning all the possible loss sources.

The research outlined in this thesis aims at characterizing non-ideal compressible internal flows of dense vapors and developing new guidelines for the design of unconventional turbines operating with organic fluids, such as those operating in organic Rankine cycle power systems.

The influence of both the complexity of the fluid molecules and the thermodynamic state on the flow field is evaluated for some paradigmatic one-dimensional flow configurations. For these processes, loss mechanisms and relevant trends in flow variables are both qualitatively and quantitatively estimated. Moreover, a detailed analysis of the viscous dissipation in turbulent wall-bounded flows of dense vapor is performed by resorting to direct numerical simulations (DNS). Results are compared against those from an in-house reduced-order model (ROM) code solving the two-dimensional boundary layer equations.

The combined effects of the working fluid, its thermodynamic state, and the flow compressibility on the flow deviation downstream of turbine cascades are then investigated by means of Reynolds Averaged Navier-Stokes (RANS) calculations on a representative geometry. The results obtained from the simulations are compared against those estimated with reduced-order physical models. Finally, an investigation of the influence of both compressibility and fluid molecular complexity on the optimal solidity of axial turbines is performed using RANS calculations. New design guidelines for the selection of optimal solidity in the preliminary design of non-conventional turbomachinery are proposed and discussed.

Results show that turbines operating with compounds characterized by a high complexity of the molecular structure are arguably subjected to higher losses in the mixing region, as well as exhibiting larger viscous dissipation at a given Reynolds number. Moreover, the fluid strongly affects the operational range of the turbine, as well as its design.