Effects of Parameter Scaling on Archimedes Screw Generator Performance

Abstract

Archimedes screws are an ancient pumping technology that has more recently found use as a technology for hydropower generation. Currently, the literature is lacking reliable data, performance predicting models, and design guidelines. Most performance models presented in the literature are theoretical or were developed and evaluated using laboratory-scale data. This paper presents novel experimental and numerical simulation data to the literature from screw generators with a wide range of sizes (laboratory to full-scale powerplant scale) and orientations. The data suggest that the components of power production (pressure-driven and viscous/friction) scale differently depending on system size, configuration, and operating conditions. So, for the robust validation of models and the development of reliable design guidance, data from a wide range of sizes and configurations are crucial. The paper presents data collected from laboratory experiments, field measurements from operating powerplants, and data from numerical simulations. The numerical simulations were evaluated for accuracy with experimental data, then used to collect performance data from a wide range of screw geometries and scales. The length-scale (diameter), number of blades, fill height of water, inclination angle, and surface roughness were all varied. The data gathered in these experiments were analyzed and used to develop back-of-the-envelope estimations for the effect of each parameter on overall system performance; the relationships are intended to serve as a useful reference for designers, though they should not be used in lieu of a design model. The length-scale and number of blades were related to power in a way that could be reasonably approximated with a constant value. The fill height, inclination angle, and surface roughness were related to power in a way that could be approximated reasonably with first-order polynomial fits. Altogether, this paper presents much-needed, novel data to the literature; the data are integral for future model development and evaluation.