Login / Signup

Wake characteristics of a freely rotating bioinspired swept rotor blade.

Asif Shahriar NafiKrishnamoorthy KrishnanAnup K DebnathErin E HackettRoi Gurka
Published in: Royal Society open science (2021)
Rotor blades can be found in many engineering applications, mainly associated with converting energy from fluids to work (or electricity). Rotor blade geometry is a key factor in the mechanical efficiency of the energy conversion process. For example, wind turbines' performance directly depends on the blade geometry and the wake flow formed behind them. We suggest to use a bioinspired blade based on the common swift wing. Common swift (Apus apus) is known to be a long-distance flyer, able to stay aloft for long periods of time by maintaining high lift and low drag. We study the near-wake flow characteristics of a freely rotating rotor with swept blades and its aerodynamic loads. These are compared with a straight-bladed rotor. The experiments were conducted in a water flume using particle image velocimetry (PIV) technique. Both blades were studied for four different flow speeds with freestream Reynolds numbers ranging from 23 000 to 41 000. Our results show that the near wake developed behind the swept-back blade was significantly different from the straight blade configuration. The near wake developed behind the swept-back blade exhibited relatively lower momentum loss and suppressed turbulent activity (mixing and production) compared with the straight blade. Comparing the aerodynamic characteristics, though the swept-back blade generated relatively less lift than the straight blade, the drag was relatively low. Thus, the swept-back blade produced two to three times higher lift-to-drag ratio than the straight blade. Based on these observations, we suggest that, with improved design optimizations, using the swept-back configuration in rotor blades (specifically used in wind turbines) can improve mechanical efficiency and reduce the energy loss during the conversion process.
Keyphrases
  • optical coherence tomography
  • machine learning
  • deep learning