Towards viable flow simulations of small-scale rotors and blade segments

Abstract

The paper focuses on the possibilities of adequately simulating complex flow fields that appear around small-scale propellers of multicopter aircraft. Such unmanned air vehicles (UAVs) are steadily gaining popularity for their diverse applications (surveillance, communication, deliveries, etc.) and the need for a viable (i.e. usable, satisfactory, practical) computational tool is also surging. From an engineering standpoint, it is important to obtain sufficiently accurate predictions of flow field variables in a reasonable amount of time so that the design process can be fast and efficient, in particular the subsequent structural and flight mechanics analyses. That is why more or less standard fluid flow models, e.g. Reynolds-averaged Navier-Stokes (RANS) equations solved by the finite volume method (FVM), are constantly being employed and validated. On the other hand, special attention must be given to various flow peculiarities occurring around the blade segments shaped like airfoils since these flows are characterized by small chords (length-scales), low speeds and, therefore, low Reynolds numbers (Re) and pronounced viscous effects. The investigated low-Re flows include both transitional and turbulent zones, laminar separation bubbles (LSBs), flow separation, as well as rotating wakes, which require somewhat specific approaches to flow modeling (advanced turbulence models, fine spatial and temporal scales, etc). Here, the conducted computations (around stationary blade segments as well as rotating rotors), closed by different turbulence models, are presented and explained. Various qualitative and quantitative results are provided, compared and discussed. The main possibilities and obstacles of each computational approach are mentioned. Where possible, numerical results are validated against experimental data. The correspondence between the two sets of results can be considered satisfactory (relative differences for the thrust coefficient amount to 15%, while they are even lower for the torque coefficient). It can be concluded that the choice of turbulence modeling (and/or resolving) greatly affects the final output, even in design operating conditions (at medium angles-of-attack where laminar, attached flow dominates). Distinctive flow phenomena still exist, and in order to be adequately simulated, a comprehensive modeling approach should be adopted.