Fluid Flow Simulation

The objective of this fluid flow simulation project was to show the influence of correctly modeling the fluid-induced motion of a generally rigid body, as well as creating an efficient design for a cheap, mass-production feasible, wind turbine model to use in urban areas.

For this CFD consultancy project, after a simple 3-bladed design was created, 2D CFD simulations were performed for both steady-state and transient models with fixed angular speed (enough imposed to generate zero torque), as well as the realistic model itself, also 2D. Initially, the steady case was used by our CFD consultants to predict a good initial mesh (particularly the inflation) and turbulence models, resulting in the selection of the Spalart-Allmaras model, a compromise between accuracy, reliability and calculation speed. A non-motion transient CFD modeling process was then carried to determine a feasible time step (even though an implicit formulation was selected, the time advancement cannot be so great as to physically distort solutions, or introduce an excessive amount of numerical dissipation).

Next, the fixed rotation motion was introduced into the computational fluid dynamics simulation, with some adjustments made to time-step in order to accommodate a dynamic remeshing process, necessary to avoid excessive mesh distortion and zero-volume elements as the body rotates.

Finally, taking on all previous results, the realistic model was carried out using the CFD simulation software, for as long as 10s (DT = 1e-4 for a small, slow turbine), with the rigid body motion driven by airflow. In all these models, a coupled pressure-velocity scheme was rather a necessity, as segregated models consistently failed to achieve a reasonable convergence or convergence rate.

Although the 3D model is still being carried out, the 2D modeling provided enough data for the computational fluid Dynamics analysis to understand the impact of correct motion prediction, particularly for lighter and faster turbines, where motion is by no means constant. On the practical side, it was possible to give an initial prediction on the power generated by a single turbine and to infer that, in an average sense, the steady-state approach is rather suitable initially, but must later be improved by either an experiment of a realistic model. This opened the question for future work and research on modeling procedures for wind turbines.

It was requested that a high lift low drag hydrofoil is designed to provide extra upwards force for a vessel to be utilized in shallow-to-medium waters, at rather slow speeds.

Based on the mathematical description of Zhukovsky for wing profiles, and the general shape of a manta ray, our team of CFD design engineers in our CFD company decided on combining those two features to solve the question, betting on 3D printer production of the difficult geometry traits associated.

To analyze the resulting body in the computational fluid analysis, initially a 2D, steady state fluid flow simulation was carried to decide on wing chord length, total extension and camber, which gave preliminary results, later compared and adjusted by a 3D, fully immersed fluid dynamics simulation.

In both cases, the trial-and-tested Spallart-Almaras turbulence model was chosen, as it is preferred for low stall wing models.

Finally, to better assess the model’s suitability to the proposal, a shallow water model (two-phase air-water model) was carried, with an introduction of numerically generated waves, as a rather qualitative tool to give proof of capability.