# CFD Simulation of complex rotor equipment using FloEFD – Part II: Apache Helicopter

2016-08-31 Karl Du PlessisThe LEGO Aero Hawk model was not a true representation of real-world helicopter rotor blades. It did however illustrate how easily FloEFD can mesh complex geometry and also demonstrated the stability of the solver for such complex models. Moving on to an application that is a little closer to reality this time – Let us perform a simulation of an Apache helicopter as in the image above. The Apache-like model was obtained from *http://www.GrabCAD.com*. This model is by no means an exact replica, especially the rotor blade geometry, i.e. blade airfoil shape, blade pitch and twist etc., which are merely a figment of the author’s own imagination. Nonetheless, this blog article aims to illustrate the ease of use of the sliding mesh rotation functionality for simulating complex rotating bodies. A validation case of the sliding mesh capability, is subsequently shown for a single rotor blade of a light helicopter model. This case study can be found in the Engineering Edge Magazine Vol. 4 Iss. 1 from Mentor Graphics Corperation** ^{[1]}**.

Figure 1: CAD geometry of Apache-like helicopter. www.grabcad.com

Setting up an analysis with rotation(s) in FloEFD is really easy and straightforward. One merely models a component or part that envelopes the body of rotation. In this particular case simple cylindrical components would suffice, shown as transparent bodies in Figure 2 below. Then during the analysis setup select the sliding rotation option and define the various *rotating regions* by selecting the cylindrical component and specifying the angular velocity, 300*RPM* in this case, making sure about the direction of rotation. This simulation consists of two separate rotating regions for the main rotor and tail rotor respectively. A further consideration would be the transient analysis time step size, since the sliding mesh approach is an inherently transient analysis. A good time step to use is calculated as follows:

timestep = T/(N*10)

where,

T = period of revolution [s],

N = Number of blades.

Figure 2: Rotating Regions. a) Top view. b) Side view.

Once the analysis is set up, the mesh needs to be defined, so let us take a quick moment to discuss the way in which FloEFD handles the CAD geometry and subsequently the meshing approach employed.

**CAD Geometry:**

FloEFD which pioneered the fully CAD-embedded CFD simulation philosophy, allows users to work within their native CAD environment, as FloEFD is fully integrated in Siemens NX, Catia and Creo, without the need to translate from native to neutral CAD formats or export geometry to the CFD simulation software interface. The entire simulation process takes place within the user’s CAD environment, from CAD generation to analysis setup to application of boundary conditions to meshing and post-processing. FloEFD automatically detects the fluid and solid domains, negating the need to perform additional CAD Boolean operations to generate separate fluid domain components upon which the meshing and boundary condition specification actions are performed. In FloEFD the boundary conditions are applied directly to the CAD geometry.

**Meshing:**

The immersed boundary Cartesian mesh approach makes use of the CAD kernel to detect the solid boundaries and cuts the cells at the intersections between the cell faces and the solid boundaries, which makes the meshing a lot less susceptible to surface defects since the solid surfaces need not be meshed separately with a surface mesher. Furthermore, mesh refinement occurs on an octree basis, which means that for each level of refinement, each cell is divided in half in all three Cartesian coordinate directions. Separate local refinement regions can easily be applied to different geometric entities, i.e. vertices, edges, faces, parts or assemblies. For example in this analysis, local mesh refinements were applied to the same rotating region components which were used to define the rotation.

The total number of cells for this analysis was 2.7million. The entire mesh setup including mesh generation time in total amounted to <20 minutes. Thus, it is clear that the meshing process in FloEFD is very predictable and saves you hours, days or even weeks compared to traditional CFD meshing approaches with a lot of meshing related challenges which have become the norm with CFD simulations. With the mesh setup complete we are now ready to start the analysis.

**Solving & Post-processing:**

Once you start solving you can preview cut-plots to monitor the development of the flow field. You can also record these previews during solving as animated videos. The image below shows the velocity field with mesh overlaid as an example of such a recording. Once the solution is finished one can visualise rotation of the rotors and the development of the flow field by introducing cut-plots of velocity with streamlines overlaid, as demonstrated in the video below. FloEFD being fully CAD embedded makes post-processing and analysing the results simple and efficient as one can directly communicate with the CAD geometry when specifying cut-plots, surface plots, plotting XY-graphs of parameters or measuring solid or fluid parameters.

See the animation video below which illustrates the rotating mesh approach along with the velocity contour-plots with overlaid streamlines.

**Validation Case:**

The following validation case is an extract form the Engineering Edge Magazine from Mentor Graphics. The case study involves the analysis of the rotor for a light helicopter model and demonstrates the accuracy of the sliding mesh capability of FloEFD as compared to experimental data, a competitor commercial CFD package and also a specialised code for solving rotational type flows. The rotor consists of one rectangular, untwisted NACA0012 rigid blade mounted on a hub with a balancing weight.

Parameters:

Rotor Radius: R = 1.2*m*

Chord Length: c = 0.15*m*

Rotor Collective Pitch Angle: ϕ = 8°

Angular Velocity of Rotation: ω = 36.5*rad/s* (~350*RPM*)

Figure 3: Rotating mesh validation case – Geometry[1]

The pressure coefficients on the surface of the rotor are presented for free stream velocities of 0 & 11.5*m/s* and for various angles of rotation, ϕ. It is clear that FloEFD closely represents the experimental data and in general performs at the least as good as another reputable competitor traditional CFD software.

Figure 4: Rotating mesh validation case – Comparison of pressure coefficients.

This blog article is Part II of a series of blog articles on the CFD simulation of complex rotor equipment.

Part I: LEGO Technic Aero Hawk

Part II: Apache Helicopter

Part III: Rotating Radar Dish

Part IV: KJ-66 Micro Gas Turbine Thrust Prediction

Part V: Diffusor Shock Waves in a Centrifugal Compressor