Propeller Models and Rotational Volume Zones
2.3. Propeller Models and Rotational Volume Zones#
Flow360 offers a variety of options to simulate propellers. Each option strikes a different balance between simplicity, ease of use and accuracy.
2.3.2. Actuator Disk (AD)#
This is the simplest modeling technique where the propeller’s effect on the flow is modeled as a momentum source. It is useful when the propeller performance is known and thus the simulation is more focused on the effects of the propeller wash on items downstream. It assumes a circumferentially uniform pressure change so it cannot be used when the propeller disk is not aligned with the incoming flow, i.e. when the thrust and torque distribution is not circumferentially uniform across the propeller disk.
The Actuator Disk (AD) implementation requires a thrust vs radius and a torque vs radius list. This information usually comes from other modeling codes. Please refer to actuatorDisks in Solver Configuration for the expected format of the input data.
The AD is a simple way to add the effects of a known propeller into a flow field (see this paper) but it does have a couple of drawbacks:
The fact that it assumes a circumferentially uniform thrust and torque distribution means it cannot simulate any asymmetry in the flow field approaching the propeller disk (e.g., due to angle of attack, angle of yaw, body upstream deforming the incoming flow, etc.)
The input values need to be updated for each case configuration. You will need to update the thrust and torque values when changing any relevant simulation parameters.
2.3.3. Blade Element Theory Disk (BET Disk)#
This is one of the most commonly used propeller modeling techniques. It captures almost all of the physics without requiring an actual propeller geometry. Input requirements are a set of radial geometric details and associated 2D Cl and Cd polar information. Please refer to BETDisks in Solver Configuration for the expected format of the input data. From that data the BET Disk implementation calculates how a propeller would affect the flow locally and assigns the correct forcing terms. This is very useful for understanding propeller behavior and how the propeller wake will affect objects downstream.
BET Disk is useful at all steps of the design cycle, from propeller design to propulsion integration for a full vehicle. It is the best simulation technique if there is no available CAD model of the propeller geometry and when reduced computational costs are required. Overall the BET Disk modeling approach strikes an appropriate balance between ease of use, cost to run, and accuracy.
2.3.4. Blade Element Theory Line (BET Line)#
This is a similar implementation as the steady BET Disk above but in a time resolving fashion. Transient effects due to individual rotating blades and vortex shedding are captured in time instead of being averaged out as with the steady BET Disk method.
It allows for accurate simulations of transient propeller effects without the need for propeller CAD. Even if a CAD model is available, the BET Line method can be less mesh intensive than meshing the propeller details and its associated rotational volume zone method.
2.3.5. Rotational Volume Zone#
Rotational volume zones allow for the movement of different regions of a multi-domain mesh relative to one another. A rotational volume zone can be any circular shape (e.g., cylinder, sphere, etc.) for which one side is stationary relative to the other that rotates. Nested rotational volume zones are possible to allow for complex cyclical motions.
For propellers, this is the most accurate of all modeling options. It is also the most costly in terms of resources because it requires an accurate mesh of the propeller geometry, adding mesh count and modeling effort. Also, relatively small time steps are usually required to accurately capture the propeller’s forces as well as extended time spans to sufficiently develop the propeller’s wake far downstream. Various solver settings and solution strategies can be applied to improve simulation efficiency. Please reference the XV-15 tutorial and this publication for example processes and techniques
The time-accurate rotating geometry option can simulate a lot more then just propellers. Anything that moves in a rotational frame of reference can be simulated: dynamic derivatives, airplane spin, on-the-fly control surface deflections are but a few examples of what can be done with this approach.
It is recommended to use volumeZones->referenceFrame instead of
slidingInterfaces to specify rotational volume zones.
For further comparison of the pros and cons of each propeller modeling technique please see this publication on the impacts of modeling approach.
2.3.6. Multiple Reference Frame (MRF)#
The MRF (Multiple Reference Frame) method is an efficient approach for simulating rotating machinery, offering significant computational advantages when contrasted with the unsteady sliding interfaces, as the flow field can be solved using the steady state flow solver. Unlike the unsteady sliding interfaces approach, where the mesh rotates alongside the rotor, the steady state MRF employs a different strategy. Here, both the mesh and the rotor remain stationary in space, while the impact of rotation is considered by incorporating centrifugal and Coriolis source terms into the Navier Stokes equations, within individual volume zones. This approach is recommended for steady simulations of rotating components such as rotors, turbomachinery, wind turbines over the SRF approach.
2.3.7. Single Reference Frame (SRF)#
In contrast to the Multiple Reference Frame (MRF) method, where both stationary and rotating volume zones coexist, the Single Reference Frame (SRF) approach involves the entire simulation domain rotating around an axis. This approach is recommended for steady simulations involving maneuvers (pitch/roll/yaw), over the MRF approach. In SRF simulations, the “velocityType” parameter under the “NoSlipWall” and “Freestream” boundary conditions offers two options: “relative” (enforcing zero relative velocity) or “absolute” (enforcing zero absolute velocity). For NoSlipWall, “relative” is used for the rotating surfaces (e.g., blades of a rotor), whereas “absolute” is used for fixed surfaces (e.g., wind tunnel walls).