Simulating vehicle wading with the SPH Lagrangian method focuses on how vehicles
interact with deep water. These simulations are characterized by a complex geometry
interacting with a free surface fluid, enabling precise modeling of complex shapes
and free surface phenomena, like waves and splashes, generated by the vehicle's
movement.
Vehicle wading simulations in nanoFluidX are always
conducted as single-phase simulations due to the high computational cost associated
with full multiphase simulations. There are two different simulation setups
available.
Moving Vehicle
This approach mimics a real-world condition. A water channel is defined, and the
vehicle drives through it. This approach is more realistic and captures the dynamic
fluid behavior during water entry and exit. It is more suitable for comparing to
physical tests. However, it is the more computationally expensive approach.Figure 1.
Static Vehicle
This approach is simpler and uses a configuration similar to a wind tunnel. The
vehicle remains static, but the road and wheels are assigned boundary conditions
which mimic the correct vehicle speed. Water approaches the vehicle from an inlet at
the front of the channel. This approach is less computationally intensive than the
moving vehicle simulation, but it cannot effectively capture the fluid behavior
during water entry and exit.Figure 2.
Geometry Cleanup and Model Organization
The vehicle model can be imported in any CAD format or using STL. The vehicle model
should be organized into assemblies which contain the vehicle body, with the front
and rear wheels as separate assemblies.
Solution Definition
For the vehicle wading solution type, the essential set of simulation
parameters must be provided. Those parameters include particle size
definition dx, selection of the frame suite (Moving or Static vehicle),
vehicle velocity, water depth, and the total duration of the
simulation.
Particle Resolution
Particle resolution (dx) for vehicle wading is relatively simple and is
normally determined by hardware available and desired runtime. The
typical range is 3mm - 10mm, outside this range accuracy is compromised
too much or the runtime is prohibitive. The recommended resolution is
5mm as a good compromise between accuracy and runtime.
Simulation Duration
This defines the start time (t_begin) and finishing time (t_end) of the
simulation. Start time only needs to be adjusted when running a restart
simulation.
Important: Currently,
there is no check to ensure that the vehicle will reach the end of
the channel when using the Moving Vehicle approach. Ensure that the
time is sufficient for the channel and the vehicle
speed.
Channel Definition and Vehicle Position
The channel is composed of three parts: the profile (cross section),
path, and STL. The solver uses the path and profile definition, while
channel STL is used for visualization. The channel is always created in
the correct orientation, with the origin at 0,0,0. The vehicle must be
aligned with the channel using the Transform tool.
Important: The initial position of the
vehicle must not cross multiple segments of the channel in a moving
vehicle type simulation.
Part Identification and Material Assignment
The materials (phase) types required for the vehicle wading simulation
are created automatically when a Wading Solution is created. Using the
Identify Parts menu to classify the bodies as wheels or car body, the
materials are assigned automatically, and motions are created for the
wheels.
To support situations where fluid forces on the car body produce
non-negligible movement and the rigid suspension model is no longer
applicable, nanoFluidX now includes a linear
spring/damper model for the car body through Double roller
3DoF motion.
Suspension Model
In double roller 3DoF, the wheels remain rigid and follow the road in
the same way as a rigid body motion model while the car body is free to
move in the Z direction and rotate about the Y axis. This provides two
more degrees-of-freedom for a total of three, hence double roller 3DoF.
Double roller 3DoF is similar to a half-car model where tire deformation
is ignored. With the addition of double roller 3DoF motion alongside its
1DoF counterpart, nanoFluidX now offers the
flexibility of choosing the appropriate suspension model in water wading
and other compatible cases.
It is recommended to use the 1DoF motion
variant where fluid forces on the body are negligible, such as shallow
cases, and to use 3DoF only when interaction with fluid becomes
important. In cases where small changes in the road path slope between
different segments is small, using the 1DoF variant is sufficient. The
3DoF variant is recommended only when the modeled linear spring/damper
response results in non-negligible difference in body position and angle
when passing between different road segments.
In addition to wheel radius (dr3dof_whl_rd), body center of mass
(dr3dof_bod_cgpnt) and front and rear axis points (dr3dof_frw_axpnt),
that are common between 1DoF and 3DoF variants, double roller 3DoF
requires body mass (dr3dof_bd_m), moment of inertia about the Y axis
(dr3dof_bod_iyy) and spring (dr3dof_frw_ks) and damper constants
(dr3dof_frw_cs) for front and rear wheels. Front and rear wheel masses
(dr3dof_frw_ms) are optional. The figure below marks the required values
on the schematic.Figure 3. Suspension Model
Domain
Definition of the computation domain is important. The domain must
extend far enough upstream and downstream of the vehicle to capture the
bow wave formed, along with the wake trailing the vehicle. The default
values for domain size (thunderbolt icon) consider that vehicle velocity
stays below 20 km/h.
Probes and Extractors
Alongside the conventional probes that are used to extract accurate
post-processing information at a particular location, it is recommended
that you use the extractors in water management simulations. For more
information, refer to Probes
and Extractors in the
Command Reference section.
Export
Prior to the export of the solver input files, it is recommended to
perform a Data Check to avoid possible inconsistencies and errors.
