|
The
main goal of the test case is to find out benefit
of grid computing under FlowGrid system and to compare
FlowGrid results with adequate experiments or other
computer simulations. Optimal test case was searched
and determined in external aerodynamics of railway
vehicles. This branch of fluid dynamics allows to
find interesting phenomena and to create several important
setups for various physical conditions for the test
cases:
- steady calculation
of flows around trains (in the open air or in a
tunnel),
- unsteady transient
calculation of generation of vortex wake behind
the train or the details on train surface,
- aerodynamics of
moving railway vehicles.
The acceptable test
case should fulfil the conditions enabling comparisons
with other simulations or experiments, having well-determined
boundary conditions and other settings, and showing
the benefit of the FlowGrid project. Several simple
and more complicated cases were setup and calculated
in order to choose only one optimal test case setup.
These calculations were made with CFD system Fluent.
The test case is a steady calculation of 3D train
model in tunnel with detailed geometry is optimal
to compare FlowGrid and Fluent results.
The SKODA test case helps to create relevant computational
task to validate and control FlowGrid computational
system. The setup of the test case has to reflect
all requirements of an industrial user of CFD methods
in the innovation process in his branch – it
is transport for SKODA. A good test case has to describe
nowadays and future requirements for up-to-date CFD
simulation.
Properties of the test
case are given by SKODA produce programme in transport
branch – rail vehicle aerodynamics. The subject
of the test case is 3D-detailed model of a train in
a tunnel (acronym for test case identification is
TRAIN). The train consists of prototype loco and first
wagon. The model for simulation is SKODA loco. It
is one CFD simulations loco shape variant of the prepared
prototype. The typical train model, which can be used
to simulation, are as in figure 3.1.
Train for CFD test case simulation
Computational domain is long and narrow volume. Train
model and tunnel profile is shown in figure 3.2 in
which is inlet (blue) and outlet (red) profile. Length
of computational domain is 88 m.
Computational domain - train in tunnel
Description
of relevant parameters to be observed
FlowGrid
application is external aerodynamics of rail vehicles.
This task is based on solving steady calculation of
3D-model of a train in a tunnel. Train ride in a tunnel
at constant speed generates several important phenomena
for CFD analyses. Requirements of the CFD industrial
user are focused on finding main aerodynamic effects
and vehicle loads and on studying the vehicle interaction
with the tunnel walls. The industrial user is interested
in the influence of the front and back loco nose shape
or in the influence of a gap between loco and 1st
wagon on the aerodynamics. Aerodynamic response of
main details in loco surface is important for designers.
Results of complex CFD analyses of train aerodynamics
should include values which describe aerodynamic load
on the vehicle surface, flow wake - behind a train,
behind a loco, behind details on the loco surface.
Expected results format should be as follows:
- pressure
field – filled/non-filled contours of pressure
(total, dynamic, static)
- velocities
- contours of velocity magnitude
- vectors
of velocity
- graphs
– flow values vs. x,y,z-coordinates or time
on independent axis
These
results should show flow situation in various selected
points, vertexes, faces of computational domain and
train model. Loco surface pressure map is the example.
The loco nose part, loco details between bogies and
gap between loco and wagon are important places for
results studying.
Available
experimental data
Train
external aerodynamics – it is not easy to find
relevant experimental results for comparison with
FlowGrid computing. The aerodynamic test tunnel measurements
of train models usually consist only of drag-coefficient
results, or measurements in the frontal section of
the train and locomotive nose for stationary non-moving
models. Results for real train or tunnel interactions
have not been found for this category of trains (trains
are going at maximal speed 200 km/h). High-speed train
measurements in the open air and in the tunnels are
possible to search in references. At this point it
appears that the feasible possibility is to compare
FlowGrid test results with third-party-software simulations
|
