Direct Numerical Simulations (no turbulence
modelling) of a starting jet. The streamlines are
coloured by the local velocity magnitude.
Computed ratio of eddy-to-kinematic viscosity
beneath waves propagating on a slope using (top)
standard and (bottom) new turbulence model.
PIV (Particle Image Velocimetry) image of the
wave kinematics in a non-linear wave (MLHV).
Investigations of the behavior of turbulence under industrially relevant
changing conditions
TRL(Turbulence Research Lab) operates on
a combined theoretical, experimental and
computational basis for optimal synergy and
with solid anchoring in the leading international
basic turbulence research community, not least
through its numerous collaborations with leading
experts in relevant disciplines in turbulence
research. Development of novel methods in
experiments, theory and computations tailored
to answering some of the most interesting
questions in turbulence to date is a hallmark of
the research activity of TRL. Industrial collaborations
are directly linked to these activities
with a particular focus on industrially important
jets, wakes and pipe flows.
On the over-production of turbulence beneath surface waves in RANS
models
Computational fluid dynamics (CFD) is increasingly
used to simulate various free-surface
wave processes, including their simple propagation,
interaction with structures, or the highly
complex phenomena of breaking waves. In
previous numerical simulations of breaking
waves, there has been a marked tendency to
severely overestimate turbulence levels, both
pre- and post-breaking. This problem can most
likely be attributed to a previously described
(though seemingly not sufficiently recognized)
instability of widely-used turbulence models,
when they are used to close Reynolds-Averaged
Navier-Stokes (RANS) equations in a region of
nearly potential flow. This results in the unphysical
explosion (exponential growth) of the
turbulent kinetic energy, and hence the eddy
viscosity. While this problem has been known
for nearly 20 years, a suitable solution has yet
to be developed. In this work the instability
problem is re-visited, and it is shown via analysis
that virtually all commonly used two-equation
turbulence closure models suffer from this
problem. Building further on the analysis, a new
formulation of the k-omega model is subsequently
developed, which eliminates the instability
problem in nearly potential flow regions,
while importantly not affecting sheared flow
regions. An example demonstrating the significant
improvement can be seen in the left figure,
which depicts the simulated eddy viscosity
beneath waves propagating on a constant slope
before ultimately breaking. The top figure
demonstrates typical polluted results computed
with the standard k-omega model, where the
eddy viscosity is several orders of magnitude
larger than the kinematic viscosity, even prior
to breaking. This problem is eliminated using
the newly developed model (bottom), where
significant eddy viscosity (hence turbulence) is
now restricted to the surf-zone, in accordance
with measurements and physical expectations.
The new model has been demonstrated to yield
good comparison with breaking wave experiments
in terms of predicted surface elevations,
turbulence, and undertow velocity profiles.
Dynamics of extreme waves and their interaction with offshore structures
Extreme storm wave events represent a great
threat to coastal and offshore structures (e.g.
offshore oil platforms, wind turbines etc.).
These events will grow in frequency and
magnitude given e.g. the anticipated increased
storminess in Northern Europe due to climate
change effects. Furthermore, older platforms
in the sea may have been exposed to subsidence
(sinking of the seabed due to oil and gas
retrieval), which make them more exposed to
wave impact. Traditionally, more open structures,
such as jacket structures, have been used
in the oil and gas industry in for instance the
North Sea. For these open structures, the inertia
forces are less important for large waves,
while other force contributions such as drag and
slamming loads become important. Traditionally,
the Morisons equation is the state of practice
when determining the forces, which takes into
account drag and inertia contributions to the
force. The breaking process is often less violent
in intermediate water compared to shallow
water. The effect of air entrainment has a major
influence on the kinematics in the upper part of
the wave and the generated forces on offshore
structures. The wave dynamics, including
the breaking processes, are studied in details
experimentally and numerically in this project.
Contact:
Clara M. Velte, e-mail: cmve@dtu.dk
Contact:
Bjarke Eltard Larsen, e-mail: bjelt@mek.dtu.dk
David R. Fuhrman, e-mail: drf@mek.dtu.dk
Contact:
Erik Damgaard Christensen, e-mail: edch@mek.dtu.dk,
Malene Hovgaard Vested, e-mail: mlhv@mek.dtu.dk
30 Fluid Mechanics, coastal and maritime engineering