| |
A properly designed
propeller is a compromise between structural and hydrodynamic
considerations; moreover it should match the boat and
engine characteristics. An incorrect design may easily
lead to performance and speed losses, as well as to
cavitation, noise, vibrations and blade erosion. A
clear understanding of propeller flow is therefore
required. Firstly, it is necessary to know how the
drag, trim and flow field around the boat change with
speed; this should be done, in theory, with model testing
in the so-called "towing tank". Second, once
the propeller has been designed, it should be analyzed
in model scale in a "cavitation tunnel".
Both tests are time consuming and expensive (much more
that the propeller itself in most cases) and for this
reason they are seldom carried out for custom applications
on small and medium size boats.
Research and innovation have
always been the key factors of our success. In the
past, a large effort has been undertaken by our company
to test and optimize families of propellers in a
cavitation tunnel. Some years ago it became clear
that, besides experimental investigations, a breakthrough
could be achieved using CFD. This is because, in
our opinion, not only the raw numbers obtainable
by the towing tank and the cavitation tunnel are
important, but also the flow visualization offered
by CFD programs allows better comprehension of the
flow phenomena, thus speeding up the "learning
curve" of the designer.
Our first objective was to use
a tool capable of computing the hydrodynamic resistance
of fast planing hulls, which represent the majority
of our clients’ applications. This is a field
where state-of-the-art "Panel Methods" fail,
mainly due to the complexity of the free surface
generated (spray, breaking waves, complex flow at
the transom stern). Among many commercial codes tested,
we found that CD-adapco’s alternative CFD solver
Comet has best fulfilled our requirements. After
a set up period, we obtained results of the same
order of accuracy as a towing tank. This enabled
us to understand the boundary conditions and the
input data needed for propeller design. Moreover,
we could look at the pressure distribution and the
flow field, detecting problems and suggesting possible
improvements to designers of hull shapes.
Besides the computation of planing
hulls in calm water, we found it useful to explore
a similar field where, once again, no other tool
was available for computation. Performances of planing
boats are strongly affected by natural sea waves,
and the best hull shape must be a trade-off between
calm water and sea-keeping qualities. Sea-keeping
simulations at any speed and wave characteristic,
are now routinely performed using Comet. The most
useful results are the level of acceleration onboard,
the added resistance to waves and the instantaneous
pressure distribution over the hull bottom, this
last result being extremely useful for structure
scantling.
Another field where the Comet
solver proved to be a valuable tool was in the computation
of Surface Piercing Propellers (SPP). At very high
speed, when cavitation becomes unavoidable and strongly
erosive, it is helpful to let the propeller work
at the interface between the water and the atmosphere,
allowing the water vapor cavities formed over the
blades to be filled with air. In this case the free
surface becomes extremely complex; a thin unsteady
pocket of air is formed over the blade surface and
a large spray is developed.
Using the VOF and "sliding
surface" capabilities of the code, it has been
possible to obtain results that match the data obtained
in the cavitation tunnel within a few percent. Submerged
propellers are designed using an "in house" developed
Panel Method; this gives a reasonable approximation
of the flow for conventional propellers working at
the design point. Unfortunately, for off design conditions
or heavily loaded propellers, unacceptably large
errors may occur. For this reason every propeller
designed is then checked in a wide range of conditions
using the Comet solver and eventually corrected before
manufacture. The large computational resources required
for these CFD computations are achieved with an SGI
Origin 2000 system, using 24 processors and 10 GB
of RAM. An effort of flexibility has been essential
to switch from the old established way of designing
marine propellers to these new techniques; the ROLLA
Research branch was created for this purpose. Its
aim is to link engineers, production and customers,
and provide them with the best tools for answering
the many questions coming up in boat or ship design.
There is no doubt that these
technologies have largely improved the quality
of all our products, reducing the sources of error
and uncertainty. For further information, visit: www.rolla.com, www.rolla-propellors.ch
|
Calculated
pressure
distribution of a Suface Piercing Propeller
CFD computation on a
Submerged Propeller (velocity field)
Sea keeping simulation for a fast
planing hull
Wave field generated by
a planing hull
|
|
Your route to this page: