The BMW Oracle Trimaran was a marvel of engineering resulting from over two and half years development. Applied Fluid Technologies played a crucial role in that development, and was in charge of the boat’s entire aerodynamic package.  Our contributions ranged from definition of the sail plan; to optimization of the individual sail shapes; to mast section shape and profile design; to development of sail trim and yacht performance targets.  AFT software and personnel proved crucial in helping designers and sailors obtain maximum performance from the rig and boat.  But AFT’s contribution didn’t end there.  Great efforts were made to reduce overall platform windage (by managing both CFD studies and a large-scale wind tunnel program), and to design aerodynamic fairings for the boat’s various structural elements.

When the idea of a rigid wing sail become reality, AFT was asked to lead the 3D analysis of various design candidates.  Its cutting edge CFD tools, verified once again by wind tunnel results, helped steer the design into its final form.  Once an extensive performance analysis of the final design was completed, and the computed forces turned over to the performance engineers, AFT’s attentions turned to the design of foresails for use with the wing.  These sails proved essential to BMW Oracle’s success in the light wind conditions encountered during racing off the coast of the Spanish city of Valencia.

Because of his extensive expertise with viscous flow, Dr. Korpus was also asked to look into the possibility of a drag reduction system.  In collaboration with Professor Steven Ceccio (University of Michigan), he developed a system to test the concept of polymer drag reduction on sail boats.  A prototype system was built and tested on a 40 foot racing catamaran, and proved highly successful during sea-trials.  A more complex production version was therefore designed for the America’s Cup race boat, and assembled by the BMW Oracle shore team.   Sea trials on the trimaran showed the system to be a huge success.

During the almost three years spent consulting for BMW Oracle Racing, Applied Fluid Technologies made significant contributions in a wide range of design areas: from above the water to below; and from computational aerodynamics to wind tunnel testing.   Each, in its way, helped BMW Oracle in its successful quest to win back the oldest trophy in sports.

AFT RANS codes have been used to investigate the mitigation performance of a wide range of concepts as shown at right.  The top figure shows boundary layer blowing applied to spar platforms, AFT solutions demonstrate that this concept has the potential to virtually eliminate VIM at all current speeds, but only by pumping potentially large volumes of water.  The second figure shows an attempt to apply the same physics without active pumping.  It utilizes a guiding sleeve to extract high-pressure flow near stagnation and reapply it near separation.  The concept was found to reduce VIM amplitude, although not as significantly as active blowing.  The bottom two figures show risers fitted with strakes and fairings.  Each of these last two concepts is found to work in some cases but not others. 

AFT’s approach towards VIM simulation utilizes our unique overset grid approach.  Grid blocks that resolve surface detail and boundary layers move with the body to maintain accuracy.  Far-field blocks remain fixed in the Earth’s inertial frame, and communication between the two is maintained by time-accurate interpolation.  Marine fouling is modeled using specially developed turbulence models to account for roughness heights comparable to boundary layer thickness.  Additional details of this work can be found in the papers: “Active and Passive Control of Spar Vortex-Induced Motions,” and “Prediction of Viscous Forces on Oscillating Cylinders by Reynolds-Averaged Navier-Stokes Solver.” 

RANS was used to quantify lift, drag, and acoustic performance of a large number of candidate designs.  A tandem wing design was ultimately selected, and tested over a wide range of design variables like camber, angles-of-attack, thickness, and relative placement of forward and aft component.  An optimization criterion was devised to identify the best design for maximizing lift across the required range of tow speeds.  Acoustic signature (due to unsteady wing surface pressures) was constrained to an acceptable level throughout the process.  The final design achieves lift areas (i.e. lift coefficient times planform area) over 1.9 square feet – an amazing value for such a small package.  Once design work was complete, RANS was applied again to predict maneuvering and stability derivatives and develop the module’s flight control algorithms. 

In fact, AFT’s RANS codes handle these types of geometries fairly easily, and even allow for design studies where the tunnel and lip are relocated or reshaped.  Stator and rotor rows are easily included using our moving mesh capability whereby some grid blocks rotate with the propulsor while other remain fixed to the hull.  Even tip leakage effects can be resolved.  The figure at right shows velocity vectors on a vertical plane through the tunnel centerline, and demonstrates that flow predictions are smooth and continuous despite the complex grid connectivity. 

AFT’s RANS capabilities have been used to aid the design of nozzles customized for minimum drag, maximum capture volume, and minimum exterior flow disturbance.  Other potential applications include quantification of sampling probabilities for biological sensors, drag prediction for optimizing placement of aircraft inlets, and minimizing pressure drop in injectors. 

AFT RANS codes handle these challenges as demonstrated by the figures at right.  Hull, outrigger, and propulsor blocks are entirely independent, and therefore easily moved relative to each other.  Changing the shape of one piece does not require regridding the entire geometry.  Both resistance and powering optimization studies are therefore easily performed within relatively short turn-around times. 

RANS circumvents this problem by performing flow simulations at full scale, and including both hull and propeller. Empirical values like “thrust deduction” are not needed because the propeller’s effect on stern pressure is already included. Empirical extrapolations like effective wake are not needed because the propeller’s impact on stern flow velocity is also already included.

The difference between RANS and the empirical approach can be startling. The first figure at right shows effective wake behind a single screw tanker obtained using the traditional approach. The lower figure shows the same result obtained from full scale RANS simulation. Neither the shape nor magnitude of the individual contours compares. A propeller designed to work with the first effective wake will not prove optimal for the second. Both propulsive efficiency and vibration performance would likely suffer.

A typical calculation is presented at right, and shows surface limit streamlines and cross flow planes of velocity near the stern of a tanker. The design shows higher than normal circumferential non-uniformity in propeller inflow, and therefore the potential for low propulsive efficiency or propeller vibration. RANS simulations were used to trace the problem back to excessive buttock line slope causing boundary layer vorticity to separate into the wake. Hull form modifications were suggested to fix the problem. If the design had reached the point where hull shape could no longer be changed, RANS could also be used to provide propeller designs for maximizing efficiency and minimizing vibration.

RANS has also been applied to design multi-component propulsors, ducted propellers, stator rows, and blade shape details. It provides the potential to truly optimize propulsor/hull coupling by simultaneously analyzing hull and propeller flow. More detail describing AFT’s approach can be found in “Hydrodynamic Design of Integrated Propulsor/Stern Concepts by Reynolds-Averaged Navier Stokes Techniques”.

Mast design entails a compromise between structural stiffness and aerodynamic efficiency. Stiffness helps with mainsail shape control, but requires longer and/or wider tubes. Unfortunately, the bigger sections disrupt flow over the mainsail, and the two requirements tend to work against each other. The figure at right demonstrates the point by showing mainsail surface streamlines on the leeward side. Streamlines close to the mast move forward and up, and show the large extent of separation.

Mast aerodynamics is extremely complex because the yacht’s heel, the Earth’s boundary layer, and the interacting sails all combine to make the flow highly three-dimensional. Sail efficiency depends on the separated mast boundary layer reattaching to the mainsail; so any useful design tool must accurate resolve this difficult trait. Gridding the complex shapes where mast meets sail and mast meets deck is difficult but crucial to solution accuracy. Finally, it is not sufficient to guide the design based solely on force predictions. Because mainsail efficiency affects the angle a boat can sail relative to the wind, a Velocity Prediction Program (VPP) must be used to rank candidate designs.

AFT meets these demands using the overset grid method demonstrated in the second picture at right. The view looks down from above, and shows mast blocks, mainsail blocks, and Genoa blocks. The bottom figure shows velocities computed from this grid, and demonstrates that despite the complicated connectivity of blocks flow is resolved smoothly. Forces are obtained by integrating pressure and shear stress, and the process is repeated across a range of wind angles, wind speeds, and heel angles. The six degree-of-freedom forces from each simulation are then fed into AFT’s VPP to judge design performance. Mast shape optimization therefore results not necessarily in the design with the most efficient aerodynamics, but with the fastest speed made good along the racecourse.