At Cape Horn Engineering we have forgone traditional towing tank and wind tunnel tests in favor of an exclusively CFD based design philosophy. Our simulations are cheaper, faster and more reliable than traditional tests. We run all simulations at full scale which eliminates the inherent error in scaled test results. Enhanced flow visualization and force decomposition give designers much greater understanding of flow phenomena.
We run all hydrodynamic and aerodynamic simulations separately. Hydrodynamic simulations are used to research hull form, yacht behavior in waves, and appendage shape and position. Early in the design process, aerodynamic simulations are used to determine the sail forces which are in turn used as input in hydrodynamic simulations. Later we return to aerodynamic simulations in order to optimize sail shapes and investigate new sail concepts.
Our hydrodynamic cases are run using full size hull models with rudders, keels and foils. The free surface is modeled using a volume fraction (VoF) method, and simulations allow for dynamic trim and sinkage. Our computational models consist of unstructured hexahedral meshes with extensive local refinement and roughly two million cells. Appendages are meshed independently, giving us the freedom to test many different appendage shapes and sizes quite easily.
In aerodynamic simulations we model all geometry above the static water plane, including the sails, mast, boom, deck and hull. The rig is tilted to the correct attitude accurately modeling heel, pitch and yaw. A varying wind profile is used to account for the boundary layer along the surface of the water. The aerodynamic computational models use automatically generated polyhedral meshes with prism layers. The aerodynamic meshes have a similar cell count but many more nodes than the hydrodynamic meshes.
For sail analysis we use parametric modeling and fluid structure interaction (FSI) codes. In parametric models we use CFD to find the optimal aerodynamic sail shape. Designers then develop a sail with the appropriate structural elements in order to achieve this optimal 'flying' shape. Parametric variation is done either by using a predefined matrix of variations, or by incorporating an optimizing algorithm in an iterative loop. In FSI models, the CFD code passes pressure forces into a finite element model which then calculates the deformed shape of the sail. The new deformed shape is trimmed by the sail designer and put back into the CFD simulation. This cycle is repeated four to five times until convergence is reached.
All of our simulations are processed in parallel on four cores; with 436 processing cores in our cluster we can run over 100 cases concurrently. Each case is part of a matrix of predefined sailing conditions. In general hydrodynamic cases are run as unsteady simulations to allow free surface waves to develop. These take between 12 and 24 hours to complete. Alternately, most of our aerodynamic cases can be run as steady state simulations which converge quickly because there is very little separation on the sails of high performance racing boats. These sail cases generally run in only a few hours. Combined, this means we are capable of running several hundred simulations per day.
In hull shape studies we use fully appended models with a reference set of rudders, keels and foils. The position of each appendage shifts to match the shape of each candidate hull. For example, the rudders will change angle slightly to remain perpendicular to the hull surface on each hull.
For appendage studies we use a set of reference hulls and change appendage concepts, shapes, positions and orientations. Each variation is evaluated by comparing the resulting forces (drag, side force, roll and yaw moments) and by comparing the flow characteristics using stream lines and other visual techniques.
The Volvo 70 Class boats present a new, very complex, design problem. Compared to America's Cup Yachts there are many more design variables; the Volvo 70's are designed to a box rule and experience sailing conditions from all over the world. Boat speed ranges from 6 knots as a displacement hull to 30 knots planing and surfing down waves. The boats also have canting keels and water ballast which drastically changes the displacement and center of gravity. All of these variables lead to very large testing matrices. Here CFD becomes very attractive; each variable usually can be changed simply by changing a number and running the simulation again.
Our hull shape research program is quite extensive. The design spiral begins with the required transverse stability and waterline beam. We then obtain accurate sail force coefficients using our own aerodynamic simulations. Different sail sets are tested for upwind and downwind sailing in light and heavy conditions. This is important, especially for very beamy boats, because the sail forces can have a large effect on the longitudinal trim which changes the drag. Instead of attempting to match a given sail set to specific sailing conditions, which is difficult and error prone, we find the center of effort for the sails and apply a force vector at that point in the hydrodynamic model. Then the hydrodynamic simulation runs so the drag on the boat matches the force generated by the sails.
The hull shape investigation continues with studies of volume distribution, prismatic coefficient, transom width and immersion, bow fullness, etc. Hull shapes are organized with parent hull shapes and their derivatives. This allows for easy analysis of trends and performance drivers and final selection of a hull shape. After the final hull has been chosen and the lines have been sent to the builder, research continues with appendages and sails. We investigate the size, shape, and position of the keel, bulb, rudders and dagger boards. We adjust the transverse inclination, longitudinal inclination (sweep), alignment of the keel cant axis, pitch of the bulb, angle of attack of the dagger boards, etc. Different solutions and details for the attachment of foils to the hull are investigated, like the recess or 'bubble' in the hull where the keel attachment (a cylinder) rotates, and the fairings between hull and foils.
Candidate boats are then run through race simulations to determine which design is the best. VPP is used to analyze trade-offs, such as stability versus drag, and to determine optimum balance. Our Router program simulates the best course for each hull using statistical weather data for relevant parts of the world, and then compares the time needed to complete the course for each hull. Thus, the overall winner of the race is found in probabilistic terms. Other design variations, such as the position of a dagger board, are more straightforward to analyse and usually it suffices to compare the amount of drag at a given side force to draw conclusions.
In many cases the design and modification of appendages and sails is given a reality check when tested on the trial boat in real sailing conditions. During this on the water training period we are given valuable feedback from the real world that keeps us motivated and focused.
Finally, a set of seakeeping simulations are performed to investigate the dynamic behavior of the final candidate hulls. These simulations ascertain that what is good in calm water does not become detrimental in waves. Due to the nature of these time consuming simulations, only regular waves are considered. To analyze a complete sea spectrum would require a statistical analysis over a large period of time, i.e. half an hour; however, the response to regular waves can be found in less than a minute of simulated time. Despite this restriction, the information gathered in regular waves is indicative of real sea state performance. Unlike a towing tank, no other simplifications are necessary; in our simulations, any wave direction is allowed, the boat is at full scale, the center of mass is in the right position and the moments of inertia will be the actual estimated values.
© 2013 Cape Horn Engineering. All Rights Reserved.