Monday, May 21, 2012

Setting Sails

It is important to realize that sails are wings and they often act like wings (this applies to almost any sail and not just rigid wing sails now popular in the America's Cup race). Therefore, when it is required to optimize the sail settings for a specific boat, we can rely on our knowledge of wing aerodynamics and our bag of accumulated tricks (experience & tools) to do an efficient job.

The primary goal of sail optimization is to increase the lift (for a good drive force component) and reduce the drag (stability must be maintained during the process).  Sail optimization is a repetitive job.  This is because, like wings, the drive force of the sail in up-wind sailing depends on the apparent wind vector (direction and magnitude), the planform (area & aspect ratio), the camber (flat or curved cross sections), the gap between the fore and main sails, the mast and other parameters (such as height above the deck, hull shape, etc.).

Stallion 3D Simulation of flat thin sails. The graphs shows pressure.

A good design does not simply mean infinite aspect ratio to reduce tip vortices (tip vortices increase drag as seen in above picture).  The heeling moment will be too big (unless, of course, you have this installed on your boat).  Large camber or deeper sails can increase the drive force, however, stability and drag from supporting underwater devices (keel and rudder) can erode this advantage. To obtain the optimal (or best we can do at the moment) specifications for a particular sail, we must test our design iterations against a number of possible sailing conditions. 

Stallion 3D Simulation of cambered thin sails. 

Deadlines are the natural enemy of testing and optimization.  Experimental setup and testing in a wind tunnel can be time consuming and costly (and might even require a minor in wood carving).  Wind tunnel and tow-tank test can be made more efficient and cost-effective only if the most promising designs are tested prior to making the final decision.

Stallion 3D Simulation of cambered thin sails with a larger angle between jib and main sail.

Computational methods can be used to test conceptual and preliminary design ideas.  However, a good understanding of the assumptions used in a particular method is required to get useful information to test in the wind tunnels and water tanks.  What is the difference between 2D sections, vortex lattice, panel methods, Euler/Navier-Stokes methods? They are all useful.   Knowing the answer and how to apply the various concepts in a concerted manner can speed up your upwind sail to the best design.

No sail optimization study is complete without the consideration of the underwater systems (keel, rudders and other appendages). Like sails, keels and rudders are wings and also behave as wings.  The goal to provide stability and lateral resistance can result in drag (induced and profile).  Also, an efficient keel for up-wind sailing can be terrible otherwise.  In short, sail and keel/rudder analysis are coupled and equal partners in sailboat optimization.

Stallion 3D simulation of a keel and bulb.

In the area of aerodynamics conceptual design, a lot of nautical mileage can be quickly covered with a tool that has a built-in set of realistic physical assumptions, automatic grid generation and cost effective and readily available computing platform.  Stallion 3D can be deployed on an ordinary Windows PC, run in multiple directories to take advantage of multi-core processing and efficiently and accurately analyze the most difficult models.

Stallion 3D simulation of sails and hull (3D model from http://turbosquid.com).
Graph shows surface speed in m/s.

More information can be found at http://www.hanleyinnovations.com/stallion3d.html.  The cost of a 3-months lease of Stallion 3D is $895.

Thanks for reading. 

Do not hesitate to email or call me at (352) 240-3658 if you have any questions.

2 comments:

  1. Hi.
    Do you have a package for the study of turbocharger rotor blade designs, or automotive oil pumps and the like?
    thanks
    Ross.mcneilage@yahoo.com.au

    ReplyDelete
  2. Thanks for your comment. We do not have a package for turbomachinery analysis as yet.

    Best wishes,
    Patrick

    ReplyDelete