Section after Section, How designers get a smooth continious 3d surface ?

Not the CFD results, but you can start with the hydrostatic characteristics and work backwards. It's basically a process of starting with top level requirements and then creating more detailed subordinate requirements.

At the top level, you probably have the displacement and the length of the boat as a result of defining the boat's mission. The mission will also determine the general type of hull.

Based on the hull type, you can start to create some parametric families of curves. You can have a family of curves that will distribute the displacement along the length of the boat, creating cross sectional areas at each station. The shape of the cross sectional area distribution will determine were the longitudinal center of buoyancy will be.

You can also define the general shape of the sections for each station. You then define functions that result in a smooth variation of the section parameters along the length. Then scale the sections to match the required cross sectional area distribution. If all your functions are of low order, that should give you a shape that is fair and exactly matches the top level requirements. Chances are, the beam was determined by the scaling of the sections, so you may want to adjust the section parameters to get a particular beam or to achieve the desired roll stability.

Once you have all the section shapes defined, you can output the offsets to a CAD program to generate the hull shape. You will probably need to add some additional treatment at the bow & stern, because the process above only covers the length between the forward and aft perpendiculars. If your CAD program supports object linking and embedding, you can implement the process above in a spreadsheet and have it create the hull in CAD automatically.

Here's an example of applying the technique.

For CFD, you'll need to automate the process of turning your CAD model into a grid for the particular CFD code. Most commercial meshers will accept CAD output, such as a step file (.stp). I don't know of any 3D inverse CFD methods for marine applications, and that's what you'd need to determine the hull shape from CFD. However, wave drag depends mostly on the cross sectional area distribution and only weakly on the details of the section shape, and skin friction depends mostly on the total wetted area. So when you choose the cross sectional area distribution and the generic section shape, you are also determining the principal hydrodynamic characteristics.

This still doesn't take the art out of hull design. It merely shifts it from drafting lines to creating the functions that define the section parameters and their variation along the length.

 

http://www.boatdesign.net/gallery/showphoto.php/photo/24823