Godzilla - stern that's not canoe shapped

I'm experimenting with designing a cruising cat for fun. I may even get it engineered and build it if turns out good, but now I'm trying to use Godzilla to get a efficient underwater shape, I get the length and beam working right, but as soon as I put anything in minimum Bt or Tt or At all the constraints are violated and it never comes up with a solution. Without those it always comes up with a canoe stern. Any ideas what I'm missing?

 

You don't have to have a canoe stern, but the physics incorporated in Michlet are not really capable of modeling an immersed transom.

You might get better results if you have the transom entirely above the waterline. Instead of the cross sectional area going to zero by shrinking primarily in beam, you can have a bottom that sweeps up out of the water so the area goes to zero more by virtue of its depth. This is also an efficient shape for a sailboat as well.

 

In Godzilla, in the input file, you need to first switch to hull type #8 and provide the f0 through f7 coef set. Can you post your input file as an attachment in MS word? There should be examples of the in.mlt file in the examples folder for a transom-sterned vessel.

 

Thanks for all the help. I figured it out with your input. Being a software designer and not a NA or engineer I was getting very frustrated with the software and almost gave up. However now I love it, super powerful and does exactly what I need and for Free - thanks Leo.

I missed the whole primary constraints section and was just putting in the secondary constraints. Once I figured out how to change the series and put in the three numbers for each of the parameters I was off and running. I ended up using a series 20 and am getting the shape I was looking for. Thanks again.

 

It can give reasonable estimates of small transoms, but you're correct that the
physics are poor for large transoms.

Flotilla does a much better job, but I haven't included an optimiser with it.

 

What would be considered a small transom?

25% of midship section? More? Less? Or, put differently, at what point do the results in Michlet become less reliable.

I have been sorting through my own cat design and if the constraints are let loose to run free(series 8), I get a full transom every time. Admittedly, I had forgotten about this limitation in Michlet, but the full transoms were still definitely bothering my sense of design.