SAIC Internship Summary

(Apologies in advance if some images don't make sense because they lack numbers.  The government is paranoid, so these are the "sanitized" images)

I just finished an awesome internship at SAIC and want to record a bit about what I did there. 

Basically, there's this this awesome trimaran boat, which means it looks really cool. It also means it's a weird type of boat that isn't built very often, so all the rules the Navy uses to build boats might not apply.  SAIC has this awesome software that lets them accurately simulate the ocean and calculate the resulting forces on boats.  So the Navy hired SAIC to run simulations of a CAD model of this boat running through decades of simulated waves and see what forces to expect and design for.  

Here's an example of three of the 15 forces calculated for, at only one point along the boat.

There are 15 forces (9 forces, 6 moments) calculated over many parts of the trimaran and for many boat states (speeds, headings) and ocean states (wave height, wave modal period) and many many simulations.  So this magical software spits out a lot of data. Not all of it is important.

I was in charge of turning that mountain of data into useful visualizations for the naval architects.  I chose to use Python with Numpy and Matplotlib to accomplish this.  The first image of this post is the maximum of one force at one point for various ocean states, but only one boat state (as fast as it can go, and 180˚ (into the waves)), so in order to show more dimensions of data, I used a graph of graphs to show that one force at one point across all boat and ocean states.  That's still only one force though, so I overlayed all of the forces on top of each other (normalized) and used the clickability allowed by Matplotlib to let the user explore which forces dominate under which conditions.  The users loved this, as it shows all of the important information in one plot and gives them a feeling of the nuances involved. So as you can see below, many of the forces are maximized in the slowest speed going headlong into the waves, but some forces are maximized going fast perpendicular to the waves.  These forces are probably bending moments and vertical shear forces acting on the boat as it rocks violently side to side. Maybe that sounds too confusing, but trust me, they loved it.

I made some other types of graphs to reveal other correlations. With all the slices next to each other you can see how certain forces change spatially.  

Another visualization was to show how one ocean and boat state affected the six most important forces and moments at various positions on the boat. 

I also visualized the Weibull plot which is the statistical method used to extrapolate the probable lifetime maximum from a limited simulation time.  In this way, our servers simulated weeks or months of ocean time, and we can confidently extrapolate the maximum force the boat would experience in 30 years of ocean time.