▲ 35 points
back
12 comments
John McPhee write an excellent article about this: "The Ships of Port Revel" by John McPhee. It was an atlantic article and is included in his book "Uncommon Carriers".
I had the opportunity to cross the Panama Canal on a 48ft sailing yacht, starting on the Atlantic side. Inside the watergates we were tied together with two other yachts, one on each side (ours had the strongest engine, so we were in the middle). The canal itself is mostly surrounded by jungle, and you get a specific time slot for crossing it because the traffic is usually quite high, and the watergates need some time to get filled. During the trip it didn't really feel like there was much going on, even though many large vessels crossed us in both directions.
Thanks for posting this and bringing back some memories.
This is one of those things that pretty much is guaranteed to be replaced with VR simulations in 10-15 years. I have to believe that the challenges associated with doing "true to life" simulations of waves, wind, and impact of tugboat, will benefit greatly from a half dozen (remaining?) iterations of Moore's law.
Though - I'm also wondering, is there an opportunity for automation here as well?
I doubt it. I visited an operational naval submarine simulator recently. It was basically the electronic guts of the actual submarine, hooked up to sensor simulators and wired into a replica of the bridge (or whatever they call it on a sub). If they wanted to replicate a ship, they could. there must be a reason they prefer to simulate it in the real world on a smaller scale.
My understanding is that simulators today aren't powerful enough to provide real-world acuity, so, that even though these smaller-scale models come with their own artifacts (including time compression, and in the case of the panama, unrealistic visual comprehension of the environment) - the scale model is still significantly more realistic than what you can do in a simulator.
They probably won't be powerful enough in 5 years, or perhaps even 10 years, but I wouldn't take the bet that in 15 years they won't be able to simulate, with sufficient fidelity for pilot training, a ship moving through the panama canal.
I'm guessing the complexity of the ship-in-panama model environment is a couple orders of magnitude less than what the first officer of an airplane has to deal with; it will be a lot longer before we'll be able to faithfully replicate that environment - if ever. (Far enough out that I can't even begin to guess if it's 50 or 100 years from now).
Now, the wild card is, that even though you might be able to do the Panama simulation in 15 years from a computing perspective, it may still be much cheaper to simply do your 25:1 replica, particularly as you don't really need to train all that many pilots in the big scheme of things...
Unless you can plug directly into the inner ear, any sailing or piloting simulation is going to be fairly poor.
It's not clear that's relevant when piloting a cargo container on the Panama canal. If anything, the 25:1 mock up is problematic, because of time compression. The pilot will get acceleration and feedback that will be totally missing in the real Panama.
Do all of the ship handling parameters scale in the same fashion as the size of the ship is reduced? The article doesn't elaborate on this, but I bet there are some interesting technical considerations involved in ensuring that everything is "true to scale" (or even in defining what that means!).
No, they would have to do some interesting things to the density or viscosity of the water for the handling to be similar. Or they could operate at 25x the speed that they currently are, but that doesn't seem feasible. See http://en.wikipedia.org/wiki/Reynolds_number
This seems perfect for illustrating my docker container slides!
Pilots?! Where are the aircrafts?