Update: I submittedthis idea to BP. That's a PDF to submit with in case you have an idea too. Hopefully it does not go directly to the round file.
===========
Check my math but this seems like a way to capture most of the oil from the BP blowout that is very doable. The key is to think buckets, not pipes, and only two buckets (upside down and big, but within the demonstrated range of feasibility).
The production volume is reported to be 200,000 gallons a day. Assume that's accurate (I know, first mistake, dealt with below). That is 26,736 cubic feet. A cube 33 feet on a side is 35,937 cubic feet, meaning it could hold 268,827 gallons, enough for one day's production. A box 35 feet on a side holds 320,000 gallons, or a day and a half's worth.
Read below to see how to fix the Gulf Gusher:
Make just two collectors in this size range. Lower one into the rising stream near the source and collect for a day. Replace it with the other one when the first is raised. Empty and clean off wax buildup in the raised one. Repeat daily for as long as needed, until the relief well is functioning, for instance, and the original can be capped.
Problem: Would it be too heavy for the cranes?
Solution: The 90 ton original box/funnel was made of about 3600 square feet of steel (not counting the horizontal "fins") and they handled that with no apparent problem. A 35x35x35 foot box would have 6125 square feet of material. That's significantly bigger. I have no idea but I bet they could. Or make them smaller/lighter and swap out more often.
Problem: Maintaining precise position is impossible for a ship at sea so the collector would move around and might hit and damage the site more.
Solution: Build a rack designed to hold two collectors and set it on the bottom. The working collector is positioned over the oil output. The other sits next to it ready to slide into place when the full one is lifted off. One is always in position to take over by the time the other is full.
With a crane off each side, one ship can handle the two containers.
The funnel/box BP has sitting idle on the ocean floor can be used to test the idea pretty easily. It's 40 feet tall and by eye from photos it looks to average half that on a side, so 40x20x20 makes it 16,000 cubic feet. That's 120,000 gallons, big enough to prove the concept without a lot of wasted time. It is designed to sit right over the well so it can just be placed there then lifted off. (I know, it's tricky to get it in place). The gaseous methane would push the oil out anyway, and with the risk of damage it might not be worth it, but a test run is an option.
Problem: What about the fact that there are multiple leaks, 2 or 3?
Solution: Make one operation like this for each leak.
Problem: Inside each collector will be a growing gas bubble as frozen methane (hydrates) turn to gas with rising temperature on the way up to the surface. This gas will push the oil out the bottom.
Solution: Let it exit or capture it. It can exit through a pipe running from the inside top of the box, down and out of the collector (to a bladder to capture it if that is feasible). As the frozen hydrates turn to gas the gas rises to the top of the box, above the oil, and in turn is pushed out by the upward pressure of the oil below it, which wants to float on the seawater beneath it. The oil will not exit the end of the hose. The level of oil in the hose will match the level in the container and stabilize there.
Problem: What if the leak is bigger than reported?
Solution: Use larger containers and/or swap them out more often. A 300,000 gallon collector (a 35 foot cube) swapped out every 8 hours is 900,000 gallons a day. If the turnaround is too short put more collectors in the loop to create more topside time to clean them out. This is a production line problem and the production line can be ramped up.
The useful insight is that the actual volume of oil is within handleable limits. Drillers think in pipes. Buckets (earlier technology) is the key. What am I missing?