I still wonder if the buoyancy is controlled by pumping water in and out of a sealed pontoon or by air pressure to force water out of an top sealed only pontoon. If water is pumped, it still seems unlikely to me that the fire could heat and melt pipe all the way down to the sea level to allow water in.
OTOH, if air pressure and pipes melt up top, I see it would be sinking very quickly. Of course that would seem like a very shortsighted design that an air leak could send the rig to the bottom. Surely that would not be the case.
I don’t see how a huge roaring fire with no controls in the amount of heat produced, tens of thousands of barrels of crude oil spilling on the deck every day was going to end differently if no fire fighting was done.
This may help give some perspective on the melting of the steel.
http://www.jwco.com/technical-litterature/p09.htm
The pressure-feed fire of a blowout will totally destroy the surrounding steel structure in minutes. Derricks have fallen-in less than 30 minutes after blowout ignition. The core temperature of a low-GOR 28 deg. F API crude oil blowout in Kuwait was measured at 1,677 deg. C (3,051 deg. F). And a radiant heat temperature of 510 deg. C (950 deg. F) was measured at ground level, 15 m (49 ft) from the base of this large vertical fire, which was estimated at 30,000 bopd. Oil well firefighters commonly see surrounding sand and stones melted and fused on large fires. Steel loses most of its strength at 500 deg. C (932 deg. F) and melts at 1,500 deg. C (2,732 deg. F).
...
Water.
The most important method to limit fire damage to structures from a blowout fire is application of water.