I would bet that boiling seawater won't let the hull reach red heat. I guess I could do some sums for heat transfer once it was below 100 C, but the conductivity of coke will be a lot lower than steel, and there won't be much radiation reaching the steel, or much convection in the horizontal direction. So the temperature gradient across the hull plates might be 100 C? Might look at the sums later. Different story for a bulkhead, if fire is next to that of course.