BRE,
from today's Japan Times:
"Boeing chose lithium-ion batteries for the 787, which uses five times more electricity than similar jets, because they hold more energy and can be quickly recharged, Mike Sinnett, chief 787 project engineer, said. In a worst-case scenario in which the batteries do burn, they are designed to do so in a way that doesn't threaten the aircraft, Sinnett said.
If the jet is airborne, smoke is supposed to be vented out of the compartment so that it doesn't reach the cabin, he said, and all of the battery cells can ignite without harming the plane's ability to stay aloft."
So that must be a heck of a fire shield...
To be fair, the structural damage caused by the battery fire that did occur was quite limited. The NTSB talked about it being "confined to the area immediately near the APU battery rack (within about 20 inches) in the aft electronics bay".
Perhaps not that big a shield, but only good venting, and thermally resistant materials in the area where the heat is going.
So maybe they did a proper risk analysis after all. The statement did not say that it would leave the aircraft undamaged, only that it "could stay aloft" and wasn't threatened by the fire. The venturi-suction system would have to be able to vent all smoke out of the E/E bay without it entering other parts of the craft. From what I've seen the system is supposedly able to maintain adequate equipment cooling during cruise, but what about a full-blown Li-Ion battery fire during descent?
The venturi system as such is fine, as it requires no moving parts to operate, however it does require moving parts (valves closing and opening, fans switching off), to get it to start working.
I wonder if they did flight-tests, or at least full-size wind-tunnel tests with actual fires.
Lyman,
Sinnett admits the likelihood they will burn, and protects the company from liability via "designed to not threaten the aircraft."
I see that rather as a positive statement, indicating that they did a risk analysis, and considered the likelihood and severity of a battery fire. Doing a proper risk management means that once a hazard has been identified and the associated risk (as far as is possible) quantified, you take measures to either reduce its likelihood, or its severity, possibly by mitigating the effects on people and equipment. If it is well engineered and the regulator agrees that can be acceptable.
I haven't seen the engineering details (and probably never will) so I cannot say that it wasn't adequately done; it seems plausible. The amount of energy that can be set free can be known, the temperature, heat, propagation and duration of a worst-case fire can be estimated, and mitigating technology can be incorporated. I think it depends in large parts on the reliability of the E/E bay ventilaton system entering "override" mode, i. e. passive overboard-venturi suction.
Bernd