With the information available to me at present, I can't disagree with you or anyone else who says that it looks like there might be a systemic issue with the 787. But I'd also find it a bit difficult to disagree with anyone who claims the 787 is fine. I simply don't have enough information to be certain either way right now.
Personally, I don't feel that the 787 is significantly more dangerous than other aircraft but then I do have the luxury of not having any travel planned this year on routes served by the 787. Maybe I'd feel differently if I did, and I'm certainly not going to berate anyone for not wishing to "take the risk".
Given that the most recent event is different in a number of respects from the first two, it looks to me that the three fire or "fire-like" issues so far can just about be accommodated within the normal bounds of probability. Once-in-a-million-flight-hours doesn't have to mean that there will be a million event free hours before the first event, nor that there must be a million hours between the first and second events. Another event happening this year, however, might make the statisticians think twice.
Just for discussions sake, if we hypothetically assume that there is a systemic issue behind these events, what could possible causes be? Condensation has already been mentioned. I can think of a few others but without detailed designs, construction records and investigation results, cannot hope to even guess at how likely, or otherwise, they might be:
1) Electrical Earth Bonding.
I've been reading recently in other threads how Airbus are investing a lot of time and money in ensuring that the A350 has a good earthing skeleton, both for lightning protection and for earth bonding of the electrical systems. I presume that Boeing did the same during 787 development and I also assume that Airbus will have learnt many lessons from what Boeing has done. Unlike a metal plane where the earthing system gets joined up simply by assembling the aircraft, with a CFRP frame, there will necessarily be many points at which the various metal parts of the earth bonding skeleton need to be electrically connected together. I can't help wondering, however, why Airbus didn't include some CFRP into the iron bird to test electrical bonding during their system tests.
Aircraft are very good at building up static charges, and I suspect that plastic aircraft are even better at it. If any of the earth bonding connections are missing due to bad design, or mis-assembly (perhaps loose, or assembled with dirt in the connection) then there will be a very good possibility of high voltage static charges building up in some parts of the electrical system, which could then, when something switches on or off, be suddenly discharged via very unpredictable routes through sensitive equipment (electronics, maybe batteries), causing damage which could trigger events such as we've seen.
2) Voltage spikes.
It's no secret on Anet that Boeing might have made some mistakes on the outsourcing of the design and development of the electrical system. I get the impression that Boeing designed the system at a high level and then delegated the detailed design and development of individual parts of the system to many different contractors and subcontractors. I can't help wondering whether this might have led to the development that of lots of subsystems that work well in isolation but maybe have issues when interacting with one another.
In any electrical system with things regularly switching on and off, voltage spikes will be created by the switching, even when using solid-state switches. These spikes are usually smoothed out using accumulator devices such as capacitors or magnetic coils. As accumulator devices contain a lot of metal, they are heavy, so I can see why some of the subsystem developers might have taken the approach of "we'll assume that the signals and power that we get will be smooth" to keep the weight of their own sub-assemblies down. i.e. assume it's someone else's problem to smooth out the spikes. And, as Boeing only concerned themselves with the high level design, maybe they didn't pay much attention to this until it was too late. Or maybe any spikes they found during testing were insignificant enough to not cause damage. Perhaps the spikes only get sufficiently large to cause damage when a flap motor, a toilet flusher, an oven and a coffee pot all switch off at the same time. So perhaps it was the coffee pot's fault after all.
And finally, as they say on the other side of the pond, the "curveball". I suspect that I'll probably get laughed at for this one but I don't think it's completely outside the realms of possibility, so bear with me on this:
And I'm not talking the sort of vibrations the aircraft manufacturers try to get rid of during flutter testing. I'm talking about the low level "humming" type vibrations caused by the engines or A/C power buses.
After the first battery meltdown, and seeing how the battery was constructed, I couldn't help wondering whether sudden shocks or vibrations might cause the electrodes or electrolyte to move around within the battery, and possibly cause a short. I then got wondering about the difference between aluminium and CFRP in terms of absorption of low level vibrations. My guess is that metal, being softer, will absorb more of the vibrations than CFRP which, being more rigid, probably passes on the vibrations to be absorbed by softer materials elsewhere like seat cushions, bodies or battery electrolyte, perhaps?
Would be interested to hear what other people's thoughts are.