The theory put forward by the NTSB comes across an inextricable problem. There are reasons if vehicules do not explode in the same way than in a bad Michael Bay movie (a pleonasm) : a tank can explode only if the reamining levels of fuel are low. If the tank is full or close, it can not catch fire, as stochiometric proportions are not reached, due to lack of oxygen. There are known cases of a fuel tank catching fire on a plane due to short circuits, sometimes with dramatic consequences, but not resulting in an explosion that could shatter a plane.
It is interesting that experts were unanimous to state inthe case of the crash of the Flight Metrojet 9268 on 31 october 2015, usually considered as the result of a bombing, that an explosion in the tank was impossible. But they would easily change their claims according to the needs of a government.
Your explanation that the separation of the fuselage or of a wing accounts for missile testimonies, is not consistent with the sightings of an object ascending and coming into contact with the plane, what they reported is really removed (and it would result in a sighting of something moving away from the plane). Witnesses also reported the plane cut in two, which indicates a tremendous explosion.
As for the possibility of a replacement of some of the wreckage by some parts of an other 747, the suggestion has been made by some.
Videos of cars exploding :
https://www.vroom.be/fr/blog/video-vw-gol-explose-banc-de-puissance
https://www.facebook.com/jonatdum/videos/10203337211731696/ (fire resulting from an explosion, its expansion giving a false feeling that the explosion was huge)
http://www.gamaniak.com/video-7233-voiture-explose-pompier.html (fire causing the explosion of a tank)
(note : it's not really an explosion of a moto, but the fuel spread on the road catching fire in an explosive way, then the fire reaching the engine)
OK, a few issues here. Firstly, your video citations here do not really apply.
Motorsport runs on petrol, or gasoline, which is high octane, highly flammable and racing in particular tends to use stuff that is closer to avgas than normal petrol pump stuff.
Aviation, and the 747 100 in particular, tends to run on Jet A1, which is closer to kerosene than it is to petrol and is a far less volatile substance. I have seen myself, when doing basic fire safety back in my spanner monkey days in the industry, the demonstration of a lit match being chucked into a bucket of Jet A1, and nothing happened. At room temperatures or less, as in outside on your average northern european day, it is not very volatile at all.
However, when pressurised, or in vapour form, Jet A1 is a very dangerous beast and very hard to tackle in fire extinguishing terms.
The other important point to make is that the flight was just 12 minutes out from take off, so you would expect it to be fully laden, but it wasn't.
It was heading to Rome, but was including a stop off in Paris. Airlines do not carry any fuel beyond contingency (for re-route or alternates) and so the plane would not have had a full fuel load, given the stop over and its full capacity.
It is entirely possible, indeed probable, that for utterly banal, operational reasons, the centre tank was no where near full. Also, 12 minutes out, coming into the cruise phase of flight, the aircraft would be trimmed, which means that the fuel levels of various tanks would be equalised to ensure the plane could be made as efficient as possible to cruise using the least amount of fuel they could. This often means using transfer pumps to pump fuel form one place to another, that can often mean through a tank that is less than full, such as the centre tank.
The fuel pumps have a relatively heavy power draw, and so a fairly heavy amperage power flow through a chaffed line in the centre tank, which was more than likely not full, was the likely explanation given for a catastrophic explosion of dense vapour in the tank.
Now, don't get me wrong, this is a highly unusual occurance, and there were more than likely several circumstances coming together to allow it to happen, as the 747 is such a common fleet pony and so has operated under so many parameters and influences. But it is still the most likely explanation. Given the facts, not unfounded speculation, this is still the best explanation.
As regards the eyewitness testimony, again, it has to be questioned.
How many commercial, or civil pilots have ever witnessed a missile strike? Having worked with many ex-service people who have done just that, it's not always how it looks.
In the final phase, of what is supersonic flight, the missile is nearly invisible as the rocket motor has stopped and it is flying on momentum only as it homes for the strike.
Also, the nature of the plane break up means that it could not have been a heat seeker, as there was no indication of an engine strike. It is true that the centre tank would be one of the densest sections for a radar guided missile to strike, but, there would be very distinct damage patterns as a result which were not found by the investigators.
It if was optically guided, it could have struck anywhere, but again, radar, or optically guided devices tend to be fairly high yield and larger than an Aim-9 Sidewinder, such as the Aim-9 Sparrow or the Am-120 AMRAAM. Again, very hard to hide the kind of damage they do. If you are talking BVR, then something like the AIM-54 Phoenix does massive damage.
If you are talking SAM, again, usually huge yields, and depending on the type of warhead used tend to leave very characteristic traces, such as where meshes or shrapnel are used.
So again, on balance, there just isn't enough evidence of any of the alternatives to support anything other than one of the oddest losses of a 747 in the history of the entire marque.
I love a good conspiracy theory as much as the next, but in this instance, it is just a very unusual, but nonetheless tragic, loss.