What I remember from one of the IET lectures given by HS2 Ltd representatives is that they say that the total capacity increase from double deck is simply not worth it given the costs involved.
And yet the only growing-passenger-count railway to try high speed double decks [SNCF] is so delighted with them that they have not ordered a single single deck TGV trailer since the first Duplex trailers arrived.
This says something.
They were abandoned in Japan because of negative passenger growth and a desire to increase top speeds on the routes they served to generate additional passenger growth rather than any real operational issues.
A single deck design could use the entire width and length of the train for passengers, putting every technical component in the unusable space above and below the passenger area. A double-deck design would still need to put these somewhere, and as the entire height of the train would be needed for the two passenger decks, it would mean displacing other passenger carrying areas. When you add in the space needed to make two passenger levels work, with stairs and more level changes (the lower deck on a TGV Duplex is below the entry door, requiring an additional step inside which then can't be used to carry passengers) the total capacity change really isn't that great.
It is however significant - but only if you get to longer train lengths.
The 'TGV Grande Capacité' studies and the V-150 seem to indicate you could get enough installed power and tractive effort into a 400m TGV Duplex formation with only 2 power heads, at which point your train capacity ends up ridiculously enormous and the benefits of MUs fall away.
If this is truly a high capacity transport system that is what you will need.
Much will depend on what this TGV Duplex successor looks like - remember we are still a decade away from orders even being placed for trains and much can happen in that time.
SNCF wants an AGV Duplex and it will get one eventually.
If double decks meant that it wouldn't be possible to reach as high a speed as quickly as a single deck train would that would add additional problems as well.
The timetable assumptions were written on the basis of the standard train performance used for the TSIs - which the TGV Duplex and its later derivatives is easily capable of obtaining.
Acceleration won't be a major issue - after all when you get to the installed power ratings we are talking about TGVs don't far so badly compared to multiple unit trains. TGV Power cars don't contain much dead weight because traction equipment for these power levels really does weigh that much.
A TGV Duplex has power to 8 of its 26 axles after all - which isn't that bad in terms of local hauled operations. The TGV Atlantique managed 12 powered axles out of 30 - which is getting towards MU power levels.
The end bogies adjacent to the power cars are quite lightly loaded (without any traction gear) and could be tapped off the locomotive traction package quite easily - so 12 axles is easily obtained without breaking axle load limits.
I don't think there will be any real difference between the classic compatible and captive train designs. The two will very likely be two variations of the same design so that there would be maintenance commonality and if both can take 550 passengers, they'll most likely have the same interior design as well. I wouldn't be surprised in the least if the two of them were so indistinguishable to most people that it would be possible to hot-swap in a classic-compatible for a captive train without the seat reservations changing whatsoever.
In that case - why even bother with a captive design at all?
It is highly unlikely that the marginal cost of more classic compatible sets will be significantly greater than ordering a seperate fleet of captive ones - and it would be a major boon to operational flexibility if every set could go everywhere.
