Reading this thread is both interesting and worrying. Even before the 2013/14 rise in incidents there was firm evidence of problems with some of the CU's on the market.
http://www.tvhsg.org/uploads/Consumer_unit_fires.pdf
In figure 2, 3 and 4 of this report it clearly shows the problem in the main neutral connection between the 2 pole main switch and the neutral terminal bar. I have always been surprised at the gauge of some of these links. Some are multi-strand and some are solid, but the CSA always looks small for a neutral line in a 100 Amp rated device. One Chint CU I have seen had a neutral link that did not look much more then 6mm CSA. Maybe 10mm as it was solid not multi-strand.
The solid core neutral link especially concerns me as it is held into the terminal bar by a single screw. The area of conducting contact in that connection is the area of the end of the screw digging into the core plus the point contact opposite the screw (the meeting of the core circumference and the hole circumference). The solid core may be flattened a bit by the screw force but I am not happy with the contact area for a 100 amp neutral line. At least a multi-strand neutral link can be squashed a bit by the screw and spread out increasing the contact area (but by the same token a multi-strand can 'settle down' and become loose without a re-tighten later).
The article also refers to the breakers that have conduction problems and create local heat spots and melt down.
So given the breaker and the neutral link is designed and supplied by the manufacturer and the installer can only trust the breaker quality and check for tightness on the neutral screw then I would say that there does need to be a review of the CU internals quality (along with installer improvement where necessary) not just incarceration in a metal box.
If we cannot trust the manufacturer to make good internals how can we be sure that they can make a non flammable metal box :-)
-------------------------
Derrick B
nanos gigantum humeris insidentes