I was just at a seminar by Mike Dunne, director of LIFE at NIF, so I should be able to answer some of the questions here. Particularly interesting is that NIF and LIFE (it stands for the rather uninformative Laser Inertial Fusion Energy, but it's actually the program to design and develop a working fusion power plant based on NIF) are way more advanced than I thought. NIF is actually aiming to reach ignition by September this year, although a couple of delays mean 2012 might be a bit of a struggle. Early 2013 is virtually certain, however. And LIFE has pretty much a full design and cost analysis done, based almost entirely on existing technology and existing manufacturing base, with funding assumed to be entirely private sector and unsubsidised (although I think the headline figure does include some subsidy that applies to anything nuclear).
At which point they undergo a thermonuclear explosion. That explosion, and all its byproducts, has to be confined and controlled.
How long are the walls of the chamber going to last after being repeatedly exposed to the various high energy particles and gamma rays that get released?
The current goal is up to 4 years. No material is capable of lasting very long in such a high radiation environment, so the reactor chamber is instead designed as a modular system which can easily be removed and replaced, and can therefore be constructed just out of steel. The chamber will be filled with low pressure (around 0.03 atmospheres) gas (mainly lead, but I think I remember seeing noble gas mentioned as well, unfortunately I don't have the slides to refer to). This absorbs essentially all the ions and most of the x-rays, leaving the chamber wall relatively unscathed. The gas will have to be cycled to remove the radioactives produced , but this can be done as a slow, constant process rather than having to vent and replace it all at once.
Neutrons pass straight through everything and are absorbed by the coolant sleeve (Again modular and replaceable) surrounding the chamber. The coolant will be liquid lithium, which not only works similarly to the liquid sodium in existing plants, it reacts with neutrons to produce tritium and so can be used to produce fuel, essentially making this a breeder reactor.
How long is the energy transfer mechanism going to last?
As above - the reactor vessel will have a lifetime of a few years, while the coolant will be constantly cycled to produce fuel by removing the radioactive part. Everything past that is just the standard heat transfer part of any power station.
What happens when the explosion is a bit asymmetric due to a defect in the pellet
This is one of the main areas of current research. Since they haven't actually achieved ignition yet, no-one's been able to investigate how sensitive it is to various parameters. The assumption to start with is that everything has to be engineered to the same level as at NIF, but the hope is that some leeway will be possible which would greatly reduce the cost of manufacturing the fuel.
As for what actually happens to a pellet that was dropped wrong or failed to get hit by the lasers or whatever, I don't know. Presumably it would just sit at the bottom of the vessel. It would actually affect subsequent fusion shots, but I don't know if it would cause problems due to funny heating on the vessel and need immediate removal or if it could just be left there.
or a problem with one of the lasers?
How often do you have to shut the machine down and fix it?
The lasers are modular and hot-swappable, using the same basic system that has been used before in previous facilities (although with different specific components obviously). I can't remember the name of the place, but >99% availability with MTBF over 1500 hours has been managed before, and if anything LIFE should be better.
How much does that cost, in joules and in dollars?
For an individual intervention, I have no idea. For overall construction and final electricity costs, it's estimated to be ~$3500/kW capacity and $45/kWh to the consumer. So approximately equal to coal, more expensive than gas, cheaper than nuclear fission, and much, much cheaper than photovoltaic. Obviously this is based on various estimated, and the assumption that NIF will actually mange ignition, but the design and estimates are all done in partnership with a whole pile of major private companies involved in energy production, distribution and engineering, so this is as realistic as possible and not just some pi in the sky numbers from scientists for their pet project.
These are "just" engineering problems, but the problems with tokamaks are just engineering as well.
The difference is that the engineering problems here are mostly solved or have solutions not too far off (one of the few main exceptions being mass production of the actual cryogenic fuel pellet). Even if ITER works perfectly, which hopefully it will, no-one really has a clue how to actually turn it into a commercially viable power plant. NIF already has pretty much the whole thing planned out, and a few bits even partially tested. It's the difference between needing to work out how to implement some of the details on an existing concept and having to come up with the entire thing from scratch.
Look at it this way. ITER is planning to first plasma in 2019, with actual ignition well after that, and DEMO to have actual designs between 2017-2024. NIF is planning to reach ignition this year, and have the first commercial power plants by the early 2020s. Assuming the actual fusion part works (possibly a big assumption, but looking very hopeful) ICF is at least a couple of decades ahead of tokamaks in terms of actually getting fusion into commercial use. Tokamaks are 5 years away from planning to have the basic idea for a design. ICF is already negotiating prices with suppliers for the parts.
They are very close to the proposed theoretical limit of ignition 1.4 megajoules (within 86%).
Actually, they're already operating regularly at the design level of 1.8MJ, have demonstrated that 2MJ is possible, and are looking at increasing it to 2.2MJ and further into the future (2017-2018) potentially as high as 3MJ.
