Remember that all forms of generation require backup. Nuclear power plants are typically off-line about 10% of the time for planned refuelings, maintenance, and unplanned outages. Ideally the planned outages occur during low load periods, but forced outages can occur any time. Here in New England, the operating reserve requirement of 1,200 MW is set at that level to guard against the potential outage of Seabrook Nuclear Generating Station (and the plant's output is capped at 1,200 MW, even when conditions would otherwise permit some incremental generation). These very large, single-shaft generators create their own reliability issues, which is why some are pushing for small-scale reactors that can be distributed more broadly on the grid.
I'm really not against nuclear, btw, but it's not the perfect solution. (Neither is wave energy, but waves are much more stable as an energy source than wind or solar.) Unfortunately, with the price of natural gas so low here in the US, the impetus for nuclear development is basically dead here. The owner of a prime site for a new nuke, at Calvert Cliffs, Maryland, told the Energy Bar Association last week that a new nuke there would be $2bn NPV cash negative by their current forecast. And that's counting all the current subsidies available to nuclear here in the US. Needless to say, they're not planning on building the new Calvert Cliffs plant.
Most water cooled nuclear plants do require periodic shutdowns for fuel shuffling and replacement of old fuel with new ones. But those activities are scheduled, taking place in periods of lower grid demand.
Canadian CANDU reactors have robotic systems that perform those operations online. Fluid fuel reactors are fueled and have their fission products removed completely online (IFR and LFTR reactors).
Still reactors do require some periodic maintenance that requires them to be shutdown (even CANDUs) but it's like less than 2 weeks every 3 years.
But all of that is old school nuclear, the stuff that was developed for nuclear submarines and improved upon very little. Canada only designed heavy water CANDU because the USA refused to give them all the tech.
The problem isn't nuclear fission technology in general it's high pressure reactors (water or gas cooled). Your typical water cooled reactor operates under 150 atmospheres of pressure, it's a huge pressure cooker.
There are much better options, namely the LFTR (Liquid Fluoride Thorium Reactor) that operates under ambient pressure, is refueled and has its fuel reprocessed for fission products, all online. And is designed to be installed in smaller reactor modules (typical water cooled reactors at 1000MWe or more, while LFTRs are planned to use 250MWe modules), so to have a 1000MWe plant there would be 4 smaller reactors, meaning that only one would be shutdown for maintenance at a time (which should be even less frequent than every 3 years).
I disagree with most of the anti-nuclear rethoric, the one factor anti-nuclear guys got right is nuclear is like GPS. We don't have GPS because the US government wanted a civilian positioning system, they wanted a military system, and decided it was good military public relations to make it available for civilian usage. 99% of nuclear research was for military usage, the remaining 1% for civilian usage was all cancelled (exactly the Thorium LFTR research). Even IFR reactor research was driven by it's capacity to turn lots of U-238 into plutonium.
Unplanned shutdowns can still happen, but they are very rare, mostly due to natural disasters (actual or predicted).
Nuclear produces power when we want it to produce. Solar and Wind produces when it has sunlight and wind, regardless of civilization needs. Coal and Natural gas plants also have maintenance shutdowns.
It's a pity we don't have a grown up conversation about nuclear. Pandora's Promise is a movie everybody needs to watch, even if you hate it and don't change your mind one bit. It has flipped millions of people from anti nuclear / on the fence to pro nuclear.
- - - Updated - - -
Nuclear is a poor choice to integrate wind and solar. These units are not designed to ramp up and down rapidly, as is needed to balance intermittent renewables. And because they have a huge fixed cost and relatively trivial marginal generation cost, it's uneconomic to ramp them below max capacity.
Hydro (pondage or pumped storage) is the best way to integrate renewables, followed by natural gas or petroleum units. Unfortunately we don't have good enough inter-area control systems to allow, say, Quebecois dams to provide inter-hour balancing for New England wind turbines. Fortunately, the western US is working on a regional balancing energy system, which should markedly help renewables integration there.
Germany's plan to go green is a one trillion euro plan. That's 16 billion EURO per GWe. New nuclear plants are being built anywhere from 1 billion USD per GWe to 4 billion USD per GWe (outside the crazy NRC nuclear over regulation). Nuclear isn't cheap, but it can be cheaper even than coal, the real problem is an egg-chicken problem. The anti nuclear lobby (the greens plus big oil) caused the NRC to stop being only a safety organization and made into an anti-nuclear monkey wrench. The Anti Nuclear folks viciously attack the Vogtle-GA reactor costs, and ignore the running count of construction interruptions due to the political decisions and NRC stupidity. How much do you think it costs of keep paying the workers to do nothing ?
First we need to clear up the safety concerns, cost is easy to solve it we start building hundreds of reactors per year, otherwise without scale and with the NRC current anti-nuclear attitude, there's just no point in building more nuclear power.
- - - Updated - - -
Actually, even if you consider the current worst case prediction for Vogtle, it's still one third per GWe than the German plan !
- - - Updated - - -
The biggest barrier to building new nukes is the uncertainty about regulation -- what standards will be applied for design and construction? where can the waste go? how many times will my application be held up and reevaluated? DOE tried valiantly to streamline the regulations, but it's not clear that you can pin down something that opponents want to make a constantly moving target.
(This is from my perspective as someone whose team is working/has worked on regulatory approvals for Vogtle, Indian Point, Calvert Cliffs, and South Texas.)
DOE tried valiantly to streamline the regulations ?
Even if one builds a new reactor that have another 10 exactly the same in full operation, you still need to repeat most of the process that applied to the first one.
I'm a pilot, and I'm appalled by the FAA certification requirements on new aircraft designs, until I found out that's small potatoes compared to new nuclear certification.
The real problem is those folks habits are rooted in US Navy nuclear stuff, which you all know, nothing in the military is done efficiently, specially when it comes to paperwork.
The NRC refuses to fully certify a nuclear reactor design, and then only inspect that the work was done within the specs.
It's like very new reactor installation is like a one of a kind reactor is all respects. And they will bill the nuclear operator at US$ 300/hr for their tens of thousands of hours of studies and inspections.
- - - Updated - - -
Uh, no. That would use orders of magnitude more energy than could ever be extracted from the fuel. Never mind that it would be unsafe.
I'm all for solar power, but it's not clear that it can ever amount to a sizable fraction of generating capacity. Certainly without grid storage it cannot.
Standard light water cooled nuclear reactors only use 0,7% of total nuclear power in the nuclear fuel (and that's not counting even for the U-238 that was removed during enrichment).
So the old school nuclear industry proposes the IFR reactors (GE S-PRISM). But there are a few things we need to know about IFR reactors.
First they require at least 10 times the nuclear material inventory inside the reactor. Lots of that is cheap fertile U-238, but at startup it requires many times the fissile U-235/Plutonium/Americium/Curium (than a light water reactor). GE literature says if we startup S-PRISM only with LWR SNF (light water reactor spent nuclear fuel) then all SNF is the world is only enough to startup 20 or 30 reactors. Starting up a PRISM with new nuclear fuel requires way more uranium than a LWR.
Second, IFR reactors are cooled by molten metallic sodium. Sodium burns in contact with air, and explodes in contact with water. Of course this doesn't happen in the course of normal operations, but it makes a rupture in the reactor a significant safety concern. I agree with GE that the anti nuclear pundits overblow this, but still it's a very undersirable characteristic.
Finally, IFR reactors are still low temperature reactors. Which means they are only about 33% efficient in thermal to electricity conversion (with the remaining 66% of the heat going into a large body of water), and it's low temperature makes it unsuitable for providing industrial process heat (that burns quite a bit of natural gas otherwise).
There's just one proposed solution that combines IFR 99% fuel burnup, that is high temperature, low pressure, uses a coolant that is as stable as possible (a salt), and uses nuclear fuel that is 4 times more common than U-238 (2000 times more common than U-235), and is essentially free. Thorium Molten Salt Reactors, specifically the LFTR reactor (Liquid Fluoride Thorium Reactor). Finally, a LFTR reactor can use as little as 700Kg of fissile material (actually much less than a similarly powerful LWR), so there less nuclear stuff inside the reactor.
Why is the fuel essentially free ? Like IFR, there is no enrichment process and solid fuel fabrication costs (LWR solid fuel costs 3x the price of uranium ore today). IFR takes in Uranium/Plutonium dioxide directly. LFTR takes Thorium Fluoride directly. Finally, Thorium is currently a very undesirable byproduct of rare earth mining, it's literally preventing rare earth mining due to insane EPA regulation on Thorium handling. Obama blames the Chinese on monopolizing rare earth production, when the reality is we stopped mining it here due to the Thorium insanity. Once we have a use for the Thorium, mining can restart, storing the Thorium becomes a problem of the LFTR operators (they will gladly take Thorium Phosphate for free and deal with it).