Myth V – But we can’t rely on renewables, because of their intermittence nature

While certainly it is true that some renewables sources have problems with intermittency, notably wind or PV this isn’t a problem with all renewable sources. Certain renewable sources, such as hydro, tidal, Solar CSP or geothermal boast much higher capacity factors (as I discuss here).

Also using simple solutions such as pumped storage or hydrogen production, or simply spreading renewable systems out over a wide enough area one could largely mitigate (thought not of course eliminate) these intermittency problems.

Indeed it is something of a red herring, as Germany recently showed by exported some of its 25 GW’s of solar energy to France last winter (the irony!). Now I’m not saying it would be easy or cheap to solve these problems, indeed I go to great lengths to explain the difficulties here and here, but it’s worth remembering that nuclear power has its own form of intermittency related issues.

Nuclear power plants like to be on all the time, but the grid demand varies considerably over the day. Typically we see a small peak around morning time (people getting up & going into work, powering up PC’s etc.) a tailing off toward mid-day, then a rapid sharp spike as we roll into the evening (usually because this is when you have all three users online at the same time, domestic, industrial, commercial and there is an increased demand for lighting, heating and cooking). Power demand then tails of gradually over the evening, with the odd spike here and there (e.g. at 20:00 when Coronation Street ends and everyone puts on the kettle or opens the fridge door).

Averaged daily electricity demand for California

With electricity grids we generally break power consumption down into three categories – baseload (the static demand that’s on all the time), Peaking load ( the many short sharp peaks and troughs in demand that can occur at certain times of the day) and Intermediate load (the “ramp” of electricity demand between peaking loads and baseload).

As the graph below demonstrates, nuclear reactors are generally deployed to meet the baseload or intermediate load electricity demand, with hydroelectric and thermal plants (typically gas fired) meeting peaking power demands. It should be noted that many nuclear power stations are unable to operate for anything other than baseload capacity, although newer designs are a bit more flexible (thought even they still can’t do peaking load power). For example the AP1000 reactor designers claim that it can undergo “± 5%/minute ramp load change within 15% and 100% power“. While impressive (of course no AP1000 has yet been built to prove this capability!), that’s still insufficient to do peaking load electricity.

The load distribution profile for Canada, note that this is somewhat dated (1980’s) and the “baseload hydro” would likely be met by wind  or solar PV now. Modern nuclear plants could technically meet some of the coal based intermittent load, although not necessarily on an economic basis

The reasons for this constraining of nuclear reactors to a specific section of the grid boils down to a number of economic and technical factors. The economics are easy enough to understand – its costs about as much to operate a reactor at 20% output as at 100% output, but obviously if we vary the electrical output, this eats into the power plants capacity factor (i.e. less kWh of electricity sold) which increases the costs per kWh sold by the plant (see Myth IV). Consider that fossil fuel or hydroelectric plants that operate in such modes tend to have much lower capacity factors in the order of 60 – 45%, although in the case of peaking power plants this can be as low as 15%.

On the technical side, nuclear reactors tend not to like surprises. Sudden movement of the control rods can produce problems with control instability (which played a minor role in the TMI accident, as the operators at times had difficulty telling what the status of the reactor was as their sensors simply couldn’t keep up). Also significant and regular variations in temperature and pressure can produce problems with fatigue and creep (as I discuss in section 3.3.1 of this post on Materials Science in Nuclear reactors).

Also when a reactor is operated in a low power state there can be issues with Xenon-135 build up. In a normal operating state Xenon (a neutron absorber) is in equilibrium (new particles of Xenon are cancelled out by the neutrons generated by the reactor), but by slowing down the reactor, it can build up to dangerous levels (to give an analogy, its like a tug of war, you’re pulling as hard as you can, but then the other party lets go), leading to yet more control problems. Indeed, one of the first (of many) safety violations committed by the operators at Chernobyl was to power down the reactor to below 1000 MWt and then (as a result of sudden demand for power from the grid), power the reactor back up again (the safety protocols said they should have continued with the shutdown process).

Reactors like the AP1000 generally try to get around these problems through a combination of computer control and modifications to the steam plant. But there are obviously limits to what such modifications can achieve.

Monthly variation in Electricity consumption in the UK over the years

Furthermore, even ignoring such practical difficulties, you obviously still need enough power stations to meet the demand. i.e. if the peak demand in a particular nation is, say 100 GW’s (for maybe an hour or two in morning and afternoon), but the baseload is just 40 GW’s, you’ll need enough reactors on the grid to meet this upper bound, with 60 GW’s of you’re reactors spending the bulk of their time idle and unused. Failure to appreciate this fact will, as noted earlier, rapidly eat into the 90% capacity factor and/or thermal efficiency levels that are typically assumed for nuclear, worsen the above assumed economic figures presented previously (again, see Myth IV).

The solution is to either export the power aboard and buy it back later, which is what the French do and why they are currently the world’s largest net electricity exporter (and one of the largest importers, see IEA 2010 report here), but that’s not an option if everyone else goes nuclear too and are trying to shed load at the same time. Another option is to have a large spinning reserve of hydroelectricity or thermal power stations on hand to even out the peaks and troughs, or alternatively have some means to store the energy somehow. …..Of course the irony is that these last 2 are the very reason cited by nuclear advocates as to why renewable don’t work!….so by they’re own twisted logic surely they have conclude that we can’t rely on nuclear power either!

Thus once nuclear output goes above 40-60% of a nation’s electricity demand, or about 10-25% of total national energy use (exactly how much depends on what type of reactors we are talking about and the level of national electricity consumption) you start to get problems with load balance, i.e. the grid suddenly demands power the nuclear reactors can’t add power quickly enough to cope. When the gird demand drops, they can’t shed load quickly enough.

In a future electricity grid (without fossil fuels), we would likely meet the peaking load electricity with either stored hydrogen (created by baseload electricity) or use renewables that are more reliable (such as hydroelectric, tidal energy, geothermal or solar CSP). It should be noted on this point, ( again as I discuss in myth IV) that renewables such as Wind energy are already cheaper for baseload electricity than nuclear power, so it would make sense to use them for hydrogen production, or whatever “stored” energy options we consider.

Also its important to consider the value of “load matching”. That is matching energy demand to energy sources. This is important to any discussion of solar energy. Air-con demand tends to rise when its sunny and hot – which is exactly when solar PV systems hit peak output.

Typical profile of Californian electricity demand, overlaid with Solar energy output [Credit: Thinkprogress.com]

This point about storing energy is also important when we consider that the vast bulk of energy consumption is of things other than electricity. As I point out in this blog post, only 20% of the UK’s Final Energy Consumption is electricity, about 37% is heating and cooling and about 36% is transport fuels.

Sankey Diagram of UK energy demand

Further the demand for energy from these latter two fluctuates considerably over the year. The peak in heating demand in the UK (or other parts of Europe) for example is in mid winter, while in the US peak natural gas demand (what we ultimately use to make most of the peaking electricity at present) is in mid summer due to air-conditioning demand. Demand for gasoline world wide also peaks about mid-summer due to the US summer driving season (when Americans drive off on holiday, or we brit’s jet off to southern Spain). Such swings in energy demand can be of the order of +50% over a year.

Annual profile of average daily gas consumption each month by sector for Serbia (from Brkić 2009)

Annual demand fluctuations of oil demand in the US (based on EIA data)

Failure to appreciate the above is what has led nuclear energy supporters into making school boy errors in their assumptions as regards the potential capabilities of nuclear power. For example, the pro-nuclear author Dr MacKay proposes (after preposterously trying to argue that nuclear energy is cheaper than clean coal, btw his figures are off by at least factor of 3 to 5) a 110 GW array of nukes to meet 60% of the UK’s energy needs. Actually, if he’d done his sums properly, he’d realize that he needs a lot more than that! To account just for the aforementioned daily and seasonal fluctuations you would need to at least double, or potentially quadruple this figure (220 – 440 GW’s installed capacity just to meet the UK’s needs! Currently we have 375 GW’s installed worldwide!). Further, many of these reactors would spend most of their service lives idle, only being turned on for a few hours each day or a few months each year, which is unlikely to prove economic.

In short the reality is of course that all energy systems come with there own particular peculiarities and problems. We need a mixed energy approach and while nuclear can play a potential role in this, it isn’t necessarily an essential component. Equally on the other side of the argument, a renewables policy of wind, wind and yet more wind isn’t going to work. Nor is policy of wind and nuclear in the absence of anything else (which on balance seems to be the current government strategy). We need a bit of everything, wind, solar, wave, tidal, hydro (more pumped storage for one thing), geothermal, some fossil fuels (with CCS) and for some larger countries with poor renewable resources, maybe a bit of nuclear too.