Tue Jul 6th, 2010 at 02:25:45 PM EST
Recently on ET, there was discussion on whether nuclear power plants can be operated in a load-following way (peak load), rather than at continuous maximum output (baseload) -- something that would be necessary both in the case of working alongside intermittent renewables (also see Wind power faulted for low prices!), and in the case nuclear energy were to provide most or all of our electricity.
In February, in Load-following and intermittency, I presented a study into the question by IGE/IER from Germany, which answered it with a tentative yes. However, being based on German experience, it was largely hypothetical, with many question marks regarding the practical feasibility of load-following operation. Something describing the experience in the country known to have researched nuclear load-following most, France, was needed.
Here I review such a study, even if its information on operation in France is still rather limited. The study itself concludes in broad terms that load-following operation is technically possible, but won't be widely realised for economic reasons.
The study in question is: Pouret, L. and Nuttall, W.J. (2007) "Can nuclear power be flexible?" Electricity Policy Research Group Working Papers, No.07/10. Cambridge: University of Cambridge. A draft version is available on-line. Hat tip again goes to Jérôme.
Like the EGA/IER study, this one "answers" other studies assuming an inflexibility of nuclear power, too; in this instance, in Britain: they quote the UK government's 2006 Energy Review saying that nuclear power "has the disadvantage that it cannot easily follow peaks and troughs in energy demand".
Technicalities of load-following operation
The most potent way to regulate nuclear power plant output highlighted in the German study was to use the control rods in pressurized water reactors (PWR). Supported by some digging, I questioned whether that would be so easy maintenance-wise: in particular, control rod manipulation leads to uneven power distribution in fuel rods, resulting in stresses. I and rootless2 also questioned whether fatigue from the wear & tear on the control rods themselves in continuous operation can be dismissed.
Well, the Pouret-Nuttall study confirms these concerns and more:
It is important to note that while most reactor types have control rods for reactor shut-down control, these rods are not optimised for controlling flexible reactor power levels... The exclusive use of control rods for output power control would have negative consequences, such as: flux distribution disturbances (see figure below), component materials fatigue, mechanical wear, and adverse impacts on fuel burn up. Many of these difficulties arise for the fact that, for instance, output electrical frequency control involves very many low-amplitude rod movements (up to several hundreds a day), which may limit the lifetime of control rod mechanisms.  [References are mostly to Framatome publications]
They go on to describe the factors of temperature inhomogeneities and xenon-135 creation (which enhances said inhomogeneities) in more detail, and mention the influence of such design features as power density and fuel enrichment.
However, there are ways to mitigate these problems, by retrofitting the reactor:
...conventional rods are known as `black rods' indicating their complete absorptive capacity for stopping the passage of fission neutrons. Flexible reactor operations are facilitated through the use of special reactor control rods known as `grey rods'. These rods do not completely absorb the fission neutrons that try to pass through them. For load-following manoeuvres, a clever management of both control rods and soluble boron is found to be optimal .
Soluble boron is a neutron absorbing material, which in addition can deal with xenon. From later discussion, it appears that boron has primacy in France at present, and "grey rods" have a full roll-out only with the EPR. For the EPR in mode G (grey rods), they claim the possibility of power output regulation at a speed of up to 2% of total output per minute, between 30% and 100% of total power, and they indicate specifically that such an operation is possible on a daily basis (the assumption that quasi-permanent load-following operation of existing plants won't have adverse maintenance consequences was one of my criticisms of the German study).
A further factor is reactor temperature.
Unlike the case of Gas Cooled Reactors, the relatively low coolant temperature range in PWRs (see table below) limits a plant's thermodynamic efficiency, its thermal stresses, and the fatigue of components.
Again looking into the future, they claim that the EPR has the same temperature in the 60 to 100% range. In their conclusion,
Given these significant improvements, one can therefore state that new build PWRs will offer operational flexibility as good as that of current fossil fuel plants .
From which it follows that even the flexibility of nuclear power as practised in France is inferior to fossil fuels. Sadly the study contains no example data or load curves, nor a consideration of exports. They only state in the economy section that
Today, nuclear accounts for more than 80% of French electricity and, therefore, most NPPs have to often operate occasionally at part-load and some plants must be sufficiently flexible to load-follow to ensure grid stability.
If stepped reductions and shutdowns are more commonly used (as elsewhere), it follows that in present practice, the contribution of load-following operation to variability is rather limited.
Now, the technical aspect is one thing. The other is: will it pay for the owner of the plant? After all, when advocates of nuclear power say that it is cheap, they are thinking of current widespread practice of operation at maximum.
As nuclear power load-following remains a rather uncommon practice, exclusive to a handful of countries, economic information is larely unavailable. Most data on nuclear power plant economics assumes baseload operations.
The analysis starts with the cost structure of different plant types from a French government study:
After a disclaimer about great variability due to uncertain factors and country differences, they state:
Despite such local differences, the cost structure of nuclear power always contains more fixed costs (especially capital costs) than fossil-fuel-based alternatives. This is the essential reason why baseload operation is generally preferred for nuclear power plants. We can safely, albeit perhaps somewhat simplistically, assume from a generating company's perspective, that, given the cost structure of nuclear power, operators would want their NPPs to operate at full-load for as much of the time as possible, in order to maximise income.
With operating hours as the primary factor, the French government study calculated these production price curves (which Pouret-Nuttall include with the disclaimer that this diagram assumes only full shutdowns for nuclear, not load-following):
The discussion so far assumes constant retail prices. However, on our brave new electricity markets, balancing is a separate spot market that can have much higher prices. Still, Pouret-Nuttall say, regarding the UK example:
Prices in the balancing market can ... be very generous, but they are insufficiently high to motivate flexible nuclear generation given its very high fixed costs discussed earlier. As the market evolves and new nuclear power plants come on-line it is not inconceivable that this situation could change and nuclear power might find a role in the balancing market...
Then again, if things aren't left to market forces and nuclear gains a high share due to government policy, load-following becomes a technical necessity. Unfortunately, Pouret-Nuttall can give few numbers or conclusions on the economics of this.
Thus far the discussion considered only the main effect from fixed costs and reduced operating hours. Pouret-Nuttall also mention some of the secondary effects of extra operational costs from load-following:
..secondary effects can arise; for instance, load-following operation may imply an increased use of soluble poison to control the reactor power. In such modes of operation much more water must then be treated and discharged, which might imply extra operating costs at the margin.
There is also the maintenance effect:
In France, NPPs have relatively low availability coefficient (about 80%) . A recent study by EDF (Electricité de France) shows that operating NPPs at their maximum load improves their overall performance. It especially reduces the unscheduled outage coefficient from 3% to 1.8% in four years. This clearly suggests that load-following reduces the availability of NPPs, mainly because of more frequent maintenance (see below). The cost of such a difference of the unscheduled outage coefficient has been estimated at several millions euros .
They also state that "no study has yet been undertaken by the French authorities to estimate" the full range of extra maintenance and lifecycle costs. Will load-following accelerate plant aging? Pouret-Nuttall say that this would be sensible to assume, but French authorities claim there is no clear evidence today.
Even if they concede that a very small number of pieces of equipment (control rods drives for instance) may be adversely affected, they still argue that proper designs and load-following procedures ensure the core components are not excessively degraded. Also, there is the possibility that EDF might concentrate load-following operations on just a few NPPs, in an almost sacrificial manner so as to avoid damaging the wider NPP fleet .
It is in the economy section that the authors tell about the relevance of exports, but only to state that:
Another method by which NPPs might operate close to the top of the merit order in a given country is to shed surplus nuclear electricity via sales to neighbouring electricity systems. France is a major exporter of nuclear-generated electricity and it would be interesting to study the relationship between electricity exports and NPP operations. Such considerations lie beyond the scope of this paper.