The topic of nuclear power is one that tends to produce black or white opinions. It’s either the electricity source whose name should never be spoken or the source without which a reliable, zero-carbon power system can never be built. I’m a bit wary of writing about it!
However I think it’s always useful to add some grey to these black vs white discussions, even if you have a strong opinion. So I’ll highlight some of the issues I see as our power systems transition, and how they impact the case for or against nuclear.
Key will be the role of nuclear within a future flexible power system. That’s because greater flexibility is unquestionably the direction of travel; required to integrate more variable renewable sources like wind and solar. The latter goal seems more desirable than ever, given these are cheap (or cheapest) sources of electricity now – and will only get cheaper in future. (Though more on cost later).
For many, nuclear simply doesn’t warrant any discussion, because of radiation fears or the abject failure to tackle waste disposal. They’re important issues – potentially deal-breaking ones – but ones I’m not going to consider here. That’s not because I don’t think them relevant, but because a question to ask first is: “even if safety risks are deemed acceptable, what if nuclear simply doesn’t make economic or operational sense?”
Characteristics of Nuclear
What arguments do people make to support nuclear as a good addition to future power systems?
- It doesn’t use a fossil carbon fuel, so it doesn’t emit CO2 (or other nasty gases) into the air during operation – good news from a climate change and air quality perspective. This is probably one main driver behind the fact that over 60 new plants are under construction with hundreds more “planned” or “proposed”, especially in Asia.
- It’s the most energy dense fuel source we have – that means we get lots of energy from a small power plant footprint.
- Fuel and operating costs are low (per MWh generated) and plant lifetimes are very long. So, once capital costs are paid off, electricity is cheap.
- It is designed to provide constant power (i.e. very high capacity factor), available regardless of the vagaries of the weather. Whether this does really support its case is discussed further below.
- Because of its conventional thermal/spinning generation characteristics, it’s considered to have low “integration costs” (at least into our existing system).
Low carbon, energy dense, weather-proof and compatible with our current grid – what’s not to like? Well, here are two key things:
- It’s very expensive to build, because £/MW costs are high and plants are big (often ‘000s MW). Estimated and quoted costs vary widely, but have in some cases gone up rather than down over time; some commentators talk of a “negative learning curve“, though not all. Before these large capital costs have been recouped, nuclear electricity (£/MWh) can prove more expensive than alternatives. Even if not, a project still requires a big chunk of up-front investment to be found and risked on a single project – at a time when energy investors have many other places to put their money for quicker returns on more varied (risk-spread) opportunities.
- It takes a long time to build (in some cases much longer than was hoped!) so isn’t a quick or adaptable solution to market changes.
The cost structure of nuclear is particularly significant when thinking about the context of system flexibility.
It is technically feasible for nuclear plants to adjust their output to meet changing demand. Modern designs can do so rapidly and any can do so over longer timescales. But the economics of nuclear depend on running constantly, at high capacity factor, for investors to recover that massive and already-paid build cost. Any levelised cost of energy (LCOE) calculation you see for new-build nuclear will assume this (i.e. a capacity factor of 90% or so). Lower the capacity factor assumption and watch the LCOE head higher.
Given a starting assumption of constant running, behaving “flexibly” within a power system likely means reducing output sometimes (as load-following gas generators do). But generating less means needing to charge more when you do generate, in order to recover your capital (= higher electricity prices) or waiting a lot longer to recover that investment (= a shrinking investor pool or = higher financing costs, which also = higher electricity prices).
In other words, from a flexibility point of view, nuclear has something in common with solar and wind – it’s not very flexible!
- Solar and wind are inflexible because they generate when the weather dictates, which may not match changes in demand. Also, like nuclear, their economics are dominated by capital recovery rather than operating costs. When the weather isfavourable and they can generate, they don’t want to be curtailed (or exposed to negative prices) if demand is low: they’d make no money but still have to keep up with their debt repayments.
- Nuclear is inflexible because, although it may be technically adjustable, its ability to raise finance or generate energy at its “advertised” cost evaporates if it must adjust (lower) its output on a regular basis. It can’t afford to be flexible, it needs to generate constantly.
Which brings us onto another highly polarised discussion…
Baseload vs Base-cost
Much like “nuclear, yes or no”, the question of baseload divides opinion: do we need it or not?
On one hand are those who argue that there is always a minimum level of demand on the system, so it makes sense to have plants capable of running steadily and predictably, all day, all night, to deliver this (or some reasonably-sized portion of it). Nuclear obviously fits nicely into this idea, especially if we favour a low-carbon system and/or have limited space to cover in solar or wind farms (so value its energy density). Then other sources of supply can be stacked on top. This top part of the stack is the bit that needs to be flexible to meet variations in demand.
Fig 1: A system with “baseload” (this and the next chart come from a previous article and accompanying video of mine, which you can find here)
On the other hand are those (in particular proponents of solar and wind) who argue it makes sense to use the cheapest sources of electricity whenever they are available, even if these are variable. Given their falling costs, these cheapest sources will be, if not already, solar and wind. Then stack flexible solutions on top to make up the difference. Nuclear doesn’t fit well into such a system, being both inflexible and (it is assumed) expensive.
Fig. 2: A system without baseload (as above, more here).
Although there’s often a focus on “filling up” the supply stack to add up to a required level of demand, it’s important to note that solar and wind are relatively low capacity factor and thus require lots of capacity (MW) per unit of delivered energy (MWh). That means any system where a high percentage of energy is from these sources will certainly need to deal with situations of over-generation too, to avoid large-scale curtailment (and its impact on the economics of solar and wind supply).
So where possible we need flexibility solutions capable of creating additional demandas well as supply – for example storage, which can soak up otherwise curtailed energy when there’s “too much wind” and release it when there’s not enough.
In practice, many current transitioning power systems are a hybrid of the two system concepts above. There are plants that still operate on a steady, “baseload” basis, while policies ensure that solar and wind are taken with priority whenever available, so they are bankable investments. Then a variety of flexibility in the system ensures demand is matched: gas, hydro, biomass, interconnectors, demand response (up or down), storage, curtailment or a combination of some or all of them.
To transition to less “baseload” and more “base-cost” means building enough capacity of these flexibility solutions – particularly those that can act as both supply and demand – to cover both the removal of the “steady” (but potentially inflexible) supply and to cope with the increased risk of oversupply or curtailment.
At this point it’s worth noting that creating more flexibility doesn’t just help us to integrate inflexible wind and solar into the system; it is helpful to inflexible nuclear too! When demand is low, rather than turn down a nuclear plant and wreck its economics, we could keep generating and put the energy into storage until demand goes up; just as we want to do to avoid curtailing excess wind. (Or export it through an interconnector, or encourage demand turn-up). So if nuclear advocates want it to be part of a low-carbon system along with “variable” renewables, they should be calling for system flexibility solutions just as loudly as wind and solar supporters are.
The challenge here is not really whether we have “baseload” or not, but how we transition from systems based on supplies which were flexible but fossil-carbon-emitting, to supplies which are zero-carbon but inflexible (hydropower is excused here). To do so is not just a change in the supply mix, but a change in the whole system mix: supply, storage, connectivity and demand.
System-cost, not Base-cost
This means we have to avoid thinking about “base-costs” simply as levelised costs of generation or energy prices awarded for individual projects at auctions, headline-grabbing though these may be.
Decisions around cost need to incorporate not just the cost of individual new sources of supply, but also:
- The cost of building and financing new infrastructure to allow this supply to be available when needed (storage, interconnectors, grid upgrades etc.),
- The possible impact a source of supply or flexibility might have on other parts of the system (“integration costs”),
- Changes to the way electricity markets operate to allocate these costs (or benefits).
The impacts on a system will depend on where it is starting from, on the current and future energy resources available to a system, and on current and projected patterns of demand. Costs in one part may be offset by savings elsewhere and will change over time. It becomes a very complicated set of interconnected variables!
In other words, every system will look different and a proper analysis of costs needs detailed, whole-system analysis, incorporating all the factors above. There are an increasing number of studies which recognise and do this, as excellently summarised recently in this recent UKERC report.
The overall message from these studies is that the costs of integrating variable renewables are low when these are a minority of supply (say less than 30%); and lowest where higher flexibility already exists in the rest of the system.
In such low-penetration cases (most markets at present), a simple “base-cost” analysis should have short-term relevance. If we add cheap solar and wind, even with additional system integration costs they may remain cheaper than those of alternatives such as nuclear. This depends, of course, on the cost gap and short-term integration costs in the particular market being considered, so it’s impossible to generalise!
However where the cost gap is large, there is certainly backing for the view that nuclear is just too expensive and inflexible to be a good choice when planning a future power system. “Base-cost” may well beat “base-load”.
However, system transition is a long-term process – and, further ahead at higher penetrations of variable renewables, modelled system integration costs rise. That’s not unexpected. However the spread of results becomes wider too – i.e. the system cost impacts become much less certain (the models vary more).
By way of example, below is one chart taken from that UKERC report, looking specifically at modelled “reserves cost” impacts; the different points indicating the results of different studies. (The penetration of variable renewables is along the x-axis and the costs of providing reserves up the y-axis).
Fig 3.: source: UKERC’s “The Costs and Impacts of Intermittency“
So the long-term reality is less clear-cut, particular in view of the uncertainty around these integration costs as we try to absorb more variable generation.
Opinions vs Policies
If you put on the spot for my own opinion on power systems, flexibility and nuclear, I would summarise as follows:
- Especially as their costs fall, we can build an awful lot of supply, storage or other sources of flexibility in the time and/or with the money spent on new nuclear (>$25bn and 10 years in the case of Hinkley Point C in the UK: a power plant that will likely enter a very different market to the one that existed when its construction contract was signed!)
- I believe there is big untapped potential in demand response, bringing significant flexibility to demand as well as supply. I’m also keen to follow how power-to-fuel, EV smart charging and power-to-thermal technologies evolve, all of which provide potentially large sources of adjustable demand.
- I predict rapid progress towards smarter, more connected power systems; for example the creation of “virtual power plants” from small, distributed generation and storage resources, optimised by clever software. Not only can they be more flexible than systems based around fewer, larger power plants, they will be more resilient too (think internet vs mainframe).
On that basis, my expectation is that as these and other flexibility solutions develop (and become cheaper), so the case for large, expensive and inflexible nuclear power plants will diminish – and that both market experiences and models will show this. All that is needed is time, will, technology and ingenuity. I also believe the time required is going to be less than many expect, with change driven at the revolutionary pace of IT and software, rather than the more sedate incremental evolution that the energy sector has been used to.
However remember that a more flexible system is better able to integrate inflexible nuclear too! Full-system models may start to show nuclear does makes long-term sense as we build more wind and solar, even if those sources look cheaper on an individual power plant level. Distributed nuclear, in the shape of small modular reactors (SMRs) could be game-changing, if developed as some governments like here in the UK are keen to encourage (and if accompanied by learning curves).
Furthermore, my expectations aren’t necessarily a sound basis for policy-making!
If I lived in a different location, I may see different challenges and priorities (or different nuclear costs and timescales: compare the UK with the UAE for example). I may be too optimistic on how quickly or extensively our power systems can transform. And policy-makers wherever they are have a balance to achieve, between what is known now and what might be true in future. While creating the environment for innovation to thrive and compete with incumbents, they can’t take for granted its future outcomes (or – crucially – when these outcomes may happen).
Faced with future uncertainty, especially over integration costs and capabilities, some policymakers may opt for nuclear as a known quantity, even if it currently seems expensive. Uncertain of which of two often polarised sets of opinions on nuclear to believe, they may decide the rational strategy is to hedge their bets. They may view it as a useful way of replacing existing “big ticket” generators in a way which requires few other system changes and which helps meets emissions targets (for example if it helps them replace coal without using gas). They too may see a future like the one I envisage, but think it is too far away to solve their immediate supply requirements reliably, especially if their demand needs are growing fast.
Nuclear went through its “golden age” of capacity growth from about 1970 to 1985, followed by a long dearth of activity until recently.
Only time will tell whether current projects represent just another brief window of activity, one which will close again once more flexible power systems evolve. Or whether they represent a sustainable re-emergence of nuclear power as an integral part of reliable, cost-optimised, low-carbon future power systems.
What’s your opinion of nuclear, in your specific market? Should it play a role in the future electricity mix or is it an expensive and hazardous distraction from cheaper, cleaner alternatives?
I’d be fascinated to hear opinions on either side!