Sunday, 13 September 2009

Can nuclear power plants be expected to load follow?

In the spring / summer of 2003 FirstEnergy, a utility in the US State of Ohio, was supposed to trim trees away from high voltage transmission lines, but failed to do so. In the afternoon of 14-August, demand (or load) began to rise sharply on the electrical grid. The high voltage wires became heated and sagged into the untrimmed trees, initiating a sequence of events that resulted in the shutdown of more than 100 power plants, loss of electrical service for 55 million customers in the US and Canada, and spread so quickly that it took an international investigative panel 6 months to issue a findings report. (before / after satellite photos)

The delicate balance of generation and load on an electricity grid continuously fluctuates, often significantly and on varying timescales as industrial and household demands ebb and flow throughout a day, a week, or a season. Generation must adapt where and when required. Morning and evening demand swings generally occur over a few hours; but there are also significant plant trips to deal with (well over 1000 MWe in a single instance). Without getting into a discussion of ‘operating’ and ‘spinning’ reserve, it is evident that utilities have developed plans to account for a number of potential scenarios - the case above notwithstanding.

Utility operators can vary the electrical output from some power plants quickly to adjust total generation to total demand. Hydro power is an example. In generation mode the water falls through a turbine generating electricity. In nearly all cases, power can be quickly changed by reducing or increasing the amount of water passing through the turbine. Some hydro stations can work in reverse; taking power from the grid to pump the water back up into a reservoir. Demand can be added by increasing the pump speed or total number of pumps in operation. Such facilities give operators the flexibility to manipulate either side of the load / generation balance. This mode of operation is referred to as load following.

A load following generator’s principal attribute is responsiveness. With respect to nuclear power plants (NPPs), responsiveness of currently available light water reactors (LWRs) is challenged by neutron poisons – in particular the isotope xenon-135 (xenon). Xenon is a powerful thermal neutron absorber (poison) and will capture neutrons otherwise available for fission of the reactor fuel. It is produced directly and indirectly from fission in all reactors.

Xenon production and removal in thermal reactors has been well understood for decades [1]. However, nonlinearities related to the xenon equilibrium equation challenge the control of power swings required to support a load following mode of operation. Xenon transients have the negative impact of significant reactivity addition or removal over the time periods required by many load following scenarios (i.e. periods of several hours). The operational challenge of an in-progress xenon transient is further exacerbated by increasing or decreasing reactor power as the terms of the xenon equilibrium equation are each impacted by neutron flux (reactor power level) to varying degree.

Xenon transients

The neutron flux or power level of a reactor determines the production rate of xenon, iodine and tellurium (xenon precursors) as well as the xenon burn up. Xenon decay and tellurium / iodine decay into xenon are purely time dependent but are constrained by different half lives. Xenon concentration will reach equilibrium after a period of steady state operation or shutdown; in the later case following an initial spike in concentration due to the decay of the remaining iodine and zeroing of the xenon burn up term following the shutdown. Xenon equilibrium is not directly proportional to reactor power level. For example, the equilibrium concentration at 25% power is more than half the equilibrium concentration at 100%.

Reactivity is the parameter used to measure and control reactor power changes. As a simple analogy, reactivity to a core is like heat to a kettle full of water. Assuming the core is already critical, adding reactivity increases power (neutron flux) just as adding heat to a kettle increases the water’s temperature inside.

Withdrawing control rods increases core reactivity. As a poison, xenon absorbs neutrons and therefore reduces core reactivity with increasing concentration. Xenon transients challenge reactor operation due to continuously changing reactivity addition or withdrawal depending on the nature of the power history and attempted manoeuvre. For example, consider a reactor start-up about one day after a reactor trip from full power where xenon concentration had been at equilibrium. At the time of start-up, xenon concentration would have already peaked from the decay of iodine in the fuel at the time of the trip. The concentration would be decreasing steadily (adding positive reactivity to the core). This is not a safety concern since the control and safety rods add more than enough negative reactivity to maintain the reactor in a safe shutdown condition.

[In some cases high post-trip xenon concentrations add enough negative reactivity to prevent operators from commencing reactor start-up until adequate xenon decay time has passed.]

Post trip Xenon-135 transient [1]

As the reactor start-up progresses, the remaining xenon continues to decay but the concentration reduction is accelerated by the increasing reactor power’s impact on the xenon burn-up term of the equilibrium equation. As xenon concentration is reduced, positive reactivity is added to the core (equally accelerated). Operators must closely monitor core reactivity and take any required mitigating action to ensure the reactor is not shutdown automatically by reactor protection systems designed to limit the rate of power increase. Adding to this challenge, many reactor designs and license commitments require discrete hold points during power manoeuvres to calibrate instruments, perform reactor physics checks, synchronise the generator to the grid and perform other tests or surveillances.

Alternatively, if power is quickly reduced from 100 to 50% the resulting xenon spike, due to reduced burn up but continued iodine decay, will add negative reactivity for several hours and then reverse, adding positive reactivity as a new equilibrium is approached. As with the start-up example above, close operator monitoring and adjustment are required to ensure plant control remains within acceptable parameters.

Xenon-135 transients from power changes between 50 - 100% [1]

Returning to load following, the time periods, frequency of adjustment and response time required are in direct conflict with the nature of xenon transients at NPPs. For this reason, most NPP operators choose not to subject their facilities to load following operating modes.

NPP economics

Other economic realities further dissuade NPP operators from subjecting their facilities to load following manoeuvres. The principal financial outlay for an NPP’s lifecycle costs is the initial capital expenditure of construction. NPP operation is typically not sensitive to fuel price. Contrary to hydro plants that are able to store sometimes scarce water for peak periods or fossil stations subject to high fuel costs, NPPs do not benefit from fuel cost savings by reducing power. NPP revenue is typically directly linked to generation. Therefore the economic case for NPPs is strongest as a base-load facility, operated at 100% power.

The designs of core fuel loading for operating cycles are planned well in advance and based on assumed fuel burn-up over several fuel cycles (typically 1 to 2 years per cycle for PWRs and BWRs with any given fuel assembly remaining in the core for 3 or more cycles). Load following operating modes would add another layer of complication and financial risk to this planning.

The reactor suppliers

The economic interests of reactor design and supply organisations favour a marketable, load following design. If a load following NPP can be made available, additional nuclear generation share can be justified for a given electrical distribution grid. However, currently available designs continue to be constrained by xenon transients and the economic business case for nuclear power.

Modern designs have incorporated technology improvements to mitigate many operational challenges, such as the reduction or elimination of the hold points described above, ability to complete anticipated maintenance tasks at full power and core designs with strong neutronic coupling [2]. These improvements simplify operations during power reductions for unplanned maintenance activities and are highly desirable regardless of an operator’s willingness to load follow.

Potential for advanced and fast reactors

Xenon is a poison for thermal (slow) neutrons only. Therefore as fast reactor systems are deployed in the coming decades, the operational challenge from xenon will no longer be relevant.

Economic and other fuel cycle challenges are currently being assessed for fast reactor design concepts [3].

The World Nuclear Association is optimistic in this area. [4][5] The WNA's Ian Hore-Lacy contributed to the Encyclopedia of the Earth. His submission explains in some detail how advanced reactor designs will better accommodate load following.

In practice

Despite the challenges identified above, operators in France - with its high nuclear share of electricity generation - do elect to load follow [6].


[1] United States Department of Energy, DOE Fundamentals Handbook: Nuclear Physics And Reactor Theory Volume 2 of 2, website:, accessed 12-September 2009.

[2] AREVA, The Path of Greatest Certainty: EPR a Generation III+ Power Plant, website:, accessed 12 September 2009.

[3] OECD/NEA, Nuclear Fuel Cycle Transition Scenario Studies, ISBN 978-92-64-99068-5, Paris, France [2009].

[4] World Nuclear Association, website:, accessed 13-September 2009.

[5] World Nuclear Association, website:, accessed 13-September 2009.

[6] World Nuclear Association, website:, accessed 13-September 2009.

Posts by others

I was able to find a few posts (one, two) by Rod Adams (Atomic Insights) that mention the topic. Also thanks to Edwad Kee below for this post by Joe Romm at The Energy Collective / Climate Progress.


  1. For molten salt reactors, xenon isn't a problem either, since it is continually removed from the reactor as it is produced.

  2. The short answer is that conventional light water reators can and do follow load in France and in other places where nuclear has a large share of the total generating capacity.

    In other regions where nuclear is a smaller part of the total capacity mix, the plants are operated as baseload units.

    The following is a repeat of a discussion in the Linked-In group "Nuclear Power - The next generation" from a little over a week ago:

    Are nuclear power plants dispatchable?
    Joe Romm's Energy Collective post of 31 Aug 2009 quotes (I think) from Craig Severance:

    "The flaw in the nuclear path, beyond its tremendous cost, long lead times, and imported fuel, is that nuclear is not actually “dispatchable” power. Nuclear plants are designed to run all the time at fairly steady output — meaning nuclear power cannot provide the “peaking power” now provided by gas turbines. Thus, a nuclear path would still rely heavily on fossil fuel power plants to “ramp up” on a daily basis to provide the power needed during these daily swings."

    This is just wrong.

    There is a very big difference between an intermittent resource (i.e., wind and solar) that do not and cannot produce energy under some conditions (i.e., no wind or no sun) and a nuclear power plant that is preferentially operated at full output due to economics.

    Light water reactors are technically capable of following load.

    The economics of a nuclear power plant (i.e., low variable cost and high capital cost) mean that this rarely happens in the US, where nuclear power is relatively small part of the total energy/capacity mix. I will save a discussion of merit order dispatch (i.e., why a low-variable-cost nuclear unit runs all the time) for another time.

    In Europe (France and parts of Germany), nuclear power is a large part of the total capacity, so that nuclear units are often operated in load following mode.

    The typical LWR is capable of operating flexibly between 30% and 100% capacity, changing load at 1% to 3% per minute (with load changes as high as 5% to 10% per minute for limited periods). A UK paper from October 2007 "Can Nuclear Power be Flexible" by Pouret and Nuttal provides a detailed discussion of this issue, concluding that:

    "We confirm that modern Generation III and III+ are technically capable of flexible operation. To explain why nuclear power is almost exclusively used as baseload generation, we look at power market economics. As a result, we conclude that despite some technical abilities, nuclear power plants are preferentially used for baseload generation for economic reasons and will continue to be used in this way for the foreseeable future."

  3. Hi,

    I just found your blog and like the discussions here, even if they are a bit too complex for my educational background.

    Being French, the electricity powering my computer is 78 percent nuclear. Hydro accounts for 10 percent and thermal and other renewable solutions account for the remaining.

    I guess hydro plays the role of load follow. Our country shows that we can get a lot of our electricity by nuclear, and I hope Australia will give its chance to this fantastic low carbon energy source ! :)

  4. Edward Kee –

    According to the IAEA PRIS database, the 5 countries with the greatest share of nuclear electricity production are France, Lithuania, Slovakia, Belgium and Ukraine. Of these only the WNA country brief for France mentions the use of NPPs for load following. The Ukraine brief does state the following:

    “In connection with the South Ukraine nuclear power plant, the South Ukraine Power Complex also consists of the 11.5 MWe Olexandrivka Hydro Power Plant on the river Pivdenny Buh, generating annually over 25 million kWh; and the 2 x 150 MWe Tashlyk Hydro Pumped Storage Power Plant commissioned in 2006-07, with total annual production of 175 million kWh. The hydro units of the South Ukraine Power Complex belong to the country's nuclear utility Energoatom, and they serve as an important regulation of the peak capacity for load-following.”

    Countries with the greatest number of operating power reactors (other than those already listed above) include the USA, Japan, the Russian Federation, Korea, the UK, Canada and Germany. The country brief for Russia mentions a desire for advanced fuel to support future load following modes of operation. I also quickly checked the country briefs for India, Sweden and Switzerland, but found nothing.

    I am not aware of any evidence to support Joe Romm’s comment “NPPs can not be dispatched”. However, and in particular in the American context from which Romm writes, there is plenty of evidence that NPP operators choose not to enter into load following modes of operation.

  5. If a nuclear power plant was built so as to use heat for either of electricity production and / or water desalination then presumably the heat could be readily switched between these tasks depending on electricity demand. Given that there is little problem in storing fresh water such a plant could readily operate in a dispatch mode even if there are difficulties on the reactor side of things.

    Of course a stand alone desalination plant powered by electricity (instead of direct thermal input) can also be used as a demand leveling option for any power source (including wind, solar etc). Other loads are obviously managable in this way also.

    The advantage of integrating desalination and nuclear relates to efficiency as heat can be tapped directly.

    See also:-

  6. TerjeP, you correctly point out that two sides of the equation do, in fact, exist. Why indeed couldn't load be made to match the power being generated. This approach is used in the case of pumped hydro - not sure about desalination.

    I don't have any experience with desalination plant operations: how long to startup, the system engineering / reliability impact of frequent shutdowns or production swings, etc.

    It seems like an idea worthy of pursuit.

  7. The combination of normal electricity supply with overnight charging of electric vehicles is another good match, and may increase the base-load contribution to overall capacity by about 35%, even if the kWh increase is only about 15%. See graphs in WNA Electricity & Cars paper. The load -following possibilities then increase if smart meters are used for demand management on that charging.

  8. The impact of a xenon oscillation is categorized according to a stability index (or in other terms, how quickly with the reactor converge to an equilibrium state). This oscillation is mitigated first by not allowing load following capability to occur past a certain point in life - near the end of a cycle. This is because the xenon plays much more of a negative reactivity contributor when it is not challenged by the presence of soluble boron. Now, if a oscillation does occur, it can be controlled by the use of control rods. From a safety and operations standpoint, load following in a nuclear reactor is a possibility in the US (read the ERPO guidelines "Extended Reduced Power Operation") and is a requirement for plants in the EU. However, economics is a different beast. Whenever you introduce an increase in negative reactivity, you are reducing the neutron economy. An increase in load following over the entire cycle length will result in an energy shortfall.