Responding to System Demand

by Will Davis

Significant discussions have occurred recently on various internet venues about “load following”—that is, the capability of a generating source to adjust its power output to match variable demands. There is a myth spreading that nuclear power plants cannot load follow, and today’s ever-changing discussion about low-GHG generating sources demands that this myth be dispelled.

One might immediately ask this question: “Haven’t we been saying that nuclear plants are best for base load power generation?” That’s a valid question. Baseload generation can be thought of as that degree of electric demand below which you never go. When compared to other generating sources, nuclear power plants have a relatively high construction cost—but a relatively low operating cost—and thus are often referred to as baseload generating assets. Nuclear power plants make steady power and steady income for the utility at a low and controlled fuel cost that isn’t subject to rapid market fluctuations or interruptions in supply—and they do this all day and night.

However, today’s energy world is evolving. We now have under consideration small modular reactor (SMR) nuclear plants that may be ‘off the grid’ and required to supply variable loads at all times instead of  baseload power as part of a larger distribution network. Further, as high-GHG generating assets are retired, nuclear will become a larger percentage of the generating mix (all else held constant) and load following becomes part of the energy mix discussion.

From a utility perspective, operating today’s large commercial nuclear power plants at reduced load isn’t economically sensible, since the same staff  is paid the same money whether the plant is at 30-percent power or at 100 percent. Of course, the overall impact is much larger than just what you’re paying the staff,  considering all the other operating expenses—that’s just a simple example. Since renewable energy sources—which have a highly intermittent output—are now being seriously discussed, the capability of nuclear energy facilities to integrate with renewable sources, which would require load following, is important to address.

The Shippingport Atomic Power Station (seen in the lower part of this photo as a longish, red, left-to-right building in front of the much larger Beaver Valley nuclear station built years later) was the first large-scale commercial nuclear plant in the United States.  Shippingport was designed not only for load following but for remote load dispatching while operating in its normal power range (the plant was originally rated 265 MWt/60 Mwe, and ‘normal’ power was considered anything over 20 MWt).  The plant was designed to accommodate the following thermal power changes while in automatic control mode:  1. +15 MW or -12 MW at a step change rate.  2. ±15 Mw at a rate of 3 Mw/sec.  3. ±20 Mw at a rate of 0.417 MW/sec.

While today we don’t allow remote dispatching to control the power level of reactors, it’s important to know that they can accommodate power changes as well. Let’s take a look at some other nuclear plant design data for plants presently in service in terms of allowed power change rates, and then we’ll compare that to published data about today’s new-build AP1000 nuclear plant.

Westinghouse Pressurized Water Reactor: This design of nuclear plant was advertised in the 1980s as being “able to follow repetitive load changes automatically throughout the range of 15 percent to 100 percent of rated power consistent with the cyclic nature of the utility system load demand.” The Westinghouse PWR was designed at that time to accommodate step changes of 10 percent rated power and ramp changes at 5 percent per minute. Further, the plant was designed to operate, if required, on the 12-3-6-3 daily load cycle; 12 hours at 100 percent power, then three hours to reduce power followed by six hours at 50 percent power, then another three hours to ramp back up to full power. Finally, the plants were designed to accept up to a 50 percent rated power load rejection without reactor trip and full load rejection with reactor trip but optionally could be equipped with extra steam dump capacity in order to accept full load rejection with no reactor trip. The plants adjust both primary coolant boron concentration and control rod position as required to follow load.

Combustion Engineering PWR: Data are at hand for early generation C-E plants like that at Palisades; design criteria for this plant included the ability to accept step changes of 10 percent rated power, or ramp changes at 5 percent per minute.

Babcock & Wilcox PWR:  B&W large commercial plants were advertised as able to accommodate transients of 10 percent step changes, or ramp changes of 10 percent per minute between 20 percent and 90 percent rated power; above 90 percent rated power, the ramp change permissible was 5 percent per minute. Load reduction rates were the same without steam dump; with steam dump, load reductions of 40 percent in a step could be handled. According to B&W literature, “The turbine bypass system and safety valves permit a 100% load drop without turbine trip or reactor trip.”

GE Boiling Water Reactor: Data on hand for the late-generation BWR/6 shows that the design originally accommodated up to a 25 percent change in rated power automatically by recirculation flow control change, with no control rod motion, “thus providing automatic load following capability for the BWR.”

As we can see, these plants are responsive in varied degrees to changing system loads—and system loads don’t generally swing wildly unless there are storms in the area. What about new build nuclear plants?

Westinghouse advertises their AP1000 as having the following characteristics pertaining to variable system load: “The plant is designed to accept a step-load increase or decrease of 10 percent between 25 and 100 percent power without reactor trip or steam-dump system actuation, provided that the rated power level is not exceeded. Further, the AP1000 is designed to accept a 100 percent load rejection from full power to house loads without a reactor trip or operation of the pressurizer or steam generator safety valves.”

The Westinghouse SMR site offers a thorough description of that reactor design’s load following scheme which is also applied, according to the site, to the much larger AP1000 just described.  Click here for details. 

The competitive Generation mPower SMR is also designed for load following. In an interview on Atomic Power Review about the mPower SMR, Generation mPower LLC’s Matt Miles said of the mPower: “Traditionally, nuclear power plants have been used for base load generation. Our plants are designed for more segmented or off grid applications and are capable of load following to accommodate this type of deployment.”

As we can see, light water cooled and moderated nuclear power plants, whether of PWR or BWR type, and whether large commercial plants or SMR designs, are capable of adjusting power output to match variable system demand. Many years’ worth of operation on many various demand schedules have proven out the technology. While today, for many considerations, large commercial plants aren’t used as load followers, there is nothing inherent in the technology that precludes them from doing so; further, it is expected that SMR plants will normally behave as load followers. I hope this article clears up the spreading misconception about light water cooled and moderated reactor plants, in order to help level the discussion about applicability of technologies to a new age in which renewables will play a larger role on the grid.

(Sources consulted for this article include “Shippingport Pressurized Water Reactor, US AEC / Addison-Wesley Publishing, 1958; advertising material from Combustion Engineering, Inc. and Consumers Power for Palisades Nuclear Power Station; “The Westinghouse Pressurized Water Reactor Plant,” Westinghouse Electric Corporation, 1984; “Steam / Its Generation and Use,” 38th ed. Babcock & Wilcox 1975; “General Description of a Boiling Water Reactor (BWR/6)” General Electric 1978; Westinghouse AP1000 advertising materials, Korea Hydro & Nuclear Power advertising materials.)


Will Davis is a consultant to, and writer for, the American Nuclear Society. In addition to this, Davis is on the Board of Directors of PopAtomic Studios, is a contributing author for Fuel Cycle Week, and also writes his own blog Atomic Power Review. Davis is a former US Navy Reactor Operator, qualified on S8G and S5W plants.

About Will Davis

Will Davis is the Communications Director for the N/S Savannah Association, Inc. where he also serves as historian, newsletter editor and member of the board of directors. Davis has recently been engaged by the Global America Business Institute as a consultant. He is also a consultant to, and writer for, the American Nuclear Society; an active ANS member, he is serving on the ANS Communications Committee 2013–2016. In addition, he is a contributing author for Fuel Cycle Week, and writes his own popular blog Atomic Power Review. Davis is a former US Navy reactor operator, qualified on S8G and S5W plants.

9 thoughts on “Responding to System Demand

  1. NGNP Alliance

    The use of HTGR technology could be even better for load following. By balancing the industrial heat needs of regional industry with electricity demands, HTGRs have the ability to ramp electricity production rapidly in response to changing demand/supply with minimal change to reactor heat output. Thus maximizing the use of the capital investment, the O&M, and the fuel.

  2. jmdesp

    France has made intensive use of the Westinghouse Pressurized Water Reactor load following ability, also on it’s own variations of the design.

    You’re not including the EPR, that can follow load at 5% per mn between 60 and 100% load, keeping the temperature constant to preserve the service life.

  3. Will Davis Post author

    The HTGR theoretically does have that advantage, although of course light water reactors can be used for process heat (and have been.) I will be interested to see, should any further gas cooled reactors get built in the U.S., what their load following characteristics actually are.

  4. Dan Meneley

    To add to you examples of load following, Electicite de France has established a full regime for PWR load following — and they use it frequently. The evidence of performance of the EdeF systems is published.

    CANDU power plants are designed to manage daily load cycling (within the 25-40% range), load following, and for frequency control. But if you ask a plant owner if he would like to operate his plant according to any of these forms of dispatching, you might well receive an impolite “NO”. Many components do not respond well to frequent load changes, to say nothing of the fact that the owner is paid for units sent out, and for nothing else.

    Nonetheless, load dispatching appears in all nuclear futures, so we had best get used to it. And we need to properly explain what we mean by “base load” so that even six-year-olds and lawyers can understand us.

    Dan M

  5. Jerry Cuttler

    The CANDU reactor (power) regulating system (RRS) is controlled by the RRS software program that “runs” in the plant’s dual, redundant digital control computer (DCC) system. In the Normal Mode of operation, the RRS controls reactor power to follow the turbine load, over the power range from 100% to 60% full power. In the Alternate Mode of operation, the RRS controls reactor power to remain at a fixed power level, which is set by the operator. Many CANDU operators feel comfortable to operate in the Alternate Mode, where they set the reactor power manually, rather than allow the RRS to swing power automatically over a 40% full power range. The Normal Mode of operation has been qualified. (It is backed up by the two independent, poised (safety) shutdown systems that will shut the reactor down, in the very unlikely event that RRS exceeds any power limit.)

  6. David Walters

    THANK YOU for this informative essay on load following. I was a power plant operator for 25 years at PG&E and Mirant. I operated conventional gas fired thermal plants.

    I was amazed to see the actual quick load following of some of these units. I’m very impressed. I knowt here are issues with neutron “poisons” (such as xenon) that damp the fission reaction if load changes too rapidly. But it’s good to know that load following is more a function of *financing* than it is engineering. To wit:

    The author notes that nuclear plants are deployed (financed) as baseload units “expected” to run at 100% *availability* (which is why actual capacity factors often equal availability for nuclear.) The ISO (electricity dispatchers) *only* cares about availability, not name place capacity.

    So what? well, in the U.S. there is no reason that the rate based figuring that goes into the size and placement of a nuclear plant can’t be financed with the staffing and commerical paper payment in mind: exactly as large peaker units using combined cycle gas turbines are financed. These units which increasingly are making up double-digit percentages of all new power plant builds in teh US are not based on revenue-per-kWh-produced as nuclear and many large coal plants are.

    It is true that while the staff of a 540GW duel united combined cycle GT might be only around 40 and a nuclear plant of double that size is more than 10 to 20 times that size staff…these are not the major costs. The major costs are paying back the interest and principle of the loan used to build the plants. if this and all the operating costs are figured into the rate-base for a new nuclear plant, then the ISO can pay not on revenue per unit of energy produced but based on contractually obligated availability. I worked at GT that was known to sit there for *months* without ever running! A similar set of creative financing can be applied to nuclear plants as well. In fact, increasingly, ISOs *pay* for the ability to due this; they pay for units to load follow; they pay for all sorts of ancillary services (voltage control, load following, VAR stability/synchronous condensing, etc etc).

    I should add that the limited lower-range load for any Rankine cycle (steam) power plant is the steam turbine. It’s ability to absorb thermal shock of rapid load changes combined with the resistance to ‘anti-motoring’ at low loads is key here.

    The mind set that nukes have to run flat out is simply wrong. It’s small-mindedness of accountants, not the creative deployment of new ways to build the cleanest, most efficient and socially useful energy technology known.

    David Walters

  7. stock

    The new build “plants” are just on paper.

    Load following is also just on paper.

    the AP series is not proven, so don’t talk about it as if it is a real thing, it is not real until it is.

    take stock.

  8. Will Davis Post author

    Load following is NOT “on paper.” Nuclear reactor plants of many various types have operated as load followers for many years – in fact, since the first successful nuclear power plant ever built (STR, later S1W.) Even the PNPF, or Piqua Nuclear Power Facility, an organic moderated, organic cooled reactor was said to have displayed excellent load following characteristics. There is no reason whatsoever to believe that the AP1000, developed after half a century of successful reactor plant design, will have anything other than the predicted load following capability. Again – not that commercial nuclear plants will be operated continuously as load followers; the point is that the capability is there.