“Building Nuclear” – A Guide for Writers

by Will Davis

Reporting this week and last on the financial problems of Toshiba has variously contained the phrases “building reactors,” “building nuclear plants,” and many others. It seems that the general press is confused (and probably rightly so) when it comes to the terminology used to describe the nuclear power plant construction business. So, here I’ll provide a guide to the process and terms used in the industry and describe the various players.

Artist's concept of Palisades Nuclear Plant.  This is a nuclear power plant, or station, with a single unit, or said another way, a single reactor.  Illustration from brochure in Will Davis library.

Artist’s concept of Palisades Nuclear Plant. This is a nuclear power plant, or station, with a single unit. In some parts of the world, the term “power block” or “block” is synonymous with the word unit, meaning a combination of a nuclear reactor and its generating system. A nuclear power plant (NPP) or, said another way, a nuclear generating station may have anywhere from one to eight units. Illustration from brochure in Will Davis library.

The Players

Owner: This part of the construction triangle is obvious—it’s the utility company that is buying the nuclear power plant. This company has to provide land for the plant, pay to have it constructed, operate it, and connect the plant to its grid. Owners of nuclear plants can be single companies or groups of companies acting together either as joint owners or separate stand-alone companies with joint ownership (either of which arrangements spreads the costs). Owners generally make a decision that they need more generating capacity first, then they conduct studies to determine which way to get that added power. If the studies show that a nuclear power plant is the best option, then a process is started with other contractors. I write more about owners below.

Architect-Engineer: This is the company that designs the power plant. In most cases, once an owner has made a decision to build a nuclear plant it hires an architect-engineer (A-E) to come on board for the project. The A-E may or may not assist the utility in selecting a particular nuclear technology. Once the nuclear steam supply system (NSSS) and other particulars have been specified, the A-E designs the power plant. This will inevitably include many, many thousands of pages of drawings, and the number of considerations, specifications, and drawings has traditionally always increased. The drawings will have to match specifications provided by both the reactor vendor and the owner as to specifics for the particular project. The A-E  needs to look at, for example, would a particular job include a cooling tower, or will a river be used? The job of the A-E is later closely tied with the next entity we’ll investigate. (1)

Constructor: This is the company that builds or supervises the construction of the nuclear power plant. Typically the constructor hires many dozens of subcontractors to complete the actual jobs of pouring concrete, fitting piping, running wires, etc. The constructor builds according to the drawings supplied by the A-E and acts based upon the availability of the parts needed for installation. Tremendous troubles can be caused by a constructor building ahead of the receipt of the plans, only to find out that part of the work was done in error once the final drawings show up. This would lead to the work being ripped out and redone. Still other times, the actual work itself may be faulty and fail inspection, which would lead to removing and replacing the work. Of course, late delivery of drawings and specifications from the A-E will also lead to significant delay.

Some of the very large firms, such as Bechtel, Stone & Webster, EBASCO, United Engineers & Constructors, and some others offered not only A-E services but construction management services as well. Some owners felt that this was best to have the same company in charge of both parts of the overall project. However, at the height of nuclear plant construction in the United States, with so many units continuously under construction, ordered, and announced, it was impossible at any arbitrary moment to guarantee that a particular selected company had the resources available to tackle both jobs on any prospective contract.

In some cases, utility companies (owners) have traditionally designed and even constructed their own power plants and facilities. This tradition carried into the nuclear power plant era when owners did not hire A-E or constructor services. Tennessee Valley Authority (TVA) and Duke Power are probably the two best known examples, but Pacific Gas and Electric (PG&E) designed and built the Diablo Canyon nuclear plant itself. In other cases, the utility did part of the work itself and contracted out the rest. For example, Detroit Edison (now DTE) designed Fermi Unit 2 but contracted out the construction of the plant. In other cases, usually disastrous, the owners have attempted to build the plants without hiring a dedicated constructor, spreading out the construction contracts themselves. For the most part, though, the triad of owner/A-E/constructor holds valid.

Hope Creek, shown here, was designed as a nuclear power station with two units.  There are thus two reactors, which are inside the can-shaped containment buildings, and there are also two turbine-generators inside the long turbine building.  Palisades, seen earlier, used 'open cycle' cooling (water coming from and returning to a lake) while Hope Creek seen here uses 'closed cycle' cooling, with cooling towers.  Hope Creek was eventually completed with only one unit.  Illustration from brochure in Will Davis library.

Hope Creek, shown here, was designed as a nuclear power station with two units. There are thus two reactors, which are inside the can-shaped containment buildings, and there are also two turbine-generators inside the long turbine building. Palisades, seen earlier, used ‘open cycle’ cooling (water coming from and returning to a lake) while Hope Creek (seen here) uses ‘closed cycle’ cooling, with cooling towers. Hope Creek was eventually completed with only one unit. Illustration from brochure in Will Davis library.

Other Important Players

Reactor Vendor: These are the companies that, frankly, usually get all the press. These are the companies that actually design and construct the nuclear reactor itself, and the very closely related NSSS. The NSSS is built inside the containment building and has many interfaces with the rest of the plant. The two most important interfaces are the steam it sends out through pipes that gets used to generate electricity, and the water that comes back in to replace that which was turned into steam. Of course, there are a large number of control and indication connections to be made as well. As important as the NSSS and reactor vendor are, they are not the entities that set the overall power plant design—that is up to the A-E and the owner. Over time, there has been an increasing tendency to standardize and to duplicate complete power plants from one location to the next. This is the norm today, so that it is appropriate to talk about a particular design in terms of the reactor design itself. However, in the “old days” (certainly regarding all the nuclear plants operating in the United States today), the general speech is not that a plant is a “Westinghouse plant,” but rather a “Bechtel plant” or a “Stone & Webster plant,” because these A-E firms set the overall design of the plant into which an NSSS and many other components must be integrated. (2)

Very recently, reactor vendor/A-E teams have appeared, such as with the Westinghouse AP1000, so that the nuclear units being built at more than one location are theoretically identical (except of course for any specific owner-ordered customizable items). Further, in some countries where the industry is state-controlled or vertically integrated, there is no option as to who will design a power plant, so that all of them will more or less be the same (perhaps with incremental changes over time). (3)

Sanmen NPP in China, shown in this preconstruction artwork, is intended to be a six unit nuclear generating station.  Artwork courtesy State Nuclear Power Technical Corporation.

Sanmen NPP in China, shown in this preconstruction artwork, is intended to be a six unit nuclear generating station. Artwork courtesy State Nuclear Power Technical Corporation.

Regulator: The last significant player to mention is the regulator, or a nation’s official body charged with ensuring nuclear plant safety. Many millions of words have been written about regulators, but for the purpose of this discussion we will just say that the regulator’s inspections of the nuclear plant construction must be passed, and it has authority to halt work. Changes to regulations while a plant is already under construction can contribute significant delay if the regulator forces the changes on the plant’s completed work.

In order to be clear on the terms—especially given the recent reporting and apparent confusion—here is one final explanatory example for clarity.

OLYMPUS DIGITAL CAMERAAbove we see, from a print in my library, a rather generic illustration meant to convey the appearance of what might have been considered in 1969 or so to have been the normal appearance of a single unit nuclear power station equipped with a Babcock & Wilcox NSSS. Such a piece of art is generally the product of an A-E, which is responsible for the overall design of the plant. What we are seeing here is not a “reactor,” since we cannot see the reactor at all because it’s inside the white, roughly cylindrical containment building. This is instead a “unit,” or if you’re European a “power block.” A nuclear power plant, also known as a nuclear generating station, may have one or multiple units. They may be built connected to each other (see the Hope Creek illustration earlier) or they may be completely stand-alone and separate (see the Sanmen illustration).

If we could see the NSSS, which includes the reactor and the steam generators (since, in this case, the NSSS is a pressurized water reactor type), it would look like the model below, also from a print in the author’s collection.

BWplant1Above we see a typical Babcock & Wilcox NSSS, known as the Babcock-177 (because the nuclear reactor core has 177 fuel elements). The center item is actually the reactor vessel, and inside is the reactor. The reactor vendor, Babcock & Wilcox, was responsible for either manufacturing or outsourcing the manufacturing of all the needed equipment. It supplied drawings and specifications to the A-E for a plant, and then the equipment was sent to the site for the constructor(s) to install. This NSSS provides steam that is piped into the non-nuclear portion of the plant, where a large turbine generator converts the energy into electricity for the grid.

The terminology here seems clear enough, but is easily and quickly garbled and easily misinterpreted. When the Toshiba president and CEO said that the company was considering halting constructing in order to isolate itself from risk, what he was referring to was the A-E and constructor functions that the company itself had built up. As explained before, traditionally the reactor vendors never had this function internal to themselves. This development places the reactor vendor in a position of exposure to construction risks, and it is the company’s intent to revert to the position of supplier of systems. That is, at least, so far as the hints given by the company’s management indicate.

Hopefully this guide will prove useful to writers of all kinds.

Footnote on construction arrangements: During the period of the U.S. Atomic Energy Commission’s Power Demonstration Reactor Program, nuclear plants were built under a dual-construction arrangement wherein the AEC owned the nuclear portion of the plant (the NSSS and reactor building or containment) and the utility owned the turbine generating portion, usually abbreviated as BOP for Balance of Plant. In this arrangement, the AEC had its own A-E and constructor for the nuclear portion, and the utility had its own A-E and constructor for the non-nuclear portion (if the plant was not a “hook-on” to an existing steam plant). Later, a couple of the reactor vendors felt that jump-starting the commercial nuclear field would be helped by offering to the utilities “turnkey” nuclear power plants, which was like buying an already built new car—just get in, turn the key, and off you go. In these arrangements, the reactor vendor hired the A-E and the constructor, which worked for the reactor vendor. A flat fee was charged to the owner, who provided the land and connections for the plant’s services and output. Only a few turnkey plants were built before the conventional owner/A-E/constructor setup took over, which was used for the vast majority of nuclear plants ordered in the United States.

For much more detail on nuclear plant construction, and cost and schedule overrun, click here.

(1) The biggest and most well-known A-E firms that took part in the build of nuclear plants in the United States were, in no order, Bechtel, Stone & Webster, United Engineers & Constructors, EBASCO, Gibbs & Hill, Gilbert Associates, Burns & Roe, Fluor, and Sargent & Lundy.

(2) The biggest and most well-known US reactor vendors were/are Westinghouse, General Electric, Combustion Engineering, Babcock & Wilcox, Atomics International (Div. of North American Aviation), General Atomic (originally a division of General Dynamics, later owned by Gulf Oil), and Allis-Chalmers.

(3) Standardized, and completely duplicate nuclear plant construction did occur in the United States with the SNUPPS project; see here. A stillborn duplicate project was Duke Power’s Project 81; see here. In a number of other cases, A-Es worked off a standard set of initial drawings to design nuclear power plants for more than one owner.

ANS member Will DavisWill Davis is Communications Director and board member for the N/S Savannah Association, Inc. He is a consultant to the Global America Business Institute, a contributing author for Fuel Cycle Week, and he writes his own popular blog Atomic Power Review. Davis is also a consultant and writer for the American Nuclear Society, and serves on the ANS Communications Committee and on the Book Publishing Committee. He is a former U.S. Navy reactor operator and served on SSBN-641, USS Simon Bolivar.

10 thoughts on ““Building Nuclear” – A Guide for Writers

  1. Will Davis

    Thanks for these great questions, Jim – hope I can answer at least a few!

    First, Toshiba itself said in a 2012 press release that it was free to seek engineers on a “project by project basis” after Shaw Group exercised a put option on its 20% holding of Westinghouse stock which was immediately bought by Toshiba. So, since that time, new engineers have been acknowledged as one possibility by Toshiba. My belief is that another A-E could come in and use the Design Control Document to design a new plant and obtain a new COL using the NSSS designed for the AP1000, and create a design for a containment that has the same function. An analogy could be made to certain Navy parts; when I was in there were components that were designed to do the same job, and which would bolt into the same place but were NOT wholly identical. They just did the same job – same “goes in,” same “goes out.” So I think they CAN go with a new A-E and constructor. I think as I have said before to some that Bechtel would be an ideal choice for both jobs. I am not alone in thinking this.

    If you read the reports being made to the Georgia Public Service Commission by its own hired experts, you’ll see a whole lot of what at least they perceive to be the problems going on. These folks believe that there are serious construction management issues which are bad themselves but would also be exacerbating any component delivery issues, or any change orders brought about by a range of factors. I think this has been confirmed by Westinghouse bringing in Fluor Corporation to take over management of the construction process. There were some problems with large components, such as the main coolant pumps (these were supplied by Curtiss-Wright, who bought and still operates what had been the Westinghouse nuclear pump operations and business) but these have been corrected so far as I am aware.

    A number of your questions may be answered by a blog post I just put up at my own blog just this morning – see here. http://atomicpowerreview.blogspot.com/2017/02/new-large-light-water-construction-usa.html

    In that piece you’ll see that I agree that the only way to keep from having this same experience in the future is to keep building, and another AP1000 but contracted for a different way is probably a must. Could be Georgia, or it could be Duke Power’s William States Lee (formerly, Cherokee) site.

  2. Jim Hopf

    Adding to what I said in the last paragraph of my previous post……

    The first generation of plants were over-budget by a much greater amount than the Vogtle and Summer projects are. And yet, we never heard of companies going out of business over it, like what we’re hearing about Toshiba today. This parent company, as well the smaller companies that it bought, i.e., Westinghouse, Stone Webster, and CB&I) has took an enormous hit, to the point of being pushed to the brink, because of overruns at just two projects, the overruns being a small fraction of overall projected cost.

    Clearly, things have changed so that the reactor vendors and construction companies are bearing a very significant fraction of the risk (as opposed to the ratepayers and even the utility companies). In this context, Tom Clements’ comments about the rate payers bearing “100%” of the risk are clearly absurd. If they were bearing all of the risk, how do you explain Toshiba’s troubles!!

  3. Jim Hopf


    When you refer to a “different power plant” and being “designed” by a different firm, but it still being AP-1000, I get confused. This may show my lack of understanding. I think of the AP-1000, having gotten it’s final design approval, as being fully designed (i.e., down to the last detail). Instead you referred to it as a “concept”.

    I was thinking/hoping that they could just go with a different A/E and constructor, and perhaps that’s still possible. Your assessment is quite negative. What are your thoughts about Southern Co.’s talk about building another Ap-1000 in Georgia? The idea being that the construction team is mobilized, and now has the experience, and that it will be better/cheaper next time (if they just move on to another project building another copy of the reactor right away). Were they just blowing smoke? Or have things changed now (recently) so that they will now not pursue it? Or is there a chance that they still will?

    This is pretty depressing given that Westinghouse said that this reactor would be simpler, cheaper, and easier to build. Far fewer pumps and valves, etc.., and having main components built in a factory and shipped to the site for assembly. But now we’re hearing that it is actually more expensive, complicated, and harder to build than the previous generations of plants. The Chinese have said as much, i.e., that their 2nd generation technology was less expensive than AP-1000.

    Are the main problems occurring during component fabrication (at the factory) or assembly at the site? Also, do you think the problem is in the design itself (which make it hard to build), or is it compliance with component fab QA requirements, that would cause problems for any design.

    Based on what I have read, I had thought that it was the latter. CB&I (who has a long history of building large structures, which kill people if they fail) stated that “they have never seen or dealt with anything remotely like it before”, with respect to the nuclear fab QA requirements. They gave that as the reason for delays in component delivery, which I thought was the main reason for the delays and cost increases in the overall project.

    One could attribute these problems, with the new generation of plants, to loss of experience in making nuclear-grade components, or too requirements that are far stricter than those that were applied many decades ago, when the current fleet was built. (The old fleet was also 1/2 to 1/3 the cost. Was it just an experienced fab industry, or escalating requirements. I tend to think it’s the latter.)

    One final thing, my understanding was that these two projects are actually not nearly as over-budget, in percentage terms, as the first generation of plants. We’re talking 10-20%, as opposed to factors of several. And yet, even an overrun that is a small fraction of these projects’ overall cost is very large compared to the total value of even a large company like Toshiba. It really does bolster the “bet the company” argument against large reactors. In earlier times, I suppose they could just have the ratepayers, or the utility company bear the risk.

  4. Will Davis

    Jim, I think if the AP1000 concept is to be built in the future as a part of another (likely different) power plant it’s going to have to be one designed by a different firm and constructed by a different firm. After this debacle I don’t see anyone actually desiring the same outcome, harsh as that may sound to say. I think what there is of Shaw Stone & Webster (WECTEC, now) is finished unless it has other design and engineering work in progress that’s non-nuclear that I don’t know about. As to who will design/build in the future, I think Bechtel would be a prime choice. Back when Shaw sold its shares of Westinghouse, Toshiba announced it was free to pursue different engineer and constructor partners on a project-by-project basis, so apparently there was at least an awareness of this possibility. What was done, if anything, toward that is anyone’s guess.

    I have begun to seriously wonder whether or not AP1000 units, of any sort whatsoever, will actually get built at Moorside. I cannot see KEPCO (Korea) buying out Toshiba’s share, as one example, and then designing and building plants that use Westinghouse/Toshiba proprietary technology although I suppose that is a possibility if the contract looks right. (Heck, Stone & Webster used to build plants designed by others, as did a number of other big firms.)

  5. Will Davis

    Tom, they’re not in the story because they don’t hold any of the construction contracts — which is primarily what this story is about. It’s about the actual players involved in the construction, and the only reason a national regulator is mentioned at all is the potential for interference in work already performed that might need to be removed and replaced.

  6. Jim Hopf

    Reading the recent press has left me asking the question, “who will build the AP-1000 now?” Or does it mean that AP-1000 is dead?

    This article gives me some hope, given that having the vendor also be in the construction business is only a recent development (and that this may just mean a return to the old model). However, Westinghouse had bought CB&I and Stone Webster (as part of their entry into the construction business). What will happen to those entities now? Will they be spun off or sold, from Westinghouse? Or will CB&I and Stone Webster also not be building AP-1000 in the future?

    If (when) the UK proceeds with its AP-1000 project, who will build it? Will they be able to find another AE?

  7. Will Davis

    Thanks, Don! The SNUPPS project is covered in the linked article. Had I the space I could have talked about the Sargent & Lundy plants you mention, or the Stone & Webster plants (Connecticut Yankee, North Anna, Surry, Beaver Valley) but I have to stop SOMEWHERE or people tune out! All stuff for future posts, for certain.

  8. Barry Butterfield

    Nicely done, sir. A pet peeve of mine has always been how newspapers refer to the Yucca Mountain Repository, WIPP, and other waste management facilities as a “dump.” Billions of dollars spent on these places, and the best the reporter can come up with is “dump?”

  9. Tom Clements

    The biggest players in allowing the AP1000 projects in South Carolina to be authorized and proceed have been the state public service commissions. Where are they in the narrative? As we have seen in SC and GA, their enabling and rate-setting role has been far more important than that of the NRC. It is primarily because state laws – Construction Work in Progress – in both states unjustly forces rate payers to pay in advance for reactor construction financing costs that the two new AP1000 projects were begun and are proceeding. Those laws are a very sweet deal for the electric utilities and amazingly put 100% of the utility risks and costs on the rate payer, removing it from the companies and shareholders. In South Carolina, rate payers are already paying about 18% of the utility bill on the new reactor construction at SCE&G’s VC Summer site, while SCE&G is authorized a 10.5% ROE in spite of massive cost overruns and delays. These yearly CWIP rate hikes will continue until the units are finished – very likely beyond the claimed dates of 2019 and 2020, so we’ll soon be paying 25% of the bill just for financing. Then, much larger capital costs will hit customers hard and it’s anyone’s guess what will happen to bills. So, if not for the CWIP laws in SC & GA and compliant PSCs the projects would not be happening. — Tom Clements, SRS Watch, Columbia, SC

  10. Don Brindle

    Just a comment on standardized plants. SNUPPS only resulted in Calloway and Wolf Creek (and Sizewell B in the UK). You didn’t mention that ComEd built Byron and Braidwood, which are essentially identical 2-unit plants, except for cooling towers at Byron and a lake at Braidwood, and few minor things. Marble Hill was supposed to be the 3rd plant, but got cancelled.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>