Who said modular construction would save money on first of a kind units?
On July 27, 2015, the Wall St. Journal published an article written by Rebecca Smith titled Prefab Nuclear Plants Prove Just as Expensive. That piece has been widely shared and discussed on social media with more than 2700 shares on Facebook and more than 150 shares on Twitter as counted from the original article. I suspect that there have been several times those numbers of additional shares and retweets, based on my own limited sample on both sites.
Though the headline and article content are welcomed by those who oppose nuclear energy and disappoint many who favor its deployment, no one should be surprised. Unfortunately, many people from both points of view seem to believe that it’s another example of the nuclear industry failing to live up to its promises. Here is a quote from the article:
Building nuclear reactors out of factory-produced modules was supposed to make their construction swifter and cheaper, leading to a new boom in nuclear energy.
But two U.S. sites where nuclear reactors are under construction have been hit with costly delays that have shaken faith in the new construction method and created problems concerning who will bear the added expense.
My headline question is a serious one; I’d really like to find the marketing and sales people who oversold the benefits of of using manufactured, prefabricated modules over on-site construction. It would also be nice to locate examples of literature or presentations demonstrating excessive promises.
I’m not denying that there are substantial long-term advantages, but it is well-known among practitioners of modularity and series manufacturing that any cost and schedule related advantages of the choice would only appear in “nth of a kind” units.
No matter which modular industry one studies, there is always a cost and schedule challenge associated with creating a supply chain, tooling up for manufacturing, shaking out quality-related issues, refining interfaces between both components and suppliers, training distributed work forces, and establishing the logistics required to put the right components and modules in the right place at the right time.
Aside: I not a modularization expert, but I was once tasked with drafting initial versions of procedures required for an effective program of using prefabrication and modular construction for a nuclear reactor project. My participation in that task spread out over dozens of meetings during a several year period, yet the program was still in the early start-up phase when I made a career change. I had the opportunity to learn from some experts in the field; it is an enormously complex task, made even more difficult in an industry with a regulator and suppliers that have little experience in the processes. End aside.
The rate at which the various issues are solved depend on the industry and its relationship with regulators, customers, partners and suppliers. The more cooperative the environment and the better the interests of the various parties are aligned, the quicker the start-up issues can be solved.
The required improvement process works best when all parties recognize that prompt solutions will result in increased future sales and profits; incentives to compromise in disputes disappear when major order cancellations increase the risk that investments and other actions to solve problems will produce results that are too late to have positive effects.
The situation at the Vogtle and VC Summer projects is not ideal. There are almost unbelievable sums of money at stake, additional customers that could benefit by problem solutions are putting their orders on hold, and suppliers are justifiably nervous about making substantial investments in manufacturing tools due to the paucity of firm follow-on orders. The initial module supplier plan had to be essentially torn up and recreated with new subcontractors. The process is complicated by the fact the AP-1000 modules are some of the largest ever attempted by any industry.
Even with all of those challenges, the people working on the project are making good progress as demonstrated by the improving quality and schedule performance on the second units at each site. The utility customers, though disappointed with the performance and perhaps justifiably upset that they are being asked to pay some of the costs of being early adopters, are benefiting from low interest rates on borrowed money and low commodity prices.
Tom Fanning — the CEO of Southern Company, the parent of Georgia Power, plant Vogtle’s lead customer — has consistently explained that the final cost of the project for customers is likely to end up being less than expected when the project was initially budgeted.
My prescription, based on several years of running a manufacturing enterprise along with a deep immersion in the history of nuclear power plant construction programs is to press forward. Keep focused on solving the problems to enable improved performance on the next projects.
I also hope that marketers and salespeople in the nuclear industry learn that this is one enterprise where it is better to “underpromise and overdeliver” than to make rosy predictions that cannot be fulfilled. Word of mouth among disappointed customers can be fatal for an already embattled industry.
I disagree with people like John Rowe, who state that nuclear energy is just a business and not a cause. If it is to succeed, it needs to be pursued with as much passion and dedication as any other innovation that has the power to save the world.
Utility customers should put down the journalists’s interpretations of the nuclear revival story and dig deeper to recognize that they cannot make multi-decade investment decisions based on short term prices of a volatile commodity like natural gas. They should take time to imagine what the natural gas price trajectory will be if the gas industry succeeds in killing off both coal and nuclear as competitive suppliers in the electrical power market.
All of us need a viable nuclear option; one necessary ingredient to enable that option is for utility customers to demonstrate a greater commitment to purchasing new plants. Vendors and their suppliers must also commit to making the investments required to shake out inevitable start-up issues.
“it is an enormously complex task, made even more difficult in an industry with a regulator and suppliers that have little experience in the processes.” I worked at Marble Hill near the end of its’ construction. One change that needs to be made is in the regulatory arena. Marble Hill was part of a group of Westinghouse plants that were all to go on line with identical USARs, Technical Specifications, and other aspects to “save” training and operating costs. That required that every change that was made at the earliest plant in the construction “pipeline” to be made to each of the others. Additionally, as latter plants had different people working in construction and testing they found different problems. These “fixes” had to be applied to the earlier plants in many cases. No one in plant construction or initial plant testing and licensing saw any savings in this process. The big problem was that each plant had a separate license, required a separate license amendment, and a separate review by a separate NRC Plant Manager. Back then the NRC was not that concerned about BOP i.e. Balance of Plant (or steam cycle and other non nuclear systems) but that has changed since then. The early stages of a “modular” reactor system is going to require a sort of living license approval, at least until all of the kinks get worked out.
There are going to be design problems. The design is going from paper, to models, to the real physical world. Yes, most of the design engineers are experts in their field. Many have provided presentations for the various engineering organizations and helped write industry standards. BUT, many also NEVER get out in the field and see or touch what they are building, or work with their design in its environment. Often during startup testing, after having extreme difficulty getting a system to function properly or meet the design specifications or requirements, the design engineer was brought out to the plant to determine the problem. On more than one occasion the engineer has told me that this is the first time that I have seen my [system] operating, etc. One even said “This is the first time I have been in a reactor building” and that is the only place his system is used.
Moral – Expect problems and revise regulations to prevent bottlenecks.
It was my understanding that a good portion, perhaps most, of the cost escalations at Vogtle and Summer have been the result of design modifications imposed by changes in the regulatory environment, specifically containment integrity in certain low-probability events. There was also the rebar issue with the one plant, I can’t recall which. Ratcheting regulatory requirements and failures in the QA/management process at the supplier/construction phase are old stories in the business and haven’t gone away with the introductory of both modular construction and the COL process. Until they do, there will be no nuclear revival, and we’re likely to see another 20-30 year gap in reactor deployment.
Wayne SW, you hit the nail on the head with the QA problem. Rich’s discussion about Marble Hill regulatory problems is accurate and contributory to its demise. But the truth is it was abandoned because they were pouring tons of bad concrete. And the same folks responsible for that were hiding the extent of condition from the utility. When the truth about extent of condition came out, the utility said we can’t afford to continue, and walked. The utility even won a law suit, but were limited to recovery by contract terms. Same story with Zimmer, 98% done and no weld QA program. NRC said do it all over, utility walked on the nuke and hooked up a boiler. Midland plants were sinking in to the ground, abandoned. You can’t blame any of that on NRC. In my opinion, it was caused by an explosive construction condition that diluted the QA savvy work force, both labor and management. Today’s QA problems, doesn’t matter how we got there, is lack of both experience and an appreciation for the significance of the problem. Rod is right, it is a fixable problem that takes time to gain experience and that can’t happen with out continued work. But everyone involved better really believe failed QA is a show stopper.
“Moral – Expect problems and revise regulations to prevent bottlenecks.”
Nope…the US Navy has brought several FOAK nuke design projects on-line without “revising regulations that caused bottleneck problems.” Technical problems require technical solutions by the technical experts, i.e. the designers and builders. In the Navy the designers/approvers were pretty much the same group of Naval Reactors experts. They have the power to solve technical problems as they arise. Then the Navy guaranteed quality of the design by verification things were built to Design, Code and Standard.
It works… what does regulation by outside folks who have to be explained every change along the way have to do with it? The current process does not work very well because it is a flawed concept. Unfortunately it is how the process is set up. But maybe “then a miracle occurs and…”
@MJD, My thought was that a modular design should have a Modular license or something like that. When all of “version 1” modules are going to end up identical, but over the time span of the first version construction, there has to be a method of Licensing the “module” and simple revision of the applicable TechSpecs/USAR and get rid of per plant reviews, reviewers, etc. That is change regulations so that the NRC would License each Version of the NSSS module and not the plant.
MJD, As a specific example the Rancho Seco NSSS was virtually identical to TMI-II and received its operating license about a year before fuel load at TMI-II. However, one (and only one) of the NRC reviewers decided that the calculations for ALL NSSS setpoints did not properly incorporate the instrument inaccuracy. After three meeting with B&W, GPU, and the NRC, the NRC issued an ultimatum – Re-calculate all setpoints verifying that they adequately incorporate inaccuracies or revise the existing setpoints to provide a margin for the inaccuracy. With fuel on the site guess what GPU (ME & and the site B&W engineer) did? We determined the new setpoints that provided for the setpoint that in essence actually incorporated the margin twice. Among other things this increased the Low pressure trip 100# and lowered the High pressure trip 100#. Thus the operating band was 200# narrower than Rancho Seco and even Davis Besse. How would you like to have operate with that operating window? How would your near TMI event have unfolded? And this happened because each plant, even though IDENTICAL had different reviewers, (some who did not know their [blank] from a hole in the ground) and GPU was smelling the green stuff and folded. Identical plants should be identical and have the same reviewer. When a change impacting the design of all plants is accepted at one plant it should be applicable and “approved” for all. Identical problems and their fix should not have separate reviewers and separate NRC plant managers making decisions as this will lead to subtle differences and who knows, another TMI-II. And each plant gets delayed by these separate reviews and Pays the NRC again for a review that was already done before.
The Zimmer debacle was a terrible tragedy and a real black eye for the industry that in many ways we haven’t recovered from. Anybody with an ounce of sense will know that if you don’t have a program to document your work as being up to standards, the regulators won’t accept it. The “take-my-word-for-it” strategy isn’t going to cut it. I’m sure someone got canned for it. Maybe they should have been shot or committed seppuku. In any case the people of the region, while they initially celebrated the demise of the nuclear Zimmer, now have to pay the public health costs of breathing the coal plant effluents.
In the airplane world we live with the same issues day in and day out.
In theory, every 777 is identical to every other one . In practice, not so much. The reality is that every airplane is delivered with hundreds of major and minor non-conformances. Each of these requires a specific fix to bring it in line with the certification standard. But that doesn’t mean that the certification standard changes every time someone in the factory makes a mistake. That’s insanity.
Naturally, it takes time and experience with a particular product to fine tune the details and hone the production process. To have a different team do the build and implement the fixes is another form of insanity. How can anyone learn anything if everything is different all the time.
But I guess I’m preaching to the choir….
I found a 2009 Power Magazine article regarding modular construction. The 3rd paragraph states that the reuse of plans, procedures, design, fabrication, project management, will allow subsequent units to be built at lower cost (not the first units). I think Henry Ford would have verified this.
So thanks to Rod for calling out the WSJ. The anti nuclear people such as NIRS now have another misleading slogan to go with Too Cheap to Meter and All Radiation is dangerous.
Also, as I searched around in places like CBI, Fluor, etc. I found that modular construction of large components such as drilling rigs, LNG plants, and so on is alive and well. It is a valid idea. I wonder if they are using it in China?
John, not to mention ships.
If you had small (200 MW) modular reactors, even the first site would be likely to experience most of the benefits of modular construction since you’d have 12-14 (hopefully) identical sets of reactors and get to the good part of the learning curve after the first one or two reactors.
The WSJ also claims one problem is that the AP-1000’s large modules take much longer to manufacture. That would be another advantage of small modular reactors.
It may be true that we need to build several more copies (of the reactors) before the benefits of modular construction begin to manifest themselves. The problem is, it’s looking more and more like the nuclear industry (in the developed world anyway) will not be given the chance to build several more copies, the lack of cost reduction for the initial units being one of the main reasons.
As a few others have said, this is an argument for SMRs. Not only is the route to a large “N” quicker, but they do not involve significant amounts of onsite assembly (and additional fabrication), that is subject to the full rigors of NQA-1 and stringent NRC oversight. The NRC, basemat rebar issue is just one example of how full, safety-related on-site construction activities can be the source of significant trouble, those being activities that do not benefit from modularity or factory fabrication.
Not only do I think that we need to move to SMRs to get most of the benefit of modularity, but I think the only thing that has a real chance of happening is for a large volume SMR assembly line to be built in China. In addition to small SMR capacity, China has a large domestic market for nuclear generation. Those two factors would combine to yield the potential for a large construction volume (i.e., getting to a large value of “N” quickly).
More generally, China is able to build things at a lower cost, low-cost labor being but one reason. It is China that has managed to build solar panels at low cost, which is much of the reason for solar’s recent success. This is true of many, if not most products. China has become the world’s “factory floor”. For most products (I-phones, etc.), fabrication of physical hardware is done in Asia. Why should nuclear be any different? In fact, given that nuclear’s main problem is its high capital costs (i.e., the cost of its physical hardware), it would seem that nuclear needs Chinese, low-cost fabrication more than anyone. It is entirely practical to do this for SMRs, which can be shipped whole (the NSSS).
The US can then just buy modules from China, at a reasonable cost. Some may not like this, but it’s the only path forward I see that has a reasonable chance of actually happening.
Mjd, et al,
So, I have to ask, when they put in those boilers (fossil plants), did all that bad concrete and those non-qualified welds go away (with the reactor), or are they still present in the fossil plant? That is, were they deemed “good enough” for a non-nuclear plant? As far as what they could have been thinking when they did “crazy” things like not have a (NQA-1?) weld QA program, could it be that they (the local utility) were applying the same standards and procedures that they always used to construct non-nuclear plants?
Sorry, but I see these comments as another example of the “nuclear exceptionalism” that Rod has often written (and complained) about; a mindset that needs to change if nuclear is to have any future. That is, the notion that nuclear’s hazards are uniquely large, and unacceptable, and that a standard of perfection is required for this one industry. It’s also a sign of how nuclear exceptionalism does not only exist in the minds of nuclear opponents, but also (and perhaps mainly) exists in the minds of people who work in the industry. It is a sign of just how hard it will be to get that mindset to change.
Event’s like Fukushima have shown how off base those notions are. Expert consensus that the only release of pollution in non-Soviet nuclear’s entire history, and absolute worst case meltdown of three large reactors, caused no deaths and will have no measurable public health impact, in a world where fossil plant pollution causes ~1000 deaths PER DAY, along with global warming. Economic costs of the accident are a small fraction of those inflicted *annually* by fossil plant pollution. The fact is that even a “shoddily built” nuclear plant is far better than a fossil plant, with respect to public health risk and environmental impact. uniquely stringent regulatory oversight and fab QA requirements are not justified.
Some think that the industry just needs to find a way to fabricate components (and assembly plants) while meeting those uniquely impeccable standards. They hope that gaining experience building large numbers of such plants/components will allow us to do that. My view is that if one industry is held to vastly stricter standards than its competitors, it is very unlikely that it will ever be able to compete. Instead, the industry needs to grow a backbone and stand up for itself, und push back against these uniquely onerous requirements.
First Model T cost $1M sold for ~$1,000 (~$17,000 Today) 1909 by 1927 ~$525.
The dark side of this article is the ambiguity of the title as modular being the
Small Modular Reactor (SMR) or the AP1000?
Ambiguity was intended. Small Modular Reactors have a challenge similar to that faced by Henry Ford; COST of the first units will be far higher than any commercial customer can or should pay. Early adopters will pay higher prices than those who get in line after volume production has been fully implemented.
The trend is well understood by those who have manufactured anything, whether you are talking about cheap plastic toys, big screen televisions, automobiles, ships, or washing machines.
I understand that many may not fully agree with my point of view, expressed above. It’s possible that SMRs may offer a compromise.
The one environment where I think it’s plausible that NRC and NQA-1 fab requirements could be economically met is on a factory floor (assembly line) where large numbers of exact copies of components (or reactor modules) are being built. Even that would only be after they’ve reached the Nth copy, and have an experienced, dedicated staff. That, as opposed to a new crew of people at each plant site that have to learn everything all over again.
Thus, while I still don’t think reactors’ potential failure consequences (especially SMRs) do not justify uniquely strict requirements, I would be willing to accept having those requirements apply to an SMR assembly line, because there is a good chance they won’t increase costs too much, once fabrication gets started in earnest. It may be politically necessary.
However, the key point is that those (uniquely) strict requirements only apply at the assembly line, i.e., to the modular NSSS. It is imperative that the standards applied (fab QA, etc..) at the site be no different than those applied to fossil plants or other standard large industrial projects. Rich mentions how NRC started to think about the BOP. Well, for SMRs, that needs to change (back).
For SMRs, especially those that can go indefinitely w/o active cooling (like NuScale) the argument is that any credible issues or component failures with the site or the balance of plant outside the NSSS have little if any potential to cause a meltdown. Also, even if a meltdown were to somehow occur, the potential release is much smaller (a few percent) than Fukushima. Thus, it is hard to make the case that any such failures outside the NSSS really pose any threat to public health and safety (which is NRC’s real mandate). Those reasons may be enough to argue that standard industrial requirements are sufficient for the BOP outside the NSSS. We better hope so, since applying full NQA-1 and full NRC oversight at the SMR sites will probably render the whole idea uneconomic.
One more thing. Rich mentioned new reviewers doing separate reviews on the same design. The way it would need to work for SMRs is that the design is licensed, ONCE. Then you can deploy that SMR design at different sites with no additional licensing. Much like how it works with dry storage cask designs, the only thing needed for each site is a relatively simple evaluation which shows that the site parameters (g-loads, etc..) are bounded by those (critical) parameters that were modeled in the (SMR design) licensing analyses).
Would you be interested in producing a guest post — or a series of posts — explaining your thoughts on ways to eliminate nuclear exceptionalism?
Manufactured plants offer the opportunity to apply real quality programs at the factory, but NQA-1 is not only unique to nuclear, it is obsolete and manpower intensive in comparison to current internationally recognized quality standards.
BTW – in the early morning hours, I often think about the kind of factory production made possible by the Adams Engine, a low pressure, high temperature, N2 or air-cooled, pebble bed reactor using TRISO coated particle fuel, directly connected to a very slightly modified conventional gas turbine power conversion unit.
Look into how normal maintenance activities and FAA mandated changes are controlled in Aviation. in a nutshell, the mechanic is certified and upon completion of a task certifies that the maintenance, changes, testing, etc. were performed in accordance with the approved procedure and FAA requirements. The process greatly reduces the need for massive, time consuming unnecessary duplicative QA verification. Violation of rules can lead to severe disciplinary measures and industry black listing. Your high paying job is gone forever. Must work, it seems no one has problems flying in planes, and more die there than from nuclear accidents. The NRC rules are so overbearing they create more problems than they prevent. E.g., in the early 70’s plants used “administrative limits” to reduce the number of NRC violations from exceeding dosage limits. The NRC “fixed” this by issuing violations for “noncompliance with procedures” when individuals exceeded the plants “administrative limits.” Thus, plants just eliminated all administrative limits. There are hundreds of other examples just as ridiculous as this in the nuclear industry.
Ever ride a train? Why is it that the relays, controls, instrumentation system fails less than similar systems/components at a nuclear power plant, considering that the vibration, G-forces, sudden impacts are far greater and the maintenance is often “Don’t fix it if it ain’t broke”? And QA is a joke. Same for most of the fossil powered generating stations in the world. I am not advocating going to that low a level, just a level that at least is equal to the same cost/benefit level as used in/for commercial airlines and better than in automobiles. Look at the safety systems in the commercial power plants, They are designed and built at a degree that is as high, not higher, than all subsafe systems. WHY? You do not need to operate a commercial NPP through many, most, accidents. All you have to do is shut it down safely. Those are the true safety systems. The remainder do not need that level of “safety,” QA, etc. You are not in a submarine a 1000 feet below water. And, IMO the ex Navy Nukes have implemented subsafe requirements on all NSSS and many BOP systems for commercial NPPs.
I used the term “subsafe” as I went through that after Thresher. Not sure what they call it now.
I may be interested in that, although my thoughts will likely be those expressed in various past posts on the ANS site (such as the public prejudice series). Not sure I have much new to add.
It may be a few weeks before I could produce a post, however. I’m out of town the rest of this week and am busy next week (with work, and also writing an ANS Café post on the EPA’s final Clean Power Plan).
As for the various alternative/advanced reactor concepts, including your own, they may help, but I don’t think that improved/different technology will be enough to solve the problem. The issue is the fact that NRC will want to apply full oversight and NQA-1 to most of the components and activities for ANY reactor design, simply because “it’s nuclear”. At a minimum, the industry will have to fight a whole lot harder than it has been to have only components that really are important to safety be classified as such.
A couple examples if this. NuScale is humbly trying to argue (w/ NRC) that some of it’s electrical equipment should be subject to full, safety-related QA, the reason being that their reactor does not require active cooling (and thus, failure of said wiring has no potential to cause a meltdown).
Another example is what’s going on at ANO. They had a fairly typical industrial accident on the NON-nuclear side of the plant (where a person died). As a result, they’ve been in QA jail for years. It’s so bad that some are even talking about shutting the plant (“handing the keys to NRC”). Any tangible potential for that event, or the equipment involved, to cause a real problem (meltdown, etc..) is a stretch at best. I suppose NRC’s argument is that it was a sign of “poor safety culture” which is intolerable anywhere within a nuclear plant (that universal catch all, eh?).
So, even reactor designs like yours (perhaps even inherently safe ones) are not immune to this. What would happen if such an industrial accident occurred in the turbine building of your plant? It’s just a black and white double standard that turns on (vs. off) if anything “nuclear” is involved. It’s not based on real risk, or credible failure scenarios, at all. I’m sure even fusion would be subject to NRC and NQA-1 (as though it weren’t implausible enough….).
Technology won’t cut it; at least not alone. A fundamental change in mindset is needed.
I’d be interested in reprinting some of your ANS Nuclear Cafe pieces. Not only is the audience here a bit different, but sometimes good ideas need repetion before they start to sink in.
I would be fine with having you reprinting some of my ANS Nuclear Café articles. If I had to pick, I would way that the (3-part) public prejudice series would be the best one. It sets the stage for other ideas, and would thus perhaps be the best ones to start with.
If you are interested in pursuing this, we could discuss further off line. Do you still have my E-mail address? If not, do you want me to post it here?
@Jim Hopf August 4, 2015 at 9:07 PM
Lets say for discussion, I buy your ideas on NQA-1 fab etc. for SMRs. I can’t really come up with any particularly strong technical arguments for dispute anyway. I will also add these are technical issues, and the US has a strong history of being able to solve tech issues. But that is not the point at all when looking at the long-term marketability of an SMR. It has to be able to make money over the long haul after the initial capital expenditure has been recouped, or nobody will buy one. And there are some particularly stinky “policy” issues affecting SMR long-term O&M budgets that have mostly not even been mentioned to date, much less solved. Because they have not even been identified in current SMR discussion, nobody is working on them. You mentioned one biggie:
” We better hope so, since applying full NQA-1 and full NRC oversight at the SMR sites will probably render the whole idea uneconomic.” Fine, who’s working on it? Is it a show stopper for a potential buyer? Who’s court is this ball in? Full NQA-1 and full NRC oversight at the SMR site is the current requirement. In your own words unless it changes it will render an SMR uneconomic. So how is this change going to happen, if not soon, ever?
Some of the problems I see remind me of the way we used to attempt refuel outages in the bad old days; everything done in series. We’d eventually get done, but it took forever.
Lets discuss some problems using the latest published NuScale schedule. NuScale plans to submit a Design Certification package at end of 2016 (<1.5 yrs away), and NRC has said about a 40 month review time (don't hold yer breath). But in theory, about mid 2019 we have an NRC certified product to buy. Here's some problems ID'd by NRC Staff as "policy" issues with the NuScale design that potentially delay the certification. Staff can't decide them, full Commission must rule on them. NuScale is proposing reduced licensed operator manning and no need for 1E power (see ML15146A088 for details). What is the likelihood the Commission rules in NuScale's favor? Without approval NuScale can throw their design away and start over. What is the full Commission track record on timely policy decisions, especially hard ones? Have they started working on it? What if the vote is split (still only four Commissioners).
Assume these rulings go in NuScale's favor and it's mid 2019 with a certified design. You are a potential buyer trying to figure your future O&M costs with the design. What's your future NRC license fee going to be? NuScale is proposing one COL for a three "module" four reactor per module site, for a total of twelve reactors. Is one fee going to fly for twelve reactors? Will it be the same for a one module four reactor site? If it's good for NuScale why is that not good for Palo Verde? So what are the odds this goes in NuScale's favor?
Lets talk insurance required by 10CFR50.54(w). Similar problem to the NRC license fee, only messier because the insurance is supplied indirectly by the insured utilities through NEIL. Will it be one insurance fee per reactor? Who's currently working on it? The issue needs to be resolved in a buyer's mind to understand future O&M budget.
This discussion is really not OFF-TOPIC. The topic originally is/was is modular construction the savior of nuclear power. Modular construction is a technical issue and the US solves those pretty well. It's not the problem restraining new nuclear development in the US. Policy issues are. And it can't all be blamed on NRC. Until they actually have a Design Certification submittal to actually consider it's all just speculation and even then it is just for one unique SMR design. I'd say our whole future nuclear development system is flawed, except we don't have one so how can it be flawed? It sure ain't about modular construction.
I definitely agree that this subject is not off topic, and that modularization itself will not be enough, unless some more fundamental things change, in the area of policy, regulation, and just how the whole industry does business. It’s also true that NRC is not fully to blame. Much of the problem lies with the industry itself, which not only does not push back against uniquely strict requirements, but has seemed to largely have bought into that mindset itself (that nuclear is unique and demands perfection).
As for who’s “working on it”, your probably right that nobody is. That’s what I’m trying to change. I’ve been trying to get the industry to understand that unless these things change, things are NOT going to be OK, and the industry will die a slow death. Many are in denial about this, and that is the one thing I’m trying to change. But frankly, I feel rather impotent; a lone voice in the wilderness.
I’m not in a position of any authority or influence. Hell, I’m not even a real expert in many of these areas (details of plant design and QA issues, etc…). On many of these issues, I’m not as much of an expert as you guys (commenters) are. Despite that, I still think I’m right about the points I’m raising, i.e., that costs need to come down significantly if the industry is to survive, and that what’s needed to bring costs down lies in the regulatory oversight/requirements/QA area, as opposed to reactor design or technology. I think I bring that one insight to the table. So, I’ve been trying to get that message out, in the hopes that someone who is in more of a position to affect change will hear it. My hope is to put these ideas into the heads of important people.
So far, my efforts only consist of writing articles on the ANS Nuclear Café site, as well as posting comments on various articles/sites, like the comments here. I have no idea if it is having any impact at all. I’m trying to think of other, more effective ways to proceed, but I haven’t thought of any…..
As for SMR fees, we either need to rethink the whole idea of the NRC budget coming entirely from user fees (do coal plants pay for the EPA’s operations?), or we need to scale the fees with capacity. Again, the inherently lower probability, and consequence, of potential releases for SMRs needs to be considered and credited. I would think that lower fees are justifiable since, theoretically NRC shouldn’t be spending as much time on licensing and oversight of such less potentially harmful facilities. What do hospitals pay? Seriously.
@Jim Hopf August 5, 2015 at 6:21 PM
“As for SMR fees, we either need to rethink the whole idea of the NRC budget coming entirely from user fees (do coal plants pay for the EPA’s operations?), or we need to scale the fees with capacity.”
It’s not just the NRC fees, although the current system is a total mess. As I understand it (reading NRC publications) once the NRC annual budget is approved, by law 90% must be recoverable from the utilities. So the total 90% number is divided by the total number of power reactor licenses, regardless of MW size. For that part of the fee, every power reactor base fee is the same. Then they have the “per work hour” fee, which in theory can give NRC incentive to put a plant on an increase work hour “watch list.” Under that structure, as time marches on and more plants retire (without replacement) the last man standing gets 90% of the annual NRC budget as “the fee.” Now NRC will say, as the licensed plant numbers decrease (or increase) the budget goes down (or up) so it is self correcting. But is it? All NRC annual budgets back to 1976 are available on-line, called annual “Information Digest.” Plot the inflation adjusted $ and staff numbers and try to make sense of it all factoring in all the cancellations and new plants licensed over time. You can’t… because it doesn’t make any sense. However there are messages, like the NRC “Research” budget peaked at $205M in 1982, well over what it is today. Tells me NRC is more interested in trying to experimentally determine the TMI2 core didn’t melt (after the fact) than it is today researching a new reg structure for new technology like SMRs. But this can’t all be blamed on NRC either. They have to gear up in advance for things like the “renaissance” and handling 20+ new license apps.
But the NRC fees are only part of the story, and not the only “fee” a nuke utility has to contend with when considering future O&M budgets. In an earlier comment I mentioned the NEIL insurance. Participation in the NEIL insurance pool requires participation in INPO to be eligible for the NEIL insurance. Is NuScale’s “walk-away-safe” reactor also INPO-Proof? If not there is trouble. From what I see something drives the total site staffing level besides MWe. I base this observation on what I see comparing site staffing levels at a Kewaunee or VY1 plant to a single unit 800-1000 MWe plant. The smaller MWe output plants are just not much different at a total site staff level.
If I am a potential buyer for a single module four reactor 200 MWe FOAK unit (like the “Consortium” NuScale is talking with), and I have to gear up my site staff to almost Kewaunee or VY1 size just to keep up with all NRC and INPO programs… here’s a news flash…YOU.WILL.LOOSE.MONEY. For example one of NuScale’s marketing points is reduced licensed operator manning (which reduces O&M budgets). For a three module twelve reactor 600 MWe full NuScale site they are proposing 3-Licensed ROs and 3-Licensed SROs per shift. It remains to be seen if this will fly. But in theory a single module four reactor site should need less licensed operators (if it is considered and approved in the Design Cert submittal). But if I am a buyer I want to know what drives the total site staff level to support operator licenses. If it ends up being an INPO certified and accredited (and maintained) operator training and requal program (including a simulator), because I need that for NEIL insurance, I don’t see a savings on O&M. My training staff will end up being larger than my licensed operator staff just to implement INPO.
Where will the INPO issue be decided? The nuke utilities collectively “own and run” NEIL. Will they “align the stars” in NuScale’s favor for a competitor?
As you correctly point out, something has to change. Discussing the pinch points for future nuclear development in the US of a single unique SMR design, relative to specific policy issues one-by-one, on a blog solves nothing. And ringing in my ears is the echo of a major policy decision made at one point in time being totally unraveled at a later point in time by a single rogue NRC Commissioner controlling a budget. If you win your argument on the NQA-1 requirement today, will it hold up later? I don’t expect an answer because there isn’t one. Our whole regulatory system is flawed. It needs to change more in the direction the navy uses. Regs don’t insure safety, people do.
Modularization is always cheaper than field fabricate.
Let me repeat that for those who didnt hear me: Modularization is always cheaper than field fabricate.
If anything, modularization for Vogtle has most likely offset additional costs elsewhere in the project.
You can repeat and bold your statement all you want. That won’t make it true.
You MIGHT be somewhat more successful if you can provide good references or describe your own experience in successful, completed major projects using modularization.
I’ve done a number of O&G projects in remote areas (Alberta and Queensland Australia) and all major equipment supply is modularized. The savings come from not having to pay a premium for labor in remote spots and from being able to fabricate indoors, year round in a controlled environment drawing upon a local pool of skilled labor instead of having to import them.
Even in a location like Georgia with access to a large local labor pool you will still save money, although not as much with a remote site.
Granted. However, I’m pretty sure that your modules were not first of a kind and I know that they were not required to undergo the kind of regulatory scrutiny applied to the first of a kind modules for the AP1000. I’m sure that they were of sufficient quality to be safe and to work as designed, but they probably would not have been torn apart and rebuilt if a few measurements were slightly different from print.
Rod, these modules were first of a kind. The QA regimes most O&G companies work under are on par with aviation and just a step below nuclear.
I think you are missing the forest for the trees with respect to my comment. Smith’s article states that prefabrication and modularization didn’t save any money. She never actually does anything resembling substantiating her argument though. Its kind of an underwear gnome argument.
1. Voglte is over budget
3. Modularization does not save any money.
She never fills in point #2. Had she been able to substantiate her argument with something along the lines of estimates of stick built systems with similar delays in construction would have cost $x less than the actual costs of modularized systems according to blah blah blah.
The WSJ piece is just an example of garbage journalism with no real informational value. Just a supposition with no supporting evidence. I can point to four large ($150 million plus) O&G jobs I’ve worked on and say ‘based on local labor rates, mobilization costs, yada yada yada …. we saved between 20 and 40% on a particular project’. There’s noting like that with Smith’s article and that’s what you should be hammering away on.
I’ve been to Plant Vogle and spoken with people who are intimately familiar with the challenges associated with building and qualifying the modules there.
The scale of your experience is about 1/10th that of Vogtle, and though the specific modules associated with your projects may have been FOAK, I’d bet they have similarities with other O&G projects. The nukes under construction in the U.S. are not only using a fabrication technique that’s new in the industry, but doing under the supervision of a regulator that has little experience in any kind of new plant construction.
It would be interesting if you could interview CBI and Southern about the trials and tribulations of the whole modular construction process so far. What worked and what required reworking. It is my understanding that everyone (industry and NRC) knew there were going to be numerous license amendments as both Vogtle and Summer worked through construction. It is also my understanding the NRC reviews of the license amendments to date have not held up any of the construction work. I have only heard that in passing so I would like to know if that is true.
I’ve tried to land on the record interviews, but many of the issues where details would be valuable lessons for others are subjects of high dollar litigation. So far, I haven’t found anyone with both knowledge and permission to talk frankly.
Too bad. Maybe you can get in a question at one of the industry conferences. I will also try but I do not get the opportunity to attend these as often as I would like.
China seems to be building reactors in about 6 years or less, basically on budget. Even Russia built the 2 Tianwan reactors in ~ 6 years. From Wikipedia:
Construction commenced on 20 October 1999 for the first unit, and on 20 October 2000 for the second reactor unit. The first reactor went critical on 20 December 2005. Construction of the second reactor finished in May 2007 and commercial operation began in August. This is the first time the two countries have co-operated on a nuclear power project.
So SMR’s may be the answer, but in other environments the large reactors are being built in a reasonable time and reasonable cost.
“…that nuclear is unique and demands perfection.”
This is part of the INPO mantra. It is stated so in the little books we have to carry around to show we have a safety culture.
Can you provide more details? What project is pictured? When was the photo taken?
Let me get this straight:
You post a picture of an engineered, deliberate test-to-destruction of a nuclear rocket motor…
… and use this to imply that all SMRs are unsafe?
I can’t tell whether you’re a master troll doing emotional manipulation as an art, or if you’re simply clueless and have no reasoning faculties.
Your link begins with the following:
Please explain, in your own words, the relevance of your post and your implication that the photo proves anything about the safety of modern SMRs, or even 1960s vintage SMRs.
One of the best reasons for going small is the ability to test to destruction to show that doomsday scenarios are ludicrous.
“The flood walls at Fukushima were not primary side components, but their failure was important to safety…”
The flood walls worked as designed. They were not damaged. They stopped their design basis flood, and nothing more.
If the flood walls were classified as “safety grade” or not is irrelevant. No QA would have saved the plant. It was badly designed, exposing critical electrical infrastructure outside and inside the basement, equipment that is sensitive to flooding, and they built this plant in a location that receives 15 meter floods once every 100-150 years or so.
I’m sure the concrete of the flood walls met all QA, ASTM and whatnot. I am sure it was good concrete. And it didn’t help a darn.
QA focus safety culture is worse than useless. Its dangerous.
QA is not part of the solution. QA is part of the problem.
Well executed quality assurance programs contribute to long term reductions in cost in construction, operations and maintenance. Paper perfect NQA-1 programs, especially those that reference ancient standards because it’s too hard to bring them up to date, add enormous layers of cost in usually uncounted ways.
“Well executed quality assurance programs contribute to long term reductions in cost in construction, operations and maintenance.”
And your evidence for this claim is? Plus evidence that this is not a spurious correlation? I know many companies with skilled workers that can produce good quality. They have QA programs, but the workers do not need it. It produces a spurious positive correlation. The companies that do well do not need QA. They are expert at what they do and have their thinking caps on all the time.
“Paper perfect NQA-1 programs, especially those that reference ancient standards because it’s too hard to bring them up to date, add enormous layers of cost in usually uncounted ways.”
I believe that this is the natural end point that all QA focussed safety programs invariably lead to. It leads to a paper reality by definition.
In additon, QA focus actively discourages innovation, and promotes entrenchment and legal-focused (as opposed to technical-focused) development.
It is bad in all of the corners of the house of cards that is QA.
I look at real accidents at real NPPs and I do not see QA showing up in the priority list. What I see is insufficient design or outright bad design, often a mix. The irony is that things like Fukushima are actively encouraged by the mountains of QA. It is an effective smokescreen, an excellent excuse to stop thinking. After all, the signatures are there, and all the boxes are checked, so we have a safe plant. Right?
Recent Comments from our Readers
The Clinton Nuclear Plant also in Illinois was shutdown essentially for almost 2 years before it was taken over by…
Good Podcast – Very informative One thing that was not discussed is how to deal with a particular fear that…
Renewables people are masters in marketing. Unreliable intermittent generators whose output is all over the place, and usually badly correlated…
Looking at their lineup, Westinghouse seems bound and determined to keep Gen IV in its “place” which is apparently the…
So they are developing a scaled down version of the AP1000, which is a scaled up version of the AP600,…