Follow us |
|
| Follow us |
|
| "Remember Your Humanity" blog |
|
Plutonium pit production: haste makes waste or, how many pit production facilities will Biden Administration policies imply -- one, or at least three? Two is not an option. Part three Dear colleagues -- With the Administration's budget request yet to be submitted it seems useful to write again on the crucial topic of the new Administration's proposed pit production requirements and policy. Two prior letters on this topic are included below the line. (Nota bene: this letter corrects a mistake in the 5/3/21 letter; see below.) The gist of this letter is, firstly, this: the requirement to make more than a de minimus number of new plutonium warhead cores ("pits") in the 2020s for training and technology preservation purposes is far more costly than meets the eye, and should be abandoned. As noted previously, present requirements are mostly blind momentum and contractor self-interest dressed up as some kind of national emergency. Secondly, the awkward effort to use Building PF-4 at Los Alamos National Laboratory (LANL) as a pit factory -- let alone as the sole U.S. pit production factory, for which it is nowhere near adequate -- is going to require the construction and operation of anywhere from one to three more additional Hazard Category II (HCII) plutonium facilities, depending on schedule and other factors, plus the Savannah River Plutonium Processing Facility (SRPPF) at the Savannah River Site (SRS), which is required in any case. At LANL, construction of other (non-nuclear) facilities, long-term property leases, and from 1,600 to 2,000 permanent additional workers will also be required. At present the U.S. has no functioning pit factory. This is not yet a problem, but it is important to avoid denial and wishful thinking as to what would constitute a reliable, adequate pit production facility. An unsafe facility is not reliable; a facility which requires two production shifts (p. 15) to produce a pit every week or two, as is apparently planned, is not really a production facility either. An old, small facility with poor access to labor markets despite being close to residences, one that also houses essentially all of the National Nuclear Security Administration's (NNSA's) other necessary plutonium missions, is never going to be an enduring, reliable production facility. I hope it is by now apparent to all that PF-4 cannot be the sole U.S. pit production facility (slide 15; slides 46-50), no matter how many billions are spent to acquire, stabilize, and increase capacity there (slides 23-31). The recent revelation that producing even 20 pits per year (ppy) in PF-4 would require 24/7 operations (p. 15), assuming it is true, is a stark reality not envisioned in any prior study. Jon Medalia's body of work at the Congressional Research Service (CRS) on pit production mentions, but does not analyze, two-shift production at LANL as a possible means of reaching 50 or 80 ppy -- not 20 ppy. Of course NNSA may be lying, for public relations purposes. (Hopefully NNSA is being candid with Congress at least.) The true aim of investing in, and initiating, two-shift production at LANL may be a higher production rate. There is a statutory requirement to reach for a higher capacity, which may be classified. This does not change the real conditions on the ground that are cited in the links above. Production at PF-4 that is a) enduring and b) does not threaten other critical plutonium programs is an impossible dream, as NNSA decided in 2017 (pp. 2, 47-48). While NNSA looked at higher production levels in 2017 there is no evidence, or any other reason to think, that production at PF-4 is likely to be reliable at any level. Recently, LANL's geographic and infrastructure constraints have become more widely apparent ("Los Alamos National Lab seeks up to 100,000 sq. ft. of laboratory space somewhere within 50 miles," May 6; "NNSA announces huge cost increase for mysterious LANL plutonium warhead project," April 28; "Third power line proposed for Los Alamos; environmental assessment process starting," April 19; "LANL leases second office complex in Santa Fe as nuclear weapons growth pushes admin staff off 'The Hill'," March 8). There are more shoes yet to drop -- not least in legacy waste management, a long-standing problem at LANL. If we assume that no pit older than 55 years will be used in a life extension program (LEP), a 30 ppy facility starting up in 2026 and running perfectly thereafter could, by itself, build a new-pit stockpile of about 540 warheads, including spares and surveillance units, by 2044. If existing pits could be used in LEPs conducted than 2044, a larger stockpile could be supported -- larger by 30 warheads for every year past 2044. (The May 3 letter below erred -- mea culpa!) Allowing pits up to 70 years old in LEPs would enable a 30 ppy facility to support a roughly 990-warhead stockpile. That being the case, and as much as some of us would prefer a much smaller stockpile than today's 3,800 warheads, an immediate commitment to a 74% to 86% cut -- which is what sole reliance on a 30 ppy facility would mean -- is not politically realistic. And again, the notion that PF-4 would or could be a reliable 30 ppy facility, let alone remain reliable over any length of time even assuming on-time, successful startup, is ridiculous. PF-4, built for other purposes in 1978, with or without augmentation by other facilities, is far from a "forever facility." SRPPF is the only adequate, timely pit production infrastructure project available, however much some may object to this fact on ideological or partisan grounds. Unlike PF-4, SRPPF is a) adequate and b) enduring. SRPPF will be expensive -- although even at $10-14 billion (B) it is no more expensive than starting up pit production at LANL, as it turns out (see slides 23-31; estimated LANL costs have increased by roughly $2 B since then). What may not be obvious to new members of Congress, or to others new to the issue, is that the "two-site" pit production strategy announced by NNSA in May 2018 actually implies more than these two pit production facilities -- at least three but possibly as many as four or more remodeled plutonium facilities. That two-site policy is an expensive mistake. Only SRPPF is needed, because pits are not needed in the 2020s. First, as noted previously, reliance on PF-4 means reliance on a PF-4 successor facility also. Augmentation (via, for example, one, two, or even three "modules") does not solve PF-4's age problem. Production, especially two-shift production, in PF-4 will only hasten its effective end-of-life. PF-4's allowable Material-at-Risk (MAR) has already decreased 31%, from 2,600 kg to 1,800 kg (p. A-2) but this has not been enough to lower the maximum dose to off-site individuals in the design basis accident to within the required 25 rem (p. 4). One of the fundamental problems at LANL is that despite LANL's 2-or-more-hour distance from the larger labor market of Albuquerque, PF-4 is no farther from residences (~3,000 ft, map) than is, say, the Superblock plutonium facility (Building B332) at the Lawrence Livermore National Laboratory (LLNL) (at the corner of 2nd St. and S. Gate Dr., map), now being managed as a Hazard Category III (HC III) facility -- not least because of its proximity to residences. Keeping up "two-site" production will thus require paying for at least two HC II production facilities at LANL -- one remodeled, and at least one other new -- plus the still-required SRPPF at SRS. This additional LANL facility or suite of facilities was not in the FY21 budget request. To retain continuity of production at LANL, budgeting for such a facility or suite of facilities needs to begin soon given the long lead-time required. Obviously such a facility or facilities need(s) to be included in life-cycle cost estimates. Notice that one or more new HC II plutonium facilities are needed both for augmentation, if pursued, and also, separately, for replacement of PF-4. If it is proposed that LANL can achieve a sufficiently high pit production capacity via augmentation, to avoid the expense of SRPPF, the augmenting facility or facilities are really needed as soon as possible. Later, a replacement for PF-4 will also be needed. Given their high price tags, NNSA may temporarily deny the need to budget for either kind of new plutonium facilities at LANL. NNSA may also deny that the early timing of any proposed PF-4 replacement is in any way related to either pit production. (The future of surplus plutonium disposition, which up to now has required oxidation of metallic plutonium at PF-4, also bears on the future of PF-4, as does the future of the heat source plutonium program. Like pit production, both programs involve significant Material-at-Risk, MAR.) This is not the end of the cost story, however. To the extent pit production and associated activities, and possibly surplus metallic plutonium oxidation, claim floor space in PF-4 otherwise needed for important lower-mass activities, some of these other activities (e.g. stockpile surveillance, prototyping) may need to go elsewhere. There are also significant risks to non-pit programs created by starting up pit production, as well as risks that quantity pit production in PF-4 may be delayed or may fail. Concern for other (non-pit-production) plutonium missions illuminated NNSA's thinking in 2016-2018, across the otherwise quite different pit production Analysis of Alternatives and Engineering Analysis. For all these reasons, and especially if a capacity of at least 80 ppy is stringently required by 2030, yet another HC II facility with pit production capability must be made available as soon as possible, one where operations are not placed at risk by construction and equipment replacement at LANL's TA-55. The only available facility for some of these purposes -- perhaps all -- is Building 332 at LLNL, where there is a long history of pit prototyping and existing pit surveillance and manufacturing expertise. Taking a step back and looking at the big picture, haste is making a lot of waste. At the moment, NNSA is rushing toward spending at least $10 billion each to start up two pit production facilities, but even this will not be enough. At LANL and quite possibly at LLNL as well, billions more will also be required. PF-4 replacement alone would be a $10 B project, assuming a place to build could be found. As we keep repeating, there is absolutely no national need to make pits in the 2020s, and therefore no need to start up industrial pit production at LANL in the present decade -- or at any other time, since SRPPF is going to be required under all circumstances. It is the rush to build pits, and therefore the imaginary need to do so at LANL, which is generating these unnecessary billions of dollars in costs. As noted before, there is no real national security value added from this rush, no matter how much value one ascribes, or doesn't ascribe, to nuclear weapons. Regardless of decisions regarding the Ground Based Strategic Deterrent (GBSD) program, we do not understand why producing the new W87-1 warhead at all, or producing the W87-1 in MIRVable quantities, or producing the W87-1 with a new pit as opposed to an existing W87 pit (which is young enough to keep in any life extension program), provides noticeable value to the U.S. under almost any set of beliefs about nuclear deterrence. The cost of producing new pits for the stockpile prior to about 2035 is meanwhile tremendous. Thank you for your attention, Greg Mello
Part two Update to 3/11/21 note re plutonium pit production: delaying significant stockpile production is the best policy
Dear colleagues -- We last wrote most of you on March 11. That letter is below this one, separated by a line. There have been a few developments since then that might be of interest. One is the press announcement of NNSA's CD-1 decision regarding the Los Alamos Plutonium Pit Production Project (LAP4, 21-D-512), which was the subject of the first press release above. Using the new upper-end cost estimate of $3.9 billion (B) (high-end estimates, we have found, are better predictors than any other official number), I am estimating that the cost increase in LAP4 is on the order of $2 B since last year's budget request. That earlier estimate was $1.752 B through FY25 only, but that appeared to encompass most LAP4 spending. The new LANL directorate just announced for "plutonium infrastructure" (see above) encompasses a good deal more than LAP4. If we now guess that LAP4 will drag on two more years, based on the strange language in the LAP4 CD-1 announcement and some other official sources, and we include plutonium program spending at LANL over these two additional years (~$2 B), plus some overhead for waste management, facility maintenance, some share of site-wide upgrades, and various GPP projects, and anything else not included (~$1 B), we get very roughly $10 B in startup costs for pit production at LANL through FY27. This is approximately the same spending profile we predicted back in September 2020 and sent to you on March 11 in the note below. The SRPPF is going to be very expensive, but it will not be more expensive than LANL's planned costs through FY2030. Under the assumption that LANL might be able to make about 43 pits per year, we got the astounding per-pit, fully-burdened cost of $38-60 million per pit (ppy) (slides 29-30). If on the other hand LANL can only make an average of 30 ppy as we now hear (30% less), the per pit cost at LANL would be substantially more than this. I suppose most of the uncertainties could be captured within a range of $35 M to $80 M per pit -- assuming, that is, NNSA is successful at LANL, which remains an open question. A 30 ppy production facility starting up in 2026 and running perfectly thereafter would, by itself, support a stockpile of about 400 warheads of 5 types, if we assume no pit older than 55 years will be used in a LEP. (Under that assumption, a 30-year-lifetime LEP would be retired with an 85-year-old pit). If on the other hand we can use 70-year-old pits in a LEP, and assume LEPs are re-done each 15 years, requiring half as many surveillance units, a 900 warhead stockpile could be maintained with a 30 ppy facility starting in 2026. The assumption that 24/7 production at LANL (two production shifts plus a maintenance shift) could ever be stable, at such an isolated site with all its problems including a production facility built in 1978 for R&D only, is either laughable or tragic. As the Institute for Defense Analyses said, it is "very high risk." Any plan involving 24/7 production in PF-4 is really no plan at all. It is little more than wishful thinking, like LANL's 1996 promises to DOE of an extremely cheap "50 ppy" and "100 ppy" pit programs. (Summarized in: So why is it proposed? Why is it necessary to build pits in the 2020s? We see this as "a very expensive way to breathe fresh life into Livermore," as one knowledgeable federal reviewer put it. Yes, it provides a new warhead for the Air Force, but why? We see no compensating value beyond what is already available in the W87-0. We see instead an unhealthy degree of Russophobia and Sinophobia in this desperate and wasteful rush. Instead of patiently investing in a proper facility that will last decades if needed, the rush to have pits right now is because of the wicked Russians (and Chinese), if congressional testimony is to be believed. The other thing we see is partisan Democratic interest. The Obama-Trump plan was, under Klotz, for one pit factory. Nobody had ever, in past decades, imagined building two of the things. New Mexico Democrats, and others, howled and that was changed into a plan for two pit factories, the more permanent one at SRS and a smaller, implicitly temporary one at LANL. This plan was meant to satisfy everybody, both politically and managerially. Democrats, led by Heinrich and Udall, wanted a permanent, large pit production mission at LANL, and Democrats didn't want to provide pork barrel spending to South Carolina. Some in the arms control community and certain watchdog groups wanted to stick pit production in the worst place for it, in the worst facilities, with a conscious view toward making it stumble. But a plan pretending to be for a single pit factory at LANL is really a plan for two pit factories at LANL, one now and one...as soon as possible, as long as it is not in the FYNSP! Is there a suitable place at LANL to put another nuclear facility? We don't know of one. Just the PIDAS for a greenfield LANL location would cost a billion dollars. By the mid-2030s, cumulative SRS production at ≥80 ppy (meaning ~103 ppy on average, p.13) would surpass LANL's cumulative production assuming a 30 ppy average (under the highly-optimistic assumption LANL could keep up a 24/7 "surge" until then without collapsing not just pit production but all plutonium programs). There seems to be a myth that LANL has a trained pit production staff. It does, to a degree, but LANL does not have the additional 1,600 to 2,000 trained staff it needs to operate 24/7. Providing workplaces for this staff, and housing, and transportation to and from work, are all significant challenges. The present two-site plan, and the arms control plan to build all the pits at LANL, are both based on the supposed need to make pits right now. There is no pit crisis -- not yet anyway. But there will be one if NNSA thinks it can rely on round-the-clock production at PF-4. Thank you for your attention. We hope you are enjoying the spring and I am looking forward to seeing you later this year, circumstances allowing. Best wishes, Greg Mello
Part one Re: LANL leases second office complex in Santa Fe as nuclear weapons growth pushes admin staff off "The Hill"; 2,000 more staff needed for "24/7" plutonium "pit" production mission for only 20 pits/year, Mar 8, 2021. (I forgot Los Alamos County as a possible location for an additional property lease to support pit production.) Dear colleagues -- We have doubts as to whether LANL is on track to meet its 2026 deadline for 30 ppy. In FY19 and FY20, the LANL management and operating (M&O) contractor Triad LLC lost some ground in pit production apparently (see FY 2020 PES, FY 2019 PER (FOIA recvd 30 Jan 2021, redacted), FY 2019 fee determination letter), though the details have not been released. I think we know that NNSA is not going to meet its 2030 pit production deadline. Fortunately, if we take the needless MIRV'ed W87-1 off the table -- along with the rushed W93 with its youthful pit (that apparently could not have been a new-pit warhead anyway because there was no way there could be manufacturing capacity for it in the early 2030s) -- there is no need to make pits until at least the late 2030s. Meanwhile, upon information and belief, GBSD is likely to be delayed somewhat for non-pit reasons. The W87-1 warhead may also be delayed for non-pit reasons. A 2035 production deadline would seemingly be possible to achieve if NNSA avoided wasteful, artificial emergencies tasks such as the "need" to build and start up two pit factories. Apart from supporting a MIRVed W87-1, all the benefits of LANL production -- developing and sustaining production technologies, and training staff -- could be captured at some nominal, small (≤19 WR) ppy rate -- i.e. at somewhere less than two production shifts. We can't be sure where the plutonium sustainment confidence and training thresholds lie in terms of WR pits, but in no case is there a need to undertake more than one production shift at LANL to capture all the benefits of LANL's expertise and facilities (such as they are). It is not even clear to me that making WR pits at LANL is necessary at all. Prior plans included a small number (≥5 each) of "development," "process prove-in," and "qualification" pits made in successive years. If quantity production is unrealistic or not cost-effective, aren't these adequate? I would think that prolonging the life of PF-4 and thereby delaying the cost and risk of (another) greenfield plutonium facility would be a "no-brainer." As NNSA found in 2017, TA-55 "modules" just can't be made large enough to matter, given the limitations of the TA-55 site in both the vertical and horizontal dimensions (the "rock" is too weak to support large buildings given anticipated seismic loads, and the site is too small). "Modules" would still depend on PF-4, which will "age out" about the time pits are actually needed if not before. |
|||
|
|
|||
|