Letter to some congressional staff re: ongoing failure of pit plans at LANL

July 8, 2021

Dear ______ --

1. There has already been a major event in pit production at LANL, really a tremendous failure, but as far as I can tell Congress is not aware of it.

Until May 2018 there was an NNSA program called "Plutonium Sustainment" which included an effort to acquire a nominal 30 pit per year (ppy) capacity at LANL. As of 2017, this was supposed to cost about $3 billion (B), including $1B in operations from FY19 – FY23, including funds both to sustain current operations and to add the new 30 ppy capacity (slide 2).

That is $200 million (M) per year, about what had been spent each year up to then (slide 34) going back to FY2000.

From this we know that the nominal 30 ppy Plutonium Sustainment program at LANL, as of 2017, envisioned a single production shift with little or no additional program costs. Some data is missing: there was no FYNSP information provided in CY17 (on March 16) for FY18 (slide 35). But the Nov. 2017 SSMP (for FY2018) contained "No substantive changes from the FY 2017 SSMP" (p. 162) except for an acceleration in the proposed (but not yet chosen) "Plutonium Modular Approach" (PMA), which, it was said, might begin in FY19 (p. 194). The FY 2017 SSMP and FY 2017 budget request did not reflect any significant net LANL hiring for plutonium operations (budget request, p.  67).

The 2017 Analysis of Alternatives (AoA) made it clear that the 30 ppy Plutonium Sustainment program was not a reliable industrial program that could simply be added to, say, a 50 ppy production somewhere else to get 80 ppy (pp. 1, 41).

All the alternatives and analysis in the AoA and Engineering Analysis (EA) envisioned single-shift operations (AoA pp. 13, 17; EA pp. 8, 31, 45, 112, 160).

In May 2018, NNSA and DoD selected an alternative which had been rejected as incredibly expensive in the AoA (Option SRS1-A, pp. 12, 39, 45-46) and not analyzed in the EA, namely to produce a reliable 30+ ppy at LANL and 50+ ppy at SRS. From the AoA, p. 12: "...the difference between a 50 WR ppy equipment set and an 80 WR ppy equipment set is within the range of error and, therefore, did not have an appreciable effect on the determination of the preferred alternatives." See the space analyses in the AoA on pp. 17-20.

All alternatives that relied on PF-4 for enduring pit production capability had been formally rejected by NNSA under Klotz (pp. 57-58). So the new Trump decision to build TWO pit production facilities, one centered in PF-4, came as a doubly unprecedented shock.

In March 2019, the IDA study explicitly warned against multiple production shifts at LANL (pp. 8-9).

Then came this shocker, the full meaning and implications of which Congress has not assimilated as far as I can tell: to reach even 20 ppy, less than the old "Plutonium Sustainment" level, LANL would need to go to 24/7 operations and hire an additional 1,600 people to do so. To get to 30 ppy, LANL would have to add yet another 400 people (Supplement Analysis, draft, Mar 2020, pp. 12,14; Supplement Analysis, final, Aug 2020, pp. 15,17).

This huge proposed increase in staff -- a doubling from 2,000 Triad staff supporting pit production as of January 2020, according to three sources -- brings with it a huge increase in infrastructure required, along with requirements for newly-leased spaces in Los Alamos, Santa Fe and potentially elsewhere.

The staffing levels for 30 ppy at LANL were not even included in the life-cycle cost estimates of the EA (pp. 74-75). They were just "disappeared" right out of the analysis.

And what happened to that "$3 B" in total costs to get to 30 ppy? That has increased to about $11 B (see slide 23; add $2 B in additional LAP4 costs not foreseen until CD-1 and add one more year to take the total through FY26). In other words, LANL costs have almost quadrupled since 2017, and risks have correspondingly increased as well. (I could and should make this part more precise, and will, but this is all the precision supported by NNSA's estimates anyway.)

Life-cycle cost (LCC) and costs per pit at LANL are likewise staggering. The notion that PF-4 can last a century (FY22 budget request, p. 209) when it is not even known when or if PF-4 can meet current safety standards does not bear serious consideration. Under such considerations LCC analysis or any defensible prediction of pit production over time is impossible. We can't get a per-pit cost less than $35 M no matter how we squint at the numbers -- $50 M is a better median guess -- and don't see an annual operating cost less than $1 B, largely because of the staffing required (slides 29-31). Replacing and/or augmenting PF-4 would be extra.

Using LANL pits in the W87-1 triples unit cost and at least doubles W87-1 program cost overall, as noted (slide 30). (These figures use a steady-state ["enduring"] 43 ppy at LANL instead of 30 ppy and are about $2 B too low for LAP4's current estimated cost and omit capital costs for small projects as well as any pro-rata share for site-wide and corridor-wide improvements, so they are conservative.)

So why in the world would anyone want to make any appreciable pits at LANL?

Finally, as mentioned here, to get the ≤20 ppy mission (LANL's formal mission from December 26, 1996 to Sept 2, 2020), LANL and DOE said acquiring a production capacity of 50 pits per year (ppy), single-shift, 5 days per week at LANL would require a $110 M capital investment, followed by $30 M/year operating costs, plus $200 M to upgrade PF-4. Production could begin in 2002 (p. 8-7). An additional $44 M would provide capacity of 100 ppy, single-shift, 5 days per week (p. 8-19). 100 ppy was thought to be the maximum LANL capacity (p. 8-2). See Analysis of Stockpile Management Alternatives, DOE, July 1996. 

In other words, LANL and DOE said LANL could do the whole mission for $354 M in capital costs then, and an additional $30 M per year -- or in today's dollars, $616 M and $52 M/year. That's very roughly 35 times less than current estimates. In the 25 years LANL has had this mission it has made roughly 1 War Reserve (WR) pit per year. To me this looks like mission failure -- which is approximately what Ranking Member Turner said, in so many words.

2. LANL has serious and unique problems.

There was a reason a mutual friend at OMB said they were joking there that it would be easier and cheaper to build a pit factory in Albuquerque and ship it to Los Alamos by rail -- including building the railroad! (I wonder if it would be cheaper to build it in Albuquerque (at SNL) and keep it here. Farther from residences (3 miles), with more stable geology, essentially no seismicity, flat topography, and lower security costs too. It is worth re-analysis given the 4x increase in cost at LANL, if keeping the work in NM is a prime unspoken directive.)

We touched on some of the following issues but I believe key members of Congress are unaware of just how serious they are. In order of importance, and keeping the list small --

1. Workforce issues, which may well be intractable both as to quantity and quality;
2. Infrastructure issues at PF-4 and elsewhere, some unpublished, of which we are aware and need not reference in detail here;
3. Offsite, non-LANL issues including water supply, housing, educational standards, unusually high rates of crime, addiction, and corruption;
4. Transportation issues; singled out because it is almost impossible to imagine operations without a new access highway, which will be bitterly opposed by powerful parties, including in Los Alamos;
5. Site topography and crowdedness (essentially every square foot of TA-55 is already full; see aerials here); TA-55 is also very close to residences; LANL as a whole is full;
6. Geology, including seismicity, which affect not just nuclear facilities but also supporting facilities, several of which do not meet federal life safety standards, as well as access roads, the hospital, etc;
7. Waste management and shipping capacity, already the subject of state legislative and executive branch concern and legal action as well as tribal concern.
   

I could expand this list but these are the top items. When Verdon speaks of "reliability" issues at LANL he is alluding to the effects of some of these issues. Frankly NNSA is ever the optimistic agency. No amount of money can fix these problems.

3.Congress needs to ask for a study of what increasing the capacity at SRPPF from 50 to 80 would cost.

According to the AoA, the difference is much smaller than the precision of the cost estimates involved. I think you can see that already from the space and equipment requirements.

4. In 2017 NNSA saw no programmatic reason why 2033 would not be good enough. So why is it a problem now?

Nota bene it was primarily the fault of the armed services committees that created an impossible 2030 statutory deadline for 80+ ppy. In 2017 NNSA said it would take until FY2033 to reach that production rate, at the earliest, if work began in FY2018 (p. 12). It didn't. Projects didn't begin until FY2019, about one fiscal year later. By 2018 Congress and NNSA already knew that 2034 was a more reasonable date to achieve 80+ ppy.

These conclude my comments for now.

Thank you for your attention,

Greg