Thursday, March 20, 2014

The Truth About Nuclear Power - Part Three

Subtitle: Nuclear power plants cost far too much to construct.  

The instant cost plus inflation, escalation, and interest on loans adds up to more than $10,000 per kW. 
Vogtle Nuclear Plant and Construction Site
photo - Wiki Commons by Charles C. Watson Jr.

One reason that nuclear power plants are uneconomic is they cost far too much to construct for the amount of power that they produce.  If one were to build a new nuclear power plant in the USA today, the final cost would be more than $10,000 per kW.   Several references support this assertion, Severance (2009), MIT (2003), and California EnergyCommission (2010).  All of these three referenced sources use $4,000 per kW as the overnight cost.

Overnight cost is the cost to construct if the plant could be built all at one time, or “over night”.  Of course, a nuclear power plant cannot be built overnight, as they require years to construct.  The added years increase the cost by escalation of materials and labor, and by interest on construction loans.   

Severance calculates the escalation for materials and labor to be $3,400 per kW, and for interest on construction loans to be an additional $3,100 per kW (figures rounded).   The total then is $4,000 plus $3,400 plus $3,100 equals $10,500 per kW.  A new, twin-reactor plant that produces 2,000 MW net electricity would then cost $21 billion to construct.   However, as indicated in Part Two of this series, Severance and the others did not include funds to make the plant operate safely if a large commercial aircraft crashes into the plant.  Not only the reactor, but the spent fuel storage area and the cooling water system must remain operable, per new NRC regulations.  This brings the cost to construct to approximately $12,000 per kW. 

How does this estimate compare to recent experience in the US?  There are two reactors under construction in Georgia, at the Vogtle plant.  Two more reactors were cancelled in Texas due to the excessive cost estimate at the South Texas Nuclear Project, STNP.   The STNP expansion project would add two reactors to the existing two, and was cancelled after a cost estimate of $17 billion was conceded by the reactor vendors to be too low.  As a result, we will never know how much that plant would cost to construct. 

The Vogtle plant is advertised as costing “only” $14.3 billion for twin reactors at 1100 MW each using the Westinghouse AP-1000 design.  However, cost overruns already incurred have increased the cost to $15.5 billion.  It is notable that Georgia changed the state law to allow the utility to bill customers in advance for construction costs.  This was an attempt to not pay finance charges on the construction loans.  In essence, rate-payers pay more money for electricity they are already using, and the utility company spends that cash for the nuclear construction.  Without this creative financing, the Vogtle plant would be right in line with Severance’s number, $20 billion more or less.   

The Vogtle plant is also plagued by delays in the construction, which would add to the cost if traditional financing were used.   At present (1Q 2014), the reactors are two years behind schedule, with four years to go for the first reactor to start up.  Many problems can arise in the next four years, which will likely add to the cost and delays.  As Severance shows, each year of delay adds approximately $1.2 to $1.6 billion in interest costs to the final cost for a twin-reactor plant.    An interesting account of the Vogtle plant’s progress can be found at

[Update 6/24/2014: Vogtle facing more delays and cost increases  see link  -- end update]

In Finland, a single-reactor Areva nuclear plant is experiencing similar cost overruns and schedule delays. 

[Update 7/16/2014:  Finland's Areva EPL reactor plant is 7 years behind schedule and Billions of Euros over budget.  Per the article linked below:

“ "Areva was ready to do anything to win the Olkiluoto deal, including downplaying project management deficiencies. They had also previously delivered and commissioned nuclear reactors but they had never undertaken an entire project end-to-end, since the main French contractor had always been the EDF Group (Électricité de France), explained Les Échos editor in chief Pascal Pogam in an interview with Yle’s A-Studio current affairs program.
Based on accounts by parties such as the Olkiluoto owner-operator, the Finnish power consortium Teollisuuden Voima or TVO, Areva is said to have lied about the possibility of constructing a nuclear reactor within the agreed schedule."   see link  -- end update ]

It is asserted that other countries can and do build nuclear power plants for approximately $2000 per kW.  As an example, China claims to build AP-1000 reactors at $2,000 per kW, according to   One must pause at that; perhaps the lower labor rate in China is the reason, perhaps lower escalation for materials, and perhaps favorable (read: zero) cost for interest on construction.   However, the same website ( states that France’s current program has reactors that cost the US-equivalent of $5,000 per kW for overnight costs.  (Euro 3,700 per kW)


Truth Number 3:  Nuclear power plants cost far too much to construct, more than $10,000 per kW

Overview of The Truth About Nuclear Power series:

The series on Truth About Nuclear Power has several main themes:
          Nuclear power operating costs are too high, cannot compete
         Nuclear power costs too much to construct, require government assistance in loan guarantees or bill current ratepayers for construction funds (Georgia).
         Nuclear power is unsafe to operate, near-misses occur frequently, disasters happen too; they must run at steady, high output to reduce upsets; this increases revenue to spread out the very high fixed costs; older reactors are more uneconomic and less safe (San Onofre leaks in new heat exchanger is a prime example)
         Nuclear power is unsafe long-term for spent fuel storage
         Nuclear power consumes far too much precious water
         New designs to overcome these failures are unlikely to work, or to be economic if they can be made to work
a.       Thorium Reactors have serious developmental issues
b.      Modularized, smaller PWR (pressurized water reactor) reactors lose economy of scale advantages
c.       High temperature gas-turbine style reactors are far from developed
d.      Fusion at high temperature e.g. in magnetic bottle, is a pipe dream

         Nuclear death spiral on the demand for power is real and present, customers have a variety of ways to self-generate (distributed generation), and alternatives become attractive as power prices increase.  Nuclear power will increase power prices, the greater the percent nuclear, the more alternatives become attractive. 

      Part One  --  Nuclear Power Plants Cannot Compete.
Part Two  --  Preposterous Power Pricing in Nuclear Proponents Prevail
Part Three -- this article
Part Four  --  Nuclear Plants Use Far More Fresh Water
Part Five --   Cannot Simply Turn Off a Nuclear Power Plant

      Roger E. Sowell, Esq.
      Marina del Rey, California


Rex Berglund said...

Congratulations on your prescient assertion that "A new, twin-reactor plant that produces 2,000 MW net electricity would then cost $21 billion to construct."

The following is from a Georgia PSC meeting last week about the Vogtle expansion:

"The cost of the new reactors, originally projected at $14 billion, is now close to $19 billion and might reach $21 billion, according to recent PSC filings."

Roger Sowell said...

Mr. Berglund, thank you for that. New PWR nuclear plants in the US will without doubt reach or exceed the $10,000 per kW installed cost.

Even in the UK, the cost is now admitted to be that high, or higher, for the proposed Hinkley Point C twin-reactor plant, as described here on SLB at this link

"What is interesting is the quoted price to build the 3200 MW Hinkley C plant, at £24.5 billion (the equivalent of US$ 39.2 billion). This equates to MORE than $10,000 per kW, at $12,250. Again, this is precisely what SLB has maintained all along - a new nuclear power plant costs far more than the $4,000 some advocates maintain. Instead, it will cost at least $10,000 per kW, and more likely $12,000 per kW. Here (in the UK) we see at least a small beginning of honesty from the nuclear establishment.

However, given the long, dismal history of nuclear plant schedule delays and cost overruns, it is to be expected that the Hinkley Point C twin reactor plant will take far longer than 10 years to startup, and will cost far more than US$ 39 billion. It will likely require 15 years or longer, and $48 billion or even more."