For SmartPlanet last week, I reviewed the recent closure of four nuclear plants in the U.S., including the San Onofre Generation Station in California, and concluded that nuclear power is headed for a long goodbye. Read it here: Nuclear’s swan SONGS

Stick a fork in U.S. nuclear power. The dream of electricity “too cheap to meter” is dead.

It died last Friday with Edison International’s announcement that it would permanently close the San Onofre Nuclear Generating Station (SONGS) located north of San Diego. The plant (pictured above) has been shut down since January 2012 due to a leak in a tube in its steam generating system.

The reason? It would cost too much to fix.

The leak stemmed from work done by Mitsubishi Heavy Industries, which replaced the plant’s steam generators in 2009-2010. Errors in its computer models, which made the replacement parts to finer tolerances than the initial design, reduced the contact force of anti-vibration barsrestraining thousands of tubes in the steam generators. This allowed the tubes to rub together, eventually resulting in one of the tubes springing a leak. Replacing the tubes thus became a monstrously expensive job. Public opposition to reopening the plant was made worse by UPI’s revelation in May that plastic bags, masking tape and broomsticks were being used to temporarily control the leak, and that internal documents reported degraded equipment with “hundreds of corrosion notifications.” Fending off continued legal challenges, which might have required seeking a time-consuming and expensive amendment to the SONGS license, finally doomed the plant, which cost about $1 million a day to keep ready for a restart.

Falling like dominoes

SONGS, with its 2,200 megawatt (MW) generating capacity, is the fourth nuclear plant to be closed this year due to economics.

Exelon just opted to pull the plug on its 44-year-old, 630 MW Oyster Creek nuclear power plant in New Jersey after workers discovered that underground pipes were leaking tritium. Replacing them with new cooling towers was too costly, Exelon decided.

Duke Energy announced in February that it would close its 37-year-old, 860 MW Crystal River plant in Florida. The plant has been shut down since 2009, when workers cracked a concrete containment building during an upgrade and refueling procedure. After unsuccessful attempts at repairs, Duke decided it was too expensive to continue trying to fix the plant.

In May, Dominion Resources permanently closed its 39-year-old, 556 MW Kewaunee plant in Wisconsin. The plant’s contracts to sell power were ending, and with the current low price of natural gas, continued operation was deemed to be unprofitable. Dominion tried find a buyer to take over the plant, but failed. The company’s CEO said the decision to close the plant “was based purely on economics.”

Even new plants still under construction are coming under fire. Southern Co.’s new reactors at Vogtle in Georgia reportedly are running over budget and recovering costs long before the plants are to begin operation, arousing the ire of locals. SCANA Corp.’s new Summer Nuclear Station in South Carolina is running over budget and incurring delays. The Tennessee Valley Authority’s new 1,200 MW Watts Bar 2 plant, on which construction was halted in 1988, is soon to be completed at a cost of $4.5 billion, 80 percent over its initial budget, the utility says.

Budget overruns and delays are the norm for nuclear plants. As a 2009 study by the Union of Concerned Scientists (UCS) shows, the actual cost of nuclear plants has routinely come in at three times their initial estimates. Cost overruns, canceled plants and stranded costs total more than $300 billion in 2009 dollars, the study said. At a final construction cost of $4.5 billion in 1984(equivalent to $10 billion in 2013 dollars), SONGS was finished at 10 times its original estimate.

Plans for new nuclear plants in Texas and Maryland have also been scrapped as costs continue to rise. And last week, MidAmerican Energy dropped its plan to build a $1 billion nuclear plant in Iowa after a poll showed that 77 percent of Iowans opposed allowing the utility to charge ratepayers up front for its construction. MidAmerican now intends to spend $1.9 billion to build new wind turbines.

Last month, Duke signaled its intent to “suspend” its application for two new reactors at its Harris station in North Carolina, because the units “will not be needed in the next 15 years” according to the utility’s forecasts.

Recovering the lost investments in the closed plants, and paying for their decommissioning, is already fraught with dispute. Florida ratepayers will be on the hook for $1.6 billion in reimbursements to Duke Energy for the closed Florida plant; decommissioning could take 40 to 60 years.

Decommissioning SONGS is also expected to take decades, largely because there is nowhere else to put its 3 million pounds of hot radioactive waste. It will be easier and safer to let the reactors sit in a mothballed state for up to 60 years to let the radioactivity decay before cleaning and dismantling operations proceed. Edison estimates the full cleanup cost will be around $3 billion, of which $2.7 billion has already been collected from surcharges on customer utility bills. Edison has said it will take a $450 million to $650 million charge on the closure. Ratepayers have already paid about $1 billion to Edison for the plant during its closure, and the remaining costs of closure and decommissioning are likely to be borne by ratepayers, not Edison shareholders, save any damages the utility is able to recover from Mitsubishi Heavy Industries.

It’s a safe bet that further litigation lies ahead, as ratepayers seek to push the costs of the failures back onto Edison.

Rising costs

It’s not surprising that reactors begin to show wear and tear after around 40 years of operation, the duration of their original license period. As I explained last year (“Regulation and the decline of coal power”), the majority of the U.S. nuclear fleet is long in the tooth; many of the plants were built in the 1970s, and most are 21 to 40 years old, according to the U.S. Energy Information Administration (EIA). More than half of the plants in the current U.S. nuclear fleet have had their licenses extended for an additional 20 years.

More maintenance issues have come to light this year. The Nuclear Regulatory Commission (NRC) has just ordered 31 older nuclear reactors — a third of the U.S. nuclear fleet — to overhaul their vent systems to prevent a buildup of hydrogen and to keep temperatures from rising in containment buildings. The order resulted from a safety review in the aftermath of the Fukushima meltdown in Japan, which resulted from rising “decay heat” from nuclear fuel. In April, the former chairman of the NRC said that all U.S. reactors suffer from the same design flaw and cannot be fixed, and speculated that the reactors currently operating under extended licenses probably wouldn’t last to 60 years of age. He resigned from the NRC last summer after a conflict over safety issues with his colleagues.

It now seems inevitable that U.S. nuclear capacity is bound to continue falling, with planned new units unlikely to make up for plant closures. As the remaining fleet approaches 60 years of age, it is highly likely that wind and solar will become the cheapest way to add new capacity, decreasing the likelihood that retired plants will be rebuilt.

Meanwhile, nuclear costs have continued to rise, along with construction costs in general. The UCS study noted that between 2002 and 2008, the cost of nuclear plants tripled to an average $9 billion per plant as construction and financing costs exploded. Early cost estimates made by consultants, government and academics to construct new plants have typically run on the low side at around $2,000 per kW ($/kW) of capacity, while utility estimates cluster in the $3,000 to $5,000/kW range, and the estimates of independent analysts and Wall Street run in the $6,000 to $10,000/kW range.

Of those ranges, I believe only the high-end estimates of Wall Street and independent analysts are close to reality, indicating a true cost of about $100 per megawatt-hour ($/MWh), or $0.10 per kilowatt-hour (kWh), of actual production. That is the estimated cost of power from the new nuclear plant under construction in the Kaliningrad region of Russia, a German lawmaker said in April. That’s also in the same zone as the EIA’s current levelized cost of new generation estimates, released in January as part of the Annual Energy Outlook 2013 report, which give an average cost of $108.40/MWh for new “advanced nuclear” plants in 2018. (Comparing the costs of various power generation technologies is a complex task involving many assumptions, but most analysts recognize levelized cost as the best basis for comparison.)

New solar plants are sharply undercutting the cost of new nuclear. As I mentioned in my analysis of carbon capture and storage costs in March, the new 50 MW Macho Springs solar plant under construction by First Solar in New Mexico will deliver power for $50.79/MWh under a 25-year power purchase agreement with the local utility. That price was confirmed last week by Greentech Media. Other U.S. solar projects have come in this year in the range of $70 to $90/MWh.

Compared to nuclear power, solar cost estimates are absolutely rock-solid. They don’t leave billions of dollars worth of decommissioning costs off the books, to be foisted externally onto ratepayers. They don’t leave the cost of managing nuclear waste hanging for decades in an uncertain future, where it will likely be pushed onto taxpayers via federal programs. They don’t require federal insurance liability protection or billions of dollars worth of federal loan guarantees. If solar developers can’t come up with the money to build a plant on budget, it simply doesn’t get built. And the power that solar plants generate is sold under a fixed long-term contract, not bumped up over time as initially lowballed costs inflate. Ratepayers are beginning to recognize that solar (and wind) costs are simply far more trustworthy than nuclear costs.

In addition to cost and safety considerations, renewables have the distinct advantage of being something utilities can build quickly, using much smaller chunks of capital. Instead of committing to spending $5 to $20 billion for a nuclear plant that will take a decade to build — and decades more to recover the investment –utilities may look on investments in renewables more favorably, particularly in the current environment of utility business model disruption, uncertainty, and the growing popularity of renewables. Renewables are also much simpler and less risky than insanely complex nuclear plants; a wind farm or a solar plant isn’t suddenly going to develop flaws that require billions of dollars to fix.

The nuclear plants recently closed or kiboshed are only the beginning, the swan SONGS if you will, of nuclear power’s demise. The cost trends are clearly in favor of renewables and natural gas (at least so long as the latter stays cheap, which is uncertain) and against nuclear. On current growth trends, according to a new analysis by Gregor Macdonald, solar will overtake nuclear generation globally by 2020.

It’s only a matter of time.

(Photo, courtesy SoCal Edison/NRC)