Darlington B – please, no AECL! June 11, 2009Posted by Maury Markowitz in AECL, nuclear.
Tags: nuclear power
Soon, likely on June 20, “Infrastructure Ontario” will announce the winner of the Darlington B contest, which is down to Areva, Westinghouse and AECL.
Yes, that AECL. The one that costs us hundreds of millions of dollars a year to keep running so they can design reactors that bankrupted Ontario Hydro while promising that our sales efforts would lead to untold riches in the international market.
I wrote this piece a while back, but it seems the time is right to get it out there. Enjoy!
The recent flap over the Maple reactors and the NRX shutdown have focused the media’s attention on AECL, but much of this focus has been on the isotopes market and a possible breakup and sale of the company’s assets. There is a larger story that seems to be slipping under the radar, the whole issue of the CANDU reactor and its follow-on, the ACR. This takes on some urgency now that it’s been announced that Ontario will be building new reactors at Darlington, although the design that they will build has not yet been announced.
A recent Globe and Mail story quoted AECL as suggesting that cancelling their new ACR reactor design would be similar to cancelling the Avro Arrow. Even decades later, the Arrow continues to generate anguish over its effects on the Canadian aviation industry – it’s the one that got away. AECL banks on the fact that most people will consider the Arrow’s cancellation to be paramount to, as one book puts it, “Selling the Canadian Dream”. A sober analysis would note that in spite of this supposed disaster, Canada is the #3 aircraft manufacturer in the world, and that most other countries of similar size (Italy, Sweden, Israel – even the UK) have all abandoned indigenous fighter production for the same reason we canceled the Arrow – it’s too expensive.
But the AECL argument isn’t about facts, it’s about emotion. It’s not so much about the Arrow as it is about what the Arrow myth exemplifies; if we cancel ACR, will we regret it?
In a word, no.
I’m going to talk about the CANDU a bit, and I think that when you hear the story -or my version of it anyway- you will likely conclude that continued development of AECL’s ACR-1000 reactor design is an obvious example of throwing good money after bad.
There was a point in time where one could argue that the CANDU design was superior to the reactors being offered by most companies, on technical grounds. The main benefit was that using heavy water allowed CANDU to use un-enriched fuel, which lowered the cost of operation. Instead of having to use various enrichment systems to get the fuel to “burn” in the reactors, with CANDU you simply refined the raw ore like any other metal – melting it down, cleaning it, and forming it into rods. In a traditional design, the “Pressurized light Water Reactor”, or “PWR”, enrichment represents over 1/2 of the cost of the fuel. Given a large nuclear economy, which everyone in the 60s expected, CANDU’s lower fuel processing costs would mean cheaper power.
Another practical advantage of the CANDU design was that it could be continually re-fueled. In a PWR the reactor core is a huge pressure vessel, so in order to re-fuel it you have to cool it down, crack it open, refuel it, seal it and start it back up again. In comparison, the CANDU had high pressure only in small tubes, which were inserted into a larger low-pressure vessel known as a “caldera”. Since there was no single pressure vessel, the CANDU could replace fuel rods while the reactor was still running. Since the vast majority of the cost of nuclear power comes from the construction of the reactors, every second of power it generates lowers the overall cost of the power generated over time. So in a PWR, the time spent refueling is lost cash, pure and simple.
Another side effect of the lack of enrichment is that it lowers the proliferation risk. With some work, the same plants that enrich fuel for conventional reactor designs can also be used to enrich fuel for nuclear weapons — not good ones mind you, but weapons nonetheless. The difference between enriching for power and enriching for weapons is not something you can tell from a reconnaissance photo, as the recent events in North Korea or Iran demonstrate. With CANDU there is no need for enrichment, so any country using CANDU that started building enrichment systems would immediately be suspect.
So in theory the CANDU offered a 1-2 sales punch; for us Canadians it meant lower operational costs and cheaper power, while its lower proliferation risk meant it could be sold to countries that would otherwise be considered “off limits” with conventional technology. Everyone had high hopes for the system.
Simply put, the nuclear economy that everyone expected never materialized. That means that the savings in fuel costs simply aren’t there. If Ontario had dozens of multi-reactor plants, which is what people expected, then the CANDU might have actually been a cost-effective solution. But with only three plants, the savings in fuel infrastructure is washed away. The downside is that there’s a cost in trying to get those savings; CANDU is larger, more expensive to build, requires extremely expensive heavy water, and extracts less power per volume. So in the end the CANDU is more expensive, not cheaper.
A lot more expensive.
When two CANDU 6’s were sold to China the total cost was estimated at $4 billion. For the same price they could build a South Korean two-unit APR-1400 that generates well over twice as much power. Startup costs on all of the newer “generation III” designs are similarly reduced; Westinghouse’s AP-1000, General Electric/Hitachi’s ABWR and the Areva EPA are all less expensive to build than CANDU, while also producing more power.
Even the operational “uptime” of the CANDU, a major selling point in the 70’s, has since long been surpassed by other designs. When CANDU was first coming online in the 1970s they generated an impressive 80% uptime, when reactors of the same generation in the U.S. and others were 70% at best, and closer to 65% on average. Since then, however, a better understanding of the reactor operation has reversed this trend. While the CANDU maintains an impressive 88% track record in Canada (lower elsewhere), the U.S.’s boiling-water fleet is running at about 92%, in spite of many of those reactors being older than the CANDU fleet.
Even the reduced proliferation risk turned out to be a bit oversold; while it remains difficult to remove useful fissile material from CANDU, it’s relatively easy to remove tritium. Tritium can’t make a bomb, but it can dramatically boost the yield of even simple weapon designs. That’s not good.
That feeling is your wallet groaning
Theory’s great, but how did it all pan out in practice? Well, look no further than Darlington A, which was five times over budget! AECL is quick to blame others for this, and it’s true that the majority of the overruns were due to labour disputes, changing governments, and rampant inflation. They’re not so quick to mention the very real technical problems they ran into though, like the ever-failing power shafts and changes to the control systems that added hundreds of millions of dollars. We Ontarians are still paying for that one, it’s the “debt retirement charge” on your electricity bill.
Outside Ontario, the majority of sales have been one-off deals of such dubious nature that it makes your stomach hurt just thinking about them… Romania under a communist dictatorship? Argentina under a military one? Pakistan and India, both of whom would go on to build nuclear weapons? China, where “loans” paid for half the cost and then they told AECL to take a leap when similar loans were not forthcoming for further builds? South Korea, who was so impressed by CANDU that they designed their own reactor that completely outperforms it?
And so CANDU ends up being essentially an Ontario design. Of the 34 CANDU’s operating worldwide, 20 of those are in Ontario, well over half of the fleet. This has cost the Canadian taxpayer an estimated 20 billion dollars [it’s more now] in AECL funding alone, ignoring the cost of actually building and maintaining the reactors. Add in the billions of subsidies offered to foreign countries and the continued offers to Turkey and Jordan, and what you have is an unarguable, unmitigated, unprecedented financial disaster.
It’s not like the government is unaware of this, they’ve had a stream of 3rd party accounting companies tell them this over and over again. They just try their best to bury it.
It shouldn’t be terribly surprising that CANDU cannot match the performance of more modern designs – that’s the definition of progress. AECL’s solution? A new CANDU, the ACR-1000. And how good is the ACR-1000? Its barely competitive with other generation III designs, and there are a bunch that totally outperform it.
A good example of a modern generation III reactor design is the Westinghouse AP1000, one of the designs in the running for Darlington B. In comparison to older designs generating the same amount of power, the AP1000 is about 1/2 the size, has 1/2 the valves, 1/3 less pumps, 80% less piping, and 85% less cabling. What does that mean? Well a basic engineering maxim is that the number of parts is the essential measure of failure rates, so the AP1000 is going to be many times safer than a generation II design like CANDU — ignoring anything else.
But they haven’t ignored everything else, they’ve improved everything else too. The AP1000, and most other generation III designs, are “passively safe”, meaning one can cut off every cable and pump and the reactor will, slowly but surely, shut itself down. CANDU simply can’t match this, and there is a growing international understanding that if nuclear power is going to move forward, that sales will have to be passively safe, low-proliferation designs.
So what about ACR-1000? Well lets see: it’s the same size as a CANDU 6 but generates 1000 MWe, so that’s good. But it’s still much larger than the other designs, so that’s bad. Passively safe? No. Still requires expensive heavy water? Yes. Still doesn’t need enrichment? No! Can run on a variety of mixed fuels? Yes. Can everyone else? Yes.
AECL trots out a shopping list of features for the ACR-1000, but every one of them can be met or bettered by other designs, ones that have already completed their design process, are less expensive, and still aren’t finding any buyers.
Arrow, my foot
So apparently we’re in the midst of this big nuclear power renaissance, or so the companies keep telling us. Canceling the ACR now, right when we need it, would be just like canceling the Arrow just before it got the Iroquois engine. Lets not be hasty and end up regretting it!
You know, whatever else AECL might want to make us think of with the Arrow analogy, there’s one thing that’s different: the Arrow was a great plane, a world-beater in some ways. The ACR is not.
The funding argument remains the same, sadly. There’s all these sales out there, and we need to be ready to enter the market. But is anyone outside Ontario actually going to buy this thing? No. That’s been made very clear. Even Hugh MacDiarmid, AECL’s CEO, has stated that the only hope for external sales is to make us buy them first. So, the pressure is on to get OPG to do it’s Darlington expansion with the ACR. And for that we Ontarian’s get to pay more for our electricity in order to help sell a federal program that we paid for. Nice.
But let’s say that happens, let’s say that OPG is politically manhandled into buying APR-1000. Will that increase sales? Good luck! GE-Hitachi, Westinghouse,, Areva and others are already way in front of AECL. Yet I can’t see the lineup of bulldozers building those designs. Even so, we’d have to sell dozens of ACR-1000’s to pay off what’s been dumped into the project even over the last couple of years, let alone the last 15 when Darlington A finally started operation. That’s just not going to happen.
Stick a fork in it
The feds are desperate to unload AECL, and who can blame them? A couple of hundred million a year and all they get for it is a stream of political scandals? I can sympathize.
But if they’re going to get a reasonable price when they dump it, then they need to have something to sell. The medical isotopes side is one disaster, and the reactor side another. The hope is that if they can cajole OPG into buying the ACR, then the reactor side won’t be a complete money pit, and maybe Areva will be willing to buy it. Good luck with that.
There’s a term for this, it’s called “throwing good money after bad”. Mr. Harper, pull the trigger already.