"What happened to the disposable, small scale pebble bed reactors that could be disposed of without removing the fuel from the reactor?"
Somebody accidentally deleted the grad student's Powerpoint presentation?
There's lots of oddball theoretical designs for power-generating reactors kicking aound, most of them the work of grad students. A few of these oddball designs have been built and tested and either failed spectacularly (sodium fires) or just failed (cracked pebbles) or turned out to be so costly to operate they weren't worth keeping in operation even for experimental reasons (most commercial breeder reactors ever built). The Germans had a real power-generating pebble-bed reactor running for a couple of years, the THTR-300 but the company operating it went bust after it broke down when a fuel pellet got stuck in the wrong place. The Chinese have the HTR, a small pebble-bed research reactor in operation but it only generates 10MW and is not intended to be commercial (i.e. capable of producing electricity at about $0.05/kWh).
"There are industrial accidents and oil and gas explosions every week that have more fatalities than the Fukushima disaster."
There was a fuel pipeline explosion in West Africa in the autumn of 2011 that killed over a hundred people (it is thought they were stealing petroleum). The incident barely made the news headlines here in the UK, I don't know if it appeared on the news in the US at all. The news here today has just reported on a gas explosion in Clacton that destroyed several houses and put a bunch of people in hospital. Compare those ho-hum incidents with the acres of breathless DOOOOOM!!! reporting from Fukushima where... there was that worker (62 years old) who died of a heart attack in the summer of 2011 and, um, well...
Reservoir storage costs money, about $200 million per GWh assuming good geography for the high and low reservoirs and lots of surplus water (a major problem for large drought-ridden parts of the US and other countries). Pumped-storage wastes about 30% of the electricity fed into it, for every GWh put in it can put out 650MWh on demand. Sabatier reactors and other electrochemical cells have poor conversion efficiency and cost money to build and operate so the round-trip efficiency of electricity in to electricity out will be a lot worse than pumped-storage.
So who's going to pay for these storage systems? They don't generate any electricity, just waste lots of it so they're not eligible for renewable subsidies and taxpayer dollars. The only source of cash for such investments are the consumers who are screeching that their energy bills are too high now.
Pebble-bed reactors have been around since the 1970s and earlier. They have problems. The German high-temperature pebble-bed AVR was shut down in 1988 and they're still trying to figure out how to completely decommission it -- they hope to develop techniques to do this over the next 60 years although that could be wishful thinking on their part. Some of the fuel pebbles are wedged in cracks in the reactor, along with bits of intensely-radioactive broken "pebbles", dust and other crap. In comparison the Japanese recently decommissioned a conventional uranium Magnox reactor (Tokai 1) in just 11 years from shutdown to greenfield status i.e. they could grow crops on the land where the reactor used to stand.
Other developers are looking at pebble-bed designs again, there's some interesting materials research going on with thorium as part of the fuel mix -- the idea is that neutrons from the enriched uranium and/or plutonium in the fuel pebble breeds the thorium into U-233 which then fissions and produces more energy. Light-water reactors are the way ahead for the next twenty years at least though.
Sodium-sulphur batteries cost about $2 million per MWh of storage. They also catch fire (not surprising when you've got sodium metal and sulphur at 300 deg C in a battery...) and it can take weeks for the fire to be put out.
Dinorwig in Wales can produce about 1.6GW at maximum output and can store about 8GWh in total. It is never run down to empty or even close as it's part of the Black Start network designed to bring the British grid back up after a major power failure.
There is another big pumped-storage station in Scotland, a repurposed hydro dam at Cruachan which can generate 440MW with a similar storage capacity to Dinorwig, about 8GWh. It is also a Black Start facility. There are plans to build a third large pumped-storage facility in Scotland but they're still in the position-papers and Powerpoint stage of the process.
The famous ORNL molten salt reactor never generated any electricity. It produced, during short runs at maximum output, 7.4MW of heat that was dumped to air. It also never used thorium fuel, not suprising as it wasn't meant to breed thorium into fissile fuel. It used expensive U-223 specially made in other isotope-producing reactors and later it ran on U-235 as an experiment.
ARM is a design operation, it produces no silicon of its own. Intel is an end-to-end company both designing and fabricating its own product with world-leading fabs, die shrinks, finFETs etc. Intel hasn't bothered much with low-power performance up till now as there wasn't a demand for it until relatively recently -- performance was for the data centre, supercomputers and workstations, low power consumption was for embedded devices and braindead phones and handhelds. Laptops were in the middle but a three-hour battery life was deemed acceptable if a powerful CPU was required.
The markets have changed and Intel is putting serious effort and expertise behind a push to lower-power versions of its stock-in-trade x86 platform while ARM is moving into unknown territory trying to wring performance out of its core designs which were always optimised for low power operation first and foremost. At the moment it seems Intel is winning.
You say "backup" as if it was your singular backup. If you only have one backup you effectively have no backups. Two backups is a good start. An extra offsite backup is better, two offsite backups is starting to live in the real world where excrement eventuates.
It needs to be thoroughly mixed to a statistical level to be used as fuel in a reactor as lumps of higher and lower enrichment within fuel pellets could cause some odd radiological and physical problems. Converting HEU to UF6 makes the mixing easier and more predictable. Of course we could stop guessing and look up how the downblending process was actually achieved...
Conversion to UF6 and then "mixed with other material". They're not specific about whether this "other material" was depleted UF6 or regular UF6 produced from minehead yellowcake. Either would be available to the downblenders and would work as well but I don't know the intricacies.
I don't know which they chose for the downblending process but if they had depleted UF6 to hand they may well have used that; it's a byproduct of centrifuge enrichment lines. It's expensive to convert the depleted UF6 back into metal unless there's a good reason. The US has about 700,000 tonnes of UF6 in storage, for example.
The Boxing Day tsunami that hit Indonesia and elsewhere in the southern Pacific in 2004 killed over 230,000 people, over ten times the number of dead resulting from the Great Tohoku earthquake and tsunami of 2011. Indeed that tsunami wasn't even the biggest natural disaster to hit Japan in a hundred years as over 100,000 people died in the Great Kanto earthquake of 1923.
You're wrong in many particulars but your errors are common ones commonly repeated so I won't bother knocking them down one by one. I'll just point out that IFR reactors cover a wide range of designs and technologies and several have run successfully for a few decades although not without major problems in some cases. The Soviet/Russian BN-350 is one such fast reactor which operated from the mid-70s providing electricity and desalination process heat. It was only shut down around 2000 when its rather specialised fuel was no lnoger available. The larger sodium-cooled BN-600 based on a similar design started up in 1980 and is still running today despite several accidents and fires as the core was not compromised and there were no radiation leaks when things did go wrong. It's been enough of a success that an even larger version (the BN-800) is being built in Russia (planned commercial startup is April 2014) and, if agreement is reached, more BN-800s or dervatives will be built in China.
The physical technology of the IFR is pretty well understood, the problems arise due to the failure of materials exposed to high temperatures, very high neutron fluxes, radiochemistry and other factors that should be solvable with enough experience of operation and seeing where the bits land after the next explosion. Baby steps baby steps.
The US paid for downblending to be done in Russia and the other ex-Soviet republics; no weapons-grade material (90% plus enriched) was actually shipped to the US.
Not particularly. Nuclear fuel is cheap per joule of electricity generated, about 0.6cents/kWh. It's the cost of operating the plant, paying off the loans to build it in the first place, licencing and regulation, insurance, paying for spent fuel disposal and funding the eventual decommissioning of the plant that brings the total generating cost up to par with coal or gas.
The recycling of the ex-Soviet weapons material has depressed the world markets for mined uranium for the past few years meaning yellowcake (U3O8) is ridiculously cheap, currently $35/lb. A number of mining operations have cut back production or closed temporarily for this reason. Even if yellowcake tripled in price it would only add a cent or two to the wholesale price per kWh for generators. If the cost of gas or coal tripled that would have a much greater effect on electricity prices.
That mercury WAS buried underground well away from the biosphere. Now it's been dug up, burned, vapourised and spread over cropland and towns and cities downwind, deposited into rivers and lakes supplying drinking water to the population before it eventually makes its way into the sea where it bioaccumulates in fish to the point where authorities recommend people don't eat too much of it because of the toxic mercury content.
You might want to look up "sequestration" sometime.
The US has a large stockpile of weapons-grade material (U-235 and also Pu-239) from decommissioned nuclear weapons produced in the 1960s when it had over 30,000 weapons ready for use. It now has about 5000 warheads, most in reserve (i.e. not ready for immediate use or kept as "junk box" units that could be refurbished given the need, will and funding). The ready-for-use warhead count is about 2,200 or so.
They don't need to divert this ex-Soviet material to make more weapons, they don't need more weapons, they don't have the launchers and platforms to carry more weapons and they don't have the facilities or funding to pay for new weapons to be built and besides the uranium arriving in America has already been downblended to fuel-level enrichment (probably 4 or 5%) from the original 90% or so of the original weapon cores.
That's how we know the US didn't use it for their own weapons.
The pollution from coal waste is permanent, it never decays unlike nuclear waste. US coal-fired power generators pump 50 tonnes of mercury into the environment every year, it never goes away or decays, it ends up in water and the soil, in the sea and seafood. Nobody cares, any attempt to reduce these sorts of emissions is a "War on Coal".
Apple didn't exist back then, it was Apple Computer. Apple Computer produced the first Mac, Apple produce the current Mac range. I forget when the rebranding occurred.
Sony, Toshiba, Fujitsu and Panasonic were Japanese electronics manufacturers making MSX-based personal computers in 1983 before the Mac was released and they're still manufacturing PCs today.
As far as I know most worthwhile computer-Go programs don't use alpha-beta for the reasons you stipulate. A lot of the in-game scoring involves trying to calculate territorial control and areas of influence, moyo in Japanese. Like chess, at the beginning of a game computer programs tend to use standard openings (joseki) and as in chess an experienced human player will break out of the book early on in the game to force the computer to start evaluating moves instead. The end-game, like chess simplifies down as there are many fewer possibile moves on each turn and ownership of large areas of the board have already been decided or conceded at which point calculating a full solution to the end is possible. Its the tricky bit in the middle that's the fun part (especially ko fights).
Small-board Go of 7x7 size has a small enough number of unique positions that its gamespace could be exhaustively evaluated today given the ubiquity of rentable cloud computing and cheap storage to hold the results. All it would cost is time and money. Creating a good Go-playing computer program which can evaluate positions and choose moves wisely is more tricky.
In Go it is assumed black going first has the advantage and in an even game white is given some points in compensation (komi).
I vaguely recall reading that Go on a 7x7 board had been strongly solved some time ago by exhaustive computer analysis and I assumed that 9x9 Go would be the next to be solved. Is anyone working on the problem at all? How much would it cost to rent enough computer power to do so? Could it be done using BOINC?
Slashdot Mirror
"What happened to the disposable, small scale pebble bed reactors that could be disposed of without removing the fuel from the reactor?"
Somebody accidentally deleted the grad student's Powerpoint presentation?
There's lots of oddball theoretical designs for power-generating reactors kicking aound, most of them the work of grad students. A few of these oddball designs have been built and tested and either failed spectacularly (sodium fires) or just failed (cracked pebbles) or turned out to be so costly to operate they weren't worth keeping in operation even for experimental reasons (most commercial breeder reactors ever built). The Germans had a real power-generating pebble-bed reactor running for a couple of years, the THTR-300 but the company operating it went bust after it broke down when a fuel pellet got stuck in the wrong place. The Chinese have the HTR, a small pebble-bed research reactor in operation but it only generates 10MW and is not intended to be commercial (i.e. capable of producing electricity at about $0.05/kWh).
"There are industrial accidents and oil and gas explosions every week that have more fatalities than the Fukushima disaster."
There was a fuel pipeline explosion in West Africa in the autumn of 2011 that killed over a hundred people (it is thought they were stealing petroleum). The incident barely made the news headlines here in the UK, I don't know if it appeared on the news in the US at all. The news here today has just reported on a gas explosion in Clacton that destroyed several houses and put a bunch of people in hospital. Compare those ho-hum incidents with the acres of breathless DOOOOOM!!! reporting from Fukushima where... there was that worker (62 years old) who died of a heart attack in the summer of 2011 and, um, well...
Reservoir storage costs money, about $200 million per GWh assuming good geography for the high and low reservoirs and lots of surplus water (a major problem for large drought-ridden parts of the US and other countries). Pumped-storage wastes about 30% of the electricity fed into it, for every GWh put in it can put out 650MWh on demand. Sabatier reactors and other electrochemical cells have poor conversion efficiency and cost money to build and operate so the round-trip efficiency of electricity in to electricity out will be a lot worse than pumped-storage.
So who's going to pay for these storage systems? They don't generate any electricity, just waste lots of it so they're not eligible for renewable subsidies and taxpayer dollars. The only source of cash for such investments are the consumers who are screeching that their energy bills are too high now.
Pebble-bed reactors have been around since the 1970s and earlier. They have problems. The German high-temperature pebble-bed AVR was shut down in 1988 and they're still trying to figure out how to completely decommission it -- they hope to develop techniques to do this over the next 60 years although that could be wishful thinking on their part. Some of the fuel pebbles are wedged in cracks in the reactor, along with bits of intensely-radioactive broken "pebbles", dust and other crap. In comparison the Japanese recently decommissioned a conventional uranium Magnox reactor (Tokai 1) in just 11 years from shutdown to greenfield status i.e. they could grow crops on the land where the reactor used to stand.
Other developers are looking at pebble-bed designs again, there's some interesting materials research going on with thorium as part of the fuel mix -- the idea is that neutrons from the enriched uranium and/or plutonium in the fuel pebble breeds the thorium into U-233 which then fissions and produces more energy. Light-water reactors are the way ahead for the next twenty years at least though.
Sodium-sulphur batteries cost about $2 million per MWh of storage. They also catch fire (not surprising when you've got sodium metal and sulphur at 300 deg C in a battery...) and it can take weeks for the fire to be put out.
http://www.ngk.co.jp/english/n...
That's not the only NGK battery fire that's been reported over the past few years even though there aren't that many installed around the world.
Dinorwig in Wales can produce about 1.6GW at maximum output and can store about 8GWh in total. It is never run down to empty or even close as it's part of the Black Start network designed to bring the British grid back up after a major power failure.
There is another big pumped-storage station in Scotland, a repurposed hydro dam at Cruachan which can generate 440MW with a similar storage capacity to Dinorwig, about 8GWh. It is also a Black Start facility. There are plans to build a third large pumped-storage facility in Scotland but they're still in the position-papers and Powerpoint stage of the process.
The famous ORNL molten salt reactor never generated any electricity. It produced, during short runs at maximum output, 7.4MW of heat that was dumped to air. It also never used thorium fuel, not suprising as it wasn't meant to breed thorium into fissile fuel. It used expensive U-223 specially made in other isotope-producing reactors and later it ran on U-235 as an experiment.
ARM is a design operation, it produces no silicon of its own. Intel is an end-to-end company both designing and fabricating its own product with world-leading fabs, die shrinks, finFETs etc. Intel hasn't bothered much with low-power performance up till now as there wasn't a demand for it until relatively recently -- performance was for the data centre, supercomputers and workstations, low power consumption was for embedded devices and braindead phones and handhelds. Laptops were in the middle but a three-hour battery life was deemed acceptable if a powerful CPU was required.
The markets have changed and Intel is putting serious effort and expertise behind a push to lower-power versions of its stock-in-trade x86 platform while ARM is moving into unknown territory trying to wring performance out of its core designs which were always optimised for low power operation first and foremost. At the moment it seems Intel is winning.
You say "backup" as if it was your singular backup. If you only have one backup you effectively have no backups. Two backups is a good start. An extra offsite backup is better, two offsite backups is starting to live in the real world where excrement eventuates.
It needs to be thoroughly mixed to a statistical level to be used as fuel in a reactor as lumps of higher and lower enrichment within fuel pellets could cause some odd radiological and physical problems. Converting HEU to UF6 makes the mixing easier and more predictable. Of course we could stop guessing and look up how the downblending process was actually achieved...
http://www.usec.com/educationa...
Conversion to UF6 and then "mixed with other material". They're not specific about whether this "other material" was depleted UF6 or regular UF6 produced from minehead yellowcake. Either would be available to the downblenders and would work as well but I don't know the intricacies.
I don't know which they chose for the downblending process but if they had depleted UF6 to hand they may well have used that; it's a byproduct of centrifuge enrichment lines. It's expensive to convert the depleted UF6 back into metal unless there's a good reason. The US has about 700,000 tonnes of UF6 in storage, for example.
http://web.ead.anl.gov/uranium...
The Boxing Day tsunami that hit Indonesia and elsewhere in the southern Pacific in 2004 killed over 230,000 people, over ten times the number of dead resulting from the Great Tohoku earthquake and tsunami of 2011. Indeed that tsunami wasn't even the biggest natural disaster to hit Japan in a hundred years as over 100,000 people died in the Great Kanto earthquake of 1923.
You're wrong in many particulars but your errors are common ones commonly repeated so I won't bother knocking them down one by one. I'll just point out that IFR reactors cover a wide range of designs and technologies and several have run successfully for a few decades although not without major problems in some cases. The Soviet/Russian BN-350 is one such fast reactor which operated from the mid-70s providing electricity and desalination process heat. It was only shut down around 2000 when its rather specialised fuel was no lnoger available. The larger sodium-cooled BN-600 based on a similar design started up in 1980 and is still running today despite several accidents and fires as the core was not compromised and there were no radiation leaks when things did go wrong. It's been enough of a success that an even larger version (the BN-800) is being built in Russia (planned commercial startup is April 2014) and, if agreement is reached, more BN-800s or dervatives will be built in China.
The physical technology of the IFR is pretty well understood, the problems arise due to the failure of materials exposed to high temperatures, very high neutron fluxes, radiochemistry and other factors that should be solvable with enough experience of operation and seeing where the bits land after the next explosion. Baby steps baby steps.
The US paid for downblending to be done in Russia and the other ex-Soviet republics; no weapons-grade material (90% plus enriched) was actually shipped to the US.
Not particularly. Nuclear fuel is cheap per joule of electricity generated, about 0.6cents/kWh. It's the cost of operating the plant, paying off the loans to build it in the first place, licencing and regulation, insurance, paying for spent fuel disposal and funding the eventual decommissioning of the plant that brings the total generating cost up to par with coal or gas.
The recycling of the ex-Soviet weapons material has depressed the world markets for mined uranium for the past few years meaning yellowcake (U3O8) is ridiculously cheap, currently $35/lb. A number of mining operations have cut back production or closed temporarily for this reason. Even if yellowcake tripled in price it would only add a cent or two to the wholesale price per kWh for generators. If the cost of gas or coal tripled that would have a much greater effect on electricity prices.
That mercury WAS buried underground well away from the biosphere. Now it's been dug up, burned, vapourised and spread over cropland and towns and cities downwind, deposited into rivers and lakes supplying drinking water to the population before it eventually makes its way into the sea where it bioaccumulates in fish to the point where authorities recommend people don't eat too much of it because of the toxic mercury content.
You might want to look up "sequestration" sometime.
The US has a large stockpile of weapons-grade material (U-235 and also Pu-239) from decommissioned nuclear weapons produced in the 1960s when it had over 30,000 weapons ready for use. It now has about 5000 warheads, most in reserve (i.e. not ready for immediate use or kept as "junk box" units that could be refurbished given the need, will and funding). The ready-for-use warhead count is about 2,200 or so.
They don't need to divert this ex-Soviet material to make more weapons, they don't need more weapons, they don't have the launchers and platforms to carry more weapons and they don't have the facilities or funding to pay for new weapons to be built and besides the uranium arriving in America has already been downblended to fuel-level enrichment (probably 4 or 5%) from the original 90% or so of the original weapon cores.
That's how we know the US didn't use it for their own weapons.
The pollution from coal waste is permanent, it never decays unlike nuclear waste. US coal-fired power generators pump 50 tonnes of mercury into the environment every year, it never goes away or decays, it ends up in water and the soil, in the sea and seafood. Nobody cares, any attempt to reduce these sorts of emissions is a "War on Coal".
"bad ass in a fight?" The F-22? Did you see how it performed in "Pacific Rim"? Pitiful.
Apple didn't exist back then, it was Apple Computer. Apple Computer produced the first Mac, Apple produce the current Mac range. I forget when the rebranding occurred.
Sony Viao L-series all-in-one desktops PCs.
http://www.sony.co.uk/product/...
Fujitsu (no longer Fujitsu-Siemens) Esprimo desktop PCs.
http://www.fujitsu.com/uk/prod...
Panasonic tablet-based PCs running Windows 8.1
http://www.panasonic.com/busin...
Sony, Toshiba, Fujitsu and Panasonic were Japanese electronics manufacturers making MSX-based personal computers in 1983 before the Mac was released and they're still manufacturing PCs today.
I'd vote for the memorial dedication to read "Next time, no more Mister Nice Guy. Capisce?"
As far as I know most worthwhile computer-Go programs don't use alpha-beta for the reasons you stipulate. A lot of the in-game scoring involves trying to calculate territorial control and areas of influence, moyo in Japanese. Like chess, at the beginning of a game computer programs tend to use standard openings (joseki) and as in chess an experienced human player will break out of the book early on in the game to force the computer to start evaluating moves instead. The end-game, like chess simplifies down as there are many fewer possibile moves on each turn and ownership of large areas of the board have already been decided or conceded at which point calculating a full solution to the end is possible. Its the tricky bit in the middle that's the fun part (especially ko fights).
Small-board Go of 7x7 size has a small enough number of unique positions that its gamespace could be exhaustively evaluated today given the ubiquity of rentable cloud computing and cheap storage to hold the results. All it would cost is time and money. Creating a good Go-playing computer program which can evaluate positions and choose moves wisely is more tricky.
In Go it is assumed black going first has the advantage and in an even game white is given some points in compensation (komi).
I vaguely recall reading that Go on a 7x7 board had been strongly solved some time ago by exhaustive computer analysis and I assumed that 9x9 Go would be the next to be solved. Is anyone working on the problem at all? How much would it cost to rent enough computer power to do so? Could it be done using BOINC?