Here’s one version. It contains the transformerless power supply (the large capacitor, large resistor, zener diode and rectifier diode) as well as the current measurement shunt on their own PCB; the PCB that also holds the mains plug and socket contacts.

The 5V regulator, op-amps, microcontroller, EEPROM and LCD display are on a second, separate PCB. Here, the op-amp chip (a LM2902, which is equivalent to the very common LM324 quad op-amp) and the EEPROM (a pretty standard 8-pin I²C serial EEPROM) are both in surface-mount SOIC packages and they’re both mounted on the front side of the PCB, along with the main microcontroller IC (which is packaged in the cheap and nasty mass produced chip-on-board or COB technique, where a naked silicon die is mounted on the PCB, connected up with wire bonds and then covered in a blob of conformal epoxy) and the pads which connect to the conductive-rubber zebra connector which connects to the LCD display.

The six wires from the transformerless power supply board to the microcontroller board bring over the 240VAC active and neutral into the voltage-measurement op-amp (configured as a differential amplifier with significant attenuation), both sides of the current shunt resistor, into a differential amplifier, and the floating positive DC rail and the device’s DC ground from the transformerless power supply.

Of course, sampling the voltage waveform is very useful as it allows the true mains voltage to be measured, along with things like the power factor.

(Please remember that the Kill-a-watt uses a transformerless power supply, and you have to be extremely careful if you’re attempting any kind of interface to external electronics. It must be optically isolated, with an independent power supply on both sides of the optoisolators.

Basically, there is a 5V potential difference between the microcontroller’s Vdd and Vss rails, but there is *not* a potential difference of 5 volts between the microcontroller’s Vdd rail and mains earth – the 5V rail inside the kill-a-watt will in fact be at about 120VAC or 240VAC relative to mains earth. All the components inside are at the live AC mains potential relative to mains earth.

So if you just bring out a digital line – I2C or serial or whatever it is – from inside the Kill-a-Watt’s microcontroller and connect it to another external microcontroller with a conventional power supply, that external microcontroller will see 120 VAC on that line, relative to its own DC ground. Which is a Bad Thing.)

There’s also a slightly different variant, where the microcontroller, EEPROM and op-amp are all on the main board, and almost all of the components on the main board are surface mount, but the op-amp is replaced with a DIP through-hole package mounted on the back side (the non-copper side) of the board.

Here’s a photo of this version.

(Photo courtesy of Limor Fried, and used (pinched) in good faith. Please check out the great products available from Adafruit Industries :) )

Now, here’s a different version. In this version, the mains input and output sockets, the transformerless power supply components, the voltage regulator components that generate the +5V and +6.2V rails, the op-amp chip and the components that are associated with the op-amp chip are all on a single board, and the LCD display, microcontroller and EEPROM are on a second board.

In this version, the LM2902 op-amp is in a through-hole DIP package, and through-hole resistors are used for the op-amp circuits. On the microcontroller board, the microcontroller is packaged in a standard package, which appears to be something like a 48-pin QFN. The chip’s label appears to be etched off, however. The EEPROM is still in a standard 8-pin SOIC package.

The six wires between the analog instrumentation board and the microcontroller board appear to be +5V, DC ground, two different analog outputs from the current measurement circuit into the microcontroller’s on-board ADC, the output from the voltage-measurement circuit into the microcontroller, and one more signal which I’m not completely sure about, but which I believe appears to be mains zero-crossing detection.

There appear to be a couple of slightly different batches of this model. One uses a X-class capacitor with a metallized polypropylene dielectric, in the moulded cuboid package, in the transformerless power supply circuit. A different variant uses a cheaper non-X-class red resin-dipped 630V polyester “greencap”.

(Images courtesy of Sparkfun, and used in good faith. Please support Nate and the team, they make nice things. :) )

Note the provision on the PCB for what appears to probably be a relay used to switch the mains load on and off, although it is not populated on the board on any of these units.

Note that the microcontroller (or ASIC) chip is labelled in the last picture, and it is branded by PRODIGIT out of China. It appears that all these devices are actually manufactured by PRODIGIT, and branded or badged under different names.

And finally, just for those of you who might be interested, here’s a schematic of the Kill-a-Watt. This is the schematic of the latter version described above, where the op-amp circuits are on the same board as the current shunt, the mains connectors and the transformerless power supply, with the microcontroller on a separate board.

This is only the schematic of the op-amp board, not the microcontroller board, and it is possible that this schematic might contain some bugs, and it might need some tweaks – check carefully! I hope it is helpful for you :)

Now… I could write a comprehensive technical deconstruction and debunking of essentially the whole lot… but I’m only one person, with a finite amount of time.

I’ll get started, just for now, by taking a look at just one particular sentence of nonsense from Caldicott.

Ask questions. Seek the evidence. Ask everybody questions, and never take anybody’s word for it. Are factual statements backed up by evidence? Are quantitative statements backed up by measurements, calculations, or derivations? Can those measurements or derivations be described and reproduced? Read everything you possibly can, and you decide.

Do people like Caldicott have the right idea? Or do people like George Monbiot have the right idea – that beneath the FUD, rhetoric and hysteria, these people have absolutely no real evidence, facts, knowledge or technical literacy at all?

HC: “Well it won’t recover. These accidents go on forever because plutonium’s half-life is 24,400 years. It lasts for half a million years. Thirty tons of plutonium got out at Chernobyl.”

Thirty tons of plutonium “got out” at Chernobyl!?

Personally, that reads many thousands of counts per minute on my baloney detector.

Let’s follow Dr. Caldicott’s favourite piece of advice… let’s read her book. Surely, just like all of Caldicott’s other “references” usually are, it’s got to be “in my book”, right?

“Plutonium is so carcinogenic that the half-ton of plutonium released from the Chernobyl meltdown is theoretically enough to kill everyone on Earth with lung cancer 1100 times, if it were to be uniformly distributed into the lungs of every human being.”

(From Nuclear Power is Not The Answer).

Hmmmm. Curious. It looks like we’ve gone from “a half-ton” in the book to “thirty tons” in this recent interview. Well, so much for “you should read my book… it’s all in the book!”

(By the way… that “kill everyone on Earth with lung cancer 1100 times…” bit is complete baloney. But that’s a story for another day.)

Reactor-grade plutonium typically consists of approximately 1.3% ²³⁸Pu, which has a half-life of 87.7 years and a specific activity of 634 GBq/g, 56.7% ²³⁹Pu, which has a half-life of 24,110 years and a commensurately far smaller specific activity of 2.3 GBq/g, 23.2% of ²⁴⁰Pu, with a half-life of 6564 years and a specific activity of 8.40 GBq/g, 13.9 % of ²⁴¹Pu, with a half-life of 14.35 years and a specific activity of 3.84 TBq/g, and 4.9% of ²⁴²Pu, with a half-life of 373,300 years and a specific activity of 145 MBq/g.

Taking the weighted sum of all the above, we find that the overall specific activity of reactor-grade plutonium is 545.3 GBq/g, predominantly due to the ²⁴¹Pu and the ²³⁸Pu content.

(Reactor-grade plutonium is considerably more radioactive than weapons-grade plutonium, due to the presence of substantial concentrations of these relatively unstable, high-activity plutonium nuclides. Weapons-grade plutonium is almost entirely ²³⁹Pu, which despite being a good fissile fuel, is more stable and less radioactive. The radiological heat output of ²³⁸Pu, gamma-radiation (from the ²⁴¹Am daughter of ²⁴¹Pu) and the high rate of neutron emission from the spontaneous fission of ²⁴⁰Pu all make these nuclides extremely deleterious and undesirable in nuclear weapon design and engineering.)

The best value determined based on the available data for the quantity of plutonium (a reactor-grade cocktail of different plutonium nuclides) released at Chernobyl is, as published in the reports of the Chernobyl Forum, 3 PBq (3×10¹⁵ Bq).

The approximate total mass, based on the best available data, of plutonium released into the environment at Chernobyl is 3 PBq divided by 545.3 GBq/g.

As the British physicist David Mackay put it, I’m not trying to be pro-nuclear. I’m just pro-arithmetic.

It’s 5.5 kilograms.

Incidentally, that’s a very small amount of plutonium compared to the amount of plutonium that has been dispersed around the environment from half a century of nuclear weapons testing. 5.5 kilograms of plutonium is, approximately, the amount of plutonium in the pit of a single nuclear weapon. A single zero-yield “fizzle” of a nuclear weapon with no fission, or a zero-yield one-point-implosion safety test, or the accidental HE explosion (without proper implosion of the primary, as in the Palomares and Thule accidents) of a single nuclear weapon will disperse a roughly comparable mass of plutonium into the environment. (But less radioactivity, since weapons-grade Pu is less radioactive than reactor-grade Pu.)

So, Caldicott has gone from exaggerating the true number by a factor of approximately 100 to exaggerating the true number by a factor of approximately 6000.

Anyway… let’s just step back a minute. 30 tons of plutonium released at Chernobyl? Let’s apply what scientists, engineers and technologists sometimes refer to as the “reasonableness test” or the “smell test”. Can you quickly “smell” the data and determine if it is roughly plausible or not?

The total mass of uranium dioxide fuel in the fuel assemblies of a fully fueled RBMK reactor is about 180 tonnes. That’s about 159 tonnes of uranium, if you take off the mass of the oxygen in the uranium dioxide. When LEU fuel is irradiated at a typical burnup in a nuclear power reactor, about one percent of the mass of the original uranium ends up as transuranic actinides, mostly plutonium, by the time the fuel is removed. So, that’s a total plutonium inventory in the Chernobyl reactor of approximately 1.6 tonnes.

So, if we make a conservative, pessimistic and entirely unrealistic assumption that 100% of the plutonium inventory in the nuclear fuel was entirely vaporised and released into the environment during the Chernobyl accident, that would be 1.6 tonnes of plutonium released to the environment. (In reality, that fraction was something more like 0.34% of the total inventory of plutonium within the irradiated uranium dioxide fuel.)

So, does “thirty tons” pass the smell test? Not by a long shot.

It doesn’t matter what you do or don’t do to the reactors or the used fuel, or what condition they’re in – at the end of the day, the radiation dose to the public is how we measure the effect of this incident on the public and its potential for harm.

To be honest, I’m really not concerned much with what the dose rates are in the plant itself.

The men and women who work there understand dose rates and health physics quite well. They routinely work in areas of elevated above-background dose, and they know how to work safely in those environments. They understand how to measure and quantify the radiation field in the working environment, and the accumulated doses that they’re personally receiving.

They understand how to manage shielding, exposure time, radiation measurement and dosimetry in order to get the work done safely and effectively.

Even with abnormally significantly elevated radiation fields in some areas as a result of these incidents, they still know how to work safely. If the radiation dose rate in some particular area is so highly elevated that it cannot be entered safely for any length of time at all, then they won’t be entering it.

It’s pointless to scare the public with elevated on-site dose rate measurements. They’re not working on the site. Leave that for the people with health physics training. I’m much more interested in off-site dose rate measurements, personally, as those are the measurements that are actually of relevance to the public.

Off-site radiological dose rates

The KEK accelerator physics complex in Tsukuba (165 km from Fukushima) has a webpage showing their real-time measurements of the environmental gamma dose rate (counted with a GM tube). They’re currently measuring 0.17 μSv/h, at the time I write this, which has been fairly constant over the last few days, except for a brief, narrow spike up to 0.6 μSv/h, which they observed on the 16th.

(NOTE: Just to avoid any ambiguity, “m” means milli, 10^-3. “μ”, or “u” if you don’t have access to the Greek alphabet in your software, means micro, 10^-6. And when I say milli I’m quite sure I mean milli, and when I say micro I’m quite sure I mean micro. Because these terms are important, personally, I make damn sure to get them right.)

Each year, a resident of the United States receives an average total dose from background radiation of about 3.1 mSv. This is the radiation dose from natural background sources; from natural radioactivity in the Earth, and cosmic rays from space. That’s equal to 0.354 μSv/h.

In practice, the average dose that a person receives each year in the United States is significantly higher than that natural background dose, about twice that, once you’ve added on the dose from medical imaging and things like that.

The radiation dose rate being measured in Tsukuba right now, after the Fukushima accident, is less than half of the average natural background radiation dose rate that a person receives in the United States. This includes all sources of radiation in Tsukuba, including natural geological radioactivity, cosmic radiation, and any radioactivity released at Fukushima, as well as any ionising radiation from the particle accelerators at KEK, which is what these sensors are actually intended to monitor.

That brief, narrow spike seen in the radiation field measured at KEK doesn’t really concern me. The radiation dose you’ll receive if you hang around in that area for an extended period of time is the area under the graph – the integral – over that period of time. For such a short, sharp spike, the overall potential dose is still quite small.

In order to quantify the potential harm from a significant release of radioactivity, it would make more sense to “filter” that dose-rate data from the detector as a rolling average, to make it simpler to make a more straightforward interpretation of the potential to receive any significant radiation dose.

KEK is also measuring the concentration of ¹³¹I and the short-lived fission product ¹³²Te in the atmosphere and reporting regular updates to this data online. The concentrations we’re looking at here are extremely small – on the order of 10 microbecquerels per cubic centimeter – but they are concentrations which they are able to accurately measure at KEK, using a high-volume air sampler and a high-purity germanium gamma-ray spectrometer.

This site gives us a continually updated log of the gamma-ray dose rate in Tokyo.

The environmental gamma-ray dose rate measured in Tokyo between 11 pm and 12 am, averaged across one hour, on March 18th, was 0.0471 μSv/h. This radiological monitor in Tokyo returned its highest reading yet on the 16th, from 05:00 to 05:59, at a dose rate of 0.143 μSv/h.

So, that most recent figure from Tokyo is 13% of the average natural background radiation dose rate in the United States. One banana dose is something like 0.1 μSv, so what we’re measuring in Tokyo at the moment comes in at just under 0.5 banana per hour. (One banana per hour, and you’re going to triple that dose rate.)

The highest figured measured at all in recent days, 0.143 μSv/h, is equal to 40% of the average natural background in the United States.

The radiation level in Saitama, outside Tokyo, is also being recorded and charted on the web. As of 21:00 on the 18th of March, they report a dose rate of 0.058 μSv/h. The maximum value reported, at 1 am on March 17, is 0.067 μSv/h. These figures are 16% and 19% of US natural background, respectively.

Two things are apparent from this data.

(b) The background ionising receive dose rate that people receive across Japan has not been elevated significantly at all, at least outside the immediate vicinity of the plant, as a result of the Fukushima damage.

There’s an extremely interesting, valuable plot of dose-rate measurements on this page. Units are μSv/h.

(ASIDE: If you can’t read Japanese – I can’t – a little bit of Google’s automatic translation goes a long way in helping you sort through this important data.)

If we look at the 5 monitoring sites closest to the 30 km radius marked on the chart, we see that the last three measurements marked on the chart, for each of those sites are 52, 52, 52, 140, 140, 150, 40, 45, 45, 8.5, 9.0, 8.7, 1.6, 1.6, and 2.0 μSv/h.

This tells us that there is detectable radioactivity which is moving in a narrow plume in the atmosphere – it is not distributed out isotropically, which is indeed exactly what you would expect from thinking about the meteorology.

This chart of compiled radiation measurements also tells us a very similar story.

At that 30 km radius, the average dose rate from that monitoring station which reports the high outlier values – the one corresponding to the location of the plume – is 143 μSv/h.

I wonder what radionuclides are present in that plume? The presence of ¹³¹I, ¹³²Te and ¹³³Xe would tell us that this radioactivity has come from a reactor, the absence of these short-lived fission products will tell us it has come from used fuel in the pool. A little bit of gamma spectroscopy, and we would have the answers.

The presence of these radionuclides as measured at KEK confirms that at least a tiny bit of radioactivity has been released from the reactors themselves.

That’s fairly high, but it’s not obviously high enough to hurt people. If you stood in the location of that plume for an entire week, you would receive 24 mSv over the course of a week – which is a dose figure which would be consistent with a relatively-high-dose nuclear imaging procedure using something like ²⁰¹Tl to make an image of a tumor, or something like that.

If we remove those three outliers corresponding to the plume location from the above set of numbers and we take the mean of the remaining values, this gives us a rough idea of the mean dose rate elsewhere along the 30 km radius, outside the location where the source term of radioactivity is passing in a plume of wind. That mean value, then, is 26 μSv/h.

If you were standing in that radiation field, 26 μSv/h, for five hours per day every day for a year, you would reach a total annual dose of 47 mSv, which is just above the allowed occupational radiation dose – above natural and non-occupational background – of 50 mSv per year, for people working around radioactivity, such as nuclear power plant employees. (This is the limit set in the United States by the NRC; I’m not sure what the corresponding dose limit is in Japan, but it will be something loosely similar.)

But it’s well worth remembering that that radioactivity that is present now, in very low levels, will not be sticking around for a whole year. It is dispersing rapidly, and it drops away exponentially as you move away from the Fukushima site. As we move further out from the 30 km radius marked on that map, the dose rates recorded are all at harmless levels, consistent with background radiation dose rates experienced by people in the United States and elsewhere across the world.

In Ramsar, Iran, the natural background radiation dose rate is unusually high, at 260 mSv per year in some places. That is 30 μSv/h, which is higher than the mean value of about 26 μSv/h measured at these monitoring stations 30 km west of Fukushima, as I described above.

The people of Ramsar experience a background radiation dose significantly above that which most other people across the world experience – but they do not seem to experience any ill health effects at all from this.

I hope all the above helps to put these dose rates in context.

Used-fuel radiochemistry

The composition of the radionuclides that are responsible for most of the radioactivity in used nuclear fuel that has been stored in a cooling pool for a few months is very different than in nuclear fuel in a reactor that is operating, or has just been shut down.

Fuel straight out of an operating reactor contains a number of rather short lived, rather high specific activity fission product radionuclides which are of the largest health physics significance in the time immediately following severe reactor accidents.

Some of these short-lived fission products include iodine-131, xenon-133, xenon-135, tellurium-131, tellurium-132, and ruthenium-105. These short-lived fission products were very significant contributions to people’s radiation doses in the environment around Chernobyl in the time immediately following the Chernobyl disaster, for example, when they were dispersed from “hot” nuclear fuel from the reactor.

However, they are not present to any significant level in stored nuclear fuel, because they decay away relatively fast, and they cannot contribute any significant source term into the environment in some sort of accident scenario involving used nuclear fuel which has been stored for a month or three post-defueling.

So, what radionuclides are present in stored fuel? The main ones of interest here are the longer-lived fission products. ¹³⁷Cs, ⁸⁵Kr and ⁹⁰Sr are the most significant ones. Of these, ⁸⁵Kr is a gas, and has the most potential to be readily released from the fuel into the atmosphere. ¹³⁷Cs accounts for most of the radioactivity of the used nuclear fuel, and it is usually the most feared radionuclide in the used fuel inventory, in terms of the potential source term released from an accident with a used-fuel pool.

Fuel-pool water evaporation

“In this house, we obey the laws of thermodynamics!”
— Homer Simpson

With the used fuel heating the water in the Unit 4 fuel transfer pool, how long will it take for all the water to be boiled away? It is actually possible to know this, without any real speculation. The physics is pretty simple.

We know that there is a full core-load of used fuel in the Unit 4 defueling pool, which was put there after the reactor was shut down for inspection on November 30.

(Assumption I’ve made here which may possibly be wrong: That that single core-load of fuel is the only fuel in the pool. Is there additional fuel in the pool? If there is, somebody needs to tell me how much and how long it has been cooling for, so we can re-run the numbers – or you can take what I’ve explained here and re-calculate it for yourself.)

That fuel has been cooling for the last 3.5 months, or approximately 9.2 * 10⁶ s.

After this time, the radiothermal power output of the used fuel is small. Looking at this decay heat chart, we read that the decay heat is approximately 2 MWt. However, this chart is for a power reactor with a thermal power rating of 3000 MW. (And I’ve done a not-really-precise job of eyeballing that chart.)

But the Fukushima-I Unit 4 reactor has an electrical power capacity of 784 MW; that’s about 2352 MW thermal. So, we need to scale back the above figure commensurately; it’s approximately 1.6 MWt, from the entire core load of fuel.

The latent heat of vaporisation of water, at 100 C, is 2260 kJ/kg. (Let’s assume, conservatively, that the water in the pool is boiling; it’s at 100 degrees C, and the only route of energy dissipation from the system is through vaporisation of the water. This also assumes that none of the energy released is stored in the water by means of a rise in the water’s temperature, because it’s already at boiling point, and that there is no functioning mechanism for otherwise cooling that water.)

1.6 MW / (2260 kJ/kg * 1000 g/L) = 0.7 litres per second. 61 cubic meters per day.

The used fuel pool at Vermont Yankee, which is also a GE BWR-4, is 40 feet long, 26 feet wide and 39 feet deep, and is normally filled with 35,000 cubic feet of water. I will make an assumption that the Fukushima I Unit 4 used fuel pool has the same dimensions.

The level of water in the used fuel pool is normally 16 feet above the top of the fuel assemblies. With the water level evaporating at the rate described above, the water level will drop by 2 feet per day.

Uncovery of the fuel assemblies will take eight days. (Beginning from the point where the water level reached boiling point, after active cooling ceased.)

(Working assumption which you may subject to some skepticism: That there is no form of leakage or other water loss pathway from the used fuel pool.)

It seems plausible that a fire hose or something can be used to add water to the pool at a rate equal to this loss rate of 700 ml per second. The Chernobyl-style helicopter drops seem like overkill, and they are doing an effective job of whipping up “Chernobyl again” fear, and images that the newspapers are having a field day with.

In the wake of the Fukushima incident, it has really helped me to understand what happened in 1979 at Three Mile Island; with nonsense all throughout the media, and FUD, and panic spreading, with good information almost impossible to find, and with the over-abundance of bad information leading to hysteria.

But this is the first time it has happened to the Facebook and Twitter generation; I’ve yet to determine whether that’s a good thing or a bad thing. We need to keep working hard to make sure it’s a net benefit for the good information, not the dodgy information.

Design Basis Godzilla

Why wasn’t the earthquake design basis set high enough?, some people ask. What if the next earthquake is magnitude 10? Magnitude 12? Magnitude 20? But where does it stop?
Where do you set the design basis? What if the reactor is attacked by Godzilla?

No matter where you set the design basis, you will always exceed it one day, eventually. And when you do, the anti-nuclearists will complain that the design basis is not set high enough.

There is always some really extreme, really catastrophic situation that you can imagine, but what is its probability in any given year?

It’s all about Probabilistic Risk Assessment.

You know what the average probability per year of a magnitude 8 earthquake in the area is, what the probability of a magnitude 9 earthquake is, what the probability of a large tsunami is, etc. I’m not sure what the probability of Godzilla attack is. Need to ask an expert. Somebody get Matthew Broderick in here; he’s the guy to ask.

And you decide what the acceptable probability of a severe core damage (“meltdown”) incident (which won’t hurt anybody but will probably write off the reactor, like TMI) for the nuclear reactor is in any given year – let’s suppose it is decided that one such failure per 50,000 reactor-years of operation is acceptable. Then, with that in mind, you design the degree of seismic hardening and safety engineering for the nuclear power plant so that you hit that target.

It is the safest way to do it which is actually realistic. You can’t say that it’s absolutely 100% resistant to any hypothetical scenario of destruction that you can imagine, because there is always something that you can imagine that is more destructive.

Seawater injection into the primary containment of Unit 1 – not into the reactor vessel itself

On 12 March, TEPCO announced that they planned to cool the Unit 1 reactor with seawater, adding boric acid to the water as a nuclear poison, to prevent any possibility of unintended criticality. The injection of fresh water and seawater into the primary containment vessel through a fire-extinguishing system line commenced on the 13th of March.

This is not an injection of seawater into any part of the nuclear reactor or the Nuclear Steam Supply System itself. It is an injection of seawater into the containment structure surrounding the reactor pressure vessel.

These reports confirm my earlier prediction that they were not talking about actually putting seawater into the nuclear steam supply system, despite the lack of any previous clear, sensible announcements in the press to this effect.

Given that the reactor pressure vessel is completely intact and the control rods were fully inserted, normally, by the Reactor Protection System at the first sign of the earthquake, and the reactor is in a completely subcritical configuration, adding additional nuclear poisons (i.e. boron) to the system is seriously overkill; there is no pressing reason why it’s really necessary, other than to be extra conservative.

ASIDE: “Going critical” isn’t some kind of catastrophe, it’s what a fission reactor is designed to do, which it normally does under normal operating conditions.

Status of Unit 2

Over March 14 and March 15, Unit 2 has been in a similar low-coolant condition to Unit 1, with coolant water being boiled off through the torus, and with the Emergency Core Cooling Systems, such as the Low Pressure Coolant Injection system, operating.

Some kind of explosion was detected in the Unit 2 reactor building on the 15th of March, in a different area to the hydrogen ignition in Unit 1 and Unit 3. This hydrogen explosion was possibly within or close to the torus (the toroidal overpressure-supression tank), which is below grade, near the bottom of the reactor building.

Although this explosion may have damaged the torus, it has not damaged any of the real containment structures, such as the primary containment structure, the drywell, the reactor pressure vessel, or the outer reactor building.

There are valves that can be closed to isolate the vents that connect the torus to the drywell. Therefore, even if the torus was damaged or breached, it can be isolated from the drywell, meaning that there’s no real pathway between the drywell (the primary containment vessel) and the reactor building, through the damaged torus.

That means that your key layers of containment are all still intact.

Status of Unit 3, and its MOX fuel

There has also apparently been a loss of some of the normal reactor coolant systems in Unit 3, and the precautionary injection of borated water into the primary containment vessel, as at Unit 1, also began on March 13 and continued through to March 14.

As with Unit 1, seawater has been being injected, in addition to fresh water, but it is into the drywell – not into the reactor vessel. There has also been a hydrogen explosion in the rooftop refueling crane space above the reactor building at Unit 3 – very similar to the explosion in the same area at Unit 1.

Fukushima I Unit 3 is currently fueled with MOX fuel, containing recycled plutonium dioxide in its initial fuel load, in addition to low-enriched uranium dioxide.

You might see some slightly different characteristics in this MOX fuel with regards to things like the void reactivity coefficient compared to ordinary LEU fuel, because the fission cross section, plotted as a function of the incident neutron energy, will look slightly different for ²³⁹Pu compared to ²³⁵U. So, if the cross-section doesn’t drop away as rapidly as you increase the incident neutron energy, for ²³⁹Pu as compared to ²³⁵U, then you might see a negative void reactivity coefficient that isn’t quite as strong. Stuff like that.

But in practice there’s no evidence that we’re seeing any neutronic behavior or thermal behavior in the fuel that is any different, in practice, as far as the emergency situation is concerned, at Unit 3 as compared to Unit 1.

There is no significant difference in the composition of the radioactive fission products, or in their potential to be released to the environment, in the MOX fuel compared to the LEU fuel in the other reactors.

What we will see, however, is a lot of “plutonium phobia” in the media with regards to this MOX fuel, which is completely independent of any rational, fact-based discussion of the science and engineering. There’s this idea in the popular consciousness that plutonium is some mysterious, almost mythical, horrifying and terrible thing, and you’re going to turn inside out and die if you ever even look at it the wrong way.

But plutonium isn’t something that’s been spawned out of the arse of a demon or something, it’s actually just another element. It’s radioactive, but less radioactive than most of the fission products formed in a nuclear reactor.

The potentially harmful radioactive materials that might be released into the environment from a LOCA scenario such as the Fukushima I-1 situation, or the Three Mile Island accident, are the radionuclides of the high-volatility fission products, such as radioisotopes of xenon, krypton, and to a lesser extent, radioisotopes of iodine. Tritium is also on the list of key radionuclides that are present within the coolant water and are volatile enougn to potentially be relased, but its relatively low specific activity compared to the hort-lived fission products and its very low beta decay energy means that tritium is not significant in terms of its potential for harm.

Uranium and plutonium, for example, are heavy metals, they’re not soluble in the coolant water, and they are nowhere near as volatile as these gaseous and volatile fission products. This means that there’s no way that these elements can escape out into the primary coolant and eventually find their way out into the reactor building atmosphere, as these more volatile elements can. This means that there’s no way that they will be released out into the atmosphere in the same way that very small quantities of tritium and gaseous fission products may be be released. Even if they could be, uranium has a negligible specific activity, and it is a negligible source of radiation dose – and plutonium is not extremely radioactive either; being less radioactive than the majority of the fission products.

The Used Nuclear Fuel

There is a small fuel transfer pool in the reactor building at each of these GE BWRs, near the top of the reactor pressure vessel, that is used for the temporary transfer of used nuclear fuel during refueling. However, the longer-term storage of the used nuclear fuel is done in a pool elsewhere on the site. Those storage pools, outside the reactor buildings, are seismically hardened and defended-in-depth, just like the reactors themselves, and there are no indications of any problems with them. Since there was no refueling going on at the damaged reactors at the time of the earthquake, there is little or no fuel in the fuel transfer pools.

That’s my understanding of the situation, anyway.

Is there actually some used nuclear fuel stored in the Unit 4 fuel transfer pool at the time of the earthquake? I don’t believe there is much, if any.

Some media reports suggest there might be; but I’m extremely distrusting of the incomplete, garbled information being filtered out through the mainstream media on this whole issue – which is why I’m trying to patiently take information in from a diverse range of sources, and to patiently, carefully and skeptically piece it together into a self-consistent picture of what we actually know about the situation, based on my knowledge of the physics and my limited but partial knowledge of nuclear engineering.

If there is some fuel stored in the transfer pool, we need to know a few things about it.

How much is stored there? How long has it been there for? When was the reactor that it was taken from powered down, prior to fuel unloading? The decay heat of used nuclear fuel drops off exponentially following reactor shutdown, as the short-lived fission products that account for most of the radioactivity rapidly and exponentially decay. After a short period of cooling, the used fuel would not be significantly heating the pool water, nor would it require active cooling.

Immediately after fission is shut down in an operating nuclear power reactor, the power output from nuclear fission ceases. However, the nuclear fuel is still extremely radioactive, at least initially, due to the presence of very short lived fission products with very high specific activities. But that decay heat output falls off exponentially over the coming hours, as the very short lived fission products decay.

Thanks to Kirk over at Energy From Thorium for preparing this excellent chart, which I’ve borrowed from the EFT blog – all credit to him. Go and check out Energy From Thorium – it’s an excellent site.

Initially, the decay heat power level – the heat emitted by the radioactive decay of the radionuclides present in the used nuclear fuel, in a subcritical configuration – is approximately 5% of the fission reactor’s normal thermal power output.

The Fukushima I Unit 1 reactor has a nameplate capacity of 460 MW_e. At a thermodynamic efficiency of approximately 35%, the reactor has a thermal power output of approximately 1.31 gigawatts.

Therefore, immediately following reactor shutdown, the decay heat power output from the reactor’s nuclear fuel will be about 5% of that, or about 66 megawatts.

Over the last 5 days since the shutdown of the reactor, however, that power output has been dropping off exponentially, and it is now somewhere down around 6 megawatts of thermal power.

So, how much fuel is presently stored in the Unit 4 fuel transfer pool, if any? I might take an educated guess and say that it’s about one third of one full core loading, since about one third of the core fuel is replaced during one refueling. Let’s take a conservative guess, and say that it has been out of the Unit 4 reactor (which was already offline for maintenance prior to the earthquake) for about a week, at the least. It’s probably longer than that.

If those assumptions are valid, then the radiothermal power output from the used fuel in the fuel transfer pool will be only about one or two megawatts at the present time, and still decaying of course.

We know what the thermal power input from the fuel is, what the volume of water in the pool is, what the dimensions of the pool are, what the nominal temperature of the pool is, and what the specific heat capacity of water and the latent heat of evaporation of water are.

Therefore, we know exactly what the temperature of the water of the pool is going to do, and exactly what the volume of water in the pool is going to do. There is no need for guessing or speculation.

The nuclear fuel is made up of pellets of uranium oxide clad in tubes of Zircaloy. (Zircaloy is the trade name of the alloy in question, which is almost pure zirconium.) UO₂ is a refractory metal oxide with a very high melting point – it is not flammable or combustible in any way at all. It does not burn. In theory, the zirconium metal cladding could burn, if it was removed from all cooling water, exposed to an oxygen-containing atmosphere, and heated to an extremely high temperature. However, it is almost impossible to burn a piece of zirconium in this fashion in practice.

The rate at which the water in the fuel transfer pool would evaporate, in the absence of active cooling system functionality and in the presence of the fuel’s remaining thermal power output, is not more than a few percent of the water volume per day.

Given that there is approximately 16 feet or more of water above the surface of the fuel assemblies in the pool under normal conditions, it will take many days to weeks for this water to evaporate to the point where the fuel is potentially exposed above the surface, assuming that no additional water was added by any other mechanism.

The outermost layer of the multiple layers of containment surrounding the reactors – the cuboid-shaped reactor buildings – have walls and a roof made of solid concrete. On top of the concrete reactor buildings, however, there is an additional part of the structure – it is not made of concrete, but it is made of steel, with steel sheets over a steel frame. This steel building on top of the reactor building houses the fuel transfer crane, and it is built on top of the concrete roof of the reactor building.

That is, the part of the structure above the concrete shield plug and the refueling platform at the top of the concrete reactor building, as shown on the diagrams below.

It is this relatively weak steel structure on top of the reactor building, which is not really part of the reactor building proper, which has been blown out by a hydrogen explosion at Fukushima I Unit 1, as I described in my previous post, and apparently at one of the other Fukushima I reactor buildings as well.

The concrete roof of the reactor building proper – which is not where the hydrogen explosion occurred – is built over the top of the temporary fuel transfer pool; the pool is protected within the concrete reactor building.

Unit 4 Explosion and Fire

On 14 March, an explosion of hydrogen gas in the steel-framed structure atop the Unit 4 reactor building occurred, similar to the earlier explosion at Unit 1. As was the case at Unit 1, there was no real damage to the reactor building proper or any of the other layers of containment within.

On March 14, a fire was reported in the Unit 4 reactor building. The fire was caused by a leak of lubricating oil in the mechanical drivetrain that drives the reactor’s recirculation water pumps. Contrary to inaccurate information which was spread widely in the media and the social-media grapevine, this small fire had absolutely nothing to do with the used-fuel transfer pool or any fuel which may have been in it. The fire was extinguished shortly after.

Detection of radioactivity off the coast

It has been reported that the crew of the Nimitz-class carrier USS Ronald Reagan, off the Japanese coast, has detected the presence of radioactivity emitted from the Fukushima nuclear power plant. Since she’s nuclear powered, the vessel is obviously carrying sensitive instruments and sensors for detecting the presence of radioactivity around the ship, and obviously they’ve detected the trace levels of radioactivity that have escaped into the atmosphere.

It’s possible to measure the presence of small traces of radioactivity fairly easily with extremely high sensitivity. Just because we’re able to accurately measure it and quantify it at very low levels doesn’t mean we’re dealing with levels of radioactivity that are anywhere close to harmful levels.

Has anyone actually reported the actual quantitative presence of radioactivity that was detected aboard the ship?

The status of the Fukushima II plant

At the time of the March 11 earthquake, all four operating units at the Fukushima II station were automatically SCRAMmed by the Reactor Protection Systems.

Although there were some reports of disruptions to the Emergency Core Cooling Systems at Fukushima II, all the Emergency Core Cooling Systems at all the reactors were restored to operational status.

Between March 12 and March 14, every one of these reactors has been bought to a safe state of cold shutdown, where the decay heat from the nuclear fuel is being completely dissipated and the fuel is being kept cool, with no rise in fuel temperature.

“Is it true they have nuke stuff inside of them?”
“Radiothermal isotopes.”
“What happens if one gets busted open? Everyone gets all mutated?”
“If you ever find yourself in the presence of a destructive force powerful enough to decapsulate those isotopes,” Ng says, “radiation sickness will be the least of your worries.”

— Neal Stephenson, Snow Crash

At the present time, the people of Japan are struggling to deal with one of the most serious natural disasters anywhere in the world in recent recorded history. My thoughts are with them.

But I’ll get straight to the point. All right, what do we actually know about the effects of this disaster on Unit 1 at the Fukushima I Nuclear Power Station?

There are 53 operational nuclear power reactors in Japan today. Most of them were operating normally at the time of the recent earthquake, and continued to operate normally, since they were relatively far away from the earthquake’s epicenter. Some of them were offline at this time for routine scheduled maintenance or refueling. Several reactors, closer to the earthquake’s epicentre, experienced normal, automatic reactor trips (“SCRAMs”, in Western nuclear engineering parlance) controlled by the RPS (the Reactor Protection System), exactly as you would expect either in the presence of ground acceleration under earthquake conditions, or due to a loss of electricity grid connectivity to the plant (which is known as a Loss of Offsite Power, or LOOP, in nuclear power engineering parlance), which is a very likely event during a severe earthquake.

For the purposes of designing safe nuclear power plants, loss of offsite power is recognized as a relatively frequent, relatively high probability event. For the purposes of designing safe nuclear power plants, especially in Japan, it is recognized that the plant can be subjected to a severe earthquake – and on the Japanese coast, to tsunami surges as well.

And of those 53 power reactors, only one is behaving in a somewhat abnormal way in its shutdown state – Unit 1 at the Fukushima I Nuclear Power Station. (Fukushima Dai-Ichi, as opposed to Fukushima Dai-Ni, which is the Fukushima II plant.) The other 52 are completely normal, either operating or behaving as predicted in a shutdown state. To see that 52 out of 53 are behaving completely normally, and many are still operating normally, generating electricity on the grid, in the wake of one of the strongest earthquakes the world has ever seen in an industrialized area, shows you just how resilient nuclear power infrastructure is in response to natural forces like this.

UPDATE: At the time I wrote this, what I above was accurate and there were no concerns about the state of any of the power reactors other than just this one. It is now the case that there are issues at a couple of the power reactors, and the above is not strictly accurate. However, the phenomena and the systems I have described below in the context of Fukushima I Unit 1 are still relevant to understanding what is happening at the other reactors.

Meanwhile, oil refineries and natural gas plants are pretty much all going up in flames in Japan’s earthquake-affected areas.

The absolute worst case scenario that we could potentially be looking at here is partial melting damage to the nuclear fuel – similar to the Three Mile Island accident. This will not harm any people or harm the environment, but it will have serious financial and political costs for TEPCO, it may write off the reactor, and it will be a significant political and rhetorical advantage for anti-nuclear activism and FUD.

If there is any real environmental damage to come out of this accident, it will come as a result of increased use of coal and fossil fuels instead of nuclear energy.

Fukushima I-1 is a General Electric BWR (Boiling Water Reactor), with a (relatively small) nameplate capacity of 460 MW. It first achieved criticality in October 1970, only 3 years after its construction commenced in 1967.

Fukushima I Unit 4, Unit 5 and Unit 6 were all already offline for maintenance and fueling operations at the time of the earthquake, and Units 1, 2 and 3 were shutdown automatically at the time of the earthquake, either by the RPS seismic sensors or by the RPS relays opening when off-site power was disrupted, completely as intended.

The control rods are all already fully driven into the reactors, and the reactors are fully subcritical. The systems are not even close to criticality and cannot reach criticality – or any measure of supracriticality – at all. This has been the case at all times following the initial RPS trips and control rod insertion at the time of the earthquake.

However, for a limited period of time following reactor shutdown, cooling of the reactor core still has to be maintained, to dissipate the decay heat of the short-lived fission products in the nuclear fuel. And that cooling, and how it is maintained, or not maintained, in the absence of offsite power, is at the root of our discussion of all this fuss at Fukushima.

Even with the reactor in a subcritical configuration, with control rods inserted, if the reactor core coolant level drops excessively and it is not replenished, over the course of the next 48 hours or so following reactor shutdown, the fuel can eventually heat up excessively from its decay heat, leading to core damage – partial melting of the fuel, which will be very difficult and costly to fix. This is not significantly dangerous for the people and the environment around the nuclear reactor. This worst-case scenario, damage to the fuel in the reactor core, is not dissimilar to the damage to the Three Mile Island Unit 2 PWR in the United States in 1979; although it is worth noting here that the TMI reactors are Pressurized Water Reactors and the Fukushima reactors are BWRs.

As the name suggests, however, the decay heat will decay away fairly rapidly – and the fuel’s thermal power output will drop below levels which are potentially problematic in the absence of proper cooling after a few days. It will take a few days for the fuel rods to stabilise their own temperature, in the absence of active water cooling, as the short-lived fission products in the fuel which are generating the heat continue to decay. The reactor will be then in what is known as “cold shutdown”. At that point, only minimal coolant injection into the reactor will be required, and preparations can be began to remove the nuclear fuel from the core.

We’re probably already not far from reaching this point, chronologically, at Fukushima I-1. The decay heat from the fission products in the fuel has been decaying constantly for the last couple of days, ever since control rod insertion at the time of the earthquake. We’re now reaching the point, over the coming few days, where the risk of further potential core damage has passed.

Following LOOP, most of a reactor’s instrumentation and emergency systems are generally transferred to a backup auxiliary power supply provided on site by diesel generators, or to batteries in the case of some systems. However, it appears that these generators at Fukushima I-1 were damaged by the earthquake. The diesel generator appeared to start up correctly, and then it stopped abruptly about an hour later.

So, what happens to a BWR, in terms of its decay heat removal, when the reactor is tripped, and offsite power is offline, and the auxiliary electric power supply from the diesel generators is offline?

To find out, we need to take a closer look at the BWR. To start with, here are a few little diagrams that basically illustrate the architecture of a typical BWR of the kind we’re discussing here. Click through for the full-sized images.

Wikipedia has a surprisingly good page on the safety systems of a Boiling Water Reactor, and I think that’s a very good place to start. It’s not too technical, since it is Wikipedia, after all, but it’s impressively good basic technically literate material for a Wikipedia article.

Like all light-water reactors, the GE BWR has a negative void coefficient. In other words, as the proportion of steam to liquid water increases within the reactor, the moderation of neutrons within the reactor is decreased, since the lower-density steam is less effective as a moderator, and the reactor’s average neutron energy spectrum hardens a little, and this causes the reactor’s neutronic power output to decrease, since the reactor is operating with an enriched uranium fuel in a typical neutron energy spectrum where hardening the neutron energy spectrum will decrease the fission power output.

A sudden increase in steam pressure within the BWR (caused, for example, by the closing of the main steam isolation valve from the reactor) will cause a sudden increase in the proportion of liquid water to steam within the reactor, which will cause an increase in the reactor’s power output, due to the negative void coefficient. Such an event is known as a pressure transient.

The BWR is specifically designed to suppress such pressure transients, by safely venting the overpressure through safety relief valves to below the surface of a pool of liquid water within the containment. This toroidal-shaped tank, known as the torus, is shown on the drawings above. There are 11 safety overpressure relief valves on the older generation of BWRs such as the ones at Fukushima, and only a couple of them need to be opened in order to completely mitigate a pressure transient.

Although a pressure transient will cause a transient in the fission power output for a brief moment, the rapid actuation of the pressure relief valves will cause the pressure to drop off rapidly, and correspondingly, the neutronic power will rapidly drop off once the valves are opened, to a level far below nominal operating power.

There is an intrinsic physical relationship between temperature, pressure and fission power output in a light-water reactor, because of the void reactivity coefficient.

The Emergency Core Cooling System, the ECCS, of a light-water reactor is made up a set of many interrelated, redundant layers of different safety systems which are designed to protect the nuclear fuel within the reactor pressure vessel from overheating in the event of the loss of coolant level, by maintaining that coolant level. To understand what’s going on at Fukushima, it is good to have a basic understanding of what these different systems are.

The Emergency Core Cooling System(s)
————————————————————

The High Pressure Coolant Injection System (HPCI) is the first line of defense in the ECCS. The HPCI is designed to inject substantial quantities of water into the reactor while it is at high pressure, and to prevent the activation of the additional, redundant low-pressure “layers” of the ECCS. HPCI can deliver approximately 19,000 L/min to the core at any core pressure above 690 kPa (100 psi). This is usually enough to keep the water levels sufficiently high to avoid activating the low-pressure “layers” of ECCS except in a major contingency, such as a large break in the makeup water line. The HPCI necessarily operates at a high pressure because it injects water into the reactor at a high flow rate against the high pressure already within the reactor, without releasing that pressure.

It’s worth noting here that whilst the Fukushima reactor may be losing coolant level at a limited rate through steam venting through the pressure relief valves into the torus, there is no pipe break, no stuck-open valve, or any other serious large-scale LOCA scenario here with a serious rate of coolant loss, which is the kind of thing the ECCS is designed to safely compensate for.

The HPCI system is powered by steam from the reactor – its operation is not dependent on off-site power, or power from the diesel generators, or battery power. It is powered by the heat remaining in the reactor itself.

It is completely plausible that a turbine trip, with sudden closure of the main steam isolation valve (MSIV) between the reactor and the turbine hall, will cause a significant power transient in the reactor, for the reasons described above, and steam venting into the relief valves as a result of that transient will cause some loss of the coolant level. The HPCI system is more than adequate to make up the reactor water level in this scenario.

The next one of the redundant components of the ECCS is the Reactor Core Isolation Cooling System, or RCIC. RCIC is also one of the high-pressure coolant injection systems, capable of injecting approximately 2000 L/min of water into the reactor core. The RCIC is able to operate with no source of electric power other than battery power, and is capable of providing decay heat removal by itself in the event of a station blackout, where off-site power is lost and the backup power supply from the diesel generators also fails.

If the water level cannot be maintained with the HPCI and/or the RCIC, and the core water level is still falling below some present point even with these systems working full-bore, then the next systems in the stack of ECCS systems respond. If, for some reason such as a large-break LOCA, the water level cannot be maintained, we then move to looking at the next layers of redundancy in the ECCS – the depressurisation and low-pressure coolant injection systems.

For the low-pressure coolant injection components of the ECCS to operate, the pressure within the reactor must be reduced, by the depressurization system. The Automatic Depressurization System (ADS) is designed to activate in the event that the reactor pressure vessel is retaining pressure, but the water level cannot be maintained using high pressure cooling alone, and low pressure cooling must be initiated. When the ADS activates, it rapidly releases pressure from the reactor vessel in the form of steam, through pipes that are piped to below the water level in the the torus, which is designed to condense the steam released into it, bringing the reactor vessel pressure below 32 atmospheres, allowing the low pressure components of the ECCS to be activated.

The low-pressure ECCS systems have extremely large capacities compared to the high pressure systems and are powered by multiple different power sources. They will maintain any required water level, and in the event of a worst-case LOCA, such as a break of a large water pipe feeding into the reactor vessel below core level, which could potentially lead to temporary fuel rod “uncovery”, they will rapidly return the water level over the fuel in the core prior to the fuel heating to the point where core damage could occur.

The Low Pressure Core Spray System (LPCS) is the first of the low-pressure ECCS components, designed to suppress steam generated by a major contingency. As such, it prevents the reactor vessel pressure from re-increasing above the LPCI coolant injection pressure, 32 atmospheres. It activates while the pressure in the reactor is still below 32 atmospheres, and delivers approximately 48,000 L/min of water in a deluge from the top of the core.

The Low Pressure Coolant Injection System, LPCI, is the final piece of the ECCS, the “heavy artillery” in the ECCS. Consisting of 4 pumps driven by diesel engines, it is capable of injecting a mammoth 150,000 L/min of water into the core. Combined with the core spray system to keep steam pressure in the core sufficiently low, the LPCI can suppress all core-cooling contingencies by rapidly and completely flooding the core with coolant. One should also note that the diesel-engine driven pumps that run the LPCI are completely independent of off-site electrical grid power, they are independent of steam power being extracted from the reactor (unlike HPCI), and they are independent of the diesel generators that provide the backup electricity supply for the plant in the event of the loss of offsite power.

The Standby Liquid Control System, the SLCS, is used in the event of major contingencies as a last-ditch measure to prevent core damage. It is not intended ever to be used, as the RPS and ECCS are designed to respond to all contingencies, even if multiple components of those systems fail, but if a complete ECCS failure occurs, it could be the only thing capable of preventing core damage. The SLCS consists of a tank containing a large quantity of water loaded with soluble nuclear poisons (such as boron) protected by explosively-opened valves and redundant battery-operated pumps, allowing the injection of the water into the reactor against any pressure within it. This water is fully capable of dissipating heat from the nuclear fuel; and the nuclear poisons in this water will send the system fully subcritical even if, somehow, insertion of the control rods has completely failed (which is not the case at present for any of these Japanese nuclear power reactors.)

The SLCS is a system that is never meant to be activated unless all other measures have failed to maintain integrity of the nuclear fuel. In the older generation of existing BWRs its activation could cause sufficient damage to the plant (due to the salts used as neutronic poisons causing corrosion and contamination of the whole nuclear steam supply system) that it could make the reactor inoperable without a complete overhaul.

There is now talk of pumping seawater into the reactor building; although the information in the press on this subject seems to be vague and confused. There is very little good, unambiguous information out there. Are we talking about spraying seawater within the reactor building, in order to condense steam and reduce the temperature and pressure? That seems to make sense. Are we talking about spraying seawater within the drywell to help cool the reactor pressure vessel, and reduce temperatures within the drywell? That also makes sense.

Surely it wouldn’t make sense to actually inject seawater within the actual Nuclear Steam Supply System, would it? This would cause significant problems with regards to contamination and corrosion of the entire nuclear steam supply system, which would be difficult, time consuming and expensive to rectify. Why would this ever be considered, when the SLCS and the ECCS systems are designed to perform the same function, safely and reliably, under adverse emergency conditions, without ruining the reactor? I do not expect that this is actually what is being planned – but again, the information that is tricking out through the hysterical mass media is so bad, it’s hard to tell.

The Emergency Core Cooling Systems and the Design Basis Accident
——————————————————————————————————–

(I’ve borrowed most of the material for this example scenario illustrated here from here.)

The Design Basis Accident (DBA) for a nuclear power reactor is the most severe possible single accident that the designers of the plant and the regulatory authorities could realistically imagine, as a contingency which the operators of the plant must be able to handle. It is, also, by definition, the accident the safety systems of the reactor are designed to respond to successfully, even if it occurs when the reactor is in its most vulnerable state.

The DBA for the BWR consists of the total rupture of a large coolant pipe in the location that is considered to place the reactor in the most danger of harm – specifically, for the older generations of existing BWRs, such as the Fukushima BWRs, the DBA consists of a “guillotine break” in the coolant loop of one of the recirculation jet pumps, which is substantially below the core waterline, and as such, has the makings of a very serious Loss of Coolant Accident or LOCA. The DBA scenario combines this large-scale loss of coolant with a simultaneous loss of feedwater to make up for the water boiled in the reactor (a loss of proper feedwater or LOFW), combined with a simultaneous collapse of the regional power grid, resulting in a loss of power to certain reactor emergency systems, or LOOP.

The BWR is designed to shrug this accident off without core damage.

The Design Basis Accident is not directly relevant to what happened to the reactor at Fukushima, but it is a good example to use to illustrate how the various different layers of the ECCS and the Reactor Protection System work under severe accident conditions, which is important background to a good understanding of what happened at Fukushima.

The immediate result of such a large-scale pipe break (we’ll call this time T+0) would be a pressurized stream of water well above boiling point shooting out of the broken pipe into the drywell, which is at atmospheric pressure. As this water stream flashes into steam, the pressure sensors within the drywell will report a pressure increase to the Reactor Protection System, within no more than 300 milliseconds; that is, by T+0.3. The RPS will interpret this pressure increase signal as the sign of a break in a pipe within the drywell. As a result, the RPS immediately initiates a full SCRAM, closes the Main Steam Isolation Valve (isolating the containment building), trips the turbines, attempts to spin up RCIC and HPCI using the residual steam, and starts the diesel-driven pumps for LPCI and the core spray.

Now, let’s assume that the LOOP occurs at this time, at T+0.5. The RPS is on an uninterruptable power supply, so it continues to function. The RPS immediately detects the loss of offsite power, however, and already enters a fully defensive state and trips the reactor and the turbine, if it has not already. Within less than a second from power outage, auxiliary batteries and compressed air supplies are starting the emergency diesel generators. Power will be restored by T+25 seconds.

(Remember that at Fukushima I, the backup diesel generator failed shortly after, but there was no real pipe break or LOCA. But never mind that, in the scenario we’re looking at here. In any case, remember that many of the ECCS sub-systems have different, redundant energy sources.)

Due to the rapid escape of coolant from the reactor core, HPCI and RCIC will fail rapidly due to loss of steam pressure, but this is immaterial, as the 2,000 L/min flow rate of RCIC available after T+5 is insufficient to maintain the water level; nor would the 19,000 L/min flow of HPCI, available at T+10, be enough to maintain the water level, even if it could work without steam, in the event of such a serious LOCA. At T+10, the temperature of the reactor core, at approximately 285 °C at this point, begins to rise as enough coolant has been lost from the core that voids begin to form in the coolant between the fuel rods and they begin to heat rapidly. By T+12 seconds from the initial LOCA, fuel rod uncovery begins. At approximately T+18, parts of the rods have reached 540 °C.

At T+40, the core temperature is at 650 °C and rising steadily; the LPCI and the pressure-regulating core spray kick in and begin deluging the steam above the core, and then the core itself. A large amount of hot steam still trapped above and within the core has to be knocked down first, or the water will be flashed to steam prior to it hitting the fuel. This happens after a few seconds, as the approximately 200,000 L/min of water these systems release begin to cool first the top of the core, with the LPCI deluging the fuel rods, and the core spray suppressing the generated steam until at approximately T+100 seconds, when all of the fuel is now subject to this deluge and the last remaining hot spots at the bottom of the core are now being cooled.

The peak temperature that is attained in the fuel elements in this scenario, even with temporary uncovery of the fuel rods, is 900 °C, well below the maximum of 1200 °C which is acceptable before fuel damage begins, at the bottom of the core, which was the last hot spot to be cooled by the water deluge.

The core is now cooled rapidly and completely by the LPCI, and following cooling to a reasonable temperature, below that consistent with the generation of steam, the core spray is shut down and the LPCI flow rate is decreased to a level consistent with maintenance of a steady temperature of the fuel rods, which will drop over a period of days due to the decay in fission-product decay power output within the fuel. After a few days, decay heat will have sufficiently abated to the point that defueling of the reactor is able to commence. Following defueling, LPCI can be shut down completely.

The Explosion
———————

On March 12, there was an explosion near the Fukushima I-1 reactor building. What happened?

There are about 5 different layers of containment which exist, in a power reactor reactor like the ones at Fukushima, between the people outside and the potentially dangerous radioactive fission products within the nuclear fuel.

The fuel rods themselves are clad in tubes of zirconium alloy, and that represents one such layer. That nuclear fuel is inside the reactor pressure vessel, which is made of steel six inches thick, and that reactor vessel is the next such layer. The reactor pressure vessel is within the primary containment vessel, the drywell, which is made of steel one inch thick, and that represents the next such layer. The primary containment vessel is within the secondary containment structure, which is made of steel-reinforced, pre-stressed concrete between 4 and 8 feet thick. The reactor building which is built around the secondary containment structure is the last of these multiple layers of containment, and it is also made of steel-reinforced, pre-stressed concrete, between 30 cm and 1 m thick.

If every possible measure standing between safe operation of the plant and severe core damage and melting of the nuclear fuel fails, the containment can be sealed indefinitely, and it will prevent any significant release of radioactivity to the outside environment occurring under any circumstances.

Now, let’s look at some diagrams of these structures. Click-through for the full resolution images.

The outermost layer of the multiple layers of containment – the reactor building – has walls and a roof made of solid concrete, and it’s roughly cube-shaped.

On top of the concrete reactor building, however, there is an additional part of the structure – it is not made of concrete, but it is made of steel, with steel sheets over a steel frame. This steel building on top of the reactor building houses the fuel transfer crane, and it is built on top of the concrete roof of the reactor building. I’m referring to the part of the structure above the concrete shield plug and the refueling platform at the top of the concrete reactor building, as shown on the first of the diagrams above.

It is this relatively weak steel structure on top of the reactor building, which is not really part of the reactor building proper, which seems to have been blown out by a hydrogen explosion.

The explosion at Fukushima I-1 does not appear to have occurred within nor does it appear to have breached any of the fundamental layers of containment structure described above.

Now, an explosion has not occured as a result of a release of nuclear energy. That is a scenario that is simply outside the laws of reality. An explosion can be caused by one of two things; a chemical explosion, such as an ignition of a hydrogen-oxygen mixture, or a sudden release of stored gas or steam pressure.

It appears that the structure has probably been damaged as a result of a hydrogen explosion. It’s probable that excessive hydrogen generation within the reactor core, either radiolytically or chemically by reduction of water in the presence of the zirconium cladding at significantly elevated temperatures, has been vented into the torus, and as temperatures and pressures have began to rise within the torus steam pressure in the torus has been vented out into the reactor building surrounding the torus. From there, the hydrogen mixed with that steam and water vapor has risen, as hydrogen does, and worked its way through the reactor building, escaping at the top of the reactor building, and accumulating at the top, in the area around the fuel transfer crane. It appears that the accumulated hydrogen has then mixed with air and exploded.

Radiochemistry and radioactivity releases to the environment
——————————————————————————————

When a light-water reactor is operating, some of the oxygen-16 in the water is activated into radioactive nitrogen-16, by the ¹⁶O(n, p)¹⁶N reaction. ¹⁶N is very short lived, with a half-life of only 7 seconds, but its specific activity is correspondingly very high. When a BWR nuclear power station is operating, the entire nuclear steam supply system, including the turbine hall, is a radiological controlled area, due to the radioactivity from ¹⁶N. However, after reactor shutdown, the ¹⁶N decays very quickly, reducing the radiation dose around the turbines to negligible levels basically immediately. This is one key difference between a BWR power plant and a PWR power plant – since the secondary coolant loop that drives the turbine in a PWR is isolated from the reactor’s primary coolant by the steam generator, the secondary coolant is never radioactive during reactor operation.

There can also be very small amounts of other radionuclides formed within the reactor coolant, for example tritium, which is formed by the fission of boron used as a soluble reactivity shim in the reactor coolant (if you really want to know: neutron capture on ¹⁰B forms an excited state in ¹¹B, which splits apart into two ⁴He and 1 ³H nuclei. A similar reaction occurs beginning from ¹¹B, with the re-emission of one additional neutron in the breakup of the excited ¹²B nucleus), and ¹⁴C, which is formed from nitrogen compounds such as hydrazine which are added for pH control and oxygen scavenging in the reactor coolant.

If excessive pressure within the torus or within the primary containment vessel is vented out into the reactor building, and from there it is allowed to escape out into the atmosphere, then small amounts of these radionuclides may be released out into the atmosphere, which is a possible scenario we might be seeing at Fukushima.

(A quick note on terminology: Radiation is not a substance, and it cannot leak, nor contaminate a person. To speak of an escape of radiation from a nuclear power station is kind of like to speak of an accidental leak of light from a lightbulb factory. What we are talking about here is a possible release of radioactivity, of a substance containing a radioactive nuclide.)

We will know, over the next few days, exactly what the true situation is regarding the composition, and quantity, of any releases of radionuclides into the outside environment. It’s very easy to detect radioactivity, to measure it quantitatively with high precision, and to discriminate the presence of different radionuclides and identify them.

What I suspect we might see from some anti-nuclearists, however, is something that we saw after Three Mile Island, and something that we still see on rare occasions up to the present day in regards to TMI – the conspiracy theories.

Some people will probably try and claim that there were actually enormous releases of radioactivity into the environment and it was never really measured or documented – or that it was measured and known that there were huge releases of radioactivity into the environment at Fukushima, but there’s a big conspiracy by big bad unethical TEPCO or by the Big Bad Nuclear Industry in a more general sense (and by the evil government and the conspirators at the IAEA, and the armies of Big Nuclear Shill bloggers, of course!) to cover it all up! We saw this once or twice after TMI, and I think we’ll see it again from those who are truly devout believers in the absolute, unmoderated evil of the Big Bad Nuclear Industry.

Of course, that’s absolute nonsense, for exactly the same reasons that it’s nonsense in the context of TMI. You simply cannot ever, in any context, release a very large amount of radioactivity into the atmosphere and cover it up or keep it quiet.

Look at Chernobyl for example. The Soviets didn’t tell the West about it immediately – they didn’t even tell their own nuclear scientists. Soviet nuclear experts found about it when radiation sensors at nuclear research sites and nuclear power plants (eg. the Ignalina plant in what is now Lithuania) across the eastern USSR started going off, and the West found out about it when radiation sensors at Sweden’s Forsmark NPP and other Swedish nuclear engineering facilities started going off. (For more on this note regarding Chernobyl, see the excellent first chapter of Richard Rhodes’ excellent Arsenals of Folly.)

Nuclear power plants and other facilities that use radioactive materials are all over the place in our society, and they all have sensors and instruments to make sure everything is safe and radioactive contamination does not occur. If a Chernobyl-style event occurs, you will detect it at any such site. Any nearby NPP. Any nearby molecular biology lab working with radiolabels. Any nearby physics lab. Any nearby clinic working with X-rays or medical imaging. Anyone nearby developing photo film.

If a person who has recently had a radiopharmaceutical medical imaging procedure walks into a nuclear power plant or physics lab, or a radiation detector installed at a border crossing or port around the USA, they’ll set off alarms.

Radioactivity is so easy to detect that in 1896 Becquerel discovered it accidentally.

I remember that there was a case, in November of 2008 I think, where a little bit of radioactive ¹³³Xe was vented from the ANSTO Lucas Heights radiopharmaceuticals facility… this was quickly detected in Melbourne by the atmospheric radiochemistry monitoring station which is part of the network being developed for the CTBTO for CTBT verification… one of a large network of such sites, which are extremely sensitive, all around the world which are used to detect any possible nuclear weapon test.

Japan, and Hong Kong and mainland China have plenty of expertise and infrastructure that they can use to, for example, perform the sensitive analysis of fission-product radionuclides in the atmosphere to monitor nuclear weapons testing and nuclear fuel processing in the DPRK… so they can also certainly analyse the presence of traces of artificial radionuclides in the atmosphere from this nuclear power plant incident.

How many nuclear power stations are there in the United States that are located relatively close to TMI, in the states geographically around TMI? What did their radiological monitors show? Anything? Photographic films from everyone around the area was collected and looked at – no radiation was recorded.

Basically, the whole idea of such an enormous cover up is just an enormous, impractical conspiracy theory – which would need to involve the state government, the federal government, the nuclear energy industry, and huge numbers of the public and huge numbers of scientists and industries – like an Apollo hoax conspiracy theory.

We will know, over the next few days, exactly what the true situation is regarding the composition, and quantity, of any releases of radionuclides into the outside environment. There are no coverups or conspiracies in this context – there simply cannot be.

I hope you’ve found this post informative. Please feel free to post comments, with any further discussion, questions, criticisms or what-have-you. I will likely follow up this post with a future post, following future developments of this issue, and responding to questions or new information as it comes to hand.

Over the last few months, one of the development tasks being undertaken by the Lunar Numbat team has been the development of embedded controllers which receive instructions from a flight vehicle’s central computer over a CAN (Controller Area Network) connection and translate these throttle position instructions into signals which control the powerful motors which drive the valves that admit propellant into the rocket engines.

This system has been developed primarily for ASRI‘s AusRoc 2.5 research rocket, however a derivative system may have other applications to other vehicles in the future, such as the descent engine on the White Label Space GLXP lander.

Here’s a picture of some of the fruits of our labors over the past few weeks:

Two of these boards will be used to control the AusRoc 2.5 rocket engine; one controlling the kerosene valve and the other controlling the LOX valve. These boards are based around an Atmel ATmega328 microcontroller, which communicates over CAN through a Microchip MCP2515 CAN transceiver. (At the moment the real linear pots aren’t installed, so a couple of conventional pots are connected for testing.)

The 24 V power supply rail from the main avionics fairing is wired down to the throttle fairing along with the CAN link, and this supplies power to the Rutex motor controller and is regulated down to 5 V to run the interface controller and AVR microcontroller.

These embedded microcontrollers measure the position of the propellant valves through a pair of linear potentiometers attached to the valve assembly and send the appropriate motion control information to a Rutex R2020 motor control board connected to each of the motors, which control the powerful motors through a H-bridge running at 50 volts.

Powerful motors and high-torque gearboxes are required, since we require very reliable, very fast but still carefully controlled opening of a fairly large, heavy metal valve, even when it is under the pressure of a full, pressurized fuel tank… or cryogenic liquid oxygen at 90 K:

One of the valves mated with its gearhead assembly:

At this point we need to write the software for the system, assemble the linear potentiometers and their associated mechanical components to measure the absolute position of the valves, and of course perform tests and then some more tests, and hopefully within the coming months we will be starting to integrate this system with ASRI’s rocket engine to perform some cold static testing of propellant flow into the engine.

So, OK, I thought I’ll write a blog post about it. The day is supposed to be about action, as the name suggests, so let’s talk about specific actions, with a view towards making a significant mitigation, in a realistic way, of Australia’s anthropogenic carbon dioxide emissions.

Australia’s brown coal (lignite) fired electricity generators have by far the highest specific carbon dioxide emissions intensity per unit of electrical energy generated, since they’re burning relatively high moisture brown coal. They are the most concentrated point contributors to the anthropogenic GHG output. Therefore, these are the “low-hanging fruit” – a very valuable target to look at first and foremost if we want to make the greatest realistic mitigation of the country’s carbon dioxide emissions in a practical way, followed by black coal-fired generators.

Australia’s total net greenhouse gas emissions in 2006 were 549.9 million tonnes of CO2 equivalent.

If we look at the three main sets of lignite-fired generators in the Latrobe valley in Victoria, they represent a very concentrated point source of CO2 output, so they’re a very good case to focus on specifically.

In 2006, Hazelwood generated 11.6 TWh of electrical energy, and 16,149,398 tonnes of carbon dioxide to atmosphere.

In 2006, Loy Yang A generated 15.994 TWh of electrical energy sent out to the grid and 19,326,812 tonnes of carbon dioxide to atmosphere.

I’ll exclude Loy Yang B from this list for the moment, since its numbers are eluding me.

In 2006, the Yallourn power station generated 10.392 TWh of electrical energy sent out to the grid and 14,680,000 tonnes of carbon dioxide to atmosphere.

If you look at the the total contribution of just those three brown-coal-fired plants combined, you’re looking at 9.12 percent of Australia’s total anthropogenic carbon dioxide emissions. If you replace those with clean technology that can deliver an equivalent electricity output, you get a 9.12 percent reduction in Australia’s CO2 emissions. (When you include Loy Yang B, I think it’s approximately 11-12%.)

That’s not a bad target for Australia to implement for the relatively short term for a real reduction in CO2 emissions. It can actually be done, if the real political will exists to do it.

Now, I’m not interested in this “100% renewable energy by 2020″ business from the extremist any-excuse-for-a-protest Socialist Alternative set, because it is nonsense.

Replacing all the coal-fired and gas-fired generators in this country inside 10 years (and presumably only using wind turbines and solar cells, not nuclear energy of course since it doesn’t fit their para-religious ideology)? That’s complete bullshit, of course, because in the real world it cannot be done.

There’s a difference between setting a challenging target and setting a nonsense target. Unless you’re only trying to implement a political bullshit stunt instead of actually trying to hit your targets.

Of course, you don’t just close down the coal-fired generators. You’ve actually got to build their clean replacements first. So what do you use that can realistically replace a coal-fired power station? Nuclear power, of course.

Now, again, to be realistic, we probably can’t build LFTR/MSR, PBMR/HTGR, IFR/PRISM or any kind of nuclear fusion based generation capacity on a large scale to generate grid-connected energy right now. That’s not to say that pilot-scale research and development on those very cool technologies shouldn’t continue, but right now, getting more nuclear energy on the grid means advanced light water reactors – or maybe heavy water CANDU-type things, or conventional sodium-cooled fast reactors maybe. The most practical thing for serious deployment in the relatively short term is advanced LWR technology. In the slightly longer term, there is certainly a place to be encouraging both Gen. IV and fusion.

To get the same amount of energy as the total output from those coal plants, as above, which we’re talking about replacing, we need 4.56 GW of installed nuclear capacity, assuming a 95% capacity factor.

With 4 x 1154 MWe Westinghouse AP1000s, with a 95% capacity factor, you’ve got 4.62 GW, which is a little more than what’s needed.

You can easily have four nuclear power reactors integrated into one nuclear power plant.

Now, how much does it cost?

On March 27, 2008, South Carolina Electric & Gas applied to the Nuclear Regulatory Commission for a COL to build two AP1000s at the Virgil nuclear power plant in South Carolina. On May 27, 2008, SCE&G and Santee Cooper announced an engineering, procurement, and construction contract had been reached with Westinghouse. Costs are estimated to be approximately $9.8 billion for both AP1000 units, plus transmission facility and financing costs.

That gives you an idea of how much a nuclear power plant costs today, in the current financial environment, in the current regulatory environment.

If we double that figure of USD$9.8 billion, it’s AUD $21.4 billion. There will be some saving since we’re considering building four reactors at one plant, not two independent two-reactor plants.

How much that saving will be, quantitatively, I don’t really know. If the cost is reduced by 30%, we’re looking at 15 billion Australian dollars.

How long would it take? If the real political will exists to do it, 10 years is heaps of time. We could probably do even more in that timeframe if we really, really wanted to. AP1000 construction takes 36 months from first concrete poured to fuel load, if you ignore any political protest rubbish.

This is really just a base-line relatively achievable “base case”. After this decade, of course, the rate of nuclear power deployment – and related GHG emissions mitigation – could foreseeably accelerate.

What about the uranium input? About 600 tonnes of natural uranium per year total, for all four reactors. Australia’s present production, off the top of my head, is something like 10,000-11,000 tonnes. Australia’s present uranium production can very, very easily provide for Australia’s total electricity production even without expansion of uranium production – again, considering the inefficient once-through use of low-enriched uranium in conventional LWRs.

What about the so-called “waste”?
Roughly 80-85 tonnes of used uranium fuel per year. 96% of that is unchanged uranium, so that 76.8 tonnes of uranium can be seperated and re-used. It’s just uranium, so it’s not going to hurt you.

The remaining 3200 kg is made up of the valuable, interesting and unique byproduct materials from a nuclear reactor – unique resources with all kinds of different technological applications, which aren’t all radioactive, which you cannot get anywhere else.

Anyway, that’s one scenario which I happen to think has a lot of merit.

Maybe you don’t agree – but if you don’t agree, I’d love to see you elucidate an alternative scenario which can deliver the equivalent greenhouse gas emissions mitigation – shown to be accurate in a quantitative way – within a comparable timeframe and within a comparable cost.

It will not be inexpensive, and it will not happen overnight – but I have yet to see any scenario which can honestly do the same job faster and cheaper, when some real quantitative analysis is applied.

I know EAGLE isn’t FOSS, so if you’re eally principled about not using any software that isn’t open, unfortunately you won’t be able to open those files. I haven’t learned to use gEDA yet. If you want to, by all means, feel free to re-draw the board and schematic layouts in gEDA, using the PNG images provided.

Schematic (moderately large .png)

PCB layout (moderately large .png)

EAGLE schematic file.
EAGLE board file.

That board design has everything routed…. it’s complete. Yay :)

i) Temperature sensor.
Here I’ve just assumed that we can use a DS18B20 to measure temperature; pretty simple really, just one microcontroller pin, and a 4.7 k pull-up resistor. We could of course have multiple sensors throughout the rocket and just bus them all together, and connect them all back to the one connector on the main PCB. There are other sensors we could use in theory, but the DS18B20 is common, convenient to use, and there is heaps of experience and documentation with regards to using it in the Arduino community.

I didn’t have a part library for the DS18B20, so I just put a 3-pin 0.1″ pin header on the board. You can simply solder the TO-92 package through that footprint on the board quite easily, or alternatively, you can stick a pin header on the board, and wire up the DS18B20(s) off the board.

ii) GPS.
Here I’ve just picked LS20031 5 Hz GPS from SparkFun, since Jon mention that’s the one he has experimented with.
It’s a 3.3V device, so we simply have a 1:2 voltage divider on its RX line to interface it to the 5V microcontroller.
I’ve just used a standard 0.1″ pin header footprint here, so flying wires can be soldered on to connect to the pads
on the GPS board.

Since the microcontroller on the Arduino only has one hard UART, a virtual soft UART needs to be used, and the performance that you programming gurus could squeeze out of that code will be the factor that limits the speed at which the GPS could be read. It’s not my department :)

iii) Light sensors.
Just two LDRs connected to two microcontroller ADC inputs, connected with a couple of resistors as voltage dividers. Pretty simple, really. The resistor values might need to be tweaked depending on the typical resistances of the LDRs used, but they should be pretty flexible, and 100 k should be about right, since we really only want qualitative information from them anyway. The LDRs can of course be mounted off the board on long flying wires, and mounted whereever they have to be mounted.

iv) Real-time clock.
I’ve just assumed we’re using a DS1307 here, and this is really just a very simple schematic which is needed to support this device. There are a couple of 10 k pull-up resistors on the I2C bus. You could use any DS1307 prototyping or development board, as available, for example, from Sparkfun or Microzed or Futurelec, to develop code for this, they’re all the same chip, and they all have essentially identical hardware.

v) Flash memory.
Mike mentioned the AT45DB61B, and that looks like the perfect device.

It has heaps of storage, it’s fast, it’s easy to interface, and it’s inexpensive. There’s nothing not to like with this suggestion. This is a 3.3 V device, however, but the datasheet says that the communications signals are 5 V tolerant, so there shouldn’t be any trouble with interfacing it. 2 Mb ought to be enough for anything (touch wood) but if it isn’t, there’s no reason why another chip couldn’t be added on the SPI bus.

Rockby sells the device in the CASON package for $2.60 AUD, and Sparkfun sells the same device in the SOIC package for $3.60 AUD. Since the latter isn’t far more expensive, and the CASON package looks like it’s a bitch to solder by hand, I have used the SOIC package :)

vi) RF telemetry.
Here, just for the sake of drawing the schematic, I’ve assumed we might be using an XBee device, although if you wanted to change the design to use, say, one of the DR3100 devices there wouldn’t be much of a change, they’re still 3 V devices, using bidirectional serial comms back to the microcontroller, so there really isn’t much difference. The XBees have a couple of LEDs to indicate their status, eg. that they’ve got a communication link to another device, so I’ve just added these in, since a few blinkenlights don’t hurt :)

Note that the XBee is a 3.3V device, so I’ve just divided the data input from the microcontroller down with a 2:1 voltage divider, which should be fine.

vii) Microcontroller
Here I’ve just assumed that a standard Arduino Duemilanove board is used. I’ve especially chosen the pins allocated for the I2C and SPI buses to correspond to the AVR’s pins for those hardware interfaces, although if these interfaces were implemented in software, any other pins could be used, really. There’s a reset button included, since the shield board blocks the one on the Arduino itself.

viii) Accelerometer and ADCs.
I’ve just assumed we’re using an ADXL330, and the LTC1298s as Mike mentioned, which is a cool idea.

I chose the ADXL330 pretty much arbitrarily, since it’s the first one I could find an Eagle library for, it seems pretty common, and SparkFun stock it. It’s the same 3-axis accelerometer as used, for example, in the Arduino Lillypad accelerometer board.

The Analog Devices ADXL345, with A-D conversion built in to the IC with only an interface to the microcontroller’s SPI or I²C bus needed, looks very cool though, and the price difference, once the cost of ADC chips is included, doesn’t seem that large.

ix) Barometric pressure sensor.
This is just a Freescale MPX4115, which seems somewhat common in UAV, model aircraft and rocketry work for altimeter applications.

I’ve put a bit of low-pass RC filtering on the output from the barometric sensor, using the component values just as given from Freescale AN146. I’m currently working out the maths to read out the altitude from the output from the barometric sensor, which isn’t that trivial.

ix) Power supply
I’ve removed the LiPo battery and SMPS power supply from the previous version, since we need a 12V battery for the video downlink anyway.

I’ve just assumed that 12 V is plugged into the Arduino’s power supply jack, and 5 V is coupled to the daughterboard via the Arduino’s pin headers. There’s a 3.3V regulator on the board supplying the 3.3V for the XBee and accelerometer and DataFlash, an LM1117 at this point, although you could swap it for some other kind of device that is available. It needs to be a low dropout regulator, though, since you’re taking a 5V input to regulate down to 3.3 V.

The Vin pin on the Arduino is connected to one of the analog input pins via a voltage divider, so that battery voltage can be measured by the Arduino. (I know, those pins are right next to each other, and this seemed like something of a sweet hack.) With those resistor values chosen at the moment, 12 V on Vin corresponds to about 4.53 V on the ADC input, so you can read it straight off. Actually, Vin won’t be exactly equal to the battery voltage, because there’s a polarity protection diode on the Arduino power input before that Vin pin connection…. so there will be a little drop across that. I don’t know exactly how much, maybe about 0.2 V for a Schottky diode, I haven’t empirically measured it.

Now, what to call this thing? “That Arduino based rocket instrumentation/datalogger project” is too much of a mouthful. So, we need a name.

How about, say, ARTEMIS. That is, Arduino Rocket Telemetry and Instrumentation System.
(You can totally see the experimental physicist in me reflected in that name, can’t you?)

That’s just my idea for a name. I’ll let the Hackerspace crew mull over it and come up with a better idea if they want to.

Bill of materials… for the design as it stands at present.

Resistors: (All resistors 0805 SMD package)

R1      4.7 k
R2      100 k
R3      100 k
R4      750 R
R5      10 k
R6      10 k
R7      100 k
R8      100 k
R9      56 R
R10     100 k

R11     56 R
R12     100 k
R13     150 k
R14     91 k

Capacitors:

C1      10 uF 16 V tantalum, through-hole                                       Jaycar RZ-6648
C2      10 uF 16 V tantalum, through-hole                                       Jaycar RZ-6648
C3      100 nF, 0805 SMD package
C4      100 nF, 0805 SMD package
C5      100 nF, 0805 SMD package
C6      100 nF, 0805 SMD package
C7      100 nF, 0805 SMD package
C8      1 uF, 25 V tantalum, through-hole                                       Jaycar RZ-6627
C9      10 nF, 0805 SMD package
C10     330 nF, 0805 SMD package
C11     100 nF, 0805 SMD package
C12     100 nF, 0805 SMD package
C13     100 nF, 0805 SMD package

ICs:

I've specified some examples of parts suppliers that I know have the relevant items, though they may not be the only choices, or the best choices.

IC1     Linear LM1117-3.3 LDO 3.3 V voltage regulator, SOT-223 package          Digikey LM1117MP-3.3CT-ND
IC2     Dallas DS18B20 1-Wire temperature sensor, TO-92 package                 Sparkfun SEN-00245
IC3     Linear LTC1298 analog-to-digital converter, DIP-8 package               Futurelec
IC4     Analog Devices ADXL330 3-axis accelerometer, LFCSP-16 package           Sparkfun COM-00730
IC5     Freescale MPX4115A barometric pressure sensor, 867-H package            Digikey
IC6     Linear LTC1298 analog-to-digital converter, DIP-8 package               Futurelec
IC7     Dallas DS1307 real-time clock, SOIC-8 package                           Futurelec
IC8     Atmel AT45DB161B 16 MBit DataFlash memory, SOIC-8 package               Sparkfun COM-00301

IC9     Arduino Duemilanove board
IC10    XBee module (any, really)
GPS     LS20031 GPS module                                                      Sparkfun GPS-08975

B1      12 mm coin cell holder, SMD                                             Sparkfun PRT-07948
plus CR1225 3 V lithium coin cell                                       Sparkfun PRT-00337
Q1      32.768 kHz quartz oscillator crystal, TC-38 through-hole package        Sparkfun COM-00540
S1      Momentary tactile pushbutton switch, through-hole                       Jaycar SP-0600
LDR1    Standard cadmium sulfide photoresistor, through-hole 0.1"               Jaycar RD-3480
LDR2    Standard cadmium sulfide photoresistor, through-hole 0.1"               Jaycar RD-3480
LED1    Standard LED; 3 mm through-hole                                         Jaycar ZD-0120
LED2    Standard LED; 3 mm through-hole                                         Jaycar ZD-0120

2 x 10-pin 2 mm header sockets for XBee module                                  Sparkfun PRT-08272
Break-away 0.1" machined pin socket strip for mounting LTC1298 ICs              Jaycar PI-6470
28-pin break-away 0.1" pin header strip for mounting shield to Arduino          Jaycar HM-3211

It’s everything that Google Calculator could have been but never was.

The out-of-pocket costs for installation of these systems range from about $3000 to $5000, although some companies offer such systems for essentially nothing, only $500 or so for the meter upgrade, after the rebate is repaid.

These guys sell their basic 1 kW system for $5000 out-of-pocket after the subsidies and rebates, and these guys sell their systems for $3000 after rebate.

For most of these systems, the average cost advertised, the out-of-pocket final cost after the subsidies have been taken off, is about $3000 depending on the quality of the system.

Personally, in the case of the systems advertised for zero overall cost, I’d be a little bit worried about the quality of the system, since they’d have to be honing the price down quite a bit to get it down to the point where they can pay for it, pay for installation, and still make a profit, just from the $9000 or so in the government handouts.

You wouldn’t want a shonky system that burns down your house, would you?

In southeastern Australia, including Melbourne, Adelaide, Sydney and everything in between, the average daily solar exposure is 15 megajoules per square meter per day.

So, that’s an average power density of 174 W/m², on average, over the whole day.

Now, if you buy a solar PV module that is rated at 180 W, or whatever power figure it is, you get that amount of power if there is 1000 W/m² of solar radiation incident onto the panel.

So, a “1000 W” array, in the real world with an average of 174 W/m² worth of incident radiation flux, will generate 174 W of power, on average. (Averaged over the full 24 hours in a day.)

Therefore, you get about 1500 kWh of total energy generation per year.

If you’re paying, say, 13 c/kWh for electricity, you save about $195 per year on the electricity bills.
(I know that’s a relatively low price for household electricity, but it’s about right for Victoria’s inexpensive electricity, powered by the Latrobe valley’s incredible combustible mud. If your electricity prices are considerably higher, you can see how to repeat the calculations for your electricity price.)

If you pay about $3000 out-of-pocket for such a system, then, it will take 16 years to pay for itself.

However, after about 10 years, the grid-connect inverter will die (These guys have a 5 year warranty on theirs), and there won’t be a subsidy paying for that, so that’s probably another $2000 or so you’ll need to shell out. So, that adds another 10 years to the payback time. You probably won’t even be able to pay it off before that second inverter reaches the end of its life.

There are installers that offer higher quality inverters with longer warranties, but they are the higher end of the price brackets for the systems – this is the catch with the extremely cheap systems.

So you’re looking at a payback time of 26 years, for a system where the solar cells are unlikely to last more than 20-25 years.

Such systems, in all likelihood, are never going to pay themselves off, even with the huge government subsidies.

All these businesses that are doing the installations are literally leeching off the huge government subsidies; if you took away the subsidies they would all disappear straight away.

With a saving of close to $200 per annum off your electricity bill, if the customers had to pay the $8000 which is subsidized by the government, just that $8000 portion alone would take 40 years to pay off, and you would never, ever even come remotely close to paying it off.

This scheme is just a huge money sink for the government; it’s completely unsustainable, and it doesn’t accomplish anything meaningful.

Customers love it, since they’re effectively getting this huge investment mostly given to them by the government.

It’s just like Krudd’s economic stimulus handouts – people are getting a generous free handout, so they think that’s fantastic, and people will very rarely stop and question whether this actually makes sense as a worthwhile thing for the good of the country.

One of these systems generates about 1500 kilowatt-hours per year.

In 2006, the electricity output sent to the grid from Loy Yang A, just as a typical example, was 15,995 gigawatt-hours.

Therefore, if you wanted to generate the same amount of energy from 1 kW solar PV installations as just one coal-fired power station, you’d need 10.7 million of these installations. That’s significantly more than the number of households in this country.

You would need just under 11 million typical household 1 kW rooftop solar PV installations – well in excess of the number of households in the country – to give you the same amount of electricity as one coal-fired power station.

Even if you could do that – which you can’t – that still doesn’t give you the means to replace the coal-fired power station, because it isn’t high capacity factor, baseload, generation. You still need that high capacity factor baseload generation to back you up when the photovoltaics are delivering less energy, or no energy at all.

If the government paid out the $8000 subsidies for 10.7 million 1 kW solar panel installations – which aren’t capable of replacing even one coal-fired station – it would cost 86 billion dollars.

This is an enormous amount of money getting flushed away to do nothing in reality, and I’m happy to see it scrapped, personally.

Here’s a contemporary real world example of people buying into “green” ideology without the ability to count kilowatt-hours, and without realising just how damned expensive silicon photovoltaics are. I really, really suggest people do their homework before handing over sums of money like that.

BRISBANE environmental lawyer Jo Bragg and her partner, Gary Kane, spent $28,000 on three roof panels to generate solar power for their home in the inner Brisbane suburb of Highgate Hill.

After receiving a federal government rebate of $8000, they hoped to recover their investment in a cleaner planet within a few years by selling excess power into the mains electricity grid.

In the three months to April, they used 1384 kilowatt hours and produced 388 kilowatt hours of excess power, for which they received the princely sum of $12.96 after taxes.

“Governments are not being serious about reducing energy consumption with lousy amounts of money like that,” Ms Bragg said.

Her family is the kind Kevin Rudd had in mind yesterday when he announced that individuals and households would be part of a revamped carbon pollution reduction scheme.

The Prime Minister said households would be able to calculate their energy use at home and pledge contributions to the $25million energy efficiency savings fund to effectively offset their emissions.

“Individuals will be able to calculate their energy use and establish the savings they could achieve with a more energy-efficient home,” Mr Rudd said.

“A household or individual could then make a tax-deductible donation to the pledge fund, which the fund would use to buy and cancel carbon pollution permits equivalent to that level of energy use.”

Ms Bragg said she hoped the carbon permits scheme would be flexible enough to allow households with renewable energy to be paid for the gross amount of power produced — not just the excess — as happened in Germany and some other countries.

“It makes sense to provide incentives to homes to make it worth their while to invest in renewable energy,” she said.

“Even if we were paid for the gross amount of power produced, it would take us eight or 10 years to recover the investment.”

Obviously, this thirteen dollar figure seems a little bit surprising, so let’s see if we can break it down a little and try and extract some more concrete information about where that figure came from.

The average Australian residential electrical energy consumption is approximately 25 gigajoules per annum, and 1384 kWh over three months is about 80 percent of that, which is plausible for a relatively small, energy-efficient household.

Based on the quoted cost of $28,000 (post-rebate), and the mention of “three roof panels” in the article, I’m going to take an educated guess here and say they have a relatively large system, with 3 kW of installed nameplate capacity.

We’ll assume that the BOM insolation data for February is representative of the average of the three month January-February-March period, meaning that the average insolation is 21 MJ/m²/day, which is 243 W/m² on average, meaning that a “3 kW” nameplate capacity installation will produce 729 W of power on average. Therefore, over three months, the system should produce about 1598 kWh.

Since the Green protagonists of our tale used 1384 kWh and sold back onto the grid 388 kWh of net energy production, the gross energy production from the system was presumably 1772 kWh, which is 111% of the theoretical 1598 kWh. So, these published figures are consistent with the numbers we theoretically expect.

Now, in Queensland, they’re getting paid a special elevated 44 c/kWh feed-in tariff for their net electricity generation from their PV installation. Therefore, their 388 kWh should have earned them 171 dollars. But, ostensibly, it did not. So where did the rest of the revenue, the seemingly missing $158, go?

The answer is that it probably went, I presume, to cover various flat-rate fees included on the electricity bill, such as a service charge, and/or an ambulance levy, or what have you; components of the electricity bill other than the per-kilowatt-hour charge for the energy use. Maybe it seems a little high for that, but we can’t be certain without seeing the actual breakdown of the bill.

Normally, before the PV installation, they’d be paying an electricity bill of about 221 dollars for their 1384 kWh, assuming a rate of about 16 cents per kWh, plus those charges of $158 or so on the bill, or a total of about $379. Now, they’re getting paid 13 dollars for this quarter, not paying any electricity bill at all. Therefore, they’re saving $392 per quarter, or actually somewhat less than that during the winter months since the solar insolation will be less. And they’re complaining about that? Even after getting the free $8000 government rebate and the 44 c/kWh feed-in handout, they’re complaining that that’s not enough and they want more of a free handout, for something that will never make any real contribution towards replacing coal-fired generation?

Let’s suppose, without calculating it accurately, that they save $300 per quarter on average, every quarter. Therefore, paying off their $28,000 system will take just over 23 years; assuming they don’t need to replace the grid-connect inverter in less than that time. They might, maybe, pay the cost of the system off within the operational lifespan of the solar cells. Maybe. That is, assuming that the 44 c/kWh elevated feed-in subsidy continues indefinitely, and they’re really lucky and their inverter doesn’t need to be repaired or replaced within that timeframe.

__________________