WNA Fukushima Accident 2011 (updated 27 April 2011)

Fukushima Accident 2011

(updated 27 April 2011)

• Following a major earthquake, a 14-metre tsunami disabled the power supply of three Fukushima Daiichi reactors, interrupting cooling and hence causing a nuclear accident. 
• The accident was made worse by used fuel storage on site losing water and in one case forming hydrogen and apparently releasing radioactivity. 
• The accident has been provisionally rated 7 on the INES scale, due to the high radioactive releases in the first few days. All four reactors are evidently written off - 2719 MWe net. 
• After two weeks the three reactors (units 1-3) were stable with active intervention but still not with proper cooling re-established. Achievement of cold shutdown is not expected for some months.  

The Great East Japan Earthquake* with magnitude 9.0 at 2.46 pm on Friday 11 March did considerable damage in the region, and the large tsunami it created caused very much more. The earthquake was centred 130 km offshore the city of Sendai in Miyagi prefecture on the eastern cost of Honshu Island (the main part of Japan), and was a rare and complex double quake giving a severe duration of about 3 minutes. Japan moved a few metres east and the local coastline subsided half a metre.
* originally, and more specifically, the Tohoku-Chihou-Taiheiyo-Oki Earthquake.
Eleven reactors at four nuclear power plants in the region were operating at the time and all shut down automatically when the quake hit. The operating units which shut down were Tepco's Fukushima Daiichi 1, 2, 3, Fukushima Daini 1, 2, 3, 4, Tohoku's Onagawa 1, 2, 3, and Japco's Tokai, total 9377 MWe net. Fukushima Daiichi units 4-6 were not operating at the time, but were affected, total 2587 MWe net (units 4-6). Onagawa 1 briefly suffered a fire in the turbine building, but the main problem initially centred on Fukushima Daiichi units 1-3. Unit 4 became a problem on day five.
Power, from grid or backup generators, was available to run the Reactor Heat Removal (RHR) system cooling pumps at eight of the eleven units, and despite some problems they achieved 'cold shutdown' within about four days. The other three, at Fukushima Daiichi, lost power at 3.42 pm, almost an hour after the quake, when the entire site lost the ability to maintain proper reactor cooling and water circulation functions due to being flooded by the tsunami.  This disabled all 13 back-up generators on site and also the heat exchangers for dumping reactor heat to the sea.
Thereafter, weeks of focused work centred on restoring heat removal from the reactors and coping with overheated spent fuel ponds. This was undertaken by hundreds of Tepco employees as well as some contractors, supported by firefighting and military personnel. Some of the Tepco staff had lost homes, and even families, in the tsunami, and were living in temporary accommodation under great difficulties and privation, with some personal risk. Media coverage of the Fukushima drama often ignored the context of the enormous natural disaster which greatly affected how it played out.
Three Tepco employees at the Daiichi plant were killed directly by the earthquake and tsunami.
An earthquake with magnitude 7.1, closer to Fukushima than the 11 March one, was experienced on 7 April, but without further damage to the plant.  The epicenter was 120 km from Fukushima but only 20 km from Onagawa, where power supply was affected.  On 11 April a magnitude 7.1 earthquake and on 12 April a magnitude 6.3 earthquake, both with epicenter at Fukushima-Hamadori, caused no further problems.
The following information is provisional, especially regarding interpretation of events, which may take some months to be agreed. Where sources conflict, NISA version has been preferred. This paper will be updated and simplified as further information and analysis comes to hand. Pressures are quoted in absolute terms, ie atmospheric plus 'gauge' unless indicated otherwise (1 kPa = 0.145 psi).
Organisational acronyms:
Tepco = Tokyo Electric Power Company
NISA = Nuclear & Industrial Safety Agency (Japan, regulator),
NSC = Nuclear Safety Commission (Japan, policy body)
METI = Ministry of Trade, Economy & Industry (Japan),
IAEA = International Atomic Energy Agency (UN body),
JAIF = Japan Atomic Industrial Forum (industry body)
ISRN = Institute for Radiological Protection & Nuclear Safety (France)

Site, earthquakes and tsunamis: background  The Daiichi (first) and Daini (second) Fukushima plants are sited about 11 km apart on the coast. Japanese nuclear power plants are designed to withstand specified earthquake intensities evident in ground motion. If they register ground acceleration of a set level, systems will be activated to automatically bring the plant to an immediate safe shutdown. In this case the set scram level was 135 Gal (150 Gal at Daini). The maximum response acceleration against design basis ground motion for both Fukushima plants had been upgraded since 2006, and is now quoted at horizontal 438-489 Gal for Daiichi and 415-434 Gal for Daini. At this level they must retain their safety functions. In 2008 Tepco upgraded its estimates of likely Design Basis Earthquake Ground Motion Ss for Fukushima to 600 Gal, and other Japanese operators have adopted the same figure. The recorded data for both plants shows that 550 Gal (0.56 g) was the maximum ground acceleration for Daiichi, in the foundation of unit 2 (other figures 281-548 Gal), and 254 Gal was maximum for Daini. Units 2, 3 and 5 exceeded their maximum response acceleration design basis in E-W direction by about 20%. Recording was over 130-150 seconds. All nuclear plants in Japan are built on rock (ground acceleration was around 2000 Gal a few kilometres north, on sediments). The design basis tsunami height was 5.7 m for Daiichi and 5.2 m for Daini, though the Daiichi plant was built about 10 metres above sea level and Daini 13 metres above. Tsunami heights coming ashore were more than 14 metres for both plants, and the Daiichi turbine halls were under some 5 metres of seawater until levels subsided.  The maximum amplitude of this tsunami was 23 metres at point of origin, about 160 km from Fukushima. In the last century there have been eight tsunamis in the region with maximum amplitudes at origin above 10 metres (some much more), these having arisen from earthquakes of magnitude 7.7 to 8.4, on average one every 12 years. Those in 1983 and in 1993 were the most recent affecting Japan, with maximum heights at origin of 14.5 metres and 31 metres respectively, both induced by magnitude 7.7 earthquakes.

Fukushima Daiichi 1-3 & 4

Reactors 1-3 were shut down in response to the earthquake, as designed, since the ground acceleration was about 500 Gal. The reactors then reverted to an auxiliary Residual Heat Removal (RHR) cooling system, since the steam is no longer being conveyed to the turbines and condenser circuit. The RHR system is driven by electric pumps. Mains power was lost due to the quake, so the emergency diesel generators located in the basements of the turbine buildings started up to drive this system and ran for 56 minutes, but then stopped when submerged by the tsunami. This put those reactors in a dire situation and led the authorities to order, and subsequently extend, an evacuation while engineers worked to restore power. About nine hours later mobile power supply units had reached the plant and were being connected. The capacity of these would be less than the main plant diesel system. Meanwhile units 1-3 had only battery power, insufficient to drive the main RHR cooling system. The batteries were apparently depleted in about 8 hours anyway. 

Reactors: background  The Fukushima Daiichi reactors are GE boiling water reactors (BWR) of an early (1960s) design supplied by GE and Toshiba, with what is known as a Mark I containment. Reactors 1-3 came into commercial operation 1971-75. Reactor power is 460 MWe for unit 1, 784 MWe for units 2-5, and 1100 MWe for unit 6. The fuel assemblies are about 4 m long, and there are 400 in unit 1, 548 in units 2-5, and 764 in unit 6. Each assembly has 60 fuel rods containing the uranium oxide fuel within zirconium alloy cladding. Unit 3 has a partial core of mixed-oxide (MOX) fuel (32 MOX assemblies, 516 LEU). They all operate normally at 286°C at core outlet under a pressure of 6930 kPa and with 115-130 kPa pressure in dry containment. NISA says maximum design base pressure for reactor pressure vessels (RPV) is 8240 kPa at 300°C, and for containment (PCV) is about 500 kPa*.  At low containment pressures hydrogen and other gases are routinely vented through charcoal filters which trap most radionuclides. * NISA gives 430 kPa for unit 1 and 380 kPa for 2-3 at 140°C as 'maximum', apparently gauge pressure, so add 101 for absolute: 530 and 480 kPa. Before venting, unit 1 RPV got to 900 kPa and PCV to 850 kPa early on 12th. The BWR Mark I has a Primary Containment system comprising a free-standing bulb-shaped drywell of 30 mm steel backed by a reinforced concrete shell, and connected to a torus-shaped wetwell beneath it containing the suppression pool. The drywell, also known as the Primary Containment Vessel (PCV), contains the reactor pressure vessel (RPV). For simplicity, we will use the term 'dry containment' here. The water in the suppression pool acts as an energy-absorbing medium in the event of an accident. The wetwell is connected to the dry containment by a system of vents, which discharge under the suppression pool water in the event of high pressure in the dry containment. The function of the primary containment system is to contain the energy released during any loss-of-coolant accident (LOCA) of any size reactor coolant pipe, and to protect the reactor from external assaults. The Japanese version of the Mark I is slightly larger than the original GE version. During normal operation, the dry containment atmosphere and the wetwell atmosphere are filled with inert nitrogen, and the wetwell water is at ambient temperature. If a loss of coolant accident (LOCA) occurs, steam flows from the dry containment (drywell) through a set of vent lines and pipes into the suppression pool, where the steam is condensed. Steam can also be released from the reactor vessel through the safety relief valves and associated piping directly into the suppression pool. Steam will be condensed in the wetwell, but hydrogen and noble gases are not condensable and will pressurise the system, as will steam if the wetwell water is boiling. In this case emergency systems will activate to cool the wetwell, see below. Excess pressure from the wetwell (above 300 kPa) can be vented through the 120 m emission stack via a hardened pipe or into the secondary containment above the reactor service floor of the building. If there has been fuel damage, vented gases will include noble gases (krypton & xenon), iodine and caesium, the latter being scrubbed in some scenarios. Less volatile elements in any fission product release will plate out in the containment. (The later Mark II containments are similar to Mark I, but both are much smaller than the Mark III and those which became standard in PWRs.) The secondary containment houses the emergency core cooling systems and the spent/ used fuel pool. It is not designed to contain high pressure. The primary cooling circuit of the BWR takes steam from above the core, in the reactor pressure vessel, to the turbine in an adjacent building.  After driving the turbines it is condensed and the water is returned to the pressure vessel. There are also two powerful jet-pump recirculation systems forcing water down around the reactor core and shroud. When the reactor is shut down, the steam in the main circuit is diverted via a bypass line directly to the condensers, and the heat is dumped there, to the sea. In both situations a steam-driven turbine drives the pumps, at least until the pressure drops to about 450 kPa (50 psig), but condenser function depends on large electrically-driven pumps which are not backed up by the diesel generators. In shutdown mode, the Residual Heat Removal (RHR) system then operates in a secondary circuit (RHR is connected into the two jet-pump recirculation circuits), driven by smaller electric pumps, and circulates water from the pressure vessel to RHR heat exchangers which dump the heat to the sea. There is also a Reactor Core Isolation Cooling (RCIC) actuated automatically which can provide make-up water to the reactor vessel (without any heat removal circuit). It is driven by a small steam turbine using steam from decay heat, injecting water from a condensate storage tank or the suppression pool and controlled by the DC battery system.  (The NISA account does not mention the RCIC, possibly treating it as part of the ECCS, though it has been confirmed that the RCIC systems played a helpful role in all three units until the suppression pool water boiled, to 11am on 12th in unit 3 and to 2pm on 14th in unit 2.) Then there is an Emergency Core Cooling System (ECCS) as further back-up for loss of coolant. It has high-pressure and low-pressure elements. The high pressure coolant injection (HPCI) system has pumps powered by steam turbines which are deigned to work over a wide pressure range. The HPCI draws water from the large torus suppression chamber beneath the reactor as well as a water storage tank. For use below about 700 kPa, there is also a Low-Pressure Coolant Injection (LPCI) mode through the RHR system but utilising suppression pool water, and a core spray system, all electrically-driven. All ECCS sub-systems require some power to operate valves etc, and the battery back-up to generators may provide this. Beyond these original systems, Tepco in 1990s installed provision for water injection via the fire extinguisher system through the RHR system (injecting via the jet-pump nozzles) as part of it Severe Accident Management (SAM) countermeasures. A frequently-voiced concern during the first week of the accident was regarding fuel meltdown. This would start to occur if the fuel itself reached 2500°C (or more, up to 2800°C, depending on make-up). At this point, the fuel rods slump within the assemblies. Conceivably, the “corium” (a mixture of molten cladding, fuel, and structural steel) drops to the bottom of the reactor vessel. If the hot fuel or cladding is exposed to cooling water en route, it may solidify and fracture, falling to the bottom of the reactor vessel. Given that the melting point of the steel reactor vessel is about 1500°C, there is an obvious possibility on the corium penetrating the steel if it remains hot enough. (In fact, in the 1979 US Three Mile Island accident, it didn't, though about half the core melted and it went 15 mm into the 225 mm thick pressure vessel steel. The pressure vessel glowed red-hot for an hour.) But the whole fuel melt scenario is much more probable with a sudden major loss of coolant when the reactor is at full power than in the Fukushima situation. Before this, cladding cracks at about 1200°C, its oxidation begins at about 1300°C and the Zr cladding melts at about 1850°C. These temperatures are quite possible days after shutdown in the absence of cooling.

 
When the power failed about one hour after shutdown, the reactor cores would still be producing about 1.5% of their nominal thermal power, from fission product decay - about 22 MW in unit 1 and 33 MW in units 2 & 3, according to calculations from well-established models. Without heat removal by circulation to an outside heat exchanger, this produced a lot of steam in the reactor pressure vessels housing the cores. The steam would be condensed in the suppression chamber under the reactor, within the containment, but their internal temperature and pressure rose quite rapidly. At 4.36 pm water injection using the Emergency Core Cooling System (ECCS), failed in units 1 & 2, less than an hour after power loss.
At 7.03 pm a Nuclear Emergency was declared, and at 8.50pm the Fukushima Prefecture issued an evacuation order for people within 2 km of the plant. At 9.23 pm the Prime Minister extended this to 3 km, and at 5.44 am on 12th he extended it to 10 km. He visited the plant soon after. At 6.25 pm on Saturday 12th he extended the evacuation zone to 20 km.
Over the first twelve hours pressure inside the containment structures increased steadily and led to venting from the suppression chamber via the dry containment to the service floor at the top of the reactor building, starting with unit 1 early on Saturday 12th, where pressure had reached twice the design level (over 750 kPa in PCV compared with 'maximum' 430 kPa - both gauge pressures).   This would damage pump seals and possibly the top O-ring, and in addition the safety relief valve on the main steam line from the RPV in unit 1 stuck open. Inside unit 1 water levels dropped dramatically early on 12th, exposing fuel, and this was addressed by pumping seawater into the reactor pressure vessels (RPV) using external pumps brought to site, starting with unit 1 at 8.20 pm on 12th. This was apparently via the Low Pressure Coolant Injection System, connected into the RHR system.  On 25th, water from a nearby dam started to be used instead of seawater.  Boron was added to the injected water, to guard against any criticality.
Meanwhile, vented noble gases and aerosols included hydrogen, produced by the exothermic interaction of the fuel's very hot zirconium cladding with steam and water. At 3.36 pm on 12th, there was a hydrogen explosion on the service floor of the building above unit 1 reactor containment, blowing off much of the roof and cladding on the top part of the building, after vented hydrogen mixed with air and ignited. On Sunday 13th, seawater injection to unit 1's Primary Containment started.  Pressure in unit 1 RPV dropped to below 500 kPa after the explosion and has remained 300-500 kPa since then. Containment pressure was about 800 kPa until the explosion, but then dropped and on 19th was low for four days, but jumped up on 23-24th and has been consistently 200 kPa below RPV pressure from 25th. (However, a second RPV reading has escalated from that time to over 1000 kPa - about one sixth of operating pressure.) Water injection has been via the feedwater line. On 6 April Tepco started injecting nitrogen into the containment vessel on account of suspected hydrogen levels there.  This raised the CV pressure slightly to 195 kPa, which was maintained.

In unit 2, as mentioned above, water injection using the Emergency Core Cooling System (ECCS) failed at 4.36 pm on 11th. However, the steam-driven Reactor Core Isolation Cooling (RCIC) system functioned until the morning of 14th. Pressure was vented on 13th and again on 15th, and meanwhile the blowout panel near the top of the building was opened. The reactor water level dropped rapidly after RCIC cooling was lost on Monday 14th, so seawater injection to the containment was started on 15th via the fire-fighting line. However, RPV pressure was very high from mid 14th and drywell pressure reached 650 kPa (gauge), well above design base maximum of 380 kPag, so that at 6.14 am on Tuesday 15th, unit 2 apparently ruptured its pressure suppression chamber under the actual reactor, releasing significant radioactivity and dropping the drywell pressure inside. At 10.30 am Tepco was told to inject water into the pressure vessel and to vent the containment, and these activities continued. Containment damage is suspected. Since about 17th RPV pressure has been atmospheric, and drywell pressure about 200 kPa (100 above atmospheric).  On 1 April Tepco discovered a crack in the wall of a 2m deep cable pit which was leaking highly-contaminated water to the sea, apparently from the reactor itself. The company eventually plugged it early on 6 April after some 4.7 PBq of radioactivity had been released.

In unit 3, the RCIC system failed at 11 am on Saturday 12th and venting the containment was done late in the day. On Sunday 13th, at 5.10 am, water injection using the Emergency Core Cooling System (ECCS) failed in this unit also, after water levels had dropped dramatically, and at 9.20 am venting the suppression chamber was repeated. At 11.55 am fresh water injection to the RPV commenced, soon being replaced by seawater through to 25th. On Monday 14th at 5.20 am venting was repeated to the service floor, though both RPV and drywell pressure remained at about 500 kPa, then at 11.01 am a very large hydrogen explosion on the service floor of the building above unit 3 reactor containment blew off much of the roof and walls and demolished the top part of the building.  This explosion, which mobilized a lot of debris, has not been explained as yet.
In reactor 3 itself, RPV water levels were up and down over 13th but settled at a low level by midday on 14th, with core half uncovered.  Early on Wednesday 16th there was a major release of smoke and/or steam from the top of the building. Because of the possibility that the Primary Containment of Unit 3 was damaged, the operators evacuated from the central control room of Unit 3 & 4 (a shared facility) at 10:45 am, but they returned to the control room and restarted the operation for water injection 45 minutes later. Pressure then built up to 320 kPa in the containment but no further venting was required. Since at least 16th, pressure vessel damage has been suspected, and some leakage is apparently confirmed by radioactivity levels in the building. Also since 20th the reactor pressure and drywell pressure decreased and then remained stable, leading TEPCO to believe that “the reactor pressure vessel is not seriously damaged.”  By the end of March internal pressure was little more than atmospheric, and by 10 April water temperature was below boiling point.
The heat from the fuel in the reactor cores was estimated by France's ISRN as 2.5 MW thermal in unit 1 and 4.2 MW in units 2 & 3 after almost three weeks. This is enough to boil away 95 cubic metres and 160 m³ of water per day respectively if all external circulation to heat sinks ceases, indicating the need for constant top up while the heat is not being dumped in the normal heat exchanger circuits. From 30th, 7-8 m³/hour was being injected into each reactor. These heat production levels will diminish only slowly from this point. The cores remained partly uncovered, and Tepco said on 15 March that 70% of the fuel rods in unit 1, 30% in unit 2 and 25% in unit 3 were damaged. At one stage unit 2 core was dry for some hours, and a German source said that unit 1 core was without water for 27 hours on 12 March, causing major fuel damage, probably melting some fuel. Units 2 & 3 were without water on 13 & 14 March for 7 hours. Unit 2 is thought to have major core damage also.
Since Sunday 20th JAIF has said that the containment vessel of unit 2 is thought to be damaged, and that of unit 3 could possibly be also, but unit 1 and 4 were intact. (Unit 4 has no fuel in it.).

 

MIT Nuclear Science & Engineering info hub  

The AC electricity supply from external source was connected to all units by 22nd, enabling more accurate monitoring of the plants and progress towards restoring the RHR cooling system of units 1-3. Power was restored to instrumentation in all units except unit 3 by 25th. Tepco said that once the control rooms are operational, water levels can be checked as well as temperatures in the fuel storage pools, and normal cooling of those pools can be resumed. However, at least some pumps have been damaged by seawater and Tepco says that radiation levels inside the plant are so high that normal access is still impossible.  It is giving priority to removing contaminated water so as to allow better access.
Tepco has said that it is inevitable that the three reactors, with unit 4, will be written off and decommissioned. On 4 April both NISA and NSC said it could take months to repair the cooling systems due to the amount of contaminated water in the turbine buildings.
Summary on 25 April: All three units have fuel damage and low water levels. The fuel remains essentially contained except for some volatile fission products vented early on or released from unit 2 in mid March, and some soluble ones which are leaking with the water, especially from unit 2, where the containment is evidently breached. Cooling still needs to be provided from external sources, using fresh water and pump trucks, while work continues to establish a stable heat removal path to external heat sinks.  Radioactive water in the turbine buildings is hampering this work.

Fuel ponds: background  Used fuel needs to be cooled and shielded. This is initially by water, in ponds. After about three years under water, used fuel can be transferred to dry storage, with air ventilation simply by convection. Used fuel generates heat, so the water is circulated by electric pumps through external heat exchangers, so that the heat is dumped and a low temperature maintained. There are fuel ponds near the top of all six reactor buildings at the plant, adjacent to the top of each reactor so that the fuel can be unloaded under water, when the top is off the reactor pressure vessel and it is flooded. The ponds hold some fresh fuel and some used fuel, pending its transfer to the central used/spent fuel storage on site. (There is some dry storage there to extend the site's capacity.) From the plant the used fuel is intended to be shipped periodically for recycling. Tepco and JAPC are building a Recyclable Fuel Storage Centre in Mutsu, operating from mid 2012 with 5000 t capacity. The JPY 100 billion facility will provide interim storage for up to 50 years before used fuel is reprocessed at Rokkasho. NISA approved this in August 2010. Until it is finished and operational in 2012 there has been a build-up of used fuel at reactor sites. The Rokkasho plant has been much delayed, and is now expected in commercial operation in October 2012. There is some storage capacity there, though this may be full. At the time of the accident, in addition to a large number of used fuel assemblies, unit 4's pond also held a full core load of 548 fuel assemblies while the reactor was undergoing maintenance, these having been removed at the end of November. The temperature of these ponds is normally low, around 30°C when circulation is maintained with the Fuel Pool Circulation and Clean-up (FCP) system, but they are designed to be safe at about 85°C in the absence of pumped circulation (and presumably with moderate fuel load). They are about 12 metres deep, so the fuel is normally covered by 7 metres of water. Unit 2, 3 & 4 ponds are about 12 x 10 metres, with 1240, 1220 and 1590 assemblies capacity respectively (unit 1 is about 12 x 7 m, 900 assemblies). Unit 4 pond contains a total 1331 used assemblies (783 plus full fuel load of 548), giving it a heat load of about 3 MW thermal, according to France's IRSN, which in that case could lead to 115 cubic metres of water boiling off per day, or about one tenth of its volume. Unit 3's pool contains 514 fuel assemblies, unit 1 has 292 and unit 2 has 587, giving it a heat load of 1 MW. There is no MOX  fuel in any of the ponds. Two of the reactor unit ponds (2 & 4) were unusually full before unit 4 core was unloaded in November, since there was little spare space (only for 465 assemblies) in the central fuel storage pond on site.  Thus there was a lot more fuel in the reactor ponds with correspondingly high heat loads and cooling requirements than might have been the case. Moving the used fuel from reactor ponds to central storage involves loading it under water into casks which are lowered down and trucked the short distance (see RH side of cutaway diagram above). It requires access from the service floor and the use of cranes which were damaged in the hydrogen explosions. The central fuel storage on site near unit 4 has a pond about 12 x 29 metres, 11 m deep, with capacity of 3828 m3 and able to hold 6840 fuel assemblies. Its building is about 55 x 73 m. (There are 6375 assemblies in the undamaged central pool storage on site, with very low decay heat, and 408 in dry cask storage - utilized since 1995 for used fuel no longer needing much cooling.)

Fuel ponds: developing problems

A separate set of problems arose as the fuel ponds, holding fresh and used fuel, in the upper part of the reactor structures were found to be depleted in water. It has not been explained why or how the low water levels came about, though elevated temperatures due to loss of cooling circulation would have been a major cause, especially in heavily-loaded unit 4.  On 13 March Tepco said it was consulting with NISA and METI regarding cooling of the ponds
In unit 4, at about 6 am on Tuesday 15th, there was an explosion in the top part of the building, near the fuel pond, which destroyed the top of the building and damaged unit 3's superstructure further. Then there was a fire and soon after the radiation level near the building reached 400 mSv/hr, apparently from this source. The fire was extinguished in three hours. Evidently the used fuel there became exposed and got hot enough to form hydrogen. At 10 pm on 15th Tepco was told to implement injection of water to unit 4 pond. The gates which give access in refueling may have been disturbed by the earthquake, or some other damage may have occurred, allowing some water to drain out. Apart from this, it appears that the water level dropped due to evaporation, if not boiling, caused by the high heat load from 1331 fuel assemblies once circulation ceased.
The focus from Tuesday 15th was on replenishing the water in the ponds of units 3 and 4, through the gaps in the roof and cladding, using seawater. On 19th, unit 3 pond was reported to be stable, and on 24th it was able to be replenished using built-in plumbing for FPC system. On 25th the same was achieved for unit 2 pond. These ponds, 12 x 10 metres, were not an easy target for ground-based fire pumps, but the arrival of a concrete pump with 58-metre boom on 22nd enabled more precise replenishment in units 1, 3 & 4 that had damaged walls. Unit 2 pond had been topped up internally. On 19th, the temperature of unit 4's pond had come back to 48°C.  On 25th, water from a nearby dam started to be used instead of seawater.
The pond at unit 4 is the main focus of concern now. It needs continual top-up with water, but at the same time there is concern about the structural strength of the building, which has been weakened either by the earthquake or the hydrogen explosion. Some 195 m³ was added to the pond on 13 April, about 20% of its capacity, and another 140 m³ on 15 and also on 17 April, by concrete pump. Another 100 m³ went in on 20 April, then 200 m³ on 22nd, 140 m³ on 23rd, 165 m³ on 24th and 210 m³ on 25th. Temperature has been up to 90°C and water level 5 metres down. It is not clear whether the main water loss is from leakage or boiling. However, Tepco reports that analysis of radionuclides in water from the used fuel pool of unit 4 suggests that some of the fuel assemblies there may have been damaged, but the majority are intact.
On 18th, Tepco made holes in each of the superstructure roofs of units 5 & 6 to allow ventilation of any hydrogen, though pond temperatures had reached only about 69°C. On 19th the residual heat removal pumps for units 5 & 6 ponds were restarted as power was restored, and temperatures then declined.  
The central spent fuel pool holds about 60% of the used fuel on site, and is immediately west (inland) of unit 4.  It lost circulation with the power outage, and temperature increased to 73°C by the time mains power and cooling were restored on 24th. The pool was topped up on 21st.
Summary on 25 April: Spent fuel ponds in units 3 & 4 still need to be topped up repeatedly, with some use of internal plumbing for units 2 & 3 and by concrete pump with boom for unit 4.  The pond heat exchangers for units 1, 3 & 4 are very damaged.  There is concern about the structural integrity of the unit 4 building supporting its pond.

Radioactivity Radioactive releases are measured by the amount of (radio)activity in the material, and quoted in Becquerels. Whether this is in the air or settled on the ground, it may expose people to ionizing radiation, and the effect of this is measured in Sieverts, or more typically milliSieverts (mSv). Exposure to ionizing radiation can also be by direct radiation from the plants and fuels themselves, though not released to the environment. This is only a hazard for those on the plant site, and the level diminishes with distance from the radioactive source. It is the chief hazard for the plant workers, who wear film badges so that the dose can be monitored. A short-term dose of 1000 mSv is about the threshold of acute radiation syndrome (sickness). The main radionuclide released from among the many kinds of fission products in the fuel is volatile iodine-131, which has a half-life of 8 days. It has been in both venting to air and in water. Iodine-131 decays to inert and stable xenon-131.  The other main radionuclide is caesium-137, which has a much longer half-life and may contaminate land for some time.  In assessing the significance of atmospheric releases, the Cs-137 figure is multiplied by 40 and added to the I-131 number to give an "iodine-131 equivalent" figure.

 Radioactive releases to air
After the hydrogen explosion in unit 1, some radioactive caesium and iodine were detected in the vicinity of the plant, indicating fuel damage. This material had been released via the venting.  Further I-131 and Cs-137 and Cs-134 were apparently released during the following two weeks, particularly following the explosion at unit 3 on 14th and the apparent rupture of suppression chamber of unit 2 on 15th. The caesium was at low levels (about two orders of magnitude less than the iodine). The hydrogen explosion in unit 4 involving the spent fuel pond on 15th apparently added to the airborne radionuclide releases.  But generally speaking, there remains a lot of uncertainty about the exact source of radioactive releases.
On 16th, Japan’s Nuclear Safety Commission recommended local authorities to instruct evacuees under 40 years of age leaving the 20 km zone to ingest stable iodine as a precaution against ingestion (eg via milk) of radioactive iodine-131. The pills and syrup (for children) had been pre-positioned at evacuation centers. The order recommended taking a single dose, with an amount dependent on age.  On 11 April the government suggested that those outside the 20km zone who were likely to accumulate 20 mSv dose should move out within a month.  It is not clear that there are hazardous levels of radioactivity in the evacuation zone beyond some limited areas, and the Cabinet Secretary has confirmed that the evacuation zone has no direct relationship to dose rates being measured. However, he said that the 84,000displaced residents will not be able to return home for at least six months, though some residents have already been returning unofficially. On 21 April the government said that "any residents who chose to remain in the zone would be required to leave because the damaged reactors have not been stabilized," and it apparently perceives an ongoing risk from them
France's IRSN has estimated that maximum external doses to people living around the plant are unlikely to exceed 30 mSv/yr in the first year. This is based on airborne measurements between 30 March and 4 April, and remains to be confirmed on the ground. It compares with natural background levels mostly 2-3 mSv/yr, but ranging up to 50 mSv/yr.
On 17 March NISA raised the statutory intervention limit from 100 to 250 mSv per person after consultation with health experts, to allow work to be carried out on the Unit 4 reactor.
Gamma radiation on site close to the reactors decreased greatly when unit 3 fuel pond was replenished with water on 19th. On Sunday 20th, levels were mostly below 3 mSv/hr and on 21st were nearly down to 2 mSv/hr about 500 m north on unit 3.
IAEA reported on 19 March that airborne radiation levels had spiked three times since the earthquake, notably early on 15th (400 mSv/hr near unit 3), but had stabilized since 16th at levels significantly higher than the normal levels, but within the range that allows workers to continue on-site recovery measures. For instance, NISA reported 12 mSv/hr dose rate on the site boundary early on 14th, then 3.4 mSv/hr mid 16th, dropping to 0.65 mSv/hr 13 hours later at the same point.  Late on 24th it was about 0.2 mSv/hr at the front gate, having been ten times that a few days earlier. Steady decrease has been maintained - on 4 April it was 0.12 mSv/hr at the front gate and 0.05 mSv/hr at the west gate.  On 17 April dose rates at eight monitoring points around the boundary ranged from 0.01 at the north end to 0.19 mSv/hr at the south.
Tepco said on 13 April that it estimated the total inventory upon shutdown as 81 EBq of I-131, of which 130 PBq is estimated by NISA to have been released from the reactors, mostly around 15 March and the two days following - 0.16% of total. In 32 days this released iodine would have diminished to one sixteenth of original activity - 8 PBq, while Tepco estimated that 440 PBq remained inside.  NISA's report to IAEA said 130 PBq of I-131 and 6 PBq of caesium-137 had been actually released, giving an "iodine-131 equivalent" figure of 370 PBq, a figure which resulted in the re-rating of the accident to INES level 7. The NSC estimated that 12 PBq of Cs-137 had been released, giving an "iodine-131 equivalent" figure of 630 PBq. This is about ten percent of the Chernobyl figure. The NSC said that most radioactive material was released from the unit 2 suppression chamber during two days from its apparent rupture early on 15 March.
Tepco has been spraying a dust-suppressing polymer resin around the plant to ensure that fallout from mid March is not mobilized by wind or rain.  It hopes to finish spraying around the reactor buildings by the end of May and around the rest of the site by late June. Tepco says that radiation levels around the plant site remain relatively low.  In addition it is removing rubble with remote control front-end loaders, and this is further reducing ambient radiation levels. Much of the debris around the former office building has been removed, and it has started clearing the rubble around the Number 3 and 4 reactors.
 Managing contaminated water
Removing contaminated water from the reactor and turbine buildings had become the main challenge in week 3, along with contaminated water in trenches carrying cabling and pipework. Run-off from the site into the sea was also carrying radionuclides well in excess of allowable levels. Over 1-6 April some 520 m³ of contaminated water from unit 2 with 4.7 PBq of activity leaked into the sea until the source was sealed. By the end of March all storages around the four units - basically the main condenser units and condensate tanks - were largely full of contaminated water pumped from the buildings.
Accordingly, with government approval, Tepco over 4-10 April released to the sea about 10,400 cubic metres of slightly contaminated water in order to free up storage for more highly-contaminated water from unit 2 reactor and turbine buildings which needs to be removed to make safe working conditions. This is the main source of contaminated water, though some of it comes from drainage pits. Tepco said that although the level of iodine-131 in the release to the sea was about 100 times the legal limit, this would be safe, and that eating fish and seaweed caught near the plant every day for a year would add some 0.6 mSv to the dose from natural background.  NISA confirmed that there was no major change in radioactivity levels in the sea as a result of the 0.15 TBq discharge.
Tepco then began transferring highly-radioactive water from the basement of unit 2 turbine hall to the holding tank just south of unit 4. This was expected to take 26 days, at 480 m³/day. The water contains 3 TBq/m³ of I-131 and 13 TBq/m³ of Cs-137. About 168 m³/day of fresh water is injected into unit 2 reactor core and it is assumed that this replenishes the contaminated water that is being removed. Tepco said it expects to complete construction of a second waste treatment facility by June to receive another 15,000 cubic metres of contaminated water from unit 2 and an additional 45,000 cubic metres of less-contaminated water from turbine buildings of units 1 and 3. The company will initially use coprecipitation methods to decontaminate the water, but by June these will be complemented by ion-exchange equipment. Decontaminated water will be used for re-injection. It has accepted a proposal from Areva to supply a further treatment plant.
Tepco has installed double-layer silt barriers on the inlet canal and in front of seawater inlet bar screens of units 1-4 to impede leakage to the sea. Also, steel plates have been installed on the inlet bar screen of unit 2.
Radiation exposure in plant and beyond

On 24 March three contractors laying cable in unit 3 received a dose of more than 170 mSv, two suffering beta radiation burns on their legs from contaminated water. The Fukushima Labour Bureau issued a stern rebuke to Tepco.  By 22 April, 29 workers had received doses over 100 mSv, and none had reached 250 mSv. There were 245 workers on site.

No radiation casualties (acute radiation syndrome) had been reported, and few other injuries, though higher than normal doses are being accumulated by several hundred workers on site.  Monitoring of seawater, soil and atmosphere is at 25 locations on the plant site, 12 locations on the boundary, and others further afield.
Government and IAEA monitoring of air and seawater is ongoing, with high but not health-threatening levels of iodine-131 being found. Environmental levels are decreasing. With an 8-day half-life, most I-131 has aleady gone. However a radiological hotpsot was identified near Iitate village 30 km northeast of the plant.
On 4 April radiation levels of 0.06 mSv/day were recorded in Fukushima city, 65 km northwest of the plant, about 60 times higher than normal but posing no health risk according to authorities.  Monitoring beyond the 20 km evacuation radius to 13 April showed one location - around Iitate - with up to 0.266 mSv/day dose rate, but elsewhere no more than one tenth of this.  The safety limit set by the central government in mid Aril for public recreation areas is 3.8 microsieverts per hour (0.09 mSv/day).
Potential radiation doses around Fukushima Daiichi to 19 April
Fukushima Daiichi 5 & 6
Units 5 & 6, in a separate building, also lost power on 11th due to the tsunami. They were in 'cold shutdown' at the time, but still requiring pumped cooling. An emergency diesel
 generator of Unit 6 was repaired on Saturday 19th, allowing full restoration of cooling for units 5 and 6. While the power was off their core temperature had risen to over 100°C (128°C in unit 5) under pressure, and they had been cooled with normal water injection, presumably with the RCIC system. They were restored to cold shutdown by the RHR system on 20th, and mains power was restored on 21-22nd.  Tepco said that local people would be consulted on whether the reactors might be restarted.

Fukushima Daini plant

Units 1-4 were shut down automatically due to the earthquake, but some interruption to cooling occurred due to the tsunami. The government ordered an evacuation within 10 km (in practice this was within Daiichi 20 km zone).
In units 1, 2 & 4 there were cooling problems still evident on Tuesday 15th. Unit 3 was undamaged and continued to 'cold shutdown' status on 12th, but the other units suffered flooding to pump rooms where the equipment transfers heat from the reactor heat removal circuit to the sea - the ultimate heat sink.
All units achieved 'cold shutdown' by16 March, meaning core temperature less than 100°C at atmospheric pressure (101 kPa), but still requiring some water circulation.
Radiation monitoring figures remained at low levels, little above background.

International Nuclear Event Scale assessment

Japan's Nuclear & Industrial Safety Agency originally declared the Fukushima Daiichi 1-3 accident as Level 5 on the International Nuclear Events Scale (INES) - an accident with wider consequences, the same level as Three Mile Island in 1979.  The sequence of events relating to the fuel pond at unit 4 was rated INES Level 3 - a serious incident.
However, a month after the tsunami the NSC raised the rating to 7 for units 1-3 together, 'a major accident', saying that a re-evaluation of early releases suggested that some 630 PBq of I-131 equivalent had been discharged. This included 130 PBq of I-131, and 6.1 PBq of Cs-137, presumably with significant amounts of noble gases. This then matched the criterion of "environmental release corresponding to a quantity of radioactivity equivalent to a release to the atmosphere of more than several tens of thousands of terabequerels (ie PBq) of iodine-131", most of this being in the first week. The IAEA quotes a more specific 50 PBq of I-131 equivalent.
For Fukushima Daini, NISA declared INES Level 3 for units 1, 2, 4 - each a serious incident.
Accident liability

Beyond whatever insurance Tepco might carry for its reactors is the question of third party liability for the accident. Japan is not party to any international liability convention but its law generally conforms to them, notably strict and exclusive liability for the operator. Two laws governing them are revised about every ten years: the Law on Compensation for Nuclear Damage and Law on Contract for Liability Insurance for Nuclear Damage. Plant operator liability is exclusive and absolute (regardless of fault), and power plant operators must provide a financial security amount of JPY 120 billion (US$ 1.46 billion) - it was half that to 2010. Beyond that amount, the government may provide coverage if damage results from “a grave natural disaster of an exceptional character”, and in any case liability is unlimited.
In mid April, the first meeting was held of a panel to address compensation for nuclear-related damage. The panel will establish guidelines for determining the scope of compensation for damage caused by the accident, and to act as an intermediary. It was established within the Ministry of Education, Culture, Sports, Science and Technology (MEXT), and is led by Law Professor Yoshihisa Nomi of Gakushuin, University in Tokyo. Although a nuclear operator (in this case Tepco) is primarily responsible for the nuclear damage caused by the operation of its reactors and other equipment, the accident at the Fukushima NPS was caused by a natural disaster. Based on indemnity agreements with the government for the compensation of nuclear damage, the concept of “financial security” will be applied, limiting Tepco's liability to JPY120 billion (USD1.46 billion).

Tepco's 17 April remediation plan
Tepco has published a 6- to 9-month plan for dealing with the disabled Fukushima reactors.
Cooling reactors
In units 1 & 3 the containment vessels will be flooded up to the level of the top of the fuel, and nitrogen injection will be continued. In unit 2 the damaged containment will be sealed with grout, before similarly flooding. New heat exchanger circuits will be built for all three units. Cold shutdown target in 6-9 months. The plan does not mention removal of fuel from the reactors.

Cooling and removing spent fuel
For all four units water injection will be improved to the spent fuel ponds in each reactor building and cooling circulation for heat removal restored, with new or restored heat exchangers. A support structure will be built under unit 4 pond. Fuel will then be removed to central storage on site.

Managing contaminated water
Further storage capacity will be installed, along with treatment plant to enable recycling. Reactor 2 is a particular focus. Slightly-contaminated water will be treated. "Full-fledged" treatment plant will follow.

Minimising release of radioactive materials to atmosphere
Dust-suppressing polymer resin will continue to be applied, and debris removed to improve working conditions on site. A light temporary structure will then be built over reactor buildings 1, 3 & 4, followed by a more substantial structure.

Preparing for return of evacuees
Monitoring will be expanded and the evacuation zone will be decontaminated where required so that evacuees can return as soon as possible.

Consortia led by both Hitachi-GE and Toshiba have submitted proposals to Tepco for decommissioning units 1-4. This would generally involve removing the fuel and then sealing them for a decade or two while the activation products in the steel of the reactor pressure vessel decay. They can then be demolished. The Hitachi-GE group includes Bechtel and Exelon, the Toshiba team includes Babcock & Wilcox and Shaw. Areva is also planning to submit a proposal.

Post-construction modifications

In the last ten years many national regulators have required licensees to go well beyond the original design basis in ensuring the safety of nuclear plants in a variety of accident and assault scenarios. It is not yet clear to what extent NISA had required Tepco to modify the original 1960s design in ways which might have helped avoid the accidents, and to what extent Tepco had accomplished changes, though it is understood that all early modifications, recommended to the 1980s at least, were carried out here as in the USA. Certainly regarding direct seismic vulnerability a lot had been done as mentioned above in Earthquake background.  Also as a result of tsunami analysis in recent decades, the design basis height was increased from 3.1 to 5.7 m and Tepco moved pump motors and other equipment higher accordingly.

Sources:
Tepco
NISA
IAEA
METI
JAIF
NSC