| The official releases, which are worth introducing for assessment from a scientific point of view, have decreased further this week. Although there are a large number of news reports related with Fukushima Dai-ichi disaster, most of them are covering political issues as well as amelioration/decontamination issues. I will continue to watch important releases in Japan, however, it may be a time for me to reduce this periodical update to "event-driven" update. In this update, I focused on my assessment of root cause of hydrogen explosion in 1F1, although it is not driven by a new information obtain this week. Currently I intend to send next update in a week or so. I. Hydrogen explosion scenario of Fukushima disaster - a recap of the proposed hypothesis 1. Background 1.1 Gross fuel meting scenario is increasingly unlikely Recently TEPCO published an updated report of internal Investigation Committee on Fukushima Disaster, describing that the IC-A (Isolation Condenser A) was re-activated by opening both MO-2A and 3A valves at 21:30 of March 11 and the timely water injection into the secondary side of IC was successful by using D/D FP (diesel-driven fire pump) followed with direct pure water injection into the RPV at 5:46 0f March 12 by employing fire engines. After reviewing this, I am increasingly confident that the gross core melt scenario is unlikely because the reactor core was kept cooled with the water level maintained at least above TAF-1700, not completely exposed the core from the water surface. The lower plenum of the reactor vessel never experienced a dry-out situation. This remarkable finding was explained in Earthquake (172, Dec 30, 2011 - Jan 6, 2012)) and Earthquake (171, Dec 23-30). I also believe that 1F1 did not experience the "long-term station blackout", as defined in the NUREG/CR-9850: Analysis of Long-Term Station Blackout Without Automatic Depressurization at Peach Bottom Using MELCOR (Version 1.8) presented by I.K. Madni Let me quote the relevant description: "In the Long-Term Station Blackout scenario, loss of all off-site and on-site AC power leads to the loss of all active engineered safety features except the steam powered emergency core cooling systems. The latter, however, require DC power for operation and would fail when the station batteries are depleted, which has been estimated at six hours after the start of the accident. Following failure of the emergency core cooling system, the primary system inventory is boiled off through the relief valves by continued decay heat generation. This leads to core un-covery, heat up, clad oxidation, core degradation, relocation, and eventually, vessel failure. This will cause further pressurization of the dry-well from steam and non-condensible gases, which may lead to containment failure. Multiple hydrogen burns can occur in the reactor building after the containment fails." The 1F1 experienced the "loss of off-site power", induced by the seismic damage of transmission lines at 14:46 of March 11 before arrival of the tsunami. The site operator manipulated the IC by remotely open/closed the MO-2A and 3A, as well as MO-2B and 3B valves, as planned. After arrival of the tsunami at 15:37, 1F1 went into a crippled "short-term station blackout," where the decay heat removal through IC was suspended, since the operator could not manipulate the valves without the instrumentation and control DC power. It has not been confirmed whether the HPCI (High Pressure Coolant Injection system) worked even partially or not, perhaps due to flooding of DC power centers. However, the decay heat removal was still automatically performed by "bleeding" steam from the reactor water through SRVs (Safety and Release Valves), where the "safety valve function" of the valves worked without the DC power, with the steam discharging into the suppression pool. At 21:30 of March 11, the operator re-opened both MO-2A and MO-3A valves to continue operation of IC-A, judging that there is a water inventory in the secondary side of IC-A. The operation became possible due to a temporal recovery of the DC power. Therefore, the "short-term station blackout lasted approximately 8 hours. During this period, the water level will be down to TAF-1000〜-2000, in my "back of an envelope" calculation. As a matter of fact, at 16:44 of March 11, the water level inside of RPV was found kept at TAF+250 mm, and again at 21:19 of March 11, at TAF+200 mm, leading the headquarter to judge that IC-A had been working. The temporary confirmed water level was approximately 1〜2 meters higher than my estimation. (The reason for this difference is unknown, however, there is a possibility that HPCI was recovered with the temporal recovery of the DC power. With its large capacity, temporal operation of several minutes should have been sufficient to recover a few meters of water level.) The crippled "short-term station blackout" further continued through 21:35 of March 11, when the D/D FP (Diesel Driven Fire Pump) started to inject fresh water into the secondary side of IC, until at 1:45 of March 12 when the D/D FP was found not running. The water level was recovered to TAF+530(A)/1300(B). At 5:46 of March 12, the direct water injection into the reactor vessel started by employing the fire engines. Although the details of the chronology may change, when TEPCO updated further, it indicate that 1F1 have stayed in the "short-term station blackout" state, with external water injection until now. It never went into the "long-term station blackout" situation with the total core "un-covery", heat up, clad oxidation, core degradation, relocation, and eventually, vessel failure. 1.2 Hydrogen generation is more likely through water radiolysis With the notion that 1F1 successfully evaded from going into the "long-term station blackout", which could have resulted in gross core oxidation, I renewed my interest on the original hypothesis that the source of hydrogen generation is through water radiolysis, where the produced hydrogen was accumulated during the station blackout (loss of AC power) and the hydrogen gas was separated from the reactor water which was kept within the RPV without circulation. I have been continue pointing out this possibility several times in the early phase of this series of update, including in Earthquake (20), sent on March 31. Let me copy and paste the relevant description below. "At the time of station blackout, an estimated decay heat should be down to around 1.3%. There are automatic pressure release and safety valves, which should have acted for ‘bleed cooling’, with battery powered ‘feeding’ until 16:36, resulting in a loss of water injection into the reactor vessel and reported an occurrence of an article #15 event. The ‘bleed cooling’ provided a cooling by releasing steam into the suppression pool water until 0:49 of March 12, until the time when the temperature of the suppression chamber reached at the boiling point of water, which made TEPCO to report an occurrence of another article #15 event. At this moment, an estimated decay heat was down to less than 1%. There is a possibility that a top portion of fuel rods could have lost its submerged condition for approximately 8 hours. Due to an increase in the pressure inside the reactor vessel, it became urgent, at 14:30 of March 12, to vent the steam into a ventilation duct of the Reactor Building Reactor Room Heating and Ventilation System. It is a system which discharges room air from the stack after filtering. Unfortunately, the H&V was out of electric power and most of the vented steam is likely released into the Fuel Handling Floor, located at the top of the Reactor Building, perhaps due to a back flow towards the inlet of H&V duct opened to this floor area. This should be confirmed by consulting with a detailed flow chart of the H&V system. Approximately one hour after the start of venting, at 15:36 of March 12, the hydrogen explosion occurred. A significant amount of hydrogen must have been accumulated at the time of venting. Whether this hydrogen is due to the Zr-water reaction or due to accumulation of hydrogen gas through water radiolysis is in the root of the differences in scenarios. Those people who support the Zr-water reaction scenario argue that a significant oxidization must have occurred by exposing the top half of the fuel rod out of the water level. However, those who argue against this scenario recalls an experimental result of fuel rod tests, partially submerged, sufficiently cooled by steam and droplets coming from the boiling water. This scenario indicates that a self-standing configuration is still maintained, even though some small fraction of fuel rods must have punctured to release rare gases and volatile species, accompanied with a small amount of cracked fuel fragments, into the reactor vessel. A question of how tough the Zr fuel clad should have been is also in the root of these arguments. A group of the Japanese fuel specialists explains that fuel clad is not so fragile to break so easily. Also, the lessons learned from the TMI accident showed that it is necessary for damaged fuels to pile up first, restricting core flow, to start melting. As a matter of fact those fuel rods located in the periphery region of the damaged TMI core kept the self-standing configuration and did not participate in the corium. " Also, in Earthquake (71) - with corrections to the H2 and O2 accumulation, I distributed a "back of an envelope" estimation of the amount of hydrogen generation through water radiolysis for 1F1 as attached. There are many new information accumulated since then, it is necessary to update the calculation made in the previous attachment. I intend to write a paper sometimes this year. Therefore, let me limit my follow on argument only from phenomenology, not by doing a detailed radiation chemistry calculation. 1.3 Some of the notable new information related to the water radiolysis scenario From a phenomenology point of view, let me recap some of the recent findings. 1.3.1 Source of oxygen for potential "internal hydrogen explosion" inside the RPV Although the radiation yield of hydrogen is well known, the yield of oxygen is not directly provided in the radiation chemistry data. The primary g-value does not provide g(O2) but only g(H2O2) and other radicals including G(e, hydrated electron), g(H), g(H2), g(OH) and g(HO2/O2-). Only through solving the detailed reaction chains by reaction rate equations, the final concentration of H2, H2O2 and O2 can be estimated. However, the hydrogen peroxide thus generated should ELECTROCHEMICALLY decompose into water and oxygen, as established in the Pourbaix Diagram (Marcel Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon Press (1966)), although the published data is limited at the ambient temperature. This reaction at the high temperature has been discussed in detail by G. Saji. Degradation of aged plants by corrosion: ’Long cell action’ in unresolved corrosion issues. Nuclear Engineering and Design 239 (2009) 1591–1613. In view of complexities, a stoichiometric volume from the hydrogen concentration was assumed, by referring to the hydrogen-explosion/pipe-rupture accident experience of Hamaoka Unit 1, as introduced in the attached one page summary of Earthquake (69). Another source of oxygen should be the environmental water (fresh- or sea water) injected into the RPV. In the Fukushima Dai-ichi accident sequence, it is likely that the top reactor cover region was filled with hydrogen gas after prolonged reactor isolation. With little oxygen concentration, the gas may not explode immediately. However, the water injection operation using environmental water may have transported oxygen in a form of saturated dissolved oxygen. in a "back of an envelope" calculation, the solubility of oxygen in air in water is approximately 0.0089 g/kg-water at 25℃. The following amount of oxygen gas was transported to the RPV through the water injection before the hydrogen explosion: Unit amount of injected environ. water before explosion amount of injected oxygen stochiometric amount of H2 1F1 80,000 kg 0.7 kg 0.1 kg 1F2 1.0E+06 kg 8.8 kg 1.1 kg 1F3 0.7E+06 kg 6.2 kg 0.8 kg With hydrogen and oxygen gas separated (by boiling) from water and accumulated at the top reactor cover plenum, there should be a risk of hydrogen explosion as occurred at Hamaoka Unit 1 BWR. In this accident, a hydrogen-explosion/pipe-rupture accident occurred during the scheduled ECCS test. Also, there were eight cases of similar hydrogen explosion accidents reported inside TEPCO at the small instrumentation pipes. Although the ignition source has not been identified, I believe a similar hydrogen explosion might have occurred inside of RPV of 1F1. The explosion induced pipe break followed with discharge of the hydrogen gas accumulated inside of RPV as well as PCV, although the leak paths are not known. The released hydrogen gas diffused widely in the reactor building through ventilation ducts. However, let me add that it was found very difficult to reproduce the spontaneous ignition in a confirmation experiment performed after the Hamaoka Unit 1 hydrogen-explosion accident. The spontaneous ignition started only with noble metals, deposited through the noble metal chemistry. With the station blackout in Fukushima Dai-ichi, it is hard to pinpoint exact ignition sources. References: (1) M. Naito et al., Analysis on Pipe Rupture of Steam Condensation Line at Hamaoka-1, (I) Accumulation of Non-condensable Gas in a Pipe, Journal of NUCLEAR SCIENCE and TECHNOLOGY, Vol. 40, No. 12, p. 1041–1051 (December 2003). (2) M. Naito et al., Analysis on Pipe Rupture of Steam Condensation Line at Hamaoka-1, (II) Hydrogen Combustion and Pipe Deformation, Journal of NUCLEAR SCIENCE and TECHNOLOGY, Vol. 40, No. 12, p. 1041–1051 (December 2003). 1.3.2 Venting and hydrogen explosion Many people suspect that the venting triggered a series of hydrogen explosions in 1F1 through 1F3, even including the hydrogen explosion in 1F4, where the reactor core was evacuated into the SFP (Spent Fuel Pool) for repair work of the RPV. I also share this view, since all of the hydrogen accidents occurred hours after venting. The H&V was out of electric power and most of the vented steam is likely released into the Fuel Handling Floor, located at the top of the Reactor Building, perhaps due to a back flow towards the inlet of H&V duct opened to this floor area. This should be confirmed by consulting with a detailed flow chart of the H&V system. The hydrogen gas diffusion mechanism, through H&V ducts, have recently been confirmed by TEPCO, who measured dose distribution along the H&V system as well as the current failure position of ventilation dumpers for 1F3 as reported in: http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_111226_01-e.pdf Although no similar measurement have been released for 1F1, the results support the possible mechanism of hydrogen gas diffusion through ventilation ducts. 1.3.3 Stratification and ignition of hydrogen gas-air mixture This issue has been introduced in Earthquake (172, Dec 26, 2011 - Jan 6, 2012) with a summary of hydrogen explosion study performed for a BWR plant in Finland:. M. Manninen, A. Silde, I. Lindholm, R. Huhtanen, and H. Sjoval, 2002. Simulation of hydrogen deflagration and detonation in a BWR reactor building, Nuclear Engineering and Design 211 (2002) 27–50 (www.elsevier.com/locate/nucengdes) The objective of the work was to investigate, whether hydrogen can form flammable and detonable mixtures in the reactor building, evaluate the possibility of onset of detonation and assess the pressure loads under detonation conditions. The safety concern is, whether the hydrogen in the reactor building can detonate and whether the external detonation can jeopardize the containment integrity. The analysis indicated that the possibility of flame acceleration and deflagration-to-detonation transition (DDT) in the reactor building could not be ruled out in case of a 20 mm2 leakage from the containment. The outcome of the performed 3D analyses of hydrogen distribution (Manninen et al., 2000) was that a significant hydrogen cloud, reaching well the flammability limits, would accumulate in the upper section of the room. The maximum amount of hydrogen in the room was 21 kg. Since the ignition energy of hydrogen deflagration is low, the reactor building room contains several candidate ignition sources, e.g. lamps and valves that may generate static electricity. However, a direct ignition of detonation in an unconfined geometry needs high energies and is not considered easily reachable (CSNI, 2000. State-of-the-art report on flame acceleration and deflagration-to-detonation-transition in nuclear safety. NEA/CSNI/R(2000)7). Thus flame acceleration was considered the greatest threat of reaching supersonic combustion. The purpose of performing 3D deflagration analyses was to study if the flame starts to accelerate in the specified cloud and room geometry. My "back of an envelope" estimations of hydrogen generation through water radiolysis indicated that the amount of hydrogen were in general not large enough to fill the the Reactor Building with a flammability limit of 4%. However, when the stratification is considered, it is a different story. The stratification should occur much easier, in Fukushima, since the heating and ventilation was stopped due to the station blackout. In addition, the initial temperature of the released hydrogen gas should have been around 100℃ to facilitate thermal stratification and specific volume differences. The condensation of steams at the building walls facilitate the hydrogen gas separation. However, actual building configuration of the Reactor Building is much more complex in Fukushima Daiichi, since many of the rooms are interconnected with ventilation ducts of a huge Reactor Building Heating and Ventilation System, through which the radioactive gas has spread by reverse flowing from the venting ducts as recently revealed in the TEPCO's dose measurement in the system: http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_111226_01-e.pdf 2. A hypothetical hydrogen explosion scenario 2.1 Intrinsic safety It is my strong belief that the magnitude of the accident and its sequence depends strongly on the intrinsic safety of the reactor, the safety provisions in place as well as operational issues. For example, the Chernobyl accident was initiated due to the vulnerability of the graphite-moderated water-cooled channel type reactors, combined with the positive void coefficient and violation of an extended low power operation by bypassing the ECCS. Among all the lessons learned from the Chernobyl accident, the most significant one is the importance of intrinsic safety characteristics as a prerequisite of safe nuclear power plants. Intrinsic safety is an essential part of prevention, in particular, the reactor should have self-limiting physics characteristics which prevent a large energy release beyond the control of even in an abnormal or accidental situation. (G. Saji, 2005. Management of Nuclear Risks by Intrinsic Safety, Siting, and Defense-In-Depth for Future Reactors: Lessons Learned from the Chernobyl Accident. In the Proceedings of the International Topical Meeting on Probabilistic Safety Analysis, PSA’05, 11-15 September 2005, USA) . Let me attach a preprint of my previous paper. Likewise, the Fukushima Dai-ichi accident has revealed the strong effects of intrinsic safety. 2.1.1 Intrinsic safety and design provisions with respect to CHC in BWRs One of the most important issues in the intrinsic safety characteristics of BWRs is in coping with the radiolytic decomposition of water. Since the water chemistry is not kept above the CHC (critical hydrogen concentration, which is the minimum concentration of dissolved hydrogen required to prevent the net radiolytic breakdown of water. Radiolysis is said to be in suppression when there is no net decomposition of the water due to the addition of excess hydrogen. ) , it is essential to incorporate active systems for removal of hydrogen during the entire operational modes, including the accidental situations. Such an active hydrogen removal system is necessary just as in the case of DHRS (decay heat removal system). During normal operation, the hydrogen is removed at the turbine condenser as an incondensable gas to keep high vacuum necessary for operation of turbine. This is performed by steam ejectors and the removed hydrogen gas is recombined into water before releasing into the environment. However, after the reactor isolation such as in the case of "loss of offsite power", the reactor is separated from the hydrogen removal system due to isolation. According to the PSAR of a BWR with Mark I containment vessel, there are "Burnable Gas Concentration Control System" and "Emergency Gas Processing System" (also called "Standby Gas Treatment System"), both of which appears to be designed to remove or purge from the PCV (Primary Containment Vessel) in an event of LOCA. The "Emergency Gas Treatment System" can vent the containment atmosphere after filtration to the stack. However, they are designed to release over-pressure of the PCV and not for removal of hydrogen from the RPV in isolation. With this negligence of safety function, the hydrogen gas produced through water radiolysis continue to accumulate inside of RPV and in the suppression chamber. With extended "station blackout", the accumulated hydrogen gas started to increase until reaching the CHC, from thereafter the hydrogen production rate should have reduced significantly. This timing appears to have started around March 22, according to the environmental monitoring data. The dose rates are reduced more than an order of magnitude from this date. I have been calling this time span as the "active period", d,ays between March 11 to 22. After the active period of ten days, spikes of environmental releases occurred on March 22, 23, 25 (twice), 28, 29 and 30, although from which unit has not been identified yet. Since the hydrogen generation should have continued, with much reduced rate, this recurrence of large release should be related with the environmental water injection into the RPV. (If the water was in circulation within a closed loop, such recurrence should not occur since the dissolved oxygen can be easily removed by irradiation.) After an "internal hydrogen explosion", the system goes into another incubation period to recover the oxygen and hydrogen concentration. This anomalous hydrogen gas oxidation appears to continue even now. TEPCO reported a mysterious gas temperature hump in the containment atmospheric temperature in http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_120106_01-e.pdf One possibility is that the hydrogen gas is burning slowly inside of the PCV. In the recent days, it is becoming more and more common to use the "hydrogen water chemistry" to mitigate IGSCC (Inter-Granular Stress Corrosion Cracking) in BWRs, however, the hydrogen concentration is limited, since it results in higher dose rate in turbine halls, due to increased N-16 concentration in the seam. Although the CHC is an important radiation chemical data, I have not found the report on this subject for BWR. For PWR, in recent studies with the goal of introducing an optimized DH level, the Studvik Nuclear team in conjunction with the Japan Atomic Power Company’s (JAPC) team performed in-pile experiments as well as model calculations for the radiolysis of the PWR-simulated primary circuit at the JAPC’s Unit 2 of the PWR Tsuruga [Takiguchi, H., Sekiguchi, M., Christensen, H., Flygare, J. Molander, A. and Ullberg, M., 2000. In-pile loop experiment and model calculations for radiolysis of PWR primary coolant, in: Conf. Water Chem. of Nucl. Reactor Sys. 8, BNES, Bournemouth; Takiguchi, H., Ullberg, M. and Uchida, S., 2004. Optimization of dissolved hydrogen concentration for control of primary coolant radiolysis in pressurized water reactors”, J. of Nucl. Sci. and Tech., Vol. 41, No. 5. P.6-1~6-9.] Their suggested value is below 6 cm3-STP/kg-1, is still debatable with only three widely scattered experimental values. I have also investigated CHC in pure water for fusion environment (G. Saji, 2012. Scientific Basis of Corrosion Control for Water-Cooled Reactors such as ITER. ICONE20-55046 20th International Conference on Nuclear Engineering, July 30 - August 3, 2012, Anaheim, California, USA (To be published) It is not known whether BWRs can be operated above the CHC to improve the intrinsic safety. If not feasible, active systems for hydrogen removal should become more important. Perhaps diesel-driven incondensable gas treatment system may become necessary to be able to prepare for extended station blackout. 2.1.2 Intrinsic safety and design provisions with respect to decay heat removal in 1F1 In the Long-Term Station Blackout scenario, loss of all off.site and on-site AC power leads to the loss of all active engineered safety features except the steam powered emergency core cooling systems. The latter, however, require DC power for operation and would fail when the station batteries are depleted. This threat was barely ameliorated by installing auto batteries to recover essential instrumentation and control power in 1F1. The plant also had a D/D FP (diesel driven fire pump) which was used for water injection into the secondary side of the IC (Isolation Condenser). The IC is a passive cooling system with a bundle of cooling tubes installed inside. The cooling water was designed to be supplied from the Reactor Components Cooling System, which became unavailable due to the "loss of ultimate heat sink." Fortunately, the IC had a large water inventory which was replenished with the D/D FP (diesel driven fire pump). Without the heat sink, the water inventory boiled off by continued decay heat generation at the core. These prudent design has obviously were effective in preventing the reactor into a worst possible situation of gross core melting. In addition to IC, the safety and relief valves automatically released steam from the RPV into suppression pool, through "feed and bleed" cooling capability. Although the tsunami induced "loss of ultimate heat sink" combined with "station blackout", 1F1 was successful in evading the worst scenario of the "long-term station blackout," thanks to prudent safety provisions and adequate operator actions. Although it is impossible to stop the decay heat, the safety features prepared for decay heat removal were effective by prolonging the time for core un-covery. II. Re-classification of evacuation zones On January 8, Yomiuri Shinbun reported the new evacuation zone classification schemes being coordinated with the local governments. The new scheme divides the previous "vigilance" zone of 10 km radius and "scheduled (and organized) evacuation" zone into: (1) "Difficult Resettlement" Zone (>50 mSv/y) Applies to those areas where more than 5 years is anticipated for the dose rate to decrease below 20 mSv/y. The Government is considering purchasing the affected properties. Estimated number of affected residents will amount to 25,000 people, 30% of the evacuees, with the following breakdown: Iidate-mura (900, 10%), Minami-Soma-shi (70, 0.1%), Namie-machi (4400, 20%), Kuzuo-mura (200, 10%), Futaba-cho (4800, 70%), Ookuma-cho (10,000, 90%) and Tomioka-machi ( 30%) (2) "Restricted Re-settlement" Zone (20-50 mSv/y) Applies to those areas where several years are expected for the dose rate to decrease below 20 mSv/y. Temporary home visits will be allowed. Re-settlement can be made when the dose rates are reduced by decontamination. An estimated number of affected residents is 30.000 people. (3) "Preparing Lift of Evacuation Order" Zone ( |
|
