Fukushima: Failure by Design

Editors note: This post originally appeared on Peter’s Blog as part of a series investigating the technical details at Fukushima Daiichi.

This post will conclude my quatrology of essays touching on peculiarities of the reactor accidents at the Fukushima I (Daiichi) nuclear power station on the northeastern coast of Japan as a result of the Tohoku-Oki Earthquake and Tsunami on Mar. 11, 2011. I will continue to update the content of the series as new information crucial to the discussed issues becomes available.

This essay once more is of technical nature, seeking to examine potential system failures that led to the fuel core melt-downs and melt-throughs at the three units that were producing power at the Daiichi station on the day of the earthquake and tsunami and suffered catastrophic failures in their wake. The fourth unit with extensive damage was shutdown at the time and the fuel was stored in its spent fuel pool. Forceful explosions resulted from excessive accumulation of hydrogen in the buildings of the four units. The hydrogen was produced by radiolysis and oxidation of cladding of superheated fuel rods. Radioactive material in amounts only superseded by the Chernobyl reactor accident in 1986 was released into the atmosphere and into the ocean with yet unfathomable consequences for public health. The lives of hundreds of thousands of people living around the plant will be profoundly affected. Fukushima Prefecture alone is home to roughly 3 million people.

The calamity has been unfolding on the heel of the utter destruction of coastal villages and towns owing to the earthquake and tsunami. According to the Report of Japanese Government to the IAEA Ministerial Conference on Nuclear Safety, page III-10, released last month, 24,769 persons were reported dead or still missing on May 30, 2011. The examination of technicalities of the reactor disaster must pale compared with the human toll exacted.  However, this essay may help shine a light on shortcomings in design that may pertain as well to other nuclear power reactors around the world and help prevent yet another calamity that added in unprecedented ways to this human tragedy of mind-shattering proportion.

BWR reactor schema with Mark I bulbous primary containment and ring-shaped torus which contains a suppression pool used to depressurize the containment when needed.

The four greatly damaged boiling water reactors, BWRs for short, at Fukushima Daiichi nuclear power station possess the same bulbous Mark I primary containment with a massive ring-shaped suppression chamber, called torus, underneath. However, major differences in reactor protection system design were implemented based on the time of construction. Unit 1 is the oldest. Indeed, the unit represents the oldest commercial nuclear power reactor in Japan. It was completed in 1970 and is considered a third generation BWR (BWR-3). Among the nuclear reactors in the US, the one at the Oyster Creek nuclear power plant, Lacey Township, New Jersey, which is considered a BWR-2, and Unit 2 at the Dresden nuclear power plant, Morris, Illinois, which is considered a BWR-3, appear to resemble the design of Daiichi Unit 1 most closely. By contrast, Units 2, 3 and 4 follow the more recent BWR-4 design, which is most common for operating BWRs in the US today.

The video below provides an overview on the layout of power stations with BWR-3 reactors and Mark I containments in the US and a good impression of the structures’ dimensions:

Though Units 1, 2 and 3 incurred loss of coolant accidents with fuel core melt-downs and, very possibly, melt-throughs penetrating their primary containments, differences in design, particularly of systems that cool reactor cores and vent gas and steam, may help explain the observed differences in event sequence, timeline, and thrust.

The reactor turbine feed loop represents the main circulation loop for steam and water during normal reactor operations, transferring the heat generated by nuclear fission to the turbine that drives the electrical power generator. Water condensed in the turbine condensers continuously replenishes coolant in the 500-m3 RPV, preventing the fuel core from overheating. In addition to serving as coolant, water is used to moderate neutron fluxes between the fuel rods, enabling a controlled, sustained nuclear chain reaction.

After a seismic SCRAM like at Fukushima, main steam isolation valves shut the turbine loop off. Though the safety rods are inserted to stop nuclear fission, radioactive decay of neutron-activated radioactive isotopes continues to produce heat which keeps the reactor water temperature above boiling. Steam and pressure build in the RPV. Once a setpoint is reached, the safety/relief valves of the automatic depressurization system (ADS) succinctly depressurize the RPV to prevent damage to the vessel, venting steam into the suppression pool. As a consequence of the depressurization, the water level in the RPV falls rapidly unless water is injected. The emergency core cooling system, or ECCS for short, consists of several interrelated standby systems that can be used to achieve replenishment with auxiliary feedwater.

Schema showing IC (BWR-3) and RCIC (BWR-4) standby cooling systems (Stoll U, 2011).

In addition to high pressure coolant injection (HPCI) and massive low pressure coolant injection (LPCI) pumps, the isolation condenser (IC) constitutes a third-line option of injecting auxiliary water into the RPV during the shutdown of BWR-3s like Daiichi Unit 1. The IC condenses steam from the suppression pool and reintroduces the water at the top of the RPV entirely passively driven by convection and gravity. No pumps are needed. According to the scram log of Unit 1, the IC was initiated immediately after the seismic SCRAM, but had run dry for unknown reasons within two hours (see previous post with the title “Fukushima: Failure of the Mind” published May 17, 2011).

In BWR-4s, by contrast, the IC was replaced with the reactor core isolation cooling (RCIC) system [USNRC Technical Training Center, Boiling Water Reactor GE BWRA4 Technology Technology Manual (Rev 0197), Chapter 2.7: Reactor Core Isolation System]. The main component of the RCIC system is a turbine-driven pump which can discharge water through the uppermost feedwater line into the RPV at 400 gallons, or roughly 1.5 m3, per minute. Compared to the HPCI at 19 m3 per minute and the LPCI at 150 m3 per minute, the RCIC’s capacity seems small. Its pump would need five hours and a half to fill the entire RPV. However, the rate suffices to replace the boil off anticipated 15 minutes after shutdown.

The RCIC pump’s turbine is fed from the main steam line with the steam produced by the reactor decay heat. The exhaust is discharged into the suppression pool. Alternating and direct current are needed for system initiation. Once started, however, the RCIC system should be capable of running controlled by battery-powered direct current alone. Alas, the pump has a distinct history of operational failures mainly because of governor valve malfunction [Boardman, JR (1994) Operating Experience Feedback Report – Reliability of Safety-Related Steam Turbine-Driven Standby Pumps. NUREG-1275, Vol. 10].

The Terry Corporation, now part of Dresser-Rand, manufactured most turbines of the kind used to drive RCIC pumps in US BWRs. Reaching full capacity within two minutes from start-up, the pumps are aligned by default to remove water from the condensate storage tank, also known as the contaminated condensate storage tank, and inject it into the RPV. If tank water is unavailable or if the suppression pool water level reaches a preset level, valves automatically direct the suction path via a 16-inch pipe to the torus.

Example of a Terry RCIC turbine.

However, the NRC post-inspection letter to Exelon Nuclear below dated Jun. 29, 2004, provides evidence that RCIC suction isolation valves have failed in US reactors in the past because of the inadequate control design.
List of Inspection Reports 2004004[964]

List of Inspection Reports 2004004[964]

Outboard piping downstream from the pump is commonly not as seismically resistant as the inboard piping from the torus which is considered existential to reactor safety. If the RCIC design at the Fukushima Daiichi reactors was the same as that mandated in the US, the inferior seismic rating of the parts of the RCIC loop downstream of the pump may have contributed significantly to the inability to provide adequate coolant to units 2 and 3 after the earthquake.

The quake may have ruptured that piping, providing a continuous escape route for coolant, steam, and non-condensible gases, e.g. hydrogen and radioactive noble gases, from the suppression pool. The suppression pool, however, supplies the coolant for the emergency core cooling system. According to Jake Adelstein and David McNeill’s post with the title “Meltdown: What Really Happened at Fukushima?” published online in the Atlantic Wire on Jul. 2, 2011, eyewitnesses report that the earthquake immediately inflicted widespread, substantial damage to piping at the plant.

In normal operation, the tall stack towers between Units 1 and 2 and Units 3 and 4 at Fukushima Daiichi nuclear power station are designed to vent filtered, decontaminated off-gases from the main steam turbine [Row TH (1973) Radioactive Waste Systems and Radioactive Effluents]. By contrast, separate shorter stacks attached to the reactor building walls provide venting for uncontaminated effluents from the buildings. If increased levels of radioactivity are detected, the effluents will be routed through the Standby Gas Treatment System to the stack. This system also filters gases from the primary containment. Compared to the towers, the stacks for uncontaminated effluents are short, ending just above the roofline. One is still visibly intact on the Northwest corner of unit 2:

Vent stack on NW corner of the reactor building of unit 2.

Roughly 29 seconds into the report below, the footage suggests that steam was already emanating from the reactor building rooflines when the first tsunami waves arrived, probably through these short stacks:

At 15:29, that is approximately the same time the footage above was recorded, a monitor about a mile from the reactor buildings sounded radiation alert according to Yuji Okada, Tsuyoshi Inajima and Shunichi Ozasa’s post with the title “Fukushima May Have Leaked Radiation Before Tsunami” published online on Bloomberg News May 19, 2011. The source of the radioactivity may have been contaminated steam from the suppression pool escaping through the short stacks via damaged valves and broken piping of standby coolant systems like the RCIC. Once the fuel rods in the reactor core super-heated because of the continuing loss of coolant, their zircalloy cladding reacted with water to produce large amounts of hydrogen that accumulated in the reactor buildings, until the gas detonated in violent explosions.

It is important to note that neither the widely-acknowledged extended offsite power outage nor the failure of onsite emergency diesel generators may have caused the standby coolant systems to malfunction. Rather, the constant leakage of coolant through shattered loops in these systems may have prevented effective reactor core cooling and may have been detrimental to further attempts of cooling with the addition of seawater during the days after the earthquake and tsunami. The loss of coolant eventually would lead to extensive uncovering of reactor fuel cores, core melt-downs and the eventual melt-throughs of highly radioactive material.

The operator of the Fukushima I nuclear power station known as Tokyo Electric Power Company, or TEPCO for short, apparently paid little heed to maintenance and improvements of emergency equipment instrumental for a successful cold shutdown after a loss of coolant accident. TEPCO’s probabilistic risk assessment predicted a temblor of the strength of the Tohoku-oki Earthquake to happen only once in ten-thousand years or less (Report of Japanese Government to the IAEA Ministerial Conference on Nuclear Safety, figure III-2-3, page III-39). Loss of coolant accidents may be met by similarly anemic standby cooling systems at any nuclear power station with such reactors in an earthquake zone.

This post would not have been impossible without the contributors to Japan Earthquake Live and the invaluable information on fukuleaks.org/web.

This article would not be possible without the extensive efforts of the SimplyInfo research team
Join the conversation at chat.simplyinfo.org

© 2011-2018 SimplyInfo.org, Fukuleaks.org All Rights Reserved Content cited, quoted etc. from other sources is under the respective rights of that content owner. If you are viewing this page on any website other than www.simplyinfo.org (or www.fukuleaks.org) it may be plagiarized, please let us know. If you wish to reproduce any of our content in full or in more than a phrase or quote, please contact us first to obtain permission.


Editor, SimplyInfo.org

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.