Accident Management for Severe Accidents at Light Water Power

Document Sample
Accident Management for Severe Accidents at Light Water Power Powered By Docstoc
					                                                 NSCRG: L-AM-II.01




                  NSCRG: L-AM-Ⅱ.01
Accident Management for Severe Accidents at Light Water
             Power Reactor Installations




             The Nuclear Safety Commission of Japan
                                                                       NSCRG: L-AM-II.01




Revision History
Authorized on May 28, 1992, 1964, by the Nuclear Safety Commission


Disclaimer
This is an unofficial translation of the Nuclear Safety Commission Document for the benefit of
readers interested. Should there be any ambiguity in phrasing; reference should be made to the
official version in Japanese.
                                                                                        NSCRG: L-AM-II.01

       Accident Management for Severe Accidents at Light Water Power
                          Reactor Installations
                                          Table of Contents                                                  page

1.     Introduction                                                                                             1
2.     Role and significance of accident management                                                             3
3.     Accident management to prevent an event from developing into a severe accident
                                                                                                                5
       (Phase I Accident Management)
4.     Accident management to mitigate the consequences of severe accidents (Phase II
                                                                                                                8
       Accident Management)
5.     Results of technical evaluation                                                                         17
6.     Conclusions and suggestions                                                                             20
Reference                                                                                                      23

1. Introduction

      In some countries, containment measures as part of severe accident(*) management have been
adopted as regulatory requirements or licensees’ voluntary acts. In light of this, the Common
Issue Committee established a working group to study accident management with a focus on
measures to maintain the integrity of the containment, using, as basic information, the latest
results of probabilistic safety analysis (PSA) performed overseas and severe accident research in
Japan and overseas, such as PSA of nuclear reactors in Japan and Severe Accidents Risks
(NUREG-1150) prepared by the US NRC. This report summarizes the results of our study.
      In recent years, it has become widely and internationally recognized that accident management
is important as one of the means of risk management for nuclear reactor facilities. As a result, in
various countries, emergency operating procedures to restore the reactor core cooling capability
and maintain the integrity of the containment have been developed; training of personnel
involved has been provided; and the development of related equipment has been considered or
implemented, assuming the occurrence of beyond-design-basis events.
      Accident management is a set of measures taken to prevent a beyond-design-basis event
potentially leading to serious core damage from developing into a severe accident, or to mitigate

(*)
   In the interim report dated February 1990 from the Common Issue Committee of the Special Committee on the
Safety Standards for Nuclear Reactors, “severe accidents” are defined as “the events significantly exceeding design
basis events and conditions in which adequate cooling of the reactor core or reactivity control cannot be provided by
the means assumed in safety design evaluation, and leading to serious damage to the reactor core.” Design basis
events are “the events that could lead reactor facilities to an abnormal state and that should be considered in the
safety design of reactor facilities and its evaluation.”
                                                         1
                                                                           NSCRG: L-AM-II.01

the consequences of the event even if it develops into a severe accident, by effectively using
dependable features other than safety margins included in current designs and safety design
features, or equipment and systems additionally installed to cope with such events. The first set
of measures based on dependable features is called Phase I Accident Management, and the
second set of measures based on additional equipment and systems is called Phase II Accident
Management.
  Phase I Accident Management consists of various operations to restore safety functions such
as the core cooling capability that have been lost from any cause. In order to ensure that these
operations are properly performed, it is important to develop measuring, display and recording
instruments designed to allow operators to easily assess the state of facilities at all stages of an
event; to develop procedures designed to allow them to perform appropriate operations based on
the displayed information on the state of the plant to restore the plant to a safe state when such a
complex event occurs that is difficult to identify the initiating event,; and to educate and train
those who implement accident management. Such display systems and procedures have been
developed in other countries. In Japan, licensees have studied and developed procedures for part
of this accident management, taking into account the relationship between accident management
and display systems, and the relationship between the procedures and the display system, and
have provided education and training to operators at the Operator Training Center and other
facilities.
  As Phase II Accident Management measures, it is being considered to restore the heat removal
capability of the reactor core or the containment to cool a damaged core; and to install a
dedicated vent line (including filtered vents) in the containment to release the containment
atmosphere containing fission products (FP) in a controlled manner when it has to be partly
released to the environment, in order to prevent the containment from being damaged by
overpressure. In the US and Europe, such operating procedures have been developed.
Particularly in Sweden, France and Germany, installing a filtered vent in the containment and
utilizing it are included in the procedures. For ice condenser containments, in the US and Finland,
a hydrogen igniter system is installed to handle the large amount of hydrogen generated. In Japan,
These Phase II Accident Management measures are being considered by licensees.
  In this report, in view of the aforementioned situation, the role and significance of accident
management in ensuring safety was first of all examined, and the status of development of Phase
I Accident Management measures in Japan and its adequacy were reviewed based on PSA of
reactors in Japan and PSA of reactors in the US provided in NUREG-1150. Then, the benefits
and costs of containment measures for Phase II Accident Management, particularly filtered
containment vents and hydrogen igniter systems were analyzed, taking into account the latest
                                                 2
                                                                            NSCRG: L-AM-II.01

results of severe accident research both in Japan and overseas, as well as the state and data of
systems considered or installed in various countries, and studied the approach to containment
measures that should be taken by Japan.

2. Role and significance of accident management

  The Law for the Regulations of Nuclear Source Material, Nuclear Fuel Material and Reactors
states, as one of the requirements for a license to establish a reactor, that the location, structure
and components of reactor facilities should be such that accidents from nuclear fuel material,
material contaminated by nuclear fuel material or nuclear reactors can be prevented safely. In
order to determine whether this requirement is met, the government checks that reactor facilities
are designed to reliably prevent accidents through proper safety management, and that even if a
design basis accident occurs in reactor facilities, safety systems are available to prevent the
accident from spreading and effectively mitigate the consequences of the accident so that
significant radiation exposure risks from reactor facilities to the public in the surrounding area
(hereinafter, radiation exposure risk to the public in the area surrounding reactor facilities are
referred to as the “risk”) can be avoided. In addition, the government requires licenses to make
safety preservation rules establishing policies for safety management before the start of the
operation of reactor facilities, and approves the rules if determined to be adequate. Furthermore,
the government makes it mandatory to periodically inspect the reactor facilities about every one
year of reactor operation.
  Although these measures sufficiently ensure the safety of reactor facilities, nuclear disaster
response measures were established under the Basic Law on Natural Disasters to deal with
unexpected events. Based on the nuclear disaster response measures, organizational structures
were developed, necessary equipment and material were prepared, and disaster drills were
defined, in order to appropriately and effectively prevent disasters from occurring and damage
from spreading when a massive release of radioactive material occurs, taking into account
knowledge about severe accidents significantly exceeding design basis events.
  Consequently, the risk from reactor facilities is sufficiently small. If appropriate accident
management is implemented when an event potentially leading to a severe accident or a severe
accident occurs in reactor facilities, the potential of leading to a severe accident will be further
reduced, or the consequences of a severe accident on public health will be mitigated, further
reducing the risk. Accident management is an action that is taken by licensees in a flexible
manner based on their technical knowledge, including case-by-case responses to actual situations.
It should therefore be recommended or expected to implement accident management as long as

                                                 3
                                                                              NSCRG: L-AM-II.01

the implementation is possible without significantly changing the components of reactor facilities
and that it reduces the risks effectively. It is appropriate to set a risk reduction target by balancing
the effectiveness of Phase I and II Accident Management, based on, for example, the quantitative
safety target (a core damage frequency of 10-4 per reactor year for existing reactors and 10-5 per
reactor year for new reactors and a probability of 1/10 of these values for the occurrence of a
massive release of fission products) provided in the Basic Safety Principles of IAEA/INSAG
(International Nuclear Safety Group).
  In order to properly implement accident management, it is important that licensees establish
procedures for actions considered to be effective and appropriate in the implementation, prepare
necessary equipment and material, establish an implementing organizational structure, and train
personnel, taking into account the latest results of severe accident research.
  When all of these matters are considered together, the following approaches to accident
management can be taken by the government.
  One approach is that the government guide licensees to develop accident management
measures from the standpoint that, as mentioned above, effective implementation of accident
management is recommended or expected, and, for example, review of measures in approving
safety preservation rules should be performed.
  Another approach is that in approving construction plans, the government should check that
accident management developed by licensees does not prevent engineering safety facilities from
being used in an appropriate manner, in order to prevent the level of protection against design
basis events from falling due to the accident management being inappropriate. More specifically,
if a filtered vent and a hydrogen igniter system are installed in the containment, it should be
examined as part of safety regulations whether they can have an adverse effect on the functioning
of engineering safety facilities against design basis events. There is also the following approach.
  Accident management is an action that is taken by licensees in a flexible manner based on
their technical knowledge, including case-by-case responses to actual situations. Therefore, if
specific means are excluded as a result of the government's prior evaluation, the effective
implementation of accident management may be prevented. However, adequate coordination
among parties involved may be necessary to effectively implement emergency response
measures. Analysis of PSA results shows that the capability of accident management to ensure
safety during severe accidents may be as effective as equipment and systems designed to deal
with abnormal conditions even if equipment or functions are used that are not expected in safety
assessment based on the review guidelines. This is an approach that requires the government to
develop a basic philosophy that licensees can use as the basis for developing accident
management measures.
                                                   4
                                                                        NSCRG: L-AM-II.01

3. Accident management to prevent an event from developing into a severe accident (Phase
I Accident Management)

3.1 Current status of development of accident management measures
  Phase I Accident Management consists mainly of various operations to restore safety functions
such as cooling the reactor core. It is being considered to develop procedures designed to allow
these accident management measures to be properly performed.
  In reactors in Japan, various Phase I Accident Management measures have been developed or
considered. Key measures are described below for each accident sequence.
1) BWR plants
    i. Station blackout
       Restoration of external power supplies or diesel generators
    ii. Anticipated transient without scram (ATWS) events
       a) Manual scram or manual rod insertion when the reactor protection system does not
       operate
       b) Manual start of the boron injection system (SLC)
    iii. Events of protected loss of heat sink during a transient
       a) Restoration of the decay heat removal system (RHR)
       b) Manual start of the containment spray system
       c) Containment vent
    iv. Injection failure after a transient
       a) Manual start of the high-pressure ECCS and isolation cooling system (RCIC)
       b) Manual start of the automatic depressurization system (ADS) and low-pressure ECCS
       c) Manual start of alternative injection systems
    In level 1 PSA in Japan, which will be described later, among the above measures, the
  manual start of the SLC, ADS and ECCS, and the restoration of the containment vent system
  and equipment were considered. For the containment vent system, it was assumed in the
  assessment that the existing system (a duct vent using the inert gas system (AC system) or the
  emergency gas treatment system (SGTS)) is used. The purpose of the vent system in this case
  is to prevent damage to the containment during a protected loss of heat sink following a
  transient.


2) PWR plants
    i. Transients
       Restoration of the main feedwater system

                                                 5
                                                                         NSCRG: L-AM-II.01

    ii. Transients and small LOCAs
       Feed and bleed
    iii. Medium and small LOCAs
       Cooling by manually opening the main steam relief valves (MSRV)
    iv. Station blackout events
       Restoration of external power supplies or diesel generators
    v. ATWS events
       Manual reactor trip and emergency boron injection
    vi. Safety injection (SI) signal malfunction
       Manual start of the ECCS at SI signal malfunction
    vii. Containment spray signal malfunction
       Manual start of the spray system at containment spray signal malfunction
    viii. Events requiring change in recirculation
       Water supply from an alternative water source to the refueling water storage tank
    (RWST)
    In level 1 PSA, which will be described later, among the above measures, feed and bleed,
  the manual opening of the MSRV, the manual start of the emergency boron injection system,
  and the restoration of equipment were considered.


3.2 Evaluation of Phase I Accident Management
  This section discusses the effectiveness of Phase I Accident Management in reactors in Japan,
based on the results of level 1 PSAs for domestic reactors performed by the industry and the
Institute of Nuclear Safety Analysis and taking account of the results of level 1 PSAs performed
overseas.
1) BWR plants
  Among PSAs for overseas BWR plants, those that can be used as a reference are the Reactor
Safety Study (WASH-1400) on the Peach Bottom reactor (BWR-4 Mark I plant) in the US,
which can be compared with domestic reactors, and NUREG-1150. In WASH-1400, major core
damage accident sequences due to internal events are ATWS events (hereinafter, referred to as
the “TC sequence”) and events of protected loss of heat sink after a transient (hereinafter,
referred to as the “TW sequence”). In NUREG-1150, accident management measures such as
cooling the reactor core after the containment vent system or vessel is damaged were considered
based on the latest research results. Consequently, the portion of the TW sequence in the
frequency of total core damage (hereinafter, referred to as the “total core damage frequency”) is
very small. The total core damage frequency is about 1/5 of that in WASH-1400.
                                                   6
                                                                            NSCRG: L-AM-II.01

  Because of the smaller capacity of Mark I containments than that of other types of
containments, the NRC decided to recommend improvements to the performance of the
containment, which include the installation of an enhanced pressure-resistant vent system. The
purpose of installing a pressure-resistant vent system is to prevent operators from hesitating to
operate the containment vent system due to their concern that the vent line might be damaged by
overpressure. The NRC also recommends that an enhanced pressure-resistant vent system be
installed in Mark I containments with an isolation condenser to directly release residual heat to
the air, as in other Mark I containments, although the TW sequence is not dominant in this case.
  For typical BWR-4 plants in Japan, some Phase I Accident Management measures mentioned
in the previous section, such as that operators can perform manual operations using the existing
vent system (duct vent), were considered in the assessment. As a result, the portion of the TW
sequence in the total core damage frequency is less than 1%, and the total core damage frequency
is less than a few portions of that in NUREG-1150.
  For BWR-3 and BWR-5, which are other typical types of BWR plants, the total core damage
frequency is of the same order as that for BWR-4. There is not much difference in containment
capacity per output power between BWR Mark I plants in Japan and the Peach Bottom Mark I
plant.
2) PWR plants
  Among PSAs for overseas PWR plants, those that can be used as a reference are the
NUREG-1150 assessment for the Zion reactor (a 4-loop plant with a large dry containment) and
the Sequoyah reactor (a 4-loop plant with an ice condenser containment) in the US. For these
reactors, the total core damage frequency is about 10-5 to 10-4/reactor year.
  For typical PWR 4-loop plants in Japan, some Phase I Accident Management measures in the
previous section were considered in the assessment. As a result, the total core damage frequency
is less than 1/10 of that in NUREG-1150. Risk Study Phase B performed by the Society for Plant
Safety and Reactor Safety (GRS) of Germany on the Biblis B plant (4-loop plant with a large dry
containment) shows about one digit difference in total core damage frequency between when
accident management was considered and when it was not considered. In the assessment
performed in Japan, the frequency is even smaller.


3.3 Summary of Phase I accident management
  In Japan, some of the Phase I Accident Management measures have been developed by
licensees as procedures called “sign basis” or “safety function basis” to supplement the
conventional event-based procedures. Furthermore, education and training have been provided to
operators at the Operator Training Center so that accident management is properly implemented
                                                 7
                                                                          NSCRG: L-AM-II.01

according to the procedures.
  A level 1 PSA, taking into account some of these Phase I Accident Management measures,
shows that the core damage frequency for typical reactors in Japan is less than 10-5/reactor year,
including uncertainties in the assessment, which meets, for example, the quantitative safety target
(a core damage frequency of 10-4 per reactor year for existing reactors and 10-5 per reactor year
for new reactors) provided in the Basic Safety Principles of IAEA/INSAG (International Nuclear
Safety Group). The core damage frequency for the plants in Japan on which PSA was performed
by the same approach is smaller than that for plants of the same type with similar systems among
the plants in the US on which PSA was performed. However, the assessment results for typical
reactors in Japan were obtained, assuming that operation management, which has been
satisfactory, would continue to be reliable and that accident management measures would be
properly implemented.
  In order to more reliably perform the restoration of safety functions as an accident
management operation, it is effective to enhance the information supply and instrumentation
systems to identify the cause of anomalies and perform response operations, as well as to
simplify and automate the operations. Furthermore, it is possible to more reliably prevent an
event from developing into a severe accident by expanding the scope of Phase I Accident
Management measures to be considered, such as enhancing the capability to substitute for safety
functions by modifying current equipment or introducing portable equipment.

4. Accident management to mitigate the consequences of severe accidents (Phase II
  Accident Management)

4.1 Current status of development of accident management measures
  The basic philosophy of Phase II Accident Management to mitigate the consequences of
severe accidents on the surrounding environment is to maintain the integrity of the containment
against various possible containment failure modes during severe accidents, using the means as
described in the Introduction section, and to prevent the abnormal release of radioactive material
to the environment by scrubbing with suppression pool water or filtering as needed in the process
of maintaining the integrity.
  In Japan, accident management measures are already in place, such as using an inert gas as the
containment gas, and installing a flammability control system (FCS) in BWRs and a containment
spray system in PWRs. Functions other than those assumed in safety assessment can be expected
for these measures. They can be considered as part of Phase II Accident Management in terms of
their results. Other Phase II Accident Management measures are under consideration by

                                                8
                                                                          NSCRG: L-AM-II.01

licensees.
  The following measures are being considered overseas to maintain the integrity of the
containment by installing additional equipment and systems, in addition to the effective use of
existing equipment and systems.
1) BWR plants
  Containment measures, as Phase II Accident Management, already in place or under
consideration in overseas BWR plants include the following:
  i. A filtered or wet well vent system
  ii. A containment water injection system
  iii. Functional enhancement of the ADS
  iv. A hydrogen control system
  As described later, the above containment measures are very effective in reducing the risk
when implemented in combination with other measures. That is, the reliability of the
containment during a severe accident can be significantly improved by implementing measures
specifically for each containment failure mode in parallel.
  Possible containment failure modes include overpressure failure, over-temperature failure,
direct containment heating (DCH) and hydrogen combustion. For example, the filtered vent
system alone can avoid overpressure failure but is little effective in avoiding over-temperature
failure. In order to prevent over-temperature failure, it is necessary to cool the molten reactor
core by injecting water into the containment, and thereby to control the core-concrete reaction as
well as lower the temperature of the containment atmosphere. Conversely, the containment water
injection system alone cannot avoid overpressure failure. If the reactor pressure vessel fails with
the internal pressure remaining high and the molten core is dispersed from the vessel, the
containment may fail due to direct containment heating by the molten core. Therefore, it is being
considered to avoid core melting at high pressure by enhancing the function of the ADS. Also,
for plants with a Mark II containment, hydrogen control systems, such as hydrogen igniter
systems to prevent containment failure due to hydrogen combustion, are in place or under
consideration.
2) PWR plants
  Containment measures, as Phase II Accident Management, already in place or under
consideration in overseas PWR plants include the following:
  i. A filtered vent system
  ii. A hydrogen igniter system
  iii. Enhancement of the containment spray system
  iv. A containment external spray system
                                                 9
                                                                          NSCRG: L-AM-II.01

  v. A containment water injection system
  The above containment measures are very effective in reducing the risk when implemented in
combination with other measures. For example, the filtered vent system alone can avoid
overpressure failure but is little effective in avoiding a containment failure due to the
core-concrete reaction. In order to prevent containment failure due to the core-concrete reaction,
it is necessary to cool the molten reactor core by enhancing the containment internal spray
system and/or injecting water into the containment, and thereby to control the core-concrete
reaction. These containment measures are in place or under consideration in plants with a large
dry containment.
  In plants with a large ice condenser containment, the containment is very likely to fail early
due to the combustion of a large amount of hydrogen generated. A hydrogen igniter system to
prevent this is in place or under consideration. Furthermore, in order to prevent the overpressure
failure of the containment, a containment external spray system to cool the containment
atmosphere by spraying water on the outside of the steel containment is being considered.


4.2 Current status of containment measures in the US and Europe
  In the US and Europe, various containment measures are being considered as a means for
accident management for severe accidents.
  A survey was conducted on containment vent and hydrogen igniter systems among these
containment measures, and on the significance of these systems in regulations in the US and
Europe.
1) Containment vent systems
  In the US and Europe, filtered vent systems as part of Phase II Accident Management are in
place or under consideration. In France, it was decided, based on the PSA performed by French
Electricity (EDF) in 1978, to reduce fission product release to the environment to a level
consistent with the emergency response plans for the surrounding area of nuclear sites even in
the event of core meltdown, as the safety goal for beyond-design-basis events. Filtered
containment vent systems are used as a means to meet the safety goal. It is required to achieve a
decontamination factor (DF) of 10 or more for fission product aerosols in specific designs.
  In Germany, the Reactor Safety Commission (RSK) recommended basic requirements for
filtered containment vent systems between 1986 and 1987. In the recommendation, filtered
containment vent systems are considered as measures to complete station emergency response
plans and are not a technical requirement to ensure safe plant operation. It is required to achieve
a decontamination factor of 1,000 or more for aerosols and 10 or more for elemental iodine in
specific designs. Currently, almost all plants (PWR and BWR plants) likely have a containment
                                                10
                                                                             NSCRG: L-AM-II.01

vent system.
  In Sweden, basic policies for severe accidents were issued by the government between 1980
and 1981. In the policies, filtered containment vent systems were required to be installed to
further reduce the release of radioactive material to the environment, even though a massive
release leading to soil contamination is unlikely, and they were installed in all plants. It is
required to achieve fission product release below 0.1% of the inventory of a 1,800MWt reactor
core in specific designs in order to prevent massive soil contamination and acute deaths.
  In Finland, filtered containment vent systems have been installed in BWR plants.
  In the UK, the use of containment vent systems is being considered, including which system to
install in PWR plants. A final decision is expected in by the end of 1991.
  In the US, as previously mentioned, the enhancement of the pressure resistance of vent
systems is required as Phase I Accident Management measures for BWR Mark I containments. A
combination of water injection and enhanced pressure-resistant vent systems as Phase II
Accident Management measures is being discussed, and filtered vent systems are not included in
items that the NRS requires to be considered. For PWR plants, containment vent systems are not
regarded as an item for consideration.
  Single failures, loss of power supply and earthquakes are not considered in the design of
containment vent systems in France and Germany, but are considered in Sweden.
2) Hydrogen igniter system
  There are no unified measures among different countries to control hydrogen during a severe
accident. There are many countries that plan to take action based on the results of research and
development that will be available in the future.
  In the US, the NRC issued recommendations and action plans based on the results of
investigation of the TMI-2 accident between 1979 and 1980, and required hydrogen control
measures for BWRs and PWRs with an ice condenser containment. Subsequently, in December
1981, rules were issued requiring that the atmosphere of BWR Mark I and II containments be
made inert by nitrogen. In January 1985, rules were issued requiring improvements to the
hydrogen control system in BWR plants with a Mark III containment and PWR plants with an
ice condenser containment. A glow plug hydrogen igniter system was installed in these plants.
  In France, the Atomic Energy Commission (CEA) has been studying hydrogen control systems
for PWRs with a dry containment, but has not decided on hydrogen control. French Electricity
(EDF) had decided to consider the installation of a hydrogen control system if its necessity arises
as a result of related experiments and research in the US, France and other countries.
  In Germany, no decision has been made on hydrogen control. However, all licensees decided
that an ignition system was suitable for hydrogen control in PWR plants with a dry containment,
                                                11
                                                                            NSCRG: L-AM-II.01

and agreed to install a hydrogen igniter system. Further detailed studies are underway.
  In Finland, a glow plug hydrogen igniter system has been installed for hydrogen control in
PWR plants with an ice condenser containment.
  In the UK, no hydrogen control system has been installed because a PSA of PWR plants
shows that loss of integrity of the containment due to hydrogen combustion is unlikely.
3) Cost of containment measures
  The cost of installing equipment and systems as containment measures varies with the
specifications, and therefore there is no typical cost. However, the estimated cost of a filtered
containment vent system ranges from about 1 to 4.5 million dollars. One cost estimate of
installing a hydrogen igniter system is about 4.9 million dollars.


4.3 Effectiveness of containment measures
  Level 2 PSA results can be used as a reference for evaluating the effectiveness of Phase 2
Accident Management measures such as containment vent systems.
  The results of preliminary Level 2 PSAs performed by the industry in Japan and the Institute
of Nuclear Safety Analysis were evaluated. In the evaluation, the containment failure frequency,
the results of a source term evaluation, the reduction of source terms by using a containment vent
system, and the reduction of the containment failure frequency by installing a hydrogen igniter
system were investigated. For BWRs, a comparison was made with the results of PSAs
performed in the US.
  The level 2 PSA evaluation results investigated were based on the knowledge gained to date
and should be revised as more knowledge is available on the phenomena dealt with in level 2
PSA.
1) Containment vent system
  [BWR]
       An evaluation of the Peach Bottom reactor in the US shows that if only a containment
     vent system (a pressure-resistant type) is available, drywell failure occurs due to
     over-temperature in all sequences. Fission product release to the environment is mainly
     from the drywell, and therefore a difference in the type of vent systems (enhance
     pressure-resistant vents, filtered vents, duct vents) did not result in a significant difference
     in the risk. It has been shown that measures such as improved ADSs, backup spray systems
     and backup pressurized containment water injection systems are not effective in reducing
     the risk when used individually, but when all of these measures are used in combination
     with a containment vent system, a significant risk reduction can be expected.
       The results of preliminary Level 2 PSAs performed on typical BWR plants in Japan are
                                                 12
                                                                     NSCRG: L-AM-II.01

summarized below.
i) For BWR-4 plants with a Mark I containment, the frequency of occurrence of total
  containment failure (hereinafter, referred to as the “total containment failure frequency”)
  was evaluated to be about 1/7 of the total core damage frequency (an analysis by the
  industry). Containment failure modes are station blackout and core cooling failure after a
  LOCA; over-temperature failure in a sequence where the containment fails after core
  damage; and overpressure failure in accident sequences such as the TC and TW
  sequences. (In level 2 PSA, over-temperature failure is defined to occur when the
  containment temperature exceeds a specified value while overpressure failure is defined
  to occur when the containment internal pressure exceeds a specified value.) The
  frequency of containment failure accidents leading to fission product release at a
  relatively early stage (within 10 hours) is about 1/10 of the total core damage frequency.
  A source term evaluation shows that in all accident sequences, the release rate of CsI and
  CsOH is higher by over one digit than that of other nuclides.
   In BWR-5 plants with a Mark II containment, the water injection function is lost due to
  the overpressure failure of the containment in the TC and TW sequences, resulting in core
  meltdown. In the Mark II plant containment, the pedestal at the bottom of the reactor
  pressure vessel fails due to concrete corrosion by the molten core. The total containment
  failure frequency was evaluated to be about 1/3 of the total core damage frequency by the
  industry and about 1/4 by the Institute of Nuclear Safety Analysis. The frequency of
  containment failure accidents leading to fission product release at a relatively early stage
  is about 1/5 of the total core damage frequency.
   A source term evaluation shows that the temperature of the molten core dispersed into
  the containment is higher than in Mark I plants, the core-concrete reaction becomes
  intense, and the contribution of Te, Sr and Mo increases. Also, the portion of CsI and
  CsOH released is generally larger than in Mark I plants because the scrubbing effect
  cannot be expected if the pedestal fails.
ii) The reduction of fission product release by a containment vent system was evaluated
  focusing on CsI, which has a significant environmental impact, and the results are
  presented below.
  In BWR-4 plants with a Mark I containment, there is not much difference in the reduction
  of CsI release to the environment between duct and enhanced pressure-resistant vents.
  Even in an accident sequence where these vents are effective in reducing CsI release, the
  release reduction factor is only about 10. This is because a sequence becomes dominant
  where over-temperature failure occurs in the drywell and CsI is released bypassing the
                                              13
                                                                       NSCRG: L-AM-II.01

    vent pipe. If the molten core can be reliably cooled by the water injection system, the
    duct or enhanced pressure-resistant vent is probably as effective in reducing fission
    product release as scrubbing, whereas the fission product release reduction factor of a
    filtered containment vent system is probably equivalent to the decontamination factor of a
    filter.
     In BWR-5 plants with a Mark II containment, neither the duct or enhanced
    pressure-resistant contribute to reducing CsI release to the environment. Instead, the
    enhanced pressure-resistant vent contributes to increasing the portion of release compared
    to the duct vent. This is because while the settling of fission products in the containment
    building can be expected for the duct vent, it cannot be expected for the enhanced
    pressure-resistant vent, which bypasses the containment building. If, as in BWR-4 plants
    with a Mark I containment, water is injected into the containment and the molten core can
    be reliably cooled, the fission product release factor of the filtered vent system is
    probably equivalent to the decontamination factor of a filter.
[PWR]
  The results of preliminary Level 2 PSAs performed on typical PWR plants in Japan are
summarized below.
  i) For plants with a large dry containment, the total containment failure frequency is about
    1/7 of the total core damage frequency (analysis both by the industry and the Institute of
    Nuclear Safety Analysis). The total containment failure frequency for plants with an ice
    condenser containment evaluated by the industry is about 1/5 of the total core damage
    frequency (without a hydrogen igniter system) and about 1/10 (with a hydrogen igniter
    system).
     The frequency of overpressure failure of the containment due to a quasi-static pressure
    increase (a slow pressure increase) is about 62% of the total containment failure
    frequency (analysis by the industry) and about 51% (analysis by the Institute of Nuclear
    Safety Analysis) for plants with a large dry containment, while it is about 30% of the total
    containment failure frequency (without a hydrogen igniter system) and about 60% (with a
    hydrogen igniter system) for plants with an ice condenser containment.
  ii) The above results suggest that in plants with a large dry condenser containment and those
    with an ice condenser containment and a hydrogen igniter system, the total containment
    failure frequency can be reduced to about 1/3 using a containment vent system with a
    water injection system installed to the containment.
     In terms of reduction in CsI release to the environment by filtered vent systems, if the
    molten core can be reliably cooled by the water injection system in an accident sequence
                                             14
                                                                           NSCRG: L-AM-II.01

       where an overpressure containment failure occurs due to a quasi-static pressure increase,
       the fission product release reduction factor of the filtered containment vent system is
       probably equivalent to the decontamination factor of a filter.
2) Hydrogen igniter system
  Due to the smaller free containment volume and lower failure pressure, the total containment
failure frequency for PWR plants with an ice condenser containment is about five times that for
those with a large dry containment. However, as shown below, the containment failure frequency
can be reduced by installing a hydrogen igniter system.
  A preliminary level 2 PSA performed by the industry shows that in PWR plants with an ice
condenser containment, the total containment failure frequency can be reduced from about 1/5 to
1/10 of the total core damage frequency by installing a hydrogen igniter system. This is because
the mode where the hydrogen generated in the reactor core before the failure of the reactor
pressure vessel is quickly combusted in the containment (this mode accounts for about 48% in
plants with an ice condenser and without a hydrogen igniter system) becomes negligible.


4.4 Issues resulting from containment measures and options
  The previous sections discussed the effectiveness of containment vent systems and hydrogen
igniter systems as part of accident management measures from the viewpoint of probabilistic
safety analysis.
  However, the use of these systems may compromise the safety of reactor facilities as a whole
due to equipment failure or malfunction or an operator’s erroneous equipment operation.
  This section identifies as many technical issues as possible that may result from the
implementation of these measures, and describes the results of a survey and evaluation of how
these issues are addressed in other countries.
1) Containment vent system
  The purpose of containment vent systems is to prevent the overpressure failure of the
containment and to maintain its integrity. Therefore, it is important in the design and operation of
a vent system to determine the set pressure to start venting. The failure threshold pressure of the
containment is said to be about three times the design pressure based on various experiments. In
many conventional vent designs, a set pressure equal to or slightly higher than the design
pressure of the containment is used.
  The possibility of safety being compromised by equipment failure or malfunction or an
operator’s erroneous equipment operation is addressed in various countries by installing two
isolation valves for normal plant operating conditions, using a rupture disc or combining these
measures. For manually operated valves with a rupture disc, it is necessary to install monitoring
                                                 15
                                                                            NSCRG: L-AM-II.01

instruments, develop procedures, and enhance operator training in order to prevent erroneous
operation.
  During the venting of the containment, (1) hydrogen combustion in the vent line, (2) cavitation
of the water injection system in BWR plants, (3) isolation valve close failure, or (4) containment
failure due to negative pressure may occur. In various countries, issues (1), (2), (3) an (4) are
addressed by installing a pressure release device (orifice) or using nitrogen gas as the
containment gas as a design option; changing the source of water supply to the injection system
or using an external water supply as an operational option; using two isolation valves as a design
option; and using a containment pressure monitoring system as a design option, respectively. In
developing procedures, it is necessary to specifically describe how to use the procedures by
sufficiently relating monitoring items and operational steps to these design options.
2) Hydrogen igniter system
  The function of a hydrogen igniter system is to reduce hydrogen concentration in the
containment during an accident in which a large amount of hydrogen is generated. There are
three types of hydrogen igniter systems: a glow-plug igniter system, a catalytic igniter system
and a spark igniter system. Hydrogen igniter systems installed in US PWR plants with an ice
condenser are of the glow plug type.
  In introducing this system, it is necessary to investigate (1) its operation timing, (2) whether it
functions in the atmosphere where a large amount of hydrogen is being generated, (3) the
possibility of localized detonation, and (4) the reliability of its power supply system, in order to
ensure its effectiveness. With regard to (1), if a glow-plug igniter system is used, the operator
manually operates it after detecting an accident (other types are automatically activated), and has
sufficient time to do so. However, appropriate procedures need to be developed to prevent
overpressure due to improper timing of activation. With regard to (2), the effect of spraying and
fogging on glow-plug igniter systems when a large amount of hydrogen is generated was
experimentally evaluated at several laboratories in the US. The effect on other types of igniter
systems needs investigation. With regard to (3), in US PWR plants with a hydrogen igniter
system, measures to avoid locally high concentrations of hydrogen are in place, such as installing
a hydrogen igniter system in each compartment of the containment. In connection with measures
to prevent local detonations, including the above measures, and their adequacy, research is being
performed on hydrogen combustion behavior both in Japan and overseas. It is important to make
a decision based on knowledge gained from these research activities. With regard to (4), it is
necessary to examine the reliability of the power supply to glow-plug and spark hydrogen igniter
systems.


                                                16
                                                                            NSCRG: L-AM-II.01

5. Results of technical evaluation

  The costs and benefits of containment vent and hydrogen igniter systems as Phase I and II
Accident Management measures were evaluated. Technical evaluation results are summarized
below.
1) Containment vent system
  i) Containment vent system as an option for Phase I Accident Management
     In the US, as previously mentioned, it is recommended to install an enhanced
  pressure-resistant containment vent system in BWR plants with a Mark I containment in order
  to prevent core damage in the TW sequence. A level 1 PSA of typical plants in Japan, taking
  into account the use of the existing vent system, (duct vent) shows that the core damage
  frequency is sufficiently low and the portion of the TW sequence in the total core damage
  frequency is small.
     In the US, an enhanced pressure-resistant containment vent system is required to be
  installed mainly in Mark I plants because of the claim that the volume of a Mark I containment
  is smaller than that of other types of containments. However, there is no difference in the
  design of containment volume per output power between Mark I and Mark II BWR plants in
  Japan. Also, there is no marked difference in total core damage frequency. It is not necessarily
  clear that an enhanced pressure-resistant containment vent should be installed only in Mark I
  containments.
     Phase I Accident Management for PWR plants in Japan does not rely on containment vent
  systems. Therefore, containment vent systems for PWRs are not discussed here.
  ii) Containment vent system as an option for Phase II Accident Management
  A typical containment vent system as Phase II Accident Management measures is a filtered
vent system, which has been installed in BWR and PWR plants in Europe. For BWRs, a
wet-well vent system is also available that is expected to be effective in reducing fission products
by scrubbing in the suppression pool. It was once considered that the aforementioned enhanced
pressure-resistant containment vent system used in the US should be used for Phase II Accident
Management.
  In current BWR Mark I and II containment designs, a filtered or wet-well vent system alone
cannot prevent the over-temperature failure of the containment, and therefore are not necessarily
effective in reducing fission product release to the environment. The use of a filtered or wet-well
vent system in combination with water injection into the containment can avoid the bypassing of
the filter or the suppression pool, and is effective in reducing fission products to the environment.
  In PWRs, a filtered vent system is effective for the quasi-static overpressure failure mode of

                                                 17
                                                                          NSCRG: L-AM-II.01

the containment. A filtered vent system alone is not effective for other failure modes. Therefore,
it is necessary to comprehensively consider the use of other measures in combination with the
containment vent system, such as enhancing the containment spray system and injecting water
into the containment, in order to reduce the containment failure frequency during a severe
accident.
  In order to determine the specifications of the containment vent system, it is necessary to
promote research on physical phenomena during a severe accident, such as DCH, the coolability
of the molten core and poor scrubbing, and to reduce the range of uncertainty.
2) Hydrogen igniter system
  Measures to prevent a large amount of hydrogen generation exceeding the design basis have
been investigated by a working group of the Special Committee on the Safety Standards for
Nuclear Reactors.
  The working group investigated problems associated with a large amount of hydrogen
generation, measures to take, the strength of the containment, regulations and research and
development activities in the US, and issues involving the application of research and
development results in the US to Japanese regulations. Subsequently, the group investigated the
safety role of the containment and how to introduce the role into the guidelines. They concluded
that even if a large amount of hydrogen is generated, current containment designs can very likely
maintain the containment function without modifications. Also, the working group investigated
hydrogen igniter systems as one of the hydrogen control measures for PWR plants with an ice
condenser containment in Japan. They concluded that if hydrogen combustion can be controlled
in a planned manner, a containment pressure increase can be limited to a minimum, but the
effectiveness of the igniter systems to burn hydrogen reliably without delay has not been
sufficiently demonstrated. Based on the evaluation results of the working group, the effectiveness
of the hydrogen igniter systems and issues and measures involving their installation were
investigated, taking into account the latest results of severe accident research both in Japan and
overseas, including PSA results, as well as the state and data of systems considered or installed
overseas. The knowledge gained to date is summarized below.
  i.    There are three ways of hydrogen control: hydrogen recombination, hydrogen ignition
        and the use of an inert gas as the containment gas. Of the three, hydrogen recombination
        is not suitable for a large amount of hydrogen generation exceeding design basis events.
  ii.   For BWR plants with a Mark I or II containment, the volume of which is smaller that of
        PWR containments, inert nitrogen gas has been used as the containment gas in Japan
        since their first design.
  iii. The use of nitrogen gas is not suitable for PWR plants with a relatively large containment
                                                18
                                                                            NSCRG: L-AM-II.01

        volume because the containment is patrolled and inspected during operation and the
        containment volume is large.
  iv.   There are three types of hydrogen igniter systems: a glow-plug igniter system, a catalytic
        igniter system, and a spark igniter system. It is necessary to examine the reliability of the
        power supply to glow-plug and spark hydrogen igniter systems. A test has not been
        completed to check the performance of catalytic and spark igniter systems under severe
        accident conditions.
  v.    A glow-plug hydrogen igniter system has been installed in PWR plants with an ice
        condenser containment in the US and Finland.
  vi. A level 2 PSA performed by the industry in Japan shows that the installation of a
        hydrogen igniter system in PWRs with an ice condenser reduces the total containment
        failure frequency to about a half.
  vii. Due to the large containment volume per output power and the high containment failure
        threshold, PWR plants with a dry containment are considered to have a large safety
        margin. A hydrogen igniter system is planned to be installed in Germany, but has not been
        installed in other countries.
  Appropriate operational procedures need to be developed to prevent overpressure when the
hydrogen igniter system is activated with wrong timing, such as after the hydrogen concentration
has become high. It is necessary to investigate the possibility of hydrogen detonation in the event
of delay in activation, even though it is low.

6. Conclusions and suggestions

  Level 1 PSAs performed in Japan suggest that in typical reactors in Japan, the possibility of a
severe accident due to factors inside a nuclear facility is sufficiently low if they continue to
perform satisfactorily, and the Phase I Accident Management measures that have been developed
under the government’s guidance can be expected to be reliably implemented. Phase I Accident
Management will be more effective in preventing severe accidents if the scope of investigation is
expanded.
  Despite uncertainties, Level 2 PSAs performed in the US and other countries and preliminary
Level 2 PSAs performed in Japan suggest that the installation of a filtered containment vent
system (BWRs and PWRs) in combination with other measures, such as water injection into the
containment, and a hydrogen igniter system in PWR ice condenser containments can be effective
as part of Phase II Accident Management.
  Based on the discussion in Section 2 “Role and significance of accident management,” it is

                                                 19
                                                                             NSCRG: L-AM-II.01

strongly recommended that accident management measures be developed and properly
implemented in the event of an accident, or it is expected that they should be done so. The
following suggestions can be made with a view to further developing accident management
measures.
(1) The Nuclear Safety Commission should make clear the nature and significance of accident
     management developed by licensees, a basic idea of the roles of licensees and the regulatory
     authority, and should set a direction and framework for the efforts by the parties involved.
(2) In this connection, it is appropriate to define accident management as follows:
     Accident management is a set of knowledge-based measures to further reduce the risk,
     which has been sufficiently reduced by previous measures, and should be implemented by
     licensees using their knowledge, preferably on a case-by-case basis and in a flexible manner
     depending on the situation. Therefore, at this moment, no regulatory action to restrict the
     installation or operation of reactors is required, regardless of whether accident management
     measures are developed or what the measures are.
(3) In this connection, it is appropriate to clarify the following:
     Licensees should strive to develop accident management measures that cover both Phases I
     and II in an effort to further reduce reactor risks. However, if reactor facilities are designed
     to eliminate the possibility of a certain accident or reduce it to an extremely low level,
     accident management measures for that particular accident may be excluded. The balance of
     the effectiveness of Phase I and II accident management or quantitative safety goals that
     some countries began to use may be used as guidelines for this judgment.
(4) In this connection, it is appropriate to include at least the following as specific issues to be
     considered in developing accident management measures.
     a) Consider the following when determining the specifics of accident management.
       • Accident management measures to be implemented

       • Development of equipment and systems for accident management (including

            measuring, display and recording instruments designed to allow operators to easily
            perform an abnormality diagnosis and assess the state of facilities)
       • Development of procedures for accident management and personnel education and

            training
     b) In developing the above accident management measures, if new equipment and systems
         are added or if when existing equipment and systems are used, provisions for operations
         not specified in current procedures are established, it should be ensured that they will
         not impair existing safety functions.
     c) Accident management measures should be implemented according to an appropriate
                                                  20
                                                                           NSCRG: L-AM-II.01

        plan, starting with those that are ready for implementation. Licensees should review the
        progress of implementation of the measures at certain intervals.
    d) Licensees should perform a PSA of individual plants based on the state and operational
        experience of the equipment and systems for accident management to re-acknowledge
        the importance of operation management, including accident management, and should
        make an effort to further reduce risks.
(5) In order to proceed, it is necessary to discuss and gain consensus on the government’s role
    in the development of accident management as soon as possible, taking into account Section
    2 “Role and significance of accident management” in this report.
(6) In reducing the risk, national research institutes and licensees should make an effort to
    reduce the range of uncertainty by investigating human factors with particularly large
    uncertainties and physical phenomena associated with severe accidents.
    It is believed that the above suggestions are urgent and important, considering the
    circumstances in Japan and overseas concerning severe accident management, and hope that
    the Nuclear Safety Commission will discuss the suggestions as soon as possible.




                                                  21
                                                                          NSCRG: L-AM-II.01

(Reference)


                                  NSC Decision of May 28, 1992 (Amended on October 20, 1997)

 Accident Management of Severe Accidents at Power Generating Light
                     Water Reactor Facilities

  The Commission established the Common Issue Committee in July 1987 to investigate
approaches to severe accidents, probabilistic safety analysis techniques, and the containment
function during severe accidents. Subsequently, on February 19, 1990, we received from the
Committee an “Interim Report by the Common Issue Committee of the Special Committee on
the Safety Standards for Nuclear Reactors” concerning knowledge about severe accidents and
some of the results of probabilistic safety analysis.
  Also, on March 5, 1992, we received from the Committee a “Study Report on Accident
Management to Address Severe Accidents – Focusing on Containment Measures” (hereinafter,
referred to as the “report”). We studied approaches that should be taken by Japan, taking into
account that measures to prevent an event from developing into a severe accident and to mitigate
the consequences of a severe accident (hereinafter, referred to as “accident management”) are
recognized as being important in further improving the safety power generating light water
reactor facilities (hereinafter, referred to as “reactor facilities”) and that containment measures
have been adopted as part of severe accident management in other countries.
  We reviewed the report and concluded that the role and significance of accident management
and the results of a technical investigation of containment measures are appropriate. We
acknowledge that the proposals by the group to further promote the development of accident
management measures are significant in that they will contribute to further improvement in the
safety of reactor facilities in Japan.
  We address the issue of severe accident management, taking into account the proposals by the
Committee and according to the following principles. We hope that licensees as well as the
government agency will make further effort according to the principles.

1. The safety of reactor facilities in Japan is sufficiently ensured by current safety regulations
    by implementing, under current safety regulations, strict safety measures in the design,
    construction and operation stages, based on the defense-in-depth concept to (1) prevent the
    occurrence of abnormal events, (2) prevent an abnormal event from spreading and
    developing into an accident, and (3) prevent the abnormal release of fission products. The
    possibility of severe accidents is sufficiently low due to these measures, to the extent that
                                                 22
                                                                            NSCRG: L-AM-II.01

   such accidents could not occur from an engineering viewpoint, and thus the risk from reactor
   facilities is considered to be sufficiently low.
   The development of accident management measures is significant in further reducing the risk,
   which is already low.
   The Commission believes that effective accident management should be developed by
   licensees on a voluntary basis and that its proper implementation in the event of an
   emergency should be strongly recommended.

2. In order to further improve the safety of reactor facilities, licensees should continue to
   develop accident management, using specific proposals made by the report as a reference.
   The government agency should define their role in the promotion and development of
   accident management, and should continue to work on the details of the role.

3. The Commission will hear from the agency, as needed, about specific policies and measures
   concerning accident management. The following are our near-term activities.
   (1) For new reactor facilities to be installed in the future, we will receive a report from the
        government agency on accident management implementation policies (specific
        measures related to equipment and systems, development of procedures, personnel
        education and training, etc.) as soon as possible after the detailed design of the facilities,
        and will review the policies. Based on the review results, licensees should develop
        accident management measures for the facilities before fuel loading.
   (2) For reactor facilities in operation or under construction, we will receive a report from
        the government agency on accident management implementation policies for the
        facilities and will review the policies
   (3) In connection with (1) and (2) above, we will receive a report from the government
        agency on a probabilistic safety analysis of the facilities and will review the analysis.

4. The authorities concerned and licensees should continue severe accident studies. The
   Commission members will update themselves on the results of their studies and will perform
   necessary reviews.




                                                  23