ML20077G497
| ML20077G497 | |
| Person / Time | |
|---|---|
| Site: | Millstone  |
| Issue date: | 08/31/1983 |
| From: | NORTHEAST NUCLEAR ENERGY CO. |
| To: | |
| Shared Package | |
| ML20077G489 | List: |
| References | |
| ENVR-830831, NUDOCS 8308040278 | |
| Download: ML20077G497 (23) | |
Text
- l MNPS-EROLS Each steam line has a two-way fast-closing isolation valve which prevents reverse flow. These valves prevent blowdown of more than one steam generator for any break location, even if one valve fails to close. l It in considered unlikely that an accident of this type will occur.
This accident is considered to be in the " infrequent incidents" category.
7.1.8.6 Large Steam Line Break (Accident 8.3a)
The assumptions used and resulting exposures for this accident analysis are the same as for the small steam line break (Section 7.1.8.5), excluding piping size.
This accident is considered to be in the " limiting fault" category.
7.1.9 Accidents Beyond The Design Bases of The Millstone 3 Plant (Class 9 Accidents)
The National Environmental Policy Act (NEPA), as interpreted by tM U. S.
Nuclear Regulatory Commission in its statement of interim policy .ssued on June 13, 1980, requires that an Environmental Impact Statement 6 prepared at the operating license stage which includes a consideration of site-specific environmental impacts from accident sequences of greater severity than design basis accidents. These accidents are commonly referred to as Class 9 Accidents and are far less likely to occur than Class 1-8 accidents analyzed in Sections 7.1.1 through 7.1.8. The main purpose of this Section is to estimate the consequences and probabilities of Class 9 accidents to address the considerations in the Nuclear Regulatory Commission's June 13, 1980 Interim Policy Statement (45 FR 40101).
7.1.9.1 Scope The Millstone 3 plant design was analyzed to quantify public risk which might result from operation of the plant if a hypothetical Class 9 accident was to occur. Section 7.1.9.2 describes initiating events that might occur which could challenge the plant's safety systems. Section 7.1.9.3 discusses how the containment was modeled and describes the dominant failure modes predicted for containment given a challenge to its integrity.
Section 7.1.9.4 describes the process used to model the public health consequences that might result from those Class 9 accidents which could lead to containment failure and subsequent radioactive releases. The results of the analysis are presented in Section 7.1.9.5 and graphically describe the relationship between the consequences of a hypothetical Class 9 accident and its probability of occurrence.
7.1.9.2 Plant Analysis Mechanical systems and the effect of operator actions at Millstone 3 were systematically studied to model the plant design and operation so that public risk from hypothetical Class 9 accidents could be quantified. This systematic study included:
6308040278 830727 PDR ADOCK 05000423 C PDR Amendment 3 7.1-14 August 1983
MNPS-EROLS
- 1. Identifying initiating events that could challenge plant systems
- 2. Modeling plant response to the. initiating events
- 3. Accounting for multiple equipment failures and human errors in the plant models
- 4. Identifying possible failure modes of plant systems If failures in plant systems could result in severe deterioration of the nuclear fuel and release of fission products, the integrity of the contain-ment structure was considered.
Initiating events were characterized as internal or external events.
Internal initiating events include accident sequences initiated by random failure of plant systems, components, power supplies or support systems essential to plant operation. External events include accident sequences initiated by phenomena external to plant systems such as earthquakes, fires, flooding, extreme winds, aircraft accidents and accidents involving hazardous materials.
Internally initiated events were grouped into distinct categories on the basis of how each initiator relates to events capable of causing core damage. A complete list of Millstone 3 initiators was derived from an extensive review of generic operating experience and an evaluation of any plant specific initiators which did not fall into one of the traditional internal event categories. The resultant categorized liet of internal initiators is shown in Table 7.1-5.
Internal initiating events are divided into loss of coolant accidents and transient events. Loss of coolant accidents are defined as any accident involving a loss of reactor coolant inventory inside or outside the containment, including failure of the reactor coolant system piping (including valves), the pressure vessel, and interconnecting systems.
Reactor coolant pump seal failures are also categorized under LOCA initiators.
Transient events include all events which challenge the core reactivity contro' function, the core heat removal function, the RCS pressure control function, or the RCS heat removal function. Ber.ause Millstone 3 is not an operating plant, no plant-specific operating data was available for incorporation into the analysis; estimates of the initiating event frequency distributions were based on domestic PWR operating experience.
Systems Analysis A complete set of internal initiating events was considered and, by similarity of plant response and ef fect, was reduced into a subset of 21 representative initiating event groups. Event trees and fault trees were used in the analysis of these initiating events.
To address the potential spectrum of initial conditions (availability of critical support system features), a support states model was developed.
Amendment 3 7.1-15 August 1983
..J
, MNPS-EROLS This model considered various combinations of offsite ac power, emergency ac power, engineered safety features logic cutput, load sequencer commands, service water, and 120 V vital ac power as potential initial conditions.
This resulted in the definition of eight logically unique support states, each of which had different impacts on system unavailability. Each event tree was then quantified using fault tree models which accounted for the eight different support states.
The event tree models employ the standard inductive approach used in probabilistic risk analysis. These event trees were constructed to determine the potential outcome from a given initiating event and combina-tions of successful and unsuccessful equipment and operator responses to the initiating event. The event trees also considered the potential of a minor event becoming a more severe challenge to the plant. As an example, the event trees have branch points addressing whether event sequences eventually involve small LOCAs, steamline leaks or breaks, or anticipated transient without scram (ATWS) scenarios. The endpoints of the event trees are categorized into unique groupings based on core damage and other parameters impacting event sequence timing, containment integrity, and 4
fission product scrubbing.
After the plant system event trees were developed, success criteria for each system modeled in the event trees were specified. Each system success criterion was defined with respect to particular accident sequences and system event trees. A majority of the success criteria were based on best estimate sa fe ty analyses; however, certain success criteria rely on the classical conservatively biased safety analyses from the Millstone 3 FSAR.
The fault tree models employ the standard deductive approach used in proba-4 bilistic risk analysis. These fault trees are developed conditional on the existence of one or more of the initial support state configurations.
Common Cause Failures l
l Common cause failures are multiple dependent type failures which can lead i to the total unavailability of redundant safety related systems. The proper treatment of such failures is essential in evaluating the unavail-4 ability of critical safety related systems. The overall common cause failure methodology used a logical framework based on detailed analysis traceable to operating plant statistics. Six basic classes of common cause failures were identified and analyzed:
- 1. Support System Failures Affecting Multiple Systems Examples include failures of ac power, service water, and
! component cooling water.
- 2. Command Failures Examples include failures in ESF actuation logic and failures in the emergency generator load sequencer.
! 3. Multiple Human Errors i
i Amendment 3 7.1-16 August 1983
l l
MNPS-ER0LS Multiple human errors leading to common cause failures include design errors, test / maintenance errors, incorrect calibration and/or operation.
- 4. Environment Temperature, excess moisture / humidity, grit, electromagnetic interference, high radiation, and excessive vibration are examples of environmentally related common cause failures.
- 5. Intersystem Dependency This is typified by RWST/ECCS interfaces and interfaces between the shared piping of the ECCS injection and recirculation systems.
- 6. External Events Seismic events and fires are examples of this class of common cause failures.
Common cause failures due to each of these generic classes were treated in the following manner: Command failures of common support systems were addressed deterministically by the support states modeling carried out in the systems analysis. Common cause failures due to intersystem dependencies or conditional dependencies were treated deterministically in the event tree structure or by use of conditional probabilities in the quantification of the event trees. Common cause failure effects due to external events were assessed by deterministic analysis for each type of external event. All other common cause failures, including multiple human errors, were assessed using a binomial failure rate model based on actual operating plant statistics. Analysis of individual system common cause unavailability was carried out for each fault tree.
Human Reliability Analysis Unlike the reliability analysis of engineered systems, the analysis of human reliability is difficult to quantify. Engineered systems can be mechanistically analyzed in detail for their ability to perform simple, well-defined functions. Human beings, on the other hand, perform multifaceted tasks and are capable of both causing and mitigating greater malfunctions than can be envisioned for hardware systems. A review of recent reactor operating experience confirms this observation. Because of these considerations, the assessment of hypothetical Class 9 accidents at Millstone 3 must consider the impacts of human error on the availability of critical safety related systems and the cognitive behavior of a well trained plant operator to diagnose a particular accident scenario and respond so as to improve the mitigation provided by automatic systems.
Amendment 3 7.1-17 August 1983
I l
?
MNPS-EROLS i
{ Conservative screening values were used for human error probabilities. The
} critical assumptions employed were:
i
- 1. A Level 1 Tagging System (as defined in NUREG/CR-1278) will be i used.
1 l 2. Checklists will be required for test and maintenance activities.
l 3. Safety related panels in the control room will be scanned, using j a checklist, at least once per shift.
These assumptions reflect actual operating procedures for the two operating Millstone units. .
j Cognitive, or decision-making errors at the event tree level were
- considered to better model the role of the operator during event sequences, i The analysis also considered the dominant factors affecting the ability of
{ the operator to perform correctly, including:
, 1. The amount of time available to make a cognitive decision and i carry out an action
- 2. The accident scenario and the likelihood the operator would l properly diagnose the scenario given the available information
- 3. The complexity of the required actions The technical reviews of the human factors analysis included a detailed j- review of assumptions by Millstone 3 operations personnel (a majority of
, whom were licensed on at least one of the operating Millstone units) and a
, review by a human factors consultant.
External Event _s The folicwing external events were evaluated:
I j 1. Earthquakes
- 2. Fires within the plant i
4
- 3. External flooding 1
- 4. Extreme winds l
l 6. Aircraft crashes-i
- 7. Transportation and storage of hazardous materials t 8. - Turbine missiles Amendment 3 7.1-18 August 1983
MNPS-EROLS Each of these potential external events was first screened for risk significance by assessing the frequency of such an event and examining the vulnerability of the plant to damage from that event. Because of specific plant design features or design criteria used in the construction of the plant, it was possible to rule out the likelihood that many of these events could cause any significant damage to the plant. As a result of this screening process, only earthquakes and plant fires were identified as requiring further detailed analysis.
Analysis of core melt frequency due to earthquakes was based on the following strategy:
- 1. The probability of ground accelerations of a given magnitude near the Millstone site due to earthquakes was analyzed
- 2. Seismic fault trees for various core damage states were developed
- 3. Seismic fragility analyses were performed
- 4. Probability distributions for fragilities were developed
- 5. Base events of the seismic core melt fault tree were quantified to develop a seismic core melt frequency and uncertainty
- 6. Seismic related containment event trees were quantified for seismic related containment failure modes Because severe earthquakes capable of damaging the seismically qualified structures and components of Millstone 3 would also damage many offsite buildings and structures and a f fect evacuation routes, conservative assumptions were used in the consequence analysis for modeling warning and evacuation times, evacuation speeds, evacuation routes, and population counting (to account for the presence of rescue personnel).
Analysis of core melt frequency due to fires within the plant was based on the followin; :
- 1. The probability of fires in certain areas was assessed based on existing operating experience
- 2. Mechanistic models of fire propagation and the effects of mitigation were quantified
- 3. Operator actions to mitigate fires and human errors were identified and quantified
- 4. The overall fire related core melt frequency and uncertainty analysis was analyzed The consequence analysis for fire-initiated events was performed in the same manner as was done for other internally initiated events.
Amendment 3 7.1-19 August 1983 1
- MNPS-EROLS i
7.1.9.3 Containment Analysis If, as a result of a low probability sequence of events, the core were to become severely damaged, numerous additional events must occur to cause i appreciable public health consequences. The Recctor Safety Study (WASH-1400) postulated in 1975 that . in all cases containment would fail given that a hypothetical core melt had occurred. The only factors a f fec ting the magnitude of public health consequences in that study were i the postulated timing of containment failure, the specific failure mode, and whether or not sprays were available to scrub the containment atmosphere of radioactive materials. More recent . studies have indicated that containment failure is not an assured event, given severe core damage.
The containment structure and systems therefore provide a major factor of risk reduction. To fully understand the magnitudejof this additional risk reduction factor, an in-depth analysis was performed of the overall containment function.
The containment response analysis used data frca previous studies and analyses of damaged core phenomenology as well as containment atmosphere transient calculations to evaluate the complex and interrelated physical processes which would occur within the coritainment building if a release of reactor coolant or core material were to take place.
Containment Event Tree t
, The containment event tree is the model. used to analyze the interrelated processes which affect containment integrity. This model permits the analysis of the containment response to a hypothetical Class 9 accident.
Possible fission product releases from containment were grouped into discrete release categories as described .in Table 7.1.-6. Each release ,
category is characterized by , variables such as the time span from the-initiating event to release, whether an energetic /' interaction of debris and-water has occurred, and whether containment sprays are operable. _
One of the most important results of the containment analysis was that for the core melt sequences analyzed involving successful operation of containment safeguards, containteent. integrity was found to be maintained.
Dominant Failure Modes "
s
~
The containment analysis identified the most likely _ ways _ in which the containment structure might fail given various reactor system; failures that could result in radioactive releases into the containment. The results of the best estimate containment ~ analyses were grouped according to. the availability of containment heat removal systems. '
- 1. No Containment Heat Removal ,
With no containment quench . spray ' available,' the reactor - cavity . is assumed to be dry at the tinie of a. hypothetical vessel failure and "
' throughout the event. ' For hypothetical large LOCA's, coupled with a complete failure of all av'dilable core cooling system's , - hydrogen Amendment 3 4 7.1-20 , . August 1983
- 14
, , , --- - .. . ,.;i - , . - _ - -
l MNPS-EROLS gene ra t. ion by Molten Core-Concrete Interaction (MCCI) would begin almost immediately. Ilydrogen burns occurring several hours after vessel failure could potentially cause containment failure because of the pressure spike from the burn superimposed on an already large containment steam pressure.
For small LOCA's and transients leading to early core melt (no emergency core cooling injection), the accumulators would be expected to discharge to the reactor cavity. The boiloff of accumulator water would require approximately 30 minutes and would render the containment steam inerted. Potential containment failure due to steam overpressure could occur after one day.
- 2. Full Containment lleat Removal The reactor cavity would most likely be dry at the time of a hypothetical vessel failure. Depending on the accident scenario, spillover of water from the containment sump into the reactor cavity would occur from tens of minutes to hours following vessel failure.
The molten debris would eventually be quenched and MCCI arrested.
Containment sprays could maintain pressure at less than 20 psig, such that if a hydrogen burn were to occur, containment integrity would not be threatened.
- 3. Partial Containment lleat Removal
- a. Quench Spray Only: Containment response up through the time of a hypothetical vessel failure and spillover of sump water into the reactor cavity would be similar to the full safeguards case.
Af ter sprays were terminated, the boiloff of water from the cavity would pressurize the containment and render it steam inerted. Eventually, sump water would cease to spill over into the reactor cavity because the volume of water available to spill over (the volume of the RWST, reactor coolant system, and accumulators combined) is only slightly greater than the volume needed for spillover. (The additional water vapor going into the containment atmosphere is sufficient to lower the level of water in the sump to below the spillover height). Containment failure due to steam overpressure could occur prior to one day,
- b. Recirculation Spray Only: Since all of the water from the RWST is assumed not to discharge into containment, spillover of water into the reactor cavity would not occur. The recirculation sprays maintain containment pressure below 25 psig so even a hydrogen burn involving the full containment volume would not threaten. containment. Basemat penetration could occur.
Containment event trees were constructed to determine the modes and timing of potential containment failure, if any, for the various accident
- sequences. Containment integrity was assured unless both the containment i recirculation spray and quench spray systems were simultaneously l unavailable. For the accident scenarios in which containment spray systems l were unavailable, the containment failures could potentially occur at Amendment 3 7.1-21 August 1983
MNPS-EROLS approximately one day due to either containment steam overpressure, for cases in which water was introduced into the reactor cavity, or to hydrogen burns for Lte dry cavity sequences. Only in the case of recirculation spray without quench spray was basemat failure found to be the dominant containment failure mode.
In all of the scenarios in which quench spray was unavailable, hydrogen generation by core-concrete interaction was found to be high, but because of the basaltic nature of the concrete in the Millstone 3 basemat, the probability of long-term overpressurization of the containment by non-condensible gases leading to containment failure was found to be low. 1 7.1.9.4 Consequence Analysis The offsite consequence analysis for Millstone 3 was classified into three sections. The " airborne pathways consequence analysis" estimates the the l potential ef fects on the population due to exposure from fission products 1 released through the airborne pathway (cloud, ground, inb lation, and l ingestion doses) as a result of a hypothetical Class 9 accident. The
! " liquid pathways consequence analysis" considers the potential consequences resulting from fission products released directly into liquid pathways.
- The " airborne rainout to fish flesh pathways consequence analysis" considers rainout of the airborne fission products from a hypothetical Class 9 accident into water bodies, contamination of fish, and consumption of the contaminated fish by the public.
7.1.9.4.1 Airborne Pathways Consequence Model This section briefly describes the methodology employed to estimate the potential consequences of airborne releases of fission products subsequent to a postulated Class 9 accident. The CRAC-2 code, an updated version of the code used in the Reactor Safety Study, was used for this purpose.
CRAC-2 models the transport of fission products that could be released from the containment. CRAC-2 considers the likelihood of relative wind directions, vertical rise of the plume (which depends on the release energy associated with a particular release category), radioactive decay of the nuclides, the likelihood of rainout, population density, and the effects of evacuation.
i Radiation doses may be received by individuals from the passing radioactive cloud (plume) and from radioactive material deposited on the ground. The cloud doses due to direct radiation by inhalation of the radioactive material suspended in the air would last only during the passage of the
! cloud over the affected population. Doses from deposited radioactive material may be received via three paths: direct radiation from the radionuclides, inhalation of resuspended material, and ingestion of contaminated food and water (a conservative assumption). The CRAC-2 code simulates all these dose paths.
Amendment 3 7.1-22 August 1983
MNPS-EROLS To assess the effect on the entire population, the individual doses were combined with the population distribution. The area within a 350 mile radius around the Millstone 3 reactor was modeled on a circular grid and the population in each area was modeled based on 1980 census data.
Several protective measures that could reduce radiation doses were also modeled in the CRAC-2 code, including evacuation of the nearby population to prevent or limit the cloud dose and ground dose, sheltering of non-evacuees to limit potential doses, long term relocation of people, and i interdiction and decontamination of land in the contaminated area.
i I Population health effects were estimated from calculating radiation doses i ar.d dose response characteristics. Ilealth effects evaluated include acute fatalities, acute injuries, latent cancer fatalities, thyroid nodules, and total population whole body dose.
7.1.9.4.2 Liquid Pathways Source and Required Failures If a hypothetical core melt were to occur, penetration of the concrete containment basemat could occur. Despite its low probability of occurrence, basemat penetration was included in the liquid pathways analysis because it is the only identified means by which any significant quantity of radioactivity might be released to the local groundwater.
, There are three possible ways in which radioactivity could be released following basemat penetration. First, fission products might be leached from the core debris as a result of immersion of the debris in groundwater.
Second, if the containment atmosphere is at an elevated pressure at the time of basemat melt-through, the atmosphere could hypothetically vent through tortuous paths around the containment foundation. Third, water in the containment sump could be released via the failed basemats this mechanism requires the additional unlikely failure of the walls which separate the containment sump from the reactor cavity. Of these potential release mechanisms, sumpwater escape was the most significant because the sumpwater could contain a large fraction of the total volatile core fission product inventory: however, as described in FSAR Section 2.4.13.2, subsurface transport of radioactive material is v(r. < "ikely.
Site Characteristics, Transport and Interdiction The groundwater regime within the underlying bedrock at the Millstone site can be characterized as follows:
- 1. Groundwater is not likely to be present in the bedrock beneath the containment basemat
- 2. If water is present, or if it is introduced (for example by a sumpwater l
release through the basemat melthole), it is not likely to be flowing nor, in the latter case, is it likely to move vary far from the point of introduction
- Amendment 3 7.1-23 August 1983 1
l
MMPS-EROLS
- 3. If water is present and is flowing, it will flow laterally to Long Island Sound (within the uppermost 100 feet) and not vertically to a deep aquifer. If there is groundwater movement, it would be very slow because the underlying rock has a very low permeability
- 4. The time required to traverse the distance from the containment to Long Island Gound would be on the order of many years at a minimum These characteristics are especially significant with respect to the possibility of successful source interdiction, i.e., measures to prevent the outward spread of fission products if they are released through the containment basemat.
If water is not present in or flowing through the bedrock, no interdictive measures would be necessary since any radioactive materials would not be able to flow out of the bedrock. If water is present and flowing, it will flow laterally and not vertically and, because of the very low permeability of the materials underlying the containment it would take many years to travel to any water bodies (See FSAR Section 2.4.13.2). Interdictive measures such as drilling a line of wells or erecting a grout curtain in the bedrock (as was done in the construction of the Millstone 3 circulating water discharge tunnel) in the path of, or to surround the radioactive materials, could be used to preclude liquid transport of radioactive materials following a hypothetical release of radioactive materials from the containment through the basemat.
7.1.9.4.3 Airborne Rainout To Fish Flesh Pathways Consequence Analysis The radiological impact of the rainout to a large water body of the radioactivity released to the air pathway following a postulated core melt accident was also evaluated. If the radioactivity were to rainout to a water body with limited circulation, potential radiological impacts to the public might prove to be a significant dose pathway. The probability of such an occurrence is very low, however, when the probabilities of a core melt accident, the applicable wind direction, and the occurrence of rain at the same time are all taken into consideration. An analysis was performed to estimate the potential impact of radioactivity released from the containment via the air pathway and subsequently precipitated into water bodies with limited circulation. The analysis included a scoping study to address isolated water bodies (such as inland ponds, reservoirs, and coastal lagoons) and a dose assessment study to address fallout to Long Island Sound. Based on previous studies, the dose pathway to the public judged the most important for this study was the consumption of contaminated fish. Thus, this dose pathway was the only one considered in this study of rainout to Long Island Sound.
Analysis of Fission Product Rainout To Isolated Water Bodies A scoping study was conducted to assess the consequencee of rainout of fission products released to the air pathway following a hypothetical core-melt accident involving containment failure. Since water bodies having good circulation or flushing would transfer their contaminants to Long Island Sound, the scope of this analysis was limited to isolated water Amendmer.t 3 7.1-24 August 1983
i 4 MNPS-EROLS ,
bodies. Initially, a survey was made of the coastline and inland areas within 50 miles of the site to identify coastal bays and lagoons and inland ponds lakes and reservoirs. Meteorological data from the Millstone station was analyzed to determine the probability of wind directions and the probability of coincident rain and wind direction. Based on this analysis, two representative water bodies (Groton Reservoir and Pachaug Pond) were [
selected for further evaluation. Probability of contamination of these two lakes was estimated and volumetric turnover rates of water in these lakes were determined to estimate flushing rates. The following paragraphs present the details of this analysis.
Survey of Coastline Near Millstone Characteristics of the coastline within 50 miles of the Millstone plant site were studied using U.S. Geological Survey maps. The primary objective I of the survey was to identify any bays and lagoons within this area that could become contaminated subsequent to a postulated core-melt accident at the plant and remain contaminated on a long term basis because of low s flushing rates.
Occurrence of rainout from a plume could cause deposition of a significant amount of fission products which could contaminate nearby water bodies.
However, tidal exchange would tend to deplete the contaminants. In the case of shallow bays and lagoons, this process would be relatively rapid (on the order of a few days) due to generally high turnover rates. For
- deep bays and lagoons, the flushing process would be slower.
Thus, emphasis was placed on identifying deep bays and lagoons within the geographic area of interest. The investigation of the coastal area revealed that the bays and lagoons in the area are generally shallow and have ready access to the Long Island Sound, auch that prolonged contamination of these water bodies is not likely.
It was thus concluded that the only bodies having the potential for experiencing long term contamination would be fresh water bodies in the region.
Meteorological Data Analysis The meteorological data at the Millstone site was analyzed to identify lakes, ponds, or reservoirs which lie in areas where the combined probability of wind direction and rain is highest. The data reviewed i included direction and intensity of wind and the amount of precipitation i
(if any) for every hour within a one year period. The frecuency of wind in the various directions was determined from the hourly data. The resulting
" wind rose" is shown on Figure 7.1-1. The probabilities are plotted at the
, mid points of each sector, i
The combined probability of wind direction and rain was also computed from the Millstone station meteorological data and plotted on Figure 7.1-2. The t
Amendment 3 , 7.1-25 August 1982
MNPS-ER0LS ranking of combined probabilities is different from that of wind direction probabilities alone because the conditional probabilities of rain are different for the various wind directions (occurrence of rain and direction of wind are not independent events).
Review of Fresh Water Lakes The sectors with the highest combined probabilities of rain and wind overall lead to the Atlantic Ocean, away from the fresh water lakes in the area. Therefore, the study considered two representative inland water bodies (Groton Reservoir and Pachaug Pond) located in overland sectors with '
the highest combined probabilities of v! id and rain. Groton Reservoir and Pachaug Pond are also representative in that one is close to the Millstone 3 site and one is further away, and one body is used for drinking water while the other is primarily for recreation.
Groton Reservoir is located approximately 7 miles east-northeast of the Millstone site in sector 4. This is downwind from the sector with the fourth highest combined probability of rain and wind. Groton Reservoir will have the highest probability of contamination of any inland water in the area. Pachaug Pond is located 22 miles northeast of Millstone in i sector 3. This sector has the fif th highest combined probability of rain l and wind. l l
Probability of Contamination The probability of contamination of these lakes in the unlikely event of a fission product release was determined.
The site meteorological data indicated that the average wind speed at Millstone is approximately 12 miles / hour. In the case of Groton Reservoir, which is approximately 7 miles from the plant, a plume released from the plant as a result of a hypothetical Class 9 accident would reach the reservoir within I hour, provided the wind was blowing in that direction.
In the absence of rain, there would not be significant contamination because most of the particulate fission products in the plume would not settle and would be carried away with the plume. If the plume were subjected to rain, most of the particulate fission products as well as iodine would be washed out, result.ing in contamination of the area below.
The probability of contamination of Groton Reservoir was therefore calculated as the probability of the coincident occurrence of rain and wind in the given direction. The combined probability of rain and wind toward sector 4 (downwind from sector 12) is 0.0070. If it is assumed conservatively that wind in the adjoining directions (from sectors 11 and
- 13) could also contaminate the reservoir when coincident with rain, to take into account the effects of potential plume meandering and collection of run-off from nearby areas, the combined probability increases to 0.026.
Thus a conservative estimate of the conditional probability of contamination of the Groton Reservoir, given a release from Millstone 3, is 0.026.
Pachaug Pond is located about 22 miles from Millstone. Since the average wind speed at the site is 12 miles / hour, the average time for a plume released from Millstone 3 as a result of a hypothetical Glass 9 accident to Amendment 3 7.1-26 August 1983
1 MNPS-EROLS l
reach Pachaug Pond is approximately 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />. In the event of rain coincident with plume release, the majority of the fission products in the plume would be washed out before the cloud reached Pachaug Pond. As in the case of Groton Reservoir, wind in the direction of the adjoining sectors j was also assumed to cause contamination when accompanied by rain. Wind directions 10, 11, and 12 were thus evaluated, and the probability of contamination of Pachaug Pond given a release from Millstone 3 was conservatively calculated as 0.010.
The probabilities of contamination estimated for the two lakes are conditional upon the occurrence of a significant release subsequent to a hypothetical core melt accident at Millstone 3. Because the probability of such a release is extremely low, conservative estimates of the probability
- of contamination of Groton Reservoir and Pachaug Pond are lower by a factor i of approximately 100.
Volumetric Turnover Rate l In addition to examining the likelihood of contamination, transient effects were also examined such as the contaminant concentration as a function of time. One of the primary determinants of depletion rate is the lake water turnover rate from the continuous flow of water into and out of the lake which will dilute the contaminant concentration in the lake. Volumetric turnover time is defined as the ratio of the lake volume to the volumetric flow rate and is expressed in units of time (days). The shorter the turnover time, the higher the depletion rate.
Groton Reservoir has a volume of approximately 2.27 x 109 gallons and an average incoming flow rate of 1.62 x 107 gallons / day which results in an average turnover time of 140 days. Pachaug Pond on a similar basis has an l
average turnover time of 25 days.
Relative contaminant concentration may be estimated using a simple model i
assuming perfect mixing. The concentration will decrease at an exponential rate with a volumetric turnover half life of 97 days for Groton Reservoir and 17 days for Pachaug Pond. Thus, the fission product concentration will decrease by a factor of 14 in one year for Groton Reservoir, and by a factor of 2 x 106 for Pachaug Pond.
Analysis of Fission Product Rainout To Long Island Sound An analytical model was developed to compute individual and population j doses from the consumption of fish assumed to be contaminated following a hypothetical core melt accident at Millstone 3. In the scenario analyzed, the airborne radioactivity released from the containment is transported to Long Island Sound. All of the airborne radioactivity (excluding noble gases) is conservatively assumed to be deposited into the Sound by rainfall I
and the radioactivity was assumed to mix uniformly in the water body.
Mechanisms for depletion of radioactivity from the Long Island Sound considered in the model are radioactive decay and water volume turnover rate due to tidal flushing and fresh water inflow.
l l
Amendment 3 7.1-27 August 1983
MNPS-EROLS It was conservatively assumed in the model that the fish in Long Island Sound remain in the contaminated water until harvested and that the fish are consumed immediately af ter being harvested. Such a pathway would not, in all likelihood, exist under actual accident conditions since contaminated fish would be identified and quarantined prior to consumption.
7.1.9.5 Results The public health consequences of hypothetical Class 9 accidents are expressed in terms of the frequency of exceeding a given consequence for each of the following four risk categories:
- 1. Early fatalities due to internal events
- 2. Latent cancer fatalities due to internal events
- 3. Early fatalities due to external events
- 4. Latent cancer fatalities due to external events The probability curves are presented on Figures 7.1-3 to 7.1-6. The risk curves are expressed in terms of best estimate (50th percentile or median) risk curves and upper bound (or 90th percentile) risk curves which include the effects of uncertainties on the assumptions, modeling, and analyses used to study the plant. The 50th percentile curves represent the best estimate of the risk, whereas the 90th percentile curve has a ninety percent probability of bounding the risk when uncertanties are included.
The curves are based on results of the analyses discussed in this section and reflect combination of the following probabilities:
- 1. Initiating event frequencies
- 2. Quantification of system failure probabilities in the system analysis (using event trees and fault trees)
- 3. Containment event tree quantification (using containment analysis techniques)
These curves demonstrate that the risk to the public from the operation of Millstone 3 is very low.
7.1.10 References for Section 7.1 Underhill, D. 1972. Effect of Rupture in a Pressurized Noble Gas Absorption Bed. Nuclear Safety, 13(6), 1972 United States Atomic Energy Commission (USAEC) 1974. Regulatory Guide 1.4, Revision 2. Of fice of Standards Development, June 1974.
Untied States Nuclear Regulatory Commission (USNRC) 1976a. Regulatory Guide 4.2. Office of Standards Development, January 1976.
United States Nuclear Regulatory Commission 1976b. Calculation of Release of Radioactive Materials in Gaseous and Liquid Effluents from Pressurized Water Reactors (PWR-CALE code). Office of Standards Development, April 1976.
Amendment 3 7.1-28 August 1983 J
MNPS-EROLS TABLE 7.1-5 INTERNAL INITIATING EVENT CATEGORIES Events i
- 1. Large LOCA
- 2. Medium LOCA
- 3. Small LOCA
- 4. Steam Generator Tube Rupture
- 5. Steamline Break Inside Containment
- 6. Steamline Break Outside Containment
- 7. Loss of RCS Flow
- 8. Loss of Main Feedwater Flow
- 9. Primary to Secondary Power Mismatch
- 10. Turbine Trip
- 11. Reactor Trip
- 12. Core Power Excursion
- 13. Spurious Safety Injection
- 14. Loss of offsite Power
- 15. Incore Instrument Tube Rupture
- 16. Special Large LOCA Initiators
, a. Interfacing System LOCA
- b. Catastrophic Reactor Vessel Rupture
- 18. Special Initiators
- a. Loss of Service Water
- b. Loss of Vital ac
- c. Loss of dc J
i i
I l
l l Amendment 3 1 of 1 August 1983
MNPS-EROLS TABLE 7.1-6 Release Categories Category Description M-1A This release category is used for core melt accident sequences where a containment bypass directly to the environment exists through the EUR system and auxiliary building. Such a pathway can result from failure of the interfacing valves separating the high and low pressure portions of systems connected to the RCS.
M-1B This release is used for core melt accident source terms where a containment bypass directly to the environment exists through a steam generator tube rupture.
M-2 These release categories are used for those accident M-3 sequences which lead to an early overpressure of the j containment with no containment sprays operational. Release category M-2 accounts for early core-melt sequences with a short warning time for evacuation. Release category M-3 accounts for late core-melt sequences with a slightly longer warning time for evacuation.
M-4 This release category is used for core-melt sequences with failure of containment isolation function.
M-5 These release categories are used for core-melt accident M-6 sequences which lead to intermediate containment failure times without containment sprays operational. Release category M-5 accounts for early melt sequences and M-6 for late melt sequences. -
M-7 This release category is used for core-melt accident sequences which lead to late containment failure times without containment sprays operational.
M-8 This release. category is used for core-melt accident sequences which lead to intermediate containment failure times with functional containment sprays.
M-9 This release category is used for core melt accident sequences which lead to late containment failure times with l
functional containment sprays.
l M-10 These release categories are used for core-melt accident i
M-ll sequences which lead to basemat melt-through. Release l category M-10 is used for the case of containment sprays nonoperational and M-ll for operational sprays.
l M-12 This release category is used for core-melt accident sequences which where containment remains intact. All sequeces in this release category have continuous spray operation.
Amendment 3 1 of 1 August 1983
[
O N
b G
@ G
@/ \
~
g\ .
g
.06
@ O
.10
.12
\
FIGURE 7.l-1 MILLSTONE WIND ROSE PROBABILITY OF WIND DIRECTIONS MILLSTONE NUCLEAR POWER STATION UNIT 3 ENVIRONMENTAL REPORT AMENDMENT 3 AUGUST 1983 L
h N
G _
W \ @
.00s
.010
.012 FIGURE 7.1-2 MILLSTONE WIND ROSE JOINT PROBABILITY OF RAIN AND WIND ROSE MILLSTONE NUCLEAR POWER STATION UNIT 3 ENVIRONMENTAL REPORT l
\
AMENDMENT 3 AUGUST 1983 i
[
i f
s.0-7 - - - - - - - -
7.0-8 :- ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
E 5.0-8 -: :
e 3.0-8 :~ ::
2.0-8 .. _.
l.0-8 .
O 7.0-9 _E i o 5.0-9 .: .
< 3.0-9 .. ..
Y 2.0-9 . _
l 6 1.0-9 _
A 7.0-10 :.E .9 :.E w 5.0-10 -: _:
y 3.0-10 :: .:
< 2 0-10 - _.
O W
w I.0-10 _ .
W 7.0-11 _E w 5 0-1i .: _ :
E 3.0-11 _: _:
2.0-11 . .
0 w
4.0 11 _.
7.0-12 x .
8 5.0-12 -: .5 ::
y J.0-12 ~: ::
u 2.0-12 __ _.
s.0-12 . ..
- N 9 m N O O O O O O O O O O O
- - N O m N O O O O O O
- m m m N O E ARL Y F Ai *.' l i l ES FIGURE 7.1.3 RISK DIAGRAM FOR EARLY FATALITIES DUE TO INTERNAL EVENTS MILLSTONE NUCLE AR POWER STATION UNIT 3 ENVIRONMENTAL REPORT AMENOMENT 3 AUGUST 1983
, ,,, ,I g 7 . . E ,, . I I I 3:8:1 -! -:
r a
a 3:8:8 :'. 9 -!
e +:g:) is o 5 0- --: -:
w 3 g:;
,5 e }: -
w -
- n. 5 .- : .:
m --
O ._8 -. -.
E 3!
8 4:R:@
5 D-9 E: -E E 3:8:8 :! ::
,t 0-9 : ;
g =,
-E 5.$-18
.o-ig 2 0-10
[. -- --
E 4 R:' 9 : '
i 8 5:D- 1 -: -E m 3.0-ii -: -
r 2 0-li u - --
+*a:5 5 0- 2
~:
. 3:8:13 i: !!
1 0-12 . --
, :..... : ; l ;;; , ; , . . , -
- N9 to A O O O OOO O O OOO O O OOO OO 00O
- N 9 OA O O O OOO O O OOO O O OOO N 9 CAO O O OOO O O OOO
- N n in s o O O C00 LATENT FATALITIES FIGURE 7.1-4 RISK DIAGRAM FOR LATENT CANCER FATALITIES DUE TO INTERNAL EVENTS MILLSTONE NUCLE AR POWER STATION UNIT 3 ENVIRONMENTAL REPORT AMENDMENT 3 AUGUST 1983
. 1 l
x w
1.0-5 -
.o 7.0-6 .5 ::
o 5 0-6 -
w . -
" 3 0-6 .. -
e w 2.0-6 . -
n_
w M
.0-6 ..
g :
o 7.0-7 -5 ~:
w -. ~
w 5 0-7 o .
x -.
.5 -
W 3 0-7 -
u.
o 2 0-7 __ --
u E e.0 7 . -
o a
w , , ,
e .:,, 1 u
1.0-10 , , , , , , , ,,, ,,, ,
N m in N o o o ooo o o ooo o o ooo ooo
- m n nn k o o o ooo o o
- m m in s o o o ooo
- m n in A o
~
E ARLY F A7 ALITIES FIGURE 7.1-5 RISK DIAGRAM FOR EARLY FATALITIES DUE TO EXTERNAL EVENTS MILLSTONE NUCLEAR POWER STATION UNIT 3 ENVIRONMENTAL REPORT AMENDMENT 3 AUGUST 1983
l.0-4 ,
- 1
- 1 7 0-5
- r -:
E, g
5.0-5 p . .
r 3.0-5
.9 o
- 2.0-5 . .
u w
5
.5 i.0-5 .. .
E 7.0-6 55 5 w
5.0-0 :: :
a 3.0-6 . .
_a w 2 0-6 w
O Y
l.0-6 . .
g . -.
o 7.0-7 .: .-
C 5.0 7 :: :
Z - -
% 3.0-7 _
o L-x 2.0 7 . .
L e.0-7 _
i.O~9 , ,, ,,, , , ,l,, l l lll ,,, , ; l ; ;';,, , l l
- am in N o oo coo oo coo oo ooo oo ooo
- Nn mA o oo ooo oo ooo oo ooo
- Nm tA N o oo ooo oo OoO
- N M LO N o oo ooo
~ "' ' " "
L ATENT FAT All TIES FIGURE 7.1-6 l RISK DI AGRAM FOR LATENT l CANCER FATALITIES l
DUE TO EXTERNAL EVENTS MILLSTONE NUCLEAR POWER STATION UNIT 3 ENVIRONMENTAL REPORT 1
l AMENDMENT 3 A UGUST 1983 l