ML19345D439
| ML19345D439 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 12/15/1980 |
| From: | TENNESSEE VALLEY AUTHORITY |
| To: | |
| Shared Package | |
| ML19345D436 | List: |
| References | |
| NUDOCS 8012150111 | |
| Download: ML19345D439 (200) | |
Text
{{#Wiki_filter:EI;CLOSGE 2 9 TENNESSEE VALLEY AUTHORITY SEQUvfAH NUCLEAR PLANT CORE DEGRADATION 2ROGRAM VOLUME 2 REPORT a
- ON THE SAFETY EVALUATION OF,THE INTERIM DISTRIBUTED IGNITION SYSTEM DECEMBER 15, 1980 h
*m esp me> am il 8012150 4
- e. AaLe. C.
_. C0!., . e... . .-
SUMMARY
I. D AS!C CONCEPT OF CONTROLLED IGNITION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I- 1 A. Amount of Hydrogen Frem Events Beycnd the Current Regulatory Requirements for Design Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I-l B. Potential I= pacts on Containment ................................I-2 C. Controlled Ignition as a Potentia'. Means of Limiting IeIact . .. ..I-3 II. INTERIM DISTRIBUTED IGNITICN SYSTEM ................... .............II-1 A. Overview ........................................................!I-1
- 3. Igniter ...................... ..................................II-1 C. Location ........................................................II-2 D. P owe r S u p p ly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I I- 5 E. General Requirements ............................................II-5 F. Igniter Tests ...................................................II-6 G. Operation .......................................................II-6 H. Surveillance Testing ............................................II-7 I. Q ua li fi c a t io n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II-7 III. EVENTS ..............................................................III-1 A. Overview ........................................................III-2 i
l B. Chara c t eriza tion o f Even t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-4 i l C. Mass, Energy, and Gas Releases ..................................III-10 l
. D. Basic Assumptions ..'.............................................III-22 E. Ca lcu la tional Me thod s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III-2 3 IV. DEFL AGR ATION AND DETONATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-1 A. Overview ........................................................IV-1
- 3. Conc e n t ra t ion E f fe c t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV-2 C. I= pact of Water Droplets ........................................IV-7
D. Flame Temperatures .......................... ...................IV-10 E. F lamma bili ty L imi t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV
- 2 F. Ignition Requirements ...........................................!Y-13 G. F law.e S p e e d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV -2 1 H. Burn Efficiency .................................................IV-27 V. CONTAINMENT PROCESSES AND DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .V- 1 A. Overview ........................................................V-1
- 3. Combustible Gas Control System ..................................V-1 C. Containment Structures ..........................................V-2 D. Ice Condenser System ............................................V-3 E. Containment and Residual Heat Femoval Spray Systecs . . . . . . . . . . . . .V-3 F. Air Return Fans .................................................V-4 VI. HYDROGEN DISTRIBUTION ...............................................VI-1 A. Overview ........................................................VI-1 B. Factors Promoting Nenuniform Concentrations .....................VI-2 C. Factors Promoting Uniform Concentrations ........................VI-3
, D. Program Plans ...................................................VI-5 i r l 1 VII. ANALYSIS OF CONTAINMENT CONDITIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .VII- 1 i A. Overview ........................................................VII-1 B. Methods, Model, Assumptions, Inputs, Environmental t Conditions , Results , and Verifications . . . . . . . . . . . . . . . . . . . . . . . . . .VII 4 I C. Containment Environmental Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . .VII 4
~!!II . E NV I RONME NT A L IMP ACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII- 1 A. Overview ........................................................VIII-1 l
B. Containment Boundary ............................................VIII-1 C. Containment Interior Structures .................................VIII-5 D. Critical Containment Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .VIII-9 E. Containment Shell Temperature ...................................VIII-11 L
IX. CCNCLUSICNS
.........................................................IX 1 A.
Ov e. r v i e .' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I X - 1 B. Ignitien Source Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX- 1 C. Contair. ment Heat Removal ........................................IX-2 C. Environmental Effects ...................................... ....IX-3 1 E. Sc o pe o f Ev en t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX 4 F. Margins Availa ble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX-5 G. Uncertainties ...................................................IX-6 4 APPENDICES A. Oc=puter Codes B. Event Definition C. A Description of Phenomena '4hich May Affect the Post Core Damage Event Distribution of Hydrogen in the Sequoyah Contain=ent D. Ncnsy==etric Containment Loads 1 1 E. Critical Components F. Ther:al Effects on Compenents G. Pressure, Shock, and Transient Leads on Components H. Contain=ent Respense to Detonations I. Contair.:ent and Associated Systems J. Contair.=ent Structural Capability Analysis K. Contain=ent Menitoring Systems L. Igniter Syste: Design l M. Igniter Tests - Endurance 2nd Acceptance Tests N. Igniter Tests - Environmental Conditions l O. (Ocitted) l P. Sensitivity Analysis
- 0. Analysis of Effects of TMI-2 Events l
I
R. Operating Procedures S. TVA Degraded Core Investigation Program T. CLASIX Program Description U. Summary of Analyses of Ice Condenser Containment Response to Hydrogen 3 urn Transients V. Progress Report on Verification of CLASIX
Si 0'fDY A safety issue that has received increased attention since the accident at Three Mile Island is the potentially hazardous production and buildup of hydrogen within the containment vessel. The design basis hydrogen production for the Sequoyah Nuclear Plant is limited, like all U. S. nuclear power plants, to that amount conservatively calculated to occur during a large LOCA. These amounts are relatively small and rate-limited and are completely mitigated at Sequoyah by redundant hydrogen recombiners. However, since operator actions during the TMI accident resulted in primary coolant system configurations that led to core damage accompanied by hydrogen production beyond the design basis spectrum, TVA and the NRC have reassessed the effect of hydrogen on plant safety at Sequoyah. Some results of TVA's initial study of the potential for hydrogen production, its effects, and its mitigation were summarized in Volume I of the Sequoyah Nuclear Plant Hydrogen Study. This included a review of design basis accidents, their analyses, and their mitigation l followed by a selection of more severe accidents, their analyses for l physical consequences as well as risk, and an evaluation of their l potential mitigators (nitrogen inerting, filtered veated containment, controlled ignition, etc.). l l The present report, Volume II of the Sequoyah Nuclear _ Plant Hydrogen l Study, addresses the potential effectiveness of one of the accident mitigation concepts recommended in Volume I: controlled ignition. l This report opens with an overview of the current safety and licensing I l
r situation and then describes the recently-installed interim i distributed ignition system (IDIS), including its operating instructions and some details of its testing. Then, the selection of
; severe core damage accidents for evaluation of the IDIS is discussed.
A survey of ,h; fundamental physical relationships in hydrogen transport and combustion follows, with an emphasis on those factors likely to be significant in the Sequoyah Nuclear Plant containment equipped with the IDIS. A descriptien of the original containment i systems and structures that would also affect the course of an accident involving hydrogen production is presented. The report concludes with a summary of the consequences predicted for severe accidents as mitigated by the IDIS and an examination of some of their environmental impacts on the plant. Various appendices detail specific topics, including results of qcntainment response analyses. In brief, the IDIS consists of approximately 45 thermal glow plugs located throughout the lower, upper, and ice condenser compartments of the containment that are designed to be energized immediately following the start of an accident tr Tt has the potential for core damage. The igniter type installed has passed preliminary functional tests, and the system is powered by a reliable ac source. Accident sequences similar to the TMI event were analyzed based on the expected i behavior of hydrogen during its release, mixing, and burning in the i
- Sequoyah containment under the influence of the IDIS and the other systems. From these analyses, TVA concluded that the IDIS as installed provides a net positive benefit to effectively mitigate a
! wide range of accidents involving hydrogen generation.
I SECTION I BASIC CONCEPT OF CONTROLLED IGNITION A. Amounts of hydrogen from events beyond the design brais accident The limiting design basis for hydrogen production at the l Sequoyah Nuclear Plant has been based en a large loss-of-coolant accident (LOCA). This is in agreement with currect Nuclear Regulatory Comission regulations regarding hydrogen control (10 CFR Section 50.44 and GDC 50 in appendix A to 10 CFR Part 59). Using conservative licensing calculations, relatively small amounts of hydrogen are predicted to be produced due to oxidation of zircaloy cladding in the reacto.' core in the short term during a LOCA and radiolysis of slump water and corrosion of various materials in the containment in the 1 ..g term following a LOCA. These design basis amounts of hydrogen are currently capable of being accommodated by j redundant hydrogen recombiners with a low-volume purge system i as backup. I The accident at Three Mile Island 2 occurred when a stuck-open relier valve vented the primary system to the containment; effectively a small LOCA. Even though the automatic safety i systems responded as designed, manual override by the operators led to the core being uncovered for longer periods than assumed in the design basis calculations. As a result, I significant core damage: occurred and more hydrogen was generated at a faster rate from the oxidized zircaloy than 1 i I-1
had been considered previously. The TMI-2 accident has caused TVA and the NRC to reassess the effect of hydrogen on plant safety at Sequoyah. Even though it is believed that the present emergency core cooling systems are completely adequate and that significant improvements in instrumentation, other system hardware, and operator training have been made since TMI, evaluations are being made of that class of events in which significant hydrogen production does result from inadequate core cooling. The NRC is planning to issue an advance notice of rulemaking on degraded core accidents shortly which will result in examination of these events and possible recommendations for their mitigation. Presently, for Sequoyah, TVA's attention is focused on events with core damage limited to approximately 75 percent since more severe degradation would almost certainly lead to complete core melt which has a different associated set of conditions to mitigate. Even with this restriction in reope, the rate of hydrogen production and release is extremely accident scenario-dependent. A representative sequence selected for a number of TVA's evaluations has been S D 2 from WASH-1400, a smali LOCA followed by loss of core injection. B. Potential impacts or containment Most of the hydrogen produced in the TMI accident was released to the containment in the first few hours. About ten hours I-2
into the event, a sudden surge of pressure of about 28 psig was recorded inside the containment by the steam generator pressure sensors. Almost immediately, several thermocouples distributed throughout the containment showed about a 50 F temperature increase. The pressure spike decayed very quickly, while the temperature fell more slowly. Although it is generally agreed that the transient was due to rapid hydrogen combustion, estimates of the amounts of hydrogen responsible for the transient vary widely. The Sequoyah containment is designed for a differential pressure of 12 psi and a peak temperature of 250 F. Due to the margins in design, TVA believes that the containment would withstand the pressure and temperature loadings actually recorded during the TMII 2 accident. However, because of the smaller volume of the Sequoyah ice condenser containment when compared to the 7MI dry containment, the addition of beat from the adiabatic burning of some of the estimated amounts of hydrogen produced at TMI would result in pressures exceeding the ultimate Sequoyah containment capability. However, a determination of the containment capability is very scenario-dependent, since the rate of energy release is at least as important as the total magnitude. In addition, the degree of conservatism vr 3 realism in the calculational technique-and assumptions used in the evaluation is extremely important to the overall conclusion. C. Controlled ignition as hydrogen nitigation I-3
TVA is convinced that the current capability of the Sequoyah containment for hydrogen control is acceptable. However, to increase this capability in the short term, several steps are being taken. Foremost among these is the installation of an Interim Distributed Ignition System (IDIS) consisting of thermal glow plugs distributed throughout the containment available after a degraded core accident to initiate combustion of any significant amount of hydrogen released. Since the containment capability is more dependant on the rate of hydrogen release and combustion than on the total magnitude, the rationale behind the ignition system is to ensure burning in a controlicd manner as the hydrogen is released instead of allowing it to collect and then be ignited by s random source. This more gradual addition of the heat of I combustion will allow the active and passive containment heat sinks to reduce the overall impact. Although the rate of , hydrogen release predicted for various accidents is very sequence-dependent, the ignition system should be effective ' for a wide range of accidents and associated release rates. ! Even though the release locations also depend on the accident scenario, mixing due to containment turbulence and proper location of efficient igniters should allow combustion of hydrogen at even lean concentratione as it is released. The containment spray systems, ice condenser, and internal l structures all serve as effective heat sinks to moderate the temperature increases due to controlled combustion. Another l advantage of controlled ignition is in reducing the potential I-4 i J
ror local detonations due to pocketing over a period of time. The only potential drawback of initiating combustion in an undesirable manner with the igniticn system is already present due to the numerous random sources of sparks in the containment. Therefore, TVA has concluded that the IDIS should result in a net benefit by providing added margin to the containment capability for hydrogen control during degraded core accidents. e i . l l l
-5
I SECTION II INTERIM DISTRIBUTED IGNITION SYSTEM A. Overview TVA has designed an igniter system that we believe will burn quantities of hydrogen at low volumetric cencentrations. The ignicer assemblies are seismically designed not to damage safety-related equipment inside centainment. They are powered i by a reliable ac source and in case of loss of offsite power will be powered by the diesels. TVA performed preliminary tests in 4 selecting the igniter to verify that it would burn hydrogen and operate for reasonable periods of time. Additional environmental and endurance tests on the entire igniter assembly have also been scheduled to assure that it is a reliable and effective design. TVA has installed the igniter assemblies throughout primary containment at Sequoyah unit 1 and we have made arrangements to periodically perform surveillance tests on the installed system to verify operability. B. Igniter The igniter that TVA has selected is a General Motors AC Division Model 7G. This glow plug is commonly used in diesel engines and is commercially available. The igniter is powered directly from a 120/14V ac transformer contained inside the igniter assembly box. The box has overall dimensions of 8"w x 8"h x 6"d and is constructed of 1/8-inch steel plate. It is covered by a spray shield and has a 1/16-inch copper heat sink mounted to the glow plug on the face of the igniter assembly. A more detailed II-1 1
description of the igniter assembly is contained in appendix L to this report. ! C. Location GQl0 The interim distributed ignition system (IDIS) has 45 igniter locations distributed throughout the lower, upper, and ice condenser compartments. Of these 45 locations, igniter assemblies I are currently installed in the 32 key, original locations. The 13 4 additional locations (all in the upper compartment) were selected F recently to assure a visible degree of confidence in the complete coverage of the centainment volume. , Igniter assemblies will be installed in these complementary locations before the first , refueling outage. Figures L-4 through L-12 contained in appendix L to this report show the igniter locations. There are a total of 4 twenty igniters in the lower compartment-, seven inside the crane wall at elevation 731.0' and thirteen outside of the crane wall in various rooms, compartments, and pipe chases. The upper compartment has three igniters which are suspended 35 feet from the top of containment. The thirteen additional locations are positioned five around the containment shell above the crane, four inside the crane wall just below the crane, one above each of the air return fans, and one on the exterior of each of the steam generator doghouses just above the operating floor. The ice condenser has nine igniter assemblies, five uniformly distributed below the ice condenser baskets and four also evenly distributed above the ice condenser but below the top deck blanket. A more detailed description is contained in appendix L. The igniter locations and quantities were chosen based en the following considerations. II-2
- 1. The IDIS is intended to provide ignition if the bulk concentration in the lower compartment, upper compartment, or ice condenser exceeds the lower flammability limits.
This is the basis for the preliminary analyses by TVA and the more recent CLASIX calculations. No credit has been taken for local ecmbustion in a smaller segment of these compartments. Therefore, igniters have been distributed in each major region in which hydrogen could be released, through which it could flow, or in which it could be concentrated to provide assurance that ignition occurs if the bulk. concentration reaches or approaches the lower flammability limit. At levels above this limit, one igniter is all that is needed to ignite a major compartment. 2 During accidents involving a degraded core, any hydrogen produced would be released in the lower compartment. For this reason, more igniters are located in the lower compartment where the initial flammable mixtures should form. An advantage to burning in the lower compartment is I the availability of the ice condenser to remove most of the
- heat, thus limiting the overall pressure increase.
! 3 However, since it cannot be shown that all of the released a hydrogen would burn in the lower compartment, several I igniters are located inside the lower inlet doors of the ice condenser below the ice bed.
- 4. Since the mixture may become more flammable as any steam condenscs in the ice bed, several igniters are located in the upper plenum and more will be just above the ice condenser in the upper compartment.
II-3
- 5. Igniters were located at or near the tops of each major subcompartment because of the tendency of hot, hydrogen-rich mixtures to rise. This allows early ignition at as low a concentration as possible and also minimizes equipment interference problems. When the height to width ratio was small, igniters were located at several places in order to assure adequate ignition as early as practicable.
I 6. Our preliminary judgment was that on the order of 20 to 30 igniters would be needed to provide adequate assurance that the objectives above would be met. Original placement resulted in 32 igniters which was in good agreement with the l l preliminary estimate. Since then, 13 additional locations have been selected to provide more visible coverage of the upper coc:partment vo.lume.
, 7. The lighting circuits were found to meet all our original requirements (power supply, number, compartments covered, spacing, height, and location within compartments). This is not surprising since these circuits are designed to provide coverage of compartments from above. Fixtures were found in l
all compartments where igniters originally were felt to be needed, although all fixtures were not used. The additional igniter assemblies will require new wiring since there are currently no light fixtures at those locations. They are capable of being supplied from the original IDIS circuits and power supply. t In phase 2 of TVA's program, we intend to study the detailed concentration profiles and the behavior of hydrogen near the combustible limit in order to take advantage of localized and II-4
~
i
partial ccmbustion in optimum locations. Such a study could suggest a change in the total number of igniters in the containment, their apportionment between ec=partments, and their location within compartments. Diversity of igniters, such as spark or flame types as well as thermal, may be desirable. Individual or area control of igniters may provide better hl utilization in some cases. A need for additional containment atmosphere sampling capability may be indicated to provide the basis for such localised igniter control. These improvements may provide additional margins of safety , but are not required in an interim system. D. Power Supply The igniter assemblies are supplied with 1207 ac from the standby lighting system. The standby lighting system, which reesives its - power through 480/2087 transformers, has alternate ac supplies from the 480v MOV shutdown boards. In case of loss of offsite power the shutdown boards are automatically energized by the diesel generators and therefore the IDIS is capable of operating in such an event. A description of the standby lighting system and the 120/14V ac transformer inside the igniter assembly box is contained in appendix L. E. General Requirements The igniter assemblies are seismically designed not to fall in the event of an earthquake and damage safety-related equipment inside containment. This is accomplished by attaching a steel cable to the igniter assembly and then anchoring the other end ! to a bolt embedded in a concrete wall er ceiling. This is I required of all components inside containment that were not II-5 L
individually seismically qualified by test or analysis. F. Igniter Tests TVA conducted a testirg program at Singleton Laboratories to obtain preliminary information about the performance of commercially available igniters. ' rom that testing prog"am it was determined that the GM AC model 7G glow plug could produce sufficiently high surface temperature (1800 F), could remain energized for prolonged periods of time (at least 148 hours), and could burn hydrogen in small sealed containers containing hydrogen concentrations of 12 percent and less. From these results TVA gained considerable assurance that commercially available igniters would be suitable for use in the IDIS. See appendix M for further details. However, TVA felt that additional testing was necessary to demonstrate that the igniter assembly will initiate a volumetric burn of hydrogen for various environmental conditions of pressure, temperature, and steam. TVA, therefore, contracted with Fenwal Laboratories of Ashland, Massachusetts, to perform hydrogen burn tests with the igniter assembly in an enclosed vessel. For a description of the igniter assembly testing see appendix N. G. Operation The IDIS is designed to be energized immediately following the start of an accident and remain energized until the unit reaches cold shutdown. In order for the operator to initiate the IDIS he must dispatch an Assistant Unit Operator to the standby lighting panel and switch en lighting circuits 10, - 11, and 12. The appropriate mcdifications to the Emergency Operating Instructions II-6
(E0I's) have been proposed. These modifications will not be incorporated until NRC approval for the system is obtained. Further details en the proposed modifications to the E0I's is contained in appendix R. H. Surveillance Testing Periodically the igniters will be subjected to surveillance i testing. Surveillance testing will consist of energizing the IDIS at the standby lighting cabinet and taking current readings of circuits 10, 11, and 12. These three circuits will have only igniter assemblies connected to their outlets; all other outlets on these circuits will have the light bulbs removed. During the post =cdification testing, once all the igniters are installed and operating, initial current readings of the three circuits will be taken. These current neadings will become the base data that will then be compared to the readings taken during the surveillance tests. The comparison of the two readings will indicate whether or not all the igniter assemblies are operational. If the readings do not compare favorably then all I the igniters will be checked visually. I. Qualification The initial testing TVA conducted at Singleton Laboratories (see appendix M) identified a commercially available igniter that TVA
- chose to use in the IDIS. However, that initial testing was not conducted in sufficient detail to provide complete confidence in the GM AC model 7G glow plug. TVA has therefore procured 300 of these glow plugs and will perform a statistical qualification progra'm on them to obtain a high level of confidence in their reliability. The qualification program will consist of selecting II-7
a statistical random sample according to Military Standard, MIL-STD-10SD, " Sampling Procedures and Tables for Inspection by Attributes," then testing the ability of this sample to withstand cycling, prolonged operation, and reponse time from ambient temperature to the minimum required surface temperature. TVA will select, from this lot of 300, the 13 additional plugs which will be installed in Sequoyah unit 1 as well as all plugs that might be installed in any future units. The 32 plugs presently installed in the igniter assemblies at Sequoyah unit 1-were procured earlier. Since they have been subjected to the same preconditioning test as the lot of 300 plugs, TVA does not plan to replace these plugs with plugs from the lot of 300. I f I I l II-8 l-
SECTION III EVENTS The igniter system has been designed to prevent the accumulation of explosive concentrations of hydrogen inside the containment following accidents exceeding the design basis. The system is not required i for any transients with severity less than er comparable to the design basis large break loss-of-coolant accident. However, it is currently planned to initiate ' operation of the system for all accident sequences with potential hydrogen generation since use of the glow plug igniter 4 system for events where they are not required will not result in unacceptable consequences. j The design basis large break 1pss-of-coolant accident is one such event that does not result in the production of excessive quantities of hydrogen. The proper operation of the emergency core' cooling systems act to prev.. .e fuel rod heatup beyond 2200 F with subsequent large cladding oxidatien rates. The redundant containment hydrogen recombiner system installed at Sequoyah has been designed with i sufficient capacity to adequately remove all hydrogen formation predicted by ECCS computer performance models in addition to that r]sulting from the assumption of conservative water radiolysis and cine / aluminum coating reaction rates. The igniters may provide some local recombination for events of this severity but will have only minimal overall impact due to the low hydrogen volumetric concentrations. Therefore, events of this and lesser magnitude have not been selected for the igniter design basis. Events where functional operability of the igniters may,be required to mitigate hydrogen accumulation are those that include.significant III-l l L
per4ods of inadequate core cooling. Production of such an event sequence requires the assumption of the failure of redundant core cooling syste=s in combination with primary system coolant loss or, as in the case of the Three Mile Island accident, operator errors which i terminate coolant injection improperly during a coolant loss event. i The Reactor Safety Study (WASH-1400) quantified the probability of degraded core events to determine which events were more likely to occur. In ahdition, mathe=atical models were created and applied to approximate the core thermal-hydraulics, containment response, and I radiation release consequences of these events. e From this study several events have been selected, based on probability of occurrence and hydrogen release potential, for the design of the igniter system (see volume I, section 5). A. Overview The event sequences selected have a low probability of occurrence but relatively severe consequences as predictea by conservative i physical models. Five accident s qu'nces were chosen to represent the spectrum of hydrogen pcoducing events from metal-water reactions. Each event results in inadequate core cooling and, if cooling is not restored, core melt. Fuel-cladding oxidation is excessive prior to-clad and fuel melt thereby resulting in the production of largo quantities of hydrogen which cannot be accommodated by the installed design basis recombiner systeu. Each event represents a significant challenge to any III-2 r
hydrogen removal scheme. For the purpose of this study, the alphameric designation of events presented in the Reactor Safety Study (WASH-1400) has been retained in tnis report. A description of these designations can be found in appendix B. The sequences selected are designated as follows:
- 1. AD - Large break LOCA with concurrent failure cf Omergency Core Cooling Injection.
- 2. 32D - Sc:all break LOCA (1/2 inch 6 D $ 2 inch) with concurrent failure of Emergency Core Cooling Injection.
3 32H - Small break 10CA (1/2 inch & D $ 2 inch) with failure of Emergency Core Cooling Recirculation.
- 4. TMLB' - Failure of the_Feedwater Delivery System (Power Conversion and Auxiliary Feedwater Systems) given Loss of Offsite AC Power with a failure to Recover either onsite or Offsite AC power within 3 hours as the initiating event.
- 5. TB9 - Transient Event of Loss of Main Feedwata followed i
l by failure of the DC Power Supply System. Each selected event sequence is characterized in the following section. A further condensation of scenarios was performed for i { the purpose of hydrogen production. The gas release rates for - these scenarios is described in section C. III-3
- 3. Characterization of Events AD - Large LOCA with concurrent failure of emergency core cooling injection.
This accident sequence uses the following event assumptions: A - Large LOCA (break area greater than that of a 6 inch diameter pipe) Z - Ice condenser functions properly B - Electric power available X - Air return fans operable C - Containment spray injection available D - Emergency coolant injection fails a E - Emergency coolant function lost due to D H - Emergency core recirculation lost due to D F - Containment spray recirculation operable G - Containment heat removal systems operable This sequence assumes the emergency cooling injection system would fail to adequately cool the core following the occurrence of a large LOCA. All containment engineered safety features would operate as cesigned. However, since failure of emergency core cooling would cause a core melt in approximately 0.5 hour after the occurrence of the LOCA, operation of the emergency core cooling system in the recirculation mode would not prevent continuation of the core melt. Note that the reactor protection III-4
1 system is not required to function early in this event since the rapid loss of moderator will force the core to become subcritical. The auxiliary feedwater system is also not required since energy removal by critical flow disch&rge of fluid at the break exceeds the energy input from decay heat. Unfortunately, without coolant injection, mass is depleted from the reactor coolant system until the core uncovers, heats, and begins to melt. Early restoration of emergency core cooling injection may i terminate this event short of complete core melt but may result in significant hydrogen generation as the fuel rods are quenched. S2 D - Small LOCA with concurrent failure of emergency core cooling injection. This accident has the following events postulated to occur: S2 - Small LOCA initiates the event Z - Ice condenser functicnal i l B - Electric power available K - Reactor protection system functional j X - Air return fans operable L - Auxiliary feedwater system available l C - Contairment spray injection available D - Emergency core injection fails E - Emergency coolant function lost due to D F - Containment spray recirculation available i G - Centainment heat removal systems available III-5 L _ -
In this sequence of events, the emergency core cooling system would fail to adequately inject water to the core following a small LOCA having a break equivalent diameter between about 1/2 inch and two inches. All containment engineered safety features would operate as designed to control containment pressure and leakage and would serve to remove radioactive iodine airborne in the containment. Since the emergency core cocling system was considered to fail in the inf t.ction mode, its operation in the
- recirculation mode would not prevent core meltdown.
The accident is characterized by a slower loss-of-coolant when I compared to the large break. This permits more time for restoration of coolant injection. Complete core melt and hydrogen generation occurs,over a longer period of time resulting in slower release rates due to the extended presence of water in l the core. The igniter system is anticipated to be more effective j for the small break scenarios where the hydrogen release rates l l are slower as exemplified by the Three Mile Island accident. S2 H - Failure of emergency core recirculation system given a small LOCA. i l This event proceeds through the following sequence: 32- Small LOCA initiates the event Z - Ice condenser functional 3 - Electric power available l III-o
K - Reactor protection system functional X - Air return fans operable L - Auxiliary feedwater system available C - Containment spray injection available D - Emergency core injection functions E - Emergency cooling functional H - Emergency core recirculation fails F - Containment spray recirculatien available G - Containment heat removal systems available In this sequence of events, the emergency coolant injection system functicns properly but in the transition to the recirculation mode, failure occurs. This event follows the same sequence as does the S 2 D case except for the above failure difference. Hence, core meltdown proceeds unobstructed during the reci-culation mode. The meltdown of the core is delayed even further in time than the S D2 case since inadequate core cooling [ does not occur until switchover to recirculation. In addition the decay heat at this time has decreased somewhat resulting in a slower core heatup. Time is available to attempt to correct the recirculation problem if external to the containment or to provide nonsafety grade water supplies to the reacter water l storage tank to reinitiate the injection mode. Containment design will limit the amount of additional water that can be i j injected. r TMLB' - Failure of the feedwater delivery systems (power conversion and auxiliary feedwater systems) given the i I ( III-7
initiating transient event of loss of offsite ac power. This event proceeds through the following sequences: TA - L ss f all a power K - Reactor protection system functional B2 - L ss f ac power occurs L - Auxiliary feedwater system fails M - Secondary side relief valves runtional P - Primary side relief valves open Q - Primary side relier valves close . Loss of all offsite ac power occurs such that neither the power conversion system nor the auxiliary feedwater system is available to remove heat from the RCS. In determining the probability of the sequence, it was considered that neither the offsite nor onsite emergency sources of ac power were recovered within about 3 hours. Since all ac power was lost and not recovered in sufficient time to prevent an excessive coolant loss through the RC3 safety and relief valves, a core melt occurs. Also, since all ac power sources were not recovered, containment engineered safety features could not operate to mitigate the radioactivity released fecm the melting core. l Failure of the ac power system also results in consequential failure of the coolant injection systems. Although no break occurs, the inability to remove decay heat from the primary system due to auxiliary feedwater failure results in primary - t ! III-8
system pressurization and pressurizer relief valve operation. The system slowly boils dry as the relief valves cycle open and closed. Core melting results when sufficient water no longer remains in the vessel to adequately cool the core. TB - Failure of the do power supply given the initiating event j of loss of main feodwater system. i This accident proceeds through the following events: T ~
""* *" 88 " ** "* # sys em a es de B ,
accident K - Reactor protection system functional B2 - AC power available D3 - DC power fails , L - Auxiliary feedwater system fails due to B j M - Secondary side relief valves functional I P - Primary side relier valves open Q - Primary side relier valves reclose D - Emergency coolant injection fails due to B 3 0 - Containment emergency safety features fail due to B 3 l l i Loss of the main feeowater system initiates a primary system transient (failure to remove primary side heat). This is-followed by a loss of de power. Loss of de power results in failures to the engineered safety features due to the deoendence i of the control systems of these features on de pm e. Hence, de power failures cause loss of ECI and the containment III-9
emergency safety features resulting in core melt and containment failure. The Three Mile Island accident sequence has also been examined in relation to WASH-1400 by the Kemeny Commission technical staff (reference III-1). It was found that the accident was roughly equivalent to the TP'QU' sequence. This is similar to other small breaks discussed above but with termination of high pressure safety injection unavailability before complete core damage resulted. TVA plans to study this accident in relation to the specifics of the Sequoyah design as discussed in appendix Q. The TMI accident would encompass the following events (reference III-1): T - Transient event , K - Scram availability P' - PORY or safety valve operation 0 -PORVorsafetyvalvestuckopen L - Secondary side cooling restored U - HPSI available U' - HPSI interrupted and PORV open for sufficient period to cause core damage U - HPSI restorec in time to avoid core damage C. Mass, Energy, and Gas Releases Each of the core melt sequences ( AD, S 2 D, S2 H, TMLB', TB, etc.) described above generates a significant quantity of hydrogen
~
III-10.
which is subsequently released to the centainment. Several of the cases are expected to yield similar hydrogen production rates perhaps merely offset in time. For exanple, the S D 2 and the S H 2 sequenci . , similar accidents, except the loss-of-core cooling does not occur in the S2 H case until the su=p water recirculation is required. This delays the onset of hydrogen production significantly in the accident. Similarly, the consequences of the TEB' and TB accident may be about the same. Therefore, due to the restricted amount of time available for this study and because of limited availability of existing data, the sequences considered have been limited to the AD, S D, and TES' accidents 2 as discussed in volume I, section 6.1, for which data was available. Other sequences may be analyzed inhouse, depending on the availability of the appropriate analytical tools as discussed in appendix Q sometime in the future. The three cases mentioned provide a wide range of mechanistic hydrogen release times ranging from very early, in the case of the AD sequence, to long term as in the case of the TEB' sequence. The cases also encompass several different release sechanisms. Both AD and 23 D accident are initiated by pipe ruptures in the reactor coolant pressure boundary. The hydrogen generated for these events should be lost through the break opening. The TEB5 accident does not postulate a break in the primary system. For this accident, mass will be lost through the pressurizer relief valves into the pressurizer relief tank. III-11
. . _ - _ -. ~ _. _ - - . . ._ _ . - - -
l 1 Hydrogen will enter the containment either through the rupture ! disk on the relief tank, or from the reactor vessel head vent line if it is used. The vent line is presently designed to discharge through a header in the same general area as the relief tank f l since it is in a relatively open area of the lower compartment and since it was already necessary to evaluate potential hydrogen releases from the relief tank. The header will be located just outside the reactor cavity wall at elevation 706' as shown in revised figure L-5 and is designed to vent at several points along its 16-foot length toward the crane wall. The three closest igniters are located about 25 feet above the header and, of these, the closest single one is about 20 feet away laterally. This arrangement should allow adequate mixing throughout the compartment during potential venting of the primary system. The I break sequences therefore represent unknown release points (all should be in the lower compartment); wherean, the release point for the TMLB' sequence is known. The release rates have been described in volume I, section 6.2 3.1. They are repeated here for convenience in figures III-1 through -9 The values were generated by Battelle Columbus i i l in an earlier study obtained by Westinghouse. They are specific i to Sequoyah. Each event may terminate prior to complete core I meltdown, provided adequate core cooling is restored. This is ! discussed in reference III-2. Several nonsafety systems may also be employed to cool the core after core melt to prevent rupture i of the containment. However cladding oxidation may be as high as 75 percent by this time. The igniters may be (22 l III-12 l t L
i l e r
................p....................,...................g...............;........ O l .. ......... ....&...
. ...... . g
- . .( . . . . . . . . . . . . . . . . .. . g o o o o . ~ ~ ~ ~ ~ t. o o . o . o o o n . o g o . ooo oooog.o.oooo o. . .
r l
. . . . . . . . . . . . . . . ..... ... ......... 6..... ................................ l...................-. . . . . . . . . . . . . . . . . . .
i . . .
.................;..................q..................4....................q.................4...................
......1............
.................4......... .
. . . . ..... t.- o . o o ~ ~ ~ o o 4.o o n o n . o o o o q.~o. . o o n o n o 4. o o n. o o .o . o. .
..... .... .......v........ .........p...................;...................p...................;...................
........q..................4...................q...................;...................
g . w l Q l ( . . . 6
. . . . . . . . . . . . . . . . . ....... .4,
. . . . . . . . . ........... . . . . ...........................1........................................t....................
...................t.......................... .. ............................... .
g .
..................g........... . . . . . . . , . ................... .
. . . . . . . . . . . . . . . . . . . .p........... . . . . . . . . ..................3....................
...................g.................. . . . . . . . . . . . . . . . . . . .i................... .
m
.....4.......................................4........................................ /
H
. . . .a
. . . . . . . . . . . . . . . . .p.............
.................... p..................... . . . . . . . . . . . . . . . . g.........o........ ....
. Og9 OMg
.................q...................
-4 :..................4.... ...................q...................
i i V 'g t, l d
. .m.4
.........;....................t..
............... .................... .t................... ga, Z
- W O i !
E.4 \ O !. :. i 1 - . en
?
?
Q,(
** ....... ........g...... ............
gd ....< ..............4....................................... ....................
.e$'O ........
........ Q ......... . . . . . . . . . . . . . . . . . . .
........ ......... y...................)...
............... .....................)..................
e
........ 9 a . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .... . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .
gg .
- . = i
. . . . i . o,,
, 1 OTI 0 0T O'S 0'9 0't O'8 0'O l
- p. (m90 ssm l III-13 1
e 2 -. O
...................i..... ............i...................;..................p...
.............i..................H
.................4...................9.................4...................9.on.oooooAot..oooo.oo...
...................6....................................6.................... ................
.................4................... ................4...................4.................4....................
.I
........4....................q..................4.................q................4...................,
i i . .
...... ... .....;...................(........................ ...... .... 6........ .......;................... .
. o . .
..........#.....4...... .........
4...
. ........... 4....................q...................;............
O
................g.................. ........ ........ .... ..................t..................>g 9...............
.. .. ..... .......t.................... .......... ........t..................... . . . . . . . . . . . . . . . . . ..... .. . . . . . . . . . . . . . . . .
.................4....................q............. . . .
4.. . . . . 11. . . . . q................-..4....................
. . . . . . . . . . . . . . . . . g...................
................ . . . p.. . .. Qm .......... ........ ..........t.................. .
. m 4................. :......d.........4.................;...................
- . t
(.
. 8
. : o : : deJ
...... . . . . . . . . . .p....................g.................. p......m...........g................p . o o oo o o o o..-
. O m:
. Os
! ! U
..................q...................g................b......~.8"........g!...................q!...................
1 .. . Cf2
%=* c; p
. . . :3
! ! ! .b
! f MN ~
. . . . . . . . .g...................t..... . . . . . . . . . . . . ...n.................t...................
O l l l ; H i b !.
. !.o !.
b l , U I. !. L ! -+. !. I "g 1 1
.........t.................... . . . . . . . . . . . . . . . . . .. ................... . . . . . . . . . . . . . . . . .
.m
. .... . .........;................:...................(...................;............ .
M4 y...................p...................p...................p................. O
........4..................... . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .
1 . e I . . t l l ........ . . . . . . . . . . l ............ ( . t . . . i 1 . . \ . . . . .
- : : : : e l
, , , i i i O y l 0000G 00938 0 0093 7098f 700GT M O'O (mW ssyn III.R
se
.................. , .< ~ O, t
...................t.................. . . . . . . . . . . . . . . . . . p...................
. . . . . . . . . . . . . . . . . . .p................ ..
...................e................... ...................q...................
..................:...................;...................:................@.3...................:....................
W* . l 4...................q...................t................. ~.................t................... ..
. 4 3
o f. : : H A. . g Q:
~
................g.....................g...................l...............%.:...................:.........
m .
- w. ..........
O . 1
. . 3 T .
g g O.
...............94.................g.................q.............g3....g...................t.................
. s .
s.
.s %
g .
. . . g .
...... .....p........ .
. ...g...................t.................g........%.........p...................
g
. g H : : : %:
- :.
- A
. : N g
s S cr> s OH s 9 QH t i : '
-O C, y *H
..........s........
...................i....................:....................&........ :
................L .. ..... .
-H v' T d
.................q....................:...................t.....................:................. t...........tq ...... W O . ...... . . . . . . . . .
....................p................... ........ ......... .............%
g
*e N
g .. . . N
........4.............................o...
! t b
-4 4......... ...................g. .
O : . 8 44 : : I
.........p...................&....................t.................&...............4 i '
1 ,
.I !* ! !
M .Oe A .........t...................:.o...............4...................:.................. M : : - 0i i
.........i...................i i
i i i i m6.............******i"*"""*""""*I"""*"""*""-
. 'd .
S. !. .
. d
.. . . ..... g o. ~ ~ ~ ~ ~ ~ o n g o.ooooooo og.o oo o o ~~~ ~ 4~ ~~.- o o o n.. .
25 :
- l8 O :
. 8 d g d g
. . . . . Oy 0JT O'OT O'S 01 GT 01r O'O
,0T. (tnq1) ssyg III-15
e O
..................i...................i...................i...................i...................i...................4
.............. 4
.....................................t...................4....................q..................4....................
.................4....................(..................4...................4...................4...................
.......4........ ..........g..................4....................q...................4....................
..... ............&......... .........p...................&....................p...................&..................
................4.......... ........g..................4..................4...................4...................
s* : 3
..................i.................... .O
.................1..................... ................. ....... ..........
.g
.................4..................g..................4..................,..................4...................
....................t.................... .
9 . . . . . . . . . . . . . . . . 9-o..-~~~.~..;-~~.""""""! H
"~~~"""".*"*"""""*""t"""""""""- .
M
. ... .. ................. ......... 4................... ... .
............4....................................... .
- : :. :. 1 -a H ..................
. . . Wt g t"""-"""""""""""""""t"""
OH
.! U Q
........-...........................'...................i...................:....................i.................... vU2 o i .
,e . . .
. 3' t
.......... . . . . . . . . . . . . . . . . . . p...................
. . . . . . . . . . . . . . . . . .p................... .
M rc , O i h O i. .
- a
? .
.g o
og M .......4.......................................................... : . . . . . . . . . . . . . . .
......... .............................i...................4.......................
. . . . . . . ..........g...................p..................;...................)..................
z8 9, e n . 4 . ca 2 . :
\ .
. =
. . . i ,
,g O
02T 0*0T 01 01 OT 02 0'0 l l pn (mq0 ssm III-16 ' -
e
.- O
................i...................i...................&..................j..................i...................)."""4
...4....................q..................4....................q................4....................
....................4........................................&................................. .....
.................4...................e4...................4....................q...................f....................
. . . . . g . o n o o.
. ..... ... ..... ... o . .o .. . .
4.... .
. . oo4.. on o.ono. o o 4..o .
o o 4. . ...o o.o..
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .p ..............................p..................4................... .
..................4................g............,.....4..............+....q..................4....................
. a .
4 :* 4 :
- : W 8
- . :. O :
., d 4 .
e
-O Q ..................j................... ...................g..................;... m .............q...................
.4 .
.g
...................t..................... . . . . . . . . . . . . . . . . .p......... .. ............. .. .. . . . . . .p.................. . .
.....g... + ..........
..................J....................!..................4.............
4.................... H
.. ....... .......g................ ..;................
9.................
............... ..p........ ......... m
. .................:.................4....................................t.............- 1 a N n
.p..................... . . . . . . . . . . . . . ..p.....................
...... . . . . . . . . . . . . . . . . .g...............
O'
. OH cH
................q................... ..................4................... ..................4................... D3 U0 d
- i :* ! !
F :. :. l : .
. o
. E.
O 3.
.;..................t....................;...................t.................... .
(,) i O i i i. e o
.H 2
4... ...................................4...................... . . . . . . . . . . . . . . . . . . .
.........r..................p.....................r......................................
i g,3 8 . . . . . . . . .
........ ......... . . . . . . . . . . . . . . . . . . . . 1 l
Q . / . . .
. g . . .
!* : . : : a l
. 4 4 6 i Og l
0'000C OTC2C 0'0093 0 096T O'00ST OV99 00 (m@ ssm III-3 7
n
* . C.4 v
).,; .
.p...................;
x.
..t.......................w..............
. L: .
p.................. .
- : : . Q :
..............o........................................t........-..
cn : :
~
. . + - . . . . . . . - . . .
t [ .
..........P.4....................g....................:...................g....b.
Lg .
. s
. . . g f,* : : : : : s
. . . . . . . . . . . . . . . . 4. . . . . . . . . . .. . . . . . . g . o n o . o . . u o m . 4 o o o o o ~~ ~ ~ ~ ~~ o o o o 4. . ~ ~ o .. ~ .N; o o .
d 4 . s s s
. . . s
...............p...................
..................6.................af-aaaaaaaa6.aa.".-
- - : : : : A
>-4 .
w e _4 .
- Or.
t O O,4 Mos
. v :3y i . . .
..... ...........g..... ........... .
w
..................g......................................q................... .
Q .. .... j M.
! I b O
Q .........p...................&...................p..................&................... . l Z Q o i
.-e hg s
.......4....................:..................4......................................
A . 8 (f) ng : : : :
. . . . . . . . . . . . . . . . . . , . . . . . . . . . ........6...................e...................................... .
w ,
. r 8
........ .........g..................4..................g.oo.oo.-ooo4oooooooo.. .
*z :
I 1 . . . \ . . . i . . . l l \ A O ! !. ! ! \ 1 . . . \ . . 1 ~
- : : : : e s
O y M N W W M M O'O
,m. (mqt) ssyn 114 %,,
l l I 1 l I o
.. .. ............,,. .... .... .......p......... .........g............. .
.....g.........
..........,,..... .... ..... ... s .O
...............4... ..............g..................4...................t........-.....*...4..*................
...................'...................;.......................................g...................:.....................
...............4....................g.... ... ......... 4.................. .g..................4....................
,..........4....................g..................4....................q..................4....................
................4..................;..................6.-.-.............a................... . -
1
................4........... ...g...................;...................q..................4....................
4 .
* . . 4
.O
,. ................... ................. .M
..................i......................................t.....................................1...................
....{................g..................4....................g....................q...........
.............t.......................................t........................................
.t.................. .
t , >-.-4 . .
. ................:................{.....................................4...................4.................. :.
. d.
3
...... ..... ..... p..................... .g..................... . . . . . . . . . . . . . . . . p...................
.... t QHH
. Q H C
.................q...................
- ..................4.................... : .
...................g..........
....... . 01 0 v4 1
5 ! to
* . -s
- : : A O
............................p......................................p................... -
.i i i i O !
i l
. . 4
. . . j e <
............................4.... .
......... .... e g ........
e4 es ........4....................:..................4....................:...................
....... g:n .........j..................j....................................!................... *
........ 8 . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . ............e.........e................
........ ......... ) ,
.t......................................t"."-~"-~~~.4..~."".--~~--
l j
........ ....... . . . . . . . . . . . . . . . . . . . ............e...... ...e............... . . . . . . . . . . . . . . . . . . . .
N . I l
. l 8
I
........e..........
8
....... ................e..
4 : 6 . l
, . . i ig C
0JT 0*0T 0% 09 07 0'3 0"O pr. (mq0 ssvn
\
TTT-10 I
us O
...................i...................c. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . , . . . . . .
g 4
.................4....................q..................4....................g..................4.................... .
....................)....................
.................4....................(..................;..................4...................4....................
i
........4....................g...................4...................q...................;...................
. .b .
. .A
..... ......... .......................................,...................p.. ...............,. .
(....................
- . : N ~ :
. . CQ .
............................. .......j................. 4...................q..n........... a .;................... .
!. !. !. !h
. O O !.
I t
. -o ..
..t.
~~4
~~.--~~ ;.-
-....--.--.g....-~~.~.~.-
.g
...................-~~~....o;-.-~~.........................-.
. . t...................
>=.-4 .................4..................9...................4...................4..................4....................
. . . . . . . . . . . . . . . . . . .g. ......... ......... . . . . . . . . . . . . . . . . . m
..t..................... . . . . . . . . . . . . . . . . .t...................
. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .......................................4....................
. c
-e 9..................... . . . . . . . . . . . . . . . . . . ..g.................... . . . . . . . . . . . . . . . . . . .
.t..................
O p
- ~ : :
O H
- : : O y 5 ..........
e :.
.......a................... ...................q................... ..................q..................
v W h ! ! ! ! ! M [g C . . . . . . . . .
.........f...................f..................h...................f...................
. k
. 9 O
I
. e O
. . . . . . . ......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .M M 2
>=H A o ............................................................................... . . . . . .
w4 o . .
........ O .......... .
M . . . . . . . . . .........!...................!g...................!...................! t >...................
.t .
O . . 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . A . .
; ; : e
.O
. . 6 . s p 0'006C O'093C O'0003 0006T O'00CT 0199 0'0 (m0 ssyn III-20
so Q (
.. .. ........42p................... . . . . . . . . . . . . . . . . . . p...................
- c. . .
.. .....................t................... .
l4 . r,3
................Ms. ................. . ...................e......
O: :. : ........................M....g.................. s . o.:
. Q:. .
.... . ......m................
m.
...g.......................................q......
%=
1..... *~ .......... . U
. g O .
t . l .
... . . ....g.... g .... . . . . ..........
. g .
U') . i
- 's s
)y ........ .......
4... ...............g....
.........g........ ..........g............ .. .A4.. ...... .........
s .
. . N.g e g
.e . g
- . s f . ; . %
. s
. . . . . . .. .. ..). .................
. . s
.- p................6-.*---*--.*.*-
*--.D..--s-......
s s g H .
- t -
l l
- l A g)
-J m s
l
. . O Hw OM O C) 7) U
..................................:...................:...................:...................:................... .g%e O .... ..... .......
W
-i
[ O . . . . . . .
.........i....................!...................!....................!,................. -
4......... . Ma g \ O M : : :
.........t...................:...................>...................:5................... - M : :.
O . w M A ........4...................:..................4...................................... H . W g A .... .
.no l
b d M i i O Q4 2; i 3O :
. . . . . . . gg2
*~'
.........g..................4...................g:..................4...................
2: " l l l
. g
. g
. . .T 4 i i i i C y
O*3T 0*0T O'S 01B 01 0'3 0*0
,0T. (mq1) ssyg III-21
effective for metal-water reactions as high as 75 percent. C? 2. Release points, as mentioned above, will not be known for the break events. For large breaks, the release may occur in either a hot leg or a cold leg. The hot leg break has the potential for releasing hydrogen near a steam generator dog house where it may collect. However, the temperature at which the hydro 6en is released will probably be in the spontaneous ignition regime for this break. Small breaks, as typified by the S D 2sequence, may occur in a wide range of locations, including the letdown line and pressurizer attached piping. Hydrogen from these sources may accumulate in the pressurizer. dog house, steam generator dog houses, or may be discharged directly into the lower compartment volume. Each of the dog house areas does not contain an igniter. Ignition in such a confined space may not be advisable. Instead, a hydrogen mixing system inlet is provided to circulate hydrogen accumulated in these areas into the air return fans where it eventually becomes mixed in lower compartment air. l l D. Basic Assumptions QL4 The three hydrogen release accidents AD, S D, and TMLB (up h 2 to the point of core melt) were the ones initially considered for the purpose of the igniter system design. However, according to conservative MARCH analyses, the AD sequence generates hydrogen I at core melt at a rate that would probably exceed the useful III-22
4 applicability of the ignition system, although a detailed, realistic analysis has not been performed. Similarly, the MARCH results for the TMLB' sequence show a hydrogen spike late in the accident. These accidents result from the gross failure of emergency core cooling systems either directly or by consequential failure as in the case of loss of ac power. Probabilities of the sequences are extremely low as discussed in Volume 1, section 5. Studies to determine accident sequence probabilities showed that a small LOCA followed by failure of emergency core cooling injection (S2 D) is one of the more probable serious degraded core accidents that might occur. As indicated in WASH 1400, a small LOCA (S } 2 is an order of magnitude more probable than a large LOCA (A). Preliminary calculations show that the sequence AD is a factor of 20 less probable than S D. 2 In addition, due to the good availability of offsite ac power, the TMLB sequence is considered to be an improbable occurrence at Sequoyah. Therefore, because of the relative probabilities of these accident sequences, the fact that S es generate sign m cant 2 quantities of hydrogen prior to core melt, and its similarity to the TMI accident, "erk thus far has concentrated on the response I of the containment to use of the igniters in the S D sequence. 2 Ql4 E. Calculational Methods Calculational techniques used to generate the gas and energy III-23
i releases mainly consist of the MARCH computer code. This tool is discussed in Volume I, section 6.2.2, and Volume I, Appendix C. The studies which defined the probabilities of each sequence 1 and resulted in the selection of the sequence for further study are detailed in Volume I, section 5, and Volume I, Appendix B. t I i 4 l ( I l r III-2'4 l n ._ __
i 4 {
References:
III-1 " Technical Staff Analysis Report on WASH-1400-Reactor Safety Study to President's Commission on the Accident at Three Mile Island," Kemeny Commission, October 31, 1979. ! III-1 " Mitigation of Small-Break LOCA's in Pressurized Water h> stor l System," NSAC-2, March 1980. 4 l 5 l-t 1 I l
~
/ t-i t i i I - III-25 \. [ -- L
SECTION IV DEFLAGRATION AND DETONATION A. Overview The behavior of the hydrogen combination process with oxygen is a complex phenomena affected by many variables. At hydrogen concentrations in air as low as 4-volume percent, burning can be sustained although burning efficiency will be low. Flame propagation will not be preferential in direction above about 9 percent by volume and burning will become essentially cc=plete. j Above 18-percent detonation can occur. Several variables
,, influence the behavior of each of these burn zones. In the course of a degraded core accident within a small volume containment, each of these burn regimes is possible; and many of the influences affecting them are found. If the burning can be l controlled within the lower concentration regimes, the resulting containment pressures and temperatures will be relatively mild and manageable.
l The success of a centrolled ignition system in preventing excessive containment environmental conditions is greatly l l dependent on the many variables which affect the burn characteristics. These include concentration effects, heat i f transfer effects caused by mixture composition, temperature and l pressure of the unburned mixture, presence of steam, and flame propagation. Each of these variables will be examined for the l i IV-1 [ -
influences exerted on the combustion behavior of hydrogen and therefore on the design of the igniter system. B. Concentration Effects Figure IV-1 (reference IV-1) shows the flammability concentration limits for mixtures of steam, air, and hydrogen. At a concentration of approximately 4-percent hydrogen by volume with approximately 95% air, combustion can be sustained but inefficiently. Apparently, below 4-volume percent the effects of heat loss away from the reacting gases is sufficient to quench the reaction. This lower concentration limit is not significantly affected by replacement of air with steam until the O concentration limit of five percent is approached. The limit is, however, sensitive-to density of the gases in the presence of steam. As indicated by the dashed line curve in Figure IV-1, an increase in the volume percentage of hydrogen is required in order to achieve flammability. For example, at 40 percent steam the volume percentage of hydrogen required has shifted from about five percent to about 12 percer.t for mixtures of constant pressure but different temperatures (75 F and 300 F, respectively, in this example) or, i.e., different gas densities, j The presence of steam is very important to the operation of an igniter from the concentration standpoint. It is interesting. to note that at concentrations of steam above about 35. percent, no detonation can occur and above 58 percent, hydrogen in the IV-2 i
a ( Flammobility Limits 100 0 l i 75 F-0 psi 9 10% reaction container
---O--- 300 F - O psig g ambient temperature 216 F
--&--- 300 F -100 psig 80 k '
20 10% recction container 100% reaction - c" cm n 216 F /
/ '
cmbient 280 F 100 */. reccCon container m ent 280 F 60 / ./ , '.' 40
\
l 8 ff Assumed ' \
- Os g
p' ! ! ! 40
/ t detonotion limits
,[
I 60 o-
/ /\ e -
f Flommcbility limits f \
- //// \
c~ . . . -
/ \,
20 [__,~.~.~ \ 80 7 l
/
00 80 60 40 20 O Hydrogen, v/o Figure IV Fla= ability and Detonation Li=its of Hydrogen - Air-Stes: Mixtures I i rI-3
mixture cannot burn. Hence, very different behaviors may be observed for different types of igniters in the various compartments of containment and for different breaks. An example of the problems which the above =ay suggest is a hypothetical mixture with a steam content arounc 'S percent and a hydrogen content around 10 percent. This would be just outside the flammability limit and a spark type igniter would not ignite such a mixture; hcwever, a glow plug may sufficiently heat up the surrounding mixture such that the local steam concentration is lowered and ignition will occur. This could propagate by virtue of preheating the surrounding mixture and changing. the concentration just ahead of the flame front. Another possibility, however, is that the hydrogen concentration in the vicinity of a glow plug =ay diffuse away from the plug sufficiently fast to render a marginally flammable mixture inflammable. Fortunately, small break studies utilizing the S D sequence and 2 specific to ice condenser plants suggest that the. concentration of steam during the long term when hydrogen could be expected to be present is sufficiently low to have little effect on the flam= ability limits and, therefore, the behavior of the igniters should not be significantly affected. The studies were performed by 'destinghouse using the LOTIC computer code. Centainment conditions were determined up to the time that significant hydrogen release is predicted to occur. The resultant steam ccacentration transient is shown in figure IV-1a as a function of 1 IV-4
l h Q
@@< C d14 we.
51 3
- _ _ . . _ . .- 1:. . ._. : .:. . . . . .
gg . -. .- . _ . . . _ . .. p hg g. - _ _ .
, .p. ....=_....:-__
i .. %. .._.4._.
- y. ._ f .L1_.__. ._ : _ .i.J ;._. =. _... L.... - _ . . . .. ._. . .
i 0
, g- ,
. . . !.g.
- _. . .. 9 l- - - 2 % ~~ 2 i - ~7 -
- ___ ___ _;
j_..
. e _ _1g
- - - _- -= ' _- -_.- - - " :-
, f '
==u2 .__ .- :- - - 3.. *-----*-:
g ' -- -- I , . . . u..____.. . .
. . ~ . _. . . _ .. = .., .... ., ~. _ .
g g . . . . , _ . _ . . .. . .
.. _..._7-._.._.i.
. _*E ::_1 . . . . .J ..._ .. ... . ..
- i. e I :: _:; m - . _ n: -- - : =. . . . . _. =L ;.1:a -
.:.: :;.- 5 _ . x;= . ._ _.; =_- -.__- 7 a . = , . - _ _ .__ r;__. . yt g ,g -
I
- - - ---4 not p.a s s.kle._.2__ ;=_ _ =.. ::
. y. ._ ; . y: .. . .. ~. w . . . ._. . . . ,;;= . . . .. ..
. _. a ..=._.:....
3 __. - .
- 1 m =: r. . __. 6.n= n r: ;-- r-- .rr 1 - =.r. L. . ; ; ;;;_. - -*
.:.=p . :; =.;.
L w a .- -.. : :; _ =.-
%~ : . :- : ;i _.
. . ~ . 7a ..-
__.__.._.:.n....g_.. . . . . . . ._-
' I '
S m t. _ _ _. g 5 _ . )
^
__ f :n : Q l j 1 , I . "- j . , l :- , "i
- _: u . _ . --"=__:=
- : :=. z - -
u:._:. . l . .. .: . . : _ ' . : ==- :.. : __^ : L. : - ' :.. . .. : - --..:------ -:=
'?
lg - -- * ==: = :_ : . = _.._ :. = . :.:-_:_ n -' . . . . - ' - = = - z.== [~~-}{~5*' t 32- a _ ._ __;. 5 : : _ _. :: --_=
~ -
3 :
- = = := == ; - - .----
l i __:: . . .. . . . . :..: :_ :a.. ..._ . _ _ . _ . ' - . . . . .
- g. t , __a : ;_ x ,
q (y . g-' 1.. I. s b'. d :,', :t- .. l : .; . y
.;'I I, . 1 3 g. ;4 2 :c.: . =. -. ._ p - - - .h y .a : .' ! . ! ... j .. . .).
g , .
^js x 2 .
.:.:. .. f . - p ....=-;g..
N: 4. I = -- M .r __ _ - m . == . . n i _. i g _. . . _ . , . . _ . __.._ - .___. __ I.'li- -j--Th;__, . . . ._ -_ u:==.: :::_ ------L~~._'. *~~ X ). 2 1:E:. 3 .:- ; 1 y e,.7; . --.;__.~.,.._._______3._.;. . _ _ _ . . . _ . _ . . _ .__g..._;_.., _ , l % i -n : .-~ v _.. . . _ - . .. . . . _ . . _ _. r .:~ . :: _.* "
- = : = :-- -:..l- .=i u 25 - r- -w .
_21-
- . m:
p #
.... ' .,:g-9, 14 k..,'.. !
I T 'N'.-%- . , l --~
.a.. A...t; _; __,g, =.
..__._...J.-._-._.:._.:_. r. ._,._..._n_.,.-,
V) g . . . _ 3 _.y i
.- ~ - ' I ^::_ .= _. . . -. _._ .
_ _ _ . _.q_. _ . 3_ . - _=_ z
.--~n.__.._.._.---
-.___.__-3;..L _.. _. _- u t u ? - . --* r;_ - i - - --;
A 1 g% - 2.- 3.n...:n::._. y;; __ . _ __- _:_ t _. ._.__ __- . n_ __:=. . . . _: =_ :. .;7 . _.: =___ . __ = =1 = - .
.....q_-.....-...._ ! -.. ----._.____.._.g'... ._,
y ;=.;, g l
,. . ~c = :-t _. .. .:_=..r_ _-.__-n . _ t . r__=_ -".._:7.;.._'.
. =_ _.;.24.:_._:....
P._=. ._; ;;..! =.. . . . _..m.. l ...:].: .~C - N. b .
.f' ..[ . : b..M[Y --
d - I . ...._...._.___.J__._...
..:M - -:2:.::
.:;. .u n_ : n:
i + :::; g . ,
- " - - - c;L - '-- J J ::=:r ~- - - -: .= 12: :
_j _ . .._._ . g .. y -
. n- r. z - .;-
,n.
t > , ;..
,; , I , -- -
- ..- w;-- :--- -- --- r _ ; ;_ ;t ; -
- : ;; ; _ ; .. -.g.-- . ['
~, l ~.. . J , . m;I . __..=P- --*-",--- -~ ;_ _ ._ _ .. . ?.- ~" ~~~- =-,
. b nr
;-.---.i-.-.._.
~ -
Z !. q . , :. c._ _ ; .==p: - ' " r. - . :-- =# 2;-- .___ . ____
..__j. - -
.: = . :q
.I. g.
.qr. 7 y, p......_.i.. . -l__--_-.
l : l t . j' i m-----; i s -l ,- . z - . 9 - .: - = . .. - - - - .4. - - - - - - . . . = =.:---g c - I . CdM/MT/IEW
; i . .i .
.I .
i . , . . . i
..;_-.y
.l d' " "i . '7t.=._. :_: q. :_;- . . . ^ U: u.;
,} .- .
i . . , , s , , , , , . . . . . . . O i' ~. I000 E Qi"if:."2eoM;i ih C -- joeM:iin.25-yo e,C.E ..hg eee 39
- =. :. .,_ . __.:._ ......:._ ___._.... _. ; , . . .
. l-. - ~...
l . . ...._._m. t
-- ' . . . ' ^ _ . . . . . . _ .2._r_._.j . . . . _ . . :. ._.: ' . ~ .:.n. . = - / : =. _.T* NE=-(.s/dbng-f ~'~.... 'L.. _l r-: r...==.. - .:.:. c _= . - - - - - - _ . . . . . ;:-- m :.: -
. ___ . u=_==.=r : -;. - - -- ~
....- u1= u .; .
- m:-- :- - = - - - :- - ---
a
- 1 = 2n . ::: :_..=. .: . :._ u . _. : ,_
m _ : _: =J.: 1 _.m 'Ti . ; _gure .m . . .IV-lf .-
- 5. .
- = --
== - (
~
. . _ . 6. : = :_ n :r:- . : .. . ,c
. ?:- - -"
x= - - .L.. . _. :., _
.r. =: : = :.- A.-..... --- -
IV-5 . , _ _ _ _ _---_- = _-_ _
i time for both the upper and lower compartments. Ihe upper compartment remains free of steam throughout the first 5000 seconds and igniter performance in that volume would not be impaired. In addition, the steam concentration in the lower compartment peaks early and drops 'to values well below the 60 percent value required to inhibit burning. MARCH results (figure III-5) demonstrate that most hydrcgen is not released in the S D scenario until 4000 seconds at which point the steam , concentration in the lower co=partment is below 30 percent (approxi=ately half the amount required to suppress i$nition). Figure III-4 indicates the steam production resulting from core boiloff during this period is decreasing (2,000 - 10,000 seconds) l l and does not increase again until vessel melt at approximately 10,000 seconds. Extrapolation of the decreasing steam concentration from LOTIC should therefore be reasonable until the time of vessel melt. This indicates that during the hydrogen generation phase of the transient, the steam concentration is low
- enough to preclude adverse impact on igniter performance for the lower as well as the upper compartment.
f p l Concentration effects caused by steam are a potential concern QS l k -
- only for the upper region of the ice condenser where the removal l
j of, steam can cause an enrichment of hydrogen. A hypothetical mixture consisting of 10-percent hydrogen, 45-percent air, and 45-l percent steam entering the ice condenser could discharge into the f l ! upper plenum near the detonable limit. However, the Westinghouse , s=all break analyses just described showed that lower compartment steam concentrations were relatively low during the period of IV-6 i
probable hydrogen release. In addition, review of the CLASIX results to date indicates that even using the conservative ignition condition of 10 volume percent .n the lower compartment, the hydrogen content in the noncendensible gases does not exceed 15 volume percent. These both indicate that the potential for significantly increasing the hydrogen concentration due to steam stripping in the ice cendenser is relatively low. In addition, air is always present in the upper plenum which will dilute this flow and make controlled burning possible since four igniters are present. f QS C. Impact of Water Droplets A number of mechanisms are known by which water droplets dispersed within a combustible mixture of hydrogen and air might act to limit combustion and/or detonation. These mechanisms include: (1) Providing a large mass of dispersed water droplets per unit volume of gas mixture is an effective method of dramatically raising the specific heat of the mixture, decrease heat i diffusion distances and times, and increase surface heat transfer area, thereby limiting temperatures and pressures should combustion occur; (2) The presence of supersaturated water vapor inhibits chain bonding in the hydrogen-oxygen reaction which may preclude detonation by slowing the kinetics of the reaction; and, i IV-7 l
(3) Many small droplets of water dispersed within a gas provide an effective particulate damping of steep-fronted pressure waves by viscous shear action, whereby the extra surface energy of the smaller droplets comes from the wave. These smaller droplets are then more effective as a heat sink in the ensuing ficw behind the pressure combustion wave. Since, at the present time, it is not possible to predict the combined effect of these suppression mechanisms acting together, various experimental programs have been, or are in the process of being performed. These experimental findings are the subject of this section. i Different modes of combustion ha've been observed over various ranges of hydrogen concentrations. An understanding of these modes is important from the standpoint of mitigative concepts designed to preclude or control hydrogen deflagrations. For J instance, it has been established that flame propagation is erratic with incomplete burning over the range from 4 to 9 volume percent H in air. I this concentration range, lesa 2 than 50 percent of the nydrogen burned. At higher i j concentrations, 20 to 24 volume percent H2 *i" *i*t"#**' detonation waves were produced but not sustained. Detonation was e well established in 28 volume percent mixtures.
. IV-8 L
In dry air, hydrogen exhibits the following characteristics in addition to those stated above: (1) A well-established flame would not propagate in 5 volume percent H
- 2 (2) An igniter which sparks 60 times per second across a 0.050
~
inch gap would not ignite air mixtures up to 9.3 volume percent H2 ; it would ignite 12 percent H 2mixture. (3) A flame will not propagate downward in less than 9 volume percent H mixture. 2 (4) Ignition of mixtures contain,ing 12 percent H2, laminar burning occurs with a flame speed of approximately 5 fps, pressures increase by a factor of 2 to 2 5 (5) Turbulent flaming occurs in mixtures containing 16 percent H , with r without water spray. 2 These dry air characteristics contrast with the following water spray results: i (1) A well-established flame would not propagate at less than 7 percent H ' 2 (2) Ignition of. mixtures containing 12 percent H2 ' **t*# "E#"I agitates the mix and may increase the flame speed up to ' approximately 50 rps. The spray also absorbs much of the heat energy and the pressure rise is limited to a factor r of 1.8. (3) Water spray was effective in quenching detonation waves. i l IV i (
In summary, water spray has been shown to be an effective mixing mechanism which suppresses the temperature and pressure rises caused by combustion or detonation. It is also effective in cooling and suppressing ignition. The containment sprays should therefore be useful to prevent detonations in the upper compartment during their operation. In addition, the sprays will serve to suppress a large temperature spike by absorbing considerable amounts of burn energy. It should be noted that 12 percent H mixtures were able to burn in the spray atmosphere; therefore, it is not expected that detonable concentrations of hydrogen can acpumui2*.e in the upper compart=ent prior to ignition by the igniter system. The igniter design uses a spray shield to prevent the direct spray of water onto the glow plug unit to ensure proper operating temperature; thus, no problems are anticipated from the concurrent use of the containment spray system and igniter system. For details of ingiter performance during Fenwal spray tests, see Appendix N. D. Fla=e Temperature Flame temperature is one of the most important factors that characterize and influence combustion behavior. Flame temperature refers to flames burning at constant pressure with no appreciable external _ heat losses or gains. The flame temperature is a function of the mixture composition, initial mixture temperature, and pressure as shown in figure IV-2 (reference IV-7). The maximum flame temperature is obtained with a slighuly rich mixture (approximately 31 percent hydrogen in air). These values drop off regularly on both sides of this IV-10
. _. _ . _ . _ ~ . ~-~ - _ - . _ - . - . . . - - , . -..
l f I i 1 4 I i ] h::= =-: .= -E 4.-ROIN==I'ilifd- Ni iPiisitiEll'E.EWW lit? 7 u- ~~i# EE-INIEWh:~'"4 J1IW!i"'ri:hMM#i- :iONM ME bbO'4':F=5f=l55f
~
d idis! =fi2W:LiL=?5?- : ==tUk U-- "?. - 4is dii:12Gsidi k:9 91 lit' & 'i?bilif4T: Mii.E=RM
iiid! Wskh-4:1-H 6 =4 kT/LjhMW pitiS .iS_iLippi ff;l iL4 @j m/; yk,y .&p3d Presswet, 'ht": N@U :?iF W Ei lYi+ ::#n' -liikE V Esn@3iih3l
,i{lpigiF 5Mi % ? dMdiI Dil'D '8=idlih b" I I Wl 'bl> MPm t
" 4 Oj:N @N Ui8h=$U4
@ ci'ji?i d!)i! %Wh! 3:ilW -
j yii :1 USiCdifE:litblii;.2.:iilHblHSC fJimi a\if '%N'?Hii? fAp. h:@C 4.t4 @NL W i' i}p tait %i: = i[ii.Siihp/r/ tt j!igN:;ht thNudm p n;W ;r,p.s :fr* : ::st: rm:artt;:: %d. .iLSE .Hjhp.di]%f
@pa; nhia:If;n/, ff.t r.4 Win.4,= bQ yiiii;_ ; ygLi;ji 3: hii-tr liidd:: 7 Eidi ';M6n flibiilhi:ii
.!!$J$ r::. !
2W @Me M *PW#iWs* WMNigh4N W FMM "" lim"iN "iW TWMd'"i#" di@iiWiM dfl j },@ tyddi ::iika: iiivfh lif :/. :ftKshiip qh.i.D 'MQukW tnip.: i!:V'thiF!:H;il; ir1! !p %!b bdishi !!4U
; G:hi E-k d f=%-d/dK'/4/ . dii:.iiM;i2TA ILME.lW-W'W l =i398-Wt' Teb40'@:fii %ld
- L: fivid 5:44L f4#f/ 8tr i/[b8 I . 'OO itgagi mg wi. i! ,
% :qI-sci M ,s z_tir=.p-id dag sii:s=, g
,1 ; !4: m;pir. .Eini.9h .6. p
{m.h M. 'li',O ,d:;W:j tii . :@y:.N'itO6j@ :Pl;w @p4 :.44rir i2!Er. m dr 4Niij: :+k/ i t.4 f 4.' , W' NhWQ'if #M M his 84tb n W ii C M!J Mfit PdtT r.!!h 8 n!P .ibi!+'/i' .1:/ if itin,N Wm si!/4 a] ner
'O "
%dillDWEM -@F iiilMNtW:i@r M ROIyD7 tile !U.N=j+.1 Eil.;3Wi:MHi; M "4! i tis "MB j fi Wbar'if og i gin It.9:W i
2200 34p;43,g;fgpgqq jgg 3;3; p g. ,gy,3=g g;q=g ;gggggy gmg_ g a gg; -
! intW.ii E!/WfWf/F/M iWNWNIW" Wir@WitWa= #=T+iif5ifi!Esi SE'.si diis NM:dit-E!=i j';gW/-!/ e: :@ us%WiWNd!5:2 N-Nlir=naidW Md a '- Ei l G
.. y~'
;swrd=0 eM= M /MNd=vMi'cM93!W :JNi- Vlaer Yii.' W-HE:5 Iia *EiriE 2 i ..
9l'#i E' dEJ//I T ill i f/1/ihliS45-9%iEs@'!!EIMIM 5 P=.Wil'MhhhM-OM=i" inst.oi Wil==[:Li## lisit i f/*lli iiiiiii:i i5dEEIO:="t-Ni 45SM:f$"M 'dMYR MEi5:N.q maNre r# eb@iiiW @if h-U4iusu5dFJ=ikiWiE st% 4W i@ebE*isW: ia+F=W "E *"!'""' I i ##id:iEJPMi;id=#iibWi =reti3 FO M5W W44+%5imsClM d '# *O n* M M e ic/.:==f'Ei U JW M i Fli
#El/rrfridi W I!e M J= h eFiriejayg m etw
+1= hil4iih=i#M=&Md T=MNM %" + dab?Eb+=E'Wiii=CeiW",g= =' i myia:-ity:NiMMNTk@W=P
~i. sLiiWMFhMIINFf N=7= 2hEMt'i sNica- M m-r iM-t rgs=tm e5 +--
s $fidadfiE5il'9[l'if @i Wi! =-55i=E ! JM dii 'IFE5055-@55Y-95 @LNJSC-i!? II"f=E =3 ..mM . I :=-l.5!! Id-d=Ef5hp.13Gg lHig ,'=, i,9
~
7E:,MM iii MgiF[3h.==~EiM@j y gd [ L I k im=41=Ei:+M!"/-F1 % W4EE :-Ri! in ! ale ENm 3 =EE W i34E-! i . := AE.= == . 2 Widi-lM TW iiM=-N E!*MnH!N! Min Wa h "Fi!WTM E!Ermit:di'm@ hi l E b MMM"W r- C.=~-Wooc i NPi /G=i: 1Ff/ -:1 ?.imisinMW3!!PJWitihPfi ir ' 1 2 ,ss!!= #.TL M#ati: i. #. cMutiisnIUP 7 Oii-W EN Je4;iWiMii= inAi: - ! I* is lirpi-51 =E(=! :i:Pri 30 M!isiR tt7iPW.MiiW 'il aW941%Fi!!aWierffNW515is!U . - 9:if/E/ if ="140W. l d=: G T- 5imE ii 6.- # d:WiNii15-.N 5#E':!ss!: lisM ia E' l'Mf5 lEf5'=$5hUl4 fi-i3F sf 3i5505-hEI EN =:Wt '!W IE 855$ hE.i' ': Y T!!"i I!i-iil@! l:i- , . , h T s'di-3:/if li'45/ ii 15ru'EP' 'i i i fs#TI-5.5il 5 J.'. #TE 4!Ii-5:.J.!* =~ V E@!d35 ' ' . MIDI 55!d' N$ , & /ipmaiW Meili.F l Pit 4":hfitsiM Jiui':Et'r f:5 iEsiisrm -iMa :=!k V rdM".N - du - f.@sf.ed -iIli#.Prit;ai ' '
!3951siinii:4E!ivisENi afiJiiisn':'! iNfatA Sit.it#NE' jib Xfoodij; N #ilini#iM!:P_r"is AsiffM=W#ist iN-=t * " ' i i MEfWIi="his *TI:9" i :
sliEl3:.Si Of!!O 63.'diliI L , . ... NEl= '"'-%= +iii i :~e59i - 4'sWiE !!hli" WAN W.U.rh i :sp.ar ny,. d \:=i,Iir=j.W.:s:y:}( i itiiTJ i;(- @f J @vHEli.i! E.- lIs. C:5
' " ~
. 6' , ;. ;. ;,.i; ' !- qsi
! id tib%.'. fjiyh%f '
! fEi p , ,,n ljE "l: j.;i;ait;g; jiii yQNfW 3;V .ti;j M i l
a ? D ! P iais~= '"9
- Ei J:E i i!Ei Mi! EB siifViie =6gg;d" .
i g;laj p.d =- ----t48p is.. LghiaI a . iriitil . isnz-a : d'%i. ;
' =;l.Q'i((y k:p.! gic M, i5il.f I.-2 A +siii8% M 7f5IIid1iI! E .' f*
iI.MI! : .*.2 . - rEi i ff1 d TJi35.@Eii:"li d'!'iii! -Ii45' i . r SSii!E =2 ~
, . .~ - l
@E5i .i dE: E i'E4 :30 iMI g T 5Ll3 $EIE/r@ EEE uMS:= -
. 5355 96Iinit$i ilUI=EI iC INN'i!"i5i-~1!' r }isliidW3I!i
@.$[ WFi+sih:MttiisiEi$%EMd:i ifi i hiliiin:_:sili=;-4.W: MEl li8!Prii ==iW=EWEnNd P i
!l - =rN !Iti3flii-iilff:iiiE l Bjd5 tut.2.r=5i siii i$$" iE=E i!!:I=J =NJ: !!@fiEiGiEif EVEN!'O af idissMiifdi S!ni iiiire 4.Wi iiis =: n '==.: WgiWM h .it : ti. r &6sa r't m i'ffiit I9.Iiin4)5- I!ME; E 5-@ filii lE -!' iZ.* ?.'*".' i liii I.~:ihi '.iill.'5Kif f:Kis f g '~- -
,'^ p ;)=-i ip e.iiif:in: i s '=j :
Mg;ijg :
- .IEF4::.=sans .= *ii g6L;ii'I: -Elj
~
H:';-M , .ff=J- - _ i'iE iiid"
'Jb- -- e.M i! Efr' lii'! =7 !E =I!?' : C 9 f.Dli
',ri!!si j- w - -
- ! , =n .
p!ig ._ q ;. Rh: m- : .- .. :a. . .- age- :47_
~
. . , 4TE 2.1. i"' P'iil =Ci.E ru iU l'iiMS=Lbr $=Le3 .
5 fbi - Ui h .
.i$j!EiMiii -! lh5 :M.E5E -!ii!N .
xE=!E n'J+ 1t=NiN il d= sW "iis w -
*#WWiii 2=e : 9+.ti eigs=piid M
.M @ ".viisc 'ptb !!dian n ilE :1 ) as%'t.f~ '.7 ':
if
@d J ;!+: 1(
j f
~
5 j "U=;;ttglA 3 ii= stir. ;i;;pw Uiii - M=ihijp= M=~l.iii -=N I=i:Mi =T3s$$i e@= 4.$Mhi:j@ ;h;it! L#i a:.6 p =;d; i 8i Lsi
- i-1:, -e no G3;,i'i
-O so 20 30 40 SO . 6C 70 80 90 roo Mydrogen in air, persen' by volume f
- j. Figure IV-2
! 0*"O *]O 3' K M" i . oo oJu 1 AIAL: f IV-11
maximum as the flammability limits are approached. In.tial mixing temperature has the effect, that except for mixtures near stoichiometric, the flame temperature increases almost linearly with initial temperature. In very rich or lean mixtures, where flame temperatures are low and the.e is little dissociation, flame temperature increases degree for degree with mixture temperature. As the composition approaches stoichiometric, however, dissociation beccmes mere important and flame temperature beccmes less dependent on initial mixture temperature. The effect of pressure is such that dissociation of the burned gas is favored by reduced pressures, so that flame t,emperature decreases as pressure is decreased. However, the size of the effect depends strongly on the general level of flame temperatures produced by a given mixture. Near-stoichiometric mixtures show a strong dependence of flame temperature en pressure, while lean and rich mixtures have little or no dependence. Mixtures that are quite lean or rich have flame temperatures too low to cause must dissociation, thus, pressure has little effect. E. Flammability Limits The flammability li=its vary, depending on whether they are measured for upward- or downward-propagating flames, because convection assists flames traveling upward. The rich limit of hydrogen is the same for both directions of flame travel, 74 percent by volume in air. The. lean limit is affected, but not in the usual way. It is 9.0 percent for downward propagation; however, for upward propagation, there are two lean limits. One is called the limit of ccherent flames; it is 9 IV-12
percent and is the leanest mixture that burns competely. Leaner mixtures down to the noncoherent limit of 4-percent hydrogen are still flammable, but the flame is made up of separated globules
' hat slowly ascend. The noncoherent flames occur because of the high diffusivity of hydrogen.
The flam=able range is widened by heating the unburned mixtures. The lean limit occurs at lower co.centrations, and the rich limit at higher concentrations as the mixture temperature is increased. This is shewn in Figure IV-3 There is a linear change in the limits with miv*.ure temperature, and the rich limit is somewhat
; more strongly affected than the lean.
The addition of inert gas to a flammable hydregen-air mixture may dilute the mixture to nonflammability. The rich limit is sharply decreased as inert gas is added, whereas the lean limit is scarcely changed. This effect varies between diluents. For instance, it takes more nitrogen than carbon dioxide to prevent l flame propagation, presumably because of the greater heat I capacity of the latter. F. Ignition Requirements Hydrogen-air mixtures ars .xtremely easy to ignite as evidenced by the mini:num spark energy required to initiate combustion i of only 0.000019 joules. This spark is not visible to the human I eye (reference IV-8). Figuri IV-4 illustrates the variation of this energy with hydrogen concentration in air. Since the l IV-13
FLAMMABILITY LIMIT VARI ATION 8 L . 13 o . m
- : Q HYDROCEN/ AIR MIXTURES U
o - l
- DOWNWARD FLAME PROPAGATION
~
b g - C 7 ..................;........ FLAMMADILITY LIMITS - - g! M o : . . : - a .- g, n 8- . . . . .;.. L1e
. . g.c Lo .
m t .. . . . . . mR q Wr, m - Q
- p,. : . .. .
(f.:)
% b 4*
d_ . . .:. . . . . .. . .. : .. . ....'. . .
.. . m M*
. c: - -
- M mo w : .
an d w e S-
. \%;p M -
g . . C 8.. m . . . . . . . . .. . . . . . . . . . . . . . . . . . . & n - t.M o. g . . t . o . . fN _
~
- g. .
5.0 13 2 21.4 20.0 37.8 46.0 542 62.4 70.6 78.8 87.0 IIYDROGEN CONCENTRATION (volume %)
l 6 . j 4 - g 2 - w
.J o PRESSURE, P, o, 1 -
1 atm - a 8 - 5 , 6 -
> 4 _
c .3 c: w . z w - z 2 - o - p .5 g N e .1 - 2 08 - D 06 - 2_. Z - 2 04 1 - l .02 - ; i i i i , i
, . 1 O . .
10 20 30 40 50 60 70 0 i HYDROGEN IN AIR, PERCENT BY VOLUME Figre IV-4 l l IV-15 l l
igniter system only uses glow plugs and not spark igniters this only is i=portant to demonstrate the mini =al energy required for ignition. It is anticipated that sparks may be abundantly available, however, particularly for the large LOCA where flooding of nonqualified electrical equipment occurs. Stall static electric charges =ay also be present. Strong sources of ignition =ay initiate deflagration or recombination in nearby mixtures even if they are outside the fla:mability limits. Reference IV-1 suggests one such source may be cc=bustion of a flan =able mixture in a partially confined area such as a duct or dead ended volume. The resultan,t strong deflagration would propagate energy into the lesser concentrated mixtures exterior to the space. Spontaneous ignition of hycrogen has also been examined (reference IV-1). Figures IV-5 and IV-6 illustrate the variation of the spontaneous igniticn te=perature with water vapor concentration and with exit velocity. -The spontaneous t ignition. temperature is measured by heating the entire mixture ! to each measured temperature, discharging the mixture into unheited air and observing whether ignition occurs. The spontaneous ignition temperature is expectec to be lower than the temperature required of a heated igniter since the igniter l must heat the surrounding mixture to the ignition temperature l and therefore must be at a higher temperature for heat transfer to exist. The spontaneous ignition temperature will influence l l whether the hydrogen immediately burns on exit from the primary l l ! IV-16 i I L
l i ( 9equagal7 Bybrogen 9tuby q
.s I . .
o M . . i 1 . l * . . . p . .. . . . . . . . . . . . c 4 W i i .
- m M
M
- . : o 4 . . . . . . . . . . . . . . . . . . . .
w e;
- g
~
N
- O i P *
} :
, D t .
l _ .s og
=
- . W O
- * * ~
O
?
E
- m. i
- i i s
- E""*
k=4 .
- : 4 Q
- cm
...'....................i.........i f
_ h ~p y i Z
- cd i
as
= D w
- [
. I ." ;
u E
>===4 4
. o
. . . .. . . . . . . . . . .. . . . . . .* . . . .........t..4........ . . .:. . . . . . . .
H ~
- C/) g>p !
i . . . . . . . . . . , , , _ m
- i D c- : - : O Q sid ; !
i ! i C1 g
. .... .. ,,. ,4 ca
. p l
Z
................p................................. . -
O
. 8 .
..............j..............:............ ......: . . . . . . .
i i
. . . . . . . . , , ,, Q .
i i i
.i 8
. 8
- p
.g *
. . 4 . .
q
. . . . O OI)e91 0'M9T O'9091 07.3 1 01M:tt OT)oti O*t9C1 0133CT 0"JS31 01NEWO331 (a.) annivuaanal FIGURE IV-5 rI-17
system. For cold leg breaks where the hydecgen must travel through the steam generator before leaving the break, it =ay be well below the spontaneous ignition temperature. Also for hydrogen release through the pressurizer relief tank, the bubbling of the mixture through the relief line sparger into the tank fluid and out of the rupture disk will cool the gas. However, for a hot leg treak concurrent with inadequate core cooling, the hydrogen-steam mixture may exit the break at temperatures well into the spontaneous ignition region. It would then mix with air and deflagrate near the break opening. If the hydrogen does not ignite ,and burn by the above mechanism, it will be important for the igniters to function to prevent the accumulation of large hydrogen concentrations. The proper functioning of the igniters requires an estimation of the necessary igniter surface temperature to initiate combustion. This temperature depends in turn on the flow velocity past the igniter. The glow plug system selected is expected to maintain temperatures between 1700 and 18000 F at the element surface. A preliminary estimate of the maximum allowable flow velocity past the hydrogen igniter glow plug has been made using l experimental data for pentane-air and hydrogen-air burns. A ratio was determined from data (reference IV-9) produced for both pentane and hydrogen atmospheres in which heated solid spheres of varying diameters were passed through the fuel gas-air mixtures at velocities of approximately 13 ft/see and i 17-19 4 r -
-1 v, -n
approximately 4 ft/see and the ignition te:peratures then measured. The ratio of the hydrogen to pentane ignition te=peratures was determined at a sphere diameter of 5 =m passing through the fuel gas-air mixture at approximately 13 ft/sec. and at close to stoichiometric fuel gas to air =ixtures for both gases, This ratio was then used to adjust a curve given for a fully heated 1/4-inch rod with a surface temperature at pentane ignition threshold plotted against stream velocity. This curve had been evaluated fo,r a stoichiometric mixture of pentane and air at an ambient te=perature of 1o0 : 10 F. The adjusted curve is shown in figure IV-7. The figure shows that-if the igniter is maintained at 1700 F, a flame will be sustained at flow velocities up to 158 ft/sec, which is well above any expected containment velocities. While this curve is for a 20 percent mixture of H in air the data used in determining the ratio indicated that the ignition temperature was much more sensitive to the stream velocity than the percentage concentration of H . This was noted in part because the 10 percent 2H mixture at approximately 4 ft/see gave lower ignition te=peratures than a 20 percent mixture at approximately 13 1t/sec. It should be further noted from figure IV-7 that at rod surface te=perature .92ch below 1500 F a flame would be hard to sustain IV-20
even at very low stream velocities. However, as noted in Appendix M, testing has confirmed that the IDIS glow plugs reach surface temperatures over 1600 F even at the minimum design voltage of 12 volts. From figure IV-7, it can be seen that an igniter with a surface temperature of 1600 F could sustain a Qll flame at a stream velocity of 100 ft/sec. Operation at ncrmal voltage should produce temperatures well over 1700 F. G. Flame Speed Laminar flame speeds are measured using several techniques, although each has sources of inaccuracy. Recent theoretical i calculations of laminar flame speed using laminar-flame-structure - models are in good agreement with experimental work, as described in reference IV.10. Hydrogen-oxygen reactions have been studied to the point where the detailed chemical reactions of importance in hydrogen-oxygen combustion are believed known, and the reaction rates are fairly well known. For hydrogen-air mixtures at atmospheric pressure, the thickness of the laminar flame front has been determined to be about imm. l Figure IV.8 shows a comparison of laminar flame speed for ! hydrogen-air mixtures. One investigator has computed the laminar j flame speed and compared his results with those of other experimenters. After corrections were applied to early work believed.to be in error, the revised data are in good agreement illustrating that theoretical understanding of laminar burning I now exists. Figure IV.9 shows the effect-of different ratios l of oxygen.to nitrogen concentration on laminar flame speed. IV-21
I I I I I I Figure IV-7. Red surface temperature at ignitien thr" nhold versus stream veloca.ty. 20% mix-ture of g in air.
%o _
I l t 1900 . g 1800 ~ 0 v a2 h-M 1700 S! 6 2 c: 8 g 16m _ _ _ _ _ _ _ _ . _ _ . . . ._ ... as 1500 , I l l 20 40 60 80 100 120 140 160 160 200- 220 240 260 STREAMVELoCITY(Pf/SEC) TV-PP
t 400 . . . . . . d b 300 - M do - t /* . cr o a O *
?
l-6 o x Gj 200 - I N -
> 8 x O , b.
-E O e
^
b 100 - D ,/ STOICHOMETRIC. , filXTURE 0.2 0.4 0.6 0.8
. HYDROGEN, PERCENT l LAMINAR Fl.AME VELOCITY OF HYDROGEN-AIR MIXTURES l T 298 K, P 100 kPa I
37 Line - Ca!culategValues, Warnitz e Jahn ( x 1. 2)
+ Scholte and Vaags ( xly)24 O Edmondson and Heap D Bartholoms a Bunvasser and Pease c Gdnther and Janisch cG Unther and Janisch
= Andrews and Bradley e Miller, Evers and Skinner
< Gibbs and Calcote l
.
- Smith and Pickering Figure r/-8
, (reference r/-10) l IV-23 l
Figure IV-9 (reference r/-9) 1200 , , , , , ,,,, 1100 - . 02 02+N2 1000 - .
.90 .
.80 / .
900 - .
.70 -
l 800 . .60 , n .50 l I E o 700 - . . M ' f b - I . L 40 g 600 -
/ .
> .35
- a - -
I Z
.30 D
500 - [ . 1 Q }
.25 l l 400 -
.21 j
.175 .
300 - t -
.15 . .
200 - -
- i. .
100 - - O *10
, d/20 30 40 50 60 70 80I 90l 100 HYDfqGEN, PERCENT
w The maximum 3 ,. tar flame speed of hydrogen-air mixtures is about 10 feet /second near a concentration of about 42-percent hydrogen. As the flammability limits are approached, the flame speed becomes much smaller. Diluents such as nitrogen reduce flame speed by reducing flame temperature, apparently by removing energy at the reaction front. Steam also reduces flame speed, but by less than the amount expected from equilibrium flame temperature considerations. Moderate changes in ambient temperature and pressure do not significantly alter the laminar flame speed. For instance, a 90 F temperature rise above room temperature, the increase in laminar flame speed is less than 0.6 ft/see for pressure change. In the range of interest for reactor containments, the expected variation of hydrogen-air flame speed with pressure will be small. The expanding laminar flame of a plane or spherical front has been shown to be unstable. The importance of the instability ' of the flat laminar flame front is that the mean speed of a laminar flame will be somewhat higher than the laminar flame speed. The laminar flame speed is the normal component of r ! velocity of the unburned gas moving into the flame front. If l the normal to the flame front is at an angle to the direction of flame propagation, tnen the propagation speed of the front is equal to tre laminar flame speed divided by the cosine of l the angle. The Mach number of the front will be extremely small, i IV-25
~
in any event, if the front stays laminar. The pressure is then 4 expected to be unifccm in containment, and to rise monotonically. The centainment structure loads will be quasi-static. However, the instability of the laminar flame fecnt can lead to turbulence, flame acceleration and possibly transition to detonation, if the sixture is within the detonation limits. ; l A laminar deflagration is likely to beceme turbulent in ! i containment. The "self-turbulisation" of laminar flames has 4 been discussed by Shivashinsky (IV-11). Many turbulent flames hav i mean flame speeds in the range two to five times the laminar flame speed. However, even at these speeds, the Mach number l of the flame fecnt will be very low. The pressure in containment will be nearly spatially uniform, and the pressure will rise monotonically. -
.. t
.1e action of obstacles in the path of flames is under study I by various researchers. The flame front is stretched and-turbulence is promoted. Increases in flame-speed have been l
observed when a f3ame fecnt passes through a field of obstacles such as wculd be expected in the lower sections of PWR containments that contain many pipes, tanks, ceams, etc. However, much of the upper portion of such containments is fairly open, particularly as in the case of the. upper compartment at Sequoyah. Unfortunately, there have been r.o experiments to study flame speed behavior whsn the flame front leaves an obstacle field and enters an open region. The consensusaof several l l l l IV-26
. , ., . ~- .-.. - . - , - , - - _ - . - -
researchers suggests that the flame speed might decrease after leaving the obstacle field. Detonation waves travel at a speed very close to that corresponding to the Chapman-Jouget point, except that " marginal detonations," (detonations in mixtures near the detonation limits) travel at a speed lower than the Chapman-Jouget speed. The pressure ratio across the detonation wave for hydrogen-air detonations will be approximately 16, the detonation velocity approximately 6500 f t/sec, and the temperature behind the detonation wave will be about 4640 F. H. Burn Efficiency , From several small- and sedium-scale laboratory experiments, it has been found that- when hydrogen-air mixtures with hydrogen concentrations in the range 4-8% were ignited with a spark, much of the hydrogen was not burned. The resultant pressure rice was far below that predicted for complete combustion. Experimental results with a spark ignition source, shown in Figure IV.10, indicate the completeness of combustion increases with increasing hydrogen concentration, and is nearly complete l at about 10% hydrogen. The range of incomplete combustion corresponds to the range in which the mixture is above the upward propagation flammability limit, but below the downward propagation flammability limit. " Separated globules" of flame have been observed in upward propagation of lean hydrogen-air flames. Even when ignition occurs at the bottom of a chamber, ! the upward propagating flame fails to burn some of the hydrogen. l i l IV-27
I i w 1[ : T
. o 'o ' *M,'o' '
('c1 f.imi 4, / h 100 13 33 g so 1.1 12 7 as is s l w - A 11 ..: ir-ri
/
O FUANO1 DATA
/
/
THECRETICAL d P FOR COMPLETE COMSUTTION
/
/
\ .
/
1
* ~
/
5 -
/
/
/ -
/
l n -
/ \
y
/ .
./ .
/
/
to - ! '
/
/
/
/ a
/ AA
. i GC4C"Aho ol ,
0 2 4 g a to 12 MYORCCEN CONCEVTR ATION TVOL %I
. Figure r/-10 PRESSURE RISE VERSUS H1pRpGEN CONCENTRATION, SPARK IGNITOR
References IV-1 McLain, Howard A., " Potential Metal-Water Reactions in Light-Water-Cooled Power Reactors," ORNL-NSIC-23, August 1968. The above document also draws from several analytical and experimental investigations listed below: IV-2 Carlson, L. W., R. M. Knight, and J. O. Henrie, " Flame and Detonation Initiation and Propagation in Various Hydrogen-Air Mixtures, With and Without 'JAter Spray," Rockwell International, AI-73-29, May 1973 IV-3 ordin, Paul M., " Hydrogen-Oxygen , Explosions in Exhaust Ducting," NACA-TN-3935, Lewis Flight Propulsion Laboratory, Cleveland, Ohio (April 1957). IV-4 Drell, I. L., and F. E. Belles, " Survey of Hydrogen Combustion Properties," NACA-BM-E-57024, Lewis Flight Prepulsion Laboratory, Cleveland, Ohio (July 1957). IV-5 Coward, H. F., and G. W. Jones, " Limits of Flammability of Gases
& Vapors," Bureau of Mines Bulletin 503, Department of Interior.
IV-6 Moyle, M. P., "The Effect of Temperature on the Detonation , Characteristics of Hydrogen Oxygen Mixtures," University of Michigan, Doctoral Dissertation IP-195 (1956). IV-7 Dre11, Isadore L. and F. E. Belles, " Survey of Hydrogen Combustion Properties," Report 1383, Lewis Flight Propulsation Laboratory. IV-8 " Hydrogen Safety Manual," NASA TMX-52454, Lewis Research Center, 1968. IV-9 Lewis, Bernard and G. Von Elbe, " Combustion Flames and Explosions of Gases," 1961 IV-29
IV-10 Sherman, M. P. , et.al. , "The Behavice of Hydrogen During Accidents in Light Water Reactors," Sandia National Laboratory and Energy Incorporated, NUREG/CR-1561 R3, August 1980. Note the above reference contains an excellent discussion of hydrogen combustion which has been used as background in this report. IV-11 Shivashinsky, "On Self-Turbuli::ation of a Laminar Flame," Acta Astron 5: pg. 569-591, 1978.
.=
o l l l 1 l 1 l l l l l r/-30
i l SECTION V j CONTAINMENT PROCESSES AND DESIGN A. Overview There are several syste=3 at Sequoyah which, by performing their normal design functions, will assist the igniters in the controlled combustion of hydrogen. Cne system, the Combustible Gas Control System, directly aids the igniters by recombining hydrogen in the containment. The other systems will act as heat sinks, serve to mix the containment atmosphere, or do both. Heat sinks, by absorbing energy generated in a hydrogen burn, reduce the containment atmosphere temperature and pressure. Mixing the containment atmosphere aids in the prevention of hydecgen pockets. Pockets of hydrogen with concentrations in i the detonation range may result in very high te=peratures and pressures if combustion were to occur. Thus heat sinks and the mixing of the containment atmosphere will reduce the temperatures ! and pressures inside containment and reduce the probability of j containment rupture. Details of various systems are given in l ! Appendix I. I i l t B. Combustible Gas Control System ? ' The Combustible Gas Control System is a safety system designed V-1 I
into the Sequoyah containment to mitigate the effects of hydrogen production. This system is composed of four subsystems which will adequately process all of the hydrogen postulated in the original design basis accident scenario with 100-percent
; margine. They include hydrogen analyzers for postaccident j monitoring, electric hydrogen recembiners for removal of l
l hydrogen, a containment mixing system to prevent pockets of hydrogen at higher concentrations than the general containment ! airspace, and a small hydrogen purge system as a backup to the l l recombiners. The Combustible Gas Control System will assist the igniters in recombining hydrogen in the containment atmosphere.
~
C. Containment Structures The containment is a large structural volume designed to contain
- the radiation, mass, and energy. released fecm a breach of the l
reactor coolant system and the nuclear fuel cladding. The Sequoyah containment is a free-standing steel structure with a free volume of 1.2 million cubic feet. The design pressure of this structure is 26.4 psia which exceeds the calculated pressures resulting from any of the present design basis events. I ( The containment is divided into three compartments: The lower compartment, the upper compartment, and the ice condenser compartment (see figures below). The lower compartment l completely encloses the reactor coolant system equipment. The l l l V-2 i
upper compartment contains the refueling canal, refueling equipment, and the polar crane used during refueling and maintenance operations. The ice condenser connects the 1 ewer compartment to the upper. In the event of a hydrogen burn, the steel containment shell, the internal concrete floors, walls, and ceilings, the internal structural steel, and the piping and equipment located inside containment .till act as heat sinks that will remove energy from the containment at=osphere. D. Ice Condenser System
- The ice condenser is a compartment inside containment enclosing large quantities of ice which is used to condense high-energy steam escaping frem a LOCA or MSLB. By condensing the steam, containment temperatures add pressures are kept relatively low.
During a LOCA or MSLB, the Icwer compartment pressurizes, forcing the lower compartment air-steam atmosphere mixture into the ice condenser. Inside the ice condenser, the steam is condensed and drained back into the lower compartment while the air is cooled and forced into the upper compartment. The ice condenser will assist controlled hydrogen combustion l not only by providing a heat sink, but also by helping to mix i i hydrogen with air as it passes through the ice condenser. { l E. Centainment and Residual Heat Removal (RHR) Spray Systems l 7-3 l
The containment spray and RHR spray systems are designed to act as containment atmosphere heat sinks by the addition of spray water into the upper compartment. Each system consists of redundant spray headers located in the top of the upper compartment. The functiening of these systems as heat sinks contributes to f controlled hydrogen combustion. Furthermore, the spray systems do a very effective job of mixing the upper compartment i atmosphere. F. Air Return Fan System The air return fan system aids controlled hydrogen combustion by assisting in the cooling of the containment atmosphere and by mixing the air inside containment. The air return fan system enhances ice condenser and spray heat 4 removal by circulating air from the upper compartment into the l lower compartment. The fans thus establish a recirculation path l l which allows more of the lower compartment air-steam atmosphere l l to be transported through the ice condenser and into the upper compartment spray area. Ducts which lead to a header on the suction side of the air j return fans draw air from the containment dome, accumulator l rooms, steam generator and pressurizer enclosures, and other j dead ended spaces where hydrogen may accumulate. (See figure i I-2.) ( l V-h L _
'4 hen both fans are in operation, there is an air circulation rate of one lower compartment volume in less than five minutes.
This kind of highly turbulent flow will also help prevent accu =ulation of hydrogen in pockets and possible detonations. 9 9 e i I V-5
e l
...)- .
W t 1. e e
.- =
for eft. ( g u,g f~****" g ,{'f7 i k;',7
" ""( f aq
.=-
t, J x cav - r
# I e
- i.c.o =n... . ou - -i
' tog 2
l .' i CamJet
*'a r
. - i ne ~
f ' i su.u., hl l , _- --.n c%,i. .teet.t . .--.-
*l l g l
i wtut _ _
- i. f .
,., hlqM.ft 3 ee . s(g ggegt J
a g ,1. i
~
su ' I w. t aA=g s y . -
- p, ,F-u.cm coneas , x ---
* , , i j = , cueu= --- -
l ' j. ........% 1 l -Mw*Mts timt g i l6 l.. 1 E -a un ar ..-
- El
- cCWit3L 83 9tre( r w l i g , J.. .. d , estnr s.it6s - 7 I .
*[
g
=,
j
- f '. ,
l
. . . . . . , .8 l 8CI S
-w. L l
..q.s. * , vw'O n.
I , , ,y , ,,_ i*
- . ,e .l' 84 'W i .
' 2- i 1
-- ,-t 'le , 1 -4 r- ** . _
.is.se
%%, l
. . .ac, s'act r =' ;fl
- ii
.. r .
' '
- 1 .
j
.um e- -
J [ 't T, .\
.- AP.
\ , . s e .:-' s g !- .
.' TJ .. i ,;
s Nv s > i, 4 y q ( , .. _. 4 . g/ 4 .--.
. eeruerc es.
&lk'rmo
. ....i ,
\ .,5.
.~
NN'? .- l g; ( N?
.. l '.
n../. .. . '..=.\- _ [' '. ' i
.jj'? .
,' !. t,*..'.-
4 3..;;:*. :.
" U. S{.Th5
* '. , .,\ . \
1
\
Q
- 1. ; , .T : , *~
- jf.?*; : .
, l ' 1,z l * ,
g1 l ! C.eou.ces..se n eu,sw .w 'a(* i
- y. ' .
.j. .t
, i.'r.N.r, '+.
- 1*.:. O opO .
A. ~@" AJ
~~ 1
- .c3 l J .s. .a l FIGUPI' V-1 l COICAIm.iE?C CROSS SECTION l V-6 L
e,, y.g-get, at toegg I l l IO s
. .as , .i, _ _ _
I* I s .- 4 !
+-e .x r, cu< ..u-.
] (i 4 I
st M*f f' ,
- ElLt>dL4MB
~"
\ _
l 1 0 l mmauis e j
;. l l
[
===.arme - l j 1 1 .
i r f._'._ oq i='-* t-
-~
gy
.a C 3.140. 0&2 8.t<t lisdj O'i O
].,,., ,
. -=
7.P..
.i
. ,(,--- '
1 gr '
- e u.s s a.
e
. ,u, l
man e=es m.m.se cou.pede e /gt-i . s' f--'f v rw ' '
... g l '
) ,. e t ._
s'4
^
il , , e / t f I e f-~ ~ ~~, I-'tg i ).. In I ome einreww
= wm
/
i - l iw, . .f 1 sat ~ s
./ ,,
,/ . - ~5 ,--.s
. LL '.')
7.< J,.. } ff E I., . . _ . _ Cg.g 9.-. t ; . yt " f I ,
-..)
- %p p / ,+ . T - .m.s ,
L. --&e . is .- . : !L. :&&-
/. .
- m. ;
'3: _..
- n.x ....
. g .
- .2 m,._.a. . Tu .
=.
. ., l.
j l .' . '
/ _ , , , . .
-...- .. 2.
;. 4 . .!.
S , . ' . ' .**, . . ;, l 4 i,y, , .: . ': * . - ,.j,:...' 3, notmz V-2 0 0 ccyr,ttmerrr cacSS SE'CHON oo od.1. _a V SECTION VI HYDROGEN DISTRIBUTION A. Overview Appendix C provides a discussion of facters that can cause or reduce nonuniform hydrogen concentrations in the containment, postaccident. Since the mass, energy, and gas release occurs in the lower ccmpartment, the hydrogen concentration tends to be higher there. Air is swept to the upper compartment, depleting the concentration of oxygen in the lower compartment. , The air return fans return air to the lower ccmpartment. 1his gross circulatica pattern is modeled in the CLASIX computer code which provides the bulk hydrogen, oxygen, nitrogen, and water concentration in each major compartment. While certain processes can establish nonuniform concentrations within major compartments, the design of the ice condenser containment provides a substantially better situation than large dry containments in regard to mixing. The ice condenser and air return fan flow provide a relatively high flow of air through the major compartments. The hydrcgen mixing subsystem provides I flow through all significant compartments. In addition, the pressure changes and temperatures associated with hydrogen b'Jrning provide both flow between compartments and thermal gradients which enhance natural convection. VI-1
As a result of these factors, there appears to be no reasonable mechanism that could cause a substantial volume of combustible gas at hydrogen concentrations significantly above the bulk hydrogen concentration of the compartment. Sufficiently detailed analyses have not been completed to determine if small local.i ted concentrations could develop, but it appears that even if such concentrations develop, they would be too small to seriously jeopardize containment. D. Factors Premoting Nonuniform Concentrations Two mechanisms have been postulated that could lead to localized hydrogen concentrations: the "pl'ume" or " jet" emanating from the RCS release point and removal of water from the air by condensation. The " jet" or plume is composed of steam and noncondensible gases. In general, one cannot get hydrogen generation and release without significant~ steam being present. Therefore, the hydrogen comes into containment diluted in steam. For this l mixture to be combustible, the effluent must be mixed in I substantial amounts 'of air (to provide oxygen). This process further dilutes the hydrogen, resulting in a concentration not t
- significantly different than the bulk lower compartment concentration. In any event, the " plume" is limited in size and is rapidly entrained in the flow through the lower compartment. Although the steam generator and pressurizer dog Lauses may provide some trapping of this plume, their low volume, VI-2
high design pressure capability, annular geometry, and high natural circulation flows minimize the impact if such concentrations were unexpectedly formed. Condensation increases the hydrogen concentration by reducing the partial pressure of water. This can occur in the ice condenser, in lower compartment coolers, and due to upper compartment sprays. The extent of such an increased concentration is small for the ice condenser and cooler effluent due to rapid mixing downstream. This limits the total energy released from combustion of the effluent; the effects are bounded by the current analysis assumptio'ns of burn throughout the major compartment. The containment sprays remove substantial water by condensation only after ice melt. For small LOCA's analyzed, this occurs after the hydrogen has been released and burned. In addition, the spray water suppresses combustion and provides a heat sink that minimizes the pressure and temperature due to combustion in the upper compartment. C. Factors Promoting Uniform Concentrations The ice condenser containment design includes many features that enhance nearly complete mixing. These include the following:
- 1. The air return fans supply recirculation flow through the Q6 accumulator rooms, through the fan rooms, and into the main area of the lower compartment where convection currents six
! it with the lower compartment air. The flow eventually i proceeds back through the ice condenser into the upper , compartment, thereby promoting rapid mixing both by induced i l VI-3 ,
turbulence within each compartment and by moving air between the compartments. With a lower compartment volume of approximately 289,000 cubic feet and a recirculation flow rate of 80,000 cubic feet per minute, the air mixing will be essentially complete. The cean residence time in the general lower compartment spaces is on the order of one to five minutes compared to the time period for hydrogen release which is 15 minutes to a few hours. The preferential air currents ?stablished by the air return fan flow are expected to dilute ;he hydrogen released frc.* the break and sweep it toward the ice condenser inlet, thereby mixing the break hydrogen and steam with lower compartment air.
- 2. The hydrogen mixing system provides auction fece all significant compartments. The flow rate is only designed to remove a slow accumulation in the upper regions of these areas. The low design flow rate (longer mean residence time in these volumes) acts to isolate these areas from the bulk of the lower compartment volume relative to the preferential flow paths created by air return fan flow. Significant flow of hydrogen from lower compartment areas into the steam generator doghouses, pressurizer dcghouses, and other confined regions cannot occur directly from the use of the skimmer system. It 1: Jesirable not to introduce hydrogen into these areas if possible, but to keep the hydrogen in the lower compartment where the igniters can combust the mixture. The potential for uncontrolled hydrogen jetting toward the doghouses and placing hydrogen directly into VI-4~
these areas is discussed in appendix C.
- 3. The air return fan recirculation flow paths also tend to isolate other dead-ended compartesntal areas in the lower containment from the release point, precluding preferential buildup in those areas.
- 4. Q6 The upper compartment sprays provide very effective mixing of the upper compartment due to spray-induced turbulence.
In addition to these mechanisms, mixing is promoted in each compartment by the pressure changes due to burning via temperature difference induced natural circulation and diffusion. The first two are so strong that they overshadow the effects of diffusion for events of interest in an ice condenser containment. ee , . t' . D. Program Plans TVA intends, as part of its Phase 2 studies of controlled ignition, to quantify the effects of mixing. These efforts include:
- Increased number of nodes in our containment analyses codes to identify more subvolumes and flow paths,
- Analytical studies and literature research to place bounds on the extent and magnitudes of local concentrations,
- Studies of the behavior of release point plumes and jets, and
- Sensitivity studies to evaluate the effects of different event and failure scenarios.
VI-5
-~
SECTION VII ANALYSIS OF CONTAINMENT CONDITIONS A. Overview An analysis was performed to determine the environmental conditions inside the Sequoyah containment which will result from igniter operation in an accident environment. The Westinghouse /0PS computer code CLASIX was used to perform calculations of containment conditions. CLASIX is a multi-compartment containment code which calculates pressure and temperature response for the individual compartments while monitoring the distribution of oxygen, nitrogen, steam, and hydrogen. CLASIX can model many of the features of an ice condenser plant such as the ice bed, ice condenser doors,
- containment sprays, and recirculation fans. Except for the ice bed, it does not, however, have the capacity to model passive heat sinks or ECCS recirculation.
Since CLASIX does not model the primary system, energy and mass data for blowdown (including hydrogen) must'be provided as input to the code. Hydrogen burning can be modeled with variable criteria for the hydrogen burn and burn propagation between compartments. Further discussion of the code is presented in appendix T. ,
==
5 VII-1
A small break LOCA followed by failure of ECCS injection (designated S 2D in WASH-1400) was used as the base case for analysis to determine the conditions inside containment for an accident plus hydrogen burn scenario. This particular accident results in large amounts of hydrogen being produced due to a lack of core cooling. The Battelle-Columbus Laboratories code MARCH was used to calculate the primary system blowdown mass and energy releases and hydrogen mass and temperature releases as a functicn of time for . an S2 D accident. Hydrogen production versus time is illustrated in figure III-5. The hydrogen mass upper limit of 1550 pounds corresponds to a 75 percent zirconium-water reaction in the core. Contatument initial conditions for about the first 3500 seconds' of this accident (tne time prior to hydrogen production) were calculated using the Westinghouse LOTIC code. This code calculates the long term ice condenser containment pressure and temperature responses. The MARCH and LOTIC calculated data was input into CLASIX for the base case analysis. Gther initial conditions and information which were input into the CLASIX calculation are shown in app.ndices T and U. For the base case, the hydrogen ignition in a given compartment was assumed to take place at an atmospheric concentration of 10 volume percent. Propagation of hydrogen burning to other compartments could only occur when the volumetric 0 VIZ-2
hydrogen concentration in the atmosphere of that compartment was 10 percent or above. Results of the CLASIX base case analysis are shown in appendix U. The results of the base case analysis indicate that the hydrogen will be ignited in a series of nine burns in the lower compartment. The time interval for the series of burns is approximately 2300 seconds. As a result of the action of engineered safety features, such as the ice condenser, air return fans and upper compartment spray, the pressure and temperature spikes were rapidly attsnuated between burns. The pressure was decreassd to its pre-burn value roughly 2 minut'es after the burn occurred. After the last
~
. ignition of hydrogen, at 6800 seconds after accident initiation', approximately 300,000 pounds of ice is left in the ice condenser section. The results of the base case analysis show an increase in containment pressure of 4-6 psi due to burning, with the containment remaining well below the estimated failure pressures. All burning in the analysis occurs in the lower compartment since this is the area of hydrogen release, thereby gaining the advantage of heat removal by the ice bed. By turning at a given concentration in the lower compartment there is also the advantage of burnirg less total hydrogen at a time since the lower compartment volume is only around 1/4 of the total containment volume. VII-3
It should be noted that passive heat sinks were not considered in the CLASIX analyses, thus neglecting the heat capacities of the containment vessel, steam generators, pressurizer, structural concrete and structural steel, and other concrete and steel materials inside containment. These masses will absorb large quantities of heat during a LOCA and hydrocen burn. If considered, these passive heat sinks would shift the calculated compartment temperature and pressure curves downward, resulting in centainment atmosphere temperatures and pressures which are lower than those calculated. ' B. Methods, Model, Assumptions, Inputs, Environmental Conditions, Results, and Verifications These subjects are described in appendices T, U, and V. C. Containment Environmental Conditions Analyses have been performed to scope the environmental conditions inside the containment following a hydrogen deflagration. The major factors influencing these conditions are flame speed, flame temperature, heat sinks, specific heats of the gases, relative humidity and the composition of the gaseous mixture. These variables affect the environmental conditions in various ways. The flame speed and flame temperature control the temperature peak of the atmosphere during the burn. The specific heats also help to determine this peak VII-4
though to a lesser extent. The composition of the mixture and the relative humidity are important in that these two factors control most of the major characteristic of the mixture (specific heat, density, etc.). The heat sinks tend to moderate the temperatures and greatly reduce the temperature peaks in the atmosphere. The flame speed information available for hydrogen mixtures at presssures near atmospheric pressure indicates a range of speed varying from 6 to 10 fps, the exact speed depending on gas concentrations. Structural heat sinks inside the lower compartment consist of over 56,000 ft 2of concrete surface area with each concrete sink having a thickness over 1 foot, and over 2 24,000 ft of steel surface with thicknesses ranging from .53 to 1.7 inches. These heat sinks act to lower the atmospheric temperatures both during and following deflagration, as they absorb the heat given off by the hydrogen burn. Both coavection and radiation heat transfer mechanisms are active during this process. Since the trend of the specific heats of the gaseous components is to increasa with increasing temp,erature this phenomena assists in keeping the environmental temperature lower. The humidity of the atmosphere is important in determining the peak temperatures of the systems. Water vapor has .e *ghest specific heat of any component gas in the atmosphere (excluding hydrogen present). It is postulated that the relative humidity of the atmosphere is at or near 100 percent due to the presence of steam escaping from the break along with the hydrogen. This water vapor lowers the adiabatic flame VII-5 J
temperature to about 670 F and removes a large amount of heat once it condenses on the cold containment surfaces. Preliminary calculations using 100 percent relative humidity and approximately 25 percent of the available concrete heat sinks (no steel) shows that temperatures are considerably below those of the S D s enario as calculated by the CLASIX code. The method 2 used involves computing the minimum natural ecnvection, based on that of a horizontal steel plate. The value of the natural convection coefficient selected is conservatively low considering the fact that considerable turbulence should exist, generated by both the air return fans and the combustion process. Although ignored, the natural convection will be augmented by the forced convection provided by air return fan operation. The radiation heat transfer coefficients used are based on radiation l calculations that are from furnace formulas developed by the International Flame Foundation. All heat produced by the hydrogen burn is assumed transferred directly into the gaseous mixture and subsequently it is transferred into the heat sinks. These calculations assume that the heat is uniformly generated in the lower compartment, that all heat is contained within that volume and there is no heat loss to the upper compartment or to the dead-ended compartments. There is 65.8 pounds hydrogen l burned in each of nine burns spaced in time according to the t burns in the S2D scenario. This is a 'Atal of 592.2 pounds of hydrogen burned (this corresponds to approximately l a 9 percent zixture of hydrogen by volume for each burn). l t I VII-6
i Using the above situation with only about 20 percent of the lower compartment heat sinks and none of the upper and dead-ended compartment heat sinks, temperatures were found to peak in the range of 1250 to 1720 F depending on the timing of the burn. Temperatures dropped rapidly between burns to below 600 F. Further studies may be appropriate. It is anticipated tnat calculations incorporating all the available heat sinks, the , ice condenser (which has potential for removing additional large quantities of thermal energy), and the heat absorbing capacities of the upper compartment and the dead-ended compartments, will demonstrate that the effective temperatures encountered following a deflagration, are expected to be on the order of those which occur for a main steam line break. This is in contrast to current studies which only include the ice condenser as a heat sink and indicate high lower compartment post-burn temperatures. An evaluation of the effects of spontaneous ignition in either the 03 upper or lower compartment has been made. Section 6 of Volume I of the TVA Rydrogen Study provided the results for a spontaneous burn in the upper compartment for an S D sequence. The results 2 show that the containment spray is very effective in removing all of the energy released during a spontaneous burn. For a burn in the upper compartment, the containment pressure and temperature remain well below the values calculated for a large LOCA. The evaluation of a burn in the lower compartment considered the ice condenser and the structural heat sinks. The structural heat VII-7
I I e i sinks considered are listed in Table VII.1. i Only the heat sinks
; located inside the crane wall were considered and the concrete )
floor was neglected. Table VII-2 provides the heat capacity of { the heat sinks for various temperature changes. The area of, t concrete used in the evaluation was reduced to account for the l differences in thermal conductivity between steel and concrete. [ Our evaluation shows that the maximum temperature rise in the heat i j sinks (in conjunction with the ice condenser) required to remove
! all of the energy released between 2200 seconds and 4800 seconds
! would be 225 F. This time period covers the onset of the release
! of superheated steam from the core to the time of peak hydrogen production prior to complete core melt. The mass flow rates'and j energy release rates were taken from the MARCH S D results. The 2
time period chosen is the most severe for heat sink response. In addition to the bulk atmospheric response, an evaluation of the radiation from a flame to a steel plate was made. The flame was I assumed to have a temperature of 4390 F and an emissivity of one. The radiative heat transfer was calculated to be 277 Btu /hr-ft at 2 10 feet from the flame. The heat transfer drops to 11 Btu /hr-ft l at 50 feet from the flame. This evaluation shows that bulk i l temperature effects are not cverly severe; however, locating equipment away from the immediate vicinity of a continuous flame or some thermal protection of ( 'jacent instrumentation may be needed. { l l i VII-8
W TABLE VII-1
~
Sequoyah Lower Compartment Structural Heat Sinks Thermal Volumetric Area Thickness Conductivity Heat Capacity 2 Material (ft ) (in) (Btu /ft-hr #F) (Btu /ft3) Concrete- 48,161 1 13 2 0.84 30.24 Steel- 29,925 2.9 27 3 59.2
- 1. In the analysis 1482 ft was used instead of 48161 ft .
4 i 1 --- 5 l i 1 l l SE:: . VII-9
TABLE VII-2 4 Heat Capacity of the Lower Compartment Structures Concrete l AT ( F) Q (Btu) 25 4.01x10 7 50 8.01x10 I 0 75 1.20x10 0 100 1.60x10 0 150 2.40x10 0 200 3 2x10
~'
A T (OF) Q (Btu) 50 2.14x10I 100 4.28x10 7 150 6.42x10 7 200 7 8.56x10 0 250 1.07x10 8 300 1.28x10 0 350 1.50x10
- 1. Based on an area of 48,161 ft . f Q3 VII-10
SECTION VIII ENVIRONMENTAL IMPACT A. Overview To fully evaluate the Sequoyah capability for hydrogen control following a LOCA, the containment environmental conditions discussed previously in Section VII =ust be applied to the plant's structures and syetems. The conditions to be evaluated are elevated pressure (including shockwave) and temperature transients. The structures include the containment shell and interior while the system components include numerous types of electrical and mechanical equipment. B. Containment Boundary , The Sequoyah containment boundary has been reviewed for its capability to withstand the environments conditions associated with the combustion of hydrogen. The containment boundary is divided into two distinct regions which have significantly different capabilities. These are the upper compartment free-standing shell and the lower compartment shell. A diccussion of the capability of these containment regions to withstand pressures and temperatures resulting from the hydrogen burn and the potential shock waves of a detonation is discussed below, section VIII.C, and appendices H and J. The lower region of ~the containment houses the primary system VIII-l
and associated emergency core cooling systems. A thick concrete wall heavily reinforced with steel which supports the large building crane surrounds the primary system separating it frem the containment barrier. In this region of containment the steel
, shell varies from a thickness of approximately one and three-eights inch to one and one-quarter inch. The space between the crane wall and the steel shell largely contains ECCS accumulators and containment cooling units. See figure 9.7, volume 1.
The 1cwer compartment communicates with the upper compartment through the ice condenser. The upper compartment is a large, relatively open volume. Its containment boundary, however, is largely hidden by the ice condenser which occupies the space between the crane wall and the steel shell. The steel shell varies in thickness also, having a minimum section of one-half inch at the containment spring line near the ice condenser top vent. Both upper and lower compartment shell have horizontal and vertical stiffeners. The horizontal stiffeners are installed on approximately 20-foot centers and the vertical stiffeners are spaced at approximately four degress. In addition, penetrations of the containment have been locally stiffened. 111 aspects of the containment shell were evaluated for the load bolring capability in order to identify the weakest element of the shell. This includes penetrations, valves, welds, and membrane materials (including effects of stiffeners). Each of
, these items were found capable of withstanding substantial pressures greater than 60 lb/in g. The one-half inch shell plate VIII-2
between the stiffeners at elevation 778'-6" and 788'-0" in the upper compartment was identified as the controlling critical area. The capability of the lower compartment region where the a membrane is more than twice the thickness has correspondingly much more capacity than the critical upper containment section. The actual yield capacity of the containment at this critical section of the upper compartment is 46 lbs/in g. See appendix J for more details on structural capability. l l l The environments which may exist in the upper and lower ccmpartment will be very different. In all events, the release
- of mass and energy occurs in the lower compartment. The temperature will be high, the atmosphere will be turbulent, and t
the pressure will be somewhat higher than the upper compartment.
. The upper compartment will be relatively cool, and the atmo' sphere will contain mostly air under pressure. The upper compartment sprays will most likely be on. In the event that significant hydrogen is generated, the hydrogen will be introduced in the lower compartment.
In the presence of igniters, the lower compartment is the most probable region for burning the hydrogen. Because of the geometry of the lower compartment and the potential for rich hydrogen release points, local regions exceeding the detonable threshold may be possible but are not probable because of the many violent mixing mechanisms occurring in this region.
'In the upper compartment the effluent from below enters via the ice condenser where the diluting effect of the steam is removed.
VIII-3 i
Here again it is possible for a pocket of highly concentrated H mixture exist momentarily in the upper plenum of the ice 2 condenser if high local concentrations exist in the lower compartment. However, mixing in the upper plenum is relatively strong and pockets of detonable gas will be quickly diluted. Effluent from the ice condenser will be further diluted rapidly as it mixes with the upper containment atmosphere under the influence of the containment sprays and the hydrogen mixing system. In summary, it is believed that the containment shell in the lower compartment may be exposed to quasi-static pressure loadings and corresponding heat loads from hydrogen burns in' addition to the steam environment from the LOCA. Also, the shell of the lower compartment could be exposed to impulse loadings occurring from localized detonations. The upper compartment, on the other hand, will be exposed to the quasi-static pulse of a possible hydrogen deflagration and the normal LOCA environment. The analyses performed by Offshore Power Systems and included i in Appendix U have predicted the pressure and temperature response due to burning within the various compartments of the containment. The calculations performed for the base case (best-estimate) showed a pressure increase'of no core than 14 psi in the upper compartment. This is far below even the yield strength of the
, weakest containment section. A recent OPS analysis (see Appendix Qf5 U) that assumed burning occurred only in the upper compartment VIII-4
showed that at least 450 pounds of hydrogen could be burned there without exceeding the nominal containment yield pressure of 46 psig. Preliminary analyses discussed in section VIII.F of the Q t5 heat loads indicate that the temperature rise in the containment walls will not be significant. While the containmcnt boundary may not withstand the pressure following a full containment volume detonation or large burn, preliminary scoping analyses indicate that sizable local detonations are within the capability of the containment. Local detonations and the structural capability analytical methods are briefly discussed in appendices H and J. C. Containment Interior Structures The principal interior structures of the containment are
. primarily the crane wall and the divider barrier separating the upper and lower compartment. These structures (see figures 9 7 and 9.8 of volume 1) are heavily reinforced concrete structures.
They were designed to take the large differential pressure loads associated with either the controlling LOCA or main steam line breaks in combination with other loads. The failure pressure of Qis the divider deck in the downward direction has been calculated to be at least 45 psi differential. Since the lower compartment I l will remain at some pressure above atmospheric during a hydrogen l l . burn transient (due to the mass and energy blowdown, even if
%\
l pressure increases from burning are neglected), the pressure 1 differential across the divider deck will be less than the pressure differential across the upper compartment containment l shell. So that segment of the containment shell is still expected l l VIII-5
to be the controlling structure for an upper compartment burn as stated in section VIII.B above. A As discussed above, the Offshore Power Systems base case studies Qlg-of burning hydrogen in the Sequoyah containment have produced a maximum differential pressure of about 14 psig. Much of these structures were designed for around 20 psig and larger if other loads in the design ecmbination were converted to differential pressure. TVA scoping studies as described in appendix D have shown that differential pressures across structures will in general not present a problem. l In addition, the air ducts at the outer perimeter of the ice condenser that separate the ice bed from its polyurethane foam insulation have been examined for structural integrity during hydrogen burn transients. The ice condenser wall air ducts and polyurethane foam insulation are shown schematically in figure l VIII-1. During normal ice condenser operation, cooling air is l l drawn from the ice condenser upper _ plenum by air handlers (fan cooler units) and discharged downward through the downcemer 1 section of air ducts adjacent to the ice bed region. At the bottom of the air ducts, the air turns 180 and continues upward l through the upcomer section of the ducts. At the top of the ducts, the air is discharged to the upper plenum. Thus the
- interior of the air ducts communicates directly with the ice condenser upper plenum but with no other portion of the ice condenser.
VIII-6 I .. .
ICE CONDENSER AfR DUCTS AND INSULATf0N I i TOP DECX UPPER PLENUM b ,= h CONTAINMENTT WALL AIR HANDLER (TYP.)
=4 I
y A INTERl4EDIATE DECK
\ -
URETHANE ] k
*}
INS ATION - [ (
' \^
N N ICE BED
\ REGION
\
\
\ >
\ e N s xt < .
N N
\ v
\
\ /LOWERSUPPORTSTRUCTURE N
N LOWER PLENUM NU -
- FLOOR FLgure VIII-1 VIII-7
l During a rapid pressurization of the ice condenser such as is calculated to occur during a large LOCA, a substantial pressure differential would exist between the ice condenser lower plenum and the ice condenser upper plenum, and hence, between the lower plenum and the air duct interior. These pressures, therefore, would impose compressive loads on the ducts and the urethane foam insulation. The ducts and roam are designed to withstand such compressive loads for the large break LOCA event. During a slower pressurization such as is calculated to occur I during long term pressurization following a LOCA or during a I hydrogen burn transient, no substantial pressure differentials will exist between various portions of the ice condenser. Therefore, no substantial compressive loads will develop across the-air ducts. (For example, during a typical hydrogen Larn, the ice condenser would experience a pressure rise of less than 1 psi /sec.) A compressive load will be applied on the foam 4 insulation during such slower pressurization events. However, the peak pressure load on the insulation during a typical hydrogen burn is 13-15 psi, a factor of two less than the minimum specified roam compressive strength of 30 psi. l As an added note, the air ducts are multilayer steel panels which totally encapsulate the foam insulation behind them. The panels are additionally attached to the steel containment wall by welded studs spaced approximately every three feet veatically and every . four feet horizontally. As a consequence of this. method of l construction, even if a very large pressure from an unspecified VIII-8
source were postulated to cause total duct collapse and roam crushing, the foam would remain on the containment wall and still be substantially protected by sheet steel from any local burning. It is also our best engineering judgment that containment internal structures would survive a substantial local hydrogen detonation with only minor concrete cracking based on the massive amount of reinforcement bar contained in the structures and on the results of high impulse loadings on similarly reinforced test specimens fecm tornado missiles. Further discussion on local detonations and structures =ay be found in appendices H and J. D. Critical Containment Systems In addition to the environmental effects of hydrogen ecmbustion on containment structures, there are several classes of electrical and mechanical components whose capability must be examined. These include various types of instrumentation, valves, cables, and protective coatings. When applying the environmental conditions discussed in section VII, it must be remembered that the CLASIX code and other state-of-the-art calculational techniques capable of analyzing ice condenser containment response to hydrogen combustion are conservative in their temperature predictions since structural heat sinks are neglected. Therefore, the peak temperatures predicted previously should be applied with caution if a reasonable evaluation is to be =ade. TVA has begun analytical efforts to predict the temperature effects of a series of hydrogen burns on VIII-9
i l I small components. In addition, work is progressing on the development of more reasonable methods to calculatelthe containment environmental response. As part of the phase 2 testing program on hydrogen ignition, TVA conducted a limited number of equipment survivability tests. Table 6 in Appendix N is a list of equipcent that was exposed to a 12 v/o hydrogen burn during the phase 2 equipment survivability tests. These components are representative of the critical components located inside containment which might be needed following a TMI-type event. The majority of the equipment did not experience any visible signs of degradation, even after as many as five burn tests. The only exceptions were some paint samples on concrete blocks which showed slight discoloration on the corners and one piece of cable which showed a couple of small (1/2x2 inch) scorch spots on the black plastic coating. Table 7 in Appendix N is a list of miscellaneous equipment which was also included in the test vessel during the testing. This list is provided to demonstrate the minor effect of the hydrogen burn environment on nonqualified off-the-shelf equipment. It does not appear that the pressure transients due to hydrogen combustion will exceed the environmental conditions of pressure for which the containment instrumentation is qualified. Shock wave effects have not been examined for small components. TVA is continuing its analysis and testing of environmental effects on critical containment components. I VIII-lO
E. Containment Shell Temperature An estimate was made of the temperature rise in the centainment shell due to hydrogen burning in the lower compartment. It was i assumed that the gas will lose heat to the containment shell by radiation and convection and to the ice condenser. Due to the relatively low temperature of the gas in the dead-ended i compartment, it was assumed that only the witer vapor emitted and absorbed radiation. The containment shell will reradiate a portion of the energy it receives from the gas, some of which vill be abscrbed by the gas. Simple finite difference equations were used to represent the heat balances for the containment shell and gas for a time increment t. The gas and shell temperatures were updated at the end of each time step and the calculation repeated until thermal equilibrium was reached. For a single burn of 100 pounds of hydrogen in the lower compartment, the average temperature of a 1" thick steel containment shell increased by approximately 8 F. Assuming l A similar temperature rise for each of the nine burns for the S2 D accident scenario, the mean temperature of the shell should increase by roughly 72 F. This corresponds to a total energy deposition in the wall of about 4.5 x 106 Btu. The one inch thick containment shell- was modeled as a one-dimensional slab.~ The total heat input of 4.5 x 10 Btu was added- to the shell over a 200 second time interval which
. corresponds to a surface heat flux of about 5370 Btu /hr-ft .
VIII-ll
The TAP-A computer program was utilized to compute the transient temperature distribution in the shell. The maximum temperature difference in the wall from inner surface to outer surface was approximately 21.4 F. The corresponding temperature difference from the inner surface of the shell to the center of the shell was roughly 15 7 F. An estimate was also made for an assumed accident scenario similar to that for the S D2base case except that no ice remains in the ice condenser after the first two burns. The calculational assumptions are similar to those used in making the previous estimate of the shell temperature rise for the S D 2 , base case. l Since ice is available for the first two burns, it was assumed that the average temperature of 1" thick steel shell increased by 8 F for each burn as previously calculated. When no ice is available in the ice condenser the average shell temperature increases by approximately 17 F per burn. There are a total of seven burns, two with ice and five with no ice. The total temperature rise in the shell is estimated to be 101 F (2 x 8 1 l l x 5 x 17). This corresponds to a total deposition in the wall of about 6.3 x 106 stu. I l It was assumed that this amount of heat is added to the containment shell over a 200 second time interval which i corresponds to a heat flux of app-oximately 7520 Btu /hr-ft 2. The TAP-A computer program was utilized to compute the transient
=
VIII-12 i L
temperature distribution in the shell. The maximum temperature difference in the shell from inner surface to, outer. surface was approximately 32 F. The corresponding temperature difference from the inner surface of the shell to the center of the shell was roughly 22 F. l r f I
/
4 l VIII-IS.
SECTION IX CONCLUSIONS A. Overview TV 45311 eves that, given a degraded core accident with hydrogen generation, ignitien of that hydrogen will occur eventually. In order to mitigate the potential gross impact of such an event, TVA has designed a controlled ignition system which we believe will burn the hydrogen at combustible levels and therefore mitigate the adverse impact of an uncontrolled combustion at high hydrogen concentration inside Sequoyah's containment. _ B. Ignition Source Effectiveness TVA conducted preliminary tests to obtain reasonable assurance that the igniters being considered would indeed ignite hydrogen at low combustible concentrations. The details of this testing are discussed in Section II and Appendix M. TVA found that the GM AC model 7G glow plus could ignite hydrogen concentrations at 12 volume percent and less. These results allowed TVA to proceed with the design of the interim distributed ignition system (IDIS) and the igniter assembly enclosure. Once the design was complete, TVA, through Westinghouse, contracted with Fenwal Laboratories to conduct larger scale tests on the g:: entire igniter assembly under various environmental conditions of pressure, temperature, steam, and flow past the igniter. These tests, which were conducted between September and Novtaber 1980, IK-1
are discussed in detail in Appendix N. Also discussed in Section II and Appendix L are the 45 igniter assembly locations in containment at Sequoyah unit 1. These igniters have been located to be able to burn hydrogen in all three containment compartments and the majority of enclosed spaces. TVA believes that the location and number of igniters is sufficient to guarantee that hydrogen will not be able to randomly collect and form high concentration pockets. To ensure that the igniters are available whenever hydrogen begins to be , released into containment, TVA is proposing to modify the emergency operating procedures (E0Is) as discussed in Section II and Appendix R. The proposed procedures instruct the operator to manually initiate the IDIS immediately after all automatic LOCA equipment. The igniters then continue to operate until the unit reaches cold shutdown. TVA has not been able to identify any
. accident scenario that becomes worse due to the operation of the igniters.
C. Containment Heat Removal j As discussed in Section V, the Sequoyah containment is equipped j with systems that actively remove heat generated inside containment and also inherently aid controlled burning by the IDIS. Although these systems were not originally designed to aid the IDIS, TVA feels that their operation adds significantly to the ability of the containment to withstand this postulated event. i-
- II-2 L
Two of the three containment heat removal systems actively remove heat frem the containment atmosphere, the ice condenser and the containment spray systems. The spray systems also promote mixing of the hydrogen in the containment volume. TVA expects good mixing of the containment air volume mainly due to the third containment heat removal system, the containment air return fans, which each take auction from the upper compartment and dead ended volumes and force the air into the lower compartment at a rate of 40,000 cfm. The mixing effect by the spray systems and the containment air return fans promote an even distribution of the hydrogen through large volumes in con'tainment. In order to monitor the hydrogen concentration inside containment, TVA has installed redundant hydrogen analyzers. These analyzers are discussed in Appendix K. They have two sampling points: one in the upper and lower compartments. TVA is not certain that the present monitoring system is adequate for the life of the plant. However, due to the mixing effect obtained from the spray system and the containment air return fans and the planned continuous operation of the igniters, we feel that the gross hydrogen concentration reading obtained from the present system is setisfactory for the interim until the decision is made to install cn improved system. D. Environmental Effects TVA has analyzed, at least in a preliminary fashion, the g e environmental effects of hydrogen burning on the containment shell, containment interior structures, critical containment II-3 ~
1 systems and critical compenents. Detailed discussions of these analyses are contained in Section VIII and Appendicqs D through J. Preliminary results indicate that the temperature rise in the containment walls will not be significant and the containment structure as a whole can withstand sizeable local detonations. TVA analyzed the effects of local detonations on interior structures (Section VIII.C) and found that the differential pressures generated by these local detonations would not be more severe thaa the major accident loadings for which these structures are presently designed.' Based on this analysis, it is our best engineering judgment that these structures would survive a substantial local hydrogen detonation with only minor concrete cracking. The effects of hydrogen burning on essential equipment
~
are much more difficult to analyze. Preliminary results indicate significant temperature rises across vital equipment. This is contradicted by the physical results experienced at TMI-
- 2. Thermocouples, which by desig'n are one of the most sensitive pieces of instrumentation, at TMI-2 experienced only a 50 F rise concurrent with the ignition of the hydrogen. To help resolve the discrepancy between analytical and actual results, TVA included representative samples of vital equipment inside the test vessel at Fenwal Laboratories during phase 2 of the hydrogen burning tests.
E. Scope of Events TVA selected five low probability, but severe consequence, l IX-4 i
^
i - , - . , , _ .
accident scenarios'to represent the spectrum of hydrogen producing events from metal-water reactions (see Volume 1, Section 5). From these five originals, three were input into the MARCH computer code. From the computer analysis (see Volume 1, Section 6), TVA selected hydrogen generation rates from the S2 D case, small break LOCA, as the representative basis for determining containment response characteristics. The Offshore Power Systems CLASIX code used the output of the MARCH code for the2S D case in their analysis of the ice condenser centainment. The CLASIX code is discussed in Appendix T. The CLASIX results for the S D2 case assuming hydrogen burning at 10 volume percent indicate that containment pressure is maintained below the containment design pressure by the use of igniters. The results of the CLASIX analysis is contained in Appendix U. F. Margin Available Inherent in any type of simplified analysis are factors which l could not be or were not included in the final results. Several factors which TVA feels provide conservative margin for the CLASIX code are: 1 The CLASIX code does not take into account the heat sinks available in the containments walls and interior structures. TVA feels that 'this causes CLASIX to produce such higher peak containment temperatures than what would be produced in an actual hydrogen burn. l l 2. The CLASIX code does not model the flow from the upper l IX-5
I l l compartment to the lower compartment created'by the cooling ) and contraction of the air in the lower compartment. The contraction of the air in the lower compartment could cause l a differential pressure much larger than that created by the air return fans. This type of pressure would create significant additional inflow of air from the upper compartment which would further aid containment heat removal. G. Uncertainties Inherent in any simplified analysis are certain factors which cannot be quantified. Throughout this entire analysis, TVA has had to make very basic assumptions as to the accident scenario and critical parameters. Several factors, however, are still fairly uncertain. The most prominent factor is that until NRC rulemaking is resolved, the accident scenario that will become f the design basis is undetermined. Until the accident scenario is specified, results obtained from the MARCH code, which assumes particular modes of core failure, may or may not give conservative hydrogen generation rates. The hydrogen generation rate is the single most critical factor in determining the structural capability of the containment and the usefulness of the hydrogen igniters. Another less central, but important uncertainty is the effect ' the heat sinks have on temperature and pressure peaks during hydrogen burns. TVA feels that the massive heat sinks help to reduce the peaks, . but this has yet to be quantified. Also, the l IX-6
i ? i effect of steam on the upper and lower hydrogen combustible limits and to a lesser degree the effect it has on flame
- propagation has not been quantified. TVA hopes to be able to 1
reduce uncertainties in all these areas through testing in the short term with the Phase 2 Fenwal tests and in the next 18 months with more definitive studies. ~ ' t 5 i k i t i l \ t t - e-1 1 1 i I t IX-7
APPENDIX A COMPUTER CODES The computer models used in the TVA hydrogen study include both primary system and containment thermal hydraulic codes. These tools are described in Volume I, sections 4.2.2 and 6.2. In addition, TVA has contracted with Westinghouse to study the Sequoyah containment with the CLASIX computer code. The CLASIX code is described in i Appendix T of Volume II; Appendix V describes code verification e f forts . Due to the importance of hydrogen distribution and deflagration, an inhouse TVA effort has also been initiated to modify an existing TVA containment code to include all physical processes believed te be important to the hydrogen problem. This effort is long term and is not expected to yield results before the first part of 1981. . i l. { l . L f i l i A-1 l
APPENDIX B EVENT DEFINITICN The potential exists for significant core damage from several hypothetical accident sequences. Each sequence involves many of the same events. For convenience, the Reactor Safety Study used alphameric designations for the major events that would alter the behavior of each accident sequence. These abbreviations have been retained and are listed below with the characteristics of each event. Event A: Large Break in the Reactor Coolant System Pressure Boundary This initiating event is a random rupture in the Reactor Coolant System (RCS) that creates a break area equivalent to that resulting from a 6-inch diameter pipe rupture and ranging up to the full area of a double-ended rupture of the largest Reactor Coolant System pipe and causes rapid depressurization of the RCS. Also included are ruptures of the reactor pressure vessel itself that place no more stringent demands on the engineered safeguards than the double-ended cold leg break. Event B: Electric Power (EP) Operability of the engineered safety features (ESP) depends on the availability of electric power. Hence, this event is generally listed soon after all initiating events because of its importance to the accident sequence. This event represents the availability of ac power to the buses that furnish power to the ESF's. The principal B-1
l l i l l i components of the electric power system are comprised'of the offsite ac network and the ensite ac diesel generacors together with the do control systems. The definition of EP failure is failure to provide sufficient ao and de power for the operation of the ESF's. required to mitigate the initiating event. EP failure affects the structure of event trees through the dependence of other systems on electric power. Hence, failure of EP may imply consequential failure of other systems. The electric power system event is subdivided into two separate events - availability of direct current and alternating current. The direct current (de) system is designated system / event B1. The buses provide control power for ESF equipment, emergency lighting, vital inverters, and other safety-related de powered equipment for the entire plant. Similarly, the alternating current (ac) system is designated B2. Ac power is also required by all ESF's. t Event C: Containment Spray Injection System l The containment spray injection system (CSIS) delivers spray to the l l containment to scavenge airborne radioactivity (iodines in particular) from the containment atmosphere during the time immediately following the RCS break. In addition, the spray aids in reducing containment pressure, thereby' reducing containment leakage to the environment and lowering the probability of containment failure. The CSIS consists of redundant spray headers and pumps that deliver water from the refueling water storage tank (RWST) which contains approximately-B-2
375,000 gallons of water. Failure of the CSIS system is considered to be failure to delivery borated water from the RWST to the containment atmosphere at a rate at least equivalent to the full delivery from one of two containment spray pumps. A schematic of the containment spray system for Sequoyah is shown in figure B-1. I Event D: Emergency Coolant Injection
~
The emergency coolant injection (ECI) system is composed of four subsystems that operate in various combinations to provide emergency coolant for a range of RCS break sizes to preclude core damage. These four subsysteel are: (1) the accumulators *(ACC), (2) the high presure injection system (HPIS), (3) the low pressure injection systcm (LPIS), an.d (4) the upper head injection (UHI) system. The accumulators 1 discharge stored borated water into the RCS cold legs when the RCS pressure drops below the pressufe setpoint of the accumulator tank. The accumulator system is a passive sytem. The HPIS injects borated water from the RWST into the RCS cold legs by using redundant, electrically driven centrifugal pumps. The LPIS injects borated water from the RWST into the RCS cold legs by using a separate set of redundant, electrically driven pumps. The UHI system is passive and discharges borated water from the UHI accumulator directly into the upper reacter vessel head when the RCS pressure drops below the UHI setpoint. ECI failure is (1) delivery of less borated uter than would result . l from the discharge of two accumulators into the RCS cold legs immediately following a large pipe break; (2) delivery of borated water at a flow rate less than the design output of either one I 3-3
f A X " AAAAAA K l AAAAAA X " AAAAAA N K N
\ N AAAAAA SPRAY HEADERS UPPER COMPARTMENT , m ICE CONDENS ER w
i-
. R.W.S.T.
I LOWER COMPARTMENT ' CORE
~ g C00 LING' -- -- -- --
CONTAINMENT RESIDUA L SPRAY MP HEAT HEAT REACTOR- + o __ __ EXCHANGERS
}
EXCHANGERS v4 ) g' )( SPRAY . I RESIDUAL PUMPS
' HEAT REMOVAL -
l PUMPS 3 r p 3 7 JL J L . Y , T I T I ( [ # ( FIGURE B-1
4 charging or one safety injection pump to the RCS cold legs (starting at about 30 seconds following a large pipe break and lasting 1/2 hour); or (3) failure to deliver UHI accumulator water into the upper head. Failure of the passive injection systems is less likely than failure of the active systems. Figure B-2 lists parameters of the Sequoyah ECCS system. Event E: Emergency Cooling Functionability Emergency cooling functionability (ECF) relates to the probability of failing to cool the core due to accident phenomenology even though the ECI functions properly. Several possibilities exist for failure of the emergency core cooling system to reflood and cool the core even though emergency coolant is delivered to the primary system cold legs. I' d are steam binding, failure of core supporting structu. resulting in an undefined coolant flow path within the vessel, and an uncoolable core geometry as a result of blowdown loads and subsequent thermal-mechanical distortion of the core. l I Event F: Containment Spray Recirculation Containment spray recirculation system (CSRS) provides for the recirculation of containment sump water through heat exchangers to
~
l spray headers inside the containment for the purposes of pressure control and the removal of radioactivity and heat from the containment. This system is composed of four trains, each of which contain a separate pump, spray header, and heat exchanger. Failure of the CSRS is considered to be a flow rate less than the equivalent normal output of two recirculation spray pumps for the B-5
, EMERGENCY CORE COOLING SYSTDi COMPONENT PARAMETERS Cc penent Parameters Cold Leg Injection Accumulators Number 4 Design Pressure, psig 700 Design Temperature. *F 300 Operating Temperature, *F 60-150 Normal Operating Pressure, psig 660 Minimum Operating Pressure, psig 600 Total Volume fc 1350 each Minimum 'a'ater Volume, f e 925 each volume N2 gas, fe 425 3cric Acid Concentration, nominal, ppm 1900 Inleakage Al' arm Sounds, ft 14 Relief Va'.ve Setpoint, psig 700 Upper Head Injection Accumulators Number 2 e Design Pressure, psig 1800 Design Temperature, *F 300 Operating Temperature, "F 70-100 Normt; Operating Pressure, psig 1350 Minimum Operating Pressure, psig 1300 3
Total Volume, ft 1800 each W eer Vol'ume, ft 1800 Volume N2 gas, t 1800 Boron Concentration, ppm min. 1900 Relief Valve Setpointi psig 1800 l ! Centrifugal Charging Fu=ps Number 2 Design Pressure, psig 2800 Design Temperature, *F 300
- Design Flow Race, gpm 150 Design Head, ft. 5800
,. ,
- Includes miniflow FIG 3tE 3-2 3-6
. - _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ .~
EMERGENCY CORE COOLING SYSTDi COMPONENT PARAMETERS Component Parameters Centrifugal
- Charging Pumps (Cont'd) Max. Flew Race, gpm 550 Head At Max. Flow Rate, ft. 1400 Discharge Head at Shutoff, psig 6000 tMotor Rating, bhp 600 Minimum Starting Time sec. 6 Safety Injection Pumps Number 2 Design Pressure,'psig 1700 Design Temperature, 'F 300 D,esign Flow Race, gpm 425 Design Hea4, ft. 2500 Max. Flow Rate, gpm .650 Head At Max.. Flow Rate, ft. 1500 Discharge Head, psig 1520 l tMotor Racing" bhp 400
_ Minimum S tarting Time, sec. 6 NORMAL OPERATING STATUS OF EMERGENCY CORE COOLING SYSTEM COMPONENTS FOR CORE COOLING Number of Safet'y Injection Pumps Operable 2 Number of Charging Pumps Operable 2 Number of Residual Heat Removal Pumps Operable 2 Number of Residual Heat Exchangers Operable 2 Refueling Water Storage Tank Volume, Cal. 370,000 Boron Concentration in Upper Head Injection Surge Tank,' min ppm l',900 Boron Concentration in Refueling Water Storage Tanks, minimum ppm 1,950 j Boron Concentration in Cold, Leg Accumulator, minimum ppm 1,900 i Boron Concentration in Upper Head Injection Accumulator, minimum ppm 1,900 Number of Accumulators 6 l Minimum Cold Leg Accumulator Pres'sure, psig 600 Minimum Upper Head Accumulator Pressure, psig 1300 Minimum Cold Leg'Accumulater Water Volume, ft 3 925 Minimum Upper Head Accumulator Water. Volume,'ft - 1800 l System Valves, Interlocks, and Piping Required for the Above Components which are Operable , All l nmnw ami ~.u \ .
- ~ _ -
first 24 hours or the normal output of one recirculation spray pump thereafter. Figure B-1, although illustrating spray injection, also shows diagrammatically the recirculation system. Event G: Containment Heat Removal System The containment heat removal system (CHRS) provides for the removal of containment heat by passing service water through heat exchangers in the CSRS. There are four heat exchangers in the plant, one for each of the CSRS trains. To successfully remove heat, service water must flow through a heat exchanger located in an operating train of the CSRS. Failure of the CHRS is the operation of less than two of the four containment spray heat exchangers during the first 24 hours and operation of less than one of the four heat exchangers thereafter. Event H: Emergency Core Recirculation System (ECRS) When ECI has succeeded, the continuity of emergency core cooling is , provided by the recirculation mode of the safety injection system, l l which is realigned to take coolant from the containment sump. Alignment into this recirculation mode of operation occurs approximately 1/2 hour after the design basis large LOCA when the RWST nears depletion. Operation action is required to switch the valve arrangements to the appropriate configuration for recirculation. The initial alignment for recirculation is for the safety injection l pump to deliver into the RCS cold legs only. However, unlike the injection phase, the delivery of fluid into the RCS hot legs is not considered as failure. The ECRS should be aligned to deliver into B-8
the RCS hot leg piping to help avoid potentiti accumulations of residue or debris in the reactor vessel, which may result from continuous boiling, within about one day after a large LOCA initiated in the cold leg piping. Failure of the ECRS is defined as failure to inject into the RCS from at least one safety injection pump. Failure to realign to hot leg delivery is also considered failure. Event K: Reactor Protection System Reactor protection system failure is conservatively defined as the failure of more than two full-length control rod assemblies to insert into the core within approximately 30 seconds after the initiating event. The failure of the required control rod assemblies to insert may be caused by electrical fauits in the signals or equipment required to gravity insert the rods into the core, or by wechanical faults that result in a hangup of more than two full-length control rod assemblies. Event L: Auxiliary Feedwater System ( AFWS) The auxiliary feedwater system supplies, in the event of less of the main feedwater supply, sufficient feedwater to the steam generators i to remove primary system stored and residual core energy from decay heat. It may also be required .to provide cooling during a small break LOCA by maintaining a water head in the steam generators. Sach system has two electric-motor-driven pumps and one turbine-driven pump. Each electric pump serves two steam generators, the turbine all four. Failure of auxiliary feedwater delivery is considered to be less than 3-9
i the full delivery from one full-size or two half-size electric-driven feedwater pumps or the equivalent ficw from the full-size steam-driven auxiliary feedwater pump. The period of demand and operation of the AFWS is about 1/2 day for a small break LOCA. Event M: Secondary Side Relief Valves The secondary side rel.ef valves vent steam generated from the water provided to the steam generators by the auxiliary feedwater system. This steam is vented to the outside atmosphere via power-operated relief valves or safety valves to augment RC3 heat removal. Failure of this system could result in steam generator overpressurization and failure or inadequate primary system heat removal for accidents where the steam line isolation valves and atmospheric dump valves cannot be opened. Figure B-3 details the valve parameters for Sequoyah. ~ Event 0: Containment Emergency Safety Features The containment emergency safety features are comprised of the containment spray injection system, the containment spray recirculation system, and the containment heat removal systems. This event is simply a general heading for coupling the three components either as a general failure or success. Event P: Primary Side Reller Valves Open This event represents the opening of the RCS pressurizer safety or safety and relief valves to limit the rise in the reactor coolant-pressure immediately following the initiating transient event. Not all anticipated transients require operability of the safety valves B-10
i PRIMARY AND SECONDARY SAFETY RELIEF VALVES DESIGN DATA Accumulation Blowdown Valve MK # Set Pressure Percent PSIC Percent PSIG Max. Set Pressure 47W400-101 1064 10.3 1174 96 1021 1075 47W400-102 1077 9.0 1174 96 1034 1088 47W400-103 1090 7.7 1174 96 1046 1102 47W400-104 1103 6.4 1174 96 1058 1115 47W400-105 1117 5.1 1174 96 1072 1132 Valve Design Considerations - Five safety valves per steam generators. Each set of safety valves should have a minimum combined capacity of 3,917,000 PPH. Nhximum actual capacity of any one safety ' valve shall not exceed 890,000 PPH l at 1,100 psia. b Valve settings consider a 21 psi pressure drop from the steam generator outlet to the most remote valve inlet. The balanced safety valve design and sonic flow limiting orifices assure set points independent of back pressures. Valve set tolerances are a maximum of +1 percent. Maximum accumulated pressure equals , 1,100 psia times 110 percent minus 21. psia which equals 1174 psig. Valve 47W400-101 has a maximum set pressure of 1075 psig, which is 10 psi below system design pressurt . Valve 47W400-105 has a maximum set pressure of 1132 psig, which provides a minimum of 3.7 percent accumulation. l I i PTGURE B-3 < B-ll
due to the surge capacity of the pressurizer to mitigate small pressure surges within the reactor coolant system.For more severe transients, such as those involving failure of the RPS to terminate core power, when required during a transient, the operability of the pressurizer safety valves would be required to prevent a rupture of the RCS from the resultant energy and pressure pulse. Typical parameters for these valves are shown in figure B-4. Event Q: Primary Side Relief Valves Reclose The RCS pressuriser safety / relief valves that open must reclose to prevent a discharge of an excessive quantity of coolant frem the RCS. Otherwise, a valve sticking open following the transient event of interest would result in a loss of coolant event siallar to the event that occurr-ed at the Three Mile Island sacility. Event R: Reactor Pressure Vessel Rupture This event considers the gross failure of the reactor vessel in which coolant loss exceeds that of a double-ended cold leg break. Event S.,: Small Break LOCA (2"< D $; 6") This event considers the range of small breaks in the RCS with a break area equivalent to that of a 2-inch diameter pipe up to and including a 6-inch diameter pipe. The separation of small breaks into two categories is due to the difference in mitigation procedures. For a break in this range, accumulator discharge and high pressure injection is required to arrest the fuel-clad temperature rise and to preven? inadequate core cooling. B-12
PRESSURIZER VALVES DESIGN PARAMETERS Pressuri:er Spray Control Valves Number 2 Design pressure, psig 2485 Design temperature, 'F 650 Design flew for valves full open, spm , 800 Pressurizer Safety Valves Number 3 Maximum relieving capacity, ASMI rated flow, Ib/hr (per valve) 420,000 Set pressure, psig 2485 (+1.0%) Fluid Saturated steam Backpressure: Normal, psig 3 Maximum during discharge, psig ' 500 Full lift pressure .
- <3% of set pressure Blowdown <5% below set pressure RCS pressure at the reactor coolan.c <110% of set pressure pump Jischarge when valve 12 at full .
lift (including pressure drop between safety valve and reactor coolant pump) Pressureizer Power Relief Valves
~
Number 2 Design pressure, psig 2485 Design temperature, 'F 680 Relieving capacity at 2350 psig, Ib/hr (per valve) 179,000 Fluid Saturated steam FIGUEE 3-4 3-13
Event S : Small Break LOCA (1/2"s;D di2") 3 mall ruptures of the RCS equivalent to the area of a 1/2-inch diameter pipe up to and including 2-inch diameter are included under this event. This event requires ECC injection via the high pressure injection system. Event T,: Transient Event Initiated by Loss of AC Power-This event can be characterized as a failure in the plant's electrical network which results in a transient being imposed on the RCS and core that (1) leads to a demand for the operation of the RPS to cause a trip of the reactor control rods to terminate power generation, and (2) requires operation of the plant normal or alternate heat removal systems to ensure adequate cooling of the reactor core. Event T3 : Transient Evrnt Initiated by Loss of Main Feedwater In this event, a failure of the main feedwater system imposes a loss of heat removal transient on the RCS. Overheating and pressurization of the RCS may result in posing a demand on the relief valve system. The same conditions apply on the RPS and alternate heat removal systems as stated in Event T
- A Event X: Air Return Fans The priaary purpose of the air return fan system is to enhance the ice cordenser and containment spray heat removal by circulating air l Trom the upper compartment to the lower compartment, through the ice condenser, and then back to the upper compartment. The air return fans also serve to mix the hydrogen generated during the design basis accident precluding concentration pocketing in the containment. The B-14
system is illustrated in figure I-2. Failure is assumed to be capacity less than the equivalent of one fan. Event 2: Ice Condenser The ice condenser containment relies on the melting of a minimum amount of ice to limit the containment pressure in the event of a large LOCA or other event that releases energy into the containment structure. The ice condenser is assumed to meet the technical , specification surveillance requirements for minimum poundage of ice. i i 5 I l l l l l
/
i
+ 9 i
B-15 2
APPENDIX C A DESCRIPTION OF PHENOMENA WHICH MAY AFFECT THE POST-CORE-DAMAGE-EVENT DISTRIBUTION OF HYDROGEN IN THE SEQUOYAH CONTAINFINT I. Purpose and Scope This report provides a qualitative description of the subject phenomena. While quantitative information is not extensive, the report does provide an evaluation of the relative i= pact of each phenomenum. The effect of hydrogen combustion inside a reactor containment is dependent on various factors, the most important of which are: (1) the total amount of hydrogen burned; (2) the rapidity of combustion (including detonation); and (3) the concentration prior to the combustion. These factors affect the transient and longer term pressures and temperatures that result from combustion. It is, of course, easiest to analyze the results if the pre-combustion concentration in the whole containment (or in a major region) is uniform over the whole volume. If concentrations are not uniform, the results of combustion may exceed or be less than those estimated for uniform
, concentrations. A nonuniform concentration (such as " hydrogen pocketing") might lead to a condition where detonation, large C-1
local pressure and temperatures, or earlier combustien might occur. On the other hand, a nonuniform concentration =ay involve distributions where a significant amount of hydrogen is in regions with a concentration less than the combustible limit leading to less total hydrogen burned than predicted by uniform distribution estimates, and ccrrespondingly reduced bulk pressures and temperatures. II. Geneaal Overview A. General Hydrogen Circulation For essentially all events of interest, hydrogen is released at high temperature (approximately 300 - 700 F) from the reactor coolant system. The RCS is located in the central area of the lower compartment. This release may be continuous over a period of time (with the major fraction of the release occurring over a time period as short as 15-30 minutes or as long as several hours), if the initiating event is a pipe break or stuck open valve. On the other hand, the release may be in rapid, sporadic bursts if the RCS is essentially intact (releases would be by way of relief and safety valves or RCS vents). In either case, the hydrogen is accompanied by steam, the amount of which depends on the particular event. { The release of these gases to the local compartment increases compartment pressure, forcing flow from the lower to upper compartment. Condensation and cooling in the ice ' condenser (when ice is present) and upper compartment (due
-C-2
to sprays) also forces flow from the lower to upper compartment. Air return fans force flow from the upper to lower compartments. In addition, the fans take suction from the top of such lower compartment subcompartments as the steam generator and pressurizer doghouses and return this flow to the general spaces of the lower compartment. Circulation within each subcompartment is quite vigorous. Phenomena causing this circulation include turbulence due to containment spray (uprar compartment) and flow through compartments; natural circulation due to different temperatures of gas, walls, ceilings, floors, and equipment; turbulence induced by jet effects from RCS breaks and vents; and, if operable, flow from fans in HVAC systems. Because of the significant pressure and temperature transients that i exist for events of interest, natural convection alone is l l sufficient to assure that no major stagnant regions exist for at least many days after the event. B. Mechanisms Affecting Distribution to Hydrogen in Containment Several phenomena have been postulated as establishing or minimizing differences in hydrogen concentration in different regions of containment. These are summarized l below and discussed in more detail in the body of the report. l c-3
- 1. " Jet" or " plume" effects - Hydrogen is released to containment mixed in staam. The concentration in the release jet or plume may be higher than the bulk concentration (at least during the time periods of most importance), leading to the jet being a region of high hydrogen concentration.
- 2. Steam condensation effects - Steam is an important constituent of the post' aent containment environment. Where massive condensation takes place (in the ice condenser, due to upper compartment sprays, or regions with significant heat sinks), the removal of steam increases the relative concentration of hydrogen.
- 3. Fan-forced convection - The air return fans force flow from the upper compartment and from lower compartment high points to the lower compartment. This flow tends to induce a counter flow through the ice condenser and bypass flow paths back to the upper compartment.
This tends to even out the hydrogen concentration in the major compartments (and provides oxygen to the lower compartment). 4 Spray-induced turbulence - The upper compartment sprays remove steam, thereby inducing flow through the ice condenser from the lower compartment. This effect Ch
is much more pronounced, of course, af ter the ice has melted out. In addition, the spray flow, by momentum transfer, thermal gradients, and condensation, induces strong turbulence in the upper compartment, assuring nearly perfect mixing of the upper compartment (this effect is supplemented by the hydrogen mixing ducts in the dome).
- 5. Natural circulation - Natural convection occurs on both large and small scales. Hydrogen / steam releases (being both light and hot) rise rapidly. Heat sinks such as structures and equipment provide cooling in i
the early stages of the event, creating strong natural convection flows in each subcompartment, promoti,ng mixing in the compartment. Therefo.*e, while natural convection causes gas with high concentrations of hydrogen to rise to the top of the compartment, the turbulence induced by this flow and by cooling from heat sinks tends to break down these concentrations. (
- 6. Diffusion - cas diffusion tends to even out the l
l hydrogen concentration within each subcompartment. Diffusien is probably not of significance however, because of the presence of turbulence which, postevent, is a much stronger mixing force. C. Initial Containment Environmental Conditions C-5 l
For loss-of-coolant accidents, significant mass and energy releases to containment occur prior to formation and release of significant amounts of hydrogen. Therefore, such important containment systems as the ice condenser, containment spray, and the air return fans have come into play prior to hydrogen release, assuring that the release is into a dynamic, turbulent environment. The lower compartment would be expected to be at a pressure slightly above that of the upper compartment, and to have an atmosphere composed of more muter vapor / steam and less air (in the early stages, the lower compartment is pressurized principally by steam and temperature, while the upper compartment is pressurized by air from the lower compartment and temperature). While the lower compartments' bulk oxygen concentration is somewhat depleted, sufficient oxygen exists in most regions to support combustion. IV. Release Point Effects For the events of interest, the release point for hydrogen is in the lower compartment. The most probable region is the main annular volume between the reactor vessel cavity wall and the crane wall. Release points night be a small hole in a large pipe, a broken small pipe, the pressurizer relief Lonk (relief, safety, and vent valve discharges), or the reactor vessel vent to containment. While there may be mechanisms which break up flow from the ; c-6 I
release point (such as release under water and impingement on structures and equipment), in general, the plume of released gas can rise relatively unimpeded to the top of the lower compartment. However, as the break jet spreads and transfers O h momentum to the surrounding mixture, it entrains air, thus diluting the initial ccncentration of hydrogen. Examination of MARCH results for the S2 D transient indicates the hydrogen concentration at the break plane peaks initially at 24 percent af ter which it drops to 16 percent and remains constant throughout the remainder of the transient until core melt and vessel failure occurs. The exit plane concentration, although near the detonable limit, contains no air, and ignition is
'therefore impossible., Air entering the jet by turbulent mixing and hydrogen leavin;; tf( jet by diffusion renders the jet well below the detonation limits within a few feet of the break.
Estimates for a turbulent jet indicate that in pure air the lower flammability limit is reached less than 30 feet from the break plane for a jet of 24-percent hydrogen in steam. If any mechanisms exist to strip only the steam, then the temperature is correspondingly decreased, reducing the jet and buoyancy forces, and preventing signficiant stratificatior. Significant mixing is therefore expected, particularly in tne margins of the pl'ame. Although the jet may be !.n the flammable limits, it is not expected to be detonable .because of the mixing mechanisms present. The gas mixture rises to the top of the compartment where it spreads laterally with a significant portion entering the ice C-7 w
condenser inlet doors. Another portion enters the deghouses (steam generator or pressurizer compartments). Since the doghouses are dead-ended volumes roughly annular in shape with a relatively low discharge flow rate into the hydrogen mixing system, the flow into the space will drop because of an initial filling which pressurizes the volume enough to resist the remaining mcmentum of any jets present. Early in the hydrogen release transient, lower concentrations than found in the lower compartment general areas will be encountered in the doghouse volumes due to the low ficw rate through these volumes and due to the initial air present which, unlike the lower compartment air, was not swept through the ice concenser by the initial break steam flow. Later, gas mixtures,in the doghouses will be at concentrations comparable to the lower compartment space as a result of natural convection mixing and any jets entering the volumes over a long period of time. This mixing interchange is relatively slow and is due to inlet turbulence, condensation, hydrogen mixing system flow, and natural circulation, with the only exception being direct jet flow which would cause a more rapid mixing. While a region of higher hydrogen concentration may be possible, its size and concentration will be limited by the mechanisms discussed above. In addition, mixing in the general lower compartment spaces due to air incoming from the air return fans and exiting through the ice condenser promotes elimination of hydrogen stratification. TVA does not believe that pockets of combustible mixtures with hydrogen concentrations significantly C-8
l above the compartment mean concentration will occur. i Q ito If the release point is below the sump water level, the steam may be at least partially removed and the hydrogen cooled. The hydrogen would then be released at the pool surface with greatly reduced buoyancy. The air flow frem the air return fans, being colder than bulk compartment air, tends to sweep this pool l surface, providin6 a strong dilution mechanism as well as breaking up vertical flows due to buoyancy. Because of the mixing effects just described on the hydrogen and Q 12. I h steam distribution in the lower compartment, the potential is considered small for preferential flow to portions of the ice bed during degraded core transients. The original design of the ice condenser considered maldistribution problems from nonuniform ice melting, small breaks, and jet impingement on the lower inlet doors which could result in a channel being burned through the i ice bed. At that time, Westinghouse performed various l calculations to model possible nonuniform ice bed melting and showed that burnthrough was t.st a problem and that considerable margin remained in the ice bed. Independent of the Westinghouse l l efforts, Battelle Northwest Laboratories evaluated ice condenser l performance after being connissioned by TVA and Duke Power and confirmed that no burnthrough would occur. l l i Recently, Offshore Power Systems evaluated ice condenser l performance giving special consideration to hydrogen burning in the lower compartment. Because the Mach number of the flame l l l c-9
front is =uch less than one, at any given time during a hydrogen burn the pressure everywhere in the lower compartment will be uniform so that the differential pressure across all the ice condenser lower inlet doors will 'e the same and so will the volumetric flow. However, they can be significant temperature differences, at any given time, in the flow entering the ice condenser if perfect mixing is not assumed. To evaluate the effect of the nonuniform temperature, a hand calculation was performed. Using the conditiens in the lower compartment just prior to the first burn in the S pD base case, a constant pressure burn was assumed with no air return fans operating. The flame fecnt was assumed to propagate across a diameter of the lower compartment so that over'the entire time of the burn, the hot gas behind the flame front always entered one of the doors and the colder air in front of the flame always entered another door with the balance of the doors having a temperature based on the position of the flame front. Assuming that both gas streams were cooled to 32 F, the amount of ice melted by the not stream exceeded that melted by the cold stream by less than 105. Although the temperatures of the two gas streams were greatly different, so were the densities, and consequently the heat capacities, so that the two effects tended to offset each other. Since 25% of the ice remained after all nine burns in the 3 2D base case analysis, there is adequats margin to overcome the potential 10% nonuniformity in ice melt. Of even greater ; i i significance, the reduced ice case in the sensitivity studies ; I l C-10 l J
(see case JV904 in Appendix U), where all the ice was gone after the second burn, indicated a peak pressure of only 26 psig compared to the containment capability of 46 psig. Thus even if burnout did occur any time after the second burn, for the base case, the containment would not fail. Q l 2. A-V. Steam Condensation Steam (or water vapor) not only dilutes the hydrogen, but also tends to suppress combustion. Therefore, if the steam is removed by condensation, the combustibility is increased. There are three strong mechanisms for such condensation in the containment: cooling in the ice beds, by the contai7 ment sprays, or by the lower compartment coolers. C} 5 The flow into the ice beds from the lower compartment g is composed of air from the return fans mixed with the RCS release flow. During its passage through the ice bed, the steam is gradually stripped out of the mixture, thus increasing the concentration of the noncondensible gases, including hydrogen. The steam is completely removed by the time the flow reaches the top of the ice bed, passes thr;ough the intermediate deck doors, and enters the upper plenum. (See Section IV.B for a discussion l l of the magnitude of this steam condensation effect.) During the relatively low flow rates through the ice bed after a small break LOCA, the mixing and dilution by volume in the ice condenser would tend to' prevent high concentrations of hydrogen i from being reached in the upper ice bed. In addition, there are no ignition sources located in the upper ice bed below the C-ll L___
intermediate deck doors. The flow out of the ice bed opens the intermediate deck doors and enters the upper plenum of the ice condenser. Here, the ice bed effluent may be distributed by some of the still-running ice condenser air handling units and is diluted in the relatively large volume of the upper plenum. Dilution, mixing, and the addition of moisture from the containment sprays occur in the upper plenum when the top deck blanket opens. To provide for controlled burning of the well-mixed effluent, four igniters are located around the upper plenum. (Note that numerous potential spark sources such as the air handling units are already located in the upper plenum which, without the igniters, might otherwise lead to uncontrolled burning.) . The flow out of the upper plenum opens the top deck blanket and passes into the upper compartment. The very large open volume of the upper compartment dilutes the ice condenser effluent while j the turbulence caused by the ecmbination of containment sprays and air return fans speeds the process. Three igniters are located around the containment dome to allow for controlled burning should the overall mixture reach combustible I concentrations. Therefore, the physical size of a potentially (gg high concentration region is limited to the ice condenser upper plenum and a few feet outside the plenum exit. Since such a region would compromisa only a small fraction of the upper compartment, since propagation of combustion into the well-sprayed upper compartment is suppressed by the falling spray C-12
droplets, and since there is little or no sensitive critical equipment in the area, strong combustions or even minor detonations will not endanger the plant. (See Appendix H.) The upper compartment sprays remove any excess moisture from the upper containment. However, if the flow through the ice bed is well cooled (the case when flow is moderate to low and substantial ice remains), the sprays can actually heat the air and add water vapor. Af ter the ice is gone, steam from the lower compartment is condensed in the upper compart=ent. While this can lead to increased concentration of hydrogen, the spray droplets plus the increased water vapor concentration at the higher temperatures involved (post-ice melt) should effectively suppress any substar.tial combustion. - The lower compartment coolers are not engineered safeguards; however, since they are believed to be capable of operating 1" a postaccident environment and their use would be desirab' 1 a degraded core event both for miring and heat removal, the
- effects are considered. The coolers will remove substantial amounts of steam, thereby increasing the hydrogen concentration.
l However, the turbulence of the exhaust promotes very rapid mixing and, within a short distance, nearly complete mixing. This has been observed when the fans were run in a steam environment. This results in limit'ed volume of higher concentration, limiting the possible effects of rapid combustion. If a sufficiently strong combustien or detonation occurred, the coolers could l suffer damage, but other equipment should not be affected. c-13 l L_
I VI. Fan-Forced Convection The air return fans exhaust to accumulator rooms from which the flow passes through the fan rooms and out through holes in the crane wall into the lower compartment main co=partment. Since this air is colder than the lower compartment, it tends to sweep the sump water surface. The major components in this region (RC pumps, RCS piping, steam generators, and RCS supports) interfere with this flow, causing turbulence and mixing, ensuring a well-mixed flow to the ice beds. The flow out the ice beds into the upper compartment is colder than the air in the upper c5feartment; therefore, it tends to move easily into
, the flow pattern established by the upper compartment sprays.
The air flow to the accumulator and fan rooms can also go (through holes in walls and floors) to the other compartments outside the crane walls (two accumulator rooms, instrument room, and raceway). These are deadended volumes whose only exhaust is through the mixing system ductwork back to the air return fans. These rooms therefore tend to contain air the same temperature and concentration as the upper compartment, but lag upper compartment trends because of the relatively low flow. Because the fan flow effectively blocks releases reca the RCS from entering the deadended volumes, there is no major hydrogen source in these rooms. The only source of hydrogen is evolution of hydrogen from sump water in the lower raceway. However, water only gets to the raceway after complecion of ECCS injection and C-lh
substantial ice melt. For small LOCA's with ECCS failure, the majority of hydrogen has been released prior to water entering l the raceway. The water that does get to the raceway has had a significant residence time in the main compartment, allowing dissolved hydrogen to be released. The hydrogen mixing system flow will keep the deadended compartment concentrations from building to high levels. The air return fans also pull flow recm the upper containment dome to prevent hydrogen buildup there. However, while the containment sprays operate, the induced turbulence is sufficient to provide rather complete mixing. The mixing system flow to the air returns is needed therefore only in the long term. The hydrogen mixing subsystem of the air return fan system also draws air from tha steam generator and pressurizer doghouses.- These volumes are not expected to be significantly different in concentration from the lower compartment as discussed previously in section C.VI. Concentrations will be below tne detonation limits but may be in the flammable regime. A burn initiated within the hydrogen mixing system would not transition to a detonation because of the low concentration. Any deflagration initiated in the skimmer system will travel through the piping exiting at one of several vent openings *r ...., the upper mixing plenum at the air return fan where it s'11 mix wit.' significant quantities of of air from the upper comparttar' Lefore passing the air return fan. assembly. This mixing will cool the combustion gases considerably since approximately 40,000 cubic C-15
feet per minute nor= ally enters from the upper compartment, whereas only 2300 feet cubic per minute normally enters from the mixing system. Studies indicate that the skitmer system piping would not be subjected to severe pressures, even if a deflagration occurred in the line. i Qlb VII. Spray-induced Turbulence The upper containment spray system consists of four subsystems, each witn its own spray header; two redundant containment spray systems and two redundant RHR containment spray subsystems. The containment spray subsystems provide 4750 gal / min each and the RHR subsystems 2000 gal / min each. The RER subsystems are intended for use in the long term, but could be used in the short i term for certain ECCS failures. The four headers are located
, between the 838' and 842' slevation, and spray the upper compartment inside the ersne wall.
The spray causes strong turbulence and mixing in the upper compartment due largely to momentum transfer between the spray droplets and the air. This turbulence has been observed during spray flow tests. Shear forces between the sprayed regior. and the relatively small unsprayed region promote mixing in the latter areas. VIII. Natural Circulation l The postaccident environment in the containment is a highly C-16
dynamic situation with pressure and temperature changes, high flows through most compartments, and heat transfer at walls, floors, ceilings, equipment, and structural elements. This leads 2mperature differences between gases and structures / equipment produce natural circulation flows. Such flows are present
- > most quiescent compartments due to the large heat capacity many of the structural elements.
This natural convection mixing is present for design basis events. For events beyond the design basis, it is augmented by the piston effect of pressure rise due to burning and by the relatively higher temperatures after burn. Since it requires realtively low temperature differences (a few degrees) to
, establish turbulent convection, strong local turbulence and -
mixing can be expected postaccident. For the major compartments, forced flow and spray-induced turbulences should predominate. For deadended volumes, natural circulation should provide sufficient mixing to prevent stratification. IX. Diffusion Diffusion would tend to reduce nonuniform concentrations. Although hydrogen has a relatively high diffusion rate, diffusion is nnt expected to have a sign.ificant effect on concentration differences because the much stronger turbulent mixing will predominate. Diffusion will enhance turbulent mixing by increasing mixinr. sithin eddies. Diffusion can play a substantial part on a microscopic scale, however, since for i hydrogen lean mixtures, it plays an important role in the combustion process. C-17
APPENDIX D NCNSYtOETRIC CONTAINMINT LOADS An evaluation of potential nonsymmetric loads on the containment shell and internal structures has been made. There are two basic phenomena that could result in such nonuniform loads. The first is due to burning in a local area. This would occur as a result of pockets of hydrogen collecting in one or mere restricted areas or a general burn that starts in a restricted area of the centainment. The second cause of local pressure perturbations would be due to a pocket of hydrogen reaching a detonable concentration and exploding. The potential for local pocketing and the effects of detonation are discussed in Appendices C and H, respectively. Discussions in this appendix will be limited to the cases involving hydrogen burning. To evaluate the potential for asymmetric loads at Sequoyah due to hydrogen burns, the containment was modeled using nine nodes. Six nodes were used to model the lower compartment inside the crane wall. The upper compartmant, reactor cavity, and an accumulator room comprised the remainder of the model. The model did not include other dead-ended volumes to reduce the overall volume in the lower portion of the containment which maximizes pressure differentials. The volumes of rooms in the lower compartment were taken from the Westinghouse TMD pressure studies for the Sequoyah Nuclear Plant containment. Because all of the burns in this study were postulated to occur in local areas of the lower compartment, these volumes considered in the analyses were reduced for conservatism. Only the , 9 upper reactor cavity was considered for free volume inside the D-1
I I biological shield wall and a dead-ended volume of 17,900 ft out of a 3 total of 94,000 ft was used. The volume of the ice condenser was neglected and the upper compartment volume was taken to be a million cubic feet (Table D-1). The actual upper compartment volume is 651,000 ft 3. The large upper compartment volume was chosen to maximise the computed differential pressures across the divider deck. Sensitivity studies were made to determine the effects of increasing the upper compartment volume on the differential pressure. The results are relatively insensitive to the upper co=partment volume. An increase of a factor of 10 had negligible effect on the differential pressure. Flow areas and loss coefficients were taken from the TMD calculations of asymmetric pressures due to large pipe breaks for the original plant design bases. Flow paths through the ice condenser were modified by reducing the available flow area out of the lower compartment by a factor of two. The asymmetric pressure loads were calculated with the SPA subcompartment pressure code. The SPA code models 2-phase, 2-component fluids and evaluates node conditions using the homogeneous equilibrium model. Superheated steam is treated as an ideal gas for temperatures outside.the range of the code's steam tables. The hydrogen burn was modeled as an energy input into the node or nodes where a local burr. was desired. The energy input was the equivalent of a 65-volume percent hydrogen mixture (which is extremely D-2
1 7 4 Table D-1 3~ Volumes Used in Appendix D Analysis Node- Description Volume (ft3) i 1 Lower Compartment 28,700 2 Lower Compartment 36,800 3 Lower Compartment 70,200 i
- 4 . Lower Compartment 38,800 5 Lower Compartment 36,800 i
6 Lower Compartment 25,100 7 Dead-Ended Compartment 17,900 l
- 8 Upper Reactor Cavity 15,300 0
. 9 Upper Compartment 10 -10 9 f i s i i i m' t m
>5 Af D.3 i
, _. , . . . ,..a. ..--v,.,--,.....
conservative) in any node studied. As shown in the figures, all of the hydrogen was assumed to burn. Steam was added to the burning co=partments in amounts that would be equivalent to the quantity produced by the combustion of the hydrogen. The air mass was not reduced to account for oxygen that would have been consumed in the reaction. hg Typical data on hydrogen burning in dry air shows the flame speed is generally accepted to range from 6 to 10 ft/sec. Velocities of this magnitude mean the energy release is relatively slow, particularly when compared to the energy releases in a large LOCA or main steam line break. This means that nonuniform loadings due to burns should not be as severe as loads already considered in the desit.: of the plant. Our evaluations show maximum differential pressures for localized lower compartment burns to be less than 11 psi. The 11 psi is based on a flame speed of 30 ft/see and assumes that the flame radiates in a spherical fashion from the center of the area. It has also been assumed that there was no hydrogen burning in other areas of the containment. These are extremely conservative l assumptions that maximize the pressure differential. If the flame speed is reduced to 10 ft/sec, the maximum differential pressure on internal structures is reduced to approximately 3 psi. In comparison, , the maximum differential pressure for a large LOCA is 14 3 psi. The 1 design values used for internal structures is 1.4 times the calculated differential pressure plus any margins required by the construction and design codes (American Concrete Institute, etc.). Figures J-1, i l 2, and 3 show the transient results for various burn velocities. l l It is concluded that nonuniform pressure ' effects for local hydrogen L D-4 i L
gahnaifu{l gifqt cifau q;nnif s o jrt
! . . : . . ... . .o m
[ . . . .
.o m .
. . . . . . . .g o
a m n . . s . m - g <n m . 38-o_ a
* . .b r
> p43m . .
os
.8c o
9 g p e:z 2 e
= <. g : .
s.c a=c e . . . . .
. o i
Z c . O M
. .e_. . .
.6o 0 -i e
m :. . m . . . .
. ap
. . . - c
' e .
..a a
n a . m
.o e . . . .
x . . . m
- : : .o Q .. . . . . . . . . . . . . .
.O
~ .
! w . : o ......,.
. a,.
2
- g C
03 72 42 12 81 51 21 9.D 60 3.0 00
)isp( ERUSSERP LAITNEREFFID i
i 1 l G-5
DIVIDER DECK DIFFERENTI AL PRESSURE c! es l m q .
- .. .Q U
- 9 NODE MODEL e
o BURN VELOCITY OF 15ft/sec C m 4 _. . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . 100% BURN FIFICIENCY - - g3
- ~T cu . : . . .
- v. . . . . .
na J .- .. ..
.........3...
g cg ; - C D o : : . U1 g _ .. .... . . . . . . . . . . . . ..: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . eg U3 p w - - Q w cg . my g, 0 - a
- . .J E g_
2
. . . . . Up
- n ca -
M oc . cal $ _' . . & ca. g c m ! ! : cn o . a . . . . . . . . . . .....
- q. . .
- a. .
g . . e' . . . . . . . 0.00 0.05 0.10 0.15 020 025 0.30 0.35 OAO DA5 0.50 TIME (seconds) FIGllRE D-2
gabttolftt{l 3766 Q 106JII q }11qif 8,
- . . 1
- l 0
art 9 1 X l c . . . w .
.t w . . . .
- e m g
% q<
m gn -
. . ny . .
.P l -
e oaP . m e 4_ q
)
gN~ O g !'i c:
.8 Pvw
- 2 =<E .-
g M i -
%' ez;Q 2g . -
e " ee- . .
.. . .dP :v $ ?
so N j m V j D .. .. .. . . . . . ... . . . . . . . .
. .Y e d 3 :
O . jrt
.. . .g l
b . . . g .
. .. ~
. : .. . . . .p d . . . . . .
m - : : j
% 5 m .
! .S D- . . . . . . . . . . .
9
'8h
. N. N t t t
t t
- t t t DLI 99 88 77 66 45 4.t4 33 22 11 00
)isp( ERUSSERP LAITNEREFFID G-L
burning are bounded by present design values. The modeling assumptions, flame speeds, and hydrogen concentrations used are extremely conservative. If hydrogen concentrations in the expected range of eight to fifteen percent had been used in the analysis, internal asymmetric pressure loads would be practically nonexistent. The localized pressure on the containment vessel was also calculated i assuming a burn in the dead-ended compartment. The energy of the burn was added to the compartment in one second. The other assumptions made were the same as discussed above. The pressure differential on qq the containment shell in the dead-ended compartment was 47 2 lb/in This is, once again, based on very con'servative assumptions. This pressure does not present a structural problem for the containment, as it is very localized. In addition, the lower compartment steel shell in the area of this pressure load has a adnimum thickness of 3/4 of an inch compared to the controlling upper compartment shell thickness of 1/2 of an inch. The pressure spike is also short in duration and, as stated earlier, extremely conservative. The effects of localized burns in the dead-ended compartments are concluded to be well within the structural capabilities of the containment. _f 5 f .! I D '
e APPENDIX E Table E-1 provides a list of components inside containment that could be required to function after a hydrogen burn resulting from a small break LOCA. 1 Shutdown logic diagrams and safety function diagrams were prepare 4 (with emphasis and detail on the functions occurring inside containment following the hydrogen burn) to determine the equipment f 1 needed for the event. The shutdown logic diagrams illustrate what functions must be achieved and what necessary or alternative paths may be available to reach cold shutdown following the event. The safety function diagrams provide further detail of what must be available to achieve the particular safety function. These diagrams were then - surveyed to identify. the equipment located inside containment which would be required to function following the hydrogen burn. The diagrams and, consequently, the listing, do not include check and manual valves in piping systems since their function is not believed to be impaired by the hydrogen burn. Also, the cables, splices, junction boxes, panels, limit switches, sensing lines, etc., associated with the listed equipment are also required (Table.E-2). Similarly, all components of the instrument loop for the-listed instruments which are inside contain' ment and necessary to, or could prevent, the function of the listed components are also required, but are not explicitly listed. Equipment. which needs to function only very early in the event (i.e., before the hydrogen burn) and cannot later prevent the function of needed equipment is not listed in the E-1
table. The table includes equipment whose function is redundant to other required equipment. Equipment whose function may only be desirable is not listed. The instrumentation and equipment is divided into four categories for evaluation: I - Temperature and pressure II - Temperature only III - Pressure only IV - No evaluation needed i l l l t E-2
Key for Table E
- H Evaluation for pressure and temperature environmental qualification required.
H Evaluation for temperature environmental qua:tfication only required. H Evaluation for pressure environmental qualification only required. H !4 - No further evaluation required, t 1 E-3 t
TABLE E-1 Equipment Function Category Justification of Category LT-3-148, -156, -164, Steam generator level input for H-1
-171, -172, -173, -174, control of AFW flow -175 -
Air Return Fans Containment pressure and H-1 hydrogen control H2 I-43-200, -210 Hydrogen analyzers H ll All equipment is located in the annulus with sensing lines into containmes.'. Installation of flame arrester will stop propagation of flame to equipment. FCV-43-201, -202, Allows air flow to hydrogen H-4
-207, -208 analyzers Valves fail open and loss of air to valve will not affect closure.
FSV-43-201, -202, Allows associated FCV to open H-4
-207, -208 ' Loss of these valves will not cause associated FCV to clor:e.
Ice Condenser doors Allows steam flow through ice i H-3 Temperatures due to hydrogen burn will bed have no e' :ect on the doors.' Require qualification for pressure only. LT-63-176, -177, Sump level for ECCS switchover H-1
- -178, -179 FCV-63-172 Closed position required for H-4 Not required to operate; must only proper ECCS flow path maintain closed position. All relays and controls for valve are outside containment. Hydrogen burn cannot cause the valve to open.
LT-68-320, -355A Pressurizer level H-1 TE-68-1, -24, -43, -65 Hot leg temperature H-4 The only equipment inside containment are the RTD's and cables inside conduits. Qualification of cables in conduits will resolve these sensors.
Equipment Function Cntegory Justification of Category TE-68-18, -41, ~60, -83 Cold leg temperature H-4 Same as TE-68-1, -24, -43, -65. Hydrogen Igniters Hydrogen control H-4 Fenwal test data and analysis have demonstrated durability. Core exit thermocouples Information for inadequate core . H-1 cooling TE-68, -373 through For reactor vessel level system H-4 Same as TE-68-1, -24, -43, -65. TE-68-386 FSV-68-394, -395, Reactor vessel vent valves H-1
-396, -397 Ice condenser seals Direct steam flow through ice H-4 Seals have been qualified for pressure requirements. Large heat sinks attached to seals will prevent damage due to high temperatures.
Penetrations X-003, Containment boundary, blind flange H-4 Penetrations have been qualified for
-111, -113. -112, -054, with 0-ring seal pressure requirements. Large heat sinks $'
-079A, -079B attached to the 0-ring seals will prevenf' damage due to high temperatures.
Electrical penetrations Containment boundary H-2 Penetrations have been qualified for pressure requirements. Airlock, Equipment Hatch, Containment boundary H-4 Large heat sinks attached to the seals and Personnel Airlock will prevent damage due to high tempera-Seals ture. Containment Isolation Containment boundary H-4 The containment isolation valves will be Valves in the required position prior to any hydrogen burn. All air supplies will be isolated and all relays and controls are outside containment with only power feeds to the valves (i.e., the valves cannot change position)~.
; TABLE E-2 Category H-1, H-2, H-3 Components Required Equipment for Equipment to Function LT-3-148, -156, -164, -171, Differential pressure. transmitter, sensing
-172, -173, -174, -175 lines, cable inside conduit, junction box Air Return Fans Flow switch, sensing lines, cable inside
- conduit, junction box, electric motor i
Ice Condenser Doors Top deck blanket, intermediate deck doors, lower inlet doors I l LT-63-176, -177, -178, Capillary tubes, differential pressure
-179 transmitter, cable inside conduit, junction box LT-68-320, -335A Capillary tubes, condensate pot, differential i
pressure transmitter, sensing line, cable inside conduit, junction box Core Exit Thermocouples Exposed cable, cable inside conduit, junction box
! FSV-68-394, -395, -396, solenoid' operator on valve, cable inside l -397 conduit, junction box ,
Electrical Penetrations Sealed penetration assembly i l ( l. l-E-6
APPENDIX F THERMAL EFFECTS ON COMPONENTS Burning hydrogen in containment produces atmospheric temperatures that' y greatly exceed the response for LOCA's and main steam line breaks. As a result of the severity of the temperature response, a realistic evaluation of this response and its effects on equipment is required. The considerable conservative margins inherent in present analytical codes used to evaluate plant response to class 9 accidents are excessive and require modification for u.9e in evaluating the true capabilities of equipment and the overall plant design. In the past, the computer codes developed to model degraded core events were used to perform consequence analyses in an attempt to determire an upper bound risk from nuclear plants. Detailed calculations and j understanding of the phenomena involved in a degracid core accident were not considered to be required ,in the consequence effort. I As a result of the intended use of the codes, the predicted core heatup rates and hydrogen generation and release rates were very conservative. Although the codes were never intended to be used for plant design, they are the only codes presently available. There has not been sufficient time since the TMI accident to thoroughly i understand all of the phenomena involved and develop new models. Development of new contaiment codes that model high temperature gas releases and hydrogen burning is underway, but completion of this work ( , is still sometime in the future. Therefore, a realistic temperature response curve based on a transient analysis for a degraded core event that is to be used for determining equipment survivability is F-1 >
unavailable since it is beyond the capability of present containment codes. , The limited amount of information presently available does offer some significant insights. While burn f.emperatures are generally high, the total amount of energy released during a burn is much less than is released for many other accidents that are part of the current plants' design bases. With moderate temperature rises, the capacity of the structural heat sinks alone is much greater than the energy released by a burn of all the hydrogen produced during a degraded core accident. For example, the TMI experience showed a temperature rise in instrumentation as a result of a hydrogen burn to-be on the order of 50 F. This is based on thermocouple readings observed during the accident. Thermocouples are at least as sensitive to temperature changes as other key equipment in the containment due to their purpose and size. The most important observation at TMI on the effects of a hydrogen burn was 'that sufficient critical instrumentation survived the event for the operators to monitor all important plant parameters and eventually cool down the unit. t Efforts are also underway to accelerate existing test programs and to l l initiate new ones that will increase the understanding of hydrogen combustion including the heat transfer processes during and after a burn. As the testing progresses and our understanding increases, we believe that realistic temperature profiles can be measured or generated by analysis and tP7se results supported by proof testing of 0 representative components. Recent testing dono at Fenwal by TVA (see Appendix N) and earlier by General Electric has shown that F-2
instrumentation, unshielded cable, and various equipment have the ability to survive repeated hydrogen burns and remain functional. Much of the equipment and instrumentation used as part of the test apparatus at Fenwal were off-the-shelf items without specici protection and qualification for severe environments. Much of the key inplant instrumentation is less vulnerable than items used at Fenwal. While additional information may show that certain modifications to components are desirable tv improve their resistance to high temperatures, it is not anticipated that major modifications will be required. t M a ?
^
q.'.1 f
.F-3
1 l APPENDIX G PRESSURE, SHOCK, AND TRANSIENT LOADS ON COMPONENTS The static pressure effects due to burning hydrogen will not prevent components from continuing to operate because the pressures to which they are qualified are greater than the accepted containment capability to withstand pressure. During a pressure transient, instruments which have a vented reference will indicate an incorrect low reading (due to the increase in their reference). Once the trantient is complete, the instruments will return to the correct reading. This scenario is confirmed by a review of steam generator pressure plots from TMI which show a low steam generator pressure during the time of the pressure transient which was initiated by the hydrogen burn. l l As discussed in Appendix H and Section C.IV, detonations are not j expedted but small pockets of higher concentrations may be postulated. 1 ( A local detonation of this type will have a fairly low impulse function (see Appendix H for a discussion of magnitude). Some- gg [g components within the detonable cloud or immediately adjacent may not withstand the local pressure effects. Equipment located beyond a few radii from the detonable cloud should be exposed to a reduced pressure l pulse. Thermal effects on components are described in Appendix F. G-1
APPENDIX H CONTAINMENT RESPONSE TO DETONATIONS The introduction of igniters into containment has greatly reduced the possibility of hydrogen buildup to detonable limits within the containment during a degraded core event. With successful operation of the igniters, only small pockets of detonable gas would be possible. The containment response to a detonation is very dependent on the duration of the shock wave in comparison with the response time of the shell. Assuming a single degree of freedom system, loads which are applied slowly are static in nature, and the containment response or deformation can be determined from static analyses (i.e., the dynamic amplification will be unity). For a suddenly applied constant load, the structural response will vibrate with a deformation twice what one would expect from a static load, i.e., a dynamic amplification of two. Thus, any load such as a blast wave which was suddenly applied to the containment and was maintained for a long time relative to the period of the structure (on the order of 40 periods) would result in the maximum deformation of about twice that expected from static analysis. If the load decays more rapidly than i { this, the deformation is reduced. If the suddenly applied load decays l very rapidly with respect to the period of the structure, the load will be impulsive. Deformation of the structure under impulsive load is not dependent on the peak load but rather on the time integral of the load. Thus, the containment can withstand a pressure peak H-1
of several thousand lbs/in 2 .f the loading on the containment is sufficiently short in duration. The dynamic a=plification would be much less than one. In reference H.1, the pressure pulse forces are calculated for several detonable mixtures and are compared to the Surry Nuclear Plant containment using the relationship: I at 1 0.32 Pd where I is the impulse pressure. At is the time duration of the detonation wave in seconds. P is the loading that produces the maximum elastic d deflection for the structure. T is the natural period of the structure. In WASH-1400 10-5 see was taken as an upper bound for the duration of the detonating wave. l l Okrent and Chan (reference H.2) applied this study to an ice condenser plant. Based on the McGuire containment, they have estimated the maximum elastic deflection for the structure. It is taken as about twice design pressure at 60 lbs/in a. Also, the natural period of the l structure was assumed to be about the same as the Surry Plant used in WASH-1400. Using this information, the ice condenser containment is expected to withstand impulses up to about 0.4 lbs/in -sec. Using a detonation pressure of 1725 lbs/in 2as given in Wash-1400, an H-2 (
impulse loading of only 0.017 lbs/in 2-see was obtained, far below the estimated containment capability (see also Appendix J). A Los Alamos evaluation for TMI-2, as discussed in the Technical Staff ANALYSIS REP 0BT ON CHEMISTRY to the PRESIDENT'S COMMISSION ON THE ACCIDENT AT THREE MILE ISLAND showed that an explosion at TMI-2 would produce a load which was less than but close to early estimates of the building's ultimate strength. It was the conclusion of the staff's report that the pulse duration of 10 microseconds in WASH-1400 is based en the thickness of the shock front. They reason that the shock front would strike the wall and be reflected, thereby producing a sudden rise in pressure on the wall. This increased pressure level
. would be sustained until rarefaction waves followed the shock and lowered the pressure. For this reason, they believe that the WASH-1400 results are probably in error, although the conclusion arrived at concerning the WASH-1400 results may be correct. The approach to the detonation problem has been oversimplified. The assumption has been a large open volume with a uniform detonable mixture. In actuality, the mixture in the containment will not be uniform; spaces are generally small containing large equipment which tend to break up and dissipate any shock waves, and in the case of the lower compartment of an ice condenser, the -steam content may be too high to support detonation.
It is difficult to identiff the impulse loading capacity of a local region of a containment structure; however, the general structure of the Sequoyah plant is on the order calculated by Okrent and Chan (reference H.2) for the McGuire Nuclear Plant. H-3
As indicated in the Kemeny Commission technical staff's report on chemistry, the key to the question of containment failuro due to hydrogen detonation is the duration of the pulse loading. The conclusion drawn by the staff's report is that based on early estimates of structural capability the containments have only marginal capability to withstand a detonation. The inclusion of an igniter system greatly improves that margin. In Sequoyah, the igniter system will present the buildup of large regions of high H2concentrations; hence, only small pockets of high concentration could exist. Detonations of these pockets would result in a : mach reduced challenge to the containment than previously considered. Because the amount of gas detcaated would be small, the duration of the pressure pulse, between the time of the arrival of the detonation wave and the ensuing rarefaction, would be considerably shorter than the few millisecond impulse duration estimated by the ! Kemeny Commission staff on chemistry. Another factor which will ' l reduce the impulse is the change that takes place at the edge of'the l detonable mixture. At this point, the detonation wave is no longer sustained, and the shock wave will begin to subside. If sufficient distance were available, the wave would eventually degrade into an acoustical disturbance. Thus, both peak and duration are reduced in local detonations. QF In our investigations using the CLASIX code, hand calculations, and Q l0 f, I qualitative examinatiens, we have not identified any areas where local detonations would be sustained. However, areas where it may be H 14
postulated to have a sufficient buildup of hydrogen would typically have an equivalent volume radius en the order of a few feet. Because of the irregular nature of these volumes, a precise detonation pressure as a function of time cannot be determined. However, we estimate that the peak pressure pulse emerging from a small detonable cloud would have a triangular pulse shape with a peak pressure on the order of 100 psig and a duration on the order of 0.5 millisecond. This peak pressure would be reduced below 40 psig within a few radii (10-15 feet) of the edge cf the cloud. No damage to major structures a would be expected from such a pulse. . Ch 10 CLE In the Sequoyah containment, the lower compartment shell is protected from detonations by the crane wall. Detonations outside the crane wall are extremely remote possibilities, but in the event of an occurrence, the duration would be extremely short because of the very small space. In the upper compartment, large detonable mixtures cannot occur because of the operation of the igniters in the lower compartment and the upper. A remote possibility of a local detonable mixture exists in the upper plenum of the ice condenser. Again, no damage would be expected because of the very short pulse duration. In addition, the many pipes exiting the ice bed and the open doors exiting the upper plenum would significantly contribute to a dispersal of the shock wave. We thus conclude that the Sequoyah containment with an operating igniter system has ample mardin to sustain detonations.
< H-5
1 i 1 Y l Beterences
- 1. WASH-1400
- 2. C. K. Chan, "On the Failure Mcdes of Alternate Containment Designs Following Postulated Core Meltdown," UCLA-ENG-7661, June 1976, Principal Investigator D. Okrent, pp. 58, 59, 60, and 93 4,
l l k i O 1 i f I 6 i \ I t 4 S 1 l. l' I
.~.a ~
3 L i r
. g_f ,
-._._.. . _ , . ... , . . - . - _ . - . . - , . - . . . . . . _ . _ . . - . _ _ . . _ . . _ . . . . . _ . _ _ . . . . - . _ _ _ . .._,.-._...-.;.-m._.. . 2._ -
i APPENDIX I CONTAINMENT AND ASSOCIATED SYSTEMS The containment system is designed to assure that an acceptable upper limit of leakage of radioactive material is not exceeded. Aiding the containment are several associated systems whose functions support the overall containment system function. In general, the containment system acts as a physical barrier to radioactive material release. The associa,ted systems function to provide containment heat removal capability and/or prevent hydrogen accumulation. (Excessive pressures inside containment recm excessive heat or hydrogen combustion could compromise containment system integrity.) . The associated systems include: (1) Ice condenser (2) Containment spray systems (3) Air return fan system (4). Combustible gas control system
- Adequate heat removal capability for the containment is provided by the ice condenser, the air return fan system, and two separate containment heat removal spray systems. One of these heat removal spray systems is the containment spray system, and the second i
is the residual heat removal spray system which is a portion of the T.1
residual heat removal system. The combustible gas centrol system works to prevent excessive hydrogen concentration inside containment. The containment and the above associated systems are discusse-i in more detail in the following sections.
- 1. Containment System The Sequoyah containment system is designed to assure an acceptable upper limit of radioactive material leakage by containing the radiation, mass, and energy released from a breach of the reactor coolant system and the nuclear fuel cladding.
The containment system is a dual containment structure composed of a primary containment and a secondary containment which encloses the former. - It is the primary contair. ment which acts to physically contain the radiation, mass, and energy resulting from an accident. For this reason in this appendix and in the other sections of this document primary containment is referred to as containment. The containment is a freestanding, welded steel vessel with a i vertical cylinder, hemispherical dome, and a flat circular base, i Free volume of the containment vessel is 1.2 million cubic feet. The design internal pressure for the containment is 12 psig, and the design temperature is 250 F. The design basis leakage rate is 0.25 percent /24 hr. The containment is subdivided into three compartments: the upper vn
compartment, the lower compartment, and the ice condenser. The lower compartment completely encloses the reactor coolant system equipment. The upper compartment contains the refueling canal, refueling equipment, and the polar crane. The ice condenser, which connects the lower compartment to the upper compartment, is discussed in a later section. The secondary containment encloses the containment (primary) and provides an effective barrier for airborne fission products that may leak from the containment during a LOCA. Gases which leak from containment are diluted with the , air enclosed within the secondary cont'ainment, held up, and filtered before being bled to the environment. The containment type used at Sequoyah was selected for the following reasons:
- 1. The ice condenser containment can accept large amounts of energy and maso inputs and maintain low internal pressures and leakage rates. A particular advantage of the ice condenser is its passive actuation not requiring an actuation system signal.
- 2. The containment combines the required integrity, compact size, and a carefully considered advanced design desirable for a nuclear plant.
3 The doub1'e-enclosur e concept affords minimal interaction
'I-3
l l l y 9%~ fQ>v 0 i. e%s% xQ $e f, Q,s ll,lp/ NNNlll/ 9,,,, /// tf 2 *Qf 1
\\\\ /
/
! NV IMAGE EVALUATION N
- TEST TARGET (M T-3) i i
l l l 1.0 gmins
! S l? Na l 1.l .![HM 1.8 l l
! 1.25 IA I 1.6
= I e 6" *
*r % ' + //h
*kYkf' ov 4)yh'fb zzzzz
$h YI//
p %p 4JS g, f %gp
\ IMAGE EVALUATION '%
TEST TARGET (MT-3)
)
l I 10 P 2n m-1 n lI! y=;, GEM l 1.i t"2 IlLE l 1.8
\_
- l.25 1.4 1.6 4
4 6" > i p% ! 44s%
- e>$e/,#
v 4.es
,6
. A th
. 4
^ 0 '- ff%gk4*h' . 4
between the containment vessel (leakage barrier) and the reactor building (protected structure), a cargin of conservatism in leakage rate from the use of two structures and the ECTS, and a reduction of gaseous and particulate radioactive releases due to annulus mixing and holdup prior to filtering and release.
- 2. Ice Condenser The ice condenser is one of the three cc=partnents inside primary contain=ent. It is designed to limit the centainment pressure below the design pressure for all reactor coolant pipe break sizes up to and including a double-ended pipe severance of the largest
=ain reactor coolant pipe.
The ice condenser concept utilizes a large = ass of ice to condense escaping high-energy steam from postulated loss-of-coolant accidents (LOCA) of steam line break accidents. The rapid condensation of steam in the ice bed keeps the maximum containment 4 pressure relatively low while maintaining the capacity to absorb a coatinuing high energy input fecm the reactor core and reactor coolant systems. The ice condenser is made up of 24 individual bays which form a 300 are inside containment. Each bay consists of three major sections: A lower plenum; an ice bed; and an upper plenum (figure I-1). The lower plenum is isolated from the lower compartment by docra in each bay that open at a differential ' pressure of 0.007 psi. The ice bed contains a minimum of 2.45 million pounds of ice. The ice is stacked in columns one foot I-4
10 8 nats I-1. CORCiORSGP my[sxp
't 1
=
-7 y
m,-c...
;gyg Bi?A ,.
e,,,e
;g,r=,,1 N ~ L,.., n.
fdhflgg
+
rmp ;/, 7 situ
'j m.
t.,,, , ,..., y r% i - lI
- statan:e E8:L 3cer I e.. , , ,
]N ~! ,: q O k,! d L'.'
W:.9_ b l W Q ~ SHELL j M' Rtikky
.j. .
pg ir o I
%. s\s( .lW-
,., + str'w uncina i _
FA8f! % ' 3 N n 5;
.f ,
$ ~N s
~ .
.. t
. .~ f
^
JMs9qn $[% .%,ff i.,, % N
/
# # q .
x .
- s. -
'ama net t %,
' =.: .
- <d(., . 'nzr oca,
,z .
M. .*' t.* * /.
/ 'lowgg
~
'l.^'
/
' ~
- It'Pf oti ^
STRL'CTUt t flock m I-5
in diameter and 48 feet high. The upper plenum contains cooling units used to stintain the Icw ice bed temperature during nor=al plant operations. The upper plenum is separated from the ice bed r.nd the upper compartment by two sets of doors that will open with a differential pressure of 0.028 psi. In the event of a LOCA, steam pressurizes the lower compartment which opens the lower inlat doors. An air-steam mixture enters the ice bed where al? the steam is condensed. The rising air then causes the top two sets of doors to open and flows into . the upper compaatment. To provide maximum use of the ice bed, air return fans are provided which circulate containment atmosphere through the ice bed condensing steam released after the initial pressurization. When all the ice has melted, the spray system located in the, upper compartment removes the remaining energy 'eleased to the containment.
- 3. Cents.inment spray systems The containment spray syste=s consist of two separate trains of equal capacity with each train independently capable of meeting 1 system requirements. Each train includes a pump, heat exchanger, ring header with noz les, isolation valves and associated piping, l and instrumentation and controls. During normal operation, all of the equipment is idle and the associated isolation valven'are closed. Upon system activation during a LOCA, adequate containment ccoling is provided by the containment spray systems whose components operate in sequential modes. These modes are:
I-6
(1) spraying a portion of the contents of the refueling water storage tank into the containment atmosphere using the containment spray pumps; (2) after the refueling water storage tank has been drained but while there is still ice remaining in the ice condenser, recirculation of water from the containment spray pumps throuF'.1 the containment spray heat exchangers and back to the containment. This spray is useful in reducing sump water temperatures and increases the effective life of the ice; (3) diversion of a portion of the recirculation flow from the residual heat removal system to additional spray headers. The latter operation occurs in the event the containment pressure starts to rise after the ice condenser has been depleted. The diversion will be by manual operation of system co=ponents. The spray water from the containment and RHR spray systems will be ret"*ned from the upper compartment to the lower compartment through two i4 inch drains in the bottom of the refueling canal. A small portion of this spray water is diverted to the floor and equipment drain sump through four 3 inch drains located on the operating deck. The primary design basis for the heat removal spray systems is to spray cool water into the containment atmosphere when appropriate in the event of a loss-of-coolant accident and thereby ensure that the containment pressure cannot exceed the containment shell design pressure of 12.0 psig at 250 F. This protection is afforded for all pipe break sizes up to and including the hypothetical instantaneous circumferential rupture of the reactor I-7
ccolant loop resulting in unobstructed flow from both pipe ends. The containment spray system supplements the ice condenser until all the ice is =elted (approximately 7,000 seconds after the LoCA) at which time it and the residual heat removal spray system become the sole systems for removing energy directly from the containment. The containment heat removal systems are de5igned to provide a means of removing containment heat without loss of functional performance in the postaccident containment environment and operate without benefit and maintenance for the duration of time to restore and maintain centainment conditions at atmospheric pressure. Although the water in the core after a loss-of-coolant accident is quickly subcooled by the emergency core cooling system the design of heat removal capability of each containment heat removal system is based on the conservative assumption that the core residual heat is released to the containment as steam which eventually melts all ice in the ice condenser. The secondary design basis for the containment heat removal spray systems is the suppressicn of steam partial pressure in the upper volume due to operating deck leakage from a small break before a full less-of-coolant accident. The requirement is that the containment spray systems be able to absorb the steam leakage through the operating deck at the maximum possible long-term deck differential pressure of one pound per square foot equivalent to the ice condenser door opening and differential pressure in the upper compartment due to deck leakage in the " double accident" situation. The " double accident" is a small break followed by a large break up to the double-ended severance of the largest I-8
pipe in the reactor coolant system. The system is designed such that both trains are autecatically started by high-high containment pressure signal. The signal actuates, as required, all controls for positioning all valves to their operating position and starts the pumps. The operator can also manually actuate the entire system from the control room. Either of the two containment spray trains containing a pump, heat exchanger, and associated valving and spray headers is independently capable of delivering the design flow of 4750 gpm. The RHR spray capacity is 2000 gpm per titin.
- 4. Combustible Gas Centrol System
~
The ecsbustible gas control, system is designed to monitor hydrogen levels inside containment and, should concentrations exceed preset levels, act to cor. trol the hydrogen gas in the containment , atmosphere. The combustible gas control system is composed of the following subsystems: (1) Hydrogen monitor (2) Hydrogen recom'ine's o r (3) Hydrogen purge The hydrogen monitor subsystem is discussed in detail in appendix K and will not be examined herein. The hydrogen recombiner subsystem consists of two electric e b hydrogen recombiner units, located in the upper containment I-9
compartment, and separate control panels and power supplies for each recombiner unit, located outside the containment in an area that is accessible following a loss-of-coolant accident. The-recombiners are completely redundant. The recombiner unit consists of a preheater section, a heater-recombination section, and an exhaust section. Containment air is drawn into the unit by natural convection, passing fir?', through the preheater section. This section consists of the annular space between the heater-recombination section duct and the external housing., The temperature of the incoming air is increased by heat transferred from the heater-recombination section. This results in a reduction of heat losses from the unit. The preheated air passes through an orifice plate and enters the heater recombination section. This section consists of a thermally insulated vertical metal duct enclosing five assemblies of metal-sheathed electrical heaters. Each heater assembly contains individual heating elements, and i the operation of the unit is virtually unaffected by the failure 1 af a few individual heating elements. The incoming air is heated to a temperature in the range of 1150 to 1400 F, where recombination of hydrogen and oxygen occurs. Finally, the air from the heater-recombination section enters the exhaust section where it is m'.4eo with cooler containment air and discharged from the unit. Tests have verified that the recombination of hydrogen and oxygen I-10
in the unit is not the result of a catalytic surface effect but occurs as a result of the increased * 'gerature of the process gases. The performance of the unit is, therefore, unaffected by fission products or other impurities which might poison a catalyst. The hydrogen purge exhaust subsystem consists of a single penetration in the primary containment wall equipped with two normally closed, remote manually operated isolation valves, one on either side of the centainment wall, and one pneumatically operated annulus purge exhaust valve located within the annulus. With these valves open, a flow path is established between t: 1 primary containment and annulus which will permit purging of the containment for hydrogen control subsequent to a LOCA. The i= pet as for flow will be provided by a differential pressure of at letst 0.5 inches of water gauge which will be maintained by . the annulus air cleanup (emergency gas treatment) system. The containment eff;uent purged for hydrogen will flow directly to the annulus where it will mix with the annulus atmosphere and be filtered by the air cleanup system prior to discharge. In the event that the hydrogen concentration exceeds 3 percent this subsystem can be used for containment purging. This subsystem is designed to provide a backup to the redundant inplace j hydrogen recombiners. Its use is not anticipated in l a design basis ace; dent even if one recombiner system fails. The actuation of this system is covered in TVA operating procedures. I I-ll (
1 Use of this system is not consistent with the use of the Interim Distributed Ignition System. The purge subsystem was analyzed assuming the design basia LOCA. For degraded core events the hydrogen release may occur faster and earlier. If the purge sub-system were used for these events, larde offsite doses would occur without appreciably reducing the containment hydecgen concentration significantly, at least in the short term. The system was required by Regulatcry Guide 1 7 and was intended for operation several days after the event. TVA's studies cf the containment response to degraded core events have shown tha'; the plant can safely withstand hydrogen concentratians in excess of 10 percent. Therefore, when use of
~
the JLIS is approved by NRC for use, TVA will modify the appropriate operating procedures to prchibit use of the purge subsystem any earlier than four days after a severe less of coo. ant accident. This period allows sufficient tiae to evaluate the nature of the event and obtain samples of the containment atmosphere, while allowing safficient time before the subsystem is needed for the design basis event (greater than 8 days).
- 5. Air Return System I-12
, n ,,.. ~
The primary purpose of the air return fari system is to enhance the ice condenser and containment spray heat removal operation by circulating air from the upper compartment to the lower compartment, through the ice condenser, and then back to the upper compartment. The cperation will take place at the appropriate time follcwing the design basis accident including loss-of-coolant accident. The secondary purpose of the system is to limit hydrogen concentration in potentially stagnant regions by ensuring a flow of air frem these regions. There are two 100 percent capacity air return fans. Each will remove air at the rate of 40,000 cfm from the upper contain=ent through a =ain duct to an accumulator room of the lower
, compartment. (See figure I-2.) The discharged air will flow from each accumulator room through the annular equipment areas into the lower compartmer.t. Any steam produced by residual heat will mix with the air and flow through the lower inlet doors of the ice condenser. The steam portion of the sixture will condense I
as icng as ice remains in the ice condenser and the air will i continue to flow into the upper compartment through doors at the top of the ice condenser. Each main duct contains a non-return damper which prevents flow from the lower compartments to the I upper compartment during the initial stages of a loss-of-coolant accident. i I Both fans will start automatically after receipt of an isolation j signal. In addition, either fan may be controlled manually from I-13
L _
}
- [ aYm@8 Y m R i I
.4 n8Y O ' c
- - 5 T I N
5 AL f*
)L E J
/ .
6 U 8 I W2 0 S ' N M O - S A U 3 L 4 E F C 1 F V C 9 N E T R A T E U A R N 5 T C U E t E M f D I G R N u- R N FI AI 5 s I T 1 - I A N N 3~? - S O O R R C O ' E D
't f 5 T f B E
- MS n 2'A E T U A L
5 e R U 57 9 - O 4 M ONB f ~ T L U N1. 2. C E S, A C "t - E S C f S A o R E s 0 3 R Y V 8 3 ~ EI T R
~ PV O f[ 80 -
PA T UC C A
\
S 1 EO MT R R w ,R E
)
OC E O DE ' b O _ $( NOTL L Z I R U o o s I E>.- > 5 N 7 E T EC S S w + U M 2
' )O M E 0 -
'0 NN IE R P,
0 I 1
- h. R T
S S I 19AG ' ., . h ( TO NR g " N I R I E R OD U CY 'v
/ G I
H F N . 5" 7 3 \/
/ R
_ O '
! 5 T (1
_ S
- 2 AL R 7 s. -
a O 3 d S 1 U 8 R S M 3 T o R C o O U 5 - E 8 T O C L 1 A T C L R C A O E E C N N L E A L N G F O E C G MN O AR N R EU E O G R Y TT O O H S E R 0 T ' R O 4 A R Y L H I U { I _ y A M z
- U q I l &- 8h C C
( A _ /
the control room. Each ran can develop sufficient head to keep the non-return dampers and ice condenser inlet doors open after blowdown is complete. A flow indicator upstream of each fan and a pressure differential indicator across each ran are provided. In '*ition, the discharge flow rate is indicated in the control room. Si=ultaneously with the return of air from the upper compartment to the lower ccmpartment, post-LOCA hydrogen mixing capability is provided by the air return fan system in the following regions of the centainment: containment dome, *each of the four steam generator enclosures, pressurizer enclosure, upper reactor cavity, each of the four accumulator rooms, and the instrument room. These regions are served by seven hydrogen collection headers which terminate on the suction side of either of the two air return fans. A schematic of this system is shown in figure I-2. a I-15
APPENDIX J CONTAINMENT STRUCTURAL CAPABILITY ANALYSIS An evaluation of the containment capability to withstand pressure has been performed. The burning of hydrogen in the centainment produces a pressure transient which is essentially static when compared to the response time of the containment vessel. Hence, the evaluation has focused on static pressure application. On the other hand, detonations =ay be possible within some areas of tt-containment. Det? nations generally have very short periris when compared to the response time of the containment and are, therefore, impulsive loadings on the containment should they occur. Some secping analysis of the impulsive loading capacity has been made. It is recognized that the actual capability is dependent on many variar.es and may be quite different at various containment locations. Influences such as pulse shape, flexibility of the containment, obstructions, etc., can have a large effect on the loading capacity. T7A plans to rsview these influences in more depth as appropriate to achieve a better understanding of the centainment capability. Dstonation Lead Capability TVA has considered the case of 100-percent metal-water reaction (about 2,000 pounds of hydrogen and about 25 percent of hydrogen by volume). We considered that tts containment was a simple cylinder (which is structurally weaker than a sphere) consisting of only a 1/2-inch-thick steel shell (the minimum shell thickness at Seqr . We used the
" impulse" loading information provided by C. K. Cnan. We have J-l
concluded that failure of the containment wall due to detonation shock wave is not expected to occur; however, even though the containment can withstand the detonation loading, due to its short duration, the resulting relatively long term pressure due to the oxidation of such a large amount of hydrogen would exceed the ultimate capability of the containment. This would be true even for smaller quantities of hydrogen uniformly distributed and completely burned in one short period. Static Pressure Capacity The calculated pressure capacity of thp SCN containment vesse? at first yield is at least 46 lb/in gauge using the actual minimum yield strength of SA 516, grade 60 material. This pressure was derived using a finite element medel of the critical area, the 1/2-inch shell between the stiffeners at elevation 778'-6" and 788'-0". An elastic-plastic analysis was performed of this aret. using the ANSYS finite element program. The shell ring stiffeners and vertical ; stringers were modeled discretely using plastic elements in ANSYS. An evaluation of all components of the centainment boundary (e.g., penetrations, the personnel locks, equipment hatch, and valves) was performed to confirm that the 1/2-inch shell plate is the controlling segment of the structure. The bellows and electric penetration inserts were evaluated by their vendors and shown not to be control.';n; items. J-2
( APPENDIX K CONTAINMENT MONITORING SYSTEM The containment monitoring system corJef its of hydrogen analyzers to mer.itor hydrogen concentrations inside containment, pressure control loops to monitor differential pressure between the annulus and the containment, arid temperature sensors distributed throughout containment to monitor containment air temperatures. Hydrogen Monitors Detailed information en the hydrogen monitoring system may be found in Volume I, section A.S.1. , of this report. In addition, each analyzer takes a well-mixed sample from .several points in bcth the upper and lower compartment for an indication of an overall average containment hydrogen concentration. Pressure Monitors Five differential pressure control loops monitor the differential pressure between the annulus and the lower compartment of the containment.- Four loops have indication ano annunciation in the main 2 control room witn a range of -1 to 15 lbs/in g as well as actuating the safeguard systems. The fiftn loop indicates and alarms in the auxiliary control room as a backup. There are also two wide-range pressure control loops with a range of 0-60 lbs/in g and 0-100 2 lbs/in 8 to monitor large pressure transients. Each wide-range K-2.
pressure control loop has trained power and indication in the main control room. i Temperature Monitors Thirty-four temperature sensors are strategically located throughout containment with indication in the main control room. Temperature sensors are located on the intake and exhaust of various containment coolers in both upper ' and lower compartments to ensure that the air tem?erature requirements for proper operation of equipment are maintained. Other sensors are used to establish an average te=perature which is mcnitored in the main
- control room to assure personnel comfort during occupancy.
o l l + K-2
APPENDIX L IGNITER SYSTEM DESIGN The interim distributed ignition system (IDIS) is designed to burn hydrogen inside the containment in the event of an accident in which excessive hydrogen is generated inside the reactor vessel and released into centainment. It is designed to ignite the hydrogen prior to it reaching a dangerously high level. This system is intended to back up the safety-grade hydrogen recombiner system, but it is not a safety-grade system itself. The system consists of 45 thermal resistance-type igniter assemblies G)(8 distributed throughout containment. Currently, 32 assemolies are installed at the original locations. The additional 13 assemblies will be installed in the upper co=partment. The locations of all these assemblies are shown in the attached figures. These assemblies (typical assembly shown in figure L'-1, attached) are located in three i different compartments inside containment (see figure L '). Lower Compartment Twenty igniters are located in the lower compartment; four at elevation 689 0', nine at elevation 700 3', and seven at elevatien 731.0'. At elevation 689 0'(see figure L-4), four igniters are lour ad on the ;utside of th1 shield wall. This area is primarily a pipe raceway. The area inside the crane wall at this elevation is the containment sump and all lighting fixtures in this area have been 8 capped and sealed against the flood. The area outside the crane wall is sealed to prevent flooding and also provides access to the T1
1II 1 , t r- . . i \ / i a.3li!!g,.,.
*: .a
. i,
/ /
i
,/ /
/
s' / M
~
/ - . .
/ l
, ~
/- \ : %
y ,/ N Y / / 0 /
#~
! 4 4 . e & I / 4
.I a Y / /
< . / e s s *
/ /
I ,# s
/
s t . / DONGAN - i TRANSFORMER I W 4
/ \
1 C.T. g PRELIllI!!ARY PIIASE I 5 Holl!1TI!1G DETAIL FIGURE L-1 j [. - .1
l l
, :, .7 _- ~%.
\f y /\ .
f 818.o rt. 1 1
,. / 809 0 ft.
7 s
-- l t.
~' -
1 11 % _ >r ir- _ __.-d -
. () r --
..-r-- I ,
3'=- .-- p 792.0 ft. g '---V '-- , \ 1/ '
/ '. 1 -1l -.
787.O et.
; , t-
[T. l . .. 3 ,- . :-] , \. l [-~.,- Tf s l 3 I I \} I I / ~-- ---
~
l . .l
- d ~ k @- ~4 ,_., -<,,_-<,g
- ~ # y .u}S.qwAi+
F. L. -
.- ..p 9. -
.:--~ .~..
i p y . (
- .n ..
ie-. - 1 - 755.0 ft. N _p,
- 3 , < .il, , , 4 ,
7h6.0 ft. s ..K 5.'... n.
.M*
6 o b.- , y-
, rh.I y:;;:c 1 7h2.O tt.
G
' 4. e e O ~~' Y O T nwp
! . .y
=,._ -N -
t t-l! . -*--'~-i
. .f , 4,,
1 y1.0 *t.
**** ~~ -
W*" t,
. i., .
.+- I.-
...+
3.j s
,- .w .
1* I- -
.a- e . e *- l }anL.,. .
* *s r-y gl. ,,,
!" .tj_ ,* fJ . r g
- -. ~ .
m , - g
.w . .e-- 411 L.31i -
f.,.... .-
- , i- - -
.~~
-'.- i.-
sr' . i ,t, : A 00*o ft _ 3.
. .m!.d .' l .f - ~- .
l W ,/
--.' 'y')
.-. N . .: . :.. p L 11,,,)f.
\
g
~ - o ~r
- n. . .. . ._ t~.,,.,_rs.
- -T l ..i . - - * ' " },! .
.- .m ';- - - _y__
- [,, =j {, --T s o $ .. y 1 ,
%689 0 rt e .,
a <
.i
,y e: ,
i.r.*-1." .:. l y a.
. , . . 669.o rt.
ig .,, ,y g -. n
}. . .
l
..'. ..ag6-w i=-
i .. , 1
.. .. ,... ... a.* .. * = , *, .-
- c. . . .
l t
***=....;. ~ . .
Centain=ent Igniter Location Elevations i
* ' Figure L-2 D** D T Tf
_e o o .1.5 = L-3 l
i l 1 KEY To FTGURES L 4 THRoUGH L-12 DenotesHydrogenIgniterLocation-h I Fixture Descriptien For=at: "X-Y" , where X is ; G- fixture containicg glewplug type igniter Y is one of the following; A- phase A electricity 3- phase B electricity C- phase C electricity Denotes H 2 sampling location: h El. 67k.78 Az. 280* ,
@ n. 716.o'A2. 31o*
@ n. 716.6 u . 230' t
! @ n. 73o.o Az. 310' l @ n. 756.o Az. 90'
@ n. 756.0 u . 315' x
FIGURE L-3 L-4
O f . O O ( - 4 d' .- 90 i- -
,f ,
.g*C OG-A Contai:::nent Igniter Locations El. 689 0' Figure L-4 l
l L-5
containment fan rooms and accumulator rooms above at elevation 693.0'. At elevation 700.3' (see figure L-5), all the igniters are located outside of the crane wall. There are five igniters located in four accumulator rooms (one each in accumulator room numbers 1, 2, and 3 and two in accumulater room number 4), one igniter each located in the two lower containment cooling fan rooms (180 degrees apart ), and two located in the area adjacent to the personnel hatch. Seven igniters are mounted inside the crane wall at elevation 731.0' (see l figure L-6). There is one igniter located above each of the four reactor coolant pu=ps, and one above the pressurizer relief tank, one between steam generators 1 and 4, and one between steam generators 2 i and 3 Ice Condenser Compartment - There are a total of nine igniters in the ice condenser area. At elevation 731.0'(see figure L-6), there are rive igniters located at the bottom of the ice condenser. At elevation 792.0' (see figure L-7), there are four igniters mounted on the crane wall above the ice condenser and below the top deck blanket. Both sets of igniters above and below the ice condenser are generally evenly distributed around containment. Upper Compartment t i There are a total of thirteen igniter locations in the upper compartment. Three igniters are located in three quadrants suspended 35 feet from the dome (see figure L-8). Five igniters will be generally distributed around the containment shell on the crane collector supports above the ice . condenser. top deck blanket at r4
O
. .. :.:,-s. . ..
e
-A 2 -
G-3 . GC O N
\.b -
[- - s .
- n. . -..
;-,' g ' .
\ ' '. .
,, g f[
; / ,h__M.i. . . .
s h C. .
%' i l
's s f M -C l
'A~-~ - -%Ta I
l t s c.1 . , .2 . . o l Containment Igniter l Locations El. 700 3' Figure L-5 l Revision 1 l L-7 l
O cs . . .
. NN G -C
. ,e . 'ss
. / \
o
'G-A .o .
\
s
< .i. ".\;f.'s, /g" \- *d. I e ', t
. f .
90 G - A h. ( /,,
- Reactor
\ T'd 'S h ' 2,A **l* Vessel ,!
Y# Vent Header' / kA i5 G -C C/ e - ( 'N / N l
,' G - B@.. i_ '
/
l
. e J
l l Centainment Igniter Locations El. 731.0' Figure L-6 1 Revision 1 l L-8 l t
m-
@ O f?h& e
-t'E se i
~
1 -
' ]l
\ ' N N ' ~
~
j G-A
,na-*;.4h a m;,,a L-9
l l 0 v e' > l
/
,- - ,. -8 p O
/ r
/ s G- / .
's o \
\
\
\
\
i f 90
-{-
[0 - 1 I
/ .s
/
/ *
/
/
/
@G -C l
I l
- t. J
' 3. ' '.
l l Containment Igniter Locations El.818.O' Figure L-b
. 10
9 elevation 809.0' (see figure L-9). There will be four igniters located inside the crane wall vall about 90 apart above the doghouses at elevatien 787.0' (c e figure L-10). One igniter will be located above each of the two air return fans at elevation 746.0' and azimuth 293 and at elevation 755.0' and azimuth 258 In addition, there will be one igniter located on the exterior of each of the two steam generator doghouses at elevation 742.0' (see figure L-11). cql8 Igniter Assembly The individual igniter assembly (shown in figure L-1) is a very simple design. Power Supply Each assembly is supplied power from one of three 120V ac standby lighting circuits. The three 120V ac circuits are supplied through transformers from the 480-volt shutdown boards which have normal and alternate ac power supplies and in 'the event of their failure are j fed from the diesel generators. i Igniter Assembly Enclosure Each igniter is partially enclosed i a 1/8-inch steel plate box which f also houses the igniter 120/14V ac transformer and all electrical connections. (In describing the box, the dimensions are oriented with respect to figure L-1.) Looking at figure L-1, the box is eight inches . long by eight inches high and six inches deep. A typical assembly weighs approximately 21 pounds. Each box is covered by a 1/16-inch thick spray shield which is 14 inches long and eight inches deep. Not shown on the FSAR figure is a copper heat sink on the face of the L-11
O
. I- _.
G-A 's,
- v. s s; .
% ~..,
s o G-B
~
s i.
\
\
. \
\
t-
.t I
i
's
_ g G,C te go o
's
- Y l .
?-
A / I e
.'. f l
i / I l
/
I / l /
/
/
1
/
G-B - l e__ _ _ _ _ - v ot' e t Contairnent Igniter Locations E1.809.o' Figure L-9 . i l .I
. i l
L-12
O
,,a 1- s.
- t' . , .
, . ~ f'
-4%* s lG-A
, / g -
i
/ 8 N
, / \
; / \
'/ \
/ \
/ \
I \ o "C - i ,_
/8 W ?
I g G-C h o,
\
,\ .
- g .
\
~
\
\
\
s
- G-B
%g I
l t ' o l' = Containment Igniter Locations E1.787.0' Figure L-10 I L-13 i
O e
. :. r '
j . N.
' 'o ,
a
..- ., , \.
3 c-s c.c @a .742.o' n.ru$o - 900
~
i 8 g I
#d, 7 ,
. O,
. - a n.755 0'
*^ @ ~
n.742.0' .' G-A @ i l e ol'
- l l l Contairment Igniter i Locations E.742.0, l n.746.0 & E. 755 0' l
Figure L-ll L-14
.(3) ma.o m.
809.o rt. e(5) I 792.0 ft.
.(4 ) e (4) .
i s I g
. a lm m i
i 755 0 ft.
, &, 74b.o tt.
.(a) .! ! 7ha.0 rt.
731.0 ft.
~
.(s ) - (7) -
- . m b# Air R5 turn
= .= r=
~ ~
e(9) ; j j 700'0 #t* i
-~
G. .] L. .i
~
e -
- 689.o it.
*(4 T 669 0 ft.
N Figure f-12 L-15
igniter assembly. This 1/16-inch thick copper plate is six inches deep and six inches high. The plate is bent outward at a 45 angle away from the box at approximately one inch to each side of the glow plug. It is secured to the glow plug and box by a washer and a nut which is screwed cnto the outer threads of the glow plug. Access to the box interior is made through one side of the box. This side is coated en the interier of the box with a rubber seal material and is attached to the box with four bolts, one in each corner. Transformer The Dongan transformer located inside the 1/8-inch steel plate box receives the 120V ac from the standby lighting circuit and steps the voltage down to 14V ac. The transformer is wound with class H insulation and is rated for operation in greater than 400 F environments. - Glow Plug The transformer supplies 14V ac to the General Motors AC Division Model 7G glow plug. In laboratory tests conducted by TVA, at 14V ac the exposed portion of the igniter glows a red-4, range and reaches 1800 F. The glow plug is basically a coiled nichrome wire surrounded by magnesium oxide inside an Inconel Alloy 601 sheath. L-16
APPENDIX M IGNITER TESTS ENDURANCE AND ACCEPTANCE TESTS TESTING C0NDUCTED AT TVA'S SINGLETON LABORATORIES I l l l l I l
t 1.0 Introduction TVA had a testing program which was conducted at TVA's Singleton Laboratory to obtain preliminary information about the performance of commercially available igniters. The purpose of these tests was to screen alternative igniters and to gain a degree of confidence that the igniters could ignite hydroren. The tests were not run under normal laboratory test conditions since the objective was to identify which igniters, if any, were most promising as subjects for more detailed testing anf evaluation. Nonetheless, TVA gained considerable infor22:icn and assurance that commercially available igniters could ignite hydrogen. The preliminary screening resulted in the choice of GM AC glow plugs for use at Sequoyah unit 1 in the interim distributed ignition system. A second purpose of the testing program at Singleton was then to perform endurance tests on a lot of 300 of these glow plugs. 2.0 Preliminary Screening A number of igniter types were evaluated, ranging from high energy spark igniters to large diameter (1-1/2" I.D.) heater coils. Although the spark plug type igniter was considered an excellent candidate for this application, it was rejected prior to preliminary testing due to potential problems with electromagnetic interference (EMI) with critical instrumentation. TVA's Electrical Engineering Branch is researching the problems associated with EMI generators, and spark type igniters may be considered at a later date for use in Sequoyah unit 2 or Watts M-1
Bar. Two other potential candidates, both coil heaters, were rejected after the first ene, a large diameter (1-1/2" I.D.) coil, ceuld not reach sufficient surface temperature, and the second one failed at the connector in less than five =inutes. Therefore, testing was restricted to diesel engine glow plugs, since they were known to be c3pable of achieving the 1500 F minime.1 surface temperature desired by TVA and because of their rugged de.ign. TVA determined that at 12 volts ac, acceptable surface temperatures could be achieved but that considering line losses, variances in system voltages, possible plug cooling due to high humidity, and other effects, TVA would need to operate the plugs at 13 volts ac 1 1 volt. Since the possibility existed that TVA could overstress the plugs by overvoltage, TVA consulted glow plug manufa:turers and identified two types of failure modes which could be expected. The first type of failure caused by overstressing would be the failure of the heater wire within the glow plug sheath. This type of failure due to the breaking of the circuit would outwardly cause the plug to discontinue glowing. The second type of failure caused by overstressing would involve offgassing of the glow plug' tip. Unlike the first type of failure after offgassing, the glow plug may continue to glow; however, the surface temperature would drop significa tly. 30 Description or Glow Plugs
- 9 I
LA O }
?
l Glow plugs =anufactured by three different cecpanies have been tested to date. They include: General Motors AC Division, Model 7G, 12 Volt BOSCH 10.5 Volt ISUSI 10.5 Volt 4.0 Endurance and Temper 1ture Tests Diesel glow plugs are not usually intended for continuous service in an air environ =ent. Therefore, TVA undertook tests to determine that the plugs:
- could reach and maintain the desired temperature;
- could withstand the effects of overvoltage and temperature; and
- that the plugs could operate for extended periods of time
( at high temperatures. 4.1 Testing Equipment For these tests, the plugs were bench mounted and operated for various periods of time. The power source was 120 V ac wall socket which was reduced by a variable transformer I (Variac, model no. Staco, Inc., type 3 P/N 1010L) to the l l appropriate voltage levels for each test. The voltage l l levels both on the primary and secondary side and at the plug were measured by a digital voltmeter (Fluke model u_o I .. __ .,.
numrer 8024A), and the current levels were =easured by an amp meter (Triplett model number 10 type 2). The surface temteraturs of each of the glow plugs was measured by either a trermocouple (type S) in centact with the surface of the plur and connected to a potentiemeter (Leeds and Northrop model number 8690-2) or by an optical pyremeter (Pyro model number 85). A total of 12 plugs have been tested to date. 4.2 Surface Temperature A GM AC model 7G plug was operated at 12, 14, and 16 volts ac. Surface temperatures as measured by the thermocouple were 1480. 1550, and 1650 F, respectively. Since the taermoccuple would be expected to increase local heat loss a.nd hence reduce the measured local surface te=perature of the thin-walled plug sheath, these values were 9robably somewhat lower than actual surface temperatures. This conclusion was supported by later readings with the pyrometer while testing another CH AC model 7G at it volts 9 ao and getting at least 1800 F. A Bosch plug has been tested at 13 volts ac. It produced a surface temperature of 1700 F as measured by an optical pyrometer. Based on these results, TVA concluded that the diesel glow plugs could produce the desired surface temperatures. Gil More recent tests on several glow plugs have been conducted to further guarantee adequate surface temperatures at the minimum design voltage of 12 volts ac. The minimum M-4
l l acceptance temperature of 1500 F was reached within 1 minute after b
- energized with 12 'tolts while all the plugs t
tested reached over 1600 F within 3 minutes. Q ll J 4.3 Voltage Tests Voltage tests have been completed on only the GM AC model 70 pluts. Based on tests on 5 GM Aq 70 plugs, reliable operat'.on at 14 volts was confirmed but two other 7G plugs
~
failed at 16 volts ac after a few minutes. Inconclusive testing on Bos7h plags resulted in two failures when operated at 14 volts act however, one Bosch plug i operated satisfac orily at 13 volts ac. In addition, one Isusi plug was tested at 14 volts ac but lasted for only
~
30 minutes. 4.4 Extended Operation Initial endurance tests were performed on two plugs for extended periods of time. A GM AC model 7G plug was
- operated continuously for 148 hours without failure and was i
later used in the hydrogen burning tests. A Bosch 10.5 volt plug was operated at 13 volts for 90 hours, then cooled down for two hours and turned back on for another 90 hours. l l After selection of the GM AC plug for use at Sequoyah, more extensive endurance testa were performed. Before performing l l l an endurance test on 50 plugs selected at random from a controlled lot of 302 plugs, a screening test was performed on all the plugs in the lot to eliminate those with l l M.c
manufacturing defects and to precondition them before the endurance test is conducted. The plugs presently installed at Sequoyah unit 1 were subjected to this screening test. The screening test consisted of energizing the plugs with 6 volts AC for 5 minutes, then increasing the voltage to 12 volts for an additional 5 minutes, af ter which the plugs were energized with 13.9 volts + 0.1 volt for 1 hour. During testing of the first 16 plugs, six failures occurred. Then the test procedure was modified to include voltage steps at 8 and 10 volts as well as those at 6 and 12 volts. Using the modified test procedure, only eight failures occurred during the testing of the remaining 286 plugs. Six of these eight failed, plugs had a manufacturing date of 0918. The overall failure rate was 4.6 percent; however, this rate dropped to 2.8 percent after adding the two additional voltage steps to the test procedure. All plugs which fail'd during the preconditioning test were eliminated from the controlled lot, and the 50 plugs selected at random for further tercing were drawn from the remaining lot of 288 plugs. The fc ther testing currently being performed on i l these 50 selected plugs consist; af cycling on-off ten times and then a 48-hour endurance run. i 5.0 Hydrogen Testing One igniter (AC 7G) was installed in a "PARR" (229HC6-T316-031579-5142) pressure vessel in order to determine feasibility of igniting hydrogen in a ser. led container. The vessel lid has M-6
a a silicone rubber sealed gas injection sampling port. Hydrogen concentrations in the vapor phase were determined before and after ignition intervals. An ignition interval is the time current ficws through the igniter circuit. The hydrogen was measured by a Perkin-Elmer gas chromatograph equipped with 3920 thermal conductivity detector and an M-2 integrator. The chromatograph was standardized with hydrogen and air mixtures prepared from research grade hydrogen and laboratory air. Temperature measurements were made with a mercury and glass (484635, ASTM 9C) thermometer. Temperatures reported are ambient for tests 1 through 3 Prior to tests 4 through 10, 100 grams of water were added to the vessel. The vessel was heated by
, a temperature adjustable hot plate to saturation temperature of .he water and maintained throughout the test. The reported temperature is the water temperature after completion of the
< test. Results of the 10 ignition tests are given in table 1. 6.0 Cenclusions and Summary i The purpose of these tests at Singleton was to select a ! con =ercially available igniter that was capable of igniting hydrogen. From the results obtained, the GM AC model 7G glow l plug produces more than adequate temperatures at a range of voltages that can be provided inside the Sequoyah containment. In addition, the plug seems capable of extended operation at ( high temperatures and has been shown in small tests to be able l to ignite 12 percent and lower volumetric quantities of hydrogen. Although it has not been tested as thoroughly, the Bosch plug M-7
appears like it my also be an optional igniter. f I e e 1 6 u_A
. . . . - - . . , - _ . . . . - - . . - . . , - _ , - ~ . . , , - . - - . - . . -
i Table M-1 HYROGEN IGNITION TESTS Initial Final Ignition Test Vessel Tgap. Hyd. Conc. Hyd. Cone. Intervals No. Contents ( F) (5 Hyd.) (5 Hyd.) (Min.) 1 Ejd., Air 90 12.5 0.1 5 2 " 80 70 0.1 5 3 80 35 0.1 5 4 Hyd. Air, Water 120 12.0 0.1 3 5 180 14.0 0.5 3 6 180 4.0 2.5- 1 I 7 a 180 25 1.5 1 8 "
, 180 1.5 1.3 1 9 180 11.0 5.0 1 10 180 5,0 2.0 1.3 i Vessel Volume 1.1 da (0.039 ft3 )
Operating Voltage 12V de l f i l l l f 9
$ b
APPENDIX N IDIS IGNITER TESTING AT FENWAL, INCORPORATED TVA, Duke, and American Electric Power, in cooperasion with Westinghouse, devised a two-phase testing program that, in general, was designed to determine the relative effectiveness of a hydrogen ignition system. The attached report is an evaluation of dhe test results generated by those tests at Fenwal, Incorporated in Ashland, Massachusetts. It contains a detailed decription of the two phases of testing, a discussion of the effects of steam, spray, and fan flow on the ability
~
to burn hydrogen, a comparison of the measured test results with the theory of hydrogen burning, an evaluatian of the igniter effectiveness, and an evaluation of hydrogen burning on equipment. i l l I I ( i i N-1 _ . _ , - _ _ , -}}