ML20246G178
| ML20246G178 | |
| Person / Time | |
|---|---|
| Issue date: | 12/21/1988 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-2613, NUDOCS 8905150213 | |
| Download: ML20246G178 (15) | |
Text
_ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _
.r OMS c4/3 M h (].Q.,[j]
[
7" ##
DATE ISSUED: - 12/21/88
,e uhilte ACRS Meeting Minutes / Summary of the Advanced Reactor Designs Subcommittee December 13, 1988 Bethesda, Maryland Purpose The purpose of this Subcommittee meeting was to review the NRC staff's draft safety evaluation report (SER) for the Sodium Advanced Fast Reactor (SAFR) conceptual design.
Attendees ACRS NRC D. Ward, Chairman J. Flack, RES C. Michelson, Member J. Wilson, RES C. Siess, Member R. Landry, RES P. Shewmon, Member R. Avery, Consultant l
J. Lee, Consultant D. Okrent, Consultant L
f M. El-Zeftawy, Staff Others G. Van Tuyle, BNL B. Seidel, ANL C. Allen, SAIC D. Wade, ANL G. Slovik, BNL R. Amar, RI B. Chan, BNL L. Felten, RI R. Hern, RI R. Rogers, RI R. Lancet, RI J. Cahalan, ANL T. Tai, Bechtel Meeting Highlights, Aareements, and Requests 1.
Mr. Ward, Subcommittee Chairman, stated the purpose of the Subcom-mittee meeting and introduced the other ACRS members and consul-tants. Mr. Ward indicated that the ACRS is expected to write a letter on this subject at the January 1989 full Committee and not the December 1988 meeting.
DESIGNATED ORIGINAL 8905150213 881221
//
"I 2
3 PDC
+
i Advanced Reactor Designs Minutes December 13, 1988
)
I
- 2..
Mr. R. Landry, NRC/RES, summarized some of the LMR international.
experience regarding the containment design and shutdown system reliability for the Superphenix and Kalkar (SNR-300) designs.
The Superphenix is a 7900 MWt LMR that was built by a consortium consisting of France, Germany, Italy, and Switzerland. The reactor is a pool-type with four intennediate loops feeding four steam generators. The control system consists of 21 primary control rods. These control rods use two separate (diverse) scram systems.
The plant also contains three articulating secondary rods which are on a separate power system and scram separately of the primary rods. Two calculation have been performed for the failure prob-ability of the overall scram system (21 primary and 3 secondary rods).
NERSA, the consortium which owns and operates the plant, has calculated the. failure probability to be less than 10-8
- CEA, the french equivalent of the AEC has calculated the failure prob-ability to be less than 10~14 The containment design for the Superphenix consists of twa vessels plus a machine dome plus a concrete building. The internal vessel (or the intermediate containment using the French terminology) consists of the reactor vessel, the closure head, and all penetrations of the closure head.
The guard vessel (or the primary containment using the French terminology) is the second vessel around the internal vessel. The internal vessel is designed to resist a release of 800 Mecajoules of mechanical energy, using unprotected loss of flow (UPLOF) transient as the basis. Also as a part of the primary containment is a machine dome which seals around the closure head. This machine dowe is designed to hase a leakage rate of less than 1%/d at 3 bar.
In addition, there ir a building surrounding the entire structure which looks from the outside as a large dry PWR containment. However, this building is not designed as high pressure containment.
It is strictly to prevent damage from aircraft impact.
Advanced Reactor Designs Minutes December 13, 1988 Another foreign plant is the Kalkar (SNR-300) in Germany.
It is
'36 MWt, three loops and three steam generators liquid metal
/
reactor. The control system consists of a primary control rods that drop by gravity, and 3 secondary articulating' rods that uses spring. insertion. The scram failure probability has been calcu-i lated by GRS to be 10-7 and by Interatom to be 10-9 The contain-ment structure consists of a double vessel designed to resist the release of 370 megajoules of energy using the UPLOF event as the l
limiting event. Surrounded by another building is a concrete building which is inerted and designed to withstand 0.3 bar pres-sure, and retain fission products for 14 days. Around that build-ing is a free standing steel shell, and then a secondary building to resist airplane crash and help retain fission products.
3.
Mr. R. Landry, RES, briefed the Subcommittee regarding the staff's review and conclusions of the SAFR conceptual design. The NRC staff reviewed a preliminary safety information document (PSID) which was provided by D0E. The Staff's review is considered a pre-application review for the purpose of providing guidance early in the design process on the acceptability of the SAFR design. The draft SER does not constitute an approval of the SAFR design, but rather documents the Commission's preliminary guidance regarding licensing requirements, including the acceptability of the DOE-proposed supporting research and development programs. The staff's review has been performed under guidance of advanced reactor policy statement, severe accident policy statement, safety goals policy statement, and standardization policy statement.
Mr. Landry indicated that the schedule for completion of internal reviews of the SER will be as follows:
i RES management by 11/88 l
ACRS 12/88 I
I
c -
Advanced Peactor Designs Minutes December 13, 1988
'NRR '
01/89
- CRGR 01/89 To Commission 01/89 The SAFR has been designed by Rockwell International (RI) in cooperation with Bechtel, Inc. and Combustion Engineering (CE).
In addition, Argonne National Lab (ANL) has provided major analytical and test support.
The SAFR conceptual design consists of one or more independent power paks.
Each power pak consists of 4 modules that produces 900 MWt(350MWe). The design consists of a sodium-cooled reactor l
system that provides heat from the primary coolant through two intermediate coolant loops to two steam generators. The power pak reactor systems employs a compact pool-type design fueled by a metallic alloy of U-Pu-Zr contained within a sodium bond in a special stainless steel cladding. The design relies on passive reactor shutdown and decay heat removal systems.
The SAFR site is proposed to have seismic specifications of SSE=0.39 and OBE = 0.19 The demographic distribution is limited to 75th percentile of existing sites per NUREG-0348. DOE has proposed a siting source term based on radioactive materials l
released from melting a single fuel subassembly rather than the traditional TID-14844 releases.
No conventional containment building and no requirements for preplanned off-site emergency evacuation are proposed.
4 I
The overall objective of the SAFR program is to develop a con-ceptual design that minimizes plant cost and maximizes inherent safety. Other objectives are minimum potential for severe acci-dents and the elimination of off-site evacuation planning by demonstrating low risk. The designers claim that the SAFR concept
[^
Advanced Reactor Designs Minutes December 13, 1988 i;
uses reduced number of systems, components and structures clas-sified as safety related in comparison with LWRs. The main control room and balance of plant (B0P) items are proposed as commercial industrial grade as well as many other items associated with-the Nuclear Island (NI) such as diesel generators and cooling water systems.
Mr. Landry outlined the major differences between the SAFR and i
Power Feactor Inherently Safe Module (PRISM) design as follows:
Feature SAFR PRISM Containment non-welded seals Welded seals large, stiff head more flexible secondary building head 100%/d leak rate none Decay heat' Turbine / Condenser Turbine / Condenser removal DRACS ACS RACS RVACS Control System Control rods Control rods (non-safetygrade)
(safetygrade)
SASS (non-safety grade)
Primary Pumps Centrifugal EM Control room inside secure area outside secure one operator / module area-one operator /
3 modules Sodium Clean-up Continuous cold shutdown cold trap trap j
l Reactor Cover Gas Continuous purge shutdown purge Turbines One per module One for 3 modules I
Steam Cycle Superheated Saturated There are two shutdown (SCRAM) systems utilized in SAFF..
Neither is currently safety grade. The automatic plant trip system (APTS)
Advanced Reactor Designs Minutes December 13, 1988 can drive in all six of the primary control rods, which have a net worth about 3.6% a K (i.e. $10.2), and can interrupt power to the electromagnetic latch and drop three secondary control rods into the core, with a net worth about 2.4% A.K (i.e. $6.7).
In addi-tion, the three secondary rods can be dropped in by the Self-Actuated Shutdown System (SASS).
SASS is based on a temperature-sensitive Curie Point magnetic material that loses its magnetic properties at a higher than normal operating temperature (i.e.
1050 F) and thereby releases the secondary safety rods whenever the core outlet temperatures exceeds that temperature. Jet pumps are utilized to divert hot sodium from adjacent fuel channels into the magnetic materials in order to assure a timely response from SASS.
In reviewing the SAFR design, the Staff developed definitions for various event categories (EC) to be used in the evaluation (which, in general, correspond to traditional LWR event categories), such as: EC-I: Abnormal Operational Occurrences, EC-II:
Design Basis Accidents, EC-III: Severe Accidents, and EC-IV:
Emergency Planning Basis Events.
It should be noted that DOE has made the selection between SAFR and PRISM for liquid metal reactor designs. The PRISM concept with GE as the prime contractor was selected.
4.
Mr. J. Flack, NRC/RES, briefed the Subcommittee in regard to the NRC staff's PRA review.
The SAFR PRA is a level three conceptual PRA which includes the systems, containment and consequence analy-sis.
I t does not consider sabotage, startup accidents and l
l 1
l lE__
e
+
1 l'
^
. Advanced Reactor Designs' Minutes December 13, 1988 L
l L
accidents at power levels other than full power, and multi-power pak interactions.
It also does not include accidents related to radioactive sources outside the reactor. vessel, external events' other than limited seismic, and normal' plant effluent releases.
l All internal initiating events were consolidated into the following five generic LMR initiators:
Loss of flow 0.2/yr
- Reactivity insertion 0.1/yr Loss of normal heat removal 0.18/yr All remaining demands on_ reactor shutdown system 1.0/yr
- Combined transient overpower and loss of flow 2.5x10-3/yr.
Mr. Flack indicated that the SAFR PRA has provided a preliminary overview of.the plant's vulnerabilities.
Seismically induced
. transient over power (T0P) and loss of flow (LOF) sequences domi-nate the early fatalities, and seismically induced loss of heat sink (LOHS) dominate the latent fatalities.
Internal events contribute less than 1% to overall risk, tir. Flack, however, j
cautioned regarding some caveats when using these risk estimates.
l
.Such caveats are:
The PRA lacks the detail and data required to substantiate the I
optimistic system reliability estimates.
External events other than seismic have not been quantified.
Three important initiating events had not been analyzed:
(a) steamgeneratortuberupture,(b)pumpseizure,and(c) withdrawal of a large number of control rods.
Source term estimates may be low for some scenarios.
\\
l Advanced Peactor Designs Minutes December 13, 1988 lack of system interactions, support systems, and power paks.
The role of the operator is not apparent from the PRA.
l Large uncertainties in the analysis could lead to mistreatment of the core response and consequence analysis.
Mr. Flack commented that deterministic engineering analyses, complemented with prototype testing on a full scale reactor module, and operational experience is essential in the determination of SAFR's safety capacity.
j i
i 5.
Mr. G. Van Tuyle, BNL, presented the results of the analyses that j
were performed by.BNL for the Loss of heat sink (LOHS), loss of flow (LOF),transientoverpower(TOP),andunprotectedsinglepump seizure accidents.
LOHS:
The feedwater pumps providing water to both of the two steam generators are assumed to lose power, causing the steam generator to dry out in 20 seconds. Heat rejection is lost. The rest of the SAFR module continues to operate as normal. BNL has performed the calculations using the SSC computer code. The SSC predicts the power to drop from rated conditions to about 6% by 400 seconds.
]
The positive feedback from the sodium is initially nullified by the negative feedback from the radial expansion. BNL assumed loss of normal cooling and the DRACS and the outside surface of reactor vessel is insulated (adiabatic heat-up conditions). With such assumptions, BNL concluded that the RACS heat removal is sufficient to prevent damage to reactor.
If RACS air flow stopped, then there is 3/4 day before major fuel damage starts to occur and It day before sodium boiling.
l i
Advanced Reactor Designs Minutes December 13, 1988 LOF: This event is initiated by an instantaneous loss of power to the primary, intermediate loop and steam generator pumps. No scram is assumed and the inertially controlled coastdown of the primary pump is characterized by an initial 6 second flow halving time. As a result the power level drops off as the flow decreases. The fuel temperatures increases because of the reduced coolant flow in the core, activating the Doppler and axial expansion feedbacks. The reduction of coolant flow in the core also causes the sodium temperatures to increase. This inserts positive reactivity because of the hardening of the neutron spectrum. The eventual result is that the reactor power level transitions to a lower level. The feedbacks which initially started out strongly negative re-established a critical state after a few hundred seconds. The negative feedbacks were able to overcome the Doppler and decrease the power, while maintaining the core at elevated temperatures.
BNL concluded, however, that the plant must be scrammed to shut the fission process down completely, but, the operator would have a sufficient time to take action.
l TOP:
Doppler feedback in a metal fuel LMR is significantly smaller l
than for an oxide core, and during an event when the power in-creases (t,g. TOP), the Doppler is not effective in stopping the power rise.
BNL analyzed two different TOP initiators. The first is 20 cent TOP with reactivity ramp rate of 0.65 cents /second. The i
second is 36 cent TOP with reactivity ramp rate of 5 cents /second.
The results showed that the radial expansion feedback is the I
largest of the negative feedbacks effects and it will be the dominant controlling factor.
Consequently, BNL concluded that no fuel damage is expected during this event.
I Unprotected Sinale Pump Seizure: This event assumes that one of the two centrifugal pumps to seize during full power operation.
The other pump continues to operate, and the plant protection
r
~
Advanced Reactor Designs Minutes December 13, 1988
{'
1 i
systems fails to scram.
DOE claims that a pump seizure at full speed is a very unlikely event.
The MINET computer code was used I
to model the flow network around the core.
The seizure of one pump I
causes a drop in system impedance, and the unfailed pump will l
experience a large flow increase up to 128% of its rated condition, and the pump will cavitate in this mode. The resistance through 1
the locked rotor directly determines the amount of flow that will bypass the core and feed back into the cold pool.
The MINET model accounted for these effects and found that 65% of the system flow will bypass the core through the locked rotor.
The remaining 35%
of the system sodium flow still continue to flow through the core.
The feedbacks reduced the power level to a point where the maximum fuel centerline temperature and the maximum sodium temperature in the core is low enougn.
BNL concluded that this event would be mitigated by the feedbacks in the core with no fuel damage or requirements for immediate operator action.
Mr. Van Tuyle overall conclusions are:
- 1) the SAFR passive cooling via RACS is comparable to PRISM and it is effective and fault tolerant, and ii) the SAFR inherent shutdown systems are similar to PRISM except SAFR runs hotter, and the SAFR control rod driveline expansion is enhanced.
6.
Mr. R. Landry, NRC/RES, summarized the SAFR design bounding events.
He stated that these events are intended to bound the LMR design basis accidents and beyond design basis accidents spectrum to account for PRA uncertainties. These bounding events are expected to provide conservatism in selecting a suitable site source term.
The bounding events that are verified by the NRC staff are:
Inadvertent withdrawal of all control rods (without scram for 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br /> and with forced cooling).
i kr.
- o L
- .
Advanced Reactor Designs Minutes December 13, 1988 I
- Station blackout for 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br />.
- Loss of forced cooling plus DRACS and RACS (with 75% RACS-I' blockageafter36 hours).
- Instantaneous loss of flow from one pump (coastdown of other pump)..
- Steam generator tube rupture.
Large sodium leak.
External events.
- Flow blockage of one fuel assembly.
Mr. Landry indiceted-.that as a result of the NRC staff's review of the SAFR design, the general safety advantages of the.SAFR design are:
1)'
Its slow response to core heatup events.
2)
The inherent beneficial reactivity feedback effects associated with fuel and core expansion effects in cores with metal fuel pins.
l 3)
The capability to demonstrate by test the significant safety features and performance of the plant over a wide range of events.
However, the Staff also indicated some potential vulnerabilities such as:
l i
L-_-_-_--__-_-__________-_-_____-___-____-_-____-----___---_______--_______-__-__-_-______-_-_____--_-__----________-,
Advanced Reactor Designs Minutes December 13, 1988 1)
The positive reactivity effect from sodium voiding in a plutonium fueled core, and 2)
The use of a relatively new metallic alloy in a sodium-bonded fuel design. This issue involves the potential-for relocation of the enriched fuel following melting or eutectic' formation that could lead to reactivity induced power excursions'.
The Staff's conclusion is that the SAFR design has the potential to achieve a 1.evel of safety at least equivalent to current generation LWRs, provided the design and research and development needs are resolved.
7.
As a result of the Subcommittee discussion, some of the Subcommit-tee's members and consultants expressed some concerns in regard to the following:
Dr. Okrent questioned the NRC staff's selection of " bounding events." He indicated that the subject events do not seem to be bounding on risk.
For example, they do not seem to include failure of the reactor vessel combined with leakage in the guard vessel.
I l
- Dr. Okrent commented that core melt can lead to a significant sodium-concrete or fuel-concrete interactions and it is ironic that the melting of "one" subassembly is taken as a point of departure for the source term. This also is magnified by the incompleteness of the PRA.
Dr. Okrent commented that the acceptable maximum flood level should be re-evaluated.
L_________.
- .?
~
- Advanced Reactor. Designs' Minutes R
13-December 13, 1988 Dr. Okrent questioned the staff's information regarding 'he t
sodium void reactivity gain and the possible reactivity gain
.due to fuel movement.
Dr. Okrent stated that the French, English,.and Germans seem to place more emphasis on core monitoring instrumentation, and wondered if the NRC staff should look into this.
- Dr. Okrent expressed concern regarding the staff's interpreta-tion of "large release" in the safety goal. policy. He'in-dicated that using the staff's definition, a delayed release of magnitude like Chernobyl would not be a "large release,"
because evacuation keeps away early fatalities.
' Mr. Ward questioned the reliability and confidence level in-the Self Actuating Shutdown System (SASS). He. indicated that the jet pumps must be able to extract enough hot sodium from neighboring fuel and blanket assemblies in a relatively short time and high efficiency.
- Mr. Ward expressed some concern regarding the Sodium density reactivity feedbacks for the SAFR design. He indicated that at nominal conditions, SAFR still has a relatively high positive reactivity sodium feedback (5.9 vs -0.7 for FFTF and
-8.7 for EBR-II).
Dr. Siess expressed some concern regarding the selection and distribution of sites for the SAFR design.
- Mr. Michelson questioned.the failure modes for the SASS. He expressed some concern on the reliability of the temperature-sensitive Curie Point magnetic material that loses its magne-tic properties at 1050'F.
. Advanced Reactor Designs Minutes December 13, 1988 Dr. Siess questioned the scenario for one pump failure and not both pumps.or a common-mode failure of some sort.
Mr. Ward questioned the role of reactor operator. He indi-cated that PRA does not give a clear description or estimate for the likelihood of certain events which may or may not require operator action.
- Mr. Michelson expressed concern regarding the definition of external events and the fire hazards from sodium (e.g., 1 inch diameter sodium pipe break).
Dr. Shewmon indicated that the staff has not defined clearly what they mean by core melt for the SAFR design.
- Mr. Michelson expressed some concern regarding the interface between safety-grade system and non-safety grade systems (e.g., non-safety grade control room to control safety grade components). Also he commented that it is not clear exactly what is the difference between the safety and non-safety grade for the SAFR design.
l
- Mr. Michelson expressed some concern in regard to the degree of assurance to scram the reactor in case of a large seismic i
event. Dr. Siess shared the same concern.
1
- Mr. Michelson commented that the steam generator integrity issue is not well quantified for this design.
Dr. Avery commented that the PRA is very limited and it is not j
very conclusive to exclude core desruptive accidents.
Advanced Reactor Designs Minutes December 13, 1988 i
Dr. Lee commented that the reactivity feedback from radial expansion is not as large as the FFTF (-9.7 for SAFR vs -22.0 for FFTF), yet some of the analysis counts on it very heavily.
Dr. Lee expressed some concern regarding the staff's treatment ofrodbowing(bowinginorbowingout).
BNL and the staff considered bowing as a very small factor and was neglected in the tre.nsient analysis.
Dr. Lee questioned BNL prediction of criticality with such precision as 0.1% ak.
Future Activities The Advanced Reactor Designs Subcommittee Chairman briefed the full Committee on December 15, 1988 regarding the Subconsnittee activities.
In addition, the NRC staff and BNL representatives have briefed the full Committee on this conceptual design.
There is an expectation to write a report to the Commission on this subject at the upcoming January 12-14, 1989 ACRS meeting.
NOTE:
Additional neeting details can be obtained from a transcript of this meeting available in the NRC Public Document Room, 2120 L Street, N.W., Washington, D.C. 20006, (202) 634-3273, er can be purchased from Heritage Reporting Corporation, 1220 L Street, N.W., Suite 600, Washington, D.C. 20005,(202) 628-4888.
i