ML20006F508

Summarizes 891103 ACRS Subcommittee Meeting in Bethesda,Md to Continue Discussion & Review of Wapwr Design & Open Items Re NRC Draft SER

ML20006F508
Person / Time
Site:	05000601
Issue date:	11/16/1989
From:	Advisory Committee on Reactor Safeguards
To:	Advisory Committee on Reactor Safeguards
References
ACRS-2675, NUDOCS 9002280102
Download: ML20006F508 (25)

Text

-_

1

? {'

Q

)Wf hbf0 ~ Ab ?$

h(jjj j(@

1. ##'N

~

DATE ISSUED:

11/16/89_

ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE MEETING MINUTES /

SUMMARY

OF THE ADVANCED PRESSURIZED WATER REACTORS NOVEMBER 3.-1989 l

BETHESDA, MARYLAND l

i PURp0SE-1 Th'e purpose of this Subcommittee meeting was to continue discussion and j

review of the WAPWR (RESAR SP/90) design. The open items related to the

[

first eight chapters of the NRC draft SER were also being discussed.

[

i ATTENDEES f

ACRS NRC J. Carroll, Chairman C. Miller, NRR I

l I. Catton, Member H; Walker, NRR C. Michelson, Member L. Donate 11, NRR D. Ward, Member T. Kenyon, NRR C. Wylie, Member

'D. Persinko, NRR M. El-Zeftewy, Staff N. Trehan, NRR l

K. Connaughton, OCM f

R. Vickrey, Region IV OTHERS B. Maurer, W R. Wilson, W.

R. Orr, W M. Shannon, W W. Schivley, W,

-C. Lewe, NUS T. Van de Venne, W B. Sodanskas Bechtel R. Turner B&W MEETING HIGHLIGHTS, AGREEMENTS, AND REQUESTS 1.

Mr. Carroll, Subcomittee Chairman, stated the purpose of the Subcomittee meeting and introduced the other ACRS members. Mr.

Carroll also stated that Dr. Shewmon is not available to attend o\\

DESIGNATED ORIGINAL g22g,g2891116 Certified By [M 2675 PDC 1

AdvancedPdRsMeetingHinutes November 3, 1989 this meeting, however, he has.some questions that deal primarily with the pressure boundary.

For examplet

• Does W has any welds in the core region? If not, how will i

they avoid it?

1 1

• What are the standards for a pipe joint design to make in-

]

spection more reliable?

i a

What are the specifications for the cast stainless steel pipe elbows, valve bodies, etc., to make inspection more reliable?

• What will be the composition of the steel in the RPV?

{

• What will be the materials of construction of the steam-generators?

Westinghouse representatives agreed to address most of Dr.

Shewmon's concerns at this meeting and perhaps at an additional later meeting if it is required..

2.

Mr. C. Miller, NRC/NRR-Branch Chief, indicated that the staff is currently reviewing the W SP/90 design for a preliminary design approval (PDA) only and not a final design approval (FDA). The-staff's goal is to complete the PDA review and document its find-ings in a final safety evaluation reports (SERs). The staff

+

recognizes the fact that there will be open items left until such a time as receipt of an FDA. There may also be some policy issues (suchassevereaccidentconsiderations)thatneedtobediscussed at a later date.

3.

Mr. L. Donate 11, NRC/NRR - Project Manager, presented the current review status for the WAPWR SP/90 design He stated that the staff l-

i i

Advanced PWRs Meeting Minutes November 3, 1989 i

I is expecting the Commission to establish a new approved priority for the SP/90 preliminary design approval (PDA).

So far, the staff has completed one draft SER regarding the PRA analysis (front-end only March 1988) ar,d two draft SERs (SRP on June 1988 and March 1989). Currently, there are 107 open items that have to be re-(

solved before the PDA is issued. There are additional 53 open items that have to be resolved before the final design approval j

(FDA)isissued.

In addition, there are 99 open items that have to be resolved before the FDA is issued and/or plant' specific applica-tion.

Mr. Donate 11 indicated that there will be two more DSERs that have i

to be issued. The first one is for the PRA (Back end portion) and expected to be issued in November 1989. The second SER is regard-ing the USIc/GSIs and severe accidents, and is expected to be issued in Fetruary 1990. Three additional ACRS Subcommittee meetingsarerequestedbythestaff(January 1990,. February 1990..

and March 1990), and at the full Comittee meeting in April 1990 with a request of a Letter. The PRA decision will be made in June l

1990, with the issuance of a supplement to the final SER (SSER).

4.

Mr. M. Shannon, W/ Licensing Manager, described the review status of l

j the NRC safety evaluation of RESAR SP/90 particularly with respect to the Standard Review Plan for DSER Chapters 3, 4, 5, 6, and 8, i

and the open issues related to those chapters. Chapters 7, 9, 10, 11, 12, 15, 17, and 18 will be covered at the next Subcommittee meeting. Mr. Shannon indicated that coverage of Chapters 1, 2, 13, 14 and 16 is not anticipated as part of the ACRS Review, i

Mr. Shannon indicated that W has categorized the 107 open issues into five categories. These categories are:

e i

e

.[

t Advanted PWRs Meeting Minutes November 3, 1989 i

i Category 1 - y has provided additional clarification and believes that the response is adequate to resolve the issue, j

There are 62 open issues under this category.

Category 2-yhasrevisedapplicationto-reflectNRC l

position.

There are 22 open issues under this category.

Category 3-yhasadoptedcurrentindustrycodesand i

standards that differ from past practice.

NRC has'not taken position on these new methods.

There are 2 open issues under this category.

Category 4 - NRC review needs to be completed to determine if j

there is any issue to be resolved. There are 13 open issues j

under this category.

• Category 5 - Potential disagreement with NRC staff. There are i

8 open issues under this category.

5.

Mr. R. Orr, y, briefed the Subcommittee regarding the open issues i

of Chapter 3 of the DSER, " Design of Structures Components, Equipirent & Systems." There are 41 open issues under Chapter 3.

There ere 20, 10, 2, 4, and 5 open issues under' Category 1, 2, 3, i

4, and 5. respectively.

j Mr. Orr indicated that he considers the 30 open issues under Categories 1 and 2 have been resolved. Under Category 5,'"Poten-tial disagreement with NRC staff," Mr. Orr outlined the open issues as follows:

• Safety Classification of Safety Related Instrument Lines - The existing regulatory guides require that instrument sensing lines which are connected to ASME Class 2 and 3 process piping-i p

,--r

---o-w v

e v-

AdvancedPNRsMeetingMinutes November 3, 1989 and are used to actuate safety related systems should be constructed to ASME Class 2 or 3 requirements. yhasproposed that supports will be designed and constructed to requirements-for Seismic Category I structures and not to ASME-NF. y believes that its position is identical to the EPRI ALWR ~

Requirements. The NRC position is not known at the present time for this open issue.

• Postulated Breaks in ASME Class 1 Piping - Standard Review Plan 3.6.2 was revised in 1987 to require that pipe breaks are postulated to occur at intermediate locations in Class 1 piping runs as follows:

... where the maximum stress range as calculated by equation 10 exceeds 2.4 Sm."

Hexpectstobe able to demonstrate leak-before-break (LBB) for all ASME Class 1 piping greater than 6 inches in diameter.

It is not ex-pected that the smaller piping will be qualified to LBB.

ybelievesthattheimpositionofthenewcriterionwill result in more pipe rupture locations.

The NRC staff are requiring use of the 1987 SRP version.

• Classification of non-ASME Class Piping - SRP 3.6.2 pennits locations in non-ASME high energy piping to be defined at intermediate locations based on the results of stress analyses including seismic loads, g'spositionisthattheANSIB31.1 piping code supplemented by dynamic seismic analyses provides a sufficient basis to predict the potential locations of pipe rupture and it is not necessary to impose full seismic Cate-gory I requirements on the piping. NRC staff is requiring Seismic Category I if credit is taken for the seismic analysis in determining pipe rupture locations.

l j

s AdvantedPNRsMeetingMinutes November 3, 1989 Other open issues under Category 5 are pipe support baseplate and anchor bolt design, and testing difficulties for inservice pump and valve testing.

For Category 4, the open issues are:

• Internally generated missiles inside and outside containment -

This open issues involves a continuing review on the part of the staff and H will take no action unless the results of this review are issued prior to final SER.

• Limited design audit of containment design - The SP/90 con-tainment is a spherical steel containment vessel. Containment design has been performed sufficiently to establish the overall dimensions and general plate thickness. These overall parameters will be incorporated in the ASME Design Specifica-tion together with the design criteria documented in RESAR SP90 and all design loadings. As identified in RESAR SP/90, itisM'sintentionthatthecontainmentbeconstructedin accordance with ASME requirements. Thus, the. design as well as the construction will be perfonned by a containment vessel supplier.

W proposes that the limited design audit be deferred until there is sufficient design infonnation to demonstrate the design configuration and details.

TheNRCstaffiscurrentlyacceptingM's' position.

• Analysis standard for time history solutions and response

.l spectrum analysis - M have submitted an application for a j

plant satisfying current codes and standards.

In particular, the application utilizes ASCE 4-86, " Seismic Analysis of

l Advan'tedPhRsMeetingMinutes

• November 3, 1989 i

i Safety Reiated Nuclear Structures," which provides require-ments for seismic analyses of structures. This new standard has not been reviewed and endorsed by the staff.- Generally, i

the requirements are compatible with those in the Standard Review Plan, but there are a few areas where the requirements i

differ from existing staff positions.

y wish to use the latest industry standard for seismic analy-sis since it provides a comprehensive set of requirements.

l i

The NRC staff is currently accepting y's position.-

Mr. Orr indicated that for Category 3, the open issue is:

i

• Review of flow diagrams showing quality group classifications ofstructuressystemsandcomponents-yhavesubmittedan application for a plant satisfying current codes and-standards.

In particular, the application refers to ANSI /ANS 51.1-1983, " Nuclear Safety Criteria for the Design of Stationary Pressurized Water Reactor Plants " which has replaced the prior standard ANS 18.2.

This new standard has i

not been reviewed and endorsed by the staff.

yclaimsthattheSP/90plantshouldmeetcurrentcodesand standards, and will continue to work with NRC staff and industry standards Comittees to resolve the issue.

The NRC staff have no resources assigned to detemining generic positions on new or revised codes and standards, and continue to review changes on a case-by-case basis.

t

l AdvancedPNRsMeetingMinutes November 3, 1989 6.

Mr.J. Miller,y,summarizedChapter4."ReactorSystem." He indicated that the SP/90 reactor design has reduced specific power (Kw/Kg), moderator control with availability for long fuel cycles, and a radial neutron reflector.

It also has the gray rods for load l

follow. The number of fuel zones is increased for same discharge f

burnup, that allows 3-zone core for 18' month cycles. This would result in an increased design margins for LOCA and DNB events.

Thedesignhasalsowaterdisplacerrods(lowneutronabsorption) that are fully irserted prior to startup displacing 13% of the core water volume. These rods remain inserted until boron concentration nearsOppm(70%ofcycle). Over the remainder of the cycle, the displacer rods are sequentially and fully withdrawn in groups.

There are three open issues for Chapter 4 All three open issues are categorized as 4.

These issues are integrated N16 and excore power density surveillance and protection system, review of crit-t ical heat flux-(CHF) tests, and departure from nuclear boiling ratio (DNBR)safetylimit. These open issues involve a continuing I

review from the staff, and y will take no action unless the results of this review are issued prior to the final SER.

7.

Mr.R. Wilson,E,summarizedChapterS."ReactorCoolantSystem."

He indicated that the key design objectives are to maximize plant availability and minimize occupational radiation exposure (ALARA).

The NSSS power is 3816 MWt, number of loops is 4, operating pres-sure is 2250 psia, design loop flow is 1J0,100 gpm, hot leg temper-ature is 625"F, and steam pressure is 1024 psia.

1 P

4 i

• Advan'ced PWR5 Meeting Minutes November 3, 1989 l

Mr. Wilson stated that the advanced PWR ste6m generator incorpo-f rates features which have been extensively tested and must have already been implemented in operating steam generators.

j i

The three dominant design enhancements over earlier SG models are:

t

• Alloy 690TT tube material

[

i

• Tube bundle sludge control

• Enhanced maintenance features.

The advanced PWR steam generator meets or exceeds all of the f

EPRI/SGOG design recommendations.

There are 11 open items under Chapter 5.

Six open items are categorizedasCategory1(clarificationprovidedbyy)and5open items as Category 2 (RESAR revised to reflect NRC staff position).

8.

Mr.T.VandeVenne,y,summarizedChapter6,"EngineeredSafety Features." He indicated that the containment type for'the SP/90-l design is a spherical steel _with a design pressure of 46.9 psig.'

There are four high head safety injection pumps (HHSI) that inject through their own RCS vessel connection to provide emergency core cooling for the LOCA events, and provide RCS makeup and boration j

for all non-LOCA events. Only one of these four pumps is required for small LOCA and feed-and-bleed cooling. There are four core i

reflood tanks with low pressure nitrogen coverage that inject into l

the RCS vessel assisting the HHSI in reflooding the core following_

a LOCA. These tanks eliminate the need for active low head SI i

i pumps.

.i 1

~

10-November 3, 1989 Advanced PWRs Meeting Minutes i

The SP/90 design has en emergency water storage tank (EWST) that

.s l

provides a means to reduce the containment cleanup resulting from discharge from the pressurizer' relief tank rupture disc and the hot l

1eg vent path, or the steam generator overfill paths.

j 4

There are also four accumulators with a tant volume of 2500 cu.

a ft. and operating pressure of 600 psig.

There are 6 open issues for this Chapter; three as Category 1, two as Category 4, and one as Category 5.

The Category 4 open issues are " minimum containment pressure analysis for performance capability studies on the ECCS." and

" containment presture 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after accident."- For these two open-issues, y does not plan to take any action unless the results of a

the staff review are issued prior to the final.SER.

'i The Category 5 open issue is " hydrogen purge and. vent system."' y's position is that in the SP/90 plant design, backup to the in-l c.catainment electric hydrogen recombiners is provided by igniters, l

a:G::h are designed to mitigate a: 100%Zr-waterreactiod.-'No separate containment hydrogen purge system is provided,'nor is this function specifically assigned to the operating purge system which o

is, therefore, not necessarily_ designed'in accordance with RG 1.7.

However, in the extremely' unlikely ' event of coincident failure of the redundant in containment electric hydrogen recombiners;and.the-hydrogen igniters, the operating purge system coul'd also be used to.

control long term hydrogen buildup.

The. staff.has indicated thatithe operating purge' system should be--

- designed to meet the requirements of RG-1.~7.

In: addition,' the__

staff has indicated that they will provide additional guidance l

regarding the need for a " hardened" venting capability.

m 61**-v~

w

-we9

+

e

i

'. Advanced PWRs Meeting Minutes November 3,:1989) 1 9.

Mr. T. Van de Venne, W, summarized Chapter 8. " Electric Power." He_

indicated that the AC power system for the SP/90 design consists of.

the following:

?

• Main generator breaker 7j

' Auxiliary and standby transformer

• ESF transformer with capability to supply one division of ESF.

d loads

[

• Two Class 1E buses, each with one Class IE diesel generator

['

• One small non-1E diesel generator to power seal injection pump snd charge batteries during station blackout..

1 The DC/ instrument AC power systems consists of the following:

n

' Four Class 1E batteries with associated Class.IE-ch'rgers, a

inverters and panels

• Two Non-Class I batteries with associated'Non-Class 1E. equip-ment l

l

' Four Class IE instrument buses j

' Two Non-Class 1E instrument buses.

l l

There are five open issues for Chapter 8.

Two open issues as i.

p Category 1, two open issues as Category 2, and one open issue as.

i Category 5.

The Category 5 open issue is, " Station blackout (USI A-44)." W's position is that the' proposed resolution for the SP/90 plant includes a small diesel generator independent of off-site and

...e.

. ~ _..

I

~

• Advan'ced PWRs Meeting Minutes November 3, 1989 1

on-site AC peser supplies, whose primary function is to power a positive displacement pump providing backup seal injection to the reactor coolant pumps. This power source can also be used to recharge the Class IE batteries, which will be depleted in_about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. Thus, continuing' operation of key instrumentat' ion and control systems is assi J.

The staff is concerned that over time, the environment in rooms containing electrical,-instrumentation,'and control equipment (e.g., emergency. control room. protection system rooms, battery _

rooms, inverter rooms, etc.) will deteriorate'due to lack of ventilation to the. point where equipment would fail.

For this reason, the staff would like to see the size of the third diesel generator increased in order to allow continuing operation of selected HVAC systems.

10. As a result of the Subcommittee discussion, some of the Subcommit-tee's members expressed concerns-in regard to the-following:-

• Dr.CattonexpressedconcernthatneithertheNRC: staff'_nor}l, has carefully analyzed the flow instabilities.and~ vibration-in-

~

the steam generators to investigate potential fluid structural ~

interaction problems.

• Mr. Carroll ques'tioned the significance of the-PDA approach.

In addition, Mr. Michelson expressed some concern regarding the definition of PDA and FDA in regard to the design and what-kind of study and analysis will be included under each defini -

tion, i

Dr. Catton questioned the use of the ill-documented computer-codes such as MAAP or THINC codes.

1

i

,c Advan~ced PWRs Meeting Minutes.

l November 3.-1989

• Dr. Catton requested W to supply him with the WCAP-reports and

]

the insights gained in the severe accidents issues.

I

• Dr. Catton requested additional infonnation regarding the-resolution of open issue No. 31, which. deals'with flow-induced vibration testing for non-prototype plants.

• Mr. Ward requested additional information regarding open-issue-No. 22, which deals with the Containment design criteria:for the SP/90 design,

' Mr. Michelson requested additional information~regarding open issue No. 16, which deals with limits of break exclusion area for ASME Class 2 piping.

}

' Mr. Michelson questioned the resolution of open issue No '41, which deals with equipment qualification, and especially W's l

categorization as Category 1.

• Mr. Carroll indicated that the interaction between W and EPRI-regarding the compliance with the EPRI requirements document I

is not very clear at the PDA stage. Mr. Michelson agreed.-

Dr. Catton indicated that throughout the SP/90 design; the che.

valves performance, piping and location' was not analyzed ;

L carefully.

l

[

1 l

• Mr. Wylie questioned the philosophy of W that after using four 1

j steam generators, four safety trains, etc., now has changed i

its-design for the SP/90 to use only two electrical: trains.

• Mr. Wylie commented that the SP/90 design does not provide any j

protection against station grounding and lightning.

- l l

l

~

j Advanced PWRs tieeting Minutes November 3, 1989' l

l

• Dr.CattonrecommendedthatWmightconsiderstudyingtheeffectofL ignitors for the SP/90 design under accident conditions. Dr.

Catton supplied W with_a report.(attached) titled, " Ignitors to Mitigate the Risk of Hydrogen Explosions - A critical Review,'" by-

]

Dr. Ing Helmut Karwat.

{

• Dr. Catton urged W to study the effects of an unexpected ~ blowdown' that results in a pipe rupture and removal of the isolation mate-

-i rials from the pipe system. Dr. Catton~ referred to a typical-damage that was experienced inside the HDR-Containment facility in November 1983.

FUTURE ACTION The Subcommittee Chairman and members decided-to conduct another Subcom-mittee meeting approximately mid-January 1990, to continue discussion of --

the subject matter.

[

NOTE:

Additional meeting details can be obtained_from'a transcript i

of this meeting available in the NRC Public Document Room, 2120 L Street, NW, Washington, DC 20006, (202) 634-3273, or t

L can-be purchased from Ann Riley and Associates, Ltd., 1612 K Street, NW, Suite 300, Washington, DC 20006,l(202)293-3950.

l I

i

Ignitors to Mitigate the Risk of Hydrogen Explosions -

A Critical Review h

Prof. Dr. Ing. Helmut ' Karwat Lehrstuhl fOr Reaktordynamik u'nd Reaktorsicherheit s

Technische UniversitEt Munchen I

l Abstract i

Risk analyses and the accident at the Three Mile Island plant (TMl-2) have shown that the formation of large amounts of hydrogen during so-vere accidents' poses a real danger to the contalnment -integrity. Proper means must be taken to prevent the occurance of global or sometimes even=

local detonations. The applicability of ignitor systems to protect-large dry -

PWR containments is. critically discussed. Although already adopted for Mark lil-BWR and Ice-condenser PWR-containments -the reliability ~ of < park.

plug ignitor systems to protect large dry containments from.'the possible '

consequences of a local or global detonation. ls neither proven.experimen-tally nor analytically. An experimental study of possible post-accident inertisation procedures deems necessary and may yield a more' convincing -

mitigation procedure.

'!a l

1 The Problem Since several years it is well known, that the formation of large amounts -

l of hydrogen accompanies severe reactor-accidents which involve core deg-d radation. The hydrogen will be released Into the containment and' may pose an early threat to the containment integrity. Depending on the se-1 quence of events and the applied analytical' methods to predict the hydro--

i gen production an early release of up to 450 kmol (approx.10000 Nm8 at standard conditions) hydrogen into the containment has been evaluated i

/1-5/.

l I

2 i

e e*

o-I Core-concrete interactions resulting from a fully developed core melt acci-

)

H dent may additionally yield the same magnitude of hydrogen, possibly en-

)

I I

riched by carbon monoxide. According to the local distribution for. com-

]

bustible gases the ignition may result in a deflagration or even detonation.

The hydrogen concentrations averaged over the free volume of the con-l tainment may' reach values between 7 and 16 percent or even more. Local _

concentrations may be much higher, in particular if steam condensation is realistically taken into account. it is concluded that within the large ge-ometries of PWR-containments a slow laminar deflagration would be very unlikely, in most cases highly efficient combustion modes must be expectact.

in view of the radiological consequences associated to an early con-

~

j tainment failure within the probabilistic risk assessment this situation is obviously not acceptable. Steel shell containments in particular a e chal-lenged not only by the possibility of. a global (eflagration or detonation.

Local detonation may result in the partial desintegration of c.ontainment '

structures inside the containment leading to the formation of missiles which may damage the relatively thin steel shell. The 'latter case-is more likely to occur, because the formation of high local hydrogen concentra-tions close to the location of release can never be excluded.

if internally generated missiles hit the steel sheel a local mechanical dam-age of the steel shell may occur. If the mechanical damage coincides with a global load caused by internal pressurization a catastrophic failure of

~

l the entire containment is possible.

l Massive pre-stressed concrete containments or concrete containments' l

which are equipped with a steel liner may be some what more favoureble in forgiving the consequences of local detonations. According to the mass ratio of concrete to load bearing steel rebars the internally generated l

missiles may only damage the liner but not necessarily cause catastrophic-failure of the steel rebars. In general it is anticipated, that concrete con-'

tainments are mainly challenged by global detonations involving 'the en-tire free volume. Furthermore the 1:6 scaled Sandia experiment has shown, that. concrete containments may indead fall in the " leak before l

rupture" mode if a global internal pressure load reaches the ' ultimate load capacity. This was not observed for the 1:8 scaled-steel vessel experi-ment which desintegrated catastrophically /6, 7/.

o 3

L...-

It' is therefore necessary, to assess the behaviour o' a containment re-s,ulting f, rom en hydrogen combustion in close consideration of the structur-al typicality of the containment. Conclusions drawn for one type of a comalnment will in general not be applicable to any other type.

For densely populated areas the possibility of an early containment failure caused by an hydrogen combustion must be eliminated. Recently, the appil-cation of spark plug or glow plug ignitors for the controlled ignition of the 'nydrogen has been proposed even for large dry steel shell contain-ments. The reliability of this method is highly questioned, because varia i

ous uncertainties exist which may turn the benefit into a negative-impact.

I The Analytical Capabilities for Combustion Predictions The application of controlled ignition requires that the combustion process must be predictable for any case of its ' activation. A prediction must.

show, that the integrity of the containment will not be challenged by any turbulent deflagration caused by the incidental or deliberate ignition of a mixture of hydrogen, air and steam. Moreover, also highly energetic lo-cal deflagrations must not damage internal structures of steel containments leading to the formation of internal missiles.

The analytical prediction of a hydrogen combustion in the large multi-compartment geometry of a full pressure containment is a very ' complicated matter. Many experiments have been made and models were d'eveloped' to i-

- stigate the ccmbustion in simplified geometries and under simple hy-draulic conditions. Silght changes of the conditions prior to or in conse-quence of the ignition and combustion may lead to considerably different results. These uncertainties stem from the fact that the:large scale com-bustion is governed by turbulence effects., The combustion creates local turbulences and the turbulences themselves enhance the combustion pro-cess, in particular for rich mixtures and within complicated geometries.

Various analytical models to calculate a laminar burn process in single volume geometries for lean mixtures are based on empirical correlations re-sulting from the interpretation of the observed burn phenomena and may re-produce such experiments. Some more sophisticated fluiddynamic chemical-models exist, which are mostly derived and tested for the application to r-i

?'

I

g ----

.,.4 4

corhbustion engines simulating the optimal turbulent combustion -inside-a small volume of simple geometry (e.g. a motor cylinder).

r I

Computer programs which may reliably predict the possible combustion modes of a hydrogen-air-mixture within the complicated geometries of a containment are not available. Several years ago first steps have been-done to develop such models. The GRS-developed code RALOC/ COMB has been published recently /8/. As a first step a. simulation procedure has been applied for laminar combustion inside a simple compartment without turbulence. Later on, an attempt was made to combine several such com--

partments into a procedure for the simulation of multi-compartment 'ar-

=

rangements. Experimental verification was never made for such conditons..

1 Many other codes employing combustion models have been generated worldwide which have the same principal deficiencies. From this situation i

it is concluded, that without further analytical development reliable simu-I lation models for the prediction of hydrogen combustion processes inside a subcompartmented PWR-containment are not at hand.

Nowadays combustion experiments inside the complicated geometries typical for containment systems are also not available. The experiments performed inside the Nevada Test Side (NTS) facility to study combustion effects have been performed with a nearly empty single room containment and at -

relatively low hydrogen concentrations /9,10/. Those results cannot be1 directly applied to the geometric conditions of a real PWR-containment and I

do not account for the possible higher local concentrations of combustible-i gases. Effects of the scale, the concentration profile and the feometry do

_ j not allow a direct extrapolation of the NTS test results to-full-size reactor conditions.

l Concerning the predictability of the distribution of hydrogen air-steam-mixtures prior to ignition the situation is more favourable. Several natural:

convection driven distribution codes are available worldwide.

l The advanced code R ALOC has been developed under BMI/BMU sponsor-ship for many years. It was verified on the basis of several experiments.

However, gas-distribution experiments have been executed only for sim-plified compartment configurations. The most -important experimental facili-ties are the HEDL (Hanford, USA) and the Battelle-Frankfurt test' rigs.

51-1 p

- The HEDL-experiments have been performed within-a containment which.

was sepa' rated into two subsections. Somewhat more-sophisticated were the -

compartment configurations of the Battelle-Frankfurt - experiments. From the latter exoeriments only the tests no.16 and 17 have been analytically assessed in detail and served as basis' for ' the verification of the

(

R ALOC-Code. Other hydrogen distribution experiments have recently been performed within the complicated geometry of the large-scale HDR-con-tainment. One experiment was closely-coupled to a preceding blowdown experiment and is now analytically assessed. A comparison between pre-test predictions and experimental evidence may be expected within the I

next future.

l.

From this situation it is concluded, that the gas distribution processes itself is considered as predictable also for -large. scale plants - and.the analytical simulation models are expected to be more and more experimen-tally assessed in the future.

Mechanical Damace of lanitors Should ignitor systems be selected to control the combustion of hydrogen it is necessary to assess the limited availability of ignitors at certain loca-tions inside the containment. Many accident sequences which. result in a i

hydrogen problem will be initiated by a ' loss-of-coolant-accident. In all-l these cases the location of the pipe rupture inside the containment will also be the location of the hydrogen release originating from the core dis-integration, it is this location, at which installed ignitors are assumed to be efficiently activated. Two problems must be addressed which may seri-ously limit the benefit of ignitors.

The formation of hydrogen during core-degradation requires the avall-i ability of sufficient water vapor inside the primary coolant system to sus-tain the chemical reaction at the hot metal' surfaces. In general, a surplus of vapor is expected which :during the blowdown period and/or during an associated feed and bleed mode of operation of the reactor systemtwill be c

released into the containment together with hydrogen. At least in the vi-cinity of the release point the partial pressure of the water vapor tempo-rarily and locally will be high enough to prevent ignition. Later and at more remote locations will condensation reduce the probability of tempo-y

i rar'y vapor inertisation and allow ignition. A careful analysis of the hydrogen-air-steam distribution will support this concern.

The second problem involves the potential for severe local mechanical da-mage to structures available in the vicinity of pipe ruptures which must i

be taken into account.

I E

A typical damage caused by an unexpected pipe rupture was experienced inside the HDR-containment facility in November 1983. An unexpected -

blowdown happened on occasion of a series of thermal shock experiments.

The blowdown caused extensive damage in the immediate vicinity of the pipe rupture and the hydraulic forces completely removed the isolation material from the pipe system, isolation material has been distributed throughout the containment, The distribution of damaged structural mate-rial and various sorts of isolation material has been studied-in detail by the USNRC. The results have been ~ documented in a specific NUREG-report /11/.

Concluding from these observations one has to assume that those ignitors, _

in particular spark plug ignitors, located in the vicinity ~of a pipe rup-j ture will be damaged and may fail to function. Hence, for most cases it is expected, that hydrogen can only be ignited-if it has been transported to ignitors which have been installed at more remote locations. There, the i

Ignition may take place, as soon as the hydrogen concentration has reached the ignition threshold. This would then result in a flame propa-i gating from lean mixtures into areas with rich mixtures. Flamefront accel-eration together with local turbulence may then increase the potential for local detonations rather than. reduce it.

It has been proposed to protect the ignitors from the mechanical impact of a local blowdown. An efficient protection of the -ignitors;however may de-Jay or exclude the ignition of the attributed local gas-air-mixtures consid-erably, if at all, ignition will be delayed and accompanied by. additional

]

local turbulence, i

Anyhow, an experimental demonstration of the benfit of ignitor systems would be necessary. If this experimental demonstration should be con ~vinc-ing, experiments should also address the above mentioned mechanical prob-o

.7 s

lems. Such e>.periments must be performed under the realistic conditions 1.*

of a pre' ceding b owdown where the' location of the rupture must.not be preconditioned against the mobilisation and distribution of-isolation materi-al and of other damaged equipment.

~

Application of lenitors in USA Within the USA two types of containments have been equipped with sys-tems for controlled Ignition. These are boiling water reactor pressure suppression system containments of the MARK lll type and some pressur-ized water reactor containments provided with ice condensers to red"M-the global pressure built-up. Both types of containments are typically -

subdivided into two main aree.

Those areas which host the primary cool-ant system are separated frw. the rest of the containment. For the' MARK lli containments it is the water pool which provides physical sepa -

j ration, for the ice condenser containment the ice box arrangements form the boundary. The areas containing the primary coolant system are de -

pleted from air during the preceding blowdown process. The air is accu-mulated downstream the water pool or the ice condenser system.

in case of the release of large amounts of hydrogen during a severe acci-3 dent sequence the hydrogen initially will accumulate Inside-- the steam inerted area. It can reach the air-enriched areas only via a predetermined flow path.

)

Thus, ignitors which have been installed downstream the Ice condensers.

respectively downstream the water pool are efficiently protected from the immediate mechanical impact of the local pipe rupture. They' may ignite reliably if inflammable concentrations have been reached, in all cases it is; anticipated, that the overall structure of the containment system has not been destroyed by the initiating loss-of-coolant-accident, an assumption.

which is also consistent with the corresponding probabilistic risk assess-ment findings.

Under such conditions the implementation of the controlled ignition ap-pears to be acceptable even if the predictability of.the activated combus-tion processes is less than vague. However, the application of controlled

i "

y 4

l Ignition 1nside the MARK lli containment has been extensively studied by-pilot experiments preformed by the Hydrogen Control Owners Group (HCOG) inside a linearly 1:4 scaled mock-up of a MARK lll containment.

Many experiments compensated the lack of reliable analytical predictions 1

for this pt~ticular containment type. Intensive scaling considerations have supported the applicability of the experimental evidence to full size reac.

tor plants. For the large dry PWR containment with ice condensers such experiments are not known so far.

By reasons explained above the acceptability of controlled ignition -inside the above mentioned US facilities is therefore not of any relevance for the acceptability of ignitors inside the large dry PWR-steel containments.

Conclusions 1

The presented arguments demonstrate that the controlled' ignition of large amounts' of released hydrogen inside a large dry PWR steel containment may not be benefical. A detailed assessment of merits and negative im-pacts would overwhelmingly show more disadvantages than merits.-

i The disadvantages are so evident, that other possibilities to eliminate the j

existing potential for a hydrogen explosion have to be studied.

i in this context it has been proposed to study in detall~ the possibilities of -

the post-accident inertisation of large dry PWR containments by the injec-tion of liquid nitrogen or carbondioxide after a loss-of-coolant-accident.

Such proposals are presently under discussion. An opportunity exists to include post-accident inertisation procedures into the phase lli -of the large scale HDR experimental program which is planned for! the next years. The proposec' experiments should be aimed to study the thermohy-i draulic effects of liquid nitrogen or_ carbondioxide injection, the distribu-tion of an efficient inertisation and the available possibilities to reduce a permanent pressure build up-inside the containment.

Post-accident injection of nitrogen or carbondioxide would result in a per :

manent increase of the containment pressure if the air initially present inside the containment has not been removed prior or during the Inertisa-J tion procedure. Therefore additional measures are necessary to compen--

P:i =

.. [i.. -'

[:*

e -

sate this disadvantage. The effects of a delayed closure of the main ve'ntilation lines must be assessed and studied as well as other methods to reduce the pressure load to the containment.

Controlled filtered venting could be utilized in all those cases which w1ll-y not develop into a severe accident sequence with strong aerosol and fis-I sion product release. According to probabilistic risk analyses 99,99 % of

. r-all initiating loss-of-coolant-accidents will proceed as designed with a min-

{

imum release of fission products and. aerosols into' the containment. A1 moderate pressure increase could easily be mitigated by filtered venting l

which again will reduce the containment pressure over..the time to accept-l ably low values. Thus, filtered venting if combined with. post-LOCA inerting could help also to mitigate the real ' threat - of a. hydrogen explosion - with the potential of a subsequent disintegration of large dry -

steel containments.

Garching, 11.01.1989 I;'

'j.-

VM l

~

l 1

i

.l

r-i.

i.'

i s.

R eferences

/1/

Allen L. Camp et.al.;

Light Water Reactor Hydrogen Manual, NUREG/CR-2726-August-l 1983

/2/

Commission of the European Communities; Hydrogen Generation, Distribution and Explosion Potential Assocl-ated with Accidents in Light. Water Reactor (A State-of the-Art Re-port), update September 1985; CEC Document Xil/107/1985 -

/3/

F.W. Heuser; I

Risikountersuchungen zu Unfullen in Kernkraftwerken GRS; Fachge-spruch 1986, S. 44 - 58, GRS-Bericht Nr. 64, MMrz 1987

-j l

/4/

Norman C. Rasmussen (Chairman);

Technical Aspects of Hydrogen Control and Combustion in Severe Light-Water Reactor Accidents; Report prepared by the Committee on Hydrogen Combustion', National Research Council; National Academy Press, Washington DC,1987 -

iq

/5/

D.F. Torgersen (Chairman);

i Source Term Assessment,- Containment Atmosphere Control. Systems and Accident Consequences; Report to CSNI by a Group of. Ex-perts, CSNi-Report 135, April 1987 l

/6/

Larry N. K6nig; Experimental Results for a 1:8-Scale Steel Model Nuclear Power Plant Containment Pressurized to Failure NUREG/CR-4216; December 1986 l

/7/

D. Horschel;

]

~

Synopsis of the Results of a Test of a Reinforced Concrete Con-tainment. Proc. 4th Containment Integrity Workshop; 1988 a

i

-r,

.. -(

/8/

T.V. Pham; C0MB - Ein phEnomenologisches Rechenmodell fur vollstEndige H '

2 Verbrennung innerhalb des Reaktorcontainments -

GRS-A-1070, MHrz 1985

/9/

A.C. Ratzel; Data Analyses for Nevada Test Site (NTS) Premixed Combustion Tests; NUREG/CR-4138, May 1985

/10/

R.T. Thompson et al. ;

Large Scale Hydrogen Combustion Experiments Vol. 1 : Methodology and Results, Vol. 2: Data Plots EPRI NP-3878 Vol.1 and Vol.2, October 1988

/11/

A.W. Serklz (USNRC);

Containment Energy Sump Performace

-f NUREG-0897 (Rev.1); October 1985-il 1

.i I

l o

1

.