ML20092C985

Technical Rept, Monitoring Hydrogen Gas in Containment During Early Phases of Severe Accident

ML20092C985
Person / Time
Site:	Arkansas Nuclear
Issue date:	12/31/1991
From:	ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY
To:
References
FACA, NUDOCS 9202120229
Download: ML20092C985 (28)

Text

,

Technical Report MONITORING HYDROGEN GAS IN CONTAINMENT DURING THE EARLY PHASES OF A SEVERE ACCIDENT

-Z_Z-Entergy operat ons Prepared by ABB Combustion Engineering Nuclear Power for Entergy Operations Arkansas Nuclear One Unit 2

December 1991 0

U/P4

{$88258SEl UbShge

o IARLL0f_CORIMIS EXCIIQM EAQE 1.0 Purposo 1

2.0 Backgrcund 2

3.0 Design Basia Accident (DBA) Considerations 3

4.0 Doyond DBA Considerations 6

4.1 Conoral Critoria 6

4.2 Survey of Measurablo Paramotors 7

4.2.1 Core Exit Temperature 7

4.2.2 Containment Radiation Levels 8

4.2.3 Ex-Core Detector Readings 10 4.2.4 Other Measurable Paramotors 10 4.2.5 liydrogon Monitoring in Containment 11 5.0 Procedural Considerations 14 5.1 Current EPGs/EOPs 14 5.2 Sevoro Accident Management 15 6.0 Summary 17 L

e APEfiDLCES FAGE Appendix A

A Study of Coro Wido Cladding A-1 Oxidation and Hydrogen Release During Donign Basia LOCAs LLSLHF TA01LS TADLE EAQE 1

Description of Accident Sequences Analyzed 18 2

Survey of Scenarios to Estimate Hydrogen 19 Concentration in Containment LLSl_OF FIGURES ELQ9BE Eh9E 1

Hydrogen Concentration in containment 20 versus Time 2

Containment Hydrogen Concentration as 21 Indicated in the Control Room 3

Core Exit Temporature vercus Time 22 4

Typical Analynis for Post Accident Dose 23 Rate Inside a cylindrical Containment l

I

4 MONITORING llYDR0 GEN GAS IN CONTAINMENT DURING Tile EARLY PilASES OF A SEVERE ACCIDENT December 10, 1991 1.0 PurRSAA The fundamental purpose of this paper is to investigate the usefulness of monitoring flydrogen gas concentration in containment during the early phases of a sovoro accident.

The timo framo considered in this study in from time zero until approximately 1.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> into the accident.

1

.J

2.0 Backaround A severo accident may be defined as an event where prolonged core uncovery has resulted in olevated temperatures and corresponding damage to the coro.

The subsequent creation of largo amounts of Hydrogen due to the Zirconium-water reaction that occurs at such temperatures is a direct consequence of such scenarios unless the core is for some reason " starved".

NUREG-0737, Item II.F.1, Attachmont 6 requires continuous indication and recording of Hydrogen conce' tration in the containment atmosphere to be functional within thirty minutes of initiation of safety injection.

Although the basis for this timo requiremont is not explicit)" provided, it may be inferred that the objective is to qui =xly provido plant personnel (operators, management, Tsc personnel, etc.)

with early indication of whether a sovere accident may be in progress.

In addition to the measuremont of Hydrogen in containment, there may be other plant paramotors that could bu more easily or more rapidly assessed on a quantitative basis by t's operators.

Use of these other paramotors may be effective in reducing the time for operator action during an accident scenario that demands prompt operator response.

The purpose of this paper is to investigate this issue and to formulate conclusions that may be used to defend a request-to the-NRC to permit initiation of Hydrogen monitoring at a more reasonable time of, say, sixty to ninety minutes.

The NRC has previously been approached to relax the thirty minute requirement based on historical DBA

. arguments where significant Hydrogen accumulation in containment occurs.only over a period of days.

Those requests have been disapproved, with the most recent correspondence citing that only DBA arguments had been prosented.

The rationale to be developed herein takes into consideration events that are wol,', beyond the design basis.

For such postulated events, measurable quantities of Hydrogen will be produced early in the accident sequence.

c However, arguments are developed to show that, in the time j

frame of interest, the measurement of these quantities of L

Hydrogen is of essentially no use in such a rapid event.

2 L

-j

l 3.O DeD10n BaAID Accident-(DDA1_C.pngldcIA011png l

ABB C-E has reviewed t i issue of Hydrogen generation and the resulting Hydrogen,oncentrations in containment that may result from varying degroos of Zirconium oxidation in a DBA.

This review, entitled, "A Study of Core Wido Cladding Oxidation and Hydrogon Roloase During Design Basis LOCAs",

was conducted based on the Arkansas plant.

Tho largo and small break LOCA ovents studied are the only FSAR events for which Hydrogen is explicitly calculated as part of the associated core uncovery.

Hence, they are excellent starting points for breaking out important phenomena concerning beyond DBA ovents that will be discussed lator in

' tis paper.

The above montioned study is provided as.

Appendix A to this paper and a synopsis is provided in the following paragraphs.

Within the design basis of a Pressurized Water Reactor there are strict licensing limits that constrain the amount of Zirconium-water oxidation that is allowed (e.g.,

10CFR Appendix K, where the maximum fraction of Zirconium oxidation allowed is 1%).

For a typical C-E coro, oxidation of all the Zirconium (and assuming that all of the Hydrogen is transported to the containment) will yield a containmont Hydrogen concentration of about 20% by volume.

The it limit for the oxidation of Zirconium referenced above thus corresponds to a 0.2% Hydrogen concentration in containment.

For a 0% - 10% Hydrogen monitor scale, 0.2% is just at or possibly slightly above the approximate threshold of visual obse rvation.

Theoretically, for a Hydrogen burn to occur in containment, localized concentrations of approximately four volume porcent would have to be reached.

In general, the Hydrogen will originate from four sources:

1.

Zirconium clad and other Zirconium in the active l

core region that reacts with water and steam 1

2.

Radiolysis of water from the decay of fission products 3.

Corrosion of other metals and materia 10 in containment 3

l 4.

The limited amount of Hydrogen icutinely present in the RCS during steady state operation.

For a DBA, the primary source of Hydrogen early in the evnnt will be from Zirconium clad oxidation (Source #1 abovo).

Data provided by Arkansas pertaining to Hydrogen buildup in containment following a DBA large break LOCA confirms that after approximately the first two hours of such an event, 85% of the Hydrogen that will have been produced originatos from the oxidation of Zirconium clad in the active fuel region.

The Hydrogen in containment data provided in tho ANO FSAH is based on a very conservatively chosen initial amount of cladding oxidation (five timos the 1% limit referenced earlier), and includes the results of reaction rate calculations for hydrolysis and corrosion of the metal surfaces in contilnment.

As shown, days into the event, the i

Hydrogen concentration approaches deflagration (burnable) levels but the use of a single recombinor is easily able to provent the minimum theoretical burn limit (=4%) from being reached.

t A single recombiner at 100 CFM and 95% efficiency can remove about 200 SCF per hour of Hydrogen from the containment at a containment concentration of 3.5%.

By contrast, the sum of all reaction rates together produces less that this (about 125 SCF per hour at the timo 3.5% Hydrogen concentration in containment is reached).

Hence, a single recombiner is adequately sized to handle the DBA event.

As will be shown, this is not the caso for the early phano of a-severe accident since the Hydrogon generation rates for those situations is much larger than the recombiner capacity, a

Figure 1 shows the concentration of Hydrogen present in containment, and Figure 2 shows the Hydrogen concentration as indicated in the control room for a number of modeled accident scenarios at Arkansas (both.DBA and beyond DBA-events).

Figure 2 includes a delay time of thirty minutes to reflect the delay expected from the time a Hydrogen sample is drawn from containment to the time that samplo reaches the instrument for analysis and indication-becomes available to the operators in the control room.

Table i provides a description of each of the casos being considered in the analysis, and includes both DBA and beyond DBA events.

4 4

.a

The two design basis cases considered were large break bOCAs with full or partial injection capability and the design basis (three out of four) Safety Injection Tank (SIT) availability.

The calculations were performed using the MAAP code for a generic c-E plant design with the data properly scaled to reflect an Arkansas containment free volume of approximately 1.86E+6 cubic feet.

The figures clearly show that, with a thirty minute time delay between the drawing of a sample from containment and the indication of sample results in the control room, the design basis loCA events (Cases 2 and 3) do not yield Hydrogen concentre ions in containment prior to 1.2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, and will not yield measurable data in the control room AL_All during the first 1.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> of the event.

Therefore, it seems clear that the NUREG-0737 time requirements for Hydrogen monitoring are based on accident scenarios that progress well beyond the DBA envelope.

Under these conditions, ABB C-E feels that Hydrogen monitoring capability is only one of many information elements present, and will be the element with the most lag time to the operators.

Therefore, it is evident that although it is important during an overall accident sequence, the ability to measure Hydrogen will not be uniquely critical during the early phases of a DBA.

For DBA events over longer time scales, Hydrogen monitoring is clearly useful to show the slower buildup of Hydrogen (over a period of days) that can determine when recombiners should be activated so as to prevent combustible mixtures from forming in the containment atmosphere.

The same may be said for beyond DBA events occurring over longer time scales.

s

t 4.0 Bevond DBA Considerations i

j.1 General Critoria With the obvious exception of events characterized by Reactor Vessel failure, core uncovery will only occur when there is a loss of RCS integrity coupled with inadequate Safety Injection flow.

These events includes i

1.

Loss of all secondary side heat removal, without once through cooling (feed and bleed).

This event will cause the RCS mass to be depleted through lifting of the pressurizer primary safety valves.

(This event is very similar to a complete Station Blackout except that RCS leakage (e.g. RCP seal leakage) was-not addressed.

Over the times of

-interest, RCS Icakage (=100 GPM for four RCP seals, per NUMARC guidelines) is only about 14% of RCS inventory, so that it is not significant here.

Hence, the event analyzed is essentially a Station Blackout).

2.

Loss of coolant with inadequate injection capability.

Large breaks were analyzed to obtain the fastest response.

Small LOCAs would provide similar results, but over longer time frames.

LOCAs beyond DBA were also analyzed to provide continuity with the previously referenced DBA LOCA results and because such severe LOCAs will easily demonstrate measurable amounts of Hydrogen in containment within 3.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.

Th

.axt section of this report contains a survey of the various parameters that will play a part during the aforementioned beyond design basis accident scenarios and will include the measurement of' Hydrogen in containment.

The purpose of this section will be to show that, even though Hydrogen may be theoretically measurable during the time frame of interest for some very low probability severe accidents, many other-measurable parameters exist that are much more relevant and timely, and are more easily obtained from more familiar control room instrumentation.

1 6

f 4.2 Survev of Measurable Parameters l

l The spectrum of measurable parameters includes Coro I

Exit Temperature, Containment Radiation bevels, Ex-Coro Detector Readings, Hydrogen Concentrations, and finally, other parameters that are directly available in the control room that can play an integral, real i

time part in determining the course of the accident.

Each is discussed in turn in the following paragraphs.

i 4.2.1 Cofo Exit Temperature In general, the single most useful paramotor in i

the early recognition of a sovere accident is tho coro exit temperature.

Excessive core oXit temperature (above, say, 700'F) is considered to be the carliost indication of the onset of core damage.

By contrast, Hydrogen monitoring requires some amount of coro damage to have occurred beforo detection is possible.

In addition, the measurement of coro exit temperaturo is a parameter that is very familiar to plant oporators.

Therefore, by choice alone, it is likely that the operator may look to this-parameter beforo examining other possibilition.

Figure 3 shows a survey of core exit temperature profiles versus time (up to ninety minutos) for the same accident sequences described in Table 1, and includes both DBA and beyond DBA events.

It is noteworthy that an indication of elevated (and increasing) core exit steam temperature will occur very early in the accident for all sequences, and before any significant fuel damage occurs.

The t

data shown was arbitrarily truncated at 2200'F' The useful range of the Core Exit Thormocouples at-ANO extends up to about 2300'F.

Henco, the technical range of the CETs is quite adequate for this purpose.- Note that in the ANO Control Room, CET indication is provided on tho' Safety parameter l

Display System and covers the range from O' to 2300*F.

l t

7

~,_. _. - _ _ _. _,. _.. _

4 4.2.2 Containment Radiation Levels Another very offective means for confirming (or flagging the strong potential for) the onnot of core damage in a severo accident is containment radiation data.

Once the core has been damaged and fission products are released to the containment atmosphers, control room monitors reading high levels of radiation are a very real-time means of determining that some degree of fuel-damage has occurred.

Figure 4 shows the generic results for radiation levels versus time and the degree of coro damage.

The figuro, plus the supporting information for the figure, provide the following data at about 1.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> into a severe accident:

APPROXIMATE CORE TEMP CONTAINMENT DEGREE OF REtDAMAGE RADIATION CORE DAMAGE STATE (F')

RAD / Hour NO FUEL DAMAGE Up to 750 10' INITIAL CLAD FAILURE

(

}

Up to 2 x 10'

(

)

2 x 10l to INTERMEDIATE CLAD (1300 - 2000) 2 x 10 FAILURE

(

)

(

)

MAJOR CLAD FAILURE

(

)

2 x 10l to

(

)

5 x 10 INITIAL FUEL OVERl! EAT

(

)

2 x 10f to

(

)

5 x 10 (2000 - 2450) 5 x 10f to INTERMEDIATE FUEL

(

)

OVERHEAT

(

)

3 x 10 MAJOR FUEL OVERHEAT 2450 - 3450 Over 3 x 10' MELTING Over 3650 Not Correlated t

8

t f

a While this information is broad based in nature and obviously relios on a number of specific assumptions regarding fission product disporsion i

and plate out, etc., it is clear that there is at least.a crudo corrolation between core outlet temperatures, fuel damago, and containment radiation lovels.

Moreover, for the coro exit temperatures shown in i

Figure 3 at one hour (at or above 2200'F for all five ovents) the corresponding fadiation lovels in containment will be at least 1p RADagHour.

The Arkansas monitors road from to 10 RADS / Hour, with a normal reading of 10,10 RADS / Hour.

As can be seen, the increasing radiation readings versus core heatup are an excellent qualitative and

~

unambiguous indication of core damage.

Unlike Hydrogon monitors, the radiation readings are continuously available, do not require operator actions to activato, and have no delay timo.

Hence, on a purely qualitative basis,

.significant radiation readings are equivalent to Hydrogen data.

Indeed, the timos at which the core exit temperatures reach the lower threshold of coro damage (= 1p00*F) which in turn corresponds to = 10 RADS / Hour, and the times at which the hydrogen. concentration in containment reaches 0.1% aro quito similar.

Stated differently, cigd rupture (at 1300 - 2000 10 RADS / Hour in containment

' ' F) produces

• 10 just as rapidly as detectable levels of Hydrogen in containment are produced.

Honco the lack of'a specific Hydrogen monitor reading early in a sovero accident is not essential to a clear understanding that coro damage has occurred.

This is not surprising since the root cause of both the Hydrogen and the release of the fission products is excessive fuel temperature.

Since.

there is only one source of fission products versus more than a single pource of Hydrogon, it is' clear that the use of containment radiation readings easily componsates for the short term 9

%Je.,

-,= sus,---J4

1. A+K.

4----h-u-44.m--

e lack of Hydrogen data.

The redundancy of plant instrumentation relative to the diagnosis of a severe accident is thus more than 11exible enough i

to accommodate the short term absence of a single component from the mjy of instrumentation and indications available.

4.2.3 Ex-Core ps_tfqqt_or_Ecmlinan Ex-Core detector readings will show levels as high as ten to one hundred times normal as the core uncovers.

At TMI, about thirty times the normal readings for post-shutdown were measured.

This ratio is easily explained in terms of the lack of neutron attenuation by the RV water as the level drops.

The TMI operators initially interpreted s

the large readings as a reactor startup.

It is now well known that these data can serve very well as a coarse, but instantaneous RV level indicator.

^

4.2.4 Additiona1 MQAngrab1e Parangj;.gra During the early stages of a DBA, or a severe accident that progresses beyond DBA " space", there are many other measurable parameters available in the control room that can heir to diagnose the plant condition in a timely f shion.

At Arkansas, much of this information is collected in the control room at the Safety Parameter Display System (SPDS), which is readily available to the operators.

Included in the SPDS is the Reactor Vessel Level Monitoring System (RVLMS) which measures fluid IcVel in the active fuel region during -a LOCA event; this information is supplemented by Ex-Coro detector readings that increase as the core uncovers as outlined above.

In addition, the operators monitor pressurizer fluid level, PCS pressure and temperature, containment pressure and temperature, and secondary water levels in the steam generators.

For a large-IDcA class of accident, the operators can casily read low pressurizer 1cvel, elevated containment prensure (approximately 50 psig), loss of RCS subcooling, and high containment 10 v-

-w

-,y..

m-,

-o m

-.m,.,-p-3

temperature.

In addition, the operators will read no unusual conditions in the steam generators.

For a total loss of feudwater scenario, the operators will see an indication of high stean i

generator pressure and, eventually, low water level.

This indicates a loss of secondary heat sink.

In addition, once secondary side heat removal is lost, the operators will read RCS repressurization as well as primary Safety Valve actuation. During the early phases of LOCA events and total loss of feedwater events, these measurements can be taken quickly and efficiently, and the operators can assess the plant condition and take whatever appropriate actions are deemed necessary.

4.2.5 jlyMrPSen_M2Ditndl19 There is no disagreement that the ability to measure Hydrogen in containment can be a useful means of assessing degrees of core damage during certain severt, accident scenarios.

This would be a TSC functicn.

From an operator's perspective, Hydrogen concentration information can be used to actuate recombiners.

This information can also be used to strategically select actions to enhance containment integrity when appropriate.

There are however, certain noteworthy difficulties associated with the use of Hydrogen indication, and with its potential usefulness to the operator during fast moving accidents where core uncovery occurs very rapidly.

1.

Hydrogen production due to Zirconium oxidation processes will only occur after high core temperatures have resulted in fluid boiloff and subsequent core uncovery.

By the time Hydrogen concentration data.'ere available, it would have been possible to use the other means discussed above to determine that core uncovery has occurred (e.g.,

core exit temperature trends, Ex-Core detector readings and containment radiation levels).

11

2.

Coro damage assessment based on. Hydrogen concentrations in containment is normally an activity conducted later in an event by the Technical Support Center (TSC).

Since t.~e TSC may not be manned untfl sixty'to s 'y minutes into the a devere accident,

'k function may not be par ticularly usefu2 2n fast moving accidento such as those discussed in this paper.

3.

During an accident sequence Hydrogen may be produced due to processes other than Zirconium oxidation.

This can complicate its use for purposes of quantitative core damage assessment.

Rac2olysis effects with water, as well as oxidatirn of other metals in containment, can both add to the Hydrogen term.

For very rapid events the early Hydrogen production is dominated by the Zirconium-water reaction.

Nevertheless,

" backing-out" tho' contributions from radiolysis and corrosion is time consuming and subject to approximations and assumptions.

4.

The method used for collecting a Hydrogen sample involves an inherent lag time.

Even if a sample.is taken at time zero in an accident sequence, it will take some time to obtain and assess the sample for use as a viable, useful data point.

During a fast moving accident sequence, other means may be more readily available,-more rapid, and therefore more useful to the operator.

S.

There is no specific operator action associated with a Hydrogen reading over the time frames of interest, except for starting the recombiners.

However, since the recombiners are sized only for DBA events, their abser.co for a short amount of time will not affect the event significantly.

For example, if 2% Hydrogen concentration is produced-over 0.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> (See Figure 1), then this corresponds to about 130,000 SCF per hour.

By contrast, at 2% concentration, a sir.gle recombiner will only remove about 120 SCF per hour under these conditions.

Hence 12

___._.m.

the ratio of production to removal is about 1,000 to 1, so that recombiners have essentially zero effect for these events and time scales.

As previously mentioned, Figures 1 and 2 include Hydrogen concentrations in containment and as measured in the control room (respectively) for three selected accidents that are well beyond the design bas!-

Specifically, these cases include two large LOCAs with no injection (Cases 4 and 5) and a total loss of all feedwater event with no-injection (Case 1).

The figures clearly show that only the worst of the two LOCA cases will progress fast enough to produce measurable amounts of Hydrogen on control room instrumentation by thirty minutes into the event (Figure 2).

Table 2 provides a summary _of the scenarios and the estimates of the concentration and timing of Hydrogen in containment and as indicated in-the control room.

ior the higher probability (but still highly unlikely) LOCA wi':h three SITS and no Safety Injection, indication in the Control Room is not initially received until one hour into the event.

13

e 5.0 Procedural Considprat.ons 5.1 Current EPGs/EQEg The ANO Emergency Operating Procedures (EOPs) are based on;the Combustion Engineering Emergency Procedure Guidelines (EPGs), CEN-152, Revision 03.

These EPGs have been approved on an interim basis by the NRC and a Safety Evaluation Report is pending.

Previous editions of CEN-152 (prior tu Rev 03) directed operators to monitor for Hydrogen in a number of

' instances.

This action had been-included in the Standard _ Post Trip Actions (;PTAs) that are used immediately after each reactor. trip (See Note 2).

Revision 03 of the EPGs however, does nQt direct this action so long as there is no evidence from other plant parameters to indicate a need to do so.

The reason for this change is that operators are heavily tasked followino a reactor trip-to obtain a comprehensive-and accur ce picture of plant safety and to support efforts to

1p diagnose the cause of the trip.

Many operators ad reported that the efforts of the Control Room staff were better focused on using Note 2:

The CEN-152 structure is entered into and is based on events that either have.an automatic reactor-trip or that are manually tripped as needed to-insure plant safety.

By-contrast, NUREG-0737 references the actuation of Hydrogen monitoring to the initiation of a Safety Injection Actuation Signal (SIAS).

It is felt that this SIAS reference is based either on a strictly LOCA orientation or possibly, on the Westinghouse orientation to entry into the EPGs.

In a C-E plant, it is not realistically possible-to have a SIAS signal-prior to reactor trip.

In this context, it is suggested that' procedural steps to monitor Hydrogen not be based on SIAS.

An obvious but specific example of this would be a Steam Generator Tube Rupture (SGTR) event that has SIAS, but does not require Hydrogen monitoring.

14

normally available instruments to perform the SPTAs, and that Hydrogen monitoring was best performed for those cases where other plant indications dictated the need for that information.

Simply stated, the current EPGs do not use Hydrogen monitoring as part of the initial event diagnostic process, as outlined abcVe.

An additional reason to support this fact is that even very severe accidents do not produce measurable quantities of Hydrogen during the time af ter trip when using SPTAs and when initial diagnostics are being performed (typically zero to five, or possibly ten minutes, depending).

Hence, in the EPGs, Hydrogen monitoring in done only for LOCAs, ESDEs, and for functional recovery unless there are indications present to show a need for monitoring.

This philosophy is quite consistent with the other reasoning presented regarding Hydrogen monitoring, which is to use the capability when it is needed as opposed to using it by a certain time.

The SER for Rev 03 has been pending for over three years; to date, there have been no comments from the NRC regarding this issue.

Also note that CEN-152 does not address the need to have the capability (at any particular time) except implicitly in that it cannot be used if it is.not available.

The ANO attempts to relax the-thirty minute requirement have been_quite correct within a-DBA context since Hydrogen accumulates significantly only over days of time _and since the only procedural guidance currently in existence is the DBA

-based EPGs/EOPs.

The implications of Hydrogen monitoring requirements for events beyond DBAs are briefly discussed in the next section.

5.2 Severe Accident Manaaement l-l Procedures (or guidance) for events beyond DBAs do not L

exist at this time for PWRs in the United States, While much research has been conducted since TMI in i

separate phenomena. the U.S.

industry is just starting to address severe accident management issues via a NUMARC initiative.

Currently, there are no NRC 6

requirements for utilities to have such guidance, and the purpose of the NUMARC initiative is to proactively work with the NRC to develop guidance that will be mutually acceptable to the NRC and implementable by the utilities.

An NRC generic letter on this topic is anticipated during 1992.

The NUMARC-initiative has currently proceeded to the point where the individual owners groups are starting their work to produce draft generic guidelines by early 1993.

The CE Owners Group schedule is commensurate with this and will start in early January, 1992.

ANO 1:s a participant in this task (CEOG Task No. 726).

.None of the NUMARC work to date has been at the level l

of detail required to address when a Hydrogen monitoring capability is functionally needed.

ABB C-E feels that this paper represents the leading edge of this issue based on extensive severe accident initiatives with NUMARC, EPRI, all other owners groups and many individual utilities.

It is also felt that the arguments presented herein adequately support the-fact that there is no overwhelming functional need to monitor Hydrogen in the early phase of any event,-given the'cther indications that are available, the time scales involved for Hydrogen generation, the non-availability of TSC guidance over these times, and the complexity of interpreting Hydrogen concentration in terms of core damage assessment.

It has been shown that significant containment threatening l quantities of Hydrogen are not present this early in any known events, that very rapid events produce Hydrogen many times faster than it can be removed by recombiners, and that there are many easier-to-use indications of a severe accident than Hydrogen early in an event.

In short,.the knowledge of Hydrogen concentration early in a severe accident is not expected to play a major role in the guidance to be developed since'this information is at best only corroborative and at worst may be superfluous or not pragmatically useful.

16

4 6.0 Symmary It:is clear that Hydrogen concentration in containment can be an important parameter in assessing post accident conditions.

The unavailability of this parameter, with its inherent. time delay, is not crucial to the activities that must be performed and the decisions that must be made in the first sixty to ninety minutes into a severe accident.

Other parameters that read out in real time, or in near real time, are of much greater utility to the operator in the time frame of interest.

Key among these activities is the operator's overriding concern to restore adequate injection to the RCS.

The existence of. measurable quantities of Hydrogen in containment,-even-if measured, will not change this priority.

The one step currently dependent on Hydrogen concentration is the starting of the recombiners.

For the beyond DBA events under consideration, recombiner operation will not significantly affect the Hydrogen concentration-in containment since they are sized for DBA events, not severe accidents.

The conclusion from these arguments is that Hydrogen monitoring early in an event does not play a crucial role in mitigating the event.

Later in the event, when this information can be used effectively, the data is important to the decisions concerning mitigating actions chosen, particularly in regard to possibly inerting the containment with steam to prevent explosive Hydrogen concentrations from forming.. During the first sixty to ninety minutes after an event however, other--parameters provide the real time information upon which the operators can assess the event and'make the appropriate decisions for its mitigation.

17

TABLE 1 DESCRIPTION OF ACCIDENT SEQUENCES ANALYZED CASE 1:

Total loss of all feedwater with no injection available CASE 2:

Five square foot cold leg break with both trains of injection available and SITS available (3 out of 4).

Recirculation mode not available.

CASE 3:

Five square foot cold leg break with one train of injection available and SITS available (3 out of 4).

Recirculation mode not available.

CASE 4:

Five square foot cold leg break with no injection capability, but with SITS available (3 out of 4).

CASE 5:

Five square foot cold leg break with no injection capability and no SITS available.

GENERAL NOTE:

These are not necessarily limiting cases as analyzed.

They are intended however, to show approximate magnitudes, time scales and trends.

The use of more limiting cases would not affect the arguments and conclusions of this paper, in particular, note that while none of the cases analyzed showed flammable concentrations in containment within 1.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> into the event, other, more limiting cases, could possibly show this.

Hence, no arguments were made herein based on flammability conditions not being reached.

The cases: analyzed do show that only very extreme and limiting cases would show flammability this early into an event.

December, 1991 18

1 TABLE 2 SURVEY OF-SCENARIOS'TO ESTIMATE i

HYDROGEN CONCENTRATION IN CONTAINMENT

-CASE-1

CASE 2 CASE 3 CASE 4 CASE 5 l

TLOF 3 SITS /

3 SITS /

3 SITS /

NO SITS /

NO SI 2 SI

_1_Sl__,

_NQ_SL NO SI

H. Mass! by 1. hr 0-0 0

295 250

(1b mass)-

Hi Mass by 1.5 hr.

26 195 140 310 '

300 (1b mass)

H,f Volume in ' -

-0.23 l'.71 1.23' 2.72 2.63 t

Containment by-1.5 hr (v/o).

• LEarliest Indication 2.0 1.9 1.9.

1.0 0.5 of_ H, in_ Control-

-Room _(hrs)~

+

Includes Zirconium oxidation, radiolytic effects, and other metal oxidation-4 Assumes 0.1: v/o instrument threshold and 30 minute instrument delay time.

December, 1991

[

19

Figure 1 Li2 % IX COXTAINMENT VS "V B H2 CONCENTRATION vol %

3%

&?

/

a j

/

a or n s' O

,/

~

l

/

5 l

t 1"n

-I i

0

/

/

/

/

0" 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 TIME (HRS)

CASE 1 CASE 2 CASE 3

+ CASE 4 CASE 5 December 1991

\\

Figure 2

-i2 % AS XJECATEJ X CONTRO1 ROOE H2 % INDICATED IN CONTROL RM 3%

l

/

- -a

~

__+

---

r--

/

gg y

/

/

/

/

/

N

/

=

l 1%

/

/

/

/

/

/

/

0%

~

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 TIME (HRS)

CASE 1 l

CASE 2

+ C ASE 3 0

CASE 4 CASE 5 December,1991 l

I

i Figure 3 COR3 3XI" " EM3 B RA"U R B VS

"~;M B CORE EXIT TEMP deg F 2300 9

'n l

/a

)

,/"'/

1800 ll

,/

n' y

/

/

/ /,

1300

./

/V P'

/

M

/

/l!,

~c l{

{-

/

/

6 i /',

i

/

i

'f' 5

300 0

0.1 0.2 0.L 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 TIME (HRS)

CASE 1

+ CASE 2

+ CASE 3

+ CASE 4

~

CASE 5 December,1991

Fl2URE 4.

TYPICAL ANALYSIS FOR POST ACCIDENT DOSE MATE INSIDE A CYLINOMICAL CONTAINMENT

-Taken from: " Development of the N1 Comprehensive Procedure Guideline for Core Damage Assessment,"

CEN-NPSD-241, Prepared for the C-E Owners Group, July, 1983.

1:108

+4, eg l

/,

t 4

1. < 0,

%y N

9

'*g,

'%o

'e 1x1@

+c

'+'e g

O 69 1

1x103 1

1O 100 100(

TIME POST ACCIDENT, HOURS CYLINDRICAL CONTAINMENT December,1991 23

j l

i APPENDIX A A Study of Core Wide Cladding oxidation and Rydrogen Release During Design Basis LOCAs A cursory review of ABB recorded calculations for licensing the Arkansas NSSS was made to survey the calculations of core wide cladding oxidation during postulated design basis IhCAs.

The

. purpose of this review is to form. a qualitative. picture of the timing and extent of Hydrogen release to the containment resulting from design basis core uncovery calculations.

As required by 10CFR50.46, the calculated core wide cladding oxidation must be less than 1% for a broad spectrum of postulated-_ brea~ks in the primary-piping.

The survey which covered both large break _and small break LOCAs showed that the Cycle 1 results have remained bounding for all future cycles, namely, Cycles 2 through 9.

Tho core wide cladding oxidation result is translated to Hydrogen released to-the containment using the following conversion:

For the SONGS 3410 Mwt FSSS, oxidapion of 100% of the core's Zircaloy produces 5.07E+5 std ft of Hydrogen _(

Reference:

CEN PS D-2 41) '.

Using total core power as a_ basis for

-similitude, for the Arkansas 2825 Mwt NSSS,3 oxidation of 100%-

of the core produces roughly 4.2E+5-std ft of Hydrogen.

Using this conversion, the 1% licensing criteripn for core wide cladding oxidation translates into 4.2E+3 std ft of Hydrogen for Arkansas.

The containment Hydrogen monitors have scales which read from 0% to 10% of ~ the containment volume.

Using rough numbers, for-a 3

containment of 2.0E+6 ft and assuming the Hydrogen monitor reads 0.1%, phe_ monitors will first detect Hydrogen in excess of 2,000 std'ft.

Therefore, at the limit of design basis licensing space for 1% core wide oxidation the Hydrogen monitors should just show a_ reading low on the scale of-roughly 0.2% of containment _ volume.

A-1

- ~ _. - --

- ~ ~. - _ -. _

Larae Break LOCA For the most limiting large break LOCA, the core wide cladding oxidation for Cycle I was calculated to be 0.617%.

This translates 3

.to 2.6E+3 std ft of flydrogen.

At this level of !!ydrogen production in design basis licensing

space, the containment monitors may just show a reading very low on tho' scale of 0.1% of containment volume.

(Core wide cladding oxidation is based on the-results of the COMZIRC code?s analysis of the 1.0 double ended guillotine break in the pump discharge leg for the limiting time-in-life for the core ' of 1000 MWD /MTU. )

The limiting large break LOCA hot rod peak cladding temperature was calculated to be 2060*F (See Note 1) at 243 seconds into the transient.

The peak local cladding oxidation for the hot rod is 9.79% of the cladding thickness.

Further analysis of the hot rod thermal _ response. shows that the highest rate of local oxidation occurs between 150 and 250 seconds and that after core reflood the local oxidation process _ reduces to a very low level before 500 seconds into the transient.

Cladding rupture of the hottest fuel rod is predicted; -therefore, the oxidation calculations _ include both sides of the ruptured cladding.

This local oxidation in the ruptured region of the fuel rod-is used in the core wide calculation with the added conservatism that all rods in the core are assumed to rupture.

Based on the hot rod thermal responso described above for the postulated large. break LOCA, if a reduction in ECCS performance were assumed beyond design basis space, either the maximum PCT limit of 2200*F or the local cladding oxidation limit of 17% would be exceeded well before the core wide cladding limit of-1%.

This means that cladding embrittlement and " shattering" of the affected fuel rods upon reflood of the core with cold ECCS inventory would-occur well before - liydrogen release from excessive core wide oxidation became a problem.

l.

i Note 1:

A temperature of 2060*F is the current value (from the Cycle 9 Analysis); a more limiting value of 2078'F (from l

the Cycle 1 Analysis) is referenced in the FSAR.

l A-2 y

-~->-p

I l

Small Break LOCA A similar survey for the small break spectrum of breaks shows that the long term core uncovery process in not as limiting as the large break blowdown and reflood process.

Even though the time of core uncovery for the limiting small break LOCA was over ten minutes, the peak local cladding oxidation of the hot rod was only 0.2046%

of the cladding thickness compared to 9.79% for the large break LOCA.

The limiting small break PCT is only 1460*F, which is more tl e large break limiting LOCA.

This PCT than 500*F less than 3

occurred for the 0.1 ft cold leg break at 760 seconds into the transient.

The core wide cladding oxidation calculation was not performed for the small break spectrum or for the limiting small break size since these results are clearly bounded by the large break calculations.

Il o w e v e r, using the peak local oxidation for the hot rod as if the entire core consisted of hot rods, the core wide oxidation is estimated to be less than 0.04% for this limiting small break LOCA.

Using the conversion, method described above, this translates into less than 200 std ft of flydrogen.

This level of liydrogen release is probably not detectable by the containment monitors.

Based on the hot rod thermal response describe above for the postulated small break LOCA, it a reduction in ECCS performance were assumed beyond design basis licensing

space, the local cladding oxidation process would reach a "run-away" condition during the core uncovery period before the ten minute point in the transient.

Excessive embrittlement of the cladding in the core would therefore occur before !!ydrogen release from core wide oxidation becomes detectable.

A-3

_