ML20196E131

Backfit Analysis in Support of Proposed Generic Ltr 'Loss of Dhr.'

ML20196E131
Person / Time
Issue date:	09/20/1988
From:	NRC
To:
References
NUDOCS 8812090276
Download: ML20196E131 (22)

Text

I L

• Q .,'

s.; .

• PACKFIT ANALYSIS IN SUPPORT OF THE PROPOSED GENERIC LETTER "LOSS OF DECAY HEAT REMOVAL" September 20, 1988

SUMMARY

A backfit analysis has been performed of the actions recomended in the pro-posed generic letter (GL). Analysis results, expressed in dollars per reduc-tion in person rem averted for each action, are as follows:

Description Cost / Benefit Ratio, $/ person rem averted 5% discount rate 107 discount rate Expeditious actions High population 170 110 Low population 950 580 Programed enhancerents .

High population 48 48 Lcw population 260 260 All GL recomendations High population 220 150 Low population 1200 800 Containment closure assuming all other actions are in place Figh population 2300 1400 Low population 10000 7000 where the high population entries refer to plants with population densities within a 50 mile radius that are represented by Zion (890 persons /sq mi), and the low population entries are for a population density of 100 persons /sq mi.

All GL recommendations are essentially justified on the basis of a cost / benefit criterion of $1000/ person-rem, although one may argue that containment closure alone is not justified if all other actions are in place. However, we caution that the uncertainty of these values is high, particularly in the cases of the expeditious actions and containment closure entries, which have a strong dependency on containment closure assumptions, and the results should be applied accordingly. We believe that assurance of containment closure is a necessary defense-in-depth response to nonpower operation cent. erns.

INTRODUCTION The Code of Federal Regulations,10 CFR 50.109 (a)(3) states Except as provid-ed in paragraph (a)(4) of this section, the Commission shall require the backfitting described of a facility)only in paragraph (c when of thisitsection, determines, that based there isona the analysis in-substantial crease in the overall protection of the public health and safety or the ccmon defense and security to be derived from the backfit and that the direct and indirect costs of implementation for .that facility are justified in view of this increased protection."

'812090276 8 8009:0 f0$

PDR MISC hl

{ GG12090276 FDC .

4

• s; . ' '

2 Paraoraph (c) states "In reaching the determination required by paragraph (a)(3) of this section, the Commission will consider how the backfit should be scheduled in light of other ongoing regulatory activities at the facility and, in addition, will consider information available concerning any of the follow-ing factors as may be appropriate and other information relevant and material to the proposed backfit:". This statement is followed by nine items which are to be addressed.

The backfit analysis described in paragraph (c) is provided below. ,

BACKflT ANALYSIS Each of the nine 10 CFR Sn.109 (c) items is addressed below:

(1) "Statement of the specific objectives that the proposed backfit is de-signed to achieve;"

The objective is to significantly reduce the expected frequency of a severe release of radioactive material from within the reactor coolant systen (RCS) to the environment due to loss of decay heat removal (DHR) sys tems.

The expeditious actions accomplish this objective in the short term by reasonably ensuring containment closure with some decrease in the likeli-hood of a core damage accident (CDA). The programmed enhancements focus on accident prevention with additional consideration given to mitigation should an accident initiate.

(2) "General description of the ectivity that would be required by the licensee or applicant in order to complete the backfit."

The GL recomends two sets of actions: expeditious actions and progranted enhancements.

The expeditious actions are primarily directed toward mitigating offsitt ,

release should severe core damage occur following a loss of DHR %e..c. '

Practical expeditious actions to reduce the likelihood of a los ' ""o event and to provide improved response to a loss of DHR event .u t occur, are also reconnended.

Progranned enhancement reconnendations emphasize improved unds. .....g, better procedures, better utilization of existing equipment and better instrumentation. These actions deal with the root cause of loss of DHR events and provide for improved operator response to a loss of DHR event if one occurs. The expeditious action reconnendations relating to con-tainment closure will be relaxed, but reasonable assurance of containment closure capability should still be provided to mitigate a severe core damage event should one occur.

4

I O  :

g,. 3  ;

1 The expeditious action recommendations are:

Recommendation 1:

Discuss the Diablo Canyon event, related events, lessons learned, and implications with appropriate plant personnel ant, provide training--

Licensees should compile material concerning these topics and discuss this information with operations, maintenance, and supervisory personnel.

Recomendation 2:

Provide procedures and administrative controls sufficient to reason-ably assure containment closure before core uncovery due to a loss of DHR event--This recomendation is for licensees to provide contain-ment closure procedures that are to be activated if a loss of DHR event initiates. Specific actions will depend upon licensee outage planning ahd plant design.

Recommendation 3:

Provid6 two temperatures representative of the core exit temperature when operating with water level reduced below the top of the hot leg nozzles (midloop operation)--The actions necessary to implement this recommendation are to keep two core exit thermocouple assemblies operating during the outace until RCS level is raised for head removal in preparation for refueling. Some operations may be affect-ed by implementation of this recorrendation.

Recommendation 4:

Provide comunication of water level information to the operators when the RCS water level is lower than three feet below the reactor vessel (PV) flange (reduced inventory condition)--This recorrenaation has no effect on licensees who display two accurate RCS level indica-tions in the control room. Other licensees should provide personriel to read level at whatever location it is available with direct comunication between that person and the control room (CR), or should provide level indication in the CR.

. Recomendation 5:

Minimize perturbations to the RCS and to systems which are necessary to maintain the RCS in a stable and controlled condition while the RCS is at a reduced inventory condition--This recomendation involves additional management attention and operator awareness of outage actions. Administrative controls should be reviewed and upgraded to ensure that maintenance and surveillance activities do not degrade RCS condition and possibly lead to a loss of DHR. Outage planning should entail more emphasis to assure that activities that may affect the RCS are not scheduled during periods where RCS inventory is reduced.

~

i

=

6 l

,9 ,

s. ,

'. 4 l 1

Recomendation 6:

Assure means of adding it,ventory to the RCS in the event of a loss of DHR event and provide procedures for their use--Licensees should identify existing equipment to mitigate a loss of DHR event, develop procedures for using this equipment and provide operator training on ,

the procedures. At a minimum, licensees should provide one high pressure injection pump and one other system, in addition to the normal DHR system pumps. They should provide procedures for purp use that include consideration of the ability to inject water so that it enters the reactor vessel. Some impact on the timing and planning for equiptrent maintenance may occur.

Recommendations 7 A 8:

Prohibit blocking of RCS hot legs unless a vent path is availat'le to prevent pressurization of the RV upper plenum--This action may entail modifications to administrative controls and outage :)lanning. t The programmed enhancements are:

Recomendatinn 1:

i Provide improved instrumentation.--The proposed generic letter (GL)

identifies (1) independent ir.dications of RCS level; (2) instrumenta-tion to measure DHR s measurements; and (4)ystem visible and performance; (3) core exit audible indicathns temperature for these instruments in the CR. This irstrumentation is needed to assure monitoring and control of the DHR function by the operator and to follow the response of the RCS following a loss of DHR event in order to assure proper operator response and verification of operator actions.

Recommendation 2:

Develop irproved procedures for normal use of the DHR systen.

Develop emergency procedures for responding to a loss of DHR

! event.--The expeditious action recommendations include development of l

procedures based on current understanding, instrumentation, and equipment for responding to a loss of DHR event. Normal DHR proce+

. dures are identified because these should cover (1) operation with an ,

RCS level in the pressurizer, (2) operation when RCS level is re- '

t duced..and (3) level transitions and operations between these re-gimes, although the recomendations specifically address only items (2) and (3). Programed enhancement recomendation 4, discussed below, provides for analy'ses to better understand RCS behavior during normal operation and following a loss of DHR event. Procedures

( should be developed, and operator training performed, using the l enhanced understanding developed from the analyses and should incor-porate the improved instrumentation identified by programed enhance-ment recomendation 1.

l

.w-- ,-,- - , ,w,v-,,, -,,--,-,-..----rn---------,n,--.---v. .---.- --- ,-,.- ,e, -

c ..-

. 5 l

Recomrnendation 3:

Provide adequate and reliable equipment for maintaining the RCS in a stable and controlled condition while in nonpower operation. Assure that additional equipment is available for responding to a loss of DHR event--No additional equipment is expected to be needed by licensees to comply with this recommendation. Modifications to existing equipment and/or administrative controls may be needed. We expect licensees to examine their DHR systens to identify support systems and to provide administrative controls to minimize mainte-nance of these support systems when operatino with a reduced RCS inventory. Additionally, we are askino licensees to consider evalua-tion of the the autoclosure interlock [ACI) for the DHR system, and we encourace its removal. Experience shows that spurious closure of these valves has caused approximately 60% of the loss of DHR events.

Since the ACI aids in preventing LOCAs outside containment (Event V),

this should be evaluated on a plant specific basis.

Recorrendation 4:

Perform analyses to understand nuclear steam supply system (?!SSS) behavior under nonpower operation--Items such as consideration of possible RCS configurations and states (e.g., nozzle dams installed, manways open, cold leg openings, decay heat level), that may exist during nonpower operation should be considered. These analyses will provide increased understanding of flSSS response to loss of CHR events, understanding of the time available and effectiveness of possible operator responses to these events, and time available for contairrent closure in the event of a severe core damage accident.

In addition, licensees should perform analyses to understand instru-ment response during nonpower operation and in response to loss of DHR events. These analyses will be used to develop the improved procedures discussed in programed enhancerent Recommendation 2.

R2correndation 5:

Identify necessary technical specification (TS) changes--The analyses and subsequent actions performed by the licensees are expected to lead to identification of TS changes which could improve the reli-ability of DHR, as well as providing enhanced operational flexibili-ty. Licensees should propose TS changes where appropriate.

3. "Potential change in the risk to the public from the accidental off-site i release of radioactive material." i

a. Probability of a Core Damage Acr.ident (CDA).

Applicable risk information is limited. Two probabilistic risk analyses (PRAs) (refs. I and 2) and a regulatory analysis (ref. 3) have been performed. The PRAs are not fully independent and the regulatory analysis is for conditions that are similar, but not identical to the proposed OL recomnendations. The PRA information is uncertain for the following reasons (in addition to those which are applicable to full power operation):

6 G .

' 6 l'

(1) Nonpower operation involves operator and other personnel inter-actions in event mitigation to a greater extent than power operation. The uncertainty associated with human interactions is high.

(2) The reported work, although recent, does not include significant information learned since the Diablo Canyon event of April 10, 1987. Of particular impact is that all of the work was based on the assumption that the shortest time to core uncovery was about four hours after loss of D4R, that the shortest time to midloop operation was three days, and that this was believed to be conservative. We now k.. " that midloop operation can be achieved in a shorter time and that RCS configurations are possible in which the core can be essentially emptied of water in approximately 15 or 20 minutes. Conversely, the reference 2 work in particular (and hence the reference 3 conclusions) gave little credit for operator flexibility in dealing with plant conditions that were outside of their experience or written guidance, anc fission product retention is not carefully ad-dressed.

(3) Only one PPA has been reported (ref. 1). Reference 2 is based upon the methodolooy of reference 1, but uses generic informa-tion. It is not a completely independent PRA, although it does contain elements of independent review and use of data.

Refereneg1reportsthattheriskofaCDAduringnonpoweroperation is 2x10' per reactor year. (We have rounded numbers tn one signifi-cant figure for purposes of thi Reference 2 reports thenonpowerCDAvalueas4x10"gdiscussion.)

per reactor year, with 85% of the CDA due to midloop operation,

b. Release Liu.ihood.

The GL 87-12 (ref. 4) review, plant visits, discussions with plart personnel, owners group meetings, and review of some plant procedures have led us to conclude that without the recorrended changes it is likely that containment will be open if a CDA accident occurs early in a refueling outage. For estination purposes, the release likeli-

. hood can be taken as essentially identical to the CDA likelihood.

c. Other applicable information.

The reference 5 letter from the Westinghouse Owners Group (WOG) supperts staff analyses that, without corrective action, there is a significant risk of core damage and release of fission products. The WOG sponsored analyses represent tne'f st relatively thorough investigation of this mode of operation.

Evaluation of plant data shows that a large number of events have occurred and continue to occur. If not mitigated, such events could lead to core damage. Many of these events have invo' Jd a loss of OPR for one or more hsurs. A number hu e resulted in boiling in the

i g _'.

. 7 core, an unanalyzed condition et many plants. Often, plant pirsonnel were unaware of the real difficulty for some time during or a'ter the event. Experience clearly substantiates that a problem exists.

Information obtained since the Diablo Canyon event of April 1), 1987 shows that many previously unrecognized mechanisms exist that exacerbate the problem, and that are not represented in the P M results. Some of these can realistically cause core uncovery ar complete core voiding in less than half an hourt significantly 'ess than the previously believed "conservative" boil down of water w'th uncovery of the top of the core in four hours.

Review of industry responses to GL 87-12 showed that most licensees were poorly prepared for reduced RCS inventory operation. Procedures were incomplete, incorrect, or nonexistent. Instrumentation was often of low quality or inaccurate, and little provision was made for using equiprent effectively. The responses established that the  :

problem was extensive. ,

d. Poter.tial risk reduction.

WeconcludethatthelikelihoodofaCDAgithcontainmentiner unclosed condition is of the order of 10" per reactor year, and that the likelihood of containment being closed durino the CDA is small, particularly if the accident initiates from a midloop operating condition. This is consistent with the conclusions presented by the Office of Nuclear Regulatory Research (RES) in their regulatory analysis (ref 3).

RES reported the CDA frequeyy as 4x10-5 due to loss of DHR and found it could be reduced to 4x10 by actions which they defined.

RES additionally provided ri:k information for a high population site *

(Zion), with a 50 mile radius population density of about 890 persons per square mile, and for a low population density site with a density of 100 persons per square mile. They further stated that the popula-tion representative of a "generic site" in the year 2000 would be about 340 persons per square mile. We will use the high and low end values to develop insight into the range behavior, essentially the '

same process as presented by RES.

RES showed that the risk for a high population site is about 1000 person rem per reactor year (PR/RY) and for a low population site they reported 200 PR/RY. A closed containment was reported to change these to about 10 and 2 PR/RY for the high and low population sites, respectively. The preventative and mitigative actions RES investi-gated, when applied without containment closure, reduced CDA likeli-hood by a factor of ten, and reduced risk to 100 PR/RY for the high population sites and to 20 PR/RY for low population sites. We note that adding the RES CDA reduction to closed containment protection has essentially no impact upon calculated risk since containment has already reduced risk to a small value. Similarly, adding containnent protection to CDA reduction accomplishes a factor of ten reduction -  ;

• ' 8 0*.

bJt in quantitative terms this will be relatively small since the CDA raduction has already removed most of the calculated risk.

Tne containment closure actions we address in the proposed GL will achieve the same risk reduction as RES reported in their regulatory analysis. The expeditious actions and programed enhancements we recomend in the proposed GL encompass the RES recomendations and consequently will achieve a risk reduction at least as great as that reported by RES in their regulatory analysis. Consequently, the RES risk reduction information is applicable to the proposed actions.

We conclude the risk reduction is roughly 900 to 1000 pR/RY for a high population site, depending upon the actions that are taken.

Applying the inforrstion te all sites assuming theg are high popula- I tion sites results in a total averted dose of 2x10 PR. The same processagsuringlowpopulationsitesleadstoatotalaverteddose of 0.010 pR.

l

'4. "Potential impact on radiological exposure of facility employees;" i Four arcas of potential impact have been identified - instrumentation installation, increased containment background radiation, mitigative action exposure which may accrue with existing procedures, and exposure due to a major release.

Only limited installation involving connections to radioactive components l 1s involved. These actions have already been taken by a number of licensees. Diablo Canyon Ursonnel incurred 6.87 rem exposure in install-ing level instrumentation. Once installed, perso W no longer need to remain near radioactive components to rea' ' " ~ ang term improvement.

Permanent installations ray also elim4 .icantly reduce dose associated with connection of temporary .n.. antation. We do not foresee 6 significant detrimental impact.

Obviously, a severe core damage accident with an unclosed containment l would cause exposure to workers in the plant and plant vicinity.

1 Finally, most of the recomended effort involves planning, analyses, procedures preparation, and training. These activities do not involve radiation exposure.

5. "Installation and continuing costs associated with the backfit, including the cost of facility downtime or the cost of construction delay;"

i No construction or plant startup delay is anticipated since the expedi-tious actions should be implemented prior to lowered inventory operation and the prograr:vned enhancements are related to refueling outages or an 18 month time span. However, the as soon as is practical philosophy con-tained in the recomenaations is a clear indication tht.t we recomend the irprovements be incorporated prior to startup where this can be reasonably accomplished without impacting the schedule. Plants presently in the planning stage should be designed to eliminate or significantly reduce the problems.

4 a C -

. . . 9 Licensee costs vary depending upon the status of each item at each facili-ty . This is particularly true for containment closure, where the calcu-lated monetary impact is high due to loss of electrical power production at the plant dur(ng the time the plant is shut down, and due to the assumption that replacement power must be purchased and charged to 'he shutdown. Cost variability is discussed further below as part of the discussion of each recor. tended action. Cost data are summarized at the end of this discussion.

The following are estimates of the impact upon licensees of the expedi-tious actions:

a. Discuss Diablo Canyon implications and conduct training Host licensees have discussed the i..1plications of the Diablo Canyon event with.their operations personnel. This recommendation involves an update and broadens coverage to include maintenance and other personnel who may be involveo with outace operations and/or manipula-tions. The cost of $75K provided in reference 3 appears appropriate for the proposed GL recommendation at each licensee's plant.

b. Provide reasonable assurance of containrent closure The primary impact of the expeditious actions is the impact on downtime for a refueling outage due to the assumed electricity replacer.ent cost. This is also an area where a wide range of esti-rates has resulted, and where a few licensees have implemented some of the recomnended actions. The trend appears to be for the actual impact to be smaller than the estimates, but there are exceptiens.

We asked representatives of each vendor owners group for their estimates of the impact of a containment closure criterion during meetinas to discuss shutdown cooling. The responses, provided without preparation or study, were estimates of less than one day to a week, with the variability reflecting licensee opinione for their plants and with about the sare range iar the three vendi/ owners groups. Based on this information, RES used 3 days per refue'irg outage as roughly representative, and increased it to 3 days /M to account for unplanned shutdowns in their cost analyses (ref. 3).

They also expected the value would decrease during future refueling outages as licensees planned for containment closure, thus "...sub-stantially reducing the industry-wide impact and, thereby, lowering the assessed cost / benefit ratio...."

The containment equipment hatch is not opened at some plants during refueling, at some this is done after fuel has been removed from the RV, and at others the hatch is removed as soon as is possible follow-ing termination of power operation. Some licensees have procedures in place to close the hatch within a few hours. Others did not at the time of their response to GL 87-12. Estimates of the impact of an equipment hatch replacement procedure on a refueling outage range fromzero(forseveralplants)toonetotwoweeks(atoneplant).

C

• 10

.: .' 1 The equipment hatch is the most visible of the pnetrations which nust be adoressed, but there are many potential openings to consider.

Personnel hatches can be closed within a few minutes if no equipment is in the way and planning should eliminate any equipnent problen.

Opening other penetrations may induce problems unless the sequence is carefully planned.

, Several plants (Diablo Canyon, Tro.ian, San Onofre) have implemented containn nt closure procedures during nonpower operation. Others (Seabrook, Farley) either don't open the equipment hatch or defuel the reactor vessel prior to opening the hatch. Still others (Arkan-sas) use an existing spare penetration for temporary cable runs ,

during shutdown operation, thus alieviating the need to run such cables throuch the equipment hatch opening, and have considered ways to close these quickly if necessary. The refueling outage at Salem that was initiated in early September 1988 uses the expeditious action recommendetions. During mid-August, a licensee representative at Salem estimater the impact of the expeditious actions as about a one day increase in plant down time.

Much of the abcve information had not been obtained at the tine of the RES work. The trend is toward less of an inpact than the prelim-inary estinates. There are additional aspects of containment closure that affect impact estimates. These include:

(1) The expeditious action recommendation for containment closure procedures is provided for an RCS level lower than three feet below the RV flangt for W and CE plants. It applies to B&W

plants for a level lower than four inches below the top of the i bot leg flow area at the connection of the hot leg with the RV.

i The level is niaintained above the four inch elevation during i

routine refueling and for many maintenance activities at B'.W

! plants. Many operations, such as draining in preparation for RV l head removal and actual head removal, can be accomplished at levels higher than three feet below the RV flange in W and CE plants. These considerations do not apply to the RES approach.

(2) We believe unplanned shutdowns are quite different from planned shutdowns insofar as containment perturbation is concerned. We consider it unlikely that a number of containnent penetrations

. will be opened in such a manner as to be difficult to close ll during most unplanned shutdowns.

' (3) The containment closure criterion is linked to the understanding of plant behavior and the time available in which containment must be closed following a loss of DHR. We specify times that are relatively short in the proposed GL. These times may be extended following suitable analyses. The WOG has funded an extensive an61ysis program which we understand provides much of this tining informatien. We have been told by other vender owners group representatives that many of their licensees have perforned calculations which aho establish tining. This is a significant reduction of the containment closure impact upon i refueling outage time.

i

.. - _ . . . 9,,_. .. - _ _ _ _ _ _ _ - _ .

e -

, . 11  :

i For purposes of cost analyses, we will use one day per refueling outage, one refueling outage each 18 months, and $400K/ day as the replacement power cost.

Procedures should be prepared to address containment closure. These should include actua.1 containment closure activities and a more extensive administrative procedure to assure that containment closure is achievable. Containment closure costs are included in programed enhancement iten b that covers procedures.

c. Provide two RCS core exit temperature indications Many plants are little affected by this recommendation. The B&W NSSSs use bottom entering thermoccuples (T/Cs) and minor replanning can allow several of these 60 remain operational during lowered inventory operation.

Some of the U-tube steam generator (SG) equipped plants are similarly equipped, with the electrical connections brought out of the RCS pressure boundary at the seal table, and at least one plant is being converted to this scheme. Such conr.ections may remain in place for a l few T/Cs or, alternatively, jumpers may be provided to allw other '

seal table work to proceed.

Plants where the RV head entry T/Cs are retained should make provi-sions for keeping a minimum of two of the T/Cs operational and should provide proceduras for th:- process.

The cost of this item is discussed in the instrumentation discussion under the procraaned enhancement effort.  ;

d. Provide two independent RCS level determinaticos Many plants already have instrumentation in place or are planning its installation. There is no impact for such plants. Others will use an operator to continuously monitor level via existing in-containment instruments. Some do this already. The cost will be essentially the labor cost plus a radiation exposure allowance. Still others will initiate the programed enhancement recomendations. Such costs are

. covered in the programed enhancement impact discussion.

e. Minimize perturbations to the RCS This reconnendation should entail planning to reorder maintenance and test actions. The cost is included in the cost of procedures prepa-  :

ration. We note that some licensees already have such procedures in '

i place (suchasDiabloCanyon),orareimplementingthem.

f. Assure equipment is available for addition of water to the RCS

ho new equipment is anticipated. This recommendation will involve

administrative procedures which reasonably assure equipment avail-r I

9

• ii.i .*

• 12 ability. No significant inpact on outage time is anticipated. Cost is included in the procedures preparation cost,

g. Do not inappropriately use nozzle dams This is a followup action since licensees should have already ad-dressed the potential problem in response to the WOG connunication (ref..r). It does not apply to the B&W plants since they (apparently) do not use nozzle dams and, for the most part, the RV vent valves eliminate the problem of pressure differential that could lead to emptying the RV. It may also be of concern only for refuel-ing outages since nozzlo dams may not be used for many unplanned outages.

We additionally note that some plants other than B&W do not use nozzle dams. 1rojan, for example, used to use them but has concluded they are not a real benefit, and no longer routinely uses nozzle dams i during refueling outages.

Most of the potential impact can be eliminated by conducting opera-tions in order to avoid an RCS configuration where there is a poten-tial problem. Where this is the case, there will be minimal cost and .

schedule impact.

! The upper bound on impact appears to be provision of a suitable vent path to eliminate a pressurization problem. This may take the form of removal of a pressurizer manway or removal of the pressurizer safety valves (as in Salen for the September 1988 refueling outage, where they were goino to do so anyway so that the valves can be tested off site). Ve estimate up to an eight hour operation with two or three personnel based upon discussion with licersee personnel.

Cost is probably small,

h. Do not inappropriately use loop stop valves Only a few plants are equipped with loop stop valves. The process of I complying with this reconnendation is similar to that of the nozzle dams, although of somewhat lesser concern since stop valves will withstand whatever pressures could be generated <turing an accident and hence accidental depressurization due to overpressure is of no 1 . concern. (Of course, an accidental o j

be addressed by suitable procedures.)pening is of concern, and must Cost is small.

The following are estimates of the impact of the programed enhancements:

l a. Provide instrumentation and alarms as described in the proposed GL i

i Several information sources were used for instrumentation cost,

) including RES (ref. 3), plant specific information, preliminary 1 estimates from the Instrumentation and Control Systems Branch (SICB),

) and judoerent based upon plant visits, discussions with licensing j personnel, and consideration of the instr.11ation to be performed.

} -

1

13 SICB estimated a complete DHR monitoring systen based upor finw rate.-

inlet and outlet pressure, noise monitoring of the pump, and pump motor current to be $100K, of which $50K was for noise monitoring.

Water level instrumentation cost was estimated by RES as $355K, in good agreement with the Diablo Canyon installed level instrumentation at a total cost of $387K, including allowance for personnel exporste and probably a small cost for procedures. We will use $375K as ti.e estimated cost.

l RES did not estimate the T/C cost directly. SICB estimated that installation of temperature monitoring capability via a seal table

! penetratien of the RCS pressure boundary would cost $5000 for materi-als plus about three days of work. To this must be added the plan-ning and quality control functions, which will be expensive since this is an RCS pressure boundary. For scoping purposes we will use j $50K, and not address the operations inpact of leaving T/Cs on the reactor vessel head since the problem can be eliminated by way of the seal table approach.

We judge the followirg licensee costs are appropriate for the instru-i mentation recorcended ii, the generic letter:

l RCS level 5375K RCS temperature 50K DHR monitoring 100K i

Total 525K

! b. Provide procedures for normal and emergency operation i

i RES (ref. 3) estimated about $900/page for simple revisions to l $3600/page for complex revisions. We estimate of the order of 100 pages at mary plants, with the complexity ranging from simple revi-sions to preparation of new, complex text. However, we anticipate the owners groups will prepare guidelines which will reduce individu-al licensee costs. We will assume the cost of guidelines as

$5000/page and that there are 100 pages from each of three owners groups for a total of 300 pages. Guidelines cost is therefore $1500K which is split among roughly 50 plants (to account for identical

- guidelines in some of the 78 plants) for a cost of $30K per plant.

Each licensee will then have to use the guidelines to achieve plant specific procedures. We will apply a $2000/page estimate to 100 pages, for a cost of $200K per licensee or 200 X 50 / 78 = $130K per plant.

Procedures costs per plant are therefore about $160K.

c. Provide reliable equipment for RCS cooling and inventory control RES estimates autoclosure ;.-+erlock removal to cost $100K to $150K per plant (ref. 3). We will un the $150K estimate.

..,' 14 The remainder of this item involves operating and administrative procedures. This is covered in the procedures section b, above,

d. Analyze plant behavior during shutdown operation The bulk of this effort for the Westinghouse plants has already been funded by the WOG and a report is expected momentarily. The CE owners group has indicated that much of the analysis effort has been performed at individual licensee facilities, although none of the information has been specifically reported to us. Analysis of the B&W plants will be less of an effort since some of the concerns are not applicable to the B&W plants.

We estimate an effort of six months duration with four people is sufficient to pe* form most of the analyses on a generic basis by each of the vendor owners groups for W and CE, and perhaps 2/3 this effort is reeded by B&W. Assuming a labor rate of $53/hr (a high value from those identified ir, reference 3), the W and CE vendor analyses will cost $220K, and the B&W effort will be $150K for a total of $590K.

(Note we are not allowing for the work that has already been accom-plished, which is significant.)

Each licensee will have to perform plant specific analyses to adapt the ceneric work to their plants. Diablo Canyon spent $42K for analyses associated with their midloop investigations. Some of these i

were more extensive than would be necessary for adaptation of generic analyses, whereas other areas may need more extensive coverage. We

estimate that $40K per plant will be sufficient for this purpose. If 1 we divide the $590K among the 78 plants and add $40K per plant, we i obtain an analysis cost of reughly $50K per plant,

e. provide Technical Specification (TS) changes The RES provided estimate of time for this work in the Generic Issue 99 resolution is 16 person weeks at $50/hr (ref. 3), or $32K. We anticipate a more comprehensive and complex change in TSs than

, perceived by RES. Consequently, we estimate the complexity to be 1

roughly three times the RES estimate, or about $100K per plant,

f. Refine avoidance of RCS perturbations l This is a more comprehensive addressing of the topic covered under i expeditious action item e. The effort has been addrassed in proce-j dures, item b, above.

i I

4 k

i

s -

6 15 Licensee costs on a per plant basis are surinarized as follows:

Discuss Diablo implications $75K Containment closure 400X/ refueling Instrumentation 525K Procedures 160K Equipment reliability 150K Analyses 50K Technical Specifications 100K Total licensee cost $1060K plus $400K/ refueling

6. "The potential safety impact of changes in plant or operational complexi-ty, including the relationship to proposed and existing regulatory re-

. quirements;"

The impact of the proposed changes in plant or operational complexity has been discussed previously. In surrnary, we believe the recomendations may have some initial impact on plant operations which in the longer term will be minimal. Programmed enhancement recomendation implementation should provide more stable RCS operation during reduced inventory conditions and a decreased frequency of loss of DHR events should result.

Several instances where other requirements may be affected and sete need for reassessment of those reauirements have been identified. These are:

a. Technical Specifications One of the proposed GL recorrendations is that licensees identify needed TS changes, and either propose new ones or show how eouivalent requirements will be incorporated into their operations. This infernation should be factored into the staff effort on new standard TSs.

~

Ve note that sore TSs may become less restrictive than those which exist.

b. Emergency procedures guidelines and emergency procedures Staff safety evaluation reports addressing emergency procedures

- guidelines contained recorrnendations that coverage be extended to nonpower operation. The proposed GL recomends implementation of part of these recomendations. Although not explicitly part of the recomendations of the proposed GL, this is an opportune time for licensees to modify their normal and emergency procedures to cover all aspects of nonpower operation, thus providing a good basis for transition to ener ncy procedures and providing complete coverage of emergency conditions,

c. Generic Letter 87-12 The proposed GL includes the staff response to inforration which industry was required to provide by GL 87-12. Further supplemental information to tnese responses would no longer be required.

s -

o .-

,.,- 16

d. Generic Issue 99 "Ir: proved Reliability of RHP Capability in pWRs" The proposed GL should provide the guidar.ce necessary to close GenericIssue(GI)99.

7. "The estimated resource burden on the NRC associated with the proposed action and the availability of such resources;"

The cost to the NRC to date consists of preparing the preposed GL, prepa-ration of the CRGR package, meetings and telephone conversations with vendor owners groups representatives, plant trips (much of which was included as observations conducted in the course of being in the plant for another purpose), and study of plant procedures. Most of the cost was incurred for other purposes, including extensive work with ersergency procedures guidelines, investigation of the Diablo Canyon and related loss

• of DHR functions, and work associated with GL 67-12. Assurance that the irrediate actions are beino properly implemented will be covered by staff review of the responses and by resident audits at each of the plants.

This effort includes review of licensee responses to the proposed GL to provide assurance that recomendations have been complied with, assurance that the licensees are pursuing an appropriate comitment to the programed enhancement recommendations, and development of a temporary instruction to guide residert investigation of whether licensees have actually implemented the proposed G!.

, recommendations. The resident inspectors are expected to verify completion of the licensee's actiers.

Review ano verification of the prograrted enhancement recomendations are expected te te performed with a combination of hRR staff, contractor, and regional resources. It is assumed that the analyses and procedural guidelines will be performed via owners' group efforts.

The proposed GL reque'sts lictrsees to inform us of their plans with respect to the recomendations. We plan to review these responses.

Roughly 50 replies will be reviewed (for the total of 78 plants), at a review rate of one per day. W( will also ask the resident inspectors to review the actions taken at each plant. This will require about two deys at each plant. The total is 206 days. Using a rate of $50/ hour, the NRC cost will be about $80K. NRR (SRXB) will prepare review guidance and temporary instructions to support the residents. This SRXB effort is estimated at I? weeks. At a cost of $50/hr, the cost is $24K. Staff effort associated with expeditious actions is roughly $105K.

Staff review of the instrumentation is not anticipated. We will. review licensee descriptions of plans, conduct plant audits, and ask the plant resident inspectors to check that the described instrumentation is actual-ly installed and that procedures are implemented for its use. Cost of these activities will be addressed below.

We anticipate that the bulk of the staff procedures review will be ars'oci-ated with the guidelines. Actual procedures will be audited and plhit .

resident inspectors will be asked to perform audits at their respective plants. Guidelines review will require 50 weeks of staff time. Each

s -

) ,.,h 17 resident inspector will spend about two days auditing procedures en a per plant basis, and complete plant audits at six licensee plants will require a team of two people about two weeks to perform at each plant. The total NRC time is 105 weeks. The previously identified temporary instructions will also provide the guidance needed by the inspectors for this activity, and were costed as part of the expeditious actior,s review. At a cost of

$2K per week, the cost of guidelines and procedures review is $210K. We also anticipate contractor assistance. With an estimated $75K for con-tractor assistance, the total NRC ccst for this item is $285K.

We anticipate generic review of the background for individual licensee changes involving the autoclosure interlock, followed by a brief review of specific plant implementation plans. The WOG has funded a generic docu-ment which supports this change, and we have already approved the change at Diablo Canyon, which we recommend in the proposed GL as a model to follow. We believe a ceneric review coupled with plant specific changes can be accomplished in no more than 24 weeks of staff effort. The cost estinate for the total is $50K.

Staff review of the analyses is anticipated, as is independent contractor calculation of plant behavior. We estinate two and a half months of staff effort ard $75K for independent contractor analyses, for a total NRC effort of abcut $100K.

The RES provided reference 3 estinate for NRC review of TS chenges is

$25K. Using a factor of two to account for additional work we envision due to more complicated changes, we estinate the NRC cost to be abeut $50K for TS changes.

The staff review effort is summarized as follows:

Expeditious actions $105K Procedures 285 Equipment reliability 50 Analyses 100 Technical Spe-ifications 50 Total staff effort $590K Sufficient resources are available to accomplish these actions.

6. "The potential impact of differenes s in facility type, design or age on relevancy and practicelity of the pioposed backfit;"

We have met with representatives of ea:n of the FWR owners groups, all of whom also had vendor rep.'*sentatives present, and we discussed the impact due to the actions recome ,ded in the rroposed s GL. Personnel were present at these meetings who w9re farailiar with some of the oldest plants (Yankee Rowe), as well as with plantf. not yet in operation. No conditions of impracticality were ider.tified wi1h respect to items included in the proposed letter.

We recognize there are nuNrous differences among plants, and have consid-ered such differences in formulating the proposed GL. We have also

d ,

?,

• 18 i.

provided a concentrated effort on minimizing plant impact consistent with obtaining our objective of reduced risk. This is reflected in relaxation of existing requirements that do not contribute to safety, in the lack of specificity regarding equipment to allow a flexible licensee response, and in the philosophy of linking actions to the state and condition of the plant.

9. "Whether the proposed backfit is interim or final and, if interim, the justification for imposing the proposed backfit on an interim basis."

Some of the expeditious actions are interim and should remain in effect entil modified via the programed enhancement program. Others should be permanent. For example, containment closure planning is recomended to accomplish an irrediate reduction in risk. Closure timing and the plant conditions leading to containment closure are snticipated to be modified as additional information is obtained, although containment closure procedures are expected to remain operational. Conversely, RCS level readings may be made from within containment with no direct reading in the control room in the expeditious actions. This is not anticipated in the longer term recorrendations.

The programed enhancement dctions are recomended as permanent changes, .

These are anticipated as implemented over several years. The overall risk reduction will be greater than obtained from the expeditious action recomendations, and a significant reduction in CDA likelihood will be obtained due to increased emphasis on prevention.

COST /EENEFIT The Office to applicable of Generic Nuclearissue Regulatory 99 (ref 3).Research There arepreviously prepared many similarities a staff analysis between that work and the staff GL recomendations. The reference 3 work will there-fore be applied to the GL recommendation:, and the sare assumptions will be used except for the changes specifically discussed in this backfit analysis report.

The ref. 3 estinated frequercl5 /RYf c re dara nonpower operation was 4.3x10 (ref. 3)ge due

, (Weto lossmore carry of DHR during significant figures than war anted by the uncertainty in this section of the report so that

. trend informat'on may be conveyed.) The stated risk due to midloop operation was 85% of one total. The risk of a release due to core damage was not direct-ly addressed, but was covered by a sensitivity study using probabilities of the containment being open during a CDA of 1, 0.01, and O. Information obtained during generation of the proposed GL recomendations has led the staff to conclude that it likely the containment will be open if a CDA occurs during midloop operation. For cost / benefit illustration purposes, we will assume the conditional probability of containment being open is 0.9, given a CDA. We will further present the results for midloop operation. The release likelihood which represents the starting point of the cost / benefit illustration is there-fore:

(4.3x100)(0.9)(0.ES) = 3.3x10-5fgy 1

O' t, f.,

19 i.

We will further assume containment closure is achieved with a likelihood of 0.99 (consistent with the ref. 3 comparison basis of a 0.01 probability of containmentbeingopen). With these assumptions the containment closure recomendations in the GL will change the probability of containment being open from 0.9 to 0.01, a factor of 90 reduction.

Reference 3 provided a risk due to shutdown cooling conditional on containment l bein open of 1014 PR/RY for plants characterized by a high 50 mile radius popu ation density)(Zion, (100 persons /sq mi . (For 890 personsa/sq comparison, median mi), and 187forPR/RY density for2000 the year a low density was ,

projected tu bo 340 persons /sq mi.) The assumption that containment will be closed 10% of tee time will reduce these values further to 913 PR/RY and 168 PR/YR, respectiv 'ely. Taking 85% of these values to limit the comparison to midloop operatior providec a risk of 776 PR/RY and 143 PR/YR for the hich and low density population plants, respectively. The assumed containment closure will reduce thast values by a factor of 90 to 8.6 PR/RY and 1.6 FR/YR, respec-tively. This is representative of the effect of the expeditious action recom-mercations in ths GL. Although there is a cost associate with recommendations other than conti.inment closure, it is negligiole in compa(ison r with the con-tainment closure impact, and we have not fully separated the other costs. '

Further, although there is a benefit from other expeditious actions than 1 containment closure, the effect of containment closure dominates and the other e'fects may be neglected.

Reference 3 reported that implementation of a number of instrumentation, procedures, and hardware changes would reduce the shutdown cooling risk to 79 PR/RY and 15 PR/YR, respectively. The GL recommendations encompass those recommendatiens, and the same values may be applied. Introducing the 85t value to account for midloop operation, and the 0.9 value to account for containment being closed some o' the time, and assuring all GL recommendations are imple-mented except for containment closure will result in a risk of 60 PR/RY and 11 PR/YR, respectively. This is representative of programed enhancement action implementation. l Similarly, the ref. 3 values applicable to implementation c' all vecomrenda-tions are 0.5 PR/RY and 0.1 PP/RY for the high and low population sites, respectively.

We estimated licensee cost of implementation of the GL recomendations as

$1060K plus $400K per refueling outage per plant. For illustration purposes, .

. we will assume one refueling outage each 18 months. Reference 3 provided the l following formula for calculation of the present value of an annual cost:

(1 + r)t -1 PV = present value = g C where: r = real discount rate t t = renaining lifetime ,

C = annual cost i 8

Using this, the $400K per refueling outage results in $3.24x10 8and $1.97x10 for 51 and 10% discount rates, respectively, for a total of 78 nuclear plants.  !

e 20 O*.

We may now coepile costs from:

1. The present value formula for Si and 10% discount rates for the cases of expeditious actions,

2. The $1.06M per plant times 78 plants for implementation of the programed enhancements, assuming certain of the costs associated with some expedi-tious actions are combined with the programed enhancement costs, and

3. The sum of items 1 and 2 for all of the actions recomended in the GL.

This results in the f'ollowing taole:

Description i;RC cost. it$ Licensee cost, $M 1

Si discount rate 10% discount rate l Expeditious actions 0.1 324 197 Programed enhancements 0.5 83 83 All GL recommendations 0.6 407 280 l

Cor,tainment closure assuming all other - 324 197 actions are in place Following the ref. 3 approach, we contrast this inferration to the worth of the

averted dose. The following dose information applies

Description Dose, PP./RY Averted Dose Averted Dose j per plant PR/RY per plant Million PR (total of7P. plants) t present conditions

! High population 776 - -

! Low population 143 - -

Expeditious actions High population 8.6 767 1.06 Low population 1.6 141 0.34 Progra med enhancements

. High population 60 716 1,73 Low population 11 132 0.32 4 All actions i High population 0.5 775 1.87 l

Low population 0.1 143 0.35 1

where the averted dose is equal to the averted dose per plant per year tines 31

! years times 78 nuclear plants.

I

[e

?. l'- 21

o. ,

Dividing cost by the averted dose provides the cost / benefit ratio, expressed in dollars per reduction in person rem averted for each action. These are as -

follows:

t Description Cost / Benefit Ratio, $/ person ren averted v 5% discount rate 10% discount rate Expeditious actions High population 170 110 Low population 950 580 Progranned enhancenents High population 48 48 Low population 260 260 All GL recommendations

. High population 220 150 Low population 1200 800 Containment closure assuming all other actic6s are in place High population 2300 1400 Low population 10000 7000 where the high population entries refer to plants with population censities within a 50 mile radius that are represented by Zion (890 persons /sq mi), and i the low population entries are for a population density of 100 persons /sq mi.

All GL reconmendations are essentially ,iustified on the basis of a cost / benefit criterion of $1000/ person-rem with the exception of containment closure with the assumption that all other actions are in place. Although one ray argue that containment closure alone is not justified if all other actions are in place, we caution that the uncertainty of these values is high, particularly in '

the' cases of the expeditious actions and the containment closure entries, which have a strong dependency on containrent closure assumptions, and the results should be applied accordingly. We believe that assurance of containment  :

closure is a necessary defense-in-depth response to nonpower operation. con-  !

cerns. -

l f

l

}

22 REFERENCES

1. "Zien Nuclear Plant Residual Heat Removal PRA," NSAC-84. July 1985.

2. T. L. Chu, et al, "Improved Reliability of Residual Heat Removal Capabil-ity in PWRs as Related to Resolution of Generic Issue 99," NUREG/CR-5015 BNL-NUREG-52121, Brookhaven National Laboratory, May 1988.

3. Spano, A. H., "Regulatory Analysis of Generic Issue 99, ' Loss of RHR Capability in PWRS '" draf t, June 1988.

4. "Loss of Residual Heat Removal (RHR) While the Reactor Miraglia, F. J.[RCS)

Coolant Syster is Partially Filled (Generic Letter 87-12)," Letter to all licensees of operating PVRs and holders of construction permits for PWRs from the Associate Director for Projects, NRR, July 9,1987

5. R. A. Newton, "Westin Operation Concerns,"letter ghouse fromOwners Group Chairman early notification of Vestinghouse Owners of fiid-loop Group to Westinghouse Owners Group Primary Representative (IL, IA), OG-88-21, May 27, 1988.

l 1 -

1 l

e

.