Text

Forwards Plant Jco for CREVS & Control Complex Habitability Envelope,Per 971208 Telcon.Rev 3 to Jco,Attached.Jco Prepared W/Guidance Provided in GL 91-18,Rev 1
	ML20203F803
Person / Time
Site:	Crystal River
Issue date:	12/15/1997
From:	Grazio R; FLORIDA POWER CORP.
To:	; NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
	3F1297-39, GL-91-18, NUDOCS 9712170467
	Download: ML20203F803 (39)
	v • d • e

m __ _ . _ .

Florid 2 4 .CORPORATION Power -

Y u eu.

w e.e ?a.o.a.'m. m -n 1

.p 1 b December 15,1997 -

3F1297-39 U.S. Nuclear Regulatory Conunission Attn: Docurnent Control Desk Washington, DC 20555-0001

Subject:

Control Complex liabitability Envelope Request for Additional Information

Reference:

FPC to NRC letter,3F1197-09, dated November 10,1997, " Control Room IIabitability, NUREG-0737 Item lil.D.3.4"

Dear Sir:

The purpose of this letter is to respond to a verbal NRC Request for Additional Information relative to the Crystal River Unit 3 (CR 3) Connol Complex liabitability Envelope (CCliE). Per telephone conversation of December 8,1997, the NRC requested a copy of the CR-3 Justification for Continued Operation (JCO) for the Control Room Emergency Ventilation System (CREVS) and the CCllE. A copy of Revision 3 of this JCO, dated December 6,1997, is attached.

l In the referenced letter, Florida Power Corporation (FPC) stated that the results of the CCilE inleakage testing and the revised calculational methodology would lA used to demonstrate operability of the CCHE and CREYS 7.rior to restart from the current outage. Consistent with the referenced letter, operability .

.of the CCllE and CREVS has been demonstrated and is documentet in the attached JCO. This JCO '

l was prepared consistent with the guidance provided in Generic Letter 91-18, Revision 1 "Information to Licensees Regarding NRC Inspection Manual Section on Reso'ution of Degraded and Nonconforming l Conditions." l The JCO states that a detailed analysis of the Steam Generator Tube Rupture (SGTC) event was performed and that the Maximum liypothetical Accident (Mil A) remains the bounding event with regard to control room habitability FPC is preparing a revised Control Room Habitability Evaluation Report to support License Amendment Request #222, " Control Room Emergency Ventilation and Emergency

Filters," and expects this report to confirm the conclusions reached in the attached JCO.

No new comniitments are made in this letter. If you have any questions concerning this response, please contact Mr. David Kunsemiller, Manager, Nuclear Licensing at (352) 563-4566.

Sincerely db3 g Robert E. Grazio, Director Nuclear Regulatory Affairs ,

REG:kdw 9712170467 971215 I

~

Attachment DR ADOCK O p

xc: : Regional Administrator, Region II ggg g gg H IHo s m n NRR Project Manager umIHlHlH

- - - ' HIH HIlH **

H Senior Resident Inspector CRYSTAL RIVER ENERGY COMPLEX: ' 15760 W. Power Ura Street a Crysta! River, Flonda 34428-6708 + (352) 795-6486 A Fkunde . Progress Company

4 ATTACHMENT TO 3F1297-39 JUSTIFICATION FOR CONTINUED OPERATION FOR TIIE CONTROL ROOM EMERGENCY VENTILATION SYSTEM AND TIIE CONTROL COMPLEX IIABLTABILITY ENVELOPE

-- - - _ = . _ _ . . . . -

s .

o 4 s DEFICIENCY REPORT INSTRUCTIONS Precursor Number: PC 97-4355 Work Request: N/A safety class: N/A Code Class: N/A-Repair, other than original design L,X_]

Use-As-Is L 3 Interim Use-As-Is r1 Expiration /re-evaluation date Renork (oniv-.apolicable for ASME Co4LQass 1. 2 or 3 cgfnponents) I .t, l- Fngineering M* ** Nx N/A r 1 Justification:

See attached disposition 3C0 to PC sbM- Shew b.41%k , Mc : M ct.d/tv 9;wCw Kurd:$%ewL., % ew4 Y bEw / '

11 W bg' tOrigtnator A40F G/ A~ R.L;t m L o, il/2 tin V Oate 9 % J e 1/ b SV , Vertf tcation Engineer Q

,L c O! /h. rasow Date

/i/24.k7 f lq 444/b

-Engtnee g bisptYvtsor

'% N }kt/ M 8 h M E's "lM/97 Date

' M/ h

l Quality Systems (as reautred) Date

-.ulp- .

l Mn_(atr ,, Date Date PRC @d1r n }hft.1:QO1.Cgg)

PRC Meet 1 'er ?~lN)

,w dL, t!.l--- -

At aP*O I2 Y-9 9 MhQ et 3EQ (0LCeQuttedl StLe sie en Page.s Q 19,2.1,.22 we-Re rwvise 4s A tese of PR C % tonpuTS o'J H/2C/M. gj g ,

p Pages A h fo A, wag gevi5g 7 A ;e s C

b tcNsm3 mme cls CP-111 Rev. 61 Il[f99@Page1/

- . . . _ ~ -

4 DEFICIhWCY REPORT Re: PC 97-4355 CHANGE RECORD CHANGE: N/A REASON: Original DATE: 11/26/97 Issue ORIGINATOR: VERIFIER: SUPERVISOR:

Robert ~J. Lane Jack Wilkinson

Harry B. Oates CHANGES Page l D(scription N/A see original cover sheet (attached) for signatures.

i

4-Changet l_ REASON:. Comments DATE: 11/28/97 from 11/26/97 PRC .

Meeting ORIGINA1DR: VERIFIRR: SUPERVISOR: ,

M. Clary Steven D. McMahan Harry B. Oates d

CHANGES Page Description 4 added ... '{ corresponding to a filter dP of 4" wg) "

19 item 2) added ". . . In addition to satisfying current ITS testing requirements ..."

item 9) changed NGRC to PRC 4

21 & 22 added information regarding spray flow rates:

Calculation I86-0002 Rev. f, 1/16/96, determined containment spray removal constants using the n,ew instrument error corrected flow values of 1397

gpm (injection phase) and 1112 gpm (recirculation phase). Spray constants associated with the lower value of 1112 gpm are used in revised dose calculations.

The instrument loop uncertainties for spray flow indication and control were being reviewed concurrent with performing the revised dose calculations. As a contingency, the revised dose calculation looked at a containment spray flow

.yate of 1000 gpm and found that it was essentially the same as the 1112 gpm case. The

- calculation concludes that containment spray rate of 1000 gpm can be tolerated.

Balliet to Widoll ltr ser NOE97-2311 dtd s 11/11/97, shows that when spray is being supplied from the RB Sump, the actual flow may be 121 gpm below the indicated flow of 1200-gpm. Thus, the lowest value may be 1079gpm.

-2

4 CHANGE: 2 REASON: Incorporate DATE: 11/29/97 Licensing Comments ORIGINATOR: VERIFIER:. SUPERVISOR: ,

Steven D. McMahan M. Clary K. Anderson for Harry B. Oates CHANGES Page Description 3 added ..." 3) CR-3 Operating License The CR #3 operating License contains a requirement to maintain Control Room habitability as specified in the post-TMI requirements of NUREG-0737. However, there is no requirement for the measurement er evaluation of inleakage in accordance with specific requirements.

4) FSAR discussion ..."

6 Added "...See page 6A, that follows..."

6a Inserted page that read ...

If a radiological accident were to occur which involved the release of radioactive material from the reactor or spent fuel storage area, the CR-3 Radiological Emergency Response Plan would be implemented. The plan provides for staffing the emergency response organization and establishing emergency response actions commensurate with the severity of the event.

Actions required in the Emergency Plan Implementing Procedures include dispatching a Health Physics Technician to the Control Complex to monitor radiological conditions, and to provide radiological and meteorological data to the Dose Assessment Coordinator. The Heahh Physics Technician will perform radiological surveys within the Control Complex, including surveys for airborne radioactivity. Dose Assessment personnel use data collected from surveys to project expected personnel doses. Provisions exist in the Emergency Plan Implementing Procedures for considering administration of potassium iodide (KI) to personnal based on projected dose. A projected dose of 25 REM to an individual has been established as the threshold for considering administration of KI.

Control room dose calculations contain very conservative assumptions 3

e .. ,

regarding cperator presence in the control room. For example, in l accordance with the Murphy-Campe methodology it is assumed that the operctor is present in the control room continuously for the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months , then 60% of the time for the next 3 days, and then 40% of his time for the remaining 26 days. Similarly, atmospheric conditions are assumed which channel the release of radioactivity toward the control complex at conditions which maximize the plume concentration, particularly in the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months of the scenario. Based on the conservatisms that exists throughout control room habitability dose calculations, no doses approaching GDC 19 limit's are anticipated in a realistic accident.

However, provisions exist in currently approved procedures to monitor actual and projected doses based on tr.easured and observed conditions, and to control personnel exposure. Through the protective features of the control complex evaluated in the control room habitability calculations, and the established emergency response procedures, protection of the control room operato.8 is assured.

4

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

o CHANGE: 3 REASON: Incorporate DATE:

NGRC Comments M (

ORIGINATOR: VERIFIER s SUPERVISOR:

W. C.2a n > a gggwd 'i I W 2.b.M k % Ms M. Clary y g,]gg& r . . .

_/ o. M 4 Page Description 1 Changed " multitude" to " number" s

6a Moveo contents of page 6A into a new section in under Justification for Continued Operation.

Now located under Additional Protective 'eatures section 3. Dose Management.

Deleted page 6A 19 Added a discussion regard g vestibules under new heading Additional Protective Features section

1. Vestibules. New section reads as follows:

1. Vestibules CCHE boundary doors represent a significant source ofleakage into the CC. There are three double doors and three single doors. Vestibules were added to the three CCHE double doors in 1996, and have proven to be etTective in reducing differential pressure across the existing doors.

Reducing the differential pressure exerted on the boundaiy doors has two benefits. First, lower differential pressure rede.;es leakage'around the doors, and second, lower differential pressure allows the door closers to perform more reliably. During this outage, vestibules were added to the single CCHE boundary doors. All of the vestibules have additionally been sealed at interfaces with the CCHE boundary making the enclosures more effective in reducing CCHE boundary door leakage.

Control-Complex tracer gas testing was performed with the vestibule doors blocked open to assure that the test was conservative. Blocking open the vestibule doors increased conservatism in two ways. First, normal access and egress was permitted during the tracer gas leakage testing which contributed to the measured leakage. Dose calculations

-5

performed in accordance with standard methods include a factor of 10 cfm of continuous leakage to account for access and egress during an accident. This factor was added to the measured inleakage in performing CR-3 dose calculations. Therefore, the effect ofleakage during access and egress was applied twice. In a~ real event the vestibules would not be blocked open. Since they function similar to an airlock, they would be effective in reducing leakage during access and egress. The second manner in which maintaining the vestibules open during the test was conservative is that the existing boundary door was the only barrier to leakage during the test. In an actual event the vestibule doors would be in their normal closed positions, and would be effective in reducing infiltration into the CCHE through the doots. This feature would be particularly effective in reducing operator dose during the MHA without LOOP scenario where leakage is induced from the Turbine Building into the CCHE due to operation of the ABVS.

19 Added a discussion regarding Auxilary Building filtration under new heading Additional Protective Features section 2. Auxilary Building Filtration. New section reads as follows:

2. Auxiliary Buildine Filtration Performing dose calculations in accordance with the requirements of NUREG-0737 Item Ill.D.3.4, Control Room Habitability, includes assuming source terms specified in Standard Review Plan 15.6.5. One aspect of this for a plant that does not have an "ESF atmosphere filtration system," is leakage of 1500 gallons of water contained in engineered safeguards piping outside of containment must be assumed 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months following the accident. This amounts to leakage into the Auxiliary Building (AB) of highly contaminrted water from the reactor building sump An ESF atmosphere filtration system is like the CR-3 ABVS with HEPA and carbon filters, however it is required that an ESF system be powered from an onsite power source. Since the ABVS is not powered from an onsite power source, CR-3 control room dose calculations include the required leakage term. This term is responsible for approximately 8 REM of the projected individual control room operator dose of 26.; REM during the MHA with LOOP scenario.

The CR-3 ABVS has redundant fans and filters which are operate.d continuously during normal operation. Therefore, the system must be maintained in good operating condition. The filters are tested and maintained in accordance with approved procedures that implement regulatory guidance on emergency filter systems. As such, filter efficiency is routinely verified by carbon filter media testing. Esaluations of the FPC power grid performed for station blackout concerns have demonstrated a 6

l

~

.4 ~

high degree of reliability. Studies have shown that- power should' be

=- restored to the site in under eight hours following a loss of offsite power i c

due to disturbances'on the grid. Therefore; there is a high__ degree of 1~

sssurance that thel ABVS will be available and capable of performing effectively _ to reduce radioactivity released from the - AB_ following an ;

event.-

-f Calculations show a near linear relationship between AB source term and operator dose. Therefore, using a conservative ABVS filtration efficiency ,

of 75% _would result in a dose reduction of approximately 6 REM (8 REM contribution X 0.75 y 6 REM- reduction) from the 26.5 REM 4 projected dose for the MHA with LOOP scenario. This reduction in dose j i- is applicable only to the MHA with LOOP analysis. 75% efficiency for

- ABVS carbon filters _ has been previously accepted for CR-3 as an interim measure during resolution of reactor building flood level issues, and is already credited in the analysis for the MHA without LOOP .

9 Added ". (NOTE:

.. The isolation dampers were tested in-place and the test boundary exhibited insignificant leakage.)..."

13 Added "...in-addition.to that measured by_the tracer gas leak test..."

13 Reworded sentence to read "...Since other DBAs might not actuate-this signal,_a review of events has-been-performed which rely on automatic radiation detection or operator action for isolation..." ,

, 21 Summary / Conclusions item 9) required DBA control Room Operator dose analysis to be

reviewed and verified prior to'NGRC approval of JCO _This action was complete so'it was deleted from list.

21 Summary / Conclusions item 7) required i

administrative controls to fill loop seals-prior j' to-Mode 4. This was complete so it was d,eleted, from the list.

a 4

0 5

d

-_4 _

7

.sm,~ ,v,,J-..~,,..a ,,---w- , ,,m, .m , , --c,,. m '

l REvislON 3 l e

JUSTIFICATION FOR CONTINUED OPERATION FOR Tile CONTROL ROOM EMERGENCY VENTILATION SYSTEM AND TIIE CONTROL COMPLEX IIABITABILITY ENVELOPE (including Mode Change to Start-up and Power Operation)

I)ESCRIPTION AND PURPOSE .

System Readiness Reviews conducted in 1997 during the CR-3 Design Basis Outage identified several issues which potentially impacted control room habitability. A number of actions have been undertaken to address t% concerns and significantly improve the level of protection provided for the control rootn operator. These include (1) modincations to reduce CCilE inleakage by improving the integrity of boundary elements (Ref.1), (2) CREVS design changes to provide alternate means of mechanical equipment room ventilation and to improve system reliability, and (3) programmatic changes to ensure that the assigned ef0ciency of the Control Complex charcoal Glters is consistent with regulatory guidance (Ref. 2).

The modi 0 cations and design changes discussed above required that the Control Room operator dose calculations be revised to align inputs and assumptions with plant d: sign. The basic methodology used in these revised calculations is consistent with that found in regulatcry guidance and utilized in previous calculations. The determination of CCHE inleakage and the application of inleakage in dose calculation differs significantly from previous methodology. These differences have been determined to constitute a USQ as stated in an informational report to the NRC on the subject of Control Room IIabitability dated 11/10/97 (ref. Docket Ixtter 3Fil97-09). This report stated that a License Amendment Request would be forthcoming from FPC to address this issue. However, as the time required for review and approval of the LAR is not expected to support the unit restart schedule, a Justi0 cation for Continued Operation is being prepared to address the safety significance of this USQ and ascertain the acceptability of restart in the interim per the guidance of Generic Letter 91-18, Rev.

1. The specific issues addressed in this JCO are (1) the operability of the Control Room Emergency Ventilation System, and (2) the integrity of the Control Complex Habitability Envelope.

SAFETY CLASSIFICATION CREVS is credited with providing environmental control for personnel comfort and equipment operation, as well as with the protection of control room personnel during radiological and toxic gas events. It is considered to be a safety related system. CCHE integrity is needed to support CREVS in the role of control room personnel rrotection, but is not considered safety related. The Control Complex structure itself is seismically qualineu and considered safety related, but many of the elements of the habitability boundary (ie., doors, penetration seals) are not.

I

l REVISION 3 l LICENSING HASIS

1) Licensimo Bi ttkyround in 1981, the NRC issued an order to FPC confirming the commitments for TMI related requirements applicable to CR #3. NUREG-0737, item lli.D.3.4, Control Room Habitability, was included in this order. In response to requirements pertaining to this item, FPC performed a compreher , .c habitability evaluation of the CR#3 control room and eventually submitted its findings in the form of the revised CR#3 Control Room Habitability Evaluation" report, dated 6/30/87. This report concluded that the Maximum Hypothetical Accident was the limiting event with regards to Control Room Habitability and that thyroid dose was the most challenging criteria. Based on methodology consistent ithh SRP 6.4 guidance, the habitability evaluation found that the MilA would result in a thyroid dose of 26.5 REM. Subsequently, the NRC issued an SER on the findings of the habitability report, stating that the 26.5 REM thyroid dose was less than the 30 REM regulatory limit, and was therefore acceptable. This 26.5 REM thyroid limit is currently taken as the acceptance limit at CR #3 (Ref. 4).

2) ITSReanirement.t Section 3.7.12 of the ITS addresses requirements pertaining to the Control Room Emergency Ventilation System, and requires that two CREVS trains shall be operable during MODES 1,2,3 & 4, as well as during the movement of irradiated fuel. The LCO associated with one CREVS train inoperable provides for restoration of the out of service train within 7 days, or be in MODE 3 within 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months and MODE 5 within 36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months . If movement of irradiated fuel assemblies is in process, then CREVS is to be placed into emergency recirculation mode immediately or movement ofirradiated fuel suspended immediately. Witis both trains of CREVS inoperable the plant must enter LCO 3.0.3.

The ITS bases for Section 3.7.12 discusses operation of RM-A5 in support of CREVS, as well as isolation during toxic gas events. It states that the Maximum Hypothetical Accident is the limiting event with regard to Control Room liabitability in MODES 1,2,3 & 4, and provides a general reference to the " Control Room Habitability Evaluation" report dated June 30,1987 in this regard. The Fuel Handling Accident is identified as limiting in MODES 5 & 6. The ITS bases also state that CREVS ensures that the , antrol room will remain habitable following all postulated design basis events, maintaining exposures to control room operators within the limits of GDC 19 of 10CFR50, Appendix A. Notably, no reference or commitment is provided in the ITS or its Bases to SRP 6.4, " Standard Review Plan for Control Room Habitability". There are no soecific requirements regarding CCHE inleakage included in either the ITS or its Bases, Beyond the aforementioned general reference to the

" Control Room Habitability Evaluation" report, no discussion or reference is given either on the quantity ofin!cakage allowable or the manner in which inleakage is determined and applied. It can be considered an implicit requirement that the integrity of the CCHE be demonstrated as adequate to support CREVS in maintaining operator exposure within regulatory limits.

ITS Section 5.6.2.12 defines requirements pertaining to the Ventilation Filter Testing Program at CR

3, and requires that the CREVS filtration units meet minimum requirements regt.rding penetration, bypass and adsorption. This program requires that inplace testing be conducted which verifies the 2

, e l REVISION 3 l performance of the CREVS filtration system at a flow rate of 43,500 cfm +/- 10%, and that the pressure drop across the filtration unit be less than 6" wg when operating in this range.

3) CR-3 Operatine License The CR H3 Overatine License contains a reauirement to maintain Control Room habitability as specifiedhlhe vast-TML reautrements ofNUREG-0737. Ilowever. there is no reautrement for the measurement or evaluation ofinleakace teaccordance with specific reautrements.

4) 1%f R discussion Section 7.4.5 of the FSAR (Rev. 23) provides information on the elements of control room habitability at CR-3. Specific items discussed include:

- licensing background, including reference to the NUREG-0737, item lli.D.3.4 (Control Room liabitalility Requirements), the June 30,1987 CR-3 Control Room IIabitability Evaluation Report, and the May 25,1989 SER on the same subject

- discussion on allowable CCllE breach area

- discussion on CCilE inleakage, including specific reference to comparison with 0.6 air change per hour criteria in addition to the above, evaluation of the radiological dose consequences of the MilA and other DBAs are found in Chapter 14 of the FSAR. The consequences of the MilA are presented in section 14.2.2.5.10 and Table 14-54, which reflect a control room operator dose consequence of 22.7 REM based on previous analyses. This section also includes specific discussion of inputs into control room operator dose calculations, including CCllE inleakage, acumptions for ECCS leakage, ABVS filtration efliciency (0%), and post-accident meteorology. Chapter 14 also provides tabulated source terms associated with other DBAs, although no other accidents are specifically assessed with regard to control room habitability.

IMPACT ANALYSIS AND RELIAHILITY CONCERNS CCilE integrity and CREVS performance are primary inputs into the evaluation of control room habitability. The 26.5 REl% limit found in the habitability report and the NRC's SER on this subject are based on assumptions regarding these parameters. No specific value for CC11E inleakage is found in the ITS or CR #3 Operating License, but it is implicit that CCHE integrity must be such that dose limits are not exceeded. ITS Basis 3.7.12 does provide a reference to the habitability eva!uation report only in identifying that the MilA is the limiting DBA from the standpoint of control room dose consequence, but does not list or reference any specific inputs or assumptions contained in the report.

Therefore, revised analyses could change inputs into the Control Room IIabitability Evaluation report such as the determination and application ofinleakage without affecting the ITS or its Bases, so long as it did not invalidate the conclusion that the MilA is the limiting DBA of rewrd.

3

.l REVISION 3 l ITS Section 5.6.2.12 defines requirements pertaining to the Ventilation Filter Testing Program at CR

3. This program requires that inplace testing be conducted which verifies the performance of the CREVS filtration system at a How rate of 43,500 cfm +/- 10%, and that the pressure drop across the filtration unit be less than 6" wg when operating in this range. Requirements for the Ventilation Filtcr Testing Program as implemented by CP.148 require that inplace testing be perfomied at How rates within 43,500 cfm +/ 5%, with the tighter limits being conservatively imposed for testing inaccuracies, Revised dose calculations have been perfonned considering a How rate of 37,800 cfm, (correspondine to a Ilher dP of./ " ud, which is well below the 41,325 cfm corresponding to the lower

- end of this range. The impact of this lower How rate on CREVS operability requirements and control room dose consequences must be evaluated. Also, the apparent disparity between the design flow rate in the ITS and CI'EVS performance with current plant configuration must be reconciled.

DESCRIPTION OF IDENTIFIED CONCERN System Readiness Reviews conducted in 1997 during the CR-3 Design Basis Outage identified several issues which potentially impacted control room habitability. The predominant issue pertained to the validity of assumptions for CCllE inleakage. CCilE integrity and CREVS performance are primary inputs into the evaluation of control room habitability. The 26.5 REM limit found in the habitability report and the NRC's SER is bard on specific assumptions regarding these parameters. The June 30, 1987 habitability repot evaluated CCifE integrity based upon calculation methods found in SRP 6.4, which detemiined that up to 355 cfm ofinleakage could be tolerated without exceeding the licensing dose limit. Of this,285 cfm was assumed to be through untiltered pathways, and 70 cfm was assumed to be filtered inleakage through inlet dampers AllD-1 and AllD-ID. This value of inleakage was applied as a constant in control room operator dose calculations. Testing identified that, contrary to the assumptions found in the Control Room liabitability Evaluation report, differential pressures across the CCllE inlet dampers was such that leakage was occurring outward, which would correspond to a like amount of unfiltered inleakage at other boundary h> cations, in addition, differenual pressures at all boundary locations differed from that assumed to the extent that leakage less the 355 cfm limit could not be conclusively demonstrated. This condition was reported to the NRC in LER 97-022.

Prior to the current analysis, inleakage was assessed on the basis of guidance found in SRP 6.4, to which FPC is conunitted "for guidance" only. Using this method, inleakage was assessed by summing the calculated leakage past CCilE boundary elements (ie., doors, penetrations) at a differential pressure of 1/8" wg, then dividing the result by two. Additional " enhancements" were made for inleakage past ventilation system boundary dampers which might be operating at a higher differential pressure and for access / egress. This inleakage value was then applied non-mechanistically into dose calculations to determine operator exposure.

bregrated icJeakage testing has been performed to provide an assessment of CCHE integrity. This integrated testing used tracer gas methods to directly measure building inleakage while operating in the post-accident alignment. If a non mechanistic baseline inleakage value were obtained by correcting the ,

tracer gas test results to 1/8" wg, and this value applied as a constant value in dose calculations (as was '

done previously), the acceptance limit for dose would be exceeded. To calculate dose based on test results, FPC developed a model which predicts inleakage under various differential pressure conditions and so can assess mechanistically induced inleakage under postulated post-accident corditions. The measurement of inleakage and its mechanistic application in dose calculations is a significant change 4

l

REVISION 3 l from previous methodology and is expected to produce a result which is more realistic, but might be less conservative. Tids issue has been determined to represent an unresolved safety question, and is the primary focus of this JCO.

LER 97-022 also identified that past modinc stions were implemented which added resistance to Control Complex Ventilation System ductwork without fully assessing the impact on recirculation Gow rate (Ref. 6). As a result, the system flow rate with clean filters is now somewhat lower than the 43,500 cfm nominal design How rate referenced in the habitability evaluation and used as an input to obtain the 26.5 REM limit. Subsequent dose calculations have addressed How rates as low as 39,150 cfm (43,500 - 10%), but evaluation of current conditions predict now rates on the order of 37,800 cfm assuming 4" wg drop across fouled filters. It is noted that the 4" filter fouling value is less than the 6" currently renected in the ITS and taken as the combined (IIEPA and charcoal) filter fouling limit. Curren' procedures constrain eperation within 43,500 cfm +/-11 cnd ensure that the 37,800 cfm minimum now rate requirement for dose calculations is met. Ilowever, the use of a filter fouling limit which is less than ITS requirements must also be assested.

Finally, numerous changes to assumptions / inputs associated with Control Room Itabitability have been implemented under 10CFR50.59 since the habitability report was issued in 1987. Generally, these changes have been assessed individually at the time of their implementation, but without updates to the habitability report to maintain this licensing document consistent with plant design on a real time basis. This JCO includes a matrix which lists the former and current values of habitability evaluation inputs, identifies the iterations associated with each parameter up to the present point in time, and provide a brief discussion for the basis and acceptability of each change.

OEERAlllLITY EVALUATION The activities associated with Restart Issue R-12 have improved the performance of CREVS and the CCIIE in assuring control room habitability. Redundant bubble-tight dampers have been installed which assure positive closure of ventilation How paths. Vestibules have been installed on all CCIIE boundary doors which provide an extra measure of leak resistance and provide defense in depth for nonnal operation and wear. An extensive penetration sealing program was conducted to improve boundary leak tightness. Leakage of the habitability envelope has been accurately measured under actual operating conditions using tracer gas test methods.

The modifications and efforts driven by resolution of R-12 required that the Control Room operator dose calculations be revised to align inputs and assumptions with plant design. The basic methodology used in these revised calculations is consistent with that found in regulatory guidance and utilized in previous calculations lie.. Ree. Guide 1.4 source terms. RB modeline ver SRP 6 5 2. meteoroloev ner Murphy Camve). Ilowever, many of the inputs found in the Control Room IIabitability Evaluation report have been superceded on the basis of actions associated by R-12, and still others have been revised under 10CFR50.59 since its submittal in 1987. A detailed comparison of significant inputs into current Control Room IIabitability analyses vs. those found in the 1987 habitability evaluation report is provided in Table A to this JCO. Notably, the methodology for determination of CCIIE inleakage and the manner in which inleakage values are applied in dose calculations is a significant change from previous methodology, including that described in the Control Room IIabitability Evaluation report.

5

REVISION 3 l Previous dose calculations were based on a model defined in NRC regulatory guidance which determines inleakage at a Control Complex pressure of 1/8" wg., then applies this inleakage as a constant value in dose calculations. This model is a good correlation for more common control room designs which rely on taking contaminated air from outside the envelope, filtering it, and then using this filtered air to pressurize the control room. In this case, pressurization of the habitability envelope to a nominal value of 1/8" is used to prevent inleakage, and is generally accepted as being high enough to overcome the effects of wind, thermal effects, and operation of ventilation systems in adjacent structures. This model does not correlate well to the CR-3 habitability envelope, which is a filtered recirculation system with iso'ation (no makeup). In this design the bulk pressure in the envelope is neutral, and the 1/8" differe1tial pressure in the NRC's model is used to drive inleakage across the envelope boundary. The qu.mtity of inleakage induced in this manner is used to develop a baseline inleakage value for use in control room operator dose calculations.

FPC has devaloped its own model for deter.nining CCHE inleakage for the purpose of input into control room operator dose calculations. This model is based on more realistic, but still conservative, considerations of the actual riotive forces which would exist for driving inleakage under postulated post-accident conditions. The' use of this model is a significant departure from previous methodology and is not described in regulatory guidance, and is considered to be an unreviewed safety question on that basis. The existence of an unreviewed safety question does not necessarily mean that an activity or system condition is unsafe. Rather, it is necessary for FPC to evaluate the safety of the condition and to assess its implications rn unit restart as provided in Generic Letter 91-18, Rey,1. As detailed within this JCO, FPC and its contractors have reviewed the treatment of CCHE inleakage iri control room habitability calculations md concluded that the change in methodology is safe and that CREVS operability and CCilE integrhy are not compromised. Ultimate resolution of the unreviewed safety question requires either approual by the NRC or placing the system in a condition which has been reviewed.

ilFSTIFICATION FOR CON TINUED OPERATION Rehtive te the issue of control ioom habitability, justification for continued operation in all operating MODES is based on establishing the operability of the CREVS and the adequacy of CCllE integrity.

Additionally, the existence of other mitigation features, such as ves;ibules at CCilE boundary doors, source term reduction by the Au(iliary Building filters, and Radiological Emergency Re>ponse Plon dose management procedures pmvide additional assurance of control room personnel protection. A discussion of the basis fbr operability of the CCilE follows:

CREVS Onerability Operability of CREVS requires l hat the specific requirements of ITS Sections 3.7.12 and 5.6.2.12 are i satisfied and that the bases for these technical specifications are not invalidated. Surveillance Activities !

required to demonstrate CREVS operability include (1) operating each CREVS train for at least 15 j l minutes each month, (2) satisfying the ventilation filter testing program, and (3) verifying that each CREVS train actuates to the emehgency recirculation mode on an actual or simulated actuation signal

! every 24 months. Of these criteria, the first and last will be included in post-moa.fication testing associated wnh damper replacement. The second criteria refers to the requirements of the ventilation filter test program defined in ITS Section 5.6.2.12. l 6

-l RiiVISION 3 l The ventilation filter test program requires that the CREVS filtration units meet minimum requirements regarding penetration, bypass and adsorption. Section 5.6.2.12 of the ITS also requires

. that the pressure drop across the combined llEPA and charcoal filters be demonstrated as no more than 6" wg when tested at a system flow rate of 43,500 cfm. - LER 97-022 identified that past modifications were implemented which added resistance to Control Complex Ventilation System ductwork without fully assessing the impact on recirculation flow rate. As a result, the system flow rate with clean filters

. is now somewhat lower than the 43,500 cfm nominal design How rate. This is evident in a recent flow evaluation of system performance which a calculated recirculation flow rate of 37,800 cfm with 4" rg across fouled filters.

The use of 4" wg filter fouling limit is less than the 6"wg value currently reflected in ITS section 5.6.2.12 and taken as the combined (IIEPA and charcoal) filter fouling limit. Since a 4" wg differential pressure across the filtration unit is expected to correspond to 37,800 cfm, it follows that the flow rate associated with the 6" fouling limit would be lower still, llowever, current surveillance procedures constrain operation to flow rates at or greater than 41,325 cfm (43,500 - 5%). Operation within this constraint has the effect oflimiting filter fhuling to a much tighter range thm even 4" wg, and ensures both that the level of filter fouling is within the 6" wg ITS limit and that the 37,800 cfm minimum flow rate postulated for dose calculations is met. License Amendment Request, LAR #222 addresses this issue in the ITS, such that CREVS flow rate and allowable filter pressure drop are reconciled.

Reduced flow rate also has the potential to degrade CREVS equipment cooling capability. Issues pertaining to CREVS cooling capability are being addressed separately by Technical Specification Change Request Notice (TSCRN) #210, and are not included in this JCO.

This LER also identified that charcoal testing has been performed in the past using non-conservative temperature and humidity conditions. Charcoal adsorption capability increases with increasing temperature and decreasing humidity. Prior to this outage, laboratory charcoal testing has been performed at a 80 C and 70% RH. Ilowever, CREVS system readiness reviews determined that this criteria was unconservative with regard to possible post accident conditions. To address this concem the ventilation filter test program is being revised to require laboratory charcoal testing at 30 C and 95% Ril. Prior to restart, laboratory testing will be performed to both the old and new test conditions.

CCHE Intecrity Not a system per se, the CCilE is the physical boundary which is credited with protectig the control room operater from the efTects of postulated DBAs or toxic gas events. The CR-3 control room ventilation system design is categorized according to SRP 6.4 as a zone isolation system with filtered, recin:ulated air. As such, it does not rely on pressurization to limit inleakage, but rather on leaktightness and filtration capability to provide the necessary level of protection. Operability of the CCHE is predicated on demonstrating a level ofintegrity such that, in conjunction with the operation of CREVS, suflicient protection is provided for the control room operator to ensure that exposure limits associated with DBAs and taxic gas events are not exceeded.

1)Inleakage detennination l

7 l

l

l REVISION 3 ,

l Previously, CCilE inleakage has been determined on the basis of methodology described in SRP 6.4, which provides a standard means by which to derive a " baseline" inteakage. The basic SRP 6.4 methodology for detennining base inleakage in a recirculation system is to measure (or calculate) the air flow required to pressurize the habitability envelope to 1/8" wg, then divide the result by two. This method ensures that all penetrations are subjcet to test pressure and provides a conservative (but relatively arbitrary) baseline inleakage value. The 1/8" value is not associated with any particular post-accident conditions, but is large enough to minimin the impact of test inaccuracies and local pressme effects. The wind speed necessary to generate this differential pressure is on the order of 15 - 20 mph, much larger than the low wind speeds associated with the 5% worst yfQ value used in SRP 6.4 to minimize source dispersion. SRP 6.4 test methods would include additional enhancements for inleakage through boundary dampers which may be subject to unusually high differential pressures, and a 10 cfm allowance for personnel access / egress during an accident. Smce FPC is committed to SRP 6.4 only as a " guidance" document, there is no commitment to adhere to this methodology, nor is there any requirement to assess building inleakage at a differential pressure of 1/8" wg for the purposes of evaluating Control Room dose consequences.

Although not prescribed by regulatory guidance, application of tracer gas technology is recognized as a means to accurately and directly measure building inleakage nnd is being increasingly utilized in the nuclear industry for this purpose. Using tracer gas test methods, it is possible to set up a test to measure inleakage under conditions which are representative of a specific postulated scenario. Determination of inleakage based on testNg under simulated post accident conditions is a departure from the existing licensing basis for CCilE inleakage, but isjustifiable given that this methodology is expected to provide a more accurate prediction of dose consequences and that CR-3 is nat committed to SRP 6.4 except as a guidance document, Tracer gas testing under post-accident conditions would have the Control Complex in its emergency recirculation mode, and treat the entire CCliE as a lumped volume. No additional penalty would be required at boundary damper locations because these dampers would be subject to the same pressures during testing as would be expected during post-accident operation.

(Note that the latter effect is inconsequential for CR-3 given that bubble tight dampers will be installed at all boundary isolation locations.) The use of a 10 cfm allowance for access / egress is still applicable, and would be incorporated into the final inleakage result.

Developing a test replicatin post-acciderat conditions requires that the scenario be postulated which provides both realistic and challenging conditions from the standpoint of exposure to source term and maximizing the differential pressure which drives inleakage. The postulated conditions may result in a variance of differential pressure across individual CCIIE-elements, and are not intended to necessarily test all penetrations to a cenain minimum differential pressure. Rather, the objective is to develop a test which realistically and conservatively gauges the overall consequence associated with

- the limiting post-accident scenario. The motive forces which might induce a significant differential pressure across the CCilE are taken as wind pressmes (assuming a loss of offsite power) and ventilation systems in adjoining structures (no loss of offsite power). Significantly higher differential pressure would be expected assum...g no loss of offsite power, but the source term would be lower given that this wo ild necessarily require the Auxiliary Building Filtration System to be in operation.

! Therefore, to fiilly assess limiting post-accident conditions requires that both scenarios be evaluated.

This is accomplished by measuring inleakage at a known differential pressure using tram gas

raethods, then analytically adjusting this value to correspond with postulated conditions.

l

2) Test Conditions 1

8 l

REVISION 3 l The tracer gas test model is established based on consideration of site layout, source terms and possible :

plant operating conditions. In the event of a MilA w / LOOP, given that the vast majority of penetrations are either on the north (Turbine Building) or south (Auxiliary Building) walls of the control complex, it follows that north / south wind directions would tend to maximize CCilE inleakage. Similarly, with no loss of ofTsite power, the Auxiliary Building supply fans would be secured by radiation monitor RM A2, causing the Auxiliary Building to develop a negative pressure and inducing leakage through the CCllE in that direction. In either case, the tracer gas test which models these conditions would utihze the Auxiliary fluilding Ventilation System te induce CCilE inleakage by creating a negative pressure in the Auxiliary Building. The following conditions were prescribed for the tracer ps testing which modeled CCIIE inleakage:

- The Centrol Room Emergency Ventilation System (CREVS) was played in emergency recirculatior, mode and operating nonnally. Both " Toxic Gas" and the "lligh Radiation" recircalation lineups were tested. All CREVS boundary damper locations were scaled " bubble tight" to duplicate post modification conditions.

(NOTE: The isolation dampers were tested in-place and the test boundary exhibited insignificant leakage.)

- All fans in the Turbine Building Ventilation Systein were secured. The turbine building normally remains well vented to atmosphere through normally open doors, roll out windows and roof vents.

- All fans in the Intermediate Building Ventilation System were secured. Note that conditions m the Intennediate Building are not deemed critical to the test in that relatively few penetrations are on the CCIIE / Intennediate Building Wall.

- The Aux. Building Ventilation was operated to test pressure of approximately 0.171" wg negative pressure vs. the Turbine Building. This value is large enough to minimize test inaccuracies and external effects and was sustained for the duration of the test.

- The tert was conducted on backshift when personnel traffic is minimized. Since a 10 cfm allowance for access / egress would be analytically applied, minimizing traffic precludes additional penalization for this effect.

- Testing was conducted with vestibule doors blocked open. This conservatively assumes no credit for the additional integrity provided by vestibules.

All loop seals penetrating the CCHE were verified to be filled prior to testing.

Controls are in place to ensure that these loop seals are periodically filled during plant operation.

Tracer gas testing conducted under these conditions measured an inleakage of 462 cfm in emergency recirculation mode at a differential pressure of 0.171" wg. Using this information, the inleakage at other differential pressures can be predicted by the use of the formula Q = C AP" , where Q = air flow in efm 9

. . - - - - - ~

IEVISION 3 l C = inleakage coefficient P = differential pressure, and n = flow exponent.

According to ASilRAE, values o nv are typically between 0,6 and 0.7. Values less than 0.171 are estimated by using n = 0.5, whic'i gives conservative results for estimating inleakage at pressures less than the ter' value. For extrapolation to conditions above 0.171" wg, the use of n = 0.5 o somewhat unconservative, and a more realistic value of n = 0.65 is chosen.

3) Analysis of MilA w / LOOP Since wind pressures are assumed to be the primary motive force under MilA w / LOOP conditions, inleakage for this scenario is determined by examining meteorologicat conditions associated with event analysis. SRP 6.4 methodology assumes post-accident meteorological conditions corresponding to the 5% x/Q value during the critical initial stages of the event in order to minimize dispersion of the radioactive plume as it is carried from the containment building to the Control Complex. The methodology then allows for three incremental increases in wind speed and direction over the duration of the accident due to the extreme improbability that these initial wind conditions would be sustained over an extended period of time. Based on these considerations, inleakage values are derived for each of the four time intervals over which x/Q values vary by correcting inleakage at the test differential pressure to the differential pressure induced by the wind speed a;sociated with that interval. Tlese -

wind induced ditTerential pressures were conservatively calculated using ASilRAE methods. Exb of p these inleakage vahies is an input into the appropriate interval in the revised radiological dose calculations such that the wind speed associated with plume dispersion corresponds to that which drives inleakage through the Control Complex boundary.

For the MilA w /1.OOP, it is noted that the use of low wind speeds provide relatively small motive force for inducing leakage throug;n the CCIIE, llowever, parametric studies show that, over the range ofinterest, increased wind speeds will tend to lower Control Room dose when it is applied uniformly to bcth x/Q values and building ditTerential pressure. It is also noted that, at these relatrsely low wind speeds, the potential effects of thumally induced inleakage becomes sig,nificant. Differential pressure across walls induced by ditTerences in inside and outside temperatures (i.e., stack effect) can be pronounced in tall structures, as its magnitude is basically a function of the difference in temperatures across a wall and the difTerence in height from a given penetration to the building's neutral pressure level. Although ihe CCilE is a relatively tall structure, the following considerations tend to mi dmize the impact of this phenomenon on control room operator dose consequence:

The temperature gradient between the Control Complex and adjacent areas at the outset of an accident would be relatively small. Given a source term model wherein the majority of exposure occurs during the :nitial stages of the event, leakage induced by the stack effect would be relatively small during this critical period.

. The neutral pressure level of a building wall tends to be towards the elevation containing the largest leakage area. or in the case of uniform leakage, at the vertical center of the building. The majority of CCllE penetrations are at or near the elevation of the cable spreading room, which is itself just below vertical center of the Control Complex elevation. Since the stack effect to

REVISION 3 l results in no appreciable differential pressure at the neutral pressure level and differential pressures which increase with distance from the neutral pressure level, the distribution of CCilE penetrations would tend to minimize the inleakage due to stack effect.

At higher wind speeds, the inleakage induced by wind pressure is dominant and stack effect pressure provides a lesser relative contribution to inleakage.

For MilA w / LOOP, the contribution of stack effect to inleakage was conservative:y considered by calculating stack effect pressures during both winter and summer conditions. A uniform temperature j of 31 *F was assumed in adjacent areas for winter conditions, while 118 F was used to assess l summertime conditions. The Control Complex itself was assumed ta remain at its design temperature l of 75 'F. These values were taken to remain constant for the duration of the 30 day accident. An average stack effect pressure was calculated, and inleakage associated with this value determined by application of relationship between building differential pressure and inleakage derived from tracer gas ,

tent results. This value was then combined to wind induced inleakage using the ASilRAE formula Oc. - (0/ + Q&'

The MilA w / LOOP analysis also gave consideration to CCllE leakage which might be induced by localized high and low pressure areas induced within the CCllE boundary by the operation of the ventilation system. Obviously, any leakage which occurs as a result oflocal high pressure areas within the CCllE would be outicakage, and of no concern with regard to dose control room consequence. It is also reasonable to assume that inleakage caused by virtue oflow pressures created within the CCilE by the ventilation system would be induced into the system return ducting, and thereby be considered Gitered inleakage. Traverse measurements taken during tracer gas testing show that significantly more supply air is directed into the Control Room elevation than that removed by return ducting, resulting in this area being slightly positive with respect to bulk CCilE pressure and, more importantly, to the cable spread room elevation immediately below it. This is significant in that the majority of CCilE penetrati<ms and leakage areas exist on the lower elevations (from the cable spread room elevation down). It follows that a relatively small percentage of CCilE inleakage occurs on the Control Room elevation, and that very little inleakage occurring on the floors below the Control Room elevation is introduced into the Control Room except by virtue of the ventilation system, where it would be filtered in the process. The assumption that a large percentage of all C" LIE inleakage is actually filtered would not be unrealistic on this basis. Since the filtration system is considered to be 95% ef6cient, neglecting these considerations is an extremely conservative treatment of CCilE inleakage, it can be readily seen that treating all CCilE inleakage as untiltered is a extremely conservative posture. In revised control room operator dose calculations, inleakage induced by CREVS operation is assessed by including an additional penalty of 125 cfm of filtered inleakage. Given that testing was performed with CREVS in operation such that this effect existed at the time, classification of any portion ofinleakage as filtered for this reason could be taken as a reduction in unfiltered inleakage.

Instead. this filtered inleakage penalty is superimposed on unfiltered inleakage due to wind, stack pressures and access / egress, and is applied for the entire 30 day duration of the event This again is extremely conservative treatment ofinleakage assumptions since the penalty for filtered inleakage is taken both directly in tracer gas test measurements and analytically superimposed again in dose calculations.

11

RhVISION 3 -

l

.4) /.nalysis of hillA without LOOP-l Given the occurrence of the hillA.withaut a loss of oftsite power, the ventilation systems in !

i adjacent buildings' are assumed to continue to operate during and aner the accident, increasing levels of radiation in the Auxiliary Building as sensed by radiation monitor Rhi A2 would result in a trip of the Auxiliary Building Ventilation System '(ABVS) supply fans, resulting in a signilhantly larger negative pressure in the Auxiliary Building. The Turbine Building is-considered to be essentially at atmospheric pressure due to the numerous large openings in that ,

structure. Under these conditions, the post accident leakage into the Control Room could be l significantly higher (especially during the early time steps) than that postulated on the basis of wind pressures (ie., hillA w / LOOP).

The release path for this scenario is based on the activity being released from the Containment and subject to initial dispersion as it travels to the Turbine Building Ventilation System intakes and into the Turbine Building. From that point it ultimately enters the Control Room as unfiltered inleakage by the differential pressure induced across the Control Complex. This release path model considers dilution into the large Turbine Building volume as well as minor decay and holdup while the activity is in the Turbine Building. .

The cvaluation of hillA without LOOP has four distinct changes from the version of the event which assumes LOOP; (1) given that the ABVS must be in operation to induce the differential pressures of concern, then filtration by the aBVS charcoal filters occurs and there is no requirement to assume an ECCS pump seal failure at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months after the accident with a leak rate of 50 gpm for 30 minutes, (2) the normal ECCS leakage which does occur is assumed to be filtered to 75% efliciency,

'(3) the activity will enter the Control Room via the Turbine Building and as such will be subject to some delay due to the buildup and decay in ti.e volume of the Turbine Building, and

' (4) inleakage will be constant for the duration of the accident and will not be affected by the wind speed used in the dose analysis. This is conservative in that the wind direction

. necessary for transport towards the Turbine Building would tend to oppose inleakage through the CCilE towards the Auxiliary Building.

Temperature effects in this scenario are assumed to be insignificant given that continued operation of adjacent ventilation systems minimizes the temperature differentials between these areas and the Control Complex. The 75% efficiency assumed for the Aux. Building charcoal filters is consistent with that recently allowed for these filters by the NRC in control room habitability analyses. (ref. discussion on the ABVS in Table A) Inleakage induced by CREVS operation is also ignored on the basis that conditions at the time of tracer gas testing are similar to those postulated under these post-accident conditions, such that this factor was present during in the tests.' As with the MilA w / LOOP, analysis of this scenario assumes that 12

. l IEVISION 3 l )

inldkage . is distributed evenly throughout _ the CCllE volume, even though test data -

substantiates that very little micakage is introduced into the Control Room from the floors below it without being subject to filtration. Given this and other conservatisms in the analysis, the treatment of MilA without 1 OOP described above is considered to be a very conservative -

treatment of this scenario.

5).Results of M'IA Analyses .

The results of this analysis shows that the bounding version of the MilA is that associated with the

' accident occurring with LOOP. Calculations show that a 26.5 REM dose limit can be maintained in 2

this scenario while allowing an additional CCllE breach area of up to 22.8 in in addition to that measured by the tracer gas leak test. The 26.5 REM vale, corresponds with that in the Control Room liabitability Evaluation report dated June 30,1987 (as referenced in the ITS Bases) and the NRC's SER in reply dated May 29,1989, and is taken as the acceptance limit at CR 3. It is concluded that, 2

given that CCilE breach areas are maintained below the value of 22.8 in , the level of CCllE integrity is sufficient to meet operability requirements pertaining to radiological consequences of the MilA.

'6) Analysis of 0ther DBAs A review of other design basis accidents for which CR #3 is licensed was performed to verify that the MilA as analyzed above is the limiting event. This review was based on (1) a review of source tenns, (2) a review of the means by which isolation of the CCliE is achieved, and (3) consideration of plant operating conditions (ie., operating MODES) at the time of the event.

This review found that the MilA source temiexceeds that associated with all other DBAs as analyzed in Rev. 23 of the FSAR. Ilowever, MilA accide:nt analysis assumes that CREVS boundary dampes are isolated essentially from the outset of the event by virtue of the 4 psi Reactor Building liigh Pressure ES signal. Since other DBAs might not actuate this signal, a review of events has been performed which rely on automatic radiation detection or operator action for isolation. Based on this review, a detailed analysis of the Steam Generator Tube Rupture event was perfonned which demonstrated that isolation of the CREVS was not necessary to maintain operator exposures less than regulatory limits. Further, given any reasonable isolation time either by the radiation monitor or operator action, the MilA remains the bounding event with regard to control room habitability.

The inputs, source terms and dose consequences of the SGTR as analyzed are presented below:

13

REVISION 3 l i

SGTR INPUT Parameter l Value l Comments Thirty I'our Minute Isolation Time Source Term (34 min. lsolation Analysis) -

Reactor Coolant Pressure 2200 psia Average Temperature of th; reactor coolant $79 F Volume of the unsprayed region I ft' Assumption for instantaneously release to atmosphere. ,

Volume of the spra; ed region I tt' Assumption for instantaneously release to atmosphere.

Projected Containment area of wind wake 1852 m' or 19,933.211' ,

Elemental lodine Fraction 0.91 Particulate lodine fraction 0.05 Organic lodine Fraction 0.04 Control Room Volume 364,922 fl' Purge flow rate to atmosphere 100 ll'/ min Assumption for an instantaneous release to the atmosphere.

Control Room tireathing Rate 3.47E 04 m'/sec

, intake (x /Q') 0-2 hour 9.0E-04 sec/m' _

0-8 hour control room clicctive wind speed 1.2 m/sec 8 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months control room effective wind speed 2.034 m/sec l 4 day control room ellective wind speed 5.320m/sec 4 30 day control room elTective wind speed 18.182 m/sec occupancy factor 1.0 incorporated into the E;Tective Wind Speeds Unlittered leakage into the control room LU cfm Caculated on CCIIE differential nressure of 220

" HE.

Control room makeup air flow 5335 ft' Assumed design goal of $700 cfm less the unfiltered leakage Recirculation of air in the control room 37.800 cfm bxline Partition Factor (O . 34 minutes) 10 Release factor through the steam relief valves lodine Dose Conversion Factors ICRP30 Gamma Correction Factor for Control Room Dose 0.0 POSTDilA Default Values.

Primary to Secondary Leakage through affected 435 gpm Steam Generator Primary to Secondary Leakage through unaffected I gpm Steam Generator Recirculation Filter Efliciency 95* for iodine species Eight flour Isolation Analysis: Uses the above input and assumptir,as unless same variable is shown below.

Eight hour isolation source term -

lodine Partition Factor (0 - 24 minutes) 10 Release factor through the steam relief valves lodine Partition Factor (34 minutes - 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months ) 10" Release factor through the condenser 14

. .. _. - - . - = . . - - . -. - .- . .

REVISION 3 l

- Steam Generator Tube Rupture Source Term (Both Analyses)

Isotope Concentration pCi/ml Kr-85m 1.54 Kr-85 8.94 Kr-87 . 0.84 Kr 88 2.69

. Xc-131m 2.40 Xe-133m 2.79 Xe-133 250.0 :

Xe-135m 0.93 Xe-135 5.96 Xe 139 0.51 1-131 3.17 l132 4.8I l-133 3.75 :

1-134 0.499 l-135 1.92 Steam Generator Tube Rupture Activity Released 1sutope Activity Released (Ci) Activity Released (Ci)

(34 min. Isolation) (8 hour9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months Isolation)

Kr-85 5.02E+02 7.08E+03 Kr-85m 8.64 E+0 : 1.22 E+03 Kr-87 4.7 I E+0 i 6.66E+ 02

- Kr 88 1.51E+02 2.13 E+03 Xe-13 im 1.35E+02 1.90E+ 03 Xe-133m 1.57E+02 2.21 E+03 Xc-133 1.40E+04 1.98E+05 Xe-135m 5.22E+ 0 i 7.37E+02 Xe 135 3.34E+02 4.72h+03 Xc-138 2.86E+01 4.04 E+02 l-131 1.78E+02 2.5 I E+03 1 132 2.70E +02 3.8 I E+03 l l-l33- 2.10E+02 2.97E+03 l l 134 2.80E+01 3.95E+02 l-135 1.08E+02 1.52E+03 15

REVISION 3 l 1

SGTR ACCIDENT CONTROL ROOM DOSE (REM) -

(without isolation)-

Thmid Tj.og Wholebody Thirty four Minute Isolation 2,23]d 8.01 E-3 _

l Eight Ilour Isolation - fLLI E 1.03 E-1 (with isolation initiatec' by RM-AS)

Thyroid Doz Wholebody Thirty four Minute Isolation. 3.20 E-2 7.31 E-3 '

Eight ilour Isolation 3.33 E-2 8.78 E-2 7). Analysis ofToxic Gas Events CCilE integrity is also required to provide protection to the control room operator in the event of a toxic gas accident. Reg Guide 1,78 provides information and assumptions for assessing toxic gas accidents with regard to control room habitability. From this document comes the basic criteria that, in the event of a toxic gas accident, appropriate toxicity limits not be exceeded in the control room 2 minutes aller initial detection in order to allow the operator adequate time to take action (ie., don an air pack) prior to being overcome. The reg guide allows for detection to be accomplished by personnel (nasal detection) or with automatic detection equipment. CREVS isolation, if required, can be attained '

either by operator action or by an automatic signal from toxic gas detectors. At CR-3, AP-513 provides the appropriate instructions for the operator in the event of a toxic gas accident, including the use of air packs.

~ Based on previous evaluations, the locations and quantities of toxic gas storage sites at CR #3 which pose a potential liability to' control room habitability and must'be specifically addressed are listed below:

Toxic Gas container Size and Location -

Toxic Gas llelper Cooling CRl/2 CR4/5 Towers Chlorine 17 tons none I ton cylinders SO 2 50 tons - 45 tons I ton cylinders 16 -

REVISION 3 l iThe most limiting source of toxic gas is a SO 2tank at CR1_which has been administratively limited to storage of 30 tons.187-0005 Rev. 2 " Control Room SO 2 Concentration from CRl/CR2 SO 2 40 Ton Tank Failures", is the current calculation of record that analyzes this event. Case 1 analyzes a 30 ton tank rupture, with 5700 cfm make-up flow and a 30 second delay between attaining the detector's set point and the dampers reaching the closed position. Inleakage only develops after damper closure. In this scenario, the event is detected and recirculation is automatically initiated. The cc.ntrol room concentration 2 minutes aRer detection is 29 ppm, vs the 36 ppm limit previously established in the Control Room liabitability Evaluation report.

The following equation - from NUREG 0570 converts gas concentrations at the air intake to ;

concentrations in the contml complex:

C,(t)) = C,(t).i) + [Co(t)) - C,(t). )](1 - EXP**)

where:

(t)) = Control Complex concentra@:. at thej* timestep (ppm)

Co(tj ) = Air intake wncentration at thej"' tme step (ppm)

W = the vcatilation/ infiltration rate before/after inlation (cfm/60sec/ min) deltaT = time steps (seconds)

V = control complex volume (ft')

When CREVs shins to recirculation, W changes from the normal make-up valuc of $700 cfm to an inleakage value which is around 10% of the initial make-up flow rate. By virtue of this effect alone, subsequent changes in control room concentration occur much more slowly given that the concentration of toxic gas in outside air remains constant. For this reason the most challenging version of the toxic ges scenario is the non mechanistic complete tank rupture (rather than longer duratica releases of lesser rates). This scenario is modeled by a " puff" for the large quantity of gas whien flashes at the time of the accident, followed by a longer tenn " plume", modeling the subsequent evaporation of the remaining spilled volume. The puff contains a much higher concentration than the plume, but is of such brief duration that most of this release has passed by the Control Complex at the two minute mark and concentration in the outside air is rapidly dropping. Using this model maximizes control room toxic gas concentrations by introducing the highest outside concentration of gas concurrent at the maximum possible flow rate. Control room toxic ges concentratien changes aner isolation occur very slowly due to the twofold eftect of much lower concentration of Wie gas in the outside air, being introduced at a flow rate which is approximately an order of magnitude less. The net effect is that the toxic gas analyses are insensitive to small variations in CCIIE inleakage.

A great deal of conservatism is inherent in the simplistic model described above. To demonstrate this, Cermak Peterka Petersen, Inc , modeled the CRl, 2 and 3 site and performed wind tunnel tests simulating a 30 ton release at the tank, empirically measuring the toxic gas concentration at the control complex intake. This data was input into Sargent.& Lundy calculation SL-9929-M-0008 with the above formula to determine the Control Complex toxic gas concentration. This analysis determined a SO2concentration in the control complex of 17.9 ppm two minutes aRer nasal detection whhout CCHE isolation, well below the toxicity limit of 36 ppm. Comparison of this result with the previous analyses which predicted A concentration of 29 ppm with automatic isolation approximately 30 seconcts into the event, it is readily apparent that a great deal of conservatism exists in the analytical model l 17 L-I' ._ _ _ _ _ _ _ _ _

. REVISION 3 l (FPC is in the process of removing the 30 ton tank from CR1 and replacing it with a pelletized system that generates sulfur dioxide gas when needed.. This activity will not be completed prior to restart, and the above analysis is valid in the interim per9d.)

The next most limiting toxic gas source is the llelper Cooling Towers. Current 1, / there is no SO2 or Cl2 stored at this location and this is ensured by a Crystal River Unit 1 " Red Tag" clearance No.1997-01543. Prior to releasing this clearance, administrative controls will be in place to limit the liciper Cooling = Towers to 8 tons of Chlorine and 30 tons of S0, 2 189-0053 Rev. 3, " Control Room liabitability llelper Cooling Tower Project", is the current calculation of record that analyzes ruptures of these tanks. The calculation analyzes the 17 ton chlorine and 50 ton SO2 tank ruptuies and found that automatic isolation was required only to meet the two minute chlorine toxicity limits. FPC has used the new Sargent and Lundy calculation (SL-9929-M-0008 Rev. 0) to evaluate the lower quantities: 8 tons of chlorine and 30 tons of SO . This analysis combined the wind tunnel results from the CR-1 SO2 tank model and traditional atmospheric dispersion mathematical modeling techniques to conclude that CREVs could remain in its nornml alignment (ie., no rCilE isolation required) without exceeding Control Room toxicity limits if up to 9 tons of Cl2 or 50 'luns of SO 2were released. Thus, CCilE inleakage is of no consequence given the new limits of 8 tons and 30 tons for Cl 2and SO2 respectively.

Sargent and Lundy calculation SL 9929-M-0008 also analyzed the Cl2 and SO; stored at CR-4/5. This calculation allows for a lion chlorine rcIcale at the CR-4 and 5 cooling towers located 3600 feet fmm the CR3 comrol complex intake. There are cip,ht I ton tanks with four in service on a sincie header at a time. The assumed accident has one tank fail and the other three leak out though the common piping.

The allowable dlct1 leak bounds this condition so no automatic detection or isolation is needed at CR3 for this source. The one ton sulfur dioxide tanks were not analyzed due to the sulfur dioxide at the llelper Cooling Towers being more limiting. The amount at CR 4 and 5 is less (50 tons versus I ton),

farther away (3400 feet versus 3600 feet), and has a larger building wake (2 versus 3). Since the calculation allows > 50 tons at the llelper Cooling Towers without automatic detection and isolation, then the CR4 and 5 sulfur dioxide will also not require the same.

Based on the above discussion, it can be seen that the current level of CCllE integrity provides adequate protection for the control room operator for postulated toxic gas events. It is also noted that the updated analyses would support operation without crediting the existing toxic gas detectors.

~

Ilowever, it is not the intention of this JCO to delete these monitors, and they will continue to be installed and surveilled as in the past as an additional conservatism.

CCilE breach margin in modes 5 & 6 is based on potential consequences of an SO2 release at Uniu 1/

2, for which previous calculations demonstrated that up to 1400 cfm of inleakage could be tolerated with the CCllE initially isolated. Since revised analyses for this event demonstrate that automatic

-isolation is no longer required, the use of considerably larger inleakage rates (up to as much as the 5.700 cfm makeup rate) could be justified. No increase in the breach margin for modes 5 & 6 is being undertaken by this JCO.

18

Kl:VisloN 3 l Additional Protective Features

1. Yestibules CCllli boundary doors represent a significant source of leakage into the CC. There are three double doors and three single doors. Vestibules were added to the three CCilE double doois in 1996, and j have proven to be effective in reducing differential pressure across the existing doors. Reducing the differential pressure exerted on the boundary doors has two benefits. First, lower differential ptessure reduces leakage around the doors, and second, lower differential pressure allows the doot closers to perform more reliably. During this outage, vestibules were added to the single CCilE toundary doors. All of the vestibules have additionally been scaled at interfaces with the CCilE beundary making the enclosures more effective in reducing CCllE boundary door leakage.

Control Complex tracer gas testing was performed with the vestibule doors blocked open to assure that the test was conservative. Blocking open the vestibule doors increased conservatism in two ways. First, normal access and egress was permitted during the tracer gas leakage testing which contributed to the measured leakage. Dose calculations performed in accordance with standard methods include a factor of 10 cfm of continuous leakage to account for access and egress during an accident. This factor was added to the measured inteakage in performing CR-3 dose calculations.

Therefore, the effect of leakage dur!ng access and egress was applied twice. In a real event the vestibules would not be bhicked open. Since they function similar to an airlock, they would be effective in reducing let - r . during acc ss and egress. The second rnanner in which maintaining the vestibules open during tac test was conservative is that the existing boundary door was the only barrier to leakage during the test. In an actual event the vestibule doors would be in their normal closed f.ositions, and would be effective in reducing infiltration into the CCllE through the doors.

This feature would be particularly effective in reducing operator dose during the MilA without LOOP scenario where leakage is induced from the Turbine Building into the CCllE due to operation of the AllVS.

2. Auxiliary lluilding Filtration Perfonning doac calculations in accordance with the requirements of NUREG-0737 Item Ill.D.3.4, Control Room liabitt,bility, includes assuming source terms specified in Standard Review Plan 15.6.5. One aspect of this for a plant that does not have an "ESF atmosphere filtration system," is leakage of 1500 gallons of water contained in engineered safeguards piping outside of containment must be assumed 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months following the accident. This amounts to leakage into the Auxiliary lluilding (All) of highly contaminated water from the reactor building sumn. An ESF atmosphere f.Itration system is like the CR 3 ABVS with IIEPA and carbon illters, however it is required that an ESF system be powered from an onsite power source. Since the AllVS is not powered from an onsite power source, CR 3 control roora dose calculations include the required leakage term. This term is responsible for approximately 8 REM of the projected individual control room operator dose of 26.5 REM during the MilA 'vith LOOP scenario.

.The CR 3 ABVS has iedendant fans and filters which are operated continuously during normal operation. Therefore, the system must be maintained in good operating condition. The filters are tested and maintained in accordance with approved procedures that implement regulatory guidance 19

. .. . ~

.. KlWISION 3 l 0

on emergency filter systems. As such, filter efficiency is routinely verified by carbon Olter media testin ;. Evaluations of the FPC power grid performed for station blackout concerns have demonstrated a high degree of reliability. Studies have shown that power should be restored to the site in under eight hours following a loss of offsite power due to disturbances on the grid.

Therefore, there is a high degree of assurance that the AllVS will be available and capable of performing effectively to reduce radioactivity released from the All following an event.

Calculations show a near linear relationship between All source term and operator dose. Therefore, using a conservative AllVS Oltration efficiency of 75% would result in a dose reduction of approximately 6 REM (8 REM contribution X 0.75 = 6 REM reduction) from the 26.5 REh!

projected dose for the MilA with LOOP scenario. This reduction in dose is applicable only to the MilA with LOOP analysis. 75% efficiency for AllVS carbon filters has been previously accepted for CR 3 as an interim measure during resolution of reactor building flood level issues, and is already credited in the analysis for the MilA without LOOP .

3. th!mhianagement if a ra<liological accident were to occur which involved the release of radioactive material from the reactor or spent fuel storage area, the CR-3 Radiological Emergency Response Plan would be implemented. The plan provides for staffing the emergency response organization and establishing emergency response actions commensurate with the severity of the event. Actions required in the Emergency plan implementing Procedures include dispatching a licalth Physics Technician to the Control Complex to monitor radiological conditions, and to provide radiological and meteorological data to the Dose Assessment Coordinator. The llealth Physics Technician will perform radiological surveys within the Control Complex, including surveys for airborne radioactivity. Dose Assessment personnel use data collected from surveys to project expected personnel doses. Provisions exist in ti,e Emergency Plan Implementing Procedures for considering administration of potassium iodide (KI) to personnel based on projected dose. A projected dose of 25 REM to an individual has been established as the threshold for considering administration of Kl.

Control room dose calculations contain very conservative assumptions regarding operator presence in the control room. For example, in accordance with the Murphy Campe methodology it is assumed that the operator is present in the control room continuously for the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months , then 60% of the time for the next 3 days, and then 40% of his time for the remaining 26 days. Similarly, at nospheric conditions are assumed which channel the release of radioactivity toward the control complex at conditions which maximize the plume concentration, particularly in the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months of the scenario, llased on the conservatisms that exists throughout control room habitability dose calculations, no doses approaching GDC 19 limits are anticipated in a realistic accident. Ilowever, provisions exist in currently approved procedures to monitor actual and projected doses based on measured and observed conditions, and to control personnel exposure. Through the protective features of the control complex evaluated in the control room habitability calculations, and the established emergency response procedures, protection of the control room operators is assured.

20

..' Rt. VISION 3 l ;

i SLLMMARY / CONCLt!SIONS !

liased on the above discussion, it is concluded that, upon completion of certain actions, CREVS ,

performance and CCllE integrity is adequate to (1) protect the control room operator in toxic gas events and Dil As for which CR #3 is licensed, such that regulatory limits are not exceeded, (2) meet !

operability requirements defined in the ITS, and (3) not invalidate the assumptions and conclusions of the 11S bases. Meeting these requirements is deemed adequate and appropriate basis for plant operation in any mode with regard to control room habitability. Specific actions which must be implemented and the timing with which they must be completed to support this evaluation prior are listed below: ;

l) Complete modifications and post modification testing associated with replacing CREVS i boundary dampers and improsements to CCllE integrity. Post modification CREVS flow !

testing and e lance must verify that the Control Room elevation remains slightly positive with !

respect to the cable spread room elevation. This action must be completed prior to h10DE 4 ;

Complete test requirements associated with the Ventilation Filter Testing program defined in 2)

ITS 5.6.2.12. In addition to .satisfrine cuvrentR5 testine reautrements. charcoal testing is to .

be perfonned at conditions of 30 C and 95% Ril, consistent with LAR # 222. This action must be completed prior to h10DE 4

3) Complete surveillance requirements associated with ITS Section 3.7.12. This action must be completed prior to hiODE 4

4) Submit LAR # 222 to request licensing changes associated with CREVS, including charcoal test requirements, surveillance requirements on CCllE integrity, breach margin and ,

reconciliation of CREVS flow rate and filter dp. This is a followup action which should be ,

completed as soon as feasible, but ior which no specified completion time is given.

5) Administrative measures must be in place to ensme that CREVS is in recirculation mode prior to fuel movement until such time as updated Fila analyses finds this is not necessary At that time, this JC0 can be revised as appropriate. This action must be completed prior to any movement of fuel.

6) Administrative measures must be in place to ensure that quantities of toxic gases stored on site do not exceed the limits evaluated herein. CR 1/2 " red tag" clearances are currently in effect.

These clearances or equivalent measures are to remain in place,

7) Adn.:nH:::he n._ ::::; ::n!:h;;:e h'ap =';:nH be upda::J ta-msta:n the a:nly W::!!:d ='" full.-- T.% =:'r =.' be =p!:::d;& " \! ODE ! (Complete). ,

8) Procedures controlling allowable CCilE breach margin must be updated to reflect an breach margin limit of 22.8 in.2 in h10 DES 1-4. Breach margin limits for h10 DES 5 and 6 will remain at their previous levels fm .ne time being. This must be completed prior to hiODE 4. 1

0) A!! ::nde: Calcu!::!ca, . :d te a;=" DBA cc,::c! rea:n c;::::ar dece nnaly= n.vt k rev!:=d n' verified. Thi :S nn/ : be een:p!:::d prier te NORC appreval of:hi DR /

KA (Complete) ,

10) All vendor Calculations used to assess Toxic gas accident control room operator exposure analyses must be reviewed and verified. This action must be completed prior to hiODE 4.

21' ;

t ., .

,. REVISION 3 l REFERENCES i

i l

1. FPC CR#3, MAR 97-07 05-01 " Control Complex Emergency Ventilation" !

2. CP 148 Rev. 2 " Ventilation Filter Testing Program," FPC CR3,8/30/96 j

3. FPC SA/USQ For Sargent & Lundy Calculation Si,9929 M 009 R1 .;

i

4. NRC letter 3N0589 25 to FPC dated 5/25/1989," Crystal River Unit 3 Control Room liabitability !

t Evaluation (NUREG 0737 Item lli.D.3.4) ( TAC No. 64805)"

5. - License Amendment Request (LAR) #222 R0, " Control Room Emergency Ventilation System and. )

Control Room liabitability" !

6. FPC CR#3, MAR 77 0411,' Control Complex Ventilation Cooling Coils AilllE" 1'

I i

--)

h

. i:

.p I

-i i

i IL 22 _.

[

i l:-

!L L: - = -. _ ._._ .__2.,... . _ _ . , _ _. _.,_. ..- _ , _ , _._ _ ._._ , __

~ '

,, RtNISION 3 l TABLE A ;

Comparison of inputs to Control Room liabitability Analyses Parameter Value in Value in Comments 6/30/87 Current Submittal Analysis kcactor 71 secoads 124 seconds Revision 3 to Calculation 186-0003 (dated 7/6/93) used a two i llullding minute Ril spray delay time based on request from FPC.

Spray Since then 186 0003 has been revised several times and uses Actuation 124 seconds as a conservative Ril netuation time. This value Time is obtained by using 120 seconds for RD spray actuation plus 4 seconds for Ril pressure to go from 0 psig to 30 psig after a 1,0CA.

More realistic values for Ril Spray initiation time are found in Calculation M94 0004 Rev. 0 (dated 1/26/94), which determined the full Rll spray actuation time from initiation, to diesel start, including block loading, pump starting, headei fill time and time to reach full Dow. Calculation shows Ril spray A reaching full Dow in 81.1 seconds and D train ,

reaching full now in 86.1 seconds. This calculation modeled the spray system completely and included all the maximum expected delay times.

Reactor 1500 gpm i112 ppm in the 6/30/87 liabitability livaluation Ril spray Dow is llullding described as full now (3000 gpm), half How (1500 ppm). No Spray Flow differentiation was made between initial injection and Rate reeirculation How rates. Reviewing GP-405 Rev. 31, Ril Spray System, which was in effect in 1987, has recirculation spray Dow set at 1150 ppm to 1250 ppm.

Calculation 190 0022 Rev. O,3/12/91, detennined that with Ril spray controller set at 1500 gpm (during initial injection), the actual Ril spray now could be as low as 1397 ppm conHdering instrumentation error. in recirculation with Rin spray controller set at 1200 gpm, the spray Dow could be as low as i112 gpm.

Calculation 186-0002 Rev. 5,1/16/96, determined containment spray removal constants using the new instrument error corrected now values of 1397 gpm (injection phase) and 1112 gpm (recirculation phase). Spray constants associated nith the lower value of1112 epm ace ystd in reviseddose calculations.

The instrwnent loon uncertainties for scrav Row indication and control were being reviewed concurrent with verforming l]1e reviseddose ca!culations As a contineencv. the revised dose calculatlan Icoked aLLLCoulainment sprav Row rate of 23

1 e, n. o

,, RINISION 3 l T "^'l 2000 epm andfunndIhar tuvas essentiallylit.iututa.Ulx lill.gutu case. The calculation concludes shalcutitalninent j 8 ,yugrate,f1000 enm can be tolerated.

, .l 5 I Balliet to Widell hr ser NOE97-2311 did 11/11/V7. Shows tliat uhen sprav is beine utppljedfrom tileRil Surnp. Ihe

l j actyal flowinav be 121 epus below the indicatedRow of1200 i t_. wtL1huL1htfu1rcatsulucituttbe lDZ922nL l Reactor 490,182 gal 343,347 gal The habitability submittal assumes the liquid sump volume lluilding (65,532 113) (45,902 ft3) was as 490,182 gallons (7.48 gallfl'or 65,532.353 ft') This Sump Volume volume was Calculation 186-0003 Rev. I, S/2/91, referenced GCl calculation DC 5515-0841.h1E, Rev. 0, dated 3/26/90 that calculated new Ril sump volumes based on eliminating Naoll tanks and switching to TSP baskets (htAR 88 05 0101), New volumes were based on cubic feet and were referenced to 130' F. New volume was determined to bc 500,718.7 gal or 66.941 fl', Calculation 186-0003 Rev. 6, 3/30/95, then switched to 45,90211' or 343,347 gallons. This figure was the output from Calculation h195 0007. An important design reference for Calculation h195 0007 was Calculation h195 0005, hiinimum llWSr Level to Prevent Vortexing Rev. O. FOP-8 swaps from 11WST to Ril sump starting at 15'. An instrument error of 1.2' was used in llWST level calculations. LOP 8 requires swapping over when llWST is less than 15' and has to be complete by 7' to prevent ilWST vortexing. (5.5' from Calculation h195 0005)

These low level considerations reduced the amount ofIlWST water going into the Ril sump significantly.

Reactor 8.5 7 7.6 The 1987 habitability evaluation repar1 contained spray fluilding solution pil Table 4.1 1, Results of Drawdown Analysis for a Sump hiinimum of 6,0 wt % Sodium liydroxide in the Storage Additive / pil Tank. This table listed five Ril spray cases with initial spray pil and time post l.OCA for spray pil to reach 8.5. The iodine removal constants were calculated using SRP 6.5.2 Rev. I.

IIAW 2044," Elimination of Containment Spray Additive",

was a 11 & W study to determine how to convert to from Naoll storage tank to TSP. With TSP, the initial Ril spray ,

pil will be around 4 5 because that is the pil of the 11WST water. Afler the water mixes with the TSP in the Rll llooded level and Ril spray is swapped to recirculation, then the Rll spray water pil increases to the range of 7 7.6. FPC installed the TSP baskets by htAR 88-05-0101.

GCl revised Calculation 186 0002, Containment Spray REhioval Constants (lodine REhtoval) to Rev 2 and calculated the CR 3 specific iodine removal constants using l 24 l

,, ~ REVISION 3 l

~

SRP 6 5.2 Rev. 2 methodology in 1991.186 0002 Rev. 5, 1/16/96, recalculated the total containment spray iodine removal constants for 1397 (1500 ppm with largest maximum negative error) and 1112 gpm (1200 ppm with largest maximum negative error). These constants r*

considered to reDect current plant design and config on, and are used in icvised Jose cales.

MilA Source liased on liased on The higher power rating was incorporated based on recent Terms llD 14844 'llD 14844 licensing activities regarding a CR #3 power uprate. This and a power and a power action has not been completed, but the post accident source leselof 2594 level of 2619 term associated with the higher power rating has been MWth M Wth incorporated into dose calculations. Since the source term is determined based on a per megawatt basis per TlD 14844, the use of the larger MWth rating sesults in a source term slightly higher than that which would be predicted with the lower power rating. This is cleariy a conservatism (not a USQ) given that the plant is still licensed to the lower value.

Auxiliary 0% ellicient 0% ellicient fly letter dated September 13, 1989 (310989 01), FPC lluilding in LOOP submitted a revised licensing basis for the CR 3 Loss of Filtration events, ZL*f Coolant Accident (LOCA) and the Makeup Sy. stem letdown dBcict2f in 1.ine Failure Accident (Ll_FA) offsite radiological events for consequences to eliminate the credit for the Auxiliary lluilding w hich power Ventilation System (AllV) due to lack of safety grade power.

is assumed to FPC re-evaluated the offsite radiological consequences of a be LOCA using the same methwlology for fission product release maintained. as that used to evaluate the CR 3 control room habitability in its June 30,1987 habitability report (3F068716), i.e. no credit for Auxiliary lluilding Olters.

During calculational veri 0 cation efforts relative to the Reactor iluilding (Ril) flooding issue, FPC identified that the control room habitability dose is adversely efTected by the change in Rll flood volume. This affect was documented in FPC letter to the NRC dated June 4,1990 (3F0690-04). The habitability repon postulates a gross failure of a passive component which causes a 50 gpm leak for 30 minutes at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . It was considered that since CR 3 does have a filtration system associated with the areas containing the Engineered Safeguards (ESP) systems and passive failures such as that postulated to cause the 50 gpm leak have not been considered as part of the CR 3 licensing basis, the gross failure of a passise component would not be postulated in the CR 3 control room habitability dose analyses. Discussion with the NRC regarding the Ril flooding issue and the adverse effect on the control habitability dose resulted in the FPC analyses

< including the postulated gross failure of a passive component causing a 50 gpm leak for 30 minutes at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months with the 25 l

- e ,

,. KliVISION 3 l o

AllV system in service with 75% cilicient charcoal tillers for lodine removal (3f0690-06 and 31^069013). The NRC documented acceptance of this in its letter to FPC dated June 21,1990 (3N069015) as an interim measure until the Ril ikxxling issue was permanently resolved. Subsequent to replacement of Sodium liydroxide spray additive solution with TSP baskets, calculations wcre performed w hich demonstrated acceptable dose sequences without the AllVS Olters and credit for their operation w as discominued.

In resised dose analyses, the AllVS filters are assumed to be operating for any event w hich assumes that the Auxiliary fluilding is at a high negative pressure. Under these conditions, the AllVS supply fans are assemed to be tripped and the exhaust fans discharging through the charcoal filtration system and out the stack. Differential pressures across the CCllE on the order of 0.20" wg would be expected, w hich would result in leakages considerably higher than that associated with MilA/ LOOP. Ilowever, given that the ABVS is assumed to be operating throughout the event, per SRP 15.6.5 no 50 gpm leak would be postulated at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months into the event, and " normal" ECCS leakage would be subject to filtration. Thus, this scenario is bounded by the MilA/ LOOP with respect to Control Room liabitability, in the event of a MilA w / LOOP, no credit is taken for AllVS filtration for the duration of the 30 day accident period.

CRliVS i low 43,500 cfm 37,800 clm 43,500 cim is the original design flow rate of the CREVS, Rate and is the value used to determine IPF in the 6/30/87 habitability report. Dose consequences were later evaluated (recire mode) at 43,500 10% (39,150 efm) corresponding to the allowable range of operation found in li S Section 5.6.2.12 relative to the filter test program. During a system readiness review it u as recognized that previous modifications had been made which reduced system flow rate without adequately assessing the effect on CREVS Revised dose analyses incorporate a value of 37,800 cfm, based on consideration of current system capabilities under dirty fiher conditions.

CRliVS 95 % 95 % t'iltration etliciency hasn't changed, but filter testing has Filtration been upgraded to utilire n ore challenging criteria. Previous Efliciency charcoal testing was performed at 80 C at 30% Ril, test program has been revised to evaluate charcoal at 30 C and 90% Ril. Criteria for inplace filter testing is penetration and system bypass of <0.05%.

26

e, ., .

.l REVISION 3 l o

CCilli/ CR 355,31 I it' / 364,922 f t' / Original solumes were based on an internal memo from Volume 85,573 ft' 88,000 fl' Gilbert. CCllE solume was estimated by calculating the volume of the entire ens clope, then subtracting 10% for internal walls and contents. Updated volumes wcre calculated based or, a room by room survey performed by S&L for use in Control Room heat up evaluations.

CRIIVS/ As described As modified 'ihe following modifications are being implemented to CCllE in the by R 12 address the concerns associated with Restart issue R 12.

Mmlincations habi' ability malineations 1:igures C 1 and C 2 provide a schematic of the pre and report post modi 0 cation configurations. Note that except as otherwisc stated, pairs of dampers replacing a single damper receive the same control signals and act in unison, such that system logic is not changed. <

e Damper AllD 99, which brings supply air to the Ventilation Ik uipment l Room is being removed and a permanent blank installed. Ncw supply and return registers shall be installed in the ductwork (164' elevation) which will now serve as the ventilation for this area. 'this will climinate AllD-99 as a potential source ofinleakage.

e thisting damper AllD 12, located in the supply duct to the CA, is to be removed and replaced with two new bubble tight dampers, AllD 12 and AllD 12D.

Existing damper AllD-2, located in the exhaust duct to the outside, is being locked open and abandoned in place. Two new bubble tight dampers, AllD-2C and AllD 2E, shall be installed in series in the exhaust path. AllD 2C will be normally closed.

The position of recirculation air damper AllD-3 will be established during the process of balancing the system for the normal operating mode.

Dampers AllD 1 and AllD lD, located in the air intake duct, are being disabled and abandoned in place. Two new bubble tight dampers, AllD IC and AllD-lE, are being installed in series on the inlet duct. Dampers AllD-lC, AllD-2C and AllD 3 will retain positioners which provide a manual override feature. This feature allows operators to position these dampers to modulate the outside airflow as required for purging smoke or other contaminants from the CCilE.

Mecht.nical Equipment Room Ventilation Air llandling Fans, AliF-21 A/B and associated dampers AllD-24, AllD-25, AllD-26, and AllD 27 are being spared in place and the associated CCllE penetration sealed. This portion of the system originally exhausted air from the Mechanical Equipment Roorn, Elevator Equipment Room, lavatory, kitchen 27

Os ,

,o RiiVISION 3 l and toilet, t his eliminates another potential source ofinleakage into the CCllE.

New supply and return registers are being insta: led in the ductwork in the Mechanical Equipment Room.

This will provide sentilation to this portion of the CCilE during both nonnal and recirculation modes.

A skid mounted air handling unit consisting of a fan and a charcoal filtration unit will be installed to ventilate the Elevator Equipment Room, lavatory, Litchen and toilet. This system is non safety and non seismic and will vent approximately 1,000 cfm by way of a field connection to a non safety related (NSR) duct.

Small bore drain pipes penetrating the CCilE are being litted with loop seals to prevent inleakage though the lines. These will be added to a queued work request in the work controls system which maintains CCllE drain line loop seals.

Vestibules have been installed over all CCllE boundary doors, and have been sealed to provide maximum leaktightness. These vestibules provide a means to test individual CCllE boundary dsor leaktightness, as well as zeducing inleakage associated with CCilE access / egress.

In addition to the abos e modifications, an extensive effort was undertaken to survey CCllE penetrations and seal as required to minimize inleakage. As a result of this work,it is concluded that conduit penetrations do not pose a significant liability to CCilE integrity. Penetrations associated with electrical cable banks were inspected and scaled to the extent feasible with existing procedures and materials, but some leakage paths remain through the interstitial spaces between individual cables. Additional work is being planned to improve the sealing of penetrations with the most significant leakage.

CCilh Estimated on Measured by See detailed discussion pertaining to inleakage elsew here in inleakage the basis of tracer gas the "JUSTil lCATION l'OR CONTINUED OPERATION" summation testing and section, this JCO.

leakage past analytically CCllE corrected to boundary predict elements per intenkage SRp6A under postulated post accident conditions 28

0,

, e

/* l RINISION 3 l r

Dose ICRP2 ICRP30 I he NRC Safety Evaluation of I PC's control room Conversion habitability is based on the " Control Room liabitability Factors Evaluation Report" submitted to the NRC on June 30,1987.

At that time,ICRP 2 methodology was used for internal dose calculations. Revised methods for calculating organ dose and relating organ dose to whole body dose were published in ICRP 30, and endorsed for use in this country by the Environmental Protection Agency (EPA)in Federal Guidance Report #11. For the radionuclides of concern, use ofICRP 30 / Federal Guidance Report #11 dose comersion factors results in the accident thyroid dose to be ~30% lower than previously calculated. CR 3 Improved Technical Speci0 cations (ITS) include specinc activity limits for primary and secondary coolant, which is measured and reported as DOSE EQUIVALENT l 131. The ITS dennition of DOSE EQUIVALENT l 131 specines that the thyroid dose conversion factors used for this calculation shall be those from ICRP 30.

Software Accident Analysis Sof tware (POS I Dil A)

Computer program POSTDilA is Sargent & Lundy proprietary software which performs radiological dose calculations and related analyses for the LOCA in a PWR or a llWR. POSTDilA was originally developed to calculate PWR con:rol room (CR) and offsite doses in accordance with requirements and recommendations of Regulatory Guide (RG) 1.4, Standard Review Plan (SRP) Section 6.4, and SRP 6.5.2., and was revised and revalidated most recently in 1994.

POSTDilA is constructed to allow the user to select the time steps and to control variable parameters for each time step.

The variables include containment spray iodine nmoval rates: pest accident source release rates (iodine and noble gases) and any iodine Oltration: x/Q changes: CR parameters (makeup, inleakage, iodine removal, breathing rates, and occupancy factors); plus the fractions of elemental, particulate, and organic iodine released to the environment.

The Orst and the following time steps can be used to vary most of the variables, and if needed, the Grst time step can be used to model a delayed release. This degree of user control allows other types of accidents to be analyzed.

Similar to POSTD11A, Computer program AXIDENT is NUS SCIENTECll proprietary software which performs radiological dose calculations and related analyses.

29

ML20203F803

Contents

Text

Subject:

Reference:

Dear Sir:

Navigation menu

ML20203F803

Text

Subject:

Reference:

Dear Sir:

Navigation menu

Search