Text

Response to Request for Additional Information Regarding Generic Letter 2004-02
	ML102160372
Person / Time
Site:	Indian Point
Issue date:	07/27/2010
From:	Joseph E Pollock; Entergy Nuclear Operations
To:	; Document Control Desk, Office of Nuclear Reactor Regulation
References
	NL-10-074
	Download: ML102160372 (38)
	v • d • e

Indian Point Energy Center 450 Broadway, GSB P.O. Box 249

-Entergy Buchanan, N.Y. 10511-0249 Tel (914) 734-6700 J.E. Pollock Site Vice President Administration NL-10-074 July 27, 2010 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, DC 20555-0001

SUBJECT:

Response to Request for Additional Information Regarding Generic Letter 2004-02 Indian Point Units 2 and 3 Docket Nos. 50-247 and 50-286 License Nos. DPR-26 and DPR-64

REFERENCES:

1) NRC letter to Indian Point Vice President of Operations, 4/29/10, "Indian Point Nuclear Generating Unit Nos. 2 and 3 - Request for Additional Information Regarding Generic Letter 2004-02 (TAC Nos. MC4689 and MC4690)"

2) Entergy letter NL-09-138, 11/19/09, "Updated Supplemental Response to NRC Generic Letter 2004-02; "Potential Impact Of Debris Blockage On Emergency Recirculation During Design Basis Accidents At Pressurized-Water Reactors""

3) Summary of June 9, 2010, Conference Call with Entergy on their Proposed Response to a Request for Additional Information on Generic Letter 2004-02 (TAC Nos. MC4689 and MC4690), dated June 18, 2010

Dear Sir or Madam:

On April 29, 2010, the NRC issued a request for additional information (RAI) (Reference 1) regarding Entergy Nuclear Operations, Inc (Entergy) updated supplemental response (Reference 2) to Generic Letter (GL) 2004-02. The NRC staff was previously provided draft RAI responses to facilitate discussion during the June 9, 2010, Category 1 public meeting. As discussed during the public meeting (Reference 3), Entergy's formal RAI responses are provided in the Attachment to this letter.

NL-10-074 Page 2 of 2 There are no new commitments being made in this submittal.

Should you have any questions or require additional information, please contact Mr. R. Walpole, Manager, Licensing at (914) 734-6710.

I declare under penalty of perjury that the foregoing is true and correct. Executed on 7- * -/ 0 Sincerely, JEP/rmw

Attachment:

1. Response to Request for Additional Information Regarding Generic Letter 2004-02 cc: Mr. John P. Boska, Senior Project Manager, NRC NRR DORL Mr. M. Dapas, Acting Regional Administrator, NRC Region 1 NRC Resident Inspector, IP2 NRC Resident Inspector, IP3 Mr. Robert Callender, Vice President, NYSERDA Mr. Paul Eddy, New York State Dept. of Public Service

ATTACHMENT 1 TO NL-10-074 Response to Request for Additional Information Regarding Generic Letter 2004-02 ENTERGY NUCLEAR OPERATIONS, INC.

INDIAN POINT NUCLEAR GENERATING UNIT NOs. 2 and 3 DOCKET NO. 50-247 and 50-286

NL-10-074 Attachment 1 Page 1 of 35 Response to Request for Additional Information Regarding Generic Letter 2004-02 The U.S. Nuclear Regulatory Commission (NRC) staff has requested responses to the following questions, which relate to the November 19, 2009, supplemental submittal, in order to continue its review of the Entergy's response to GL 2004-02:

NRC Reauest 1 Head Loss and Vortexing In RAI 12 of the staff's letter dated November 19, 2008 (ADAMS Accession No. ML083230054),

the staff requested information thatprovides traceabilitybetween the test results presentedas final values in the supplemental response and the raw test data. The RAI requested that the licensee provide the methodology for deriving the final values and the assumptions used in the evaluation. In its response letter dated November 19, 2009, the licensee provided descriptions of the tests that linked each break case to one or more tests that were used to evaluate the strainerperformance for each particularscenario. This answered the staff's question regarding the ability to determine how each break was covered by the test program. However, the staff could not determine how the test cases were extrapolatedto the plant conditionslisted in Tables 3f. 10-13 and 14. Pleaseprovide the methodology (the equations used) and assumptions used to extrapolate the test cases to each plant case listed in Tables 3f. 10-13 and 14.

Entergy Response Page # in parenthesis refer to the November 19, 2009 Entergy response.

The Entergy updated supplemental response shows total head losses for a variety of Emergency Core Cooling System (ECCS) alignments (Tables 3f.10-13 for IP-2 (pg 119) and 3f.10-14 for IP-3 (pg 120)). As a point of clarification, only certain of these alignments are permitted and supported by the Emergency Operating Procedures (EOPs). Due to the requirement for both units to initiate Hot-Leg recirculation by 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months , the only applicable alignments after that time are those with an analyzed maximum ECCS flow of 1350 gpm. In contrast to the aforementioned Tables, Tables 3g.16-7 (pg 156) and 3g.16-8 (pg 157) provide the minimum margin head loss results, and only for alignments permitted and supported by the EOPs.

The NRC staff requested information on how the results in Total Head Loss Tables 3f.10-13 and 3f.10-14 provided in response to question 3f.10 were obtained. The following explains the adjustments made using the 1 st case listed (IP-3 / IR SUMP LBLOCA or RC LBLOCA / 5263 GPM) in table 3f.10-14 (page 120) as a representative example. All the cases would follow the same methodology.

All the Alion test data is normalized from the specific test conditions to: 70°F and/or 204.70 F, and either 400 GPM or 155 GPM, depending on which is applicable for the plant sump flow rate being considered. (The lower 155 GPM flow is typically applied to analyses for the low flow cases of VC Sump LBLOCA & RC-LBLOCA case after 24-hours of Recirculation Sump operation.) Normalizing places all the data on a relative basis so the various test head losses can be compared to select the bounding cases. The resulting head losses of this 1 st normalization are provided in the last three columns of Tables 3f.10-1, -3, -4, -5, -7, -9, and -11.

NL-10-074 Attachment 1 Page 2 of 35 After matching up the applicable debris loads (C-3, E-3, & F-3), from the relevant available test cases data, the highest normalized head loss values are selected from each table for:

Conventional, Conventional & Chemical, and 30 Day Extrapolated Conventional & Chemical debris loads. These are depicted as the highlighted values in the last 3 columns of Table 3f.10-7 (page 116), specifically: 3.07 ft, 10.91 ft, and 12.32 ft. These particular values will not be used after this point since they were only to allow a relative comparison between all the test cases.

Actual test data is taken from the IPEC scaled prototypical strainer array testing performed by Alion in accordance with the 2008 protocol. The data selected from the applicable tests for the above case for "Conventional" (Test F), and "Conventional & Chemical" (Test E-3) debris loads were: 2.42 ft, and 8.43 ft, respectively. Testing instrument inaccuracy is already conservatively accounted for (added or subtracted as applicable) in these reported values. The "30 Day Extrapolated Conventional & Chemical" value of 9.52 ft is based on the curve fit formula as applied to the test data and further explained in 3o.2.17.i (pages 203 & 204). For the example case being discussed here: the test data, the extrapolation formula, and resulting value of 9.52 ft is presented in Figure 3o.2.17-17 from Test E - 3 rd Chemical Addition (page 221).

In making adjustments from raw test data to the normalized values, the laminar/turbulent fraction of the flow is considered, and variations in the test flow rate are adjusted for. For further discussion on the adjustments, see the response to question 3f.13 (page 121-122). For the values presented in the final tables, the raw test data is adjusted to the specific plant conditions using the equations below. The various equations were used to ensure conservative corrections:

.Decreased flow and temperature case:

AH!2 -- A/l 0.97Q.A Q 2J1U 2 Increased flow and decreased temperature case AH2 2

fAH 2 [L frac 2 0.97xQ1 T

(O9) 22 (0.97 Q1 2 J Decreased flow and increased temperature case A6H Q2 2 =AH10.97xQ1 FL 2+ T A2 Lifac fra, A Increased flow and increased temperature case l Lfra 'L Q2 +TAP2 Q2 1

,H2 O.97xQ1 fr.A (o.97) 2 Q2 Note: The 0.97 value in the denominator is a reduction in screen area to compensate for an assumed void fraction of 3%.

Additionally, the viscosity of water at various temperatures was calculated using the following equation:

NL-10-074 Attachment 1 Page 3 of 35 10 85

/* = 0.074T-h Where:

=U Absolute (dynamic) viscosity (lb-m/ft/s)

T = Temperature ('F)

For this case, the calculated maximized plant accident flow rate of 5263 gpm through the Recirculation Sump Strainer (-3121 sq-ft surface area) compares to a scaled flow rate of -192 gpm through the Test Array Strainer (-114 sq-ft surface area) based on a ratio of the full and scale strainer areas.

Note: CSHL = Clean Strainer Head Loss and is given as 0.325 feet in the sections below.

For the Conventional Debris Load only (Test F data)

The measured head loss (2.42 ft), at measured flow rate (391.67 gpm), and a temperature of 83.67°F are used to arrive at the corrected head Loss at 70°F (1.49 ft). The equation for decreasing temperature and flow rate is used (note that the equation is based on an assumption of 100% laminar flow for conservatism):

192gpm 6.66 10-4 tibm HL at 7 0 °F = 2.42 ft ft = 1.49ft 0.97(391.97gpm{5.47.10-4f s)Ibm Then, the total head loss is determined by summing:

corrected head loss at 70F + RMI head loss + CSHL = (1.49) + (0.0) + (0.325) = 1.815 1.82 ft-water The test parameters outlined above, the laminar & turbulent fractions (0.37% & 0.63%), and the equation for increasing temperature and decreasing flow rate was used to obtain a corrected head loss at 204.7 0 F (0.91 ft).

2.03 m 59.99 bft3 HLat 204.7 0 F= 2.42ft 192+gpm (037) ftIbmgpm2(037)bft/sflb S +0.63 =0.91ft 0.97(391.97 gpm) 5.47.10-4 Ibm 6 L ft* s Densities where obtained from ASME steam tables (1967).

Then, the total head loss is determined by summing:

corrected head loss at 204.7F + RMI head loss + CSHL = (0.91) + (0.0) + (0.325) =1.235

=> 1.24 ft-water For the Conventional & Chemical Debris Load (Test E-3 data)

The measured head loss (8.43 ft), at measured flow rate (404.36 gpm), and a temperature of 89.92°F are used to arrive at the corrected head Loss at 70°F (5.46 ft). The equation for

NL-10-074 Attachment 1 Page 4 of 35 decreasing temperature and flow rate is used (note that the equation is based on an assumption of 100% laminar flow for conservatism).

Then, the total head loss is determined by summing:

corrected head loss at 70F + RMI head loss + CSHL = (5.46) + (0.0) + (0.325) = 5.785 =

5.79 ft-water For the 30 Day Extrapolated Conventional & Chemical Debris Load (Test E-3 data extrapolated)

The measured head loss (8.43 ft) was extrapolated (9.52 ft), and then using measured flow rate (404,36 gpm), and a temperature of 89.92°F are used to arrive at the corrected head Loss at 70°F (6.16 ft). The equation for decreasing temperature and flow rate is used (note that the equation is based on an assumption of 100% laminar flow for conservatism).

Then, the total head loss is determined by summing:

corrected head loss at 70F + RMI head loss + CSHL = (6.16) + (0.0) + (0.325) = 6.485 =

6.49 ft-water NRC Request 2 Head Loss and Vortexing Pleaseprovide the results of an evaluation of the potential effect of voids (possiblyresulting from deaerationof coolant)on the pumps' net positive suction head required(NPSHR) values as discussedin Regulatory Guide 1.82, Appendix A, and adjust the NPSHR values as described in that guidance. Please explain how the results of the evaluation affect the NPSH margin calculation.

Enterpy Response In Entergy's updated supplemental response to GL 2004-02, the void fraction values were provided in response to issue 3f.8 (page 104). The maximum void fractions downstream of the strainers were reported as:

0.31% and 0.67% for the IR and VC sumps, respectively for IP-2

0.41% and 0.60% for the IR and VC sumps, respectively for IP-3 These maximum values are reached during recirculation after extended periods of time (approximately 115 days), well beyond the GL 2004-02 mission time of 30 days. The void fraction varies over the 30 day time period and is often zero, or significantly below the maximum values quoted for long periods of time.

For conservatism, the above void fractions were calculated:

from the top surface of the top screens in the assembly.

The lower screens and an increasing water level above them would result in lower calculated void fractions; therefore, the true overall void fractions would be less than those reported above.

NL-10-074 Attachment 1 Page 5 of 35 assuming a constant minimum small break LOCA water level throughout the transient.

The small break water level is bounding and as safety injection and containment spray continue to draw down the RWST, containment water level will increase above this minimum level.

Re-evaluation of Void Fractions Regulatory Guide 1.82, Appendix A, provides guidance on an adjustment to NPSHR based on

% void fraction at inlet to the Emergency Core Cooling System pumps that perform the recirculation function. The void fractions reported in the supplemental response are too conservative for application at pump suction. Therefore, in order to obtain more realistic void fraction values for input into the NPSHR adjustment, credit was taken for:

" the smaller void fractions generated by strainers in the lower elevations of the sump

the compression of voids, in accordance with the ideal gas law, between sump outlet and pump inlet In determining the gas absorption into the fluid before it passes through the strainer, the void fraction calculation conservatively considers the pressure and temperature conditions of the sump fluid and the atmosphere above it. The void fractions just downstream of the strainer elements have been calculated for each level of the strainer assemblies and then averaged to obtain the overall value of the assembly. Lower levels of the strainers have a greater static head of water above them, thus reducing the overall void fraction.

The IP-2 and IP-3 IR and VC Sump strainers all consist of only horizontally mounted Top-Hat cylinder strainer elements connected to plenum boxes. The IP-2 and IP-3 IR strainer

-assemblies have nine levels of Top-Hats in a sump pit.

The IP-2 VC Sump strainers are horizontally mounted on one side of a plenum box. The strainer extension (outside of the sump pit) has a plenum leading to and then connecting to the plenum in the VC Sump pit. The strainers in the VC sump pit are stacked in levels with one removed in the top level due to the extension connection. The extension is about 1-3/4 feet above the sump pit plenum connection point and about 15 feet away.

The IP-3 VC Sump strainers are horizontally mounted and connected on both sides of a plenum box. The strainers in the VC sump pit are stacked in levels with the lower two rows on one side shortened for interferences and four removed for the connection to the suction line. There is no extension for IP-3 outside the sump pit.

In calculating the void fraction, the water level above the strainer appropriate for the time and applicable event was employed. In some cases, instead of applying the maximum structural limit of the strainer, actual strainer head loss values were used.

Any void bubbles that may develop as flow passes through the individual Top-Hat strainers would flow into the plenum assembly of the strainer. There they will have an opportunity to rise in the plenum and exit the strainer assembly through multiple high point vent holes, however, no credit is taken for release through high point vents. If the void bubbles continue downward towards the pump, there will be additional pressure head that will reduce the void size by compression.

NL-10-074 Attachment 1 Page 6 of 35 The re-evaluated void fractions are provided in Table 1. These are the maximum void fraction values. At other times during the 30 day mission time the values are less, and are sometimes zero.

Table 1 Void Fractions Void Fractions %

Strainer Pump Unit Sump As Reported(1 ) Assembly t 2 2

Average( ) suction IP-2 IR 0.31 0.0 0.0 VC 0.67 0.098 0.064 IP-3 IR 0.41 0.087 0.084 VC 0.60 0.134 0.114

1. Maximum values attained after 115 days of recirculation

2. Maximum values during the 30 day mission time at the specified location The void fractions, after compression due to available head at the suction of the pumps, are minimal and will have a negligible impact on pump performance as demonstrated below.

Regulatory Guide 1.82 Evaluation Applying the guidance from Regulatory Guide 1.82, Appendix A, the following adjusted NPSHR values were determined.

Table 2 RG 1.82 Evaluation RG 1.82 Evaluation As RG 1.82 Unit Sump Reported As Reported RG 1.82 NPSH NPSH NPSHR ft NPSHR ft Margin(1 ) ft Margin ft IP-2 IR_(2) - - - -

VC 2.51 10 1.032 10.32 2.19 IP-3 IR 0.01 9.25 1.042 9.64 -0.38 VC 0.83 10 1.057 10.57 0.26

1. Minimum NPSH margin cases

2. Not applicable, ,8=1 due to a zero void fraction.

Table 2 shows that there is adequate NPSH margin to accommodate the adjusted NPSHR values for the IP-2 IR and VC sumps and the IP-3 VC sump. The IP-3 IR sump calculated

NL-10-074 Attachment 1 Page 7 of 35 NPSH margin is -0.38 feet. However, the IP-3 IR pump is operable based on the following considerations.

IP-3 IR Pump Performance Capabilities In support of IR pump performance capabilities the vendor and manufacturer of the IR pumps (Flowserve) reported that these pumps are low energy pumps, (i.e. impellor tip speed less than 60 ft/sec). Flowserve concluded that pump reliability will not be affected by resulting cavitation in the short term (limited to 7 days) while operating under deficient available NPSH conditions for flows greater than 3000 gpm due to the self limiting capability of the pump. For flows less than 3000 gpm, a 1 foot deficiency in NPSHA could be sustained for long-term operation without damage.

For flows greater than 3000 gpm the pump reliability is determined by the predicted reduction in developed head, not by a specific NPSH deficiency limitation. The vendor supports a reduction in developed head of up to 20%, which is ultimately controlled by the reduced NPSH that is available. A 1 foot deficiency in NPSHA would not result in a reduction greater than 20%

developed head, therefore, short and long term vendor restrictions are bounded.

The 0.38 ft NPSH deficit, as a result of void fraction considerations, occurs prior to switchover at a pump flow rate 4149 gpm. This flow rate would occur for no more than 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months , after which the flow is reduced via termination of recirculation containment spray, and then to switchover to hot leg recirculation at 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months after the break. The 0.38 ft deficit is momentary and will be less than this value at all other times due to changing containment conditions. The 0.38 ft deficit is well below the 1 foot deficit described above for short and long periods of pump operation.

While deaeration is admittedly a somewhat different phenomenon, cavitation is assumed to be a more severe condition to operate a pump under. Therefore, any intermittent periods of time where a small void fraction potentially exists do not present a challenge to pump performance.

Additionally, the IR pump is a vertical pump with the suction on the bottom, discharge on the top. This configuration for the IR pump does not support air collection in the pumps.

Regulatory Guidance In addition to the guidance provided in RG 1.82 additional regulatory guidance is provided as follows.

NUREG-0897 NUREG-0897, Revision 1 describes typical pumps configurations for PWRs in section 2.1 as single stage centrifugal of low specific speed generally rated at 3000 GPM and heads of 300 feet, and requires about 20 feet of NPSH. The IR pump has a design point of 3000 GPM, a corresponding head of about 350 feet, and require less than 20 feet of NPSH at the design point. The pumps are constructed of stainless steel, and although the IR pump is a 3-stage vertical pump, it is a low specific speed design. The IR pumps operate under 1200 RPM.

Therefore, the following NUREG-0897 statements apply to IPEC's pumps:

"Pump impeller materials are generally highly resistant to erosion, corrosion, and cavitation damage."

NL-10-074 Attachment 1 Page 8 of 35

"...pumping performance is only slightly degraded when air ingestion is less than 2%.

This value would be a conservative estimate for acceptable performance..."

NUREG-0897 supports the reliability of these pumps under conditions significantly worse than would be postulated with the minimal void fractions reported above.

Generic Letter 2004-02 The NRC Safety Evaluation related to NRC GL 2004-02 (Reference 1, below), in Section 6.4.6 states the following "... for Region II analyses, the GR states that limited operation without an NPSH margin is acceptable if it can be shown that the pumps reasonably be expected to survive during the time period of inadequate available NPSH. The suggested technical justification for this statement would include vendor information in the form of test data or engineering judgment derived from test and/or operational events". Although IPEC does not currently apply the less restrictive Region II allowances, the methodology is relevant. Region II includes LBLOCA events and therefore encompasses the IP-3 LBLOCA IR pump limiting NPSH case evaluated.

The Region II methodology supports the use of a vendor assessment of reliability of the IR pumps under inadequate available NPSH conditions.

Generic Letter 2008-01 Generic Letter 2008 "Managing Gas Accumulation in Emergency Core Cooling, Decay Heat Removal, and Containment Spray Systems", requires that the impact of gas voids on

.ECCS recirculation pump performance be evaluated. In Revision 2 to NRC staff criteria for gas

movement in suction lines and pump response to gas (draft) (ML090900136) up to a 1% void would be well within the capability of pumps to accept and pass without detriment to the pump.

There is no upstream piping for the IR pumps as they are located adjacent to the strainer plenum. The sump and strainer are a low point and self vent as the Containment floods up from the break flow. Therefore, there can be no existing gas voids present prior to the accident. The calculated void, for the IP-3 IR pump due to deaeration (0.084%) is well below the staff accepted 1% void fraction for not jeopardizing pump operability. In addition, the GL 2008-01 criteria document referred to above discusses the RG 1.82 NPSHR correlation and refers to the correlation as including substantial conservatism.

Conclusion As reported in the Entergy supplemental response of November 1 9 th, 2009, void fractions were conservatively determined immediately downstream of the top strainer and no credit for strainers at lower elevations, nor void compression due to the elevation head between this point and the suction point of the IR or RHR pumps was taken.

A new evaluation shows that the additional static head, minus any applicable frictional losses, reduce any voids to a point where they would be expected to travel through the pump without affect. There is adequate NPSH margin to accommodate the adjusted NPSHR values for the IP-2 IR and VC sumps and the IP-3 VC sump. The adjusted NPSHR value for the limiting IP-3 IR sump results in a negative NPSH margin.

NL-10-074 Attachment 1 Page 9 of 35 The IP-3 IR sump calculated void fraction is considered to be negligible at 0.084% and the RG 1.82 NPSHR adjustments are considered to be overly conservative for void fractions of this magnitude. Nevertheless, even with the RG 1.82 NPSHR adjustments made, the pump vendor supports operation of the IR pumps under these conditions. In addition to RG 1.82 other regulatory guidance is available. NUREG 0897, the NRC SER on GL 2004-02 (Region II analysis), and NRC guidance on GL 2008-01, all support pump operation with limited void fractions that envelope the predicted plant void fractions evaluated in this response.

In summary adequate NPSH margin is available to accommodate the adjusted NPSHR values for the IP-2 IR and VC sumps and the IP-3 VC sump. The IP-3 IR sump can also accommodate the adjusted NPSHR values when credit is taken for pump performance capabilities under the predicted gas void conditions.

It is important to note that IPEC currently does not credit any Containment accident over pressure in performing NPSH calculations. This very pressure, which is considered in the void creation, represents margin that is not credited in void mitigation, but by necessity would be present during the postulated event.

.Reference

1. Safety Evaluation by the Office of Nuclear Reactor Regulation Related to NRC Generic Letter 2004-02, Nuclear Energy Institute Guidance Report (Proposed Document Number NEI 04-07), "Pressurized Water Reactor Sump Performance Evaluation Methodology,"

Issued December 6, 2004 NRC Bulletin 2003-01, "Potential Impact of Debris Blockage on Emergency Sump Recirculation at Pressurized Water Reactors," dated June 9, 2003 NRC Request 3 Coatings In RAI 20 the staff asked the licensee to provide the rationalefor using a 4.28 diameter (D) zone of influence (ZOI) for inorganiczinc (IOZ). In attachment 3 of the response, the licensee response noted the 4.28D ZOI has been evaluated, design verified, and the data (new data presented in response) conservatively applied. However, in Section 3h.5, the licensee response noted that due to the applied thicknesses and densities of the various Indian Point coating systems, the current approachof 100% epoxy at 4D is bounding with respect to the IOZ 5D cases. It is unclear to the staff which approachthe licensee is using (4.28D for IOZ or a bounding 4D epoxy case). In addition, the staff has become aware of issues with the testing intended to supporta 5D ZOI for IOZ (WCAP- 16568P). Westinghouse provided revised ZOI analyses for both epoxy and untopcoated inorganiczinc coatings in a letter dated March 24, 2010 (Accession Number ML100880023). The staff has not accepted the Westinghouse-sponsored confirmatory testing and analysis for untopcoatedinorganiczinc coatings. Therefore, the NRC staff no longer finds a 5D ZOl acceptable for untopcoatedinorganiczinc. This conclusion will be documented in a near-term revision to the staff's review guidance on this subject. Instead, licensees may rely on the staff's prioracceptance of a 1OD ZOl for untopcoatedinorganiczinc as documented in its SE for NEI 04-07. Please clarify andjustify the ZOI for untopcoated IOZ without the use of the WCAP- 16568P reduction, or describe impacts on strainerperformance as a result of a decision to use a largerZOI.

-- J:' .

NL-10-074 Attachment 1 Page 10 of 35 Enterqy Response The current Debris Generation calculations, as reported in Entergy's updated supplemental response, used the following design inputs: 4D ZOI for Epoxy coatings and 4.28D ZOI for untopcoated Inorganic Zinc (IOZ). However, the calculation notes that using a 4D ZOI consisting of an Epoxy only system results in a greater volume of coatings debris because the thickness of the Epoxy coatings is much greater than the thickness of the IOZ coatings. The bounding values for the 4D ZOI Epoxy system were reported in the calculation and used in subsequent analysis and testing.

Subsequent to Entergy's updated supplemental response, the NRC no longer endorsed the 5D ZOI from WCAP-16568-P and recommended the use of a 10D ZOI as noted above (ML100880023).

Entergy has evaluated the impact of applying the 10D ZOI by revisiting the debris generation calculation. Because walkdown documentation indicates the presence of Epoxy and untopcoated IOZ, a 4D ZOI for Epoxy coatings and the NRC recommended 10D ZOI for the untopcoated IOZ were applied. The new calculations determined the volume of coating debris for the applicable surface areas of untopcoated IOZ and Epoxy coating systems by utilizing more realistic, yet still conservative, quantities of the different coating systems.

The results of the bounding RCS loop break case which includes the coatings of the Pressurizer supports are presented below for the original and revised coating debris volumes:

Unit Original quantities in Revised quantities in Debris Generation Debris Generation (4D Epoxy) (4D Epoxy and 31OD IOZ)

IP-2 5.32 ft3 5.20 ft IP-3 4.67ft3 4.63 ft3 These results illustrate that the original 4D Epoxy coating debris volumes are still bounding and conservative. The Debris Generation calculations will be updated to address the issue raised by this RAI, however, the strainer qualification calculations remain bounding.

NRC Request 4 Chemical Effects In RAI 23 the staff asked the licensee to submit the revised chemical effects test results and analyses. The licensee responded that chemicalprecipitateswould not occur for the first seven hours following a loss-of-coolantaccident (LOCA) at Indian Point Units 2 and 3. This conclusion is based, in part, on the following statements from this letter (page200 of 243):

Based on plant-specificaluminum concentrationsand pH, aluminum is predicted to precipitate at 118 0F for IP-2 and 121 OF for IP-3 [Ref. 46]. The minimum possible temperature at 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months after a LOCA is 122°Fand 123 OF for IP-2 and IP-3, respectively. While the temperatures shown above for the minimum containment temperature and the predictedprecipitationtemperature indicate that there is only a small amount of margin, both values contain significant conservatisms that make the

NL-10-074 Attachment 1 Page 11 of 35 actualmargin larger than the calculated margin. The minimum containment temperature is based on a model which used simplifying assumptions to minimize temperature rather than a model using refined inputs to achieve an exact result.

Since the minimum containment temperaturesare lower than the calculated threshold temperaturefor precipitationbased on an equation developed by Argonne NationalLaboratory, the staff does not understandhow the temperatures shown above demonstrate any margin.

Additional statements indicated that the post-LOCA temperatures will not reach the stated minimum values; however, there is no discussion to quantify how these assumptions result in unreasonablylow calculatedminimum temperatures. Pleasejustify the conclusion of no precipitation during the first seven hours following a LOCA.

Enterqy Response In Entergy's updated supplemental response to GL 2004-02, the values for the minimum possible temperature of the post-LOCA containment sump pool provided for IP-2 and IP-3 were incorrectly listed as 11 0°F and 11 60 F, respectively. The following sentence taken from page 200 of 243 of Entergy letter NL-09-138, dated November 19, 2009 is revised below to indicate the correct temperatures:

"The minimum possible temperatureat 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months after a LOCA is 122 OF and 123 OF for IP-2 and IP-3, respectively[Ref. 48]."

Based on the corrected temperatures above, the precipitation point temperatures (1 180 F for IP-2 and 121OF for IP-3) are below the minimum calculated sump pool temperatures; therefore, precipitation is not expected to occur earlier than 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months into the LOCA for either unit.

IP-2 IP-3 Precipitation Point Temperature 118 0 F 121 OF Minimum Sump Pool Temperature 122 0 F 123 0 F The minimum sump pool temperatures were determined in calculation IP-CALC-09-00128, "IP Maximum Post-LOCA Containment Cooling Prior to Hot Leg Recirculation for Units 2 and 3",

that was prepared to obtain a conservatively low temperature in the post-LOCA sump pool at 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months into the accident. These calculated temperatures were compared to the results of the conservative Argonne Nation Laboratory (ANL) method for determining a precipitation point temperature to show precipitation would not occur prior to the switchover to hot leg recirculation, which is required to be completed no later than 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months into the LOCA timeframe. After the procedurally required switch to the hot leg recirculation pathway, the pump flow rate is significantly reduced, and consequentially, so are the head losses for the strainers.

This calculation, together with the associated assumptions, was formulated to result in the maximum cool down rate of the post-LOCA sump pool to conservatively determine if the precipitation point is reached within 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months .

NL-10-074 Attachment 1 Page 12 of 35 The following conservative assumptions were made in the cooldown calculation:

Containment Fan Cooler Units (CFCUs) operating at maximum performance capabilities.

The CFCUs directly cool and condense the air and steam atmosphere of post-LOCA containment on metal tube and fin coils cooled directly by the Service Water System (Hudson River water). The heat transfer data is taken from a calculation in which a fouling factor of zero is assumed for the coils. The CFCUs are normally in operation throughout the year as the primary means of cooling the Containment. Some degree of tube fouling would be expected from the river water, reducing the predicted maximum cooling performance used in the calculation.

Containment Fan Cooler Units operatinq with maximum heat transfer rate CFCU performance was taken from existing Westinghouse Peak Clad Temperature (PCT) analyses which are designed to maximize heat removal at the units and apply minimal Service Water temperatures. Any increase in Service Water temperature directly reduces the delta T for the CFCUs, and therefore, decreases overall transfer rate and quantity of heat removed from Containment.

Heat transfer through the Containment wall only considers concrete and steel liner resistance The sump pool fluid, in addition to the cooldown provided by the Residual Heat Removal Heat Exchangers (RHRHXs), transfers heat to the its surroundings. The Containment wall starting just below the concrete floor at 46 foot elevation, and going up to at least the 68 foot elevation (and higher in some locations), has rigid thermal insulation with a metal jacket against and attached to the wall to reduce the transfer of post-LOCA sump pool heat into the concrete. The insulation protects the concrete by forming a thermal barrier between the sump fluid and the Containment steel liner and concrete. Therefore, the insulation directly reduces heat transfer from the hot sump fluid at this critical location. The effect of the Containment liner insulation is not considered in the calculation; therefore, the magnitude of the cool down rate is conservatively increased.

Heat transfer through Containment uses larger outer building diameter Outside dimensions were applied in calculating the heat losses through the Containment wall to maximize the area and heat transfer to the outside environment. This over predicts the heat transfer since the actual heat input into the walls is physically on a smaller surface area based on inside dimension of the Containment.

Entire outer surface of Containment assumed equal to minimum design basis external air temperature This assumption is conservative because the Containment Reactor Cavity, and much of the Containment Building below 46 foot elevation, where the sump fluid is, is either below ground or shielded. Sub-surface ground temperatures are generally constant and would be expected to be closer to 40 to 50 0 F, well above the postulated minus 50 F ambient air design value applied. Some of the lower portions of the containment are also shielded by and adjacent to heated structures (e.g. Primary Auxiliary Building, Auxiliary Feed Pump/Steam Bridge, and Fuel Storage Building) that reduce the temperature differential and heat transfer to the assumed minus 50 F outside air temperature.

NL-10-074 Attachment 1 Page 13 of 35 The Component Cooling Water System temperature is assumed to be at, and remains at, its lowest possible value of 70°F The CCWS is a closed loop cooling system that accepts heat from the Residual Heat Removal Heat Exchanges (RHRHXs) (and other plant loads), and then rejects it to the Service Water System (SWS), and ultimately to the Hudson River. The CCWS loop temperature is procedurally maintained between a minimum of 70°F and 11 0°F for other equipment served by the loop and as part of the implementation for the resolution to GL 2004-02. As discussed in the Ultimate Heat Sink Analyses for IP-2 and IP-3, during abnormally high system heat loads, which occur in a LOCA where the RHRHXs are the primary means for sump pool fluid temperature reduction, the CCWS will initially increase above this defined minimum temperature (70 0 F) during recirculation mode. The CCWS may also be at a higher temperature before the accident, as 70°F is a minimum value. The increase in CCWS temperature would reduce the heat transfer effectiveness (delta T) and consequently the sump cooldown rate.

Sump fluid density assumed to be at 125 0 F for RHRHX calculations throughout analysis The density of the sump fluid was assumed to be at a low temperature of 125 0 F (as opposed to the high end temperature, or a more realistic average of the range) in the RHRHX heat transfer calculations throughout the entire duration of the accident. This higher applied fluid density effectively maximizes the heat removal performance of the RHRHXs by increasing mass flow rate in the tubes, and results in a larger Logarithmic Mean Temperature Difference (LMTD) between the hot (tube) side and cold (shell) sides. In implementing this conservatism, the heat capacity (cp) is taken at the same temperature (125 0 F). This is reasonable, since it will vary only by about 2% from the range of interest (125 0 F to 2500 F) and has a negligible net effect on sump temperature.

Maximized flow applied during recirculation spray mode (suction from sump)

Flow rates for this period of post-LOCA recirculation spray operation for IP-2 use a beyond design basis condition of two spray headers in operation. An extra spray header in operation during this time increases sump and RHRHXs flow rates, thereby increasing heat transfer and Containment cooling. Additionally, both the IP-2 and IP-3 cool down evaluations use flow rates which are from hydraulic calculations which purposely maximized flow. This was done by selecting enhanced pump performance curves, minimizing system resistances, and maximizing Containment water level. The resultant flows are conservatively high, thus increasing heat transfer in the RHRHXs and the cool down rate.

Minimum sump temperature calculated for 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months into the event Hot-leg recirculation, and the associated beneficially reduced flow rates through the strainers, is procedurally initialized at no later than 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months . The final temperature results presented are calculated for a 7 hour8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months heat removal period, 1/2 hour longer than permitted.

This is conservative because the additional time allows further cool down of the sump fluid.

Refueling Water Storage Tank (RWST) temperature is taken at Technical Specification minimum The Technical Specification minimum temperature for the RWST fluid is 40°F for IP-2 and 350 F for IP-3. Both storage tanks have a steam coil heating system to ensure the temperature is maintained above these minimum values during colder periods. Although IP-3 has a lower allowable, the heating system is set to, and capable of maintain at least a minimum 40°F temperature. These heating systems will typically maintain a higher

NL-10-074 Attachment 1 Page 14 of 35 temperature to cover instrument inaccuracy, and for the majority of the year during warmer periods (spring/summer/fall), the RWST will be closer to outdoor ambient temperatures, which are above the 40°F temperature. Since the injected/sprayed fluid (over 300,000 gallons) from the RWST forms a majority of the fluid in the sump pool; this is an additional, unaccounted for heat load in the calculation that must be removed.

Conclusion Based on the above conservatisms, coupled with the ANL methodology that is a known conservative predictor of precipitation point temperatures for the IP conditions, it can be concluded that there is a larger margin between the precipitation and sump pool temperatures than provided in the corrected response above.

NRC Request 5 Debris Transport In RAI 1, the NRC staff asked the licensee to provide an adequate technical basis to support the assumption that some percentage of small pieces of fibrous debris will be capturedon gratings in the upper containment. The licensee's response stated that the retention percentage assumed 50 percent holdup of small pieces on gratingas an input. This assumption in turn was based on drywell debris transportstudy information (NUREG/CR 6369) which showed for each test case the washdown fraction was less than 50 percent. The response continued with a detailed discussion of associatedassumptions, testing, and plant-specific information. The staff considers the licensee response did not adequately address this issue for the following reasons:

a. Forboiling water reactors (BWRs), most debris may be blown downward to the suppressionpool and capturedon the upper surface of gratings, whereas in pressurized water reactors(PWRs), most debris may be blown upward and captured on the underside of gratings. Washdown occurs more readily when debris is capturedon the underside of gratings. Thus, the BWR washdown capture data likely overestimates the PWR condition.

b. A substantialfraction of the debris blown to the upper containment may be blown through gratings. This is unlike the BWR configuration, wherein debris subject to washdown may be blown downward and trapped on the upper side of grating without having first passed through other grating. The licensee stated that this effect is conservatively accounted for by assuming debris trapped on the underside of gratings would fall back to the containmentpool. However, there are additionalconsiderations.

The staff would expect that full considerationof this design difference would have resulted in significantly fewer small pieces being blown to upper containment than assumed by the licensee (59 percent). In addition,NUREG/CR-6369 shows that debris thatpasses through one or more levels of grating during blowdown is more like fines that would tend not to be retainedby gratingsthan like small pieces. Thus, the BWR retention data cited by the licensee for the small pieces in the uppercontainment of a PWR would not be applicable to a significant fraction of these pieces that would be significantly smaller. The staff expects that washdown for PWR debris would be significantly higher than the BWR washdown data, due to the fact that the pieces that reachedupper containment would likely be smaller and more like fines.

NL-10-074 Attachment 1 Page 15 of 35

c. Blowdown testing has shown that substantiallyless capture is observed on the second gratingin a series due to the smaller debris size distribution. The licensee's model lacked considerationof this factor when crediting the second grating in series, instead having a capture fraction equal to the first.

d. Debrisin the NUREG/CR-6369 washdown and erosion testing was piled up and packed together much more than the NRC staff would expect for the PWR case, which would not blow down directly onto gratings to the same extent. It is not clear that such packing would exist for the PWR configuration, except at boundaries where debris is washed off of solid floors or surfaces and is exposed to concentratedflow. More washdown will occur for a less-packed debris configuration or with the presence of concentratedflow.

e. Concentrateddrainagewas not consideredin the Indian Point evaluation. The licensee determined a flow flux of 0.4 gpm/ff2 , apparentlyassuming uniform drainageacross the containment cross section. This value is significantly lower than the value used in the BWR testing. However, the Indian Point containment drainagewould likely be more concentratedat locations where large debris masses are trapped on gratings, since water and debris typically transporttogether during washdown. Solid flooring and obstacles will lead to significant non-uniformity in the debris and water drainage distribution that includes flow through gratings. The staff questions the conservatism of the licensee's assumption of 50% pass through of small pieces through grating since neither testing nor evaluation has adequately consideredthe effect of non-uniform drainage. It is unclearthat a low uniform dispersed flow representspotential plant conditions in that local conditions where washdown and erosion would occur are not accounted for.

f. Of all the tests done with sprays, only tests of 30 minutes were done for small pieces, with one 60-minute test for medium pieces. NUREG/CR-6369 concludes that a transportfraction of 1.0 is appropriatefor debris smaller than gratings for either break or spray flows. The licensee'sassumption of 50 percent small piece retention on gratings is inconsistent with the conclusions of the document from which the data is taken.

g. Although the licensee correctly stated that NUREG/CR-6369 indicates that the majority of washdown occurred during the first 15 minutes of testing, it is clear that the NUREG did not conclude that washdown ceases after this time. Without having run tests prototypicalof plant conditions (finer debris that is more spread out and potentially lower spray flows), the staff does not agree with the licensee's determination that washdown will effectively cease after 15 or 30 minutes.

h. The Utility Resolution Guidance indicates transportfractions of 1.0 for Mark I and Mark Ill BWR containments; the correspondingSE modified the Utility Resolution Guidance position to recommend a 1.0 transportfraction for Mark II containmentsas well.

Therefore, no hold up credit for gratings was permitted by the approved analysis methodology used by the BWRs. The staff notes that the licensee's discussion (response Page 80) using the BWR Owners Group washdown data relied on a method the staff did not consideracceptable for BWRs. The staff did not consider the BWR Owners Group washdown testing conservative for its intended use. This also supports the staff's interpretationthat NUREG/CR-6369 concluded that no retention should be credited underspray-only conditions.

L The licensee's discussion on gratingsin series does not appearto account for the reason the debris was washed down. The NRC staff considers that debris in the washdown tests did not pass through the gratingsbecause it lined up correctly with the openings in the grating; the debris was piled up on top of the gratings more or less randomly. The more likely reason it passed through the gratings was due to the flow

NL-10-074 Attachment 1 Page 16 of 35 interactingwith, or breaking up, or realigning,or forcing the debris through the openings.

Therefore, the staff does not consider it appropriateto credit multiple gratingsin series with the same capture fraction based on a simplified geometric argument that does not address the size distributionchanges discussed above, nor the associatedmechanisms by which the debris could pass through the initial grating.

j. The licensee considereddebris retention on solid floors an uncreditedconservatism.

Although the predicted flow velocities on such floors typically would exceed the incipient tumbling velocity for certain fibrous small pieces, the licensee considers that the debris pieces would be saturatedwith water and thus transportvia partiallysubmerged tumbling. However, the NRC staff's view is that, at more than several linearL/D from the break, debris pieces, while wetted, would not likely be fully saturatedwith water by the jet. The staff does not consider it conservative to assume debris pieces will be soaked when determining transportacross containment floors. Pieces of fiber would likely still be partiallyfloating, particularlyin cooler spray water that would constitute the water on solid floors in upper containment. This latter effect was not consideredby the licensee and could significantly increase transport. Even if the debris were not floating, there would still be no way to assess whether the debris continued to house trappedair that would change significantly the frictionalforce felt by the tumbling debris per the licensee's analyticalmethodology. The staff considers it appropriatefor the licensee to consider limiting fluid thermodynamic conditions when assessingdebris wetting and saturationwith water. The staff considers the licensee's analyticalderivation of transportmetrics for fully liquid saturateddebris transportingunder partiallysubmerged conditions to lack adequatejustification. The licensee took significant credit for this unvalidatedmethodology, deriving incipienttumbling metrics 20 times higher than the accepted measuredvalues for submerged conditions. Measurements of actualdebris transportvelocities would be needed to validate the licensee's analyticalmethodology.

Therefore, the staff did not consider this analysis to show a significantconservatism.

k. The licensee stated that retention of inertially captured debris would realisticallyoccur but was conservatively not credited. The staff considers it unlikely that a significant mass of inertiallycaptured debris will be retained in the long term. Whether by sprays, condensate, or gravity, much of this debris will release from vertical surfaces or the underside of horizontalsurfaces.

I. The uncertainty with blowdown and washdown transportis very high due to a lack of testing. The behavior of debris in response to these mechanisms is not well understood,as is discussed in NUREG/CR-6369.

Pleaseaddress the above issues to justify the holdup credited,or otherwise consider the impact of reduced holdup of small debris on gratingsand other features above the containmentpool.

Enterqy Response This response will evaluate the impact of revising the debris transport methodology to use a higher washdown fraction for small fiberglass debris.

Debris Transport Fractions for 100% Washdown To obtain the "revised" debris transport fractions for small fiberglass pieces due to 100%

washdown from upper containment, the debris transport logic trees from the Debris Transport

NL-10-074 Attachment 1 Page 17 of 35 Calculations were modified. The IP2 debris transport logic tree for Nukon to the IR sump is shown below as an example. The original logic trees for each sump and each debris type required a few changes. The value shown in Figure 1 for "Retained on Structures" under

'Was~hdown Transport' (0.40) was changed to zero, while the 'Washed Down Annulus" value was increased to 0.53. Those values were then carried forward through the logic (to the right on the tree) to obtain the revised transport fractions. A summary of the results is shown in Table 1.

The revised transport fractions for small fiberglass pieces are approximately twice the original values.

Blwdown Washdown Pool Fill CID Fraction Of Debris Site Transport Transport Transport _rinulsport Erosion Debris at Sum Transport 0.40 Erodes to Fines Relained on O09 Structures Remains Intact 0.10 0.000 Transport 0A7 0.51 Washed Down 0.1( 0.024 Upper Steam 1.0 Erodes to Finor Containment Generator Sediment 0.90 Compartments Remains Intact 0.91 0.060 Transport 0.13 Washed Down 0.10 0.001 Anuulus OI19 Erodes to Fines Small Sediment 0.9 0 Fiberglass Remains Intact Debris 0.01 0.010 G-:eneration 00 Past IR Sump

_______000 _

Flow Barriem 1.0Q " 0.00 Transport 0.000 Active Pool 1.004 0.10 0.040 t,00 Erodes 1o Finies Sediment 0.9(]

0.40 Remaies Inat Lower Containment 0.0] 0.000 Sump SL,'eeon3 0.00 rIaL'i

vhe Pool Sum: 0.145 Figure 1: Original IP2 Nukon, Transco Blanket, and fiberglass small pieces debris transport logic tree (IR sump)

NL-10-074 Attachment 1 Page 18 of 35 Table 1: Small Piece Debris Transport Fractions Original, 50% Revised, 100%

Washdown from Washdown from Upper Containment Upper Containment IP2 IR Sump Nukon 0.15 0.330 IP2 IR Sump Temp-Mat 0.15 0.352 IP2 VC Sump Nukon 0.09 0.115 IP2 VC Sump Temp-Mat 0.14 0.343 IP3 IR Sump Nukon 0.13 0.284 IP3 IR Sump Temp-Mat 0.15 0.352 IP3 VC Sump Nukon 0.08 0.107 IP3 VC Sump Temp-Mat 0.14 0.343 Debris Load Comparisons for 100% Washdown To determine the "revised debris load" at the sump, the debris generation values for a LBLOCA, a reactor cavity LBLOCA, a 6 inch LOCA, and a reactor cavity 6 inch LOCA (4 cases for each unit), are multiplied by the transport fractions. Then, for comparison to the existing test inputs, the bounding debris load at the sump was determined and multiplied by the test scaling factor according to the methodology documented in the Test Debris Amounts Calculation. A summary of the revised test debris loads is shown in Tables 2 and 3. It should be noted that only the fiber loads are affected by this change in methodology.

Table 2: 1P2 Revised Test Debris Loads Total Surrogate Amounts for Test (lbs)

Cal- W A Unit Sump Break Nukon Temp-Mat Min. Wool Sil Sil-Co-Sil Dirt (NaAISi)

(Ibs) (Ibs) (Ibs) (Ibs) (Ibs) (Ibs)

IP-2 IR LB 35.12 7.59 0.00 7.49 73.60 6.26 3.12 IP-2 IR RCLB 1.10 4.33 0.00 0.00 69.73 6.26 3.02 IP-2 VC 6" 15.79 0.00 0.00 6.16 85.34 19.07 5.87 IP-2 VC RC 6" 3.37 3.25 0.00 0.00 77.82 19.07 6.54

NL-10-074 Attachment 1 Page 19 of 35 Table 3: 1P3 Revised Test Debris Loads Total Surrogate Amounts for Test (lbs)

UiSup BekCal- WCAP Unit Sump Break Nukon Temp-Mat Min. Wool Sil Si--Co-Sil Dirt (Ibs) (Ibs) (Ibs) (Ibs) (lbs) (Ibs) (NaAISi)

IP-3 IR LB 25.23 31.36 1.04 9.66 52.26 7.75 3.74 IP-3 IR RCLB 1.37 4.17 0.00 0.00 42.99 7.75 2.76 IP-3 VC 6" 4.78 38.59 0.00 9.57 27.42 24.63 6.49 IP-3 VC RC 6" 4.36 3.26 0.00 0.00 20.66 24.63 6.42 As shown in the table above, all of the revised debris loads for the Reactor Cavity breaks have a fiber mass of less than 8 lbs (after assuming 100% washdown from upper containment), which is less than that required to form a thin-bed (see Table 4). Therefore, the full load for these Analytical Debris Generation Cases are bounded by the thin-bed tests and do not require further evaluation.

In the sections below, each Analytical Debris Generation Case (non-reactor cavity) will be discussed individually. The revised debris (fiber) load is compared to the prototypical array test data collected in Spring 2009 using the "Test-for-Success" methodology and conforming to the requirements of the March 2008 Guidance. This methodology is similar to that performed in the original Strainer Head Loss Calculation to determine the bounding head loss for each Analytical Debris Generation Case.

Indian Point's GSI-191 analysis is based on no chemical precipitation prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months into the accident. The head losses for conventional debris and conventional plus chemical debris are evaluated independently below because they have different acceptance criteria. The conventional debris head losses are evaluated against the NPSH margins and the minimum flow criteria prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months when the sump pool temperature and flow rates are relatively high.

The conventional plus chemical debris head loss is compared to the NPSH margins and structural limit after 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months when the sump pool is cooler and the flow rates are lower. Both head losses are presented below because it is important to evaluate the margins for both time periods.

IP2 IR Sump LBLOCA The revised transport fraction results in a total fiber load for the IP2 IR sump of 42.71 lbs. Test C was conducted with a total of 41.39 lbs of fiber and produced a conventional head loss that was 0.27 ft (9%) lower than the thin bed head loss for conventional debris and 4.86 ft (43%)

lower than the thin bed head loss with chemical effects. The final head loss margin for this Analytical Debris Generation Case is 0.52 ft (44%) for conventional debris (prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months ) and 6.97 ft (450%) for conventional plus chemical debris (after 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months ). This results in an "effective" margin of 53% for prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months and 493% after 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months . Since the revised fiber load is 3.2% higher than Test C and the effective head loss margin is approximately 53%, it is not expected that the full load for this Analytical Debris Generation Case will challenge the NPSH margin.

NL-10-074 Attachment 1 Page 20 of 35 The revised insulation debris loads for this case together with the applicable debris loads and normalized head losses from each test stage are shown in Table 4. The highlighted cells indicate the bounding tests that were used to determine the final head loss margins (NPSH, minimum flow, or structural margins).

Table 4: Debris Loads for IP2 IR Sump LBLOCA and Prototypical Tests with Applicable Test Stage Head Losses Test Debris Loads *Measured Head Loss Extrapolated Temp- Min. Sil-Co- Conventional Conventional &

Nukon Mat Wool Cal-Sil Sil Dirt WCAP Conventional & Chemical Chemical (lbs) Olbs) Olbs) (lbs) (lbs) (lbs) I(Ibs) I ft-water) (ft-water) (ft-water)

C-I1 18.94 21.75 0.70 9.66 51.09 7.75 3.38 2.80 5.05 6.51 E E-2 (Thin Bed) 14.25 0.00 0.00 9.66 80.49 15.57 3.12 1.10 9. 11,37 F-2 (Thin Bed) 8.55 0.00 0.00 9.57 23.37 24.63 3.12 3.07 9.37 10.17 IP2 VC Sump 6 Inch Break LOCA The revised transport fraction results in a total fiber load for the IP2 IR sump of 15.79 lbs. This

.revised load can be compared to Test Al which contained approximately 5% less total fiber.

The 5% difference in fiber amount is unlikely to cause a significant head loss increase for the revised load because it is equivalent to a 1/32" increase in debris bed thickness. Additionally, Tests B and C were conducted with drastically greater fiber loads and did not produce bounding head losses. Therefore, it is concluded that the revised debris loads would not cause the full load to become limiting for this Analytical Debris Generation Case.

The revised insulation debris loads for this case together with the applicable debris loads and normalized head losses from each test stage are shown in Table 5. The highlighted cells indicate the bounding tests that were used to determine the final head loss margins (NPSH, minimum flow, or structural margin).

Another mitigating factor that provides assurance that this case will not challenge the operation of the recirculation pumps is the final head loss margin. The margin for the conventional debris head loss is 3.13 ft (81%), while the margin for the chemical effects head loss is 2.92 ft (33%).

NL-10-074 Attachment 1 Page 21 of 35 Table 5: Debris Loads for IP2 VC 6 Inch Break LOCA and Prototypical Tests with Applicable Test Stage Head Losses Test Debris Loads I *Measured Head Loss Extrapolated Temp- Min. Cal- Sil-Co- Conventional Conventional Nukon Mat Wool Sil Sil Dirt WCAP Conventional & Chemical & Chemical (lbs) (lbs) (lbs) (lbs) (lbs) (lbs) (lbs) (ft-water) (ft-water) (ft-water)

A (Thin Bed) 13.54 0.00 0.00 9.66 80.49 15.57 1.83 Al-5 (Thin Bed) 15.13 0.00 0.00 9.66 80.49 15.57 5.87 0.96 7.19 8.93 B-4 25.19 5.43 0.00 7.49 72.13 6.26 6.54 1.38 8.78 11.73 C-4 18.94 21.75 0.70 9.66 51.09 7.75 6.49 2.80 5.80 7.36 E-4 (Thin Bed) 14.25 0.00 0.00 9.66 80.49 15.57 5.87 1.10 15.87 20.60 F-4 (Thin Bed) 8.55 0.00 0.00 9.57 23.37 24.63 5.87 ' 307

0 13.98 19.89 IP3 VC Sump 6 Inch Break LOCA The revised transport fraction results in a total fiber load for the IP3 VC sump of 43.37 lbs. Test C was conducted with a total of 41.39 lbs of fiber and produced a conventional head loss that was 0.27 ft (9%) lower than the conventional debris thin bed head loss and 14.52 ft (66%) lower "han the chemical effects thin bed head loss. The final head loss margin for this Analytical Debris Generation Case is 0.83 ft (71%) for conventional debris (prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months ) and 2.92 ft (34%) for chemical effects (after 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months ). This results in an "effective" head loss margin of 80%

prior to 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months and 100% after 7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months . Since the revised fiber load is 4.8% higher than Test C and the effective head loss margin is 80%, it is not expected that the full load for this Analytical Debris Generation Case will challenge the NPSH margin.

The revised insulation debris loads for this case together with the applicable debris loads and normalized head losses from each test stage are shown in Table 6. The highlighted cells indicate the bounding tests that were used to determine the final head loss margins (NPSH, minimum flow, or structural margin).

NL-10-074 Attachment 1 Page 22 of 35 Table 6: Debris Loads for IP3 VC Sump 6 Inch Break LOCA and Prototypical Tests with Applicable Test Stage Head Losses Test Debris Loads *Measured Head Loss Extrapolated Temp- Min. Sil-Co- Conventional Conventional Nukon Mat Wool Cal-Sil Sil Dirt WCAP Conventional & Chemical & Chemical (lbs) (lbs) (lbs) (lbs) (lbs) (lbs) (lbs) (ft-water) (ft-water) (ft-water)

C-4 18.94 21.75 0.70 9.66 51.09 7.75 6.49 2.80 5.80 7'46 E-5 (Thin Bed) 14.25 0.00 0.00 9.66 80.49 15.57 6.54 1.10 ; 1,, <t2,88; F-5 (Thin Bed) 8.55 0.00 0.00 9.57 23.37 24.63 6.54 3.0 14.7319.0 I I \1 t~ T"r IP3 IR Sump LBLOCA The revised transport fraction results in a total fiber load for the IP3 IR sump of 57.63 lbs. Since the revised fiber load is 39.2% higher than Test C, the prototypical array test data from 2009 cannot be used to evaluate this case, but it is evaluated using previous head loss testing below.

The revised insulation debris loads for this case together with the applicable debris loads and normalized head losses from each test stage are shown in Table 7. The highlighted cells indicate the bounding tests that were used to determine the final head loss margins (NPSH, minimum flow, or structural margin).

Table 7: Debris Loads for IP3 IR Sump LBLOCA and Prototypical Tests with Applicable Test Stage Head Losses Test Debris Loads *Measured Head Loss V V V 9 V V Extrapolated Temp- Min. Sil-Co- Conventional Conventional Nukon Mat Wool Cal-Sil Sil Dirt WCAP Conventional & Chemical & Chemical (lbs) (lbs) (Ibs) (lbs) (ibs) (lbs) (lbs) (ft-water) (ft-water) (ft-water)

B-4 25.19 5.43 0.00 7.49 72.13 6.26 6.54 1.38 8.78 11.73 C-3 18.94 21.75 0.70 9.66 51.09 7.75 3.74 2.80 5.28 6.65 r E-3 (Thin Bed) 14.25 0.00 0.00 9.66 80.49 15.57 3.74 1.10 10.91 12.32.K;'

F-3 (Thin Bed) 8.55 0.00 0.00 9.57 23.37 24.63 3.74 3.07 10.02 10.78 Analytical Debris Generation Case Matched Test Stases IP3 IR Sump LBLOCA Comparison to Previous Head Loss Testing Indian Point conducted prototypical array testing in the Fall of 2007, prior to the release of the March 2008 Guidance. Test #1 of the series was conducted to bound the full debris load for the

NL-10-074 Attachment 1 Page 23 of 35 various sumps and Analytical Debris Generation Cases. As shown in Table 8, Test Step #1 D included debris amounts that bound the revised loads for the IP3 IR Sump LBLOCA break case.

Therefore, this test data can be used to estimate the head loss resulting from the 100%

washdown assumption. The stabilized head loss from Test Step #1 D was 2.05 ft-water at 398 gpm and 82.50 F. For comparison to the head losses presented in the section above, the head loss for Test #1 D is normalized to 400 gpm and 700 F. The flow sweeps for Test #1 indicate that the maximum turbulent fraction is 6.3%, but 10% is used in the normalization for conservatism.

The methodology developed in the Strainer Head Loss Calculation and data given in Table 9 is used for normalization as presented below.

Table 8: Debris Loads for IP3 IR Sump LBLOCA and Test #1 D Test Debris Loads Normalized min. Sil-Co- Conventional Nukon Temp-Mat Wool Cal-Sil Sil Dirt WCAP Head Loss 1ibs) (ibs) (lbs) (lbs) (Ibs) (lbs) I(lbs) (ft-water)

Test#1 D 26.42 39.00 1.04 16.20 57.98 24.00 N/A 2.56 Analytical Debris Generation Case Table 9: Data for Normalized Head Loss Correction Test Data Normalized Data Flow Rate (gpm) 398 400 Temperature (°F) 82.5 70 Water Viscosity 5.56*10-4 6.67*10-4 (Ibm/ft/s) I Turbulent Fraction 10% N/A Equation for Increased flow and decreased temperature case (note the conservatism in evaluating the head loss as fully laminar for the temperature decrease and partially turbulent for the increased flow):

"I2 = A"I 2 1 A U Ifrn 0.97xQ2 Q, +Tfa Q22

"- (0.97)2QI2 AH2 = 2.05ft 6.67 .10-4 5.56 .10.9 400gpm 0.97x398gpm (400gpm)2 (0.97)2 (398gpm)2 1 2 The normalized head loss for Test #1 D (Fall 2007) is calculated to be 2.56 ft, which is less than the head losses measured in the most recent tests for thin-bed (Test F: 3.07 ft) and full load (Test C: 2.80 ft) (see Table 7). Test #1 D was conducted with only minor differences in debris preparation of fine debris as compared to Tests C and F. The fine debris preparation described in the March 2008 guidance is more critical for thin bed tests, while the increase in debris discussed here only applies to full debris loads. Therefore, Test #1D is considered acceptable to evaluate head loss for the IP3 IR Sump LBLOCA case, and the comparison indicates that the revised debris loads will not have an adverse effect on the full load head loss. Because the debris loads and conventional debris head loss for Test #1D were similar to those of Test C, it is

NL-10-074 Attachment 1 Page 24 of 35 expected that the effects of chemical precipitates would also be similar. The extrapolated chemical effects head loss for Test C is 6.65 ft, which is 5.67 ft lower than the thin-bed test (Test E: 12.32 ft) that was used to qualify the strainer. Therefore, it is reasonable to assume that the revised full load would not become bounding for conventional or chemical debris head losses.

Additionally, the final head loss margin for chemical effects head loss for the IP3 IR LBLOCA is 7.83 ft. (477%), which provides further confidence that the revised debris load would not adversely affect operation of the recirculation pumps.

Alternate Methodoloqy: Determination of Limiting Washdown Fraction An alternate method of evaluating the impact of a higher washdown fraction is to determine the maximum washdown fraction while maintaining a total fiber load less than Test C (41.39 Ibs),

which is the maximum fiber tested in 2009. The fractions for the four Analytical Debris Generation Cases are determined by iterating the methodology described above, and a summary of the results is shown in Table 10. The IP3 IR Sump LBLOCA case resulted in no change in washdown fraction because the highest fiber load test (Test C) was designed to contain the exact debris load for this case based on 50% washdown (see previous section for evaluation of this case).

Table 10: Maximum Washdown Fractions to be Bounded by 2009 Prototypical Head Loss Testing IP2 IP3 IR Sump LBLOCA 95.5% 50%

VC Sump 6" LOCA 100% 93.5%

Conclusions The washdown fraction used in the Debris Transport Calculation was intended to include debris that is captured on equipment, piping, walls, floors, grating, inactive areas, and other miscellaneous items. The justification used for the washdown value was based on testing conducted with small fiberglass pieces on grating with flow rates that were significantly higher than Indian Point's flow rate. The test results should be conservative when considering only the grating as a debris transport restriction and certainly conservative when considering the upper containment area as a whole. RAI 5 focuses solely on washdown through gratings, but does not consider that the debris must also transport to the grating past equipment, piping, walls, floors, inactive areas, and other miscellaneous items before it can be washed down. Therefore, the 100% washdown assumption evaluated above is an extremely conservative methodology. It is difficult to imagine that nothing would be captured in the upper containment area and it would be completely fiber-free following a LBLOCA.

The transport fractions used to determine debris loads in this evaluation contain considerable conservatism. The erosion fraction is assumed to be 10%, which was demonstrated to be conservative based on the March 2010 erosion test results. As discussed above, the 100%

washdown fraction is very conservative. Finally, the recirculation transport is based on maximum flow rates for the beginning of recirculation, 3,700 gpm for the VC sumps, 7,100 gpm for the IP2 IR sump, and 5,400 gpm for the IP3 IR sump. The alignment is switched to the HHSI

NL-10-074 Attachment 1 Page 25 of 35 pumps no later than 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months and the flow rate is reduced to 1,350 gpm (or less) for the remainder of the 30 days.

As shown in the sections above, the 100% washdown assumption for small fiberglass pieces from upper containment has a significant effect on the debris loads that are predicted to arrive at the sump. However, for each of the four Analytical Debris Generation Cases, test data exists for comparable fiber loads which indicate that the thin-bed debris loads would continue to produce the bounding head losses rather than the revised full debris load amounts determined in this analysis. Note that only four cases are evaluated because the 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months time-dependent cases are evaluated in a separate response, and the reactor cavity cases are completely bounded by the thin-bed tests and don't require further analysis.

The alternate methodology shows that 3 of the 4 Analytical Debris Generation Cases can use a washdown fraction near (or exactly) 100% to obtain a fiber quantity that is bounded by the latest "Test-for-Success" data. The other case, using the 100% washdown fraction, is bounded by a previous test that was conducted prior to the March 2008 Guidance. In summary, using a 100%

washdown fraction for debris from upper containment versus the 50% does not change the bounding debris head loss and the existing analysis used to certify the strainer is valid.

Finally, each case has other mitigating factors, including final head loss margin, that provide confidence that adequate NPSH will be maintained for the recirculation pumps (see Table 11).

Table 11: Mitigating Factors for Each Analytical Debris Generation Case IP2 IP3 IR Sump Significant conventional and Bounded by previous Test #1 D.

LBLOCA chemical effects head loss Significant chemical effects head margin, loss margin.

Bounded by Tests B and C. Significant conventional and VC Sump Significant conventional and chemical effects head loss margin.

6" LOCA chemical effects head loss margin.

NRC Request 6 Debris Transport In RAI 2, the staff asked the licensee to provide an adequate technicalbasis to support its assumption of ten percent fibrous debris erosion in the containmentpool The response stated that this was a reasonableassumption and providedjustification. To further support the assumption, the licensee is participatingin an industry program to generateadditionalerosion test results, with a reportexpected in April 2010. Please provide a description of the test and the test results once completed in order to demonstrate the adequacy of the assumed erosion percentage.

NL-10-074 Attachment 1 Page 26 of 35 Enterqy Response Entergy participated in the Alion testing program to determine an erosion fraction for small pieces of Low Density Fiberglass (LDFG). After a 5 day pre-test was completed, the primary test was performed for 30 days on prepared LDFG samples, and then a follow-up confirmatory test was performed for 13 days on a second group of samples. The latter test provided information on repeatability and to confirm that the erosion rate tapers off early on during the recirculation phase. The LDFG test samples were arranged in the test flume such that the velocity and turbulent kinetic energy (TKE) values of the flow stream bounded the values in the IP-2 and IP-3 Containment Debris Transport models.

The 30 day test determined an erosion fraction value of 6.19%, while the confirmatory 13 day test determined an erosion fraction of 4.91%. Although the resultant tested values are less than the currently assumed 10% erosion fraction, the report concludes that a conservative value of 10% should be used for both small and large pieces of LDFG. As 10% is the erosion fraction that was applied in the IPEC analyses prior to this confirmatory testing, no changes to Debris Generation or other design bases analyses are required.

A copy of the Alion test report: ALION-REP-ALION-1006-04, Revision 0: "Erosion Testing of Small Pieces of Low Density Fiberglass Debris - Test Report," is available for NRC staff review, see NRC ADAMS Accession Number ML101090490.

NRC Request 7 The licensee is creditingtime-dependent debris transportfor qualificationof the vapor containment (VC) sump. In RAI 5, the NRC staff asked the licensee to provide adequate technicaljustification that the time-dependent model is conservative. The response provided an analysis that noted the effect of each staff concern was quite small, and that only a small fraction of the debris would remain in the pool after one day (0.5 percent for IP2 and 0. 7 percent for IP3). After reviewing this information, the staff still has questions concerning the adequacy of the head loss test assumption of less than 5 percent fiber transportto the VC sump. The staff considers that the licensee's response did not adequatelyaddress this issue for the following reasons:

a. The licensee assumed debris all washed down prior to switchover, minimizing transportto the VC sump. The RAI response stated if washdown were delayed to 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months , the transportof fines to the VC sump would be about 4 percent. The NRC staff finds this response inadequatein that, when a realisticdelay is assumed, in conjunction with the other items noted below, it could lead to greater than 5 percent transportto the VC sump.

b. The NRC staff noted that the fiberglass erosion curve presentedby the licensee was based on data from Alion that anomalouslyshowed significantly less cumulative erosion for long-term tests than for short-term tests. The staff also noted that the curve fit is not consistent with data seen from tests better suited to assessing time-dependence, and does not seem consistent with the most recent test data that Alion is collecting concerningerosion and its time dependence.

, *NL-10-074 Attachment 1 Page 27 of 35

c. The licensee indicated that the strainershave bypass eliminatormesh installed, which would significantly reduce the quantity of fibers and some other types of debris that may pass through the strainers. The staff questions, however, whether the 100 percent captureassumption for fine particulate(e.g., 10-micron diameter)is realistic, as simulations have shown that 10-15 pool turnovers are needed to filter out fine particulatefor a debris bed of representativeporosity.

The staff does not agree with the licensee that after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months for a single-train case less than 5 percent of the fine particulate debris would remain in suspension based on the times associatedwith forming a debris bed with high filtration efficiency and subsequently to achieve 10-15 pool turnovers.

d. The licensee stated little debris bed movement was observed during Indian Point plant-specific testing, that check valves would prevent significant reverse flow into the internal recirculationstrainers,and the debris bed would not be easily broken down due to agglomerationof constituent debris pieces. The licensee stated that releaseddebris would not easily transportdue to being in a pit that is physically separatedfrom the VC sump. The NRC staff did not fully agree with these statements. The staff has observed that accumulated aircould result in significant debris release; has seen considerable debris bed movement following pump stoppage; and if only the top row of top hats releases debris, and only a tenth of this debris is releasedand transported,about 1 percent of the total internal recirculationsump debris load could be on the VC sump strainer.

e. The licensee stated that debris would not be directed toward the VC sump during blowdown. The NRC staff generally agrees that significant transport would not occur during blowdown based on the barriersthe licensee installedthat were describedas preventing blowdown transport. However, for pool-fill, although a significant part of the fines may still be in upper containment, blowdown transport is chaotic and difficult to predict. Therefore, the staff expects that pool-fill would resultin the transportof a fraction of the fine debris to the VC sump. Although difficult to predictaccurately,it is not clear to the staff that a non-recirculation transport(i.e., primarily through pool-fill) fraction less than a percent or two can be justified for fines (as an order of magnitude), which would pass through the perforatedbarriers.

The NRC staff questions the 5 percent assumption given the items identified above. Please justify the assumption in light of the items above, or else please provide a descriptionand results of an evaluation of how the plant's system response would be affected by potentially greater debris transportto the VC strainerbased on these considerations.

Enterpy Response Delayed Washdown (RAI 7, Item a)

The original Debris Transport Calculations assumed that all debris was washed down to the sump pool prior to the start of recirculation. While it is a reasonable assumption based on test data discussed in NUREG/CR-6369, this assumption was identified as a potential non-conservatism in RAI 7.

The previous RAI response presented an example where washdown was delayed for four hours. This was a hypothetical scenario postulated to demonstrate that the transport to the VC sump would be below 5%; however, the data in NUREG/CR-6369 indicates that a 4 hour4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months delay

NL-10-074 Attachment 1 Page 28 of 35 is excessive. A more accurate estimation of the total transport fraction can be developed by using a more realistic delay of washdown based on a conservative interpretation of the washdown time data in NUREG/CR-6369. The tests were run for 30 minutes and yielded washdown fractions of less than 50% for 1.5 inch fiberglass pieces. It was also noted in the observations that the majority of washdown occurred in the first 15 minutes. Therefore, a time-dependent washdown curve can be estimated using a data point of 50% washdown at 30 minutes, which predicts a conservatively late washdown. As shown in Figure 1, the majority of washdown occurs in the first 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months , which is a realistic delay for washdown. Instead of using the previous assumption that all debris is in the pool at the beginning of recirculation (-40 min), it is assumed that all debris enters the pool at 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months after an accident, or approximately 50 minutes after the start of recirculation (based on IP3 parameters). This reduces the amount of debris collected by the IR sump because of the reduced number of pool turnovers in the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . The effect of any small amount of debris that enters the pool after 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months is offset by assuming that no debris collects on the IR sump prior to 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months .

Additionally, no washdown would occur after 4-6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months into the LOCA scenario because sprays are highly likely to be terminated during this time period in accordance with the Emergency Operating Procedures.

1.20 -

1.00 CL CL 0.

c 0.80 "

E

.C* 0.60

,,,0 0.40 W

0.20 0.00 . .....................

0 0.5 1 1.5 2 2.5 3 3.5 4 Time (h)

Figure 1: Washdown as a Function of Time Based on NUREG/CR-6369 Assuming that the debris in the pool is uniformly distributed and the water entering the pool is clean, an exponential decay model (see Equation 1) based on pool turnovers can be used to estimate the quantity of debris in the pool at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . This methodology was used in the Debris Transport Calculations to determine the time-dependent debris transport fractions.

Using an initial fraction of one and the same conservatively low flow rates (for IP3) and maximum sump pool volumes as the Debris Transport Calculations, Equations 2 - 4 show that debris depletion from 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months to 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months will leave 0.9% of the total debris in the sump pool.

This transport fraction will be used in calculating the total debris transport. Note that three equations are needed due to different flow rates for the various time periods after an accident.

This assumption and methodology is conservative because it includes all debris (i.e. smalls and fines), not just the debris that is washed down from upper containment.

NL-10-074 Attachment 1 Page 29 of 35 x(t) = xie(2CJ Equation 1 Fraction of debris remaining in the pool after 4 hour4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months s:

2.5hr *60 min *3,517gpm 0.1337 ft3 hr gal = 0.24 X(tl.5 4 hr )-=1.0" exp - 3 Equation 2 49,998ft Fraction of debris remaining in the pool after 5.5 hour5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months s:

1.hr. 6 0 min 1,402gpm.0.133ft 3 X(t4 _5.5 hr )= 0.24. exp

hr 3 gal =0.17 Equation 3 49,998ft Fraction of debris remaining in the pool after 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months s:

. ft3 18.5hr 60

min- 1,000gpm 0.1337 hr gal 0.009 X(t5.5 24 hr)=0. 17.exp - 3 Equation 4 49,998ft Fiberalass Erosion (RAI 7, Item b)

The Debris Transport Calculation methodology included 10% erosion of the small and large pieces of fiberglass in the sump pool at the beginning of recirculation. Ninety-five percent (95%)

of the transportable debris is collected on the IR sump prior to VC sump recirculation. Thus, 5%

of the total 10% erosion occurs after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months , which is 0.5% (0.1*0.05) of the full amount of small and large fiberglass debris generated. The value was rounded up to 1% for conservatism in the Debris Transport Calculations. It should be noted that erosion does not apply to fine debris because it is assumed to be 100% transportable (or 5% to VC sump after time-dependent reduction).

The most recent erosion testing by Alion provides a time-dependent profile of the erosion process for small fiberglass pieces. It shows that the flow erosion fraction for small pieces is 3.59% for 30 days, while the flow erosion fraction before 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months is 0.7%. Using a simple ratio, 81% of the total erosion occurs after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . However, the average velocity in the testing was much greater than the average velocity around non-transporting debris at Indian Point.

The containment recirculation CFD model was used to analyze the fluid velocities over the areas in the containment buildings where small piece debris does not transport. Using the IR sump maximum flow rates, the average fluid velocities were determined inside the crane wall at the 46' elevation, in the annulus, and in the in-core instrumentation tunnel/under vessel area.

The in-core instrumentation tunnel/under vessel area for IP3 has the highest average fluid

NL-10-074 Attachment 1 Page 30 of 35 velocity, 0.0897 ft/s. The CFD models show that approximately 85% of area inside the crane wall has sufficient velocity to transport small fiberglass pieces to the tunnel, while approximately 25% of the area inside the crane wall has sufficient velocity to transport large fiberglass pieces to the tunnel. However, the velocity in the tunnel is insufficient for either size debris to transport out of the tunnel. Therefore, a large portion of small and large pieces is predicted to remain within the under vessel area or in the tunnel. For the purposes of this evaluation, a velocity of 0.0897 ft/s will be used for conservatism. For the VC sump LBLOCA 24 hr case being analyzed here, the flow rate is only 1,350 gpm, and the flow barriers are assumed to be blocked such that the full flow is directed through the in-core instrumentation tunnel. The fluid velocity is likely to be less than analyzed here because the flow barriers will be partially blocked. Fluid velocities for VC sump recirculation can be approximated using a ratio of the flow rates. A conservative safety factor of 2 is included to obtain a scaled velocity of 0.045 ft/s ([0.0897*1350/5400]*2).

The industry erosion testing was conducted at a fluid velocities ranging from 0.119 ft/s to 0.229 8 884.

ft/s, and a correlation was developed to relate erosion fraction to velocity, 1.0674*vl Because the velocity around eroding debris at IP (0.045 ft/s) is lower than the average velocity in the erosion testing (0.162 ft/s), the correlation can be used to determine a more accurate and bounding erosion fraction for Indian Point. Using a velocity of 0.045 ft/s, the erosion fraction can be estimated as 0.31% (1.0674*0.04518884). However, this value is based on an erosion data that produced an erosion fraction of 3.59%, which is lower than the assumed (and NRC accepted) value of 10%. For this reason, the value is scaled up using a ratio of 10/3.59 for the total erosion, then multiplied by the 0.81 fraction (81%) that was previously determined to erode after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . Using this methodology, the erosion fraction of small debris after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months for Indian Point is 0.70%, which is less than the 1% value used in the Debris Transport Calculations. Therefore, the existing time-dependent analysis for erosion is bounding.

Strainer Bypass (RAI 7, Item c)

The time-dependent methodology used total particulate transport fractions that are significantly higher than the 5% of fine particulate debris quoted in RAI 7. All un-qualified coatings (outside the ZOI) and all fines due to erosion of Cal-Sil/Asbestos are not included in the debris that collects on the IR sump in the time-dependent debris transport analysis. Instead, 100% of these quantities are conservatively assumed to transport to the VC sump. Additionally, it is important to compare these values to the transport fractions for the IR sump to put the values in the proper perspective. The transport fraction for Cal-Sil to the IR sump is not 100% because a portion of the Cal-Sil is destroyed as small or large pieces and doesn't transport to either sump (see Table 1). The transport fraction for coatings debris to the IR sump is not 100% because a portion of the unqualified epoxy paint fails as chips and doesn't transport to either sump (see Table 2). These values skew the transport fractions when comparing the values for the IR sump and VC sump. As shown in Tables 1 and 2, the difference in IR sump transport fraction and the VC sump time-dependent transport fraction is not as large as it initially appears.

NL-10-074 Attachment 1 Page 31 of 35 Table 1: Com parison of Transport Fractions for Cal-Sil*

Time-Dependent Transport of Transport of Cal-Sil to VC Cal-Sil to IR SumD SumD I39% 1 59%

Percentages are based on the total amount of Cal-Sil generated Table 2: Com parison of Transport Fractions for Coatings Time-Dependent Transport of Transport of Coatings to VC Coatings to SumD IR SumD IP2 33% 83%

IP3 7% 1 59%

Percentages are based on the total amount of coatings generated The capture of particulate debris is dependent on full coverage of the strainers by fibrous debris.

If the strainers are fully covered, the fiber acts as an effective filter, and if the strainers have clean screen area, the particulate simply recirculates or settles. Both scenarios will be discussed in detail below.

In order to credit a fully covered strainer for filtering particulate, there needs to be reasonable assurance that sufficient fiber will collect on the strainer. Several scenarios exist that have the potential to deposit a small amount of debris on the IR strainer. First, a small amount of debris could be generated; this scenario is evaluated in the clean screen section below. Secondly, debris transport to the IR sump could be significantly less during the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . For this scenario, it is reasonable to expect the transport to the VC sump to be less than the 5% used in the analysis, especially since the flow rates will be less than 1/3 of those for the IR sump (7,100 gpm or 5,400 gpm for the IR sump versus 1,350 gpm for the VC sump at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ). Finally, the transport of debris to the IR sump could be delayed. The Delayed Washdown Section demonstrates that 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months is a reasonable delay for debris washdown, and a delay of this magnitude would result in 99% transport of fibrous debris to the IR strainers. Therefore, all plausible scenarios are being covered by this evaluation.

Strainers with Clean Screen Area The prototypical array testing for the IP2 and IP3 VC sump time-dependent (after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ) debris transport cases demonstrates that significant clean screen is present due to the low debris loads. The total fiber load for the IP2 time-dependent case is 5.54 lbs (Table 3, Test G:

4.47+1.07), and the total fiber load for IP3 time dependent case is 8.02 lbs (Table 3, Test D:

3.66+4.23+0.13). The thin-bed tests required fiber amounts of 8.55 lbs (Table 3, Test F) to 15.13 lbs (Table 3, Test Al) to fully cover the screens. Without fully covered strainers, the particulate would simply recirculate (or settle), rather than collect on the screens and contribute substantially to head loss. Additionally, the prototypical testing in 2009 included tests with a

NL-10-074 Attachment 1 Page 32 of 35 wide range of particulate amounts, and the head loss results do not indicate that they are strongly dependent on particulate amounts (see Table 3). For example, Tests D and G were conducted to address the IP2 and IP3 time-dependent debris transport cases. The head loss results were essentially equal (5.81 ft versus 5.42 ft) for the two tests which included 22.34 lbs and 77.26 lbs of particulate. In summary, test results and theory indicate that any increase in particulate load is unlikely to increase the debris head loss.

I The clean screen scenario also applies to the IR sump. At first glance, it would seem that this would lead to a case where the particulate does not collect on the IR sump but then collects on the VC sump, The IR strainer surface area for IP2 and IP3 is approximately 3 times that of the VC strainer, and the debris transported to the VC sump is less than 1/3 that of the IR strainer (fine debris: 100% transport to IR vs. 5% transport to VC). Therefore, for a case where insufficient fiber is generated to cover the IR sump strainer, there would less fiber per strainer area for the VC sump which would also be insufficient to cover the screen.

Table 3: Summary of Prototypical Array Test Debris Loads and Results Test Debris Loads *Measured Head Loss Extrapolated Temp- Min. Sil-Co- Conventional Conventional Nukon Mat Wool Cal-Sil Sil Dirt WCAP Conventional & Chemical & Chemical (Ibs) (lbs) (lbs) (lbs) (lbs) (lbs) (lbs) (ft-water) (ft-water) (ft-water)

A (Thin Bed) 13.54 0.00 0.00 9.66 80.49 15.57 1.83 AI-6 (Thin Bed) 15.13 0.00 0.00 9.66 80.49 15.57 6.54 0.96 8.54 9.98 B-4 25.19 5.43 0.00 7.49 72.13 6.26 6.54 1.38 8.78 11.73 C-4 18.94 21.75 0.70 9.66 51.09 7.75 6.49 2.80 5.80 7.36 F-D-1 3.66 4.23 0.13 4.85 16.26 1.23 9.33 0.13 3.49 5.42 E-5 (Thin Bed) 14.25 0.00 0.00 9.66 80.49 15.57 6.54 1.10 18.45 21.88

F-5 (Thin Bed) 8.55 0.00 0.00 9.57 23.37 24.63 6.54 3.07 14.73 19.02 G-1 4.47 1.07 0.00 3.01 73.30 0.95 6.84 0.10 3.45 5.81

Measured Head Losses are scaled to 400 gpm or 155 gpm (which ever is applicable) and 70TF Fully Covered Strainers The prototypical array testing in 2009 also demonstrates that the full debris load on the IR sump is a very effective filter for particulate debris as evidenced by sharp decreases in turbidity after each debris addition. To evaluate this information, the number of pool turnovers required to filter the majority of particulate is determined. Pool turnovers is the appropriate metric to compare the test with the plant because it accounts for the differences in scale. Since very little particulate is required to increase turbidity, it is assumed that turbidity values less than 100 NTUs indicate that very little (<5%) particulate remains in the recirculating fluid. This is reasonable considering that the 25% of particulate added in the first debris addition for Test C increased the turbidity to over 600 NTUs (see Figure 2). For the 4 debris additions in Test C, the time to reduce the turbidity to 100 NTUs ranged from 36 minutes to 14 minutes (see Figure 2). This corresponds to an average of 3.7 pool turnovers, while the minimum number of pool turnovers in the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months for the Plant is 5.3 and 5.0 for IP2 and IP3, respectively. These pool turnover values are based on the minimum flow rates used in Equations 2 - 4 in the delayed washdown section, and the actual pool turnovers in the Plant could be two or more

NL-10-074 Attachment 1 Page 33 of 35 times greater. Therefore, a fully covered strainer in the Plant is expected to be an effective filter for particulate in the first 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months .

_7 7777 T.......... T7T .... .

47400 I-300 ------------

2006 1T00 ii0.

iA 22:48 0:00* 1:12 2.24 :3 .48 6:00 7:12 Time Figure 2: Turbidity Data for Test C Debris Movement after Securina Pumi (RAI 7, Item d)

The updated Generic Letter 2004-02 Response provided to the NRC included a practical explanation as to why debris would not be resuspended when the IR pumps are secured. The reasons included:

i very little (if any) debris bed movement when pumps were secured during testing,

the presence of check valves in close proximity to the recirculation pumps,

agglomeration of debris observed for all prototypical tests,

the strainers are situated in a pit,

" the strainer pit is located in a room that is not in the flow path,

the flow paths would not promote transport from the IR sump to the VC sump.

However, for the purposes of this evaluation, a conservative non-zero value will be considered for material released from the IR sump debris bed. The Indian Point IR strainer design consists of an array where each column has 9 Top Hats. Consequently, approximately 11% (1/9*100%)

of the strainer surface area is located in the top row. It is assumed that only the strainer surface area in the top row would be susceptible to debris movement and that approximately 20% of the fine fiberglass collected on these screens could be released and transported when the IR

NL-10-074 Attachment 1 Page 34 of 35 pumps are secured. This release is equivalent to 2.2% of the total fine fiberglass debris, which will be used to evaluate total transport in Table 4. The small volume of particulate released from the'top row of the IR sump after the IR pumps are secured will not appreciably contribute to head loss (see Strainer Bypass Section) and will not be further considered by this evaluation.

Pool-fill Transport to VC Sump (RAI 7, Item e)

The entrance to the IR and VC sump pits are located on the 46-foot elevation. The flow of water from a LOCA is first directed into the reactor cavity, below the 46-foot elevation, where debris will most likely settle. The water must travel up through the in-core instrumentation tunnel, through the crane wall, along the annulus, and to the sumps. Therefore, the reactor cavity and tunnel are filled first during the pool fill-up phase of an accident. The volume of the VC sump pit (conservatively neglecting the strainer and strainer support structures) is 232 ft3 (5'6"x5'6"x7'8")

for IP2 and 496 ft3 (8'x8'x7'9") for IP3. Although some fine debris could be transported to the VC sump during the pool fill-up period, the pit volume is small compared to the minimum water volume. At the beginning of recirculation, the total pool volume for an LBLOCA is at least 35,080.32 ft3 and 29,527.72 ft3 for IP2 and IP3, respectively. Consequently, the ratio of VC sump pit to minimum water level is 0.7% and 1.7% for IP2 and IP3, respectively. A transport value of 1.7% is used for conservatism in calculating the total debris transport. An additional conservatism not included in this transport value is the fact that the VC sump and IR sump cavities would be filled simultaneously early in the event. At this time, a large percentage of the fine debris would be in upper containment, so the fraction of fines in the pool that are available to transport to the VC sump and IR sump during the pool fill-up phase would actually be less than 1.7%.

Total Time-Dependent Transport Fraction The sections above evaluate each of the five potential non-conservatisms identified by the NRC in RAI 7 and estimate a conservative transport fraction for each individual item. For fine debris, the cumulative effect of the applicable items is estimated to be the sum of the four individual items. As shown in Table 4, the total time-dependent debris transport fraction for fine debris, using the conservative methodology presented above, is 4.8%. Therefore, the 5% transport determined for fine debris in the Debris Transport Calculations is considered bounding. In addition to the 5% transport of fine debris, the small and large fiberglass pieces are subject to erosion. Erosion is evaluated above and determined to be less than the 1% utilized in the Debris Transport Calculations.

Table 4: Total Time-Dependent Debris Transport Fraction for Fine Debris Item Transport Fraction (%)

Delayed Washdown 0.9 Strainer Bypass 0 Debris Movement after Securing Pump 2.2 Pool-fill Transport to VC Sump 1.7 Total 4.8

NL-10-074 Attachment 1 Page 35 of 35 Conclusions The transport fraction determined in this evaluation contains considerable conservatism considering that the number of pool turnovers is calculated using very conservative inputs of minimum flow rates and a maximum pool volume. Because the pool volume will be lower and the flow rate will be higher, the fraction of fine debris remaining in the pool will be lower than 4.8%. Additionally, the recirculation transport for the VC sump is based on maximum flow rates for the beginning of recirculation, 3,700 gpm for the VC pumps. The alignment is switched to the HHSI pumps by 6.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months , and the flow rate is reduced to approximately one-third (1,350 gpm or less) for the remainder of the 30 day mission time.

After applying a conservative methodology for each of the 5 RAI items, the transport fraction for fine fibrous debris is less than the values used in the Debris Transport Calculations. Indian Point's prototypical array testing demonstrates that the majority of particulate will collect on the IR sump prior to 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months . However, if the particulate amount is greater than that determined in the analysis, test data indicates that the head loss would not be significantly affected due to the presence of clean screen area. Finally, because of the low velocities around non-transporting debris at Indian Point, test data indicates that the erosion fraction will be less than the 1% used for post-24 hours in the Debris Transport Calculations. Therefore, the existing analysis, which includes NPSH margins of 2.51 ft for IP2 and 4.41 ft for IP3, is justified as bounding.

NL-10-074, Response to Request for Additional Information Regarding Generic Letter 2004-02

Contents

Text

SUBJECT:

REFERENCES:

Dear Sir or Madam:

Attachment:

Navigation menu

NL-10-074, Response to Request for Additional Information Regarding Generic Letter 2004-02

Text

SUBJECT:

REFERENCES:

Dear Sir or Madam:

Attachment:

Navigation menu

Search