Text

Response to Request for Additional Information Re Fukushima Flood Hazard Reanalysis Report, Dated April 25, 2014
	ML16070A290
Person / Time
Site:	Oconee
Issue date:	04/25/2014
From:	Batson S; Duke Energy Carolinas
To:	; Document Control Desk, Office of Nuclear Reactor Regulation
Shared Package
ML16070A295	List: ML16070A284; ML16070A287; ML16070A288; ML16070A290; ML16272A216; ML16272A217; ML16272A218; ML16272A220; ONS-2015-028, Supplemental Information Re NRC 2008 and 2012 Request for Information Pursuant to 10 CFR 50.54(f) Pertaining to External Flooding, Dated March 6, 2015 (Redacted);
References
	FOIA/PA-2016-0071
	Download: ML16070A290 (120)
	v • d • e

ML14120A431 Scott L.Batson DUKE Oconee Nuc fear Station SENERGY ueEeg ONOlVP I 7800 Roche sterHwy Seneca, SC 29672 ONS-204-0630: 864.873.3274 ONS- 14063f. 864.873. 4208 Scotf.Batson@duke-energy.com April 25, 2014 ATTN: Document Control Desk 10 CFR 50.54(f)

U.S. Nuclear Regulatory Commission Washington, D C 20555 Duke Energy Carolinas, LLC (Duke Energy)

Oconee Nuclear Station, Units 1, 2 and 3 Docket Numbers 50-269, 50-270, 50-287 Renewed License Numbers DPR-38, DPR-47, and DPR-55

Subject:

ONS Response to Request for Additional Information Regarding Fukushima Flood Hazard Reanalysis Report

References:

1. NRC Letter, Request for lnformation Pursuantto Title 10 of the Code of Federal Regulations 50.54(f) Regarding Recommendations 2.1, 2.3, and 9.3, of the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi Accident, dated March 12, 2012, (ADAMS Accession No. ML12053A340)

2. Duke Energy Letter, Flood HazardReevaluation Report in response to NRC letter;,

"Request for In formation Pursuantto Title 10 of the Code of FederalRegulations 50.54(f)

Regarding Recommendations 2.1, 2.3, and 9.3, of the Near-Term Task Force Review of Insights from the Fukushima Dai-ichi Accident, "dated March 12, 2013

3. NRC Letter, Oconee Nuclear Station, Units 1, 2, and 3, Request for Additional Information Regarding Fukushima Lessons Learned Flood Hazard Reevalution Report (TAC Nos. MFI012, MF1013, AND MF1014), dated March 20, 2014, (ADAMS Accession No. MLML14064A591)

Ladies and Gentlemen, Oconee Nuclear Station (ONS) submitted its Flood Hazard Reevaluation Report (HRR)

(Reference 2) to the NRC on March 12, 2013 Pursuant to the NRC's 10 CFR 50.54(f) letter (Reference 1).

By letter dated March 20, 2014 (Reference 3), the NRC staff requested additional information regarding the Flood HRR. Answering several of the RAl questions will require Duke Energy to engage industry subject matter experts for support. The current demand on seismic, geological and hydrological expertise in the industry and the complex aspects of certain RAIs, will result in a delay to respond to RAl 1 and 14. The response to the balance of the RAls are contained in the enclosure to this letter. The enclosure also provides the dates for which Duke Energy plans to provide the responses to RAI 1 and 14.

fL.. ... . . .. J.... ..1 ...... . ~ni~r rzmzin ~ lthoo ir -iiu1'~~

ug4_.

Duke Energy Response to RAl Regarding Flood HRR April25, 2014 Page 2 This letter does not create nor revise any Regulatory Commitments.

This letter contains Security Sensitive information and is requested to be Withheld from public disclosure in accordance With i0OCFR 2.390(d).

Should you have any questions concerning this letter, or require additional information, please contact David Haile at (864) 873-4742.

I dectare under penalty of perjury that the foregoing is true and correct. Executed on April 25, 2014.

Sincerely, Scott L. Batson Vice President Oconee Nuclear Station Enclo.sure -

1. Oconee Nuclear Station Units 1, 2, and 3, Duke Energy response to RAI regarding Fukushima Flood Hazard Reevaluation Report (HRR)

,~, ~j~rn . 1 ~ ~ t U.. T - Formution 'Yithht~"f '"-

ru'D 2 W~(i) u;LA ,n =;' u; .. . . .. .. ~ ~... W....... Z.... L......... . ,, cni.

Duke Energy 2014 Response to RAI Regarding Flood HRR April 25, Page 3 cc:

Mr. Victor McCree, Regional Administrator U.S. Nuclear Regulatory Commission - Region II Marquis One Tower.

245 Peachtree Center Ave., NE Suite 1200 Atlanta, Georgia 30303-1257 Mr. Eric Leeds, Director, Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission One White Flint North, Mailstop 13-H16M 11555 Rockville Pike Rockville, MD 20852-2738 Mr. James R. Hall, Project Manager (ONS)

(by electronic mail only)

U.S. Nuclear Regulatory Commission 11555 Rockville Pike Mail Stop O-8B1 Rockville, MD 20852 Mr. Eddy Crowe NRC Senior Resident Inspector Oconee Nuclear Station

Duke Energy Response to RAI Regarding Flood HRR April 25, 2014 Page 4 bxc:

T.P. Gillespie (EC07H)

T.D. Ray (ON01 VP)

R.H. Guy (ON01 VP)

T.L. Patterson (ON01 VP)

D. A. Baxter (ON03PC)

C.T. Dunton (ON01 El)

0. C. Jones (ON01 El)

D. M. Hubbard (ON03PC)

C.J. Wasik (ON03RC)

D.C. Haile (ON03RC)

M.C. Nolan (EC05P)

J. A. Olivier (EC2ZF)

C.J. Thomas (ECO1T)

D.H. Llewellyn (EC09E)

G.D. Robison (EC09E)

P. F. Guill (EC01T)

ONS Master File (NON2DM, File OS 801.01)

ELL (EC2ZF)

Enclosure I Oconee Nuclear Station Units 1, 2, and 3 Duke Energy response to RAI regarding Fukushima Flood Hazard Reevaluation Report (HRR) w=

ll!

.v

.~I i... -

oR iN BII*R I

-i

_ = no iiL ui uu

- Iv In . -I

...oon rtijmp:pi u~ tno :nuonmorno. trio ~nuiou~ro in ~nnon:roiIou I I ........... I -- R LL-- i 6-- -- L .... L--

Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pg Page 2

Background

On March 12, 2012, the U.S. Nuclear Regulatory Commission (NRC) issued a ,Request for Information pursuant to Title 10 of the Code of Federal Regulations (10 CFR) 50.54(f) regarding Recommendations 2.1, 2.3, and 9.3, of the Near-Term Task Force review of insights from the Fukushima Dai-ichi Accident" (ADAMS Acc¢ession No. ML12053A340) to all power reactor licensees and holders of construction Permits in active or deferred status. By letter dated March 1:2, 2013, (ADAMS Accession No. ML13079A227), Duke Energy Carolinas, LLC (Duke Energy, the licensee) submitted the "Oconee Flood Hazard Reevaluation Report," (FHRR) dated March 6, 2013,-~(ADAMS Accession No. ML13240A016), for the Oconee Nuclear Station, Units 1, 2, and 3 (ONS).

The NRC staff has determined that it.needs additional information in order to complete its review and has submitted 15 questions (RAIs) to Duke energy. The responses to RAIs 2- 13, and 15 is provided below.

Duke Energy is engaging multiple industry subject matter experts in support of answering PAls I and 14. Due to the current demand for seismic, geological, and hydrological expertise in the industry and the complexity of the PAls, the subject matter experts have determined that approximately 1 month is needed to achieve an appropriate response for RAIs Ic and 14, and approximately 6 months is needed :tO for an appropriate response to RAls Ila and l~b..

Duke Energy plans to provide responses to the remaining RAls as follows:

Response to RAIs Ic and 14 - by June 13, 2014

Response to PAls Iaand lb - by November 7, 2014 Responses to RAIs 2 - 13, and 15:

RAI-2: Local intense Preclipitation and Associated site Drainager(Model Documentation and InputlOutput Files)

The licensee performed the. local intense precipitation flooding analysis using the modeling software Info Works CS (IWCS). A report provided in the electronic reading room (ERR) states that the model construction is described in a separate report, dated November 2012 and titled "ONS Local Flooding Analysis Hydraulic Modeling Report Yard and Roof Drainage Local Flooding, Current Licensing Basis," This November 2012 report was not provided in the ERR. Therefore, the NRC staff requests the following:

a) A detailed description of how the IWCS software was implemented for this site, including:

1. Identification of the specific IWCS modules that were used and description of the methods implemented by these modules (e.g., 2-D depth averaged, etc.)

2. Description of treatment of model's boundary conditions .(inflow, outflow, T

'rmiii pn f hu - . tn j, jjj'li :..

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pg 3 Page surface boundary conditions), initial conditions, and convergence criteria.

3. The NRC staff also requests electronic input and output files for this model including the AVI files listed In Appendix D of the technical document provided in the ERR.

Rese~onse to RAI 2a Subpart 1 and r2:-A report *titled "ONS LOcal Flooding Analysis. Hydraulic Mo*deling RePort,

Yard and Roof Drainage Local Flooding, Current Licensing Basis," dated November 2012 is provided as Attachment A and is also available ih the Electronic Reading *Room (ERR), The/

rePort contains information specific to the, IWCS model construction and treatment of boundary conditiOns. The electronic AVI files listed as attachments to that~report are provided on DVD as Attachment B to this document.

Subpart 3,, The ,requested data, AVls. and model inputs/outputs, are provided as electronic files on DVD (Attachments B and C). Attachment B contains the AVIs and model Input/Output data associated with Current Licensing Basis. Attachment C contains the AVIs and model Input/Output data associated with Beyond Licensing Basis.

b) Information on the methods used to calibrate the IWCS model against known conditions and solutions (I.e.,, the results of this calibration, but not the methods, are provided in Appendix C of the technical report provided in the reading room).

Response to RAI 2b The IWOS model was calibrated against known rainfall events and, established elevations provided by Duke Energy at several locations included within the IWCS model. This was

,done by running the IWCS model with historicrainfall data and sending resulting elevations to Duke Energy for comparison. If elevations were too high or too low modifications were made to the IWCS model for more definition in the two-dimensional (2-D) grid and the process was repeated until calibr'ation results were considered to be satisfactory by both HDR Engineering, Inc. of the Carolinas (HDR) and Duke Energy., The storms that were calibrated .are:,

. Historic Rainfall 713011991

Historic Rainfall 8/20/1995

Historic Rainfall 9/11/199.5 - Historic Rainfall 9/22/2003

Historic Rainfall 06/28/2006 c) Information on the approaches used to assure conservation ,of mass associated with the Solution transfer from the roof to the site grid and from rainfall to runoff in the IWCS model.

Response to RAJ 2c The 2-D engine within IWCS includes conservati~on of mass and momentum within its solution*.

Each 2-DSimulation is accompanied by a Volume Balance Report in the simulation Log file,.

fl....O A f-l.-l*t-tl.)'. A iElt,,,.UR 1 _ly *. A....rn,.l...-

IUiI~lIr*lIU¥£I1LtL~i . 1 ci.ii=.

Encloure Duke Energy I Response to RAI Regarding Flood HRR Pg Page 4 As part of the model development and Quality Control (QC) process, HDR reviewed the Volume Balance Reports for each simulation to ensure that the mass error balance (%) was within an acceptable range (i.e., essentially zero). An example of a Volume Balance Report is provided below:

Volume Balance Report 2-0 Zone Name: ONS 2-D Basin Initial Volume (in3 ): 0.0000 Net Inflow (ms): 104893.5533 Inflow (mn3 ): 121625.4200 Total Volume in the surface (in3 ): 227543.6455 Volume in the 2-D zone (in3 ): 10586.8866 Volume out of the 2-D zone (in 3): 216956.7589 Rain volume in the 2-D zone (mn3 ): 122650.0922 Volume lost in the 2-D zone (in 3 ): 0.0000 Mass error balance (%): 0.0000 Effective area (ha): 19.8212 Flooded area at the end of the simulation (ha): 10.0110 Maximum flooded area (ha): 18.7074 (Volume Balance Report from Current Yard Drainage Blocked Simulation Log file) d) Clarify what physical aspect is modeled using the 1-D method for IWCS Response to RAI 2d Physical aspects of the ONS site are specified in the "ONS Local Flooding Analysis Hydraulic Modeling Report, Yard and Roof Drainage Local Flooding, Current Licensing Basis," dated November 2012 and located in the ERR.

The three physical aspects that were modeled using the one-dimensional (1-D) methods within IWCS are as follows:

1, Offsite hydrology and hydraulics (Green features in Figure 2-1)

Offsite hydrology uses the Soil Conservation Service Curve Number (SCS CN) methodology to route flow to specific locations. These locations are either a 1-0 hydraulic node or a loading point on the 2-D mesh as appropriate. Where the hydrology loading point is a 1-0 hydraulic node, this node is linked via conduits that represent open channels, culverts, or roadways to route the flow towards a natural outlet or a 2-0 mesh element based on the hydraulics. For example: regarding the west side of South Carolina Highway 130, the southern basin is loaded into a node on the upstream side of the culvert under the highway with a split to an open channel. When the flow starts to exceed the capacity of the culvert, the excess flow is directed down the open channel as the invert is at a higher level than the culvert. Flow through the culvert passes through open channels on the downstream side to the 2-0 mesh.

TI~._.

i Jtt.::I. ...... L. Cr *.. .........- tl*m : *. x . ia,,, . , w*j.I j i

r. ... ""~; "" L

*... v*... . ... ... iu cr' . r*390(=*

, I;.#v LII I *l U i

U~ *I ,*U zT .IUI;

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pg Page 5

2. Roof hydrology (Pink and lime features in Figure 2-1)

Each onsite roof uses the Storm Water Management Model (SWVMM) rainfall-runoff methodology for transforming rainfall into runoff. There are three roof categories:-

Parapeted roof with roof drainage system connected to ONS Yard (Yard) drainage system (Lime). Rainfall storage on roof occurs if the capacity of inlets or pipes is exceeded until the crest of the parapet is overtopped and roof Overflows to 2-D mesh.

Roof with gutter and downspouts to Yard (Pink.). Rainfall runoff is directed to the gutter that drains via downspouts to 2-D mesh or overflows directly to 2-0 mesh.

Roof with direct runoff to Yard (Pink). Rainfall runoff is directed to,2-D mesh.

3. Piped system (Blue features in Figure 2-1.)

The piped system iscomprised of the Roof land Yard Drainage. Pipes are connected to*

either a 1-D roof element or a 2-D mesh element via a designated Catch-basin Node.,

Figure 2-1 below, shows the l1oD model elements.

vf ll,l*jl,Jlu *l Villa i*ulurllllk, Ulla*l*ll*UJ I* UJlUl*l j ....

JI U L%.f 3. 3 .,

,.,: :1 J~ tL Att....L....L. tL.. ~ Z ..... ,4. JLJ.

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pg Page 6 Figure 2-1 IWCS 1-D MODEL ELEMENTS U IIr nflintuInuIrlIL xiU r.. IIIUU --- >I----

y 10 Cfl :.Pe(d)

- ...I i'lL .t1LLzturti, thi ~h~- ~

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pg Page 7 RAi-3: Local Intense Precipitation and Associated Site Drainaae (Choice of Methods and Technical Rationale)

The NRC staff requests information on the watershed delineation, the role of the various drainage structures, and onsite local intense precipitation flood hazard modeling.

Specifically, the staff requests the following:

a) Technical rationale for the selection of the delineation of the boundary between the offsite sub-basins and the onsite area and reasoning for the two different treatments, i.e., Soil Conservation Service (SCS) runoff method for the offsite sub-basins and two-dimensional modeling for the onsite sub-basins.

Repose toRAI 3a Watershed delineation was performed based on topographic divides of each* external contributing runoff area. This model approach was taken since the external contributing areas have Similar physiographic features (e.g., Steep slopes with defined channels) and precise flooding elevations were not of interest (located outside of security area) but total contributing volume was. The SOS method was selected for these areas since it is a standard of industry methodology. Two-dimensional (2-D) modeling was performed in the ONS Yard (Yard) since there was little topographic relief (i.e., relatively flat parking lot runoff surface) with little to no channeling resulting in unconfined flow and 2-D modeling was appropriate to simulate model flood elevations in this area. The connections (i.e., location of channel flow contributing to the 2-D zone) of the external contributing areas modeled in one-dimension (l-D) were veified using a full 2-D model. However due to the model Size, the 1-D/2-D modeling approach was facilitated to allow for significantly faster simulation runtimes while maintaining the required level of accuracy. Current model run times are in excess of approximately 20 hours2.314815e-4 days 0.00556 hours 3.306878e-5 weeks 7.61e-6 months .

Buildings were modeled with the :I-D method rather than the 2-b method since the hydraulics describing them behave in a I1-D manner. The hydraulic performance of the roof drainage system is best described as 1-D. due to the following characteristics"

Rainfall collecting on roofs with parapets and draining into the drainage system or overtopping the parapets and spilling onto the 2-D surface; u Rainfall running directly off a roof without parapets or via a combination of downspouts and overflowing gutters onto the 2-D surface.

b) Additional information on the modeling approach for local intense precipitation (LIP) and the conservatism of the analysis. More specifically, the NRC staff requests the technical rationale and documentation for the SCS analysis and the SCS curve number (ON) values assigned to each of the offsite sub-basins. The NRC staff also requests technical rationale for not treating the offsite sub-basins as impervious and the rationales for selecting CN values and determining travel times.

"1 n jLLUa t i u WI II t III C u I, *

,L:I u , u1 ,

I L UuI1 1,LU 1 U.o. -- - ,U

.A U - yy , a. n g . . . .. i . l,

,1E I ,,III Ji.,,,u. ....

. ,..J .

p,*.. .. ; . 1.. R .J9~.. '^...-^"..."*"l d

Enclosure I Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pg Page 8 Response to RAI 3b An external peer review was performed by an independent consultant (Wilson Engineering),

which verifies that the SCS curve number values and travel times used in the model meet current-industry standards. The supporting Quality Control (QO) documentation is titled-

"Independent Technical Review of Modeling of PMP evenlts for Oconee Nuclear Station" by Wilson Engineering and dated November 30, 2012 and has been posted to the ERR for NRC review.

c) Additional detailed representation of the location and type of drainage structures and anyr wall structures that may exist.

Response .to RAI 3c In addition to the terrain adjustment detailed in the FHRR, the 2-D mesh included, the following elements to help define the hydraulically relevant surface features:

Break lines that act as soft break lines for determining, mesh development (Green features in Figure'3-!);

Porous walls or hard break lines that act as hard edges in 2-D mesh element creation (Pink features in Figure 3-1);

.Polygons set as voids within the 2-D mesh for areas of no flow - i.e. buildings =(Grey features in Figure 3-1);

Mesh areas that are used to set specific ground elevations (Not shown in figur'e).

Each catchbasin has a mesh polygon to ensure the correct elevation of the 2-D mesh .element for the i -D / 2-D. interaction between the surface and pipe system and for flow in general across the Yard. These c;atchbasin polygons have their areas set to the minimum defined mesh element area.

Figure 3-1 below, shows the line locations for defining the 2-D model mesh.

.,~~~, I ~~~ ~1 :: 't..2I;..I'I.-.O(l. 't~I z ut :zz-r-n I LI (13'D - oI2flA

Enclosure I Duke Energy Response to RAI Regarding Flood HRR Page 9 FIGURE 3-1 IWCS 2-D MODEL MESH

SUMMARY

I Mb fl ~ U~hA l ,tttpti"I. . ... ..

vr

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae1 Page10 d) Technical rationale to explain how the overtopping of any drainage structure was handled in the modeling process for the offsite and onsite areas.

Response to RAI 3d" The offsite drainage is-modeled as i-D elements as it falls into the guidelines defined for 1-D modeling, mostly on well-defined channels and flow paths; For areas that flow directly onto-the 2-0 mesh, contours Were used to determine locations where flow naturally accumulated and flowed onto the Yard and set as inflow locations from the 1-D hydrology to the 2-D mesh.

An example of this is shown on Figure 3-1. The offsite areas not .directly connected to the 2-D Yard were hydraulically routed along their natural flow paths. These included paths .flowing through various culverts and down open channels. Spill elements were added to culverts to allow for overtopping of the culvert when .capacity was exceeded. This included the roadway and possible side or lateral spilling of the flow to other flow paths. For example; the natural flow path is down the storm ditch on the western side of South CarOlina Highway 130 northwards. The flow then passes through a culvert to the east under South Carolina.

Highway 130 and continues down a shallow drainage ditch to a culvert under the sider access.

roadway and off the site. At each culvert, an overtopping weir was. added for the roadway bed. In.addition to the last culvert in this system going offsite, open channels representing, the

.car parking area and the access roadway .were added to allow overflow routes for the flow when the culvert capacity was exceeded. These flow paths take the overflow from the storm

.system to the Yard.

2-D elements Overtopping of the 2-D elements is handled by the 2-D hydraulic simulation engine. As described in the "ONS Local Flooding Analysis Hydraulic'Modeling Report, Yard and Roof Drainage-Local Flooding, Current Licensing .Basis," dated November 2012, the guard rails are modeled as 2.25 foot high walls with a 1.-D opening added at .regular intervals to simUlate the flow underneath the structure. Flow that backs up high enough on the upstream side of a guard rail will spill over it once the depth reaches 2.25 feet based on the simulation engine hydraulics. Raised curbs have been defined by adding lines to force the 2-0 meshing process to create elements along these features. A similar process was used for cable duct and trench lids or other significant above-ground structure other than buildings. The edges of the structure'were either modeled by using a line along the edge Or a meshing polygon to ensure they were included in the 2-D mesh. Figure 3-1 .shows a summary of the lines used in the 2-0 meshing process..

- -- -, j 1

I pE* , ,

w , 1 ...... a* j ,.-.A. . I.r. - . * ,* .__ o,'.: __. *.. - cj-L_. iJ * .... _.

S...II,,.i 11 .lim adil .... tipl 3

. 5 n.. m..pincllra 1*

fAS O Y-.*"n'aS ic *inO..itlrd~hli*Ol

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae1 Page 11 RAI-4: Local Intense Precipitation and Associated Site Drainage (.Choice of Methods and Technical Rationale)

The NRC staff requests additional clarification on how the analysis for site drainage and flow at the site were performed in the IWCS and SCS models. More specifically, the NRC staff requests information on:

a) The termination point for the drainage that leaves the Yard area, and the assumptions associated with the conditions at the termination Point and its effect on the operability of the underground drainage system.

Response to RAI 4a Termination points were developed based on site information (i.e., physical characteristics bf conveyance) and boundary conditions were set to free discharge.

The ONS Yard (Yard) drainage system exits into an open channel. The outlet pipes are steep and have security grills/grates at the downstream opening. Details of these grates were provided and input into the model accordingly as screen elements. 'To ensure that the flow influence associated with the model flow boundary condition was far enough downstream that the possible flow solution influence it produced would not affect the results at the study area, 11!6 feet of open channel was modeled downstream of the screens to move the normal flow boundary conditions away from the pipe security screens.

b) Whether and how the coupled model considers drainage, flow from the two-dimensional grid to the sub-basins, including identification of which sub-basins flow onto the site and which receive drainage from the site. More specifically, technical discussion on the possibility of backflow from onsite to the sub-basins.

Response to RAI 4b Flow entering from off-site areas is generally from steep terrain transitioning to flat terrain at the intersection with the two-dimensional (2-D) mesh that was used to modelfiow in the Yard.

Back~flow between the 2-D mesh node and the one-dimensional (i-D) channel flow is not simulated in the model. Back flow influences, between the !-D and 2-D boundary nodes would be considered less conservative than what was-modeled as it would slow the flaw of the water to the Yard area where the buildings and systems; structures, and components (SSCs) are locatedi. Backwater elevation increases in the 1-D flow routing channels is not Considered significant as it does not border critical buildings or equipment at ONS. The St. Venant equation is used to account for off-site drainage that is routed, through 1-D hydraulics.

c) How the drainage flow leaving the modeled sub-basin, "Offsite 5" in Figure A-7-A is accounted for in the model. More specifically, provide description of backflow or backwater effects with due consideration for the topographic configuration of "Offsite 5" and outside of the sub-basins.

IT;,* 4 fi uU WLthUu ~© I lifl,-,,,l,,* 11,13 . ... .u..

liliu.. ....... .. : U... .

,,*~lt.. o; 1u CFI 13,?.O(,I,). *u, ,*,,,uu, u, *;,- ~ ,.uu,

Enclosure ! Encloure I Duke Energy Response ,to RAI Regarding Flood HRR Pae1 Page 12 Response to RAI 4c Drainage flow in the 1-D sub-basins (e.g., Offsite 5, etc.) is modeled~as follows:

Hydrology is based on the Soil Conservation Service iCurve Number (SCS ON) and channel routing is computed as a fully dynamic node - link. The Sti venant eqUation is used for hydraulic routing through .open channels, culverts, and weirs for roadwaY overtopping. One-dimensional flow paths have been. accounted for Within the IWCS model. These elements are=

shown in Figure 2-1 from the response to RAI-2, The lnfoWorks CS (IWOS) model process.

includes the-following steps:

1. Sub-basin runoff is calculated and applied to the appropriate 1-D nodes (e.g., channels, culverts, weirs, etc., relating, to ithe delineated area).

2. Flow passes through Culvert ifcapacity allows or backs up until it can bypass either down an attached defined oPen channel or over a roadway spill .(weir) based on hydraulics.

3. Flow routes down loD elements .until it either enters, the 2-D mesh, leaves the system, at a culvert outlet, or overtops the roadway based on. i-D hydraulics.

d) Site terrain and .drainage patterns (by providing a more. legible and better resolution copy of .the contour drawing initially received, including legible elevation numbers).

Response to RAI 4d An electronic version of the contour drawing from the ERR is. provided on CD (Attachment D).

Viewing an electronic version will produce the resolution necessary for contour labels and.

other details to be legible.

RAI-5: Local Intense Precipitation and Associated Site Drainage (Choice of Methods and Technical Rationale)

The report on .the IWCS flow model results does not include a discussion on impacts of velocity distribution in the Yard. The-NRC staff requests a technical discussion regarding the effects of flow velocities, hydrodynamic forces, and any debris loading in and around the Yard .site.

Response to RAt 5 The LIP model results incliuded in the document titled "ONS Local Flooding Analysis Hydraulic Modeling Report, Yard and Roof Drainage Local Flooding, Current .Licensing Basis," dated November 2012, contains velocity results kin the yard. iThe Yard terrain* surroundring the systems, structures, and components (SSCS) is relatively flat (parking lot) with no sloped drainage channels-conveying flows across the Yard and past SSCs. The flat-terrain was.

modeled using a two-dimensional (2-D) mesh to simulate ponding and flow velocities. One-dimensional (ID) runoff basins channelized flow transitions to 2-D mesh where channel flow up:o.. r-emovairorue Atuwcn,-n.*, N L. ,,iVU*.I*La UJ~~~

Ih*(. uI;l.-.

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood *HRR Pae1 Page 13 quickly loses velocity as flow spreads and terrain flattens to minimal drainage slopes across the Yard. Due to the location of the critical buildings and equipment toward the interior of the 2-0 mesh and away from the boundary 1-D runoff areas, 2-D modeling simulation results for velocity in the Yard as generally low around buildings and equipment except at the extreme north end of the Yard bordering the site access roadway and 1-D hillside and at the south end of the Yard between the Interim Radwaste/Shredder buildings and the toe of the 0conee Intake Dike.

Debris and hydrodynamic loading impacts on SSCs in the Yard were not evaluated for the FHRR. Further discussion of debris and hydrodynamic loading can be found in the response to RAI #15 RAI-6: Local Intense Precipitation and Associated Site Drainage (DOcumentation)

Background:

Definitions for design basis and Current licensing basis can be found in 10 CFR 50.2 and 10 CFR 54.3, respectively. The NTT[F Recommendation 2.1 response Flood Hazard Reevaluation* Report (FHRR) Section 3 Table 11 presents the current design, basis.

flood elevation due to a local intense precipitation as 798.17 ft MSL. The NTTF Recommendation 2.3 response walkdown report does not identify 798.17 ft MSL as the design-basis elevation. The NRC staff also compared the information in the current revision 22 of the Oconee Nuclear Station Updated Final Safety, Analysis Report (UFSAR),

and the information is not available in the UFSAR. The comparison between the design basis and the reevaluated hazard is key for determining which hazards, ifany, should be evaluated in the Integrated Assessment Report. -

Request: The licensee is requested to provide a clarification regarding the apparent discrepancy between the FHRR (NTTF Recommendation 2.1 response) and the Walkdown Report (NTIF Recommendation 2.3 response) with respect to the design-basis flood elevation.

Response to RAI 6 At the time .the NTTF Recommendation 2.3 report was written, no approved calculation had ever been performed at Oconee to. determine a flooding *depth as a ,result from the local design basis rain fall event. This issue is described in the 2.3 report section 5.1 .1.1: "A new [vendor]

calculation for the Licensing Basis Probable Maximum Precipitation (PMP) which estimates the local flood heights around the ONS site is in the draft stage. However, at the time of compiling the NTTF Recommendation 2.3 report, the [vendor] calculation was not Complete and is not incorporated into the Current Licensing Basis (CLB). (Note: Mitigating actions for the preliminary calculation results'are discussed in Section 5.2 of the 2.3 report."

By the time the FHRR report was written in March 2013, the [vendor] *local PMP calculation had been checked, approved, and peer reviewed. The timing and completion of the local PMP-flooding depth analysis is the reason the flooding, depths are described in one report and not,the other. The draft version of the [vendor] analysis was used to determine interim actions and was later validated by the completed vendor calculation.

V=rj -- m'- i :lTnl L "-

U :..z::t

.. 11 ItI:,.......Il

Enclosure I Encloure I Duke Energy ResponSe to RAI Regarding FloodHRR Pae1 Page14 RAI-7: Streams and Rivers (Choice of Methods and Technical Rationale)

The FHRR Section 2.2.2 describes rainfall amount, duration, and location of the storm only for the Jocassee reservoir. The NRC staff requests clarificatiorn On the rainfall amount, duration, and location of storm for the Keowee reservoir. In addition, the staff requests Clarification of the maximum water surface elevation of 809.4 ft identified in the 1966 study in Section I.E of the report, "Report on theAnalysis to Determine the PMF -

March 29, 1995," and how it compares with the value of 808.0 ft in FHRR Tables i and 1!.

Response to RAI 7 The maximum reservoir elevation (based on the site's design basis PMP event) for the Keowee Reservoir was reported as 808.0 ft in Table i and' Table 11 of the FHRR as taken from ithe.

UFSAR, Section 2.4.2.2. Section 2.4.2.2 further states: "Studies were also made to evaluate effects on reservoirs and, spillways of maximum hypothetical precipitation Occurring over the entire respective drainage areas. This rainfall was estimated to be 26.6 inches within a 48-hour*

period. Unit hydrographs were prepared based on a distribution* in time of the storms of October 4-6, 1964, for Jocassee and August !3-15, 1.940, for Keowee." Results are summarized as follows:

Keowee Jocassee 147,800 70,500 Maximum spiliway discharge, cfs 808.0 1,114.6 .Maximum reservoir elevation 7.0 ft. 10.4 ft. Freeboard below topof dam The maximum reservoir elevation (based on the Federal Energy Regulatory Commission

[FERCJ guideline probable maximum flood [PMF]) for the Keowee Reservoir was reported as 808.9 ft. in section 2.2.1! of the FHRR. The maximum reservoir elevation value is found in the document titled, "Report on the AnalYSiS to Determine the PMF March 29, 1995.'" The F.ERC PMF determination document is used as the basis calculation for the document "*Keowee Development (FERC #2503-0.1) Supporting Technical Information (STI), Revision No. 1,1" dated January 2012. STI Section 6.2.6 states: "The PMF peak reservoir (headwater) is approximately 808.9 ft. msr".

The March 1995 FERC PMF determination report ("'Report on the Analysis to *Determine the PMF") references (Section 1I.E) the original 1966 FERC license filed PMF study and the maximum reservoir elevation as 809.4 ft. The Probable Maximum Precipitation (PMP) used in the 1966 study was obtained from the U.S. Weather Bureau's Hydrometeorological Report (HMR) No. 33. FERC required the Keowee PMF be updated ut~ilizing HMR~s 51/52, which is the reason for the change as described in the "*Report on the Analysis to. Determine the PMF- March 29, 1995."

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pae1 Page15 RAI-8: Streams and Rivers (Choice of Methods and Technical RatiOnale)

The NRC staff's review of the FHRR did not find discussions on the tailwater level below Keowee Dam as a result of PMF. The NRC staff requests information regarding the taiiwater level below Keowee resulting from the PMF discharge (see FHRR Section 2.,2.1) and discussion of the consideration given for any effects that the taliwater and possible backwater elevations may have on the site flooding hazard. _

Response to RAt 8 The minimum tailwater elevation at Keowee Dam with no hydro units operating is 660.0 ft msl as presented in the Keowee STI document, Section 6.2.2. The minimum tailwater elevation corresponds to the full pond elevation of Hartwell Dam located downstream of Keowee Dam.

The ONS plant yard elevation is identified as 796.0 ft. msl in the UFSAR, Section 3.4.1.11. The maximum tailwater elevation at Keowee Dam for the Keowee PMF is 748.6 ft msl, based on the FERO approved documentation, "Keowee-Toxaway Project, FERC Project: No. 2503, Emergency Action Plan - AppendiX A, Dam Breach Flood Analysis, January. 2012." There are no backwater consequences at the poWer block elevation from the Keowee tajiwater during the Keowee PMF discharge due to the significant elevation difference (approximately 47 ft) between the Keowee tailwater elevation and ONS plant yard elevation.

RAI-9: Streams and Rivers (Choice of. Methods and Technical Rationale)

The NRC staff reviewed PMF levels and timing Presented in Sections 2.21t and 2.2.2 of the FHRR and the associated documents provided in the electronic reading* room.

During the review and after subsequent discussions with the licensee on March 12, 2014, the. staff noted that only one PMP distribution was used. The staff requests information on the temporal distributIons of precipitation on the watershed. More specifically, the staff requests discussion and technical rationale that explains the considerations given to other PMP distribution typeslscenarios, in order to demonstrate the selection of a conservative PMP distribution for the PMF analysis.

Response to RAI 9 The existing Federal Energy Regulatory Commission (FERC) approVed Probable Maximum Flood (PMF) developed for the Keowee Development in March 1995 (Keowee Hydro Project, Keowee & Little Rivers, S.C., FERC Project No. 2503, Report on the Analysis to Determine the PMF maximized the storm runoff by moving the storm center within the watershed in order to produce the greatest reservoirelevation at Keowee Dam. This maximization process was performed using a single temporal *distribution which was developed in accordance with Hydrometeorological Report (HMR) 52 recommendations. Tihe temporal distribution placed the peak 6-hour rainfall between the 36*" and 4 2 nd hour (an example of which is provided below in Figure 9-1). An independent review, of the existing PMF study found ithat the study was performed in accordance with* generally accepted engineering practice prior to submittal and T~L;. AAa*-I_..A pz t:_ E*-*.1.

_.*t.zh ... _z*L -L ---'.--:-. -- *t.t.-*-; L f-::----..at.::-- "'4ttllllTl*LU r .... U.Ll: i _:=.i.-

l. ,Iiml r .... E

... ............................... * ....... *---tlhll1" '

Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pae1 Page 16 acceptance from the Federal Energy Regulatory Commission (FERC). For these reasons, additional analysis of the selected temporal distribution was not considered.

FIGURE 9-1 EXAMPLE TEMPORAL DISTRIBUTION FROM EXISTING KEO WEE PMF ANALYSIS (KEO WEE HYDRO PROJECT REPORT ON THE ANALYSIS TO DETRMINE THE PMF, MARCH 29, 1995) -

HEC-1 PI DECK FOR SUB-BASIN ETC01 6

13 12 0 1 1 0 2 4 6 S 10 12 1* 16 18 20 22 24 26 28 3*0 32 34 36 38 40 42 44 46 48 '.0 52 54 56 58 60 623 64 66 68 70 72 How RAI-10 Streams and Rivers (Choice of Methods and Technical Ratonlale)

As described in Section 2.2.2, the Kirpich method was used to estimate time of concentration. The Kirpich method was derived based on data from agricultural watersheds with basin areas between I to 112 acres (0.004 to 0.45 square-kilometers) and topographical characteristics based on these watersheds located in Pennsylvania and Tennessee. The NRC staff requests technical rationale to determine suitability of the Kirpich method for determining the time of concentration to be consistent with the assumptions upon which the method was derived.

Res~ons. to RAI 10 A study approved by the Federal Energy Regulatory Commission (FERC) is discussed in the report titled 'Probable Maximum Flood Hydrologic Analysis for the Jocassee Drainage Basin,"

dated January 1991. The calculations and supporting report were prepared by Law Environmental, Inc. The discussion of the selection of hydrograph lagtimne can be found starting on page 6 of the Law Environmental report, Section 3.4. The report documents the comparison of the Soil Conservation Service (SCS), Snyder and Kirpich methods for developing the lag time iI ~IC i, ...... ~.. ~u. u . ,.u .,. .;.~. . ... .L -.. - i.A

- . ............. LI.

i .l *.............. J.

,t - - - r .....

. *gm ~uv z INt ,hettacments, mue r.ncIosure us uncontrolled.

Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pae1 Enclosure I Page17 for each sub-basin in the Jocassee drainage basin. The technical rationale for selecting between the methods tested is presented below:

"There are no continuous recording rainfall or stream gages in the Jocassee Lake watershed that would allow computation of a unit hydrograph for the basin. This fact required the use of synthetic unit hydrographs to model the watershed rainfall/runoff response. Law Environmental examined a number of different methods to develop synthetic unit hydrographs. These methods were applied to four representative sub-basins from among the sub-basins .composing the Jocassee Lake watershed. The five different empirical methods for computing lag times for the synthetic unit hydrographs were: Travel Time and SCS methods (U.S. Department of Agriculture, 1972), Snyder's and Gray's methods (Viessman, et. Al, 1977) and Kirpich's method (U.S. Department of Interior, 1977). The~watershed characteristics necessary for estimating the lag-time and/or time of concentration for these sub-basins for these

.differentformulae were derived from the study of the 7.5 minute series USGS topographic maps for the area. The four sub-basins selected, along with their calculated values of the times of concentration and average velocities, are shown below in Table 2.

Comparison of the results show that the SOS and Snyder's method yield the longest times, of concentration and hence the lowest average velocities. Gray's method yields wide ranges of times of concentrations and average-velocities whose averages are comparable to the results obtained from the Travel Time method.

Application of the Kirpich method was based upon Design of Small Dams (U.S.

Department of the Interior, 1977) and used the Tc adjustment for timber covered watersheds east of the 105th meridian. The applicability of each synthetic formula will be discussed and a justification of the selected method given.

The Travel Time method would at first seem to possibly be the most rational method.

In it, longest hydraulic flow path is divided into reaches with representative bottom slopes, cross sections, and roughness coefficients. Velocities are either assigned or computed (based upon normal depth) for each reach, and the result is divided into

the reach length to obtain the reach travel time. All reach travel times are summed to obtain a sub-basin time of concentration. The disadvantages of this method are:

1) it is very time consuming and impractical to apply to many sub-basins; 2) average velocities found in standard texts are. probably too low for the very steep watersheds found in this region; 3) computed velocities based on normal depth are extremely high (30-40 ft/sec). Current research suggests that natUral steep channels do no actually flow supercritical for other than very short reaches, but without calibration data it is not possible to assign -an appropriate roughness coefficient to realistically model the hydraulics.

The SCS method was developed .forwatersheds up to 2000 acres in size. Although it is often the preferred method, it is heavily weighted on the overland flow i n* acnments to LnClOSUre i. contain *ecuriiy *)enslinVc lnrormailn - yinaiuiii puuztuv u~bcmsurv~unucr UT-on- rtrm^tl fif lh AU4h.k rth.&knu*I 4.

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae1 Page18 component of the lag time as evidence* by inclusion of the runoff Curve Number in the denominator of the equation. For the Jocassee watershed, the selected curve number (CN=55) caused computed times of concentration to be unrealistically large, yielding average velocities in the 2.0 to 2.5: ft/sec range. We believe that these are unrealistically low given the steep watershed slopes and extreme rainfall intensities.

Gr'ay's method is attractive for moderately-sized watersheds 'up to 94 square miles and yields fairly reasonable times of concentration for the Jocassee watershed. It has the disadvantages of requiring two regional parameters which have only been developed in mid-western states...

Snyder's Method was developed for large (10-10,000 square miles) Watersheds in the Appalachian Highlands. It was not chosen because of the unrealistically large computed times of concentration.

The Kirpich method was developed in the 1940's for small watersheds in'Kentucky, but has gained wide acceptance all -over the' country. Although it is recognized as generally yielding conservatively small times of concentration, when coupled with the Tc adjustment for timbered watersheds (U.S. Department of the Interior, 1.977) very reasonable average velocities were computed for each of 'the four test sub-basins.

The Kirpich method was therefore used to define the hydrograph lagtime in all of the Jocassee Lake watershed sub-basins."

RAI-1 1: Streams and Rivers (Choice of Methods and Technical Rationale)

The NRC staff requests information on ,the upstream boundary conditions and flow velocity distributions for the dam breach model described in FHRR Section 2,3.3.

Specifically, the NRC staff requests information pertaining to any considerations given for sensitivity runs made for various upstream locations for the inflow boundary condition and an increasingly refined grid. The NRC staff also requests a technical discussion on how an updated flow distribution and a grddindependent solution could be expected to affect the results of the stepped breach progression of dam failure using the SRH-2D (Sedimentation and River Hydraulics)'model.

Response to RAI 11 Based on HDR Engineering, Inc. of the Carolinas (HDR's) modeling experience the upstream boundary condition is far enough upstream from the area of interest that any possible err~ors

-assOciated with discrepancies between the one-dimensional (1-D) and two-dimensional (2-D) models will not impact the results in areas of interest. As described in the 2-D Model Report dated March 2013, the 2-D boundary location was selected since it is a shared location with an actual 1-,D model cross section (i.e., HEC-RAS cross section is based off of the Digital Elevation Model [DEMI and not an interpolated cross section) and for its relatively simplistic geometric shape and deep depths which further removes any concerns over possible errors and

Enclosure I Duke Energy Response to RAI Regarding Flood HRR Page 19

.discrepanc;ies between the 1-D and 2-D models. In addition to reducing the chance for errors between the 1-D and 2-D model, another consideration for placing the upstream boundary was to place it far enough away from the ONS site that the velocity distribution develops into a realistic distribution as it approaches the ONS site, completely independent from what was prescribed at the upstream boundary. The independence of the velocitY distribution from the upstream boundary was confirmed by an inspection during the model development phase.

Given the current level of detail in the breach meshes HDR does not. believe that a more refined mesh would be beneficial (i.e., add substantial accuracy) given the required time needed to perform simulations. Current modeling methodologies represent a balance between modeling efficiency and accuracy/resolution and the mesh resolution is reasonable when considering the resolution of the bathymetric data used to create the DEM as well as the Velocity gradients encountered in the simulations. special attention was. given to and .refined mesh elements were applied to areas of high interest/high velocity gradient. The dam crest and embankment, the ONS site, the Connecting Canal, the intake canal and embankment, and the channel immediately downstream from the dam comprise 75-percent of the total mesh elements:

RAI-12: Failure of Dams and Onsite Water Control/Storaae Structures (Choice of Methods and Technical Rationale)

The Jocassee-Keowee Dam Failure Assessment t-D HEC-RAS Model Report described in FHRR'Section 5 indicates that sensitivity runs including different expansion and' contraction coefficients and Manning's roughness coefficient n were made to the I -D model to match the 2-D model results. The NRC staff requests the technical rationale for making changes to the I .D model based on the 2-D model While the 2-D model Itself has not been independently calibrated and validated. Particularly, the NRC staff requests the technical rationale for determining that the 2-D model is producing appropriately conservative results and how the model 1is validated to appropriately address the effects of the canal: restriction on the flow.

Response to RAI 12.

Section 5 of the FHRR is titled UAdditional Actions"* and does-not discuss. sensitivity analyses related to the models. The 1-D and 2-D modeling performed to support the FHRR is discussed in Section 2.3.3 of the FHRR.

The approach of using two models to evaluate the impacts of uPstream dam failure on flooding at complex downstream sites is a standard practice in dam failure modeling based on hydraulic Conveyance. The HEC-RAS 1-D dynamic routing model is efficient and well tested for simulating channel flow and can be Used with branches to simulate multiple tributary flows.

HEC-RAS was used to compute flows and water surface elevations for a very large study area that included Jocassee, Keowee, and Hartwell developments. Model efficiency allowed multiple scenarios to be developed to analyze hydraulic-results based on variable input parameters in a relatively short period of time. Although this model was able to perform dam= breach simulations upunu rcmnuvmu uz ,Lhi.*t1acl,uzcz,(, Luc ILU.IubLuLr ;l uutuiru~p~u.

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae2 20 Page and compute hydrographs (stage and discharge) over this large area, it is not capable of revealing the more spatially varying 2-D velocity field that may be important in the geometrically complex area near the ONS.

The complex hydraulic flow distribution around the ONS site was judged to be best evaluated using a 2-D model due to the ability of flow to be simulated as it is conveyed across representative channel and land topography (computational mesh). This is a better evaluation tool than a 1-D model which computes average velocity and water surface elevation in the

.channel cross section. Thus, the hydraulic characteristics during the hypothetical flood event at and around the ONS site are not able to be described using the 1-0 model approach.

Consequently, the SRH-2D model was used to perform very detailed calculations over a relatively smaller area near Keowee Dam and the ONS, with the results from the 1-D model being used to supply, boundary conditions to the 2-D model (Wilson Engineering, Independent Technical Review of HEC-RAS and SRH-2D Modeling, March 2013).

The 2-D analysis was performed to add detail to the HEC-RAS analysis due to the complex flow routing between the Keowee and Little River arms of the reservoir, the Oconee Intake Canal, and to model the potential inundation of theSafe Shutdown Facility (SSF) in the ONS Yard (Yard). The 2-D model domain includes the area immediately surrounding the stat'ion. For purposes of 2-0 model performance, the domain was restricted in size to facilitate Scenario simulation within a time frame that would support multiple runs that could be performed in the study. Model development required the establishment of boundary conditions at three locations (Lake Keowee-Keowee River basin arm upstream of Keowee Dam, Lake Keowee-Little River basin arm downstream of ONS Intake Canal entrance, and Keowee River downstream of Keowee Dam). Routing results from the 1-0 model analysis were extracted and utilized as boundary conditions for the 2-0 analysis. Ultimately, the 1-0 model was utilized to inform the 2-D model at the boundaries. During development of the 2-0 model, there was an expectation that the 2-0 model would produce more realistic but similar results to the 1-0 model at the boundaries, including flow distribution and timing at the two outflow boundaries, performance through the Connecting Canal, timing of breach initiation, timing of peak flows, and flow over and through the dam during breaching. Once the 1-D and 2-D models reached reasonable agreement at the boundaries, the 2-0 model results were used to inform the external flood evaluation at the ONS.

A relatively simple means to perform a Quality Assurance/Quality Control (QA/QC) check between the HEC-RAS and SRH-2D model is to create pseudo-steady simulations for each model and compare the water surface profiles. This was performed by taking the unsteady models and applying constant value flow and stage boundary conditions for a very. long time period and running the models until equilibrium was established. This was performed for flows representative of the range of conditions for which the unsteady simulation was undertaken. By comparing pseudo-steady state runs for each model, a verification of appropriate geometry input and roughness coefficients without the complications of unsteady flow simulations and possibly numerical stability issues was performed. As a result, unsteady expansion/contraction loss coefficients were added to the Connecting Canal and portions of the Keowee Dam

.. ... '* ' , " . . . . . ... ) ... .. "..... ... . I J)

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pae2 Page 21 discharge reach. These adjustments to the 1-0 model resulted in improved correspondence between the two models and improved confidence in the results of the simulations. As a result, the head losses through the Connecting Canal were similar in magnitude, as were the Keowee Dam tailwater elevations (Wilson Engineering, Independent Technical Review of HEC-RAS and SRH-2D ModelingMarch 2013). -

The discussion provided in both the 1-D and 2-D model reports developed in support of the FHRR summarizes the process of performing sensitivity runs with the two models for simulating flow through the Connecting Canal that joins the two basin portions of Lake Keowee (Keowee River basin and Little River basin). The flows modeled are the result of a breach of the Jocassee Dam and are orders of magnitude larger than any natural flooding event for the drainage basin and therefore beyond any ability to calibrate to a *historic event. Based on hydraulic modeling experience, and review of the flow SimUlation between the 1-0 and 2-D models for the numerous model sensitivity simulations performed; the dam failure flow Convenience through the COnnecting Canal was judged to. be more appropriately represented by 2-0 model flow vs. the limitations associated with the 1-0 model. (HEC-RAS).

RAI-13: Failure of Dams and Onsite Water Control/Storaae Structures (Model Documentation and InputlOutput Files)

The NRC staff requests additional information on the geometry of the reservoir near the Keowee Dam (including the Little Arm, the cove area, and the canal restriction) and their treatment in the l-D and 2-D models described in FHRR Section 2.3.3. More specifically, the NRC staff requests:

a) Electronic input and output files for the HEC-RAS 1-D and SRH-2-D model runs.

Response to RAI 13a HDR Engineering, Inc. of the Carolinas (HDR) has compiled the HEc-,RAS 1-D model"input and output files to support the March 2013 FHRR. These data files are provided as Attachment E on a DVD.

HDR has compiled the SRH2D model input and outpUt files to Support the March 2013 FHR.

These data files are provided as Attachment F on~a DVD.

b) Information explaining how the storage effects of the Little River Arm and the "cove" are modeled, in the I -D HEC-RAS and how the assumptions are considered.

Response to RAI i3b The Keowee Development reservoir is comprised of two different watersheds consisting of Keowee River and Little River and is located on the boundary of Oconee and Pickens Counties in South Carolina. Keowee Dam and Powerhouse are located appr~oximately 8 miles north-northwest of Clemson ,and approximately 8 miles north-northeast of Seneca, .South

,to*. cm.,a='.*.**~

l] .......... ,vl ,,* ,..J.=[Jt~a l.t.n.. [,apnIflonrP a.* alnl~flnl~rrnlhfl_

,It'r

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae2 22 Page Carolina. Little River Dam is located approximately 3 miles north, northeast of Seneca, South Carolina, and immediately upstream from Newry, South Carolina. Keowee Dam is located approximately 12 miles downstream of Jocassee Dam.

Lake Keowee is formed by two adjacent drainage basins that are connected by a man-made ..

2,000-foot-long canal approximately 100 feet deep. The connecting canal is located approximately 0.5 mile north-of the ONS site and is generally aligned in an East-West direction. Keowee Dam impounds the portion of the reservoir that lies in the Keowee River drainage basin. Little River Dam (located approximately 5 miles south-southwest from Keowee Dam) is the principal dam that impounds the portion of the reservoir in the Little River drainage basin. Four dikes, identified as Saddle Dikes A, B, C, and D, are located within approximately 1.5 miles north of Little River Dam and form the eastern rim of the Little River portion of the reservoir. In addition, the ONS Intake Canal Dike impounds the ONS Intake Canal (man-made) that is located off the main stem of the Lake Keowee.-Little River basin arm drainage basin bringing cooling water to ONS. Lake Keowee serves as the lower reservoir for Duke Energy's Federal Energy Regulatory Commission (FERC)-licensed Jocassee Pumped Storage Project.

The 2013 ONS 1-D model (Model) is a sub-model of the FERC required 2012 Duke Energy Jocassee-Keowee Emergency Action Plan (EAP) model (2012 EAP Model) andwas utilized in the ONS external flooding evaluation study. As shown in Figure 13-1, the FERC 2012 EAP Model consisted of the reservoir-riverine system downstream of Lake Jocassee including Lake Keowee, Hartwell Lake, Richard B. Russell Lake, J. Strom Thurmond Lake, Stevens Creek Dam, and the Augusta City Lock and Dam with an approximate overall length of 217 miles.

The Model consists of the reservoir-riverine system downstream of Lake Jocassee including the two drainage basins that form Lake Keowee along with Hartwell Lake. The model extent.

that was judged to have direct impact of dam breach failure inundation for the ONS is the sub-model that ends at Hartwell Lake Dam.

The Model extent is shown in Figure 13-2. The overall length of the Model, including main stem river-reservoir and tributaries between Jocassee Dam and Hartwell Dam is approximately 148 river miles and is comprised of 3,810 cross sections. Additional model geometry information is found in Section 3.5 of the Oconee Nuclear Station, External Site FloodingEvaluation, Fukushime Study, Jocassee-Keowee Dam FailureAssessment, 1-D HEC-RAS Model Report, March 2013.

ct,'*.i" z*.no,1,.iw,,~

ta

vA i,/t/ -'==,,mr =* II ,-I

Enclosure I Duke Energy Response to RAt Regarding Flood HRR Page 23 FIGURE 13-1 LOCATION OF MODELED RESERVOIRS IN THE FERC 2012 EAP MODEL A

- ~A

'1,*

K

~ke South CeroUem oeoq

-a \

,. . . . .. . . . .. . . .. ! - -- ; - -- [ . . .*;- . . .*A -*-

r-I. . .

.~t..

ii .,v- -

T:hL....._

2 L L -*r . . ! .-- - .. .

Enclosure I Duke Energy Response to RAI Regarding Flood HRR Page 24 FIGURE 13-2 2013 ONS MODEL EXTENT

.. . ... . . . . . . . au. u....... . c . ...... .. .~ .... *, I JUWA

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae2 Page 25 FIGURE 13-3 HYDRAULIC PERFORMANCE OF CONNECTING CANAL AT SOUTH CAROLINA HIGHWAY 130 BRIDGE 1I t

'?* mC A

~....2 S C.....~.. . I... qwI - vi ,quIU him Iuf-ul umnau uI I ~~

Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pae2 Page 26 HDR utilized HEC-GeoRAS (version 4.2.93) to develop the Model geometry independent of HEC-RAS using available Geographic Information Systems (GIS) data from the State of South Carolina (with some overlap into Georgia) to create Digital Elevation Model (DEM) electronic files for Jocassee, Keowee, and Hartwell reservoir systems. The completed electronic geometry files were then imported into HEC-RAS (version 4.1). The hydraulic response of three reservoirs comprising 17 river/tributaries are incorporated in the Model.

The Model geometry for the above water terrain, Lake Keowee bathymetry, and Keowee River adjoining the ONS site is defined by 1,942 cross sections. The designated body of water and corresponding developed cross sections serving the Lake Keowee components are:

Lake Keowee (Keowee River basin arm) and immediate Keowee River downstream of Keowee Dam - 1,414;

Lake Keowee (Little River basin arm) - 402;

Connecting Canal between the two Lake Keowee drainage basins - 27; and

ONS Intake Canal and flow path downstream of ONS Intake Canal Dike - 99.

The upstream cross sections (Keowee Dam, Little River Dam, and ONS Intake Canal Dike) developed to define Lake Keowee are used to determine the reservoir storage Capacity in the Model. The percent reservoir storage capacity variance between the Model reservoir storage capacity simulated through a series of cross sections and design reservoir storage capacity for the three reservoirs modeled in the 2013 ONS Model are provided in Table 13-1. The design reservoir storage capacity for Lake Jocassee and Lake Keowee are on file with the FERC. The modeled reservoir volume for Lake Jocassee and Lake Keowee was extended beyond the normal water levels by increasing each input cross section above an elevation that would be needed by the modeling of the simulated breach of the Jocassee and Keowee dams. Modeling experience was used to locate cross sections and select end points. This includes the development of cross sections at branches in the HEC-RAS model that were used to simulate the dynamic transfer of water between the Keowee River basin and Little River basin (Connecting Canal) and the Lake Keowee-Little River basin and the ONS Intake Canal.

TABLE 13-1 MODELED VOLUME COMPARISON Flood. Pool Volume Comparison Development Elevation

(.ft msl) Storage Curves (ac-ft) HIDR (ac-ft)

HIEC-RAS Difference (ac-ft) Percent Difference Jocassee 1, 115 1,247,096 1,247,096 0 0%

Keowee 805 959,266 965,796 6,529 1%

Hartwell 665 2,832,083 2,886,998 54,915 2%

The transfer of water in the Model occurs at the junction of each of the river reaches. The water path approach angle at the junction is used in the Model to determine if the Energy I.

mT *

i - -. .t_._x. c..._ :t c*J.. _. ....

I=f rl*.... ,.* ?':1 ~l.i__gi r.*

.. .... . t.L1 l . L*

J: .. . . .....

1fo crnn*o"i'*

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pae2 Page 27 Equation or Momentum Equation is applied to determine potential losses .across the junction.

Significant angle changes that approach 90 degrees warrant the Momentum Equation application in order to account for energy losses associated with the abrupt change in direction. This is the case for the water surface areas adjoining the ONS site. The Model utilizes the Momentum Equation at Junctions K9, K16,. K17,.and K19 (reference is given to the HEC-RAS Input/Output files) which defines the water surface area around the four sides of the ONS site.

The transfer of water in the Model is evident in Figure 13-3 and illustrates the dynamic hydraulic performance of the Connecting Canal between the two Lake Keowee basins during a hypothetical failure of Jocassee Dam and cascading failure of Keowee Dam. The Connecting Canal cross section (3731). is identified with .the South Carolina .Highway 130 bridge crossing. The dynamic transfer of water through the Connecting Canal is described as:

Initial Jocassee Dam breach wave arrives at Connecting Canal at approximately 1430 hours0.0166 days 0.397 hours 0.00236 weeks 5.44115e-4 months on 11/1/2008 (fictitious model simulation date);

A portion of the Jocassee breach discharge flows from the Keowee River basin arm of Lake Keowee through the Connecting. Canal and into the Little River basin arm of Lake Keowee, achieving a peak discharge transfer of approXimately 634,800 cubic feet per second (cfs) and peak stage of approximately 817.5 feet above mean sea level (ft msl);

The water level osraof Keowee Dam continues to rise obtaining a peak stage of 818.4 ft. msl atxLjF on 11/1/2008 with a combined approximate discharge (overtopping+ initial breach+ spillway gates) otI~b*cTF) ,

.As the breach develops at Keowee Dam the combined discharge increases from

As the breach develops at Keowee Dam the flow of water thtrough the Connecting Canal is reversed. .Water transfer is now from the Little River basin arm Of Lake Keowee through the Connecting Canal and into the Keowee River basin arm of Lake Keowee (Keowee Dam breach), The reverse flow through the Connecting.Canal peaks at approximatei]lIZxF) .

As Lake Keowee continues to drain through the Keowee Dam breach theConnecting Canal stage elevation decreases from approximately 817.5 ft msl to. 730 ft mnsl over an 18 hour2.083333e-4 days 0.005 hours 2.97619e-5 weeks 6.849e-6 months period.

The total volume of water transferred through the Connecting Canal during the Jocassee Dam-Keowee Dam breach process is the result of the. rise in the Little River basin arm due to water from Lake Jocassee being forced through the Connecting Canal and creating a dynamic backwater impact on the cross sections representing the Lake Keowee-Little River basin arm reaches of the Model and the following reversal of flow once Keowee Dam is overtopped followed by the breaching process.

. .. . " i*a*,*'u* a

0" - * **eIl =-* -'in

. ..- ! - al~dl~L1^0 1C&=fl .1-?Ol.tt~n

UVIL 4*mBi*VMl V *U

Enclosure I Encloure I Duke Energy Response to RAI Regarding Flood HRR Pae228 Page RAI 15: Input to integlrated assessment: Flood height and associated effects Back~ground: The March 12, 2012, 50.54(f) letter, Enclosure 2, requests the licensee to perform an integrated assessment of the plant's response to the reevaluated hazard if the reevaluated flood hazard is not bounded by the current design basis. Flood scenario parameters from the flood hazard reevaluation serve as the input to the integrated assessment. To support efficient and effectiverevaluations under the integrated assessment, the NRC staff will review flood scenario parameters as part of the flood hazard reevaluation and document results of the review as part of the staff assessment of the flood hazard reevaluation.

The licensee has provided reevaluated flood hazards at the site including local intense precipitation flooding, probable maximum flooding on contributing watershed, flooding in streams and rivers, and flooding from breach of dams. The local intense precipitation flooding is reported to exceed the current licensing basis and subsequently the licensee has committed to perform integrated assessment.

Request: The March 12, 2012, 50.54(f) letter, Enclosure 2, requests the licensee to perform an integrated assessment of the plant's response to the reevaluated hazard if the reevaluated flood hazard is not bounded by the current design basis. The licensee is requested to provide the flood height and associated effects (as defined in Section 9 of JLO-ISG-201 2-O5) that are not described in the flood hazard reevaluation report for mechanisms that trigger an Integrated Assessment. This includes the following quantified information for each flooding mechanism (as applicable):

Hydrodynamic loading, including debris,

Effects caused by sediment deposition and erosion (e.g., flow velocities, scour),

Concurrent Site conditions, including adverse weather, and

Groundwater ingress Response to RAI 15 The mechanisms that exceeded the current licensing basis/design basis and triggered the integrated assessment at Oconee Nuclear Station are Local Intense Precipitation (LIP) event and the sunny-day failure of Jocassee Dam and the cascading failure of Keowee Dam.

Local Intense Precipitation Hydrodynamic Effects Section 2.1 and Appendix C of the FHR provided detailed results (flood depths and durations of inundation) at various locations throughout the site. Flow velocities from the 2-dimensional model were reviewed at relevant door openings to safety-related structures to determine whether hydrodynamic loading is of concern at any of the critical locations. The results indicate that maximum velocities are generally below 1 ftsec, with occasional exceedance at locations where flow is constrained between two buildings. Furthermore, the velocities reported by the T.... 'L, n~nnL tV E.ju:,m v ., U4 II l y 2,.vLUI~ E(* ,IIl:mtIIW J~l~

ILII IiIIE UUI -- UVIIIJI.UIU ,)Jll1 JUU i Ui*)iU:*UU[V UflUI*F Sr;a.........,-,..:. , ... *....... :._ .: 1..:.......... .....

Enclosure I Encloure Duke Energy I Response to RAI Regarding Flood HRR Pae2 Page 29 model do not represent velocities at the maximum flOod stage and the velocity vectors are generally not orthogonal to. the doors. Since hydrodynamic loads are a function .of flow velocity and flood depth, these loads are expected to be minimal, and well within the margin of safety provided for the respective flood protection features.

ASCE/SEI 7-10 standard provides a recommended approach for estimation of dynamic effects of moving :Water with flow velocities below 10 ft/sec.' Based on this approach, dynamic effects of moving water can be Converted into equivalent hydrostatic head by increasing, the design flood elevation by an equivalent surcharge depth, dh, equal to aV2 dh=

Where V = average velocity o~f water in ft/s;ec g =acceleration, due to gravity, 32:2 ft/sec 2 a = :coefficient of drag The average maximubm velocity rat ONS critical structures is 1.;13 ft/sec. Per FEMA 259, the most conservative coefficient :of drag Lfor building width/flood depth ratio is 2. Using these extremely conservative values, the equivalent surcharge depth is equal only to 0.04 ft.

Debris Effects The-areas within the protected area that could potentially provide, a sourcefor debris are either paved or covered with gravel or paved surfaces with little vegetation or loose materials available. The protected area is .also surrounded by vehicle barrier system and security fences which would significantly minimize the potential for any debris to impact safety related structures. In addition, relatively loW velocities Would minimize the movement of debris throughout the power block. Therefore, debris effects at Oco~nee Nuclear Station were considered negligible.

Effects caused by sediment deposition and erosion As described previously, the average maximum velocity throughout the power block is 1.13 ftlsec, with asingle highest velocity of 6 ft/sec. Since most areas within the power block are paved, no erosion is expected because~maximum values of flow velocity that can be sustained without significant erosion are an order of magnitude higher than the .average maximum =velocity.

The local intense precipitation event is a localized flooding event, which is not .expected to carry significantamount of sediment typical for riverine flooding. Therefore, isediment deposition at Oconee Nuclear Station was considered negligible.

Concurrent site conditions The meteorological events that could potentially result in significant rainfall .ofthe LIP magnitude are squall lines, thunderstorms with capping inversion, and mesoscale convective systems.

........................................... *=*mm^*

Encloure Duke IEnergy Response to RAI Regarding Flood HRR Pae3.30 Page These meteorological events are typically accompanied by hail, strong winds, and even tornadoes.

Groundwater ingress The LIP is a localized, Short duration event, which is not expected to' increase, groundwater levels on site. Furthermore, Oconee Nuclear StatiOn is protected against groundwater ingress.

Sunny-day Failure of Jocassee Dam Hvdrodynam.ic and Debris Effects The sunny-day failure of Jocassee Dam and the cascading failure of Keowee Dam results in flooding hazard elevation of 790.4 ft in the Keowee tailrace. The flooding hazard elevation remains below the nominal elevation of the power block (796 ft) and, therefore., safety-related structures throughout the power block will not be affected by the associated effects of the dam failure. The results of the .2-dimensional modeling indicate that the flow velocities in the Keowee tailrace immediately downstream *of the breach approach 40 ft/sec. These high veloCities are limited to areas downstream of the breach opening; however, they will result in significant hydrodynamic forces and debris carried the breach wave will result in impact forces on structures located within the breach inundation zone. Because the exact flow paths for debris cannot be predicted, any structures located within the inundation zone will be considered lost due to hydrodynamic and debris impact loads. Conversely, flow velocities adjacent to the Intake Canal dike and the east slope of the power block are significantly lower anddo not exceed 4 ft/sec.

Concurrent Site Conditions The results of the flooding hazard re-evaluation indicate that all access routes from the southeast direction will be inundated during the event and existing infrastructure will likely be severely damaged dueto high velocities and impacts from debris and mud flow.

Grudwater n~ress Groundwater levels on site will not increase due to the sunny-dlay breach. Furthermore, Oconee Nuclear Station is protected against groundwater ingress.

TL AII...L..L .v, .. J. ......-

,. 1 ~ ~ ,nmg iau usu in urn UICcsisr ner v..f(..-. =.= *la x+/-i

ciOSUre t is-uncontroll-ed..

Enclosure I Encloure Duke EnergyI Response to RAI Regarding Flood HRR Pae3 Page 31 Attachments A. HDR ENGINEERING, INC. OF THE CAROLINAS report: "Oconee Nuclear Station Local Flooding.An~alysis Hydraulic Modeling Report Yard And.R~opf Drainage. Local Flooding Current Licensing Basis," dated November 2012 B. Data Files. on DVD:. AVIs and model Input/Output data associated with Current Licensing Basis. The HDR disc is titled: "Oconee Nuclear Station Fukushima 2,1 FHR IWCS Hydraulic Modeling Local FloodAnalysis, Current licensing Basis Input & Output Files, November 2012" C.. Data. Files on DVD: AVIs and model InPUt/OutPut data associated with Beyond LiCensing Basis, The HDR disc is titled: "oconee .NuclearStation Fukushima 2.1* FHR /WCS

-HydraUlic Modeling.Local Flood Analysis, Beyond licenSing BasisInput & Output Files, February2013" D. Electronic drawing on CD: An electronic version of an Oconee drawing with topographical contours.

E. Data Files on DVD: HDR Input and Output data from the HEC-RAS 1-D model supporting the March. 2013. FHRR. The HDR disc is titled.: "Oconee Nuclear Station Fukushima 2.1 FHR 1-D Hydraulic.Modeling., BEP LE FinalMarch 2013 Model/Input & Output.Files" F. Data Files on DVD: HDR Input and Output data from the SRH2D model suppor'ting the March 2013 FHRR. The HDR disc is titled: 0Oconee Nuclear Station Fukushima 2.1 FHR SRH2D Hydraulic Modeling, BEP-LE SRH2D March 2013 Model lnput & Output Files~ dated March 2014 Note: The CD/DVD attachmentsare not.documents intended for printing. They are intended for transmittingelectronic data to be used.by NRC reviewers.

h*wt _- ......... $*J..-.... , v*... *,,v* muu rlUlu De sn* wernnela fromrpbi

--UrL:;

INC,, the E~nclosure l .1IR1C~tmen5, is uncontrolled.