ML20136J071

Informs of 850719 Discussions W/Ebasco & BNL Re Program for Addl Confirmatory Analysis of Basemat Structural Strength. Observations & Comments Resulting from Discussions Which May Be Subj of Future Meeting Encl

ML20136J071
Person / Time
Site:	Waterford
Issue date:	08/12/1985
From:	Joshua Wilson Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
NUDOCS 8508200603
Download: ML20136J071 (5)

Text

.. ,

AUG 121985 MEMORANDUM FOR: Docket Files Distribution De Ret J11e.>

FROM: James H. Wilson, Project Manager LB#3 Reading Licensing Branch No. 3 PDR Division of Licensing LPDR JLee THRU: George W. Knighton, Chief JHW11 son Licensing Branch No. 3 Division of Licensing

SUBJECT:

DISCUSSION CONCERNING LP&L'S CONFIRMATORY ANALYSES OF COMMON FOUNDATION - BASEMAT AT WATERFORD 3 A license condition in the full-power operating license for Waterford 3 requires that Louisiana Power and Light Company conduct additional confirmatory analyses of basemat structural strength. LP&L has submitted and the staff has accepted a program plan containing five areas of investation for confirmatory analysis.

The five specific areas are:

1. Dynamic effects on lateral soil / water loadings.

2. Dynamic coupling between the buildings and basemat for vertical seismic loading.

3. Artificial boundary constraints used in the finite element model.

4. Fineness of the basemat finite element mesh.

5. Calculation of internal basemat forces during the construction sequence so that the source of internal basemat cracks could be evaluated.

Discussions of the detailed program and progress on the various tasks were held on July 19, 1985 between EBASCO, consultant for LP&L, and BNL, consultant for NRC.

EBASCO was represented by J. Costello, A. Wern, D. Nutta and P. C. Lu, while M. Reich, C. Constantino and C. Miller represented BNL.

Enclosure 1 contains some observations and coments by BNL that came out of the discussions and which may be the subject of further conversation or a future meeting. The observations / comments address each of the specific areas of the program plan outlined above.

p/

James H. Wilson, Project Manager Licensing Branch No. 3 Division of Licensing

Enclosure:

n As stated 8508200603 850812 .

PDR ADOCK 05000382 cc; w/ enclosure l P PDR J. Knight fV D. Crutchfield S. Turk S\

q DL:L JHWil es.

D B#3 GWKnighton D E ight ADha h DCrdItEhfield 8/7/85 8/%5 8/ /85 8/( /85 -

Enclosure 1

. . 1. Evaluation of Dynamic Side-Wall Pressures i .

7

_EBASCO,ispheparkn t uc'tiuY1 model of the EPIS which is to include the capability for treating in an approximate manner the influence of vertical side wall flexibility on lateral soil pressures sustained during a seismic event. This model is being prepared for the FLUSH computer program and consists of a two-dimensional vertical finite element mesh extending through the soil and into structural basemat. Vertical stick models are included for FHB, RB and RAB, all sitting atop the basemat. To the south of the NPIS, and uncoupled from it, a small finite element model of the Turbine Building will be included. The stick models for the FH3 and RAB are to have additional  ;

vertical' beam elements included in the areas in contact with the foundation soils. The sticks are to be sized to yield the proper mass and stiffness of the side walls of the NPIS. The model, when completed, will then be subjected to a NS ehrthquake input. The model to be used for an EW earthquake input is as yet undefined.

I EBASCO is currently studying three separate it' ems which will impact on the numerical results obtained for this problem, na'mely, (a) cut-off frequency required for the calculation (b) element mesh size criteria .

(c) depth to artificial bottom boundary of mesh EBASCO has currently completed a series of one-dimensional vertical column studies (SHAKE runs) which indicate that the soil column transmission frequency is in the range of 2-3 cps. Thus a cutoff frequency of 8 cps in the FLUSH calculation will provide adequate accuracy in the numerical calculation, leading to a much reduced computer run time (and cost). We agreed with this  !

assessment, and in fact strongly suggested to E8A560 that the 8 cps cutoff be  ;

used in future calculations. This would then allow for the use of larger l elements in the mesh, which in turn, will allow for a much deeper mesh to be

{

l used in the model. It is our belief that to adequately allow the NPl$ model

to rock due to lateral seismic inputs, the bottom mesh boundary must be placed ,

- d

eper than was initially being considerd by E58ASCO. l i

1

• 2 Some additional items associated with this prodlem were discussed and the ;

following . suggestions niade by BNL: *

(a) Include additional elements between the ' lateral transmitting boundaries and the side walls of the NPIS to better ascertain side wall pressures.  ;

(b) Determine total lateral forces acting on the side walls as a function l of. time to ensure that the sidewall soils do not separate from the l

NPIS during the design seismic event. -

l (c) It was agreed that the sidewall water forchs could be calculated from the Hononobe-Okabt approximation.

(d) It is also suggested that the element stre'sses in those basemat elements in the vicinity of the mat- cracked zone be monitored during the computer run and presented together with the sidewall pressure output.

2 Dynamic Coupling i

for the vertical seismic problem, the FE modei described above will ag,ain be used, but this time inputting the v'ertical seis51c design pulse. The objective is to determine the magnitude of seismic response and compare these with the design values. It was suggested that the " soil" properties used in these runs be chosen such that the P wave velocity corresponds to the wave

velocity in water and the S wave velocity correspohds to the velocity in the soil.

l i

3 Artificial Boundary Constraints In the original NPIS study, internal basemat forces were evaluated using a finite element model of the basemat and structur'al elements framing into the mat. The model was vertically supported on foundition springs representing the soil stiffness.

Horizontal lateral soil pressures act on the exterior walls of the '

foundation. If the distribution of these forces through the basemat is to be j evaluated, care must be exercised in selecting the type of horizontal boundary i i

o l

l

constraints used in the model. Rigid supports were provided along the N-S l centerline of the mat with no Justification given to show that these supports l i

didnotinfluencethedistribdtionofloadsinthedasemat.

The applicant has proposed to use horizontal shrings (in both the E-W and >

N-5 directions) to connect the center of the basemat to the soil. The spring <

~

constants will be selected to represent the basemat' base friction and a risulting subsoil deformation. Several spring constants will be used to evaluate the significance of the value used.

The models have been prepared and they are reaBy to make the computer runs. .The approach taken is reasonable and should resolve the concern.

4 Fineness of Mesh .

In several locations in the mat the finite eNment mesh is rather coarse. A comparison of the shear and bending mome'nts computed in adjacent elements indicates that a finer mesh would yield bdtter results. A finer mesh was proposed by 'the applicant and discussed at the meeting.

This refined mesh has been reviewed by BNL in' light of the various concerns regarding mat cracking that were raised during the plant licensin'g.

Itisrecommendedthatsomeareasofthemeshshouldberefinedinadditionto those already proposed by the applicant. In particular regions outside the RC8 walls where the primary E-W cracks occurred wehe areas where mat shear and moment values ware examined in detail. It is recoinnended that the mesh be refined in two areas. The first is the area east of the RCB and contained

~

l within column line R-N-11A-9M. The second is west of the RCS and is contained within column line R-N-7A-4A. These areas are now modeled with nodes spaced in a grid of about 25' x 25'. It is recomended that additional nodes be placed such that elements in the grid are reduced to about 1/2 of the current size.

3

e .,

'~

5. 'Const rdct i oh~ 5Egliente Toiid's ~

~ '

i ihe applicant proposes to perform a finite elenlent analysis of the mat which simulates theconstruction. sequence. The mat,will be modeled with about 6-9 elements per construction block. Blocks of eledents will be added to the model as construction progresses to simulate the construction sequence.

Properties of the concrete will be taken as variable with concrete age. The fonn of this variation is being selected based on available project experimental data with extrapolations made, where rsquired, using " textbook" dat. Concrete creep effects are also included in the modeling.

Measured displacement are used as input to the model. Since settlement data are only available at the corners of some of the blocks, interpolation of the measured data must be made to fill in data for 'other nodes.

EBASCO is in the process of developing this model. They will run one strip first to evaluate the model. It is recommended that the data used by the applicant be reviewed after these results are available, and prior to the final full analysis. '

. ,' i O

i t

. _ -4 . __ _ _ _ _ _ , _