ML20202F964

Suppl SER 3 Accepting Changes in Anchorage Criteria as Updated in GIP-3.SSER 3 Constitutes Acceptable Method of Implementation of Resolution of USI A-46 as Specified in GL 87-02

ML20202F964
Person / Time
Issue date:	12/04/1997
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20202F947	List: ML20202F942 ML20202F964
References
REF-GTECI-A-46, REF-GTECI-SC, TASK-A-46, TASK-OR GL-87-02, GL-87-2, NUDOCS 9712090269
Download: ML20202F964 (9)

Text

. -- . . . _ . - . . _ . . . - . - . . . - . . - . . . --

I i

SUPPLEMENTAL SAFETY EVALUATION REPORT NO. 3 i

ON SEISMIC QUALIFICATION UTILITY GROUP'S -

GENERIC lMPLEMENTATION PROCEDURE, REVISION 3, ,

UPDATED MAY 17,1997 FOR IMPLEMENTATION OF GL 87 02 (USl A 46),

yielFICATION OF SEISMIC ADEQUACY OF EQUIPMENT lbLQLDER OPERATING NUCLEAR PLANTS 6

Enclosure 9712090269 971204 i PDR REVOP EROSOUG PDR .. -

SUPPLEMENTAL SAFETY EVALUATION REPORT NO. 3 ON SElSMIC QUALIFICATION UTILITY GROUP'S GENERIC IMPLEMENTATION PROCEDURE, REVISION 3, UPDATED MAY 18,1997 FOR IMPLEMENTATION OF GL 87 02, USl A 48 PROGRAM VARIPICAclON OF BEISMIC ADEQUACY OF EQUIPMENTJN OLDER OPERATING NUCLEAR Pl. ANTS

1.0 INTRODUCTION

in response to the U.S. Nuclear Regulatory Commission's (NRC's) Generic Letter (GL) 87-02, the seismic Qualification Utility Group (SQUG) submitted on behalf of its members the Generic implementation Procedure, Revision 2 (GlP 2), as corrected on February 14,1992 (Referonos 1), for the resolution of Unresolved Safety issue (USl) A 46, " Verification of Seismic Adequacy of Mechanical and Electrical Equipment in Operating Reactors." The NRC staff completed its re-view of GIP 2 and Sued Supplement No.1 to GL 87-02, along with Supplemental Safety Evaluation Report No. 2 (SSER No. 2), on May 22,1992 (Reference 2).

To maintain the GlP with up to-date, new information, practical experfona , and lessor,s loamed during implementation of the USl A 46 program, SQUG submitted the CO, RevisM 3, dated July 31,1995 Pre-GlP 3, Reference 3), for review by the NRC stan. SQUG also submitted Revision 2 of the " Procedure for Revising the GlP " dated July 31,1995 (Referonos 4). The staff completed its review of these two submittals and issued a safety evaluation on August 6,1996 (Reference 5).

4 Subsequently, SQUG submitted *GlP, Revision 3, Updated May 16,1997, (GIP 3), and Procedura for Revising the GlP, Revision 3,'(Reference 6). The stars evaluation of Reference 6 is documented in this Supplement Safety Evaluation Repor1 No. 3 (SSER No. 3). SSER No.

3 superceeds the stafs safety evaluation dated August 6,1996 (Reference 5).

2.0 REVISION 2 AND REVISION 3 OF THE " PROCEDURE FOR REVISING THE GlP" 2.1 BACMOROUND in a letter from NRC to SQUG dated June 23,1993 (Reference 7), the NRC staff provided a poneral comment and specific comments, as well as staff psitions, concoming certain items in Revisions 0 and 1 of the " Procedure for Revising the GIP." SQUG responded to NRC's comments with a letter dated August 30,1993 (Reference 8).

Along with Pre-GlP 3 (Reference 3), SQUG also submitted Revision 2 of the " Procedure for Revising the GIP," dated July 31,1995 (Reference 4), for the stars review.

2.2 EVALUATION OF REVISION 2 l (1) As stated in Reference 8, in response to NRC Comment 1 in Reference 7 SQUG does not intend to include staff SSERs as appendices to the GIP.- However, as stated in i.

. Reference 7, the NRC staff considers SSER No. 2 to be its position for supplementing :

t. l I

the SQUG GIP 2 and believes that GIP 2 and SSER No. 2 are sufficient for implementing the resolution of USl A 46 at affected facildles. The staff believes that GIP 3, when finalized, should refer its users to NRC staff positions delineated in SSER No. 2, in l addstion to SSER No. 3, after it is issued.

(2) NRC staff positions regarding the time period for NRC review and approval of changes to l 4 the GlP as discussed in Comments 2 and 7 in Reference 7 remain unchanged. The staff j 1

maintains that the first paragraph on pegs 3 and the second paragraph on page 5 in l Reference 4 should be revised as indicated in Comments 2 and 7 in Reference 7.  ;

2.3 EVALUATlON OF REVISION 3 In Revision 3, SQUG has revised the

• Procedure for Revising the GIP' in accordance with item .

2.2(2) above. Therefore, staff considers Revision 3 acceptable.

l 3.0 CHANGES INCLUDED IN GlP, REVISION 3. UPDATED MAY 18.1997 The stats evaluation of changes included in GIP 3 are categorized in three groups based on '

i the GlP 3 submittels: evaluations of changes that impose additional restrictions on the user, editorial changes, and typographical changes.

3.1 . EVALUATIONS OF ADDITIONAL RESTRICTIONS .

3.1.1 Anchorage Criteria ,

3.1.1.1 Background SQUG has revised the GlP 2 anchorage criteria as a result of additional anchor bolt testing

, conducted by SQUG/ Electric Power Research Institute (EPRI) (Reference 12). The following Sections 3.1.1.2 and 3.1.1.3 provide the background information relating to the basis of the stars approval of anchorage criteria in GlP 2. Section 3.1.1 A describes the proposed change in GIP 3 and Section 3.1.1.5 provides the stats evaluation of the changes.  ;

i 3.1.1.2 Allowable Expansion Anchor Loads (Capacities) With Ample Test Data i

in response to GL 87-02, SQUG submitted for the stars review a proposed expansion anchor allowable load guideline (Reference g) prepareci by URS Corporation / John A. Blume and Associates (URS/Blume) for EPRl. Reference g sets the allowable load of expansion anchors

- as the overall mean strength test value for the bolts of a particular diameter divided by 2 '

(mean/2). The staff found that these allowable loads for expansion anchors in Reference g were unconservative and notified the SQUG that they were unacceptable.

t Following rejection of the anchorage criteria, the staff met with SQUG members regularly to ensure that the anchorage criteria to be developed could be accepted by the staff for resolving ,

USl A 46 issues. URS/Blume, a consultant to SQUG, believed that the allowable loads in Reference g were reasonably conservative because less than 2 percent of the laboratory test -

t data of static loads fell below the mean/2 allowable loads, and such a probabilistic approach for l

- . _ . , , .e.,.~ ~_ _ _ . , . , , - - - ~ - _-- - 4,-, _ . , . - ...- - - . - - - . - - - . - - - - - , _, ~ , , , , , , ..m,- _ m ,---

t 3

anchorage safety was documented in Reference 10. However, at that time, the staff did not find 1 that the approach used by URS/Blume was adequate because such an approach can only address the scattering phenomenon of expansion anchor test data in the laboratory. The approach does not include other important safety considerations, such as field installation variations and difficulties, seismic load effects on anchors, decrease of anchor capacity due to multi anchor effects in an anchorage system, and the loss of restraining force on anchors when the concrete cracks after being subjected to environmental and operating load conditions. The

- staff presented to SQUG representatives in February ig8g its views and safety conoems j regarding the impact of the above effects on expansion anchor allowables. A copy of the staffs -

presentation (Reference 11) was attached to the staffs August 6,1998, letter (Reference 5),

in Reference 11, the staff assembled the safety factors for expansion anchors recommended by I anchor manufacturers or specified by codes. Some of the safety factors in Reference 11 were derived from probabilistic approaches to address the scatterir g effect of test data. For example, )

Hiltl, an anchor manufacturer, also used the 2 percent non exceedance criterion (less than 2 <

percent of laboratory test data below the specified allowable ar'chor lead), with a coefficient of

. variation of 12.5 percent for its expansion anchor test data. Hitti arrived at a required safety i factor of 4 (the allowable load for anchors equal to mean/4). Another example in Reference 11

, is the German practice for anchor design. The Germans use a 5-percent non-exceedence '

criterion that results in a required safety factor of 3.85 for anchors, with a coefficient of variation i of 12.5 percent, and a safety factor of 4, with a coefficient of variation of 15.0 percent. l Reference 10 used the values of mean strength /3 as allowable loads for anchors with ample test i data. This approach was based on the data set that EPRI compiled of static tension and shear tests on expansion anchor bolts. These data were used to justify allowable strength criteria '

based upon the everall sample mean divided by 3. For each diameter of concem (e.g.,3/8", %",

5/8", etc.) the allowable load was taken as the overall mean for the bolts of that diameter divided by 3. These allowable loads are presented in Table C.21 of GlP 2 as allowable anchor .

capacity, in the SSER No. 2 of May 22, igg 2 (Reference 2), the staff concluded that the use of mean strength /3 in determining the a!!owable load for a single expansion anchor was adequate.

However, further reduction from the allowable load of a single anchor would be calculated -

separately because of the effects of multi-anchors, edge distance, concrete cracking, and other factors affecting the anchor capacity.

. 3.1.1.3 Allowable Expansion Anchor Loads (Capacities) Without Ample Test Data

-- Some types of expansion anchors whose test data were not represented in the data pool that i formed the allowable anchor capacities in Table C.21 of GlP 2 usually had only three to five l tests for each bolt diameter. For these types of anchors without ample test data, Reference 10 i provides ground rules for establishing capac".y reduction factors (CRFs). For these anchors, the l CRFs are multiplied by the allowable anchor capacities in Tabio C.21 of GlP 2 to obtain allowable capacities. -The CRFs for non-database anchors are listed in Table C.2 2 of GlP-2. In i order to establish a CRF, a capacity ratio (CR) for each anchor is first calculated by using the .

- average' ultimate strength from published data divided by the EPRI mean strength for concrete, ,

with strength equal to or greater than 4,000 pel for both po!:out and shear strengths that very depending on bolt diameter. The CRF is then establMeed based on the concept of using the

- - - - e s.a - e.+.-,--.--.-~~w,. .r-n-v<--w --~,nw.,..- -

e- ww-w,-~m-,~-,-wme,n---m---w--., -,---,w.-am-evn--+~. v.,w, ,- , -- m,._w-e-,-r v,,-,m,,-,wr

1 1 t 4

f smallest CR (CR minimum) among the CR values of the relevant data in accordance with the ,

j following rules:

j CRF = 1.0 for CR minimum equalto or greater than 0.8 i

CRF = 0.75 for CR minimum less than 0.8 but equal to or greater than 0.6 CRF = 0.6 for CR minimum less than 0.6 but equal to or greater than 0.48 in GIP 2, only one CRF was assigned to each type of anchor for all the different diameters of bolts and for the pullout and shear strengths as listed in Table C.2 2. Consequently, the ,

smallest CR among all the diameters of bolts within each type of anchor and the smaller value #

between the pulloet and shear strengths would be chosen as the CR minimum value for j assigning the CRF for that type of anchor. The staff approved the CRFs presented in Tcble C.2-i 2 of GlP-2. -

1 1 3.1.1.4 Proposed GlP, Revision 3 The proposed GIP 3 assigns different CRFs for pullout and shear strengths for anchors by l 4

adding another column to Table C.2 2. In addition to this change in Table C.2 2, the CRFs for  ;

! anchors of Star Stud, Star Steel, WeFit Stud, and WeFit Wedge were revised; also, Hitti Sleeve and Unknown anchors were subdivided into two categories of diame:ers. References 12 and 13 were submitted as supporting materials to justify the changes.  !

i 3.1.1.5 Evaluation 3.1.1.5.1 Mean/3 as Allowable Capacity and Determination of CRFa i

The staff had reviewed and approved the ground rules, as stated above, used f . establishing the allowable capacities listed in Table C.21 of GIP 2 for anchors with arSle test data and the

- CRFs listed in Table C.2 2 of GIP 2 for anchors without ample test data. ~. s most important i ground rule for expansion anchors is that "for each diameter of concem ( , V8", %", 5/8", etc.)

the allowable load was taken as the overall mean for the bolts of that dianm livided by 3," as i

- stated en page 1 of Reference 12, when there are ample test data. This ground rule establishes the basis of acceptable strength of expansion anchors and is accepted by the staff. In order to extend the use of the allowable capacities listed in Table C.21 of GIP-2 to those expansion anchors whose test date were limited to between three and five tests for each bolt diameter, the CRFs in Table C.2 2 of GlP.2 were introduced. - Since the test data in the range of three to five tests are not considered to be statistically adequate, the concept of CR minimum was introduced ,

as a ground rule to establish the CRFs. The staff considers this ground rule conservative and i n acceptable.

3.1.1.5.2 Different CRFs for Allowable Pullout and Shear Capacities and for Different 1 Diameters of Bolts 1

~ instead of using oia CRF for both pullout and shear strengths in GlP 2, the proposed change in I categonzation, which has different CRFs for pullout and shear strengths in GlP 3, is acceptable .

.i 1

. i 5 I because References 10 and 12 stated that these values were established based on the same  !

ground rules as mentioned above. The staff also approves the proposed use of two subsets of CRFs for two categories of diameters of bolts for Hitti. Sleeve and Unknown anchors because '

the concept of CRFs was based on test data for each diameter of bolts.

3.1.1.5.3 The Proposed Revision of CRFa in GlP 3 l The staff does not agree with either the revised CRFs in Table C.2 2 of Pre-GlP 3 or the safety i factor used in Reference 12 to reach these revised numbers, in the following discussion, the staff used data from the 42 tests of Star Stud anchors listed on page 12 of Reference 12 as an example to demonstrate the inadeq :acy of the revised CRF of 0.75 in the proposed Pre-GlP 3 from the original value of 0.6 in GlP-2.

l In Reference 12, there are 10 test data values for pullout strength for 3/8-inch-diameter bolts,13  ;

test data values for Sinch-diameter bolts, g test data values for 5/8 inch-diameter bolts, and 10 test data values for 3/4-inch- diameter bolts.

The average pullout strength from these tests, the allowable pullout strength (average pullout strength divided by 3, based on the concept of ample test data), the allowable pullout capacities from Table C.2 of GIP 2, and the CRFs (the allowable pullout strength divided by the allowable pullout capacity from GIP 2) for each diameter of bolts are calculated as shown in the table below.

s Diameter of Bolt 3/8"  %" 5/8" 3/4" Parameters Average Pullout Strength in 2.56 4.10 7.58 10.g4 Kips From Test Allowable Pullout Strength in 0.85 1.37 2.53 3.65 Kips From Test Allowable Pullout Capacity in 1.46 2.2g 3.17 4.6g Kips From GIP-2 i CRF 0.58 0.6 0.79 0.78 Since GIP 2 and Pre-GlP 3 chose only one single CRF for all kinds of diameters of Star Stud l

bolts, the value of 0.6 in GIP 2 appears to be reasonable, and the revised number of 0.75 in Pre- i i

GlP 3 is unacceptably high for 3/8-inch and inch diameter bolts.  !

The CRF of 0.75 in Pre-GIP-3 for Star Stud anchors was achieved through the use of a probabilistic method, which is contained in Reference 12. The staff reviewed the method and found in Reference 12 that the average pullout strength divided by 2 instead of 3 was used as ,

allowable pullout strength in the process of deriving or using the probabilistic method. This was

! - the cause for the high CRF value of 0.75.

[

k i

i

-,...~~....,m_.._,._____.... -,.m. . . - - _ -__ _w.,.- , - , . _ . _ . - , . . - . . . . , . , ,

1, I ,

6 ,

A conference call between the NRC staff and URS/Blume was held on April 2g, igg 6, to resolve the defference in opinions on the revised CRFs. In the conference call, the staff stated the  ;

. reasons for its objections to the revised CRFs, c. Yated above. The representative of  ;

URS/Blume stated that the higher CRFs in Pre CIP-3 were justWied because the new test data i

has lower coefficients of variations than those obtained from the test data previously used to

! ottain allowable anchor capacities in Table C.21 of GlP-2, and the mean strength /2 values i 1

should be used as allowable anchor capacities because the new test data had met the 2 percent l' non-exceedance critorion. However, both the arguments relating to coefficients of variations of test data and that of the 2 percent non-exceedance concept had been addressed in earlier ,

discussions with SQUG, which resulted in the conclusion to use the mean strength /3 as the allowable capacity for a single anchor. As a result, the NRC staff does not consider the 4

arguments presented by the representative of URS/Blume to be valid. Therefore, as stated in

- Reference 5, the NRC staff does not agree with either the revised numerical value of CRFs in

, Pre-GlP 3 (Reference 3) or the safety factor used in obtaining these values as described in -

Reference 12.-

In response to staffs evaluations in Reference 5, SQUG submitted the GIP-3 (Reference 6) and  !

revised the previous CRFs for some anchors. .

3.1.1.8 Conclusions and Recommendations The staff agrees with the proposed change in categorization to have different CRFs for pullout  :

4 and shear strengths, to use the ground rules in Reference 10 to calculate the CRFs for shear strength, and to assign two subsets of CRFs for two categories of diameters of bolts for Hitti-

Sleeve and Unknown anchors.  !

~

The revised pages for GIP 3, dated May 16,1997, have replaced the original proposed value of 0.75 in Pre GlP 3 with 0.60 for anchors of Hilti sleeve (for bolt diameter between % to 5/8 inch),

, Star steel, Star Stud. WEJ IT stud, and Unknown (for bolt diameter less than 3/8-inch). The staff found the value of 0.6 to be adequately conservative and acceptable, 3.1.2 in Cabinet Amplification Factor GlP 2 has been revised to tighten the restrictions on the use of a 4.5 in-cabinet amplification factor for substructures that are part of a c:ontrol panel or a benchboard. The required minimum ,

frequency of these substructures has been raised from 11 Hz to 13 Hz. The tightened restrictions also apply to control panels and benchboards evaluated using Reference 33 (GENRS computer code) of GIP 2,- The staff considers these changes in GlP 3 to be 1

, acceptable.

3.1.3 Anchorage Capacity for Horizontal Tanks and Heat Exchangers ,

GlP 2 has been revised to correct the equations for determining the allowable anchorage capacity for horizontal tanks and heat exchangers. This correction makes GIP-3 agree with

. EPRI Report NP-5228, Rwision 1. Volume 4, Guidelines for Tanks and Heat Exchangers."

The staff considers these corrections to be acceptable.

~w e. ...w.w-.-,-.a. ,---,c..m..n----- -1 w-s,r._ _ . ,rmv+.-r.w.,.sw- e. e .- .-4,.[..%&r m. . . -w-,-..-r.-,r-,- = - w,-wr _

_ . - - ,s.,

. l i

s 7 l 1

3.2 EDITORIAL CHANGES The staff agrees that the proposed additional sentenes on page 8 39 will enhance the screening f

= guidelines to evaluate the seismic adequacy of cable and conduit receway systems and accepts the proposed addition. Also, the staff agrees that the proposed change on page 8-28 will make  :

It clear how to calculate the scaling ratio and accepts the proposed change.

The staN finds that the proposed changes on pages 8 32,8-61,8 62,8-64 and 8-65 will . ,

enhance the screening guidelines to evaluate the seismic adequacy of cable and conduit  !

raceway systems, and accepts the proposed changes. -

The editorial changes clarify and interpret the intent of GIP 2. The stan has reviewed these changes and finds them accepteNo, 3.3 TYPOGRAPHICAL CHANGES I The typographical changes may appear to have no e#ect on the meaning or intent of GIP 2.

The staffs evaluation does not include the acceptance or approval of the references or the implied criteria.

4.0 CONCLUSION

S AND RECOMMENDATIONS' The staff has reviewed GlP 3 for Seismic Verification of Nuclear Power Plant Equipment, updated May 16,1997. The staff finds that the changes in the restrictions on the use of the in-cabinet amplification factor, the corrections in equations for determining the allowable anchorage capacity for horizontal tanks and heat exchangers, and the editorial and typographical changes ,

are acceptable. The staff also finds the changes in anchorage criteria as updated in GIP-3 are acceptablei Therefore. the staff concludes that, in general, GlP-3, supplemented by SSER No.  ;

2 dated May 22,1992, and the staff evaluations stated herein (SSER No. 3) constitutes an acco,m:able method for the implementation of the resolution of USl A 46 as specified in GL 87 02.

In spite of the conclusions reached in this very focused review of the GlP-3 revision, the staff has identified, during its review of licensees' documentations of A-46 implementations, certain -

areas which could result in misapplication of GlP 2 due to lack of specificity and the potential for misinterpretation of GlP 2 provisions relating to these areas. These issues are being resolved with SQUG separately.- The staffs position on these issues is discussed in the response to

- SQUG dated December 2,1997, and is applicable to GlP-3 as well as GIP-2 Consequently, the staff recommends that future revision of the GIP should include, as applicable, any necessary  ;

clarifications to the provisions, to reflect the staffs position, in areas identified as ambiguous or -

susceptible to misinterpretation by GIP users.

I

1.  ;

I l

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

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

C' 8

5.0 REFERENCES

1. " Generic Implementation Procedure (GlP) for Seismic Verification of Nuclear Power Plant -

Equipment," Revision 2, as corrected February 14,1992, SQUG, February 1992.

2. GL 87-02, Supplement 1, " Transmittal of Supplemental Safety Evaluation Report No. 2 (SSER No. 2) on SQUG Generic implementation Procedure, Revision 2, as corrected on February 14,1992 (GIP 2)," NRC, May 22,1992.

3. " Generic Implementation Procedure (GIP) for Seismic Verification of Nuclear Power Plant Equipment," Revision 3, SQUG, July 31,1995.

4. " Procedure for Revising the GIP," Revision 2 Enclosure 2 to SQUG (N. Smith) letter to NRC (R Zimmerman), dated July 31,1995.

5.

• Evaluation of Revision 3, dated July 31,1995, to tN Generic Implementation Procedure for Seismic Verification of Nuclear Power Plani Q.voent," NRC, August 6,1996.

i

6. ' Generic Implementation Procedure (GIP) for Seismic Verification of Nuclear Power Plant Equipment, Revision 3, Updated May 16,1997, and Procedure for Revising the GIP, Revision 3, dated May 16,1997.

7. NRC (W. Butler) letter to SQUG (N. Smith), dated June 23,1993, "NRC Response to the Seismic Qualification Utility Group, Revision 1 to Procedor for Revising the GIP Dated March 26,1993 (TAC M84091)."

8. SQUG (N. Smith) letter to NRC (W. Butler), dated August 30,1993, "SQUG Comments on NRC Response to SQUG Procedure for Revising the GIP."

9. EPRI Report No. NP 5228," Seismic Verification of Nuclear Plant Equipment Anchorage, Vol.1: Development of Anchorage Guidelines," Prepared by URS/Blume for EPRI, May 1987.

10. EPRI NP 5228 SL," Seismic Verification of Nuclear Plant Equipment Anchorage,"

Revision 1, Volume 1, Project 2925-1, Final Report, June 1991.

4

11. NRC Staff Presentation to the SQUG Staff on Anchor Bolts, by John S. Ma, January 30, 1989.

12. EPRI NP 5228 V1R1, " Addendum to Seismic Verification of Nuclear Plant Equipment Anchorage," Project 2925-01, Final Report, June 1994,

13. " GIP Revision 3 Proposed Knock Down Factors for Selected Expansion Anchor Bolts,"

Memorandum from R. Starck of MPR to SQUG Steering Group, April 11,1996.

Principal Contributors: Pel-Ying Chen, NRR John Ma, NRR ,

Yong Kim, NRR Date: December 3, 1997

_ _ _ _