ML20207P646

Rev 0 to Evaluation & Resolution of Generic Technical Issues for HVAC Sys (Including Ducts & Duct Supports)

ML20207P646
Person / Time
Site:	Comanche Peak
Issue date:	12/15/1986
From:	Harrison P, Hettinger F EBASCO SERVICES, INC., TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC)
To:
Shared Package
ML20207P636	List: ML20207P646 ML20207P648 ML20207P652 TXX-6148, Forwards Revs 0 to Evaluation & Resolution of Generic Technical Issues for HVAC Sys..., SAG.CP23, Seismic Design Criteria... & SAG.CP24 as Part of Facility Response Team Cable Tray & Conduit Support Program
References
1682R, NUDOCS 8701160360
Download: ML20207P646 (194)

Text

ENCLOSURE B i I I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

PREPARED BY:

I TEXAS UTILITIES GENERATING COMPANY GLEN ROSE, TEXAS EBASCO SERVICES INCORPORATED NEW YORK, NEW YORK I

REVISION 0 DECEMBER 15, 1986 1682P 0701160360 061223 P OR h % 0*50 oo s% q b

3%

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES I

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

REPORT APPROVAL COVER SHEET This report was prepared by Ebasco Services Incorporated. The signatures below verify the accuracy of this report.

Il l

I I

I I

l l

l Prepared l

Checked l Approved l l

l Revision l By l

By l

By l Date l Pages Affected l I

I I

I I

I I

l 0

lF H lR S AlexandrulR t

2/15/861 l

l l

l l

l l

l l

l I

I l

1 l

l

~

l l

1 I

I 2

I I

I I

I I

I I

I I

i 1.

I I

I I

l 3

I I

I I

I

)l l

l l

l l

l l

I I

I i

i l

I 4

I I

I I

I I

I I

I I

I l

I' 1

I I

i l

l 5

I I

I I

I I

I I

I I

I I

I I

I I

I I

l 6

l l

l l

l l

I I

I I

I I

I I

l l

l l

l l

l 7

l

\\

l l

l l

1 l

l l

l l

l ll I

1 8

I I

I I

I l

l l

l l

l l

I I

I I

i 1682R I

TEXAS UTILITIES GENERATING COMPANY CONANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF CENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

TABLE OF CONTENTS PAGE TITLE PAGE Re,0RT A,,s0 vat c0vER SeeET 1

TABLE OF CONTENTS 11 LIST OF GENERIC TECHNICAL ISSUES 111 I

INTRODUCTION 1

II BACKGROUND 2

III ISSUE RESOLUTION PROCESS 3

IV REPORT ORGANIZATION AND RESOLUTION PROCESS 4

V ABBREVIATIONS 6

VI REFERENCES 7

APPENDICES A1.1 - A42.3 I

I I

I I

I g

u 1682R

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION I

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

LIST OF GENERIC TECHNICAL ISSUES APPENDIX ISSUE TITLE 1

Controlling load Oise For Design 2

Seismic Response Combination Method 3

Anchor Bolt Design 4

Design Of Compression Members 5

Vertical And Transverse Loading On Longitudinal Type Supports 6

Support Frame Dead And Inertial Loads 7

Design Of Angle Braces Neglecting loading Eccentricity 8

Dynataic Amplification Factors (DAF) 9 Reduction In Member Section Propetsies Due To Bolt Holes 10 System Concept 11 Validity Of NASTRAN Models 12 Working Point Deviation Study 13 Reduced Spectral Accelerations 14 Non-Conformance With AISC Specificetions 15 Member Substitution 16 Weld Design And Specifications 17 Embedded Plate Design 18 System to Support Connections 19 FSAR Load Q)mbinations I

m g

1682R

L f

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION I

L EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

F L

LIST OF GENERIC TECHNICAL ISSUES (Cont'd)

I APPENDIX ISSUE TITLE L

20 Differences Between Installation And Design / Construction 7

Drawings Without Appropriate Documentation L

21 Design Control

[

22 Design Of Support No. 3136, Detail "5", CTH Drawing 2323-S-0905 23 Loading In STRESS Models 24 Design Of Flexural Members 25 System - Structural Qualification E

26 Base Angle Design

{

27 Support Quclification By Similarity 28 Critical Support Configurations And Ioadings 29 Q mulative Effect Of Review Issues 30 System Damping Values 31 Modeling of Boundary conditions 32 Concrete Voids

{

33 Gaps Betwe?n Duct And Duct Support 34 Attachment Of Transverse Supports To Ducts 35 Integrity of Duct Joints 36 Effects Of Openings In Ducts 37 Nonlinear Response Of Fire Damper Sleeve 38 Buckling Of Cantilever leg Of Base Angle 3

iv 1682R

{

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION s

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

[

LIST OF GENERIC TECHNICAL ISSUES (Cont'd)

APPENDIX ISSUE TITLE l

L 39 Anchor Bolt Perpendicularity Requirements 40 Measurement Of Embedment From Top Of Conctete Topping L

41 Bolt Hole Tolerance And Edge Distance Violation 42 Determination of Heat Ioads For Equipment Sizing Note: Appendicer 1 to 29 address CYGNA generic cable tray technical issues 1 F

to 29 presented in CYGNA's " Cable Tray Supports Review Issues List,"

L Revision 12 dated November 20, 1985 transmitted to TUGC0 by CYGNA letter No.

84056.095 dated November 26, 1965.

[

Appendices 30 and 31 address generic cable tray technical issuet 30 and 31 raised by CASE during the course of testimony relate.1 to varloas disciplines of CPSES design.

Appendices 32 to 39 address HVAC inrernall.y identified generic technical issues.

Appendices 40 and 41 address CYJNA generic conduit technie.a1 issues No.

4 and No. 5 presented in CYGNA's " Conduit Supports Review Issues List,"

Revision 3, dated November 20, 1985.

Appet. dix 42 addresses TERA's IRR No. DAP-E-M-504 and is the only issue presently known which 3o related to system's mechanical functional

{

performance.

E 4

E

/

.F

[

l v

1 1602R

[

l

~

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF CENERIC TECHNICtL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT GUPPORTS)

I I.

INTRODUCTION Texas Utilities Generating Company (TUGCO) has retained Ebasco Services Incorporated (Ebasco) to verify the design adequacy of HVAC systems for I

both Units No. 1 and No. 2, and to verify the design adequacy of 100 percent of the existing Seismic Category I HVAC ducts and their supports for Units No. 1 and No. 2, and to design and engineer any of the HVAC ducts and their supports that have yet to be installed in Unit No. 2 at I

the Comanche Peak Steam Electric Station (CPSES).

Project criteria and procedures have been developed for design and I

design verification by Ebasco in conformance with the CPSES Final Safety Analysis Report (TSAR) (Reference 12). These documents, along with various special studies and testing program resulta, provide the technical basis for the effort.

he criteria and procedures not only reflect good engineering practice, but have also been developed to specifically address any pertinent generic technical issues raised bf external organizations in their review of the HVAC and other CPSES I

systems (cable trays, electrical conduits and piping).

In addition, issues identified by the project (TUGC0/Ebasco) during the performance of this work are also considered.

The majority of the generic technical issues of relevance to this report applicable to ducts and duct supports were raised in an Independent Assessment Program (IAP) conducted by Cygna Energy Services (CYGNA).

I Rese issues were primarily focused on analysis and 6esign assumptions and methods, control of design documents, and differences between "as-designed" and "as-built" CPSES cable tray and conduit systems.

In I

addition. TUCCO undertook a thorough review of all hearing transcripts from 1982 to date; minutes of saeting.s between the NRC, CYGNA, the Intervenor (CASE), and the hird karty; documents denerated by the NRC, I

hird Party, and CASE; results of Third Party (ERC) effects frem Issue Specific Action Plans (ISAP's) of the CPRT Program Plan; and NCP's and potentially rapo-table 10CFR50.55(e) findings (SDG's) to determine if generic technical issues other than the CYGNA issues exist for HVAC I

ducts and their support system.

The generic nonstructural technical issues, pertinent to the HVAC I

systems, identified to date, were raised in the CPRT Design Adequacy Program conducted by the CPRT Third Party. Rese issues are summarized in the Issuo Resolution Reports (IRR's) prepared by TERA (Reference 20).

TUGC0 initiated the HVAC p:ogrcm to systematically and independently identify and resolve all pertinent issues.

Ebasco is responsible for implementation of the hVAC program, including "as-designed" and I

"as-built" data collection, analysis and design verification.

The HVAC program activities are being reviewed by the CPRT Third Party who was retained by TUGC0 to ensure that all issues have been clearly identified and resolved.

I 1726R

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

I This report contains a comprehensive summary of the Ebasco evaluation and resolution of all identified generic technical issues as they may apply to HVAC systems, and HVAC ducts and duct support. Specific references to the appropriate section of criteria, procedures, special I

studies or test results are provided for each generic technical issue.

II.

BACKGROUND To better understand how each of the issues discussed in the appendices has been evaluated and resolved, it is necessary to describe the HVAC program as it applies to the overall HVAC systems and the duct and duct I

support systems.

For a detailed discussion of the systems program see Reference 20.

In summary, however, the program includes all safety related HVAC systems and also systems which are important to power generation. As part of the program, Ebasco will develop design criteria for each system, prepare as built heat load calculations, perform consistency review of all design and licensing documents, and prepare I

design modifications, if required. All external issues and the hird Party issues will be resolved as a part of the program. The design criteria document will be based on the FSAR commitments, required system interf aces and the requirements of applicable codes and regulatory I

authorities.

he criteria document will provide a controlled basis for the execution of all de. sign verification and engineering activities.

I he as-built calculations will be performed utilizing as-built cooling /

heating loads, as-built flow rates and the design performance of installed systems / equipment.

Le calculations will consider all I

applicable heat sources corresponding to the system / plant mode of operation and will include heat transmission across space boundary, occupancy heat loads, thermally hot piping and equipment, and ele.:trical heat losses from lighting and cables.

The consistency review will encompass all design, procurement and licensing documents. Resolution of all identified inconsistencies will I

be appropriately reflected in the affected documents.

If the resolution of the inconsistencies should require design modifications, they would be performed by Ebasco as a part of the program.

For the duct and duct support systems structural evaluation, there is a difference in the engineering approach taken for the Unit No. 1 and Unit No. 2 systems.

Le HVAC systems in both units have safety and I

non-safety category types of supports.

In Unit No.1, all CPSES HVAC ducts and duct supports have been seismically designed.

In Unit No. 2, the HVAC ducts and duct supports have been installed based on their I

similarity to the Unit No. 1 design.

However, no seismic qualification has, as of yet, been generated for Unit No. 2.

This qualification will be performed by Ebasco as part of its HVAC design verification program.

In addition, some ducts and their supports in Unit No. 2 have tot yet I

been installed.

I I

2 1726R

TEXAS UTILITIES GENERATING COMPANY COMAhCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SY31EFS (INCLUDING DUCTS AND DUCT SUPPORTS)

I This report addresses all the HVAC ducts and duct s<spports in both Unit I

No.1 cod Unit No. 2.

The general approach taken in the resolution of the issacs covered in the appendices differ for the two units in the following way:

1.

In Unit No.1, the design of the HVAC duct and duct support system is goserned by CCL Report No. A-424 (Reference 1).

This report provides a detailed support-by-support qualification package which I

includes a computerized analysis for a majority of the HVAC duct supports. The support analyses in the A-424 report are predominantly based on "as-built" support " field sketches" generated I

by the duct fabricator / erector, the Bahnson Service Company (BSC).

HVAC duct routings were obtained from "as-built" general arrangement drawings which depict all CPSES HVAC ducts. These drawings were also generated by BSC.

No isometric drawings were created for the HVAC systems for either installation or inspection, i.e., the HVAC systems were field run and then "as-built." and qualified.

In Unit No. 1 therefore, the general approach is one of verification I

of adequacy of both the original "as-built" program and the analyses based on that program.

2.

In Unit No 2, the general approach is different.

Since the Unit I

No.

2 design and engineering have barely begun, the approach to resolve the generic technical issues is to build the resolution of all the issues into the upcoming engineering process.

III.

ISSUE RESOLUTION PROCESS I

The following steps were taken to evaluete and resolve each of the generic technical issues associated with the structural aspects of the HVAC system, i.e., ducts and duct supports:

1.

The CPSES cable tray, conduit, and piping program Generic Issues Reports (GIR's) were reviewed to determine which of the generic technical issues of each of these programs were pertinent to HVAC.

l 2.

Documentation was reviewed as required to fully understand the l

issue. Review of the IAP summary reports issued by CYGNA was

! g sufficient to accomplish this for most issues. More detailed 5

reviews of source documentation were re9uired to fu117 underatand other issues. The understanding of each issue was then summarized.

For the HVAC system's other aspects, similar steps have or will be I

taken to evaluates and resolve each of the technical issues as described in 3 below.

g 3.

Documentation has been or will be reviewed as required to fully 5

understand the issue. Review of the IRR summary reports issued by I

3 1726R c

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

I TERA may be sufficient to accomplish this for most issues. More I

detailed reviews of source documentation may also required to fully understand other issues.

4.

Then an action plan was or will be developed to resolve all issues I

requiring, on occasion, a specific study or test to resolve a specific issue. For the HVAC system, the action plan is briefly described in Paragraph II above.

For the ducts and duct supports, I

the action plan consists, for Unit No. 1, of verification of the original "as-built" design dc cument against the external issues and the design criteria commitmer.ts, procedures development, special I

study definition and implementation and test program implementation.

For Unit No. 2, the design of each HVAC duct and duct support system will incorporate the resolution of all the issues into the engineering process.

The appendices describe the action plan as it specifically applies to cach issue.

5.

The action plan has been implemented to complete the issue resolution process by preparing design criteria and procedural documents which address all of the identified generic technical I

design issues, to design verify and/or to design and engineer both Units' HVAC systems, to complete design verification of Unit No. 1 duct and duct support systems, c2d to design and engineer Unit No. 2 I

duct and duct support systems.

The design criteria and procedural documents are periodically updated to reflect results of generic studies and tests.

IV.

REIORT ORGANIZATION AND RESOLUTION PROCESS The body of this report describes the background and approach used by I

Ebasco to evaluate and resolve each of the identified generic technical issues for design verification of HVAC systems, and HVAC ducts and duct supports. An individual appendix has been developed for each generic I

technical issue identified as pertinent to the HVAC systems, and HVAC ducts and duct supports.

Each appendix includes a summary of the issue background. For Issues I

No.1 through 29, the background is taken verbatim from the CYGNA Cable Tray Review Issues List (Reference 8), and for Issues 40 and 41 the background is taken verbatim from the CYGNA Conduit Review Issues List

('leference 5), and thus, is specific to systems other than HVAC.

References noted in th' background sections of the appendices refer to the documents listed in Section 4.0 of the appendix.

For additional I

clarity, the words " cable tray" were added in a few locations in the background descriptions, and the background for Issue No. 20 was abbreviated because of its direct applicability only to cable tray hanger design details. Appendix 42 relates to the HVAC system as a I

whole and reflects the concerns raised in the process of the DAP review by the Third Party.

I 4

1726R

g TEXAS UTILITIES GENERATING COMPANY g

COMANCHE PEAX STEAM ELECTRIC STATION I

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

I Each appendix also explains the understanding of the issue as it applies to the system for which it was originally identified (cable tray or conduit).

In Appendices Nos. 1 through 29, the understanding of the issue, as it applies to cable tray systems, is extracted verbatim from I

I the Cable Tray System GIR (Reference 19).

In Appendices Nos. 40 and 41, the understanding of the it. sue, as it pertains to conduit, is extracted verbatim from the Conduit GIR (Reference 10). Appendix 42 applies directly and solely to liVAC systems.

In addition, each appendix, except Appendix 42; also outlines the understanding of the issue as it may or does apply to HVAC systems for I

Unit No. 1 work already performed. Ris is followed by the action plan to resolve the issue.

I A complete list of relevant documents reviewed by CYGNA or other reviewers (TERA) is also included in each appendix.

For Appendices Nos.

1 through 29, this list is taken verbatim from the Cable Tray System I

GIR.

For Appendices Nos. 40 and 41, it is taken verbatim from the Conduit GIR. For Appendix 42, the list is taken verbatim from the Third Party IRR.

The implementation of the resolution for each issue is also outlined in each appendix.

The implementation section contains detailed references to appropriate sections and revisions of criteria, procedures, special I

studies or test program documentation which resolve the generic technical issue. These references are listed in Section VI below.

This report has been prepared, reviewed and approved by Ebasco. This I

report will be revised on a periodic basis, as needed. Each revision will identify affected sections and will be reviewed.ind approved by Ebasco.

I I

I I

I I

5 1726R

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RJ.3OLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

I This report has been prepared, reviewed, and approved by Ebasco. Bis I

report will be revised on a periodic basis as needed.

Each revision will identify affected sections and will be reviewed and approved by Ebasco.

V.

ABBREVIATIONS The following abbreviations have been used in this report I

ACI:

American Concrete Institute AISC:

American Institute of Steel Construction I

AWS American Welding Society B&R:

Brown and Root BSC:

Bahnson Service Company CASE:

Citizens Association for Sound Energy I

CCL Corporate Consulting and Development Company, Ltd.

CMC:

Component Modification Card CPRT:

Comanche Peak Response Team I

CPSES: Comanche Peak Steam Electric Station CTH:

Cable Tray Hanger CYGNA:

CYGNA Energy Services I

DAF:

Dynamic Amplification Factor DAP:

Design Adequacy Program DCA:

Design Qiange Authorization DIR:

Discrepancy / Issue Resolution Report ERC:

Evaluation Research Corporation FSEG Field Structural Engineering Group FSAR:

Final Safety Analysis Report I

GIR:

Generic Issues Report HVAC:

Heating, Ventilation, and Air-Conditioning IAP Independent Assessment Program IRR:

Issue Resolution Report I

ISAP:

Issue Specific Action Program LOCA:

Ioss of Coolant Accident M0P:

Manual of Procedures MRM:

Multimode Response Multiplier NCR Nonconformance Report NRC:

United States Nuclear Regulatory Commission I

OBE Operating Basis Earthquake QC :

Quality Control SDAR:

Significant Deviation Analysis Report SRSS:

Square Root Sum of the Squares I

SSE:

Safe Shutdown Earthquake TERA:

Consulting Firm responsible for nird Party review TUGCO: Texas Utilities Generating Company VWAC:

Visual Weld Acceptance Criteria I

g 1ee2R

r L

TEXAS UTILITIES GENERATING COMPANY COMANCllE PEAK STEAM ELECTRIC STATION

[

EVALUATION AND RESOLUTION OF GENERIC TECIINICAL ISSUES FOR llVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS) r L

VI.

REFERENCES 1

Corporate Consulting and Development Company, Ltd., Report Number r

L A-424-81, " Seismic Qualification Report of Seismic Category I Ductwork and llangers for Comanche Peak Steam Electric Station,"

Revision 10, January 18, 1985.

2 Ebasco Services Incorporated, Document No. SAG.CP23, " Seismic Design Criteria for liVAC Duct and Duct Supports for Comanche Peak Steam Electric Station No. 1," Revision 0, December 15, 1986.

3 Ebasco Services Incorporated, Document No. SAG.CP24, " General Instructions for !!VAC Duct and Duct Support Analysis for Comanche

[

Peak Steam Electrical Station No.1," Revision 0, December 15, 1986.

4 Texas Utilities Generating Company Comanche Peak Steam Electric

{

Station, " Field Verification Method Procedure for Seismic Duct Hanger As-Built Verification in Unit 1 and Common Areas",

TNE-FVM-CS-029, Revision 1.

[

5 CYGNA, " Conduit Supports Review Issues List," Revision 3, dated November 20, 1985.

{

6 Ebasco Comanche Peak Steam Electric Station llVAC Duct and Duct Support Volume I.

The books included in Volume I contain documents consisting of the project design criteria, general instructions, design aids, studies, computer data and usage information.

7 Corporate Consulting and Development Company, Ltd., Report Number A-414-81, " Duct Test Evaluation Report for the Bahnson Company,"

(

February 19, 1982.

8 CYGNA, " Cable Tray Supports Review Issues List," Revision 12, dated

{

November 20, 1985.

9 Ebasco letter, Memo DSG-0039 from C R Levine to G W Kralik, dated September 17, 1986.

[

[

[

7

[

1682R

I TEXAS UTILITIES GENERATING COMPANY C0HANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

VI.

REFERENCES (Continued) 10 Texas Utilities Generating Company, Comanche Peak Steam Electric Station, " Evaluation and Resolution of Generic Technical Issues for Conduit and Conduit Supports," Ebasco Services Incorporated, Revision 0, September 1986.

11 Gibbs and Hill Specification 2323-SS-30, " Structural Ihbedment,"

Rev. 2, June 16,1986.

12 Texas Utilities Generating Company Comanche Peak Steam Electric Station Final Safety Analysis Report.

13 NCIG-01, " Visual Weld Acceptance Criteria," Revision 2, May 1985.

14 QI-QP 11.10-9 R5 (Unit 1).

15 QI-QP 11.10-2A R9 (Unit 2).

16 Bahnson Service = (;npany Letter Number 3231 f rom J. Phillips to F.

Thomas (CCL), September 25, 1986.

17 AISC, "Hanual of Steel Construction," 7th Edition, including Supplements Number 1, 2, and 3.

18 Ebasco CPSES Manual of Procedures (MOP). Appendix K.

19 Texas Utilities Generating Company, Comanche Peak Steam Electric Station, " Evaluation sud Resolution of Generic Technical Issues for Cable Tray Hangers," Ebasco Services Incorporated and Impe11 Corporation, Revision 0, August 8,1986.

20 Texas Utilities Cencrating Company, Comanche Peak Steam Electric Station, " Task Description HVAC Systems Engineering & Design" Ebasco Services Incorporated, TNE-TD-EB-028, Revision 0, October 1986.

21 Texas Utilities Generating Company, Comanche Peak Steam Electric Station TERA IRR No. DAP-E-M-504 " Determination of Heat loads for HVAC Equipment Sizing, " October 3,1986.

22 Texaa Utilities Generating Cospany, Comanche Peak Steam Electric I

Station, Comanche Peak Response Team Program Plan and Issue Specific Action Program, "ISAP VII.C".

8 1682R

_s_

u COMANCHE PEAK STEAM ELECIRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 1 1SSUE No. 1, CoNTmetL1No Leio c SE,em ES1oN g

I I

I I

I I

I I

I g

I I

I u.1 I

I APPENDIX 1 I

l l

ISSUE NO. 1: CONTROLLING LOAD CASE FOR DESIGN l

1.0 BACKGROUND

Gibbs & Hill used the equivalent static method to design the cable tray j

supports. In all load cases, the equivalent static acceterations used in designing the supports for SSE events are less than 160 percent of I

the corresponding accelerations for 1/2 SSE (OBE) events. Based on this finding and citing Section 3.8.4 of the CPSES FSAR which allows a 60 percent increase in allowables for structural steel between OBE and SSE I

events, Gibbs & Hill determined that the design was governed by the OBE event (see Reference 3).

I To validate this conclusion, the 60 percent increase in allowables must be liberally interpreted to be applicable to all support components rather than applicable only to structural steel as specified in the CPSES FSAR.

Catalog items such as Richmond Inserts and HILTI Kwik-bolts I

do not have increased allowables for SSE events. By designing these catalog components to the OBE event, the manufacturer's design factor of safety is not maintained for the SSE event.

Furthermore, for the design of structural steel, the 60 percent increase in allowables is acceptable for axial and strong-axis bending stresses in structural members.

The 60 percent increase cannot be applied to I

certain other allowable stresses. For example, the maximum increase in base plate stresses may be 33 percent, at which point the material yield is reached. A limit on maximum allowable stress is not provided in the I

FSAR.

These limitations were not considered in the selection of the governing seismic load case.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS OBE (1/2 SSE) was used as governing load for all cable tray supports, I

based on the ratio of corresponding SSE to OBE accelerations and Section 3.8.4 of the CPSES FSAR which allows a 60 percent increase in allowables for structural steel between the OBE and SSE events.

To validate this conclusion, the 6') percent increase in allowables must be interpreted to be applicable to all support components rather than only to structural steel. However, this increase is not applicable to I

some components, anchor bolts for example, and even to some structural steel stresses, base plate stresses for example.

I I

I A1.2 1682R

APPENDIX 1 ISSUE NO. 1: CONTROLLING LOAD CASE FOR DESIGN (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS I

In general, the HVAC design methodology for each component used maximum load values (SSE) with minimum allowable acceptance criteria (OBE) to determine structural adequacy per Section 3.3 of Reference 1.

If this approach proved too conservative, then SSE loads were compared to the I

SSE acceptance criteria and OBE loads were compared to OBE acceptance criteria.

Thus, this issue does not impact the HVAC system design.

3.0 ACTION PLAN TO RESOLVE THE ISSUE In general, both OBE and SSE conditions are separately design verified.

However, the original design methodology (see Section 2.2 above) is I

acceptable and may be used for the design verification of the HVAC systems. If the resultant stresses and loads for the SSE condition are within the OBE condition allowables, then the structural component will meet both the SSE and OBE condition requirements.

If the resulting stresses are within the SSE allowables but exceed the OBE allowables, then a separate analysis must be performed with OBE loads to establish I

OBE condition qualification. Appropriate OBE and SSE allowables for all HVAC system components including structural steel, welds, anchorages, and catalog items are considered.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED of CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-1010, Set 5, Sheets 16-20, Revision 5.

2.

Communications Report between P. Huang, S. Chang (Gibbs & Hill) and J. Russ and W. Horstman (CYGNA) dated November 13, 1984.

3.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 5, Sheets 1-7, Revision 1.

4.

CPSES FSAR, Sections 3.8.3 and 3.8.4.

t 5.0 IMPLEMENTATION OF THE RESOLUTION All HVAC supports and their components are design verified for the effects of OBE and SSE loads by either conservatively comparing SSE I

conditions to OBE allowables or by separate evaluation of the SSE and l

l OBE conditions.

This is performed in accordance with the loads, load I

combinations, safety factors, and acceptance criteria specified in I

Section III.4 and Section IV of Reference 2 and Attachment F of Reference 3.

I Al.3 1682R

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 2 I

2 '

""'""""c""^"""""

I I

I I

I I

I l

l I

I

'I I

A2.1 I

I a2

I APPENDIX 2 I

ISSUE NO. 2: SEISMIC RESPONSE COMBINATION METHOD

1.0 BACKGROUND

A.

Closely Spaced Modes (10% Modal Combination) in Spectral Analysis In the response spectra analyses performed for the Working Point Deviation Study (Reference 2), CYGNA noted that modal responses were not combined considering closely spaced modes as required by References 1 and 3.

B.

Inclusion of Dead Loads in SRSS Combination In all Gibbs & Hill design calculations, the acceleration due to dead weight is combined with the seismic accelerations using the SRSS method. A 1.0g dead weight acceleration is first added to the vertical seismic acceleration. The sum is then combined with the two horizontal seismic components using the SRSS method.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

Response spectra analyses for the working point deviation study did not combine responses considering closely spaced modes as required by the CPSES FSAR (Reference 12) and NRC Regulatory Guide 1.92.

B.

In all Gibbs & Hill design calculations, the acceleration due to dead weight is combined with the seismic accelerations using the I

SRSS method. A 1.0g dead weight acceleration is first added to the vertical seismic acceleration. The sum is then combined with the two horizontal seismic components using the SRSS method.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

In response spectra analyses, NRC Regulatory Guide 1.92 cites the I

requirements for combining modal responses.

The HVAC design calculations incorporated those requirements per Section 7.4.1 of Reference 1.

Thus, this issue does not impact the HVAC system design.

B.

In HVAC design calculations, 1.0g dead weight load was added absolutely to the SRSS of the seismic loads per Sections 3.2.1 and I

3.2.2 of Reference 1.

I A2.2 1682R I

L F(

A,PP,ENDIX 2 ISSUE NO. 2: SEISMIC RESPONSE COMBINATION METHOD (Cont'd) l' 3.0 ACTION PLAN TO RESOLVE THE ISSUE p

A.

For all response spectra analyses, the response combination method L

recommended by NRC Regulatory Guide 1.92 for closely spaced modes is used.

(

'B.

Dead weight is not included within any SRSS sum but is added separately to the SRSS combination of seismic effects due to the three orthogonal seismic components.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

CPSES FSAR Section 3.7B.2.7.

2.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Sets 2-6.

3.

US NRC Regulatory Guide 1.92, Revision 1.

4.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Design Review Questions," 84056.031, dated August 31, 1984.

5.

Gibbs & Hill Calculation in response to IAP Phase 2 questions, CYGNA Technical File 83090.11.2.1.50.

5.0 IMPLEMENTATION OF THE RESOIDTION A.

Section IV.3.c of Reference 2 specifies that NRC Regulatory Guide 1.92 is used for calculating modal responses for dynamic analysis performed using the response spectra method.

B.

Attachment F of Reference 3 specifies load combinations used for design verification of HVAC systems. Dead weight load is not included within the seismic SRSS but added separately in these combinations.

t

[

A2.3 1682R r

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN I

I I

.I I

I I

I I

I I

I I

n.1 E

lI 1682 1

APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN

1.0 BACKGROUND

I L

A.

Frame Connection Point and Anchor Bolt Pattern Centroid Eccentricity In the design for the anchor bolts, Gibbs & Hill did not properly

{

account for the eccentricity between the frame connection point to the base angle and the anchor bolt pattern centroid.

The moment due tc the eccentricity may cause the base angle to rotate about its p

longitudinal axis, resulting in:

(1) a compressive force along the L

toe of the angle section, and (2) additional tension in the anchor bolt (s).

(

B.

Safety Factor on HILTI Expansion Anchors at SSE Levels Gibbs & Hill's cable tray support designs employed a safety factor of 4.0 for HILTI expansion anchors for the 1/2 SSE load level. As discussed in Issue 1, the 1/2 SSE event was assumed to govern the support designs, without consideration of the reduced factor of safety on HILTI expansion anchors for the SSE event. The safety

[

factor for the SSE event will range from 2.5 to 3.0, depending on the installed location in the plant.

{

C.

Inconsistent Application of ACI 349-76, Appendix B Gibbs & Hill has used the provisions of Reference 12 to qualify several designs. Examples include the qualification of anchorages for Detail "11" (Gibbs & Hill Drawing 2323-S-0905, Reference 2) and the use of code provisions as justification for the factors of safety used for Richmond Inserts.

However, the designs do not

[

comply with other sections of ACI 349-76, Appendix B.

For example, Section B.7.3 states:

"A single expansion anchor used to anchor an attachment shall be designed for one-half of the design strength defined herein."

s For any of the cable tray

• support designs employing a single expansion anchor connectidn, this code provision would require a major reduction in the expansion anchor capacity.

(

CYGNA believes that the philosophy of the entire code appendix should be considered, rather than employing selected portions of the code.

[

t D.

Factor of Safety on Richmond Inserts

(

Gibbs & Hill's cable tray support designs employed a safety factor

(

of 3.0 for Richmond Inserts for the 1/2 SSE load level. As discussed in A3.2 1682R

)

h APPENDIX 3 r

ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd)

Issue 1, the 1/2 SSE event was assumed to govern the support designs, without consideration of the reduced factor of safety on Richmond Inserts for the SSE event. The safety factor for the SSE event will

[

range from 1.8 to 2.0, depending on the installed location in the L

plant. See Item 1.0C above for a discussion of ACI 349-76 as it has been applied to Richmond Inserts.

E.

Richmond Insert Design

1. Prying action was not considered in the original design of

(

Richmond Insert connections for cable tray supports. To qualify those connections which use Richmond Inserts, Gibbs & Hill performed calculations which reference the results of the f

Richmond Insert testing program performed at the CPSES site L.

(Reference 3).

These calculations showed that 1-in. diameter Richmond Inserts, originally designed with Ta=10.1 kips and Va=9.5 kips, were not the controlling anchorage type, but rather that the HILTI expansion anchors were the limiting case. CYGNA has the following comments regarding these calculations:

E The calculations do not account for the instances where the L

allowable values for 1-in. diameter Richmond Inserts taken from Gibbs & Hill Specification 2323-SS-30 (Ta=Va=ll.5 kips) p may have been used without the prying factor.

This situation L

could occur whenever a new design was performed after the issue of this specification or a CMC /DCA allowed a change which affected the Richmond Inserts used in a support

(

installation. Although Gibbs & Hill has stated that their engineers were instructed to include the prying factor, CYGNA could not locate any supporting documentation.

F CYGNA has concerns on the use of the site testing of Richmond Inserts to justify higher allowable loads than considered in the original design. See Pipe Support Review Issues List, Item 3, for additional detail.

2. The original design calculations for concrete connections using Richmond Inserts employed allowable values of tension (Ta=10.1 kips) and shear (Va=9.5 kips). With the issuance of Gibbs & Hill Specification 2323-SS-30, restrictions were placed on certain Richmond Insert allowables.

Decreases in allowable tensions and shears were provided for Richmond Inserts in cluster arrange-ments, Richmond Inserts embedded in the sides of concrete beams,

{

A3.3

(

1682R L

I APPENDIX 3 I

ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd) and Richmond Inserts used in spacings less than those originally considered in Gibbs & Hill designs. Since these restrictions I

were imposed after the original design of the Richmond Insert connections was completed, CYGNA is concerned that cable tray supports installed using Richmond Insert clusters or Richmond I

Inserts in the sides of concrete beams may not have been evaluated for the required reduction in allowables.

I In discussions with TUGCO, CYGNA was told that the Richmond Inserts in clusters were reserved for pipe whip restraints.

Authorization to attach to these clusters should have been E

obtained from the responsible TUGC0 group, and a corresponding g

evaluation of the installation should have been performed.

However, CYGNA could not locate any TUGC0 Quality Control instructions or procedures regarding the use of these Richmond Insert clusters (Reference 10).

F.

Connection Designs

1. The cable tray support designs use angles or plates at base connections.

The design drawings and associated design change documents (i.e., CMC /DCAs) specify anchor bolt spacing and member placement tolerances. However, these tolerances may be outside the origi2a1 design limits. Gibbs & Hill has not fully evaluated the effects of all possible installation tolerances on the base member st resses or the anchorages.

CYGNA's Phase 2 Observations CTS-00-05 and CTS-00-07 respectively addressed the design of base connections for Detail "E" supports I

with three-directional loadings and Details "A-D" base plate designs (Drawing 2323-El-0601-01-S). These support connection designs must also be reviewed to assure that the above concerns I

are addressed. For several additional support types considered in CYGNA's Phase 4 review, the installation tolerances allowed by the design drawings were not considered in the design calculations.

2. For most support types, the design drawings allow the use of either HILTI expansion anchors or Richmond Inserts for their I

anchorage to the concrete. For support types A, A, A s 1

2 4

Dy, D, Detail "A" (Drawing 2323-El-0700-01-S), and Detail 11 2

(wawing 2323-S-0905), the design calculations evaluate the I

attachments for HILTI expansion anchors, but not for Richmond Inserts.

I A3.4 1682R I

i APPENDIX 3 1

ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd)

G.

Justification of Prying Factor In response to Reference 11, Gibbs & Hill support designers used a factor of 1.5 to account for the effects of base angle / plate flexibility on anchor bolt tensile loads. The value of this factor I

is dependent on the applied load, bolt pattern geometry, and angle thickness.

Justification for the use of this factor has not been provided.

H.

Anchor Bolt Substitutions for Detail 1/1H and Details B, C, and D For Detail 1H (Gibbs & Hill Drawing 2323-S-0909), " Hanger Connection I

Using HILTI Bolts for Regular Cable Tray Supports," a substitution of Richmond Inserts for HILTI expansion anchors is allowed by Note 14d (Gibbs & Hill Drawing 2323-S-0901):

E

" Detail 1H (Drawing 2323-S-0909) Any HILTI bolt may be substituted with existing 1-in, diameter or 1-1/2-in. Richmond Insert except for the 1-1/4 x 13-1/8-in. Super Kwik-bolt which may be substituted only with 1-1/2-in. diameter Richmond Insert."

Additional information on the allowable bolt substitutions is I

provided in DCA 2103, Revision 0:

" Question: When only one Richmond Insert is available for a I

two-bolt hanger connection, may a combination of one Richmond Insert and one HILTI bolt be used? If so, what is the minimum ard maximum distance between the bolts, and what is the allowable to.7 erance?

Answer: Yes, combinations of Richmond Inserts and HILTI Super Kwik-bolts may be used. Minimum and maximum spacing between bolts shall be the same as used for the "a" dimension shown in

" Detail 1H, Two Bolt Hanger Connection," and the "a" and "b" dimensions shown in "Two Bolt Beam Connection." Tolerances shall be as shown in " Detail 1H," and in "Two Bolt Beam Connection."

I I

I A3.5 1682R

a d

W APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) u

1.0 BACKGROUND

(Cont'd) f The DCA expands the scope of the substitution to include the "Two N

Bolt Beam Connection" (Details B, C, and D on Gibbs & Hill Drawing 2323-S-0903), and does not include the restriction on the use of a p

1-1/2-in. diameter Richmond Insert as a substitute for the 1-1/4 in.

L x 13-1/8-in. HILTI Super Kwik-bolts.

These substitutions are inconsistent with several aspects of the r

}

cable tray support design calculations. The minimum bolt spacings are 12 in.,15 in., and 16 in, for 1-in. diameter HILTI Kwik-bolts, 1-1/4-in. diameter HILTI Super Kwik-bolts, and 1-in. diameter

[

Richmond Inserts, respectively. The tolerances specified for the L

connections employing only HILTI expansion anchors are different from the tolerances for the equivalent connection detail employing only Richmond Inserts. For moment loads on the base connections, the tensile load in each anchor is calculated by dividing the u

applied moment by the minimum bolt spacing. The tensile load y

distribution due to direct pullout is calculated based on the

[

allowed connection eccentricity. By substituting a Richmond Insert for a HILTI expansion anchor at the HILTI spacing and eccentricity, the tensile load in the Richmond Insert may be greater than the previously calculated load.

The effect of this substitution on Richmond Insert tensile loads has not been considered in the cable tray support designs.

In addition, since DCA 2103 does not limit the size of the Richmond Insert to be substituted for a 1-1/4-in. x 13-1/8-in. HILTI Super Kwik-bolt in the beam connection, a 1-in.

Richmond Insert, which has a lower capacity than the indicated Kwik-bolt, could be used as a substitute.

Gibbs & Hill /TUGC0 were not able to provide the design verification documentation for DCA 2103 (Reference 13).

I.

Base Angle Boundary Condition Assumptions p

For trapeze type supports, Gibbs & Hill has assumed that the hanger L

connections employing two-bolt base angles are free to rotate about the strong axis of the hanger. Since both the welds between the hanger and its base angle and the base angle itself have significant flexural stiffness, this assumption requires that the connection allow the calculated rotation without base connection failure.

Gibbs & Hill has not justified such connection behavior (see Review Issue 26).

[

[

A3.6 1682R E

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd)

J.

Installation of Expansion Anchors in Diamond Cored Holes Section 3.1.4.2.3 of Reference 4 discusses the reinstallation of an expansion bolt in an empty but " pre-used" hole. Paragraph (a) of that section states:

I "The bolt being replaced has been removed from the concrete using a Diamond core bit of the same nominal outside diameter as the I

replacement expansion bolt. The replacement bolt shall be one diameter size larger than the bolt being removed."

The HILTI " Architects and Engineers Design Manual" (Reference 5)

I addresses the bit type used in drilling holes for HILTI Kwik and Super Kwik-bolts. On page C-4, Note 6a states:

I "All of the technical information pertaining to Kwik-bolts herein (e.g., pullout.i shear data) was accomplished using HILTI masonry carbide bits. Before installing the Kwik-bolt using I

another means of drilling (e.g., Diamond Core), contact your local HILTI Field Engineer for advice and proper procedures."

On page C-1 of Reference 5, a footnote to the installation process I

description states:

"To obtain maximum published holding values, use only HILTI carbide bits."

In discussions with HILTI, Inc., CYGNA learned that HILTI expansion I

strengths that are less than those published in the HILTI Design anchors installed in core-bored holes will provide ultimate Manual. Primarily, the strength reduction is due to the diameter of the core bore bit itself. It has been HILTI's experience that core I

bore bits are intentionally supplied at a larger diameter than the nominal size to account for the progressive reduction in bit diameter over its life.

Thus, at the initial bit usage, the bit I

diameter will be larger than that required for the bolt hole. It is this hole oversize which causes the reduction in expansion anchor l

capacity.

I I

A3.7 1682R I

L

(

APPENDIX 3

{

ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd)

In order to avoid any such strength reductions, careful control on the bolt hole diameter must be established.

Control may be established by measuring the core bit diameter or the hole

(

diameter. CYGNA has not observed any QC procedures which impose such control. Additionally, CYGNA did not observe any procedures which require craf t or QC to docuser.t which expansion bolts were

{

installed in diamond cored holes.

K.

Reduced Allowable Loads for 1-in. Diameter HILTI Kwik-Bolts

(

Based on expansion anchor capacity tests performed by HILTI, Inc. in 1980, HILTI, Inc. issued a letter giving reduced ultimate capacities for 1-in. diameter Kwik-bolts. In response to this letter, TUGC0

(

issued a Significant Deficiency Analysis Report (SDAR) (Reference 6) to evaluate the effect of the reduced anchor bolt capacities for support installations at CPSES.

The resolution of this SDAR was to

[,

accept all existing designs employing 1-in. diameter Kwik-bolts by allowing a reduced safety factor of 3.41, and require that all future design efforts use the reduced capacity. The US NRC accepted this resolution (Reference 8).

['

For the review of cable tray supports where the cable tray load with Thermo-Lag exceeds che design load, Reference 7, Section 3.2.2.1,

{

paragraph (b) states:

"All hangers shall then be evaluated for actual loads. During this evaluation, all pertinent design changes shall be taken into account. Consideration shall be given to use of actual tolerances, weld undercut-undersize,1-in. diameter HILTI Kwik-bolt revised criteria and actual field "as-built" configuration."

However, CYGNA's review of the subject Gibbs & Hill calculations and

[

a discussion with TUGC0/Gibbs & Hill (Reference 9) verified that the original (unrevised) HILTI Kwik-bolt allowables had been used.

TUGCO/Gibbs & Hill felt that the use of the original allowables was warranted, since the calculations reviewed an existing design. This

[_

is not consistent with the requirements of Reference 7.

A3.8 1682R E_.

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE I

2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

I Design calculations for anchor boles did not address eccentricity between frame connection poin;; and anchor bolt pattern centroid, which may lead to compressive force along toe of angle section and additional anchor bolt tension.

B.

Design safety factor for HILTI expansion anchors may be violated at specific locations, considering the governing load case (OBE vs.

SSE).

C.

Criteria from ACI 349-76 Appendix B were used for design of some I

anchorages (anchorage for Detail 11, Richmond Insert safety factor) but were violated in other designs, for example, anchorages with single expansion anchors.

I D.

Design safety factor for Richmond Inserts may be violated at specific locations, considering the 3overnig load case (OBE vs.

SSE).

E.

Prying action was not considered in original design of Richmond Inserts. Subsequently, affected anchorages were qualified by calculations based on test data, which showed HILTI anchors were the controlling case rather than Richmond Inserts.

Calculations did not consider that Richmond allowables may have been used without prying factors af ter issue of Specification 2323-SS-30.

CYGNA has concerns on the use of site testing of Richmond Inserts to justify higher allowables.

Specification 2323-SS-30 places restrictions on Richmond allowables for certain installations; original designs for these installations I

may not have been reevaluated, for example, supports attached to Richmond Insert clusters without TUGC0 authorization.

I Support design options (base angles vs. plates, HILTI anchors vs.

F.

Richmond Inserts) and tolerances on anchor bolt spacing and member placement were not addressed in evaluation of member stresses and anchor loads.

G.

Prying factor, used for effect of base angle flexibility on anchorage tensile loads, was used without technical justification.

I Factor depends on applied load, bolt pattern geometry, and angle thickness.

I I

^3 9 1682R I

h APPINDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE ISSUE AE IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

H.

Fcr certain details on design drawings, drawing notes and design changes allow substitution of Richmond Inserts and HILTI expansion

(

anchors, and mixtures of the two; may be inconsistent with design calculations due to different minimum spacings, different tolerances, and different embedaent lengths for each type. As a

{-

result, allowable tensile loads in Inserts and HILTI bolts may be overestimated.

I.

For trapeze type supports with two-bolt base angles, design

[

calculations assumed free rotation of base angles about strong axis of hanger, ignoring stiffness of welds and base angle.

(

J.

Construction procedure allowed installation of HILTI expansion bolts in pre-used holes. Anchor strength may be decreased due to oversize core bore bit used to remove previous bolt. Bit diameter was not controlled during installation. No record exists as to which bolts

[

were installed in core-bored holes.

K.

Revised (reduced) allowables on HILTI bolts were not considered in

(

reevaluation of existing supports for addition of hermo-Lag insulation. Designs performed since revision should address reduced allowables.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

h e HVAC design calculations did not consider the additional

[-

compression along the toe of the angle section or the additional anchor bolt tension due to the eccentricity between the frame connection point and the anchor bolt.

B.

The HVAC design calculations used the SSE load combination as the governing load case. HILTI expansion anchors were evaluated using f

the SSE load case and the OBE safety factor of 5 and SSE safety L

factor of 4.

C.

ACI 349-76 was not used for concrete anchor bolt design for HVAC

(

supports.

hus, this issue does not impact the HVAC system design.

D.

he HVAC support design conservatively evaluated Richmond Insert

{

load capacities by using HILTI bolt allowable loads.

Thus, this issue does not impact the HVAC system design.

[

A3.10 1682R

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS (Cont'd)

E.

The HVAC design calculations did not include prying action in the design of Richmond Inserts. Some supports may have been attached to Richmond Insert clusters without TUGC0 authorization.

I F.

HVAC design calculations were based on the "as-built" support configuration.

"As-built" anchor bolt and base angle placement were I

considered in evaluation of bolt loads and member stresses. Thus, this issue does not impact the HVAC system design.

G.

The HVAC design did not calculate prying forces on anchor bolts.

It I

nas assumed that adequate conservatism existed in design calculations to account for prying effects. Technical justification was not provided for this assumption.

H.

The HVAC design calculations were based on "as-built" support configuration. Allowable anchorage loads were based on actual anchor type, location, and embedment. Thus, this issue does not I

impact the HVAC system design.

I.

The HVAC design calculations modeled each anchor bolt as fixed in I

translation and released for rotation. Welds between the support and the base angle, and the base angle stresses were not evaluated for the rotations permitted in the released direction.

J.

HVAC support construction did not install HILTI bolts in pre-used holes. Drill bit records, as described in Reference 16, show that no diamond core drill bits were used for HVAC HILTI bolt I

installation. Thus, this issue does not impact the HVAC system design.

I K.

The HVAC design calculations consistently used the reduced ultimate capacities for 1 in. diameter HILTI Kwik-bolts published in 1980.

Thus, this issue does not impact the HVAC system design.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

Offsets and eccentricities due to the assemblage of various types of I

structural members and transmission of loads are analytically considered in design verification. The eccentricity between the frame / base angle connection and the anchor bolt pattern centroid is I

included in the design verification. The analysis addresses the effects of the compressive force along the toe of the angle on both base angle and concrete stresses.

I I

A3.11 1682R

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd)

B.

The HVAC design verification evaluates the load capacity for HILTI expansion anchors based on a factor of safety of 4 for SSE load conditions and a factor of safety of 5 for OBE load conditions.

C.

ACI 349-76 is not used for the design verification of concrete anchor bolts for HVAC supports. Analytical techniques are used to I

design verify anchorage assemblies.

D.

The HVAC design verification evaluates the load capacity for I

Richmond inserts based on a factor of safety of 2.0 for SSE load conditions and a factor of safety of 3.0 for OBE load conditions.

I E.

Prying action is considered in the design verification of Richmond Inserts and HILTI bolts. This verification is performed based on "as-built" conditions and using Gibbs & Hill Specification 2323-SS-30.

I F.

Design verification of HVAC systems is based on "as-built" information. A study to address the effects of "as-built" valkdown tolerances has been performed.

G.

See Item 3.0E above.

H.

Seismic design evaluation of anchorages has been based on information shown on "as-built" support drawings.

Thus, any anchor substitutions and spacing violations are taken into account in design verification.

I.

Base angle boundary conditions are affected by several factors.

I These include angle size and thickness, bolt edge distance, bolt span, post locations, and concrete gap size and location.

Base angle assemblies are modeled at each bolt location. Assemblies I

consist of a short section of base angle with a given anchor bolt size and edge distance. Spring rates and allowable loads for each assembly are analytically determined.

Base angles spanning between anchor bolts are included as part of the support model.

J.

None K.

Design verification is performed using an anchor bolt design criteria which incorporates the reduced HILTI bolt allowables.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculations, " Evaluation of Detail 1, Single-Bolt Connection," CYGNA Technical File 84056.11.1.259.

A3.12 I

1682R

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd)

I 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont'd) 2.

Gibbs & Hill Calculation Binder SCS-212C, Set 7, Sheet 4-11, I

Revision O.

3.

Gibbs & Hill Calculations, " Justification of the Adequacy of 1" I

Richmond Inserts for the Effects of Prying Action," CYGNA Technical File 84056.11.1.219.

I 4.

Brown & Root Procedure CEI-20, " Installation of HILTI Drilled-In Bolts," Revision 9.

5.

HILTI, Inc., " Architects & Engineers Anchor and Fastener Design I

Manual."

6.

TUGC0 SDAR CP-80-12. " Reduced Allowable Loads for HILTI Kvik-bolts."

7.

TUGC0 Instructions CP-EI-4.0-49, " Evaluation of Thermo-Lag Fire Barrier Material on Class 1E Electrical Raceways," Revision 1.

8.

US NRC Inspection Reports 50-445/81-14; 50-446/81-14, dated 10/27/81.

9.

Communication Report between R.M. Kissinger (TUGCO), B.K. Bhujang et al. (Gibbs & Hill) and W.R. Horstman, et al. (CYGNA), dated 10/10/84.

10. N.H. Williams (CYGNA) to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 21, 1985.

11. United States Nuclear Regulatory Commission Office of Inspection and Enforcement, Information Notice 79-02.

12. American Concrete Institute, " Code Requirements for Nuclear Safety-Related Concrete Structures (ACI 349-76)."

13. Gibbs & Hill Interoffice Memo, T.D. Hawkins to M. Strange, dated 7/25/84.

5.0 IMPLEMENTATION OF THE RESOLUTION A.

Section IV.1.a of Reference 2 specifies that the anchor bolt tension due to offsets and eccentricities is considered in the HVAC design verification. Anchorage assembly design verification is performed as specified in Attachment G9 of Reference 3.

The basis of Attachment G9 will be documented in Reference 6 Book 3.

B.

Section IV.l.f.ii of Reference 2 specifies that safety factors of 4 and 5 are used for SSE and OBE load conditions, respectively.

I I

A3.13 1682R

I APPENDIX 3 ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 5.0 IMPLEMENTATIfN OF THE RESOLUTION (Cont'd)

C.

References 2 and 3 do not make reference to ACI 349-76.

Attachment G of Reference J and Section IV.1.f.ii and Appendix 2 of Reference 2 specify all information necessary for anchor bolt verification.

D.

Section IV.1.f.ii of Reference 2 specifies that Richmond Inserts are design verified for both OBE and SSE load combinations using the I

appropriate safety factors.

E.

"As-built" information is used for the design verification of I

Richmond Inserts as specified in Section IV.1.f.ii of Reference 2.

This verification is performed using the criteria established in Reference 11.

Section IV.1.f.ii of Reference 2 also specifies that prying action is considered in the design verification of Richmond I

Inserts.

Prying action is incorporated into the design verification procedures specified in Attachment G of Reference 3.

The basis of Attachment G will be documented in Reference 6 Book 3.

F.

Design verification of HVAC systems it, based on "as-built" information as specified in Sections III.1 and III.2 of Raference

, E 2.

Based on the study documented in Reference 6, Book 16, the 3

effects of the "as-built" walkdown tolerances are considered in the design verification per Attachment R of Reference 3.

G.

See Item 5.0E above.

H.

"As-built" information as specified in Section III.2 of Reference 2 is used to design verify cuprorts and their anchorages.

Inaccessible anchorage shall be evaluated par Attachment X of Reference 3.

I.

Section III of Reference 3 specifies the boundary condition modeling criteria for design verification. Anchorage assembly stiffness and allowable load values are specified in Attachment G9 of Reference 3 and Appendix 2 of Reference 2.

Prying action effects are an integral part of the anchorage assembly stiffness and allowable load determinations which will be documented in Reference 6 Book 3.

I J.

None K.

Appendix 2 of Reference 2 specifies the allowable loads for 1-in.

I diameter HILTI-Kwik bolts.

I I

A3.14 g

1ee2R

I APPEND 1X 3 I

ISSUE NO. 3: ANCHOR BOLT DESIGN (Cont'd) 5.0 IMPLEMENTATION OF THE RESOLUTION (Cont'd)

H.

"As-built" information as specified in Section IIT.2 of Reference 2 is used to design verify supports and their anchct ges.

I.

Section III of Reference 3 specifies the boundary condition modeling criteria for design verification.

Anchorage assembly stiffness and I

allowable load values are specified in Attachment G9 of Reference 3 and Appendix 2 of Reference 2.

Prying action effects are an integral part of the anchorage assembly stiffness and allowable load determinations which will be documented in Reference 6 Book 3.

J.

None i

K.

Appendix 2 of Reference 2 specifies the allowable loads for 1-in.

diameter HILTI-Kwik bolts.

I I

I I

I I

I

' I I

I A3.15 1682R

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 4 1SeeE No... DESIeN e,CoxPEESSzeN mE,ERS g

I I

I I

I I

I I

I I

I A4.1 1682R

I l

APPENDIX 4 ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS

1.0 BACKGROUND

A.

In the design of compression members for trapeze type support I

frames, Gibbs & Hill did not consider the entire unsupported length of the channels to calculate the slenderness ratios (Reference 1, Sheets 11 and 18 for support types A4 and B, respectively). If 4

the correct unsupported lengths and pinned end conditions are I

assumed, the slendernese ratio of these members for bending about their weak axis will exceed 200. AISC Specification Section 1.8.4 limits the slenderness ratio for compression members to 200.

B.

In calculating the slenderness ratio of the compression members for trapeze-type supports, Gibbs & Hill did r.ot check the effectiveness of the in-plana sidesway restraint for the various support designs.

C.

In the design of the compression member for cantilever type supports (e.g., SP-7, Details E, F, G, and H on Drawing 2323-El-0601-01-S,

I etc.) Gibbs & Hill used the distance from the face of the concrete to the centerline of the cable tray as the cantilever length. The correct length should be from the concrete face to the clamp in the far side of the tray.

A value of K=1.0 was used to calculate the minor axis slenderness ratio, rather than the value of K-2.0 for cantilevers. A value of I

K-1.0 is based on the assumption that the tray will provide lateral bracing at the clamp location.

The validity of this assumption is pending on the resolution of Review Issue 18.

D.

For the trapeze type supports, Gibbs & Hill has not considered the effect of weld undercut on the section properties of compression members at the point where in-plane braces are attached to the I

channel web.

As shown in the Working Point Deviation Study (Reference 2), high stresses exist in the region of the brace attachment and may increase if the reduced section propertier, are considered.

E.

The design of compression members assumed that the applied axial I

load was parallel to the member axis. Gibbs & Hill Installation Specification 2323-SS-16b allows an installation tolerance of 2 degrees from plumb for vertical members. CYGNA was unable to locate calculations considering the effect of this tolerance. See Reference 5 for a discussion of this issue.

F.

For trapeze type supports in the Working Point Deviation Study I

(Reference 2), Gibbs & Hill reduced the unsupported length of the hangers by 5 in.

This appears to be due to an assumption that the outstanding leg of the L5x5x3/4 base angle is rigid with respect I

A4.2 1682R

APPENDIX 4 ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS (Cont'd)

1.0 BACKGROUND

(Cont'd) to the C6x8.2 hanger. However, the minor axis moment of inertia for I

the L5x5x3/4; therefore, the buckling hinge would occur within the the C6x8.2 is greater than the corresponding moment of inertia for base angle rather than at a point in the hanger below the base angle, and the reduction in unsupported length is unwarranted.

G.

For the design of braces in compression, the axial force is a function of the brace slope. Gibbs & Hill designs provide a range I

of allowable brace slopes.

In some cases, Gibbs & Hill calculations check the brace for the slope which results in the largest axial load without considering other cases which have lower loads, but also have reduced capacity due to a longer member length.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

Design calculations for slenderness ratio of channel section compression members in trapeze type supports used unrealistic lengths and end conditions.

B.

Calculations for slenderness ratio of compression members in trapeze I

supports did not verify the effectiveness of in-plane sidesway restraint for various designs.

I C.

Design calculations for cantilever supports used a distance from concrete face to tray centerline for cantilever length, instead of the distance to the outside clamp.

Calculations for slenderness ratio of members in cantilever supports assumed the tray provides lateral bracing at clamp location; needs justification. This assumption requires justification if used.

D.

For trapeze supports, calculations did not consider reduction of section properties at in plane brace attachment points for weld undercut.

E.

Installation upecifications allow tolerance on plumbness of vertical members. Effect on axial load on compression members was not I

considered.

F.

In the Working Point Deviation Study, a reduced unsupported length I

for trapeze supports was used based on invalid assumption of rigidity of leg or base angle relative to hanger.

I A4.3 1682R

I APPENDIX 4 ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (0)nt'd) 2.1 UNDERSTANDi'NG OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

G.

Calculations for braces in compression find the highest load as a function of brace slope, and check for this slope without considering cases with lower loads where capacity is reduced due to longer member length.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

Design calculations for HVAC support members did not include consideration of AISC recommended maximum slenderness ratios.

B.

Slenderness ratios for compression members were not calculated for HVAC supports. No criteria was established for the effect of in-plane sidesway restraint on the effective length factor.

C.

The HVAC design calculations did not include slenderness ratio checks for support members.

D.

The HVAC design calculations did not consider weld undercut effects on member section properties.

E.

No allowable tolerance on plumbness of vertical members was used for HVAC supIort design. Effects of out-of-plumbness were not considered.

F.

The HVAC support design did not assume a reduction in unsupported member length due to the stiffness of the base angle leg for the purposes of checking member stability.

G.

The HVAC support design was provided on an as-built" basis.

Each member was individually evaluated for deaige loads based on the I

"as-built" length.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

Slenderness ratios (Kl/r) for compression members are limited to 200 in accordance with AISC Specification Section 1.8.4.

Member lengths are determined based on centroidal intersection points of connecting members. K and r values depend on duct attachment.

I A4.4 1682R

APPENDIX 4 I

ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd)

HVAC supports which have positive duct attachment are prevented from sidesway by axial and rotational stiffness of the duct.

The K value for this type of support is taken as unity in accordance with AISC Specification Section 1.8.2.

The r values for post members can be based on the orthogonal properties of the post angle since the shear I

strength of the duct wall prevents the post from buckling in the plane of the duct wall.

I Appropriate K values for HVAC supprts unattached to the duct are

{

developed on a case-by-case basis if required. The r values for unattached posts are based on the principal properties of the post angle unless the capturing steel of the support touches the duct at I

three corners of the duct. In this case, the torsional resistance of the duct allows for r values based on orthogonal properties.

Under certain conditions, HVAC vertical posts can be classified as tension members. A maximum slenderness ratio (1/r) limit of 300 is i

applied to these members. If a vertical post member is subject to I

static tension and if the combined static plus dynamic load is less than 50 percent of the design compression strength, the member is classified as a tension member. Regardless of the member classification, a full compressive stress check is performed in I

accordance with AISC Specification for any member subject to a compressive load regardless of the amplitude of the load and regardless of whether it is a static or dynamic load.

B.

The in plane sidesway restraint of HVAC supports attached to duct is provided by a combination of diagonal bracing, frame action of the capturing steel members, and torsional stiffness of the duct. Since I

the duct is a large closed section, it has hir,h torsional stiffness. This creates a " rotation fixed and translation free" type of end condition which has an AISC recommended design K value I

of 1.2 and a theoretical K value of 1.0.

The torsional stiffness of the duct in combination with the translational stiffness of the duct, diagonal bracing of the support, and frame action of the I

capture steel justifies a classification of "sidesway prevented" and allows the application of Section 1.8.2 of the AISC code.

The effectiveness of in plane sidesway restraint for unattached HVAC supports is considered on a a case-by-case basis if required.

C.

Design verification procedures are established which specify that the appropriate unsupported length of support members is used to l

calculate compressive member slenderness ratios.

Calculatione for i

A4.5 1682R

I APPENDIX 4 ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd) slendernese ratios include the effects of lateral bracing (i.e.,

out-of-plane bracing) from the duct.

D.

Reference 13, the Nuclear Construction Issues Group WAC, is I

incorporated into the CPSES HVAC Support Design Verification Program by TUGCO. Welds not meeting the WAC undercut requirement are not considered in the HVAC support design verification process since the members are identified in accordance with Section 3.1.2.B.15 of Reference 4, and subsequently dispositioned. Welds not meeting size and length WAC requirements are shown with actual size and length on "as-built" support drawings for evaluation of design adequacy.

E.

1hc effect of out-of-plumbness on HVAC supports is being evaluated.

Small to moderate deviations in plumbness have negligible impact on HVAC support load distribution and need not be included in design I

verification. Large deviations in plumbness may have some impact on design loads and are accounted for in design verification.

I F.

In the support design verification, the appropriate lengths of all types of members are used.

G.

The design adequacy of bracing members is evaluated using "as-built" I

information.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

1.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 1.

2.

Gibbs & Hill Calculation Binder 2323-SCS-215C, sets 2-6.

3.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.022, dated August 17, 1984, I

Question 4.

4.

Timoshenko and Gere, "Iheory of Elastic Stability," 2nd Edition, page 99 and 100.

5.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.041, dated February 12, 1985.

I I

I

..e 1e m

APPENDIX 4 ISSUE NO. 4: DESIGN OF COMPRESSION MEMBERS ((bat'd) 5.0 IMPLEMENTATION OF THE RESOLUTION I

A.

Design verification of support members is performed in accordance with AISC Specification 1.8.4.

The K values for supports attached I

to ducts are taken as unity in accordance with AISC Specification

1. 8. 2.

Effective K values used to calculate member slenderness ratios of supports unattached to duct will be developed on a case-by-case basis if reqitired.

B.

Same as Item 5.0A above.

I C.

Attachment El of Reference 3 specifies that the appropriate unsupported length of compressive members is used for calculating slenderness ratios. The K values used in design verification take into account the duct out-of plane resistance.

D.

Reference 13 specifies a visual criteria for acceptable weld size and length and undercut depth (base metal defect). Welds not I

meeting the size and length requirements have the actual size and length noted on the "as-built" drawing and are evaluated for design verification as described in Section IX of Reference 3.

Welds not I

meeting the undercut requirement will be identified per References 4,14 and 15 and subsequently dispositioned.

E.

The study to be documented in Reference 6, Book 10, describes the I

degree to which out-of-plumbness affects HVAC support loads.

Attachment E2 of Reference 3 specifes that out-of plumbness greater than two degrees must be considered in design verification.

I F.

As specified in Attachment El of Reference 3, member lengths to the f ace of the concrete are used in support design verification.

G.

As specified in Section III.2 of Reference 2 and Attachment El of Reference 3, "as-built" bracing configurations are used for design verification.

I I

I I

I 1682R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIT 5 ISSUE NO. 5: VERTICAL AND TRANSVERSE LOADING ON LONGITUDINAL TYPE SUPPORTS I

I I

I I

I 2

I

~

l I

~

I I

I I

A S.1 I

1682R I

l I

APPENDIX 5 ISSUE NO. 5: VERTICAL AND TRANSVERSE LOADING ON LONGITUDINAL TYPE SUPPORTS

1.0 BACKGROUND

Longitudinal trapeze type cable tray supports (e.g., L-A, L-A,

1 4

I L-C, etc.) were assumed to act independently of the transverse 4

supports (see Reference 4). Calculations for these longitudinal supports (Reference 1) consider only longitudinal loads in the design of frame members and anchor bolts. Since these supports are rigidly I

connected to the cable trays with " heavy duty clamps," a tributary tray mass will be associated with these supports. It is CYGNA's belief that these supports must be designed for vertical and possibly transverse I

seismic loads similar to the transverse supports (see References 2 and 3).

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS Longitudinal trapeze supports were assumed to act independently of the transverse supports.

Calculations for these supports considered only i

longitudinal loads in the design of frame acabers and anchor bolts.

I Trays are rigidly connected to these cupports with " heavy duty" three directional clamps. As a result, these supports should be designed for vertical and transverse seissic tray loads as well as longitudinal tray load.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS l

Longitudinal HVAC supports must withstand not only longitudinal loads, I

but also vertical and transverse seismic loads. Iongitudinal, vertical, and transverse loads were evaluated in the design of HVAC supports as described in Section 4.2 of Reference 1.

3.0 ACTION PLAN TO RESOLVE THE ISSUE Longitudinal HVAC supports are design verified for the simultaneous application of three orthogonal direction seismic-induced loads.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill calculation Binder 2323-SCS-1010, Set 2.

2.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.025, dated Augu,st 21, 1984, questions 3 and 4.

3.

R.E. Ballard (Gibbs & Hill) letter to N.H. Williams (CYGNA), GTN-69437, dated September 10, 1984, with attached calculations.

I l

AS.2

I 1682R I

I APPENDIX 5 ISSUE NO. 5: VEKfICAL AND TRANSVERSE LOADING ON LONGIIUDINAL TYPE SUPP0KIS (Cont'd) 4.0 LIsr OF RELEVAhT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont'd) 4.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 5.

5.0 IMPLEMENTATION OF THE RESOLUIION I

Attachment B3 of Reference 3 and Section IV.1.e of Reference 2 specify that simultaneous application of three orthogonal directions of seismic loading be considered for each longitudinal HVAC support.

I I

I I

I l8

~

I

I l I I

I A5.3 l $

l 5 1682R I

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 6 1SSee me. e, SU,,e,T,, axe EAD AN,I E.,IAL lea,S g

I I'

I I

I I

I I

I I

1 I

I A 6.1 I

1682R

I APPENDIX 6 ISSUE NO. 6: SUPPORT FRAME DEAD AND INERTIAL IDADS

1.0 BACKGROUND

A.

Out-of plane inertial loads (i.e., loads in the direction parallel to the cable tray) were not considered in the design of two-way I

cable tray supports. Such loads should, as a minimum, be considered in the design of base connections and anchorages. Assuming that tray clamps are able to transmit the loads from the two-way supports I

to the cable trays, out-of-plane inertial loads from the two-way supports must also be considered in the member and anchorage design of longitudinal supports (also see Review Issue 18).

B.

Gibbs & Hill did not consistently consider support dead loads. The support design calculations considered support weight in one of the following ways:

1. Support weight was not considered.

I

2. Support weight was considered as a surcharge on the tray, in addition to the tray and cable weight (usually, this value was given as 5 psf).

I

3. The support weight was calculated by considering the actual weight of each of the support's frame members.

I

4. A dead load equal to one half the support weight was used as required by Reference 1, Sheet 3.

Method (2) also led to other problems in the support design.

I Initially, the tray unit weight was considered as 35 psf. When the

" effective" support weight of 5 psf was added to the cable tray unit weight the result was a total assumed tray design load of 40 psf.

I At a later point in time, when design changes were issued against the supports or a revised analysis was required, the designer reduced the design weight from 40 psf to 37.5 paf, or even 35 psf, to remove some " conservatism" from the design loads in order to I

qualify the support. By doing so, the designer removed a portion of the support weight.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

Out-of plane inertial loads (parallel to tray) were not considered in design of two-way supports, e.g., in base connection and anchorage design. Out-of plane loads transmittsd from two-way I

supports through trays were not considered in member and anchorage design for longitudinal supports.

A6.2 1682R I

I APPENDIX 6

)

ISSUE NO. 6: SUPPORT FRAME DEAD AND INERTIAL LOADS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

I B.

Support loads due to support dead weight were not calculated consistently; e.g., support weight was ignored, treated as additional load in tray, calculated based on actual member weights, I

applied as dead load equal to one-half of weight. When design changes occurred, designers arbitrarily reduced the weight used in analyses.

2.2 JNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

Out-of plane inertial loads for transverse HVAC supports were I

calculated and included in the design of nearby longitudinal supports. The clamping action of the duct to the transverse support was assumed to be sufficient to transmit this load through the duct to the much stiffer longitudinal support.

B.

Dead weight loading was applied consistently for the HVAC design.

The support weight was included in the support model. Duct weight I

was calculated from the "as-built" span lengths and lumped on the model. The duct weights were calculated using linear weights developed from maximum duct joint segment length. Use of these I

linear weights may underestimate the total weight for duct segments less than the maximum length. Reduced section lengths require additional joining angles or flanges which increase the duct linear weight.

3.0 ACTION PLAN TO RESOLVE THE ISSUE I

A.

Out-of plane inertial loads of transverse HVAC supports are

~

considered for all aspects of the ductwork/ support systems.

i I

Design verification for HVAC systems includes dead weight loading of B.

the support.

Duct weights are calculated to include sheet steel, l

companion angles or flanges, reinforcing angles, and any insulation. Dead weight of duct accessories such as dampers and turning vanes are lumped in the model at appropriate locations.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 5, " Cable Tray Supports (Design Criteria and Reference)."

I I

^'

I 1

I APPINDIX 6 ISSUE NO. 6: SUPPORT FRAME DEAD AND INERTIAL LOADS (Cont'd) 5.0 IMPLEMENTATION OF THE RES0GTION A.

Sections I and II of Reference 3 describe the criteria for positive load transfer between supports and ductwork and also describes the I

methods to be used for distributing loads among the supports on a duct run. Se.: tion IV of Reference 2 specifies the analysis procedures to be used for design verification.

B.

Attachments B1, B2, and B3 of Reference 3 specify that the dead weight of the support and the dead weight of ducts and duct accessories are included in the design verification. A study will I.

be documented in Reference 6, Book 7, justifying the determination of duct dead weight loads used.

I I

I I

I 3

I I

I I

I 1682R I

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 7 ISSUE NO. 7: DESIGN OF ANGLE BRACES I

i NEGLECIIN7 LOADING ECCENTRICITY I

I I

I I

I I

lI llI iI

.I

'I I

A 7.1 I

1682R I

lI APPENDIX 7 ISSUE NO. 7: DESIGN OF ANGLE BRACES NEGLECTING LOADING ECCENTRICITY

1.0 BACKGROUND

A.

Longitudinal cable tray supports typically use angle sections as I

bracing to resist the longitudinal loads (e.g., SP-7 with brace, L-A, L-A, etc.).

For the member design, loads were assumed to 1

4 produce only axial stresses. The induced bending stresses due to I

the eccentric end connections were not considered. Neglecting these flexural stresses can result in members which are under-designed.

For certain longitudinal supports, double angles are required. The design assumes that the angles behave as a composite member.

I However, no intermittent filler plates are provided as required by AISC Specification Section 1.18.2.4.

Thus, the double angles must be considered to act independently.

B.

Transverse and longitudinal cable tray supports typically use angle sections as in-plane braces to resist transverse loads and provide bracing points on the vertical members (e.g., A3 A,B,B, I

4 3

4 L-A, etc.).

For the member design, loads were assumed to produce 4

only axial stresses. The induced bending stress due to eccentric end conditions were not considered. Though it is not explicitly I

stated in the AISC Specifications, it is standard practice (Reference 3, Sheet 3-59) to consider the bending strecses due to end connection eccentricity and check the interaction ratio considering the principal axes section moduli.

C.

Single longitudinal braces are typically connected to the frame by welding along the legs of the angle. Some brace designs provide I

welding on only one angle leg at one end of the brace; while, at the other end of the brace, welding is provided on the opposite angle leg. Such end conditions may lead to failure -by twist buckling at load levels below the critical value for Euler buckling.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISS AS IT APPLIES TO CABLE TRAY SYSTEMS A.

Induced bending stresses in longitudinal angle braces due to I

eccentric end connections were not considered. Double angles without filler plates were imprcperly considered as a composite member.

B.

Induced bending stresses in in plane angle braces due to eccentric end connections were not considered. Design calculations did not follow standard practice of checking interaction ratio considering I

principal axes section moduli.

I I

1,,,

I

I APPENDIX 7 ISSUE NO. 7: DESIGN OF ANGLE BRACES NEGLECTING LOADING ECCENTRICITY (Cont'd) 2.0

_ UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

C.

Some designs for single longitudinal angle braces provide welding only on one leg of section, at opposite ends of brace, potentially I

resulting in failure due to twist buckling.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS I.

A.

The HVAC design calculations did not include effects of eccentric end conditions on longitudinal angle braces. Double angles were not used.

B.

The HVAC design calculations did not include effects of ' eccentric end conditions on transverse angle bracec.

I C.

The HVAC design did not restrict the welding to only one leg section at opposite ends of an angle brace. Therefore, some designs for I

single longitudinal acgle braces provide welding only on one leg of section, at opposite ends of brace, potentially resulting in failure due to twist buckling.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

The effects of eccentric end connections for all angle braces are considered in support design verificatt< t.

B.

See Item 3.0A above. In addition, the difference between principal I

axes cection moduli and geometric axes section moduli are considered in checking bracing member stress interaction ratios.

C.

Except for large angle sections with short lengths, flexural I

buckling of angle sections is more critical than torsional buckling. A procedure has been developed to address torsional buckling in the design verification.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CART.R TRAY SYSTEMS 1.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray I

Support Review Questions," 84056.025, dated August 21, 1984, questions 3 and 4.

I 2.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.027, dated August 27, 1984, question 2.

I A7.3 1682R I

APPENDIX 7 I

ISSUE NO. 7: DESIGN OF ANGLE BRACES NEGLECTING LOADING ECCENTRICITY (2nt'd)

I 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA POR CABLE TRAY SYSTEMS (Cont'd) 3.

AISC Specification, 7th Edition, Sections 1.15.2 and 1.18.2.4.

4.

Gibbs & Hill Calculation, " Cable Tray Support Type SP-7 with Brace.

Brace Eccentricity Calculations." CYGNA Technical File 84056.11-1.228.

5.

Gibbs & Hill Calculation " Verify the Adequacy of Brace L3x3x3/8 of the Governing Support Case C." Gibbs & Hill Calculation Binder 3

I 2323-SCS-101C, Set 1 Revision 1, dated November 16, 1984.

6.

Gibbs & Hill Calculation " Justify the Use of Two L3-1/2x3-1/2x3/8 I

Angles to Take the Appropriate Load and Homent Individually in the Longitudinal Tray Supports at the Lower Brace." Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 2, Revision 6, dated September 15, 1984.

5.0 IMPLEMENTATION OF THE RESOLUTION I

A.

Sections I.b, I.d, IV, and Attachments E2, H, I, and J of Reference 3 provide direction regarding the consideration of eccentricities on all support members, including bracing.

B.

See Item 5.0A above. Also, the use of principal axes section moduli to check bracing member stress interaction ratios is standard practice and is noted in Attachment E2 of Reference 3.

C.

Instructio_s on how to address torsional buckling of brace angles are presented in Attachment V1 of Reference 3.

I I

!I

.I

~

l I

A7.4 1682R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECIL'iICAL ISSUES i

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPCRTS)

APPENDIX 8 ISSUE NO. 8: DYNAMIC AMPLIFICATION FACTOR (DAF)

I I

I I

I i

I I

I I

~

I I

I I

I A 8. i I

1682R I

r APPENDIX 8 ISSUE NO. 8: DYNAMIC AMPLIFICATION FACTORS (DAF)

1.0 BACKGROUND

Gibbs & Hill performed cable tray support designs using an " equivalent static analysis" to account for seismic loads.

The tray dead load on a

.g support was calculated by the tributary span method. The tray seismic g

load was the product of tray dead load and the peak spectral acceleration for the given buildings elevation. A dynamic amplification factor (DAF) was not included as required by Reference 7 (see also Issue 2 5.A).

An additional factor to be considered is the ratio of the static I-reaction for a continuous beam to the reaction calculated by the tributary span method. This ratio depends on the relative stiffness between the trays and supports, the relative stiffness between different support types and the number of continuous spans.

I 2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS Supports designed by equivalent static method (seienic loads) were not reevaluated for increased dynamic amplification factor (DAF) and for I

effect of ratio of continuous tray reactiun to tributary tray reaction.

DAF was not determined or justified for supports used with nonuniformly supported spans.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS

~

The HVAC support design which was performed by equivalent static methods used a dynamic amplification factor equal to 1.0.

No consideration was given to the relative stiffness between the duct and supports and the I

relative stiffness between adjacent supports.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A Multimodat Response Multiplier (MRM) of 1.5 on the acceleration corresponding to system frequencies to the right of the peak, and 1.5 times peak acceleration for system frequency to the left of the peak is I

used in " equivalent static analysis" of HVAC supports. This factor is appropriate, based on US }!RC Regulatory Guide 1.100.

For supports with a minimum system frequency greater than the cut-off frequency of 33 Hz, an MRM of 1.0 is used.

I

.I A8.2 I

1682R I

I APPENDIX 8 ISSUE NO. 8: DYNAMIC AMPLIFICATION FACTORS (DAP) (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd)

Seismic qualification is performed using either the Equivalcat Static Method, Equivalent Static Method with system type load redistribution I

effect considered, or Response Spectrum Method. The relative stiffnesses between duct and supports and between adjacent supports are automatically considered in the latter two methode using string models I

of the system.

Tributary span loads are used in '.;he Equivalent Static Method. The tributary span method is only used for simple, regular runs of duct supported by similar supports. Relative stiffness differences are small for this type of run.

4.0 LIST OF RELEV/RT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

Gibbs & Hill Report, " Justification of the Equivalent Static Load 1.

Method Using a Factor of 1.0 Times Peak Spectrum Acceleration for the Design of Cable Tray Supports; Comanche Peak Units 1 and 2."

2.

Communications Report between J. Jan (Gibbs & Hill), G.

Bjorkaan (CYGNA) dated October 4, 1984, 4:00 p.m.

I 3.

Communications Report between J. Jan, P. Huang, J. Pier (Gibbs &

Hill), N. Williams, G. Bjorkman (CYGNA) dated September 13, 1984, 3:30 p.m.

4.

Cbamunications Report between J. Jan, J. Pier (Gibbs & Hill), G.

Bjorkman (CYGNA) dated October 12, 1984, 10:00 a.m.

I 5.

Communications Report between J. Jan (Gibbs & Hill), G. Bjorkman (CYGNA) dated October 18, 1984.

I (basunications Report between J. Jan et al. (Gibbs & Hill), H.

6.

I4 vin (TERA), R. Kissinger, et al. (TUGCO), N. Williams, et al.

(CYGNA) dated October 31, 1984.

7.

CPSES, FSAR, Section 3. 7B.3.5.

5.0 IMPLEMENTATION OF THE RESOLUTION Procedures are specified in Sections IV.1.c and IV.2.c of Reference 2 regarding use of the 1.5 MRM in equivalent static analysis.

The I

position paper justifying the MRM is documented in Reference 6, Book 9.

I l

I A8.3 1

1682R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 9 I

DUE TO BOLT HOLES I

I I

I I

I I

~

I I

I I

I A9.1 I

1682R I

I IPPENDIX 9 I

ISSUE NO. 9: REDUCTION IN MEMBER SECTION PROPERTIES DUE TO BOLT HOLES The AISC Specification (Reference 3), Section 1.10.1 states:

" Riveted and welded plate girders, cover plated beams and rolled or w,1ded beams shall in general be proportioned by the moment of I

irertia of the gross section. No deduction shall be made for shop or field rivet or bolt holes in either flange, except that in cases where the reduction of the area of either flange by such holes, I-calculated in accordance with the provisions of Section 1.14.3, exceeds 15 percent of the gross flange area, the excess shall be deducted."

CYGNA found instances where the area of bolt hoics, used for the tray clamp bolts, exceeded 15 percent of the gross flange area, and the required reduction in moment of inertia had not been considered in the design calculations.

ll. 0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS Design calculations for channels did not properly consider reduction in moment of inertia due to bolt holes, per AISC. Effects of tray placement, hole tolerances, and unused holes were not considered.

I UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS 2.2 Bolted connections were used infrequently in HVAC designs. These I

connections were used primarily between the base angle and concrete surf aces and between support steel and duct. Unused bolt holes were not considered in the design.

3.0 ACTION PLAN TO RESOLVE THE ISSUE The reduction in beam section properties due to used and unused bolt holes is considered in the HVAC design verification.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYCNA FOR CABLE TRAY SYSTEMS 1.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray and Conduit Support Review Questions," 84056.015, dated August 6, 1984, Attachment B, question 2.

I A9.2 1682R I

~

I APPENDIX 9 ISSUE NO. 9: REDUCTION IN MEMBER SECTION PROPERTIES DUE TO BOLT HOLES (Cont'd) 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont'd)

I 2.

Gibbs & Hill letter GTN-69371, dated 8/23/84, Csiculation SCS-111C, Set 8, Sheets 34-39.

3.

AISC Specification for the Design, Fabrication and Erection of I

Structural Steel for Buildings, 7th Edition.

5.0 IMPLEMENTATION OF THE RESOLUTION Bolt holes identified in "as-built" data as specified in Sections 3.1.2.B.11, 3.1.2.B.12, and Attachment 12 of Reference 4 are considered I

according to the AISC Specification as required by Reference 3, Attachment E2.

I I

I I

I I

I I

I I

^

I 1,S m I

.I TEXAS UTILITIES GENERATING COMPANY 1

COMANCHE PEAK STEAM ELECTRIC STATION I

i EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCIS AND DUCT SUPPORTS) f APPENDIX 10 ISSUE NO.10: SYSTEM CONCEPT I

I I

I I

I I

I I

I I

I I

I A10.1 I

1682R I

I I

AP'PENDIX 10 ISSUE NO. 10: SYSTEM CONCEPT

1.0 BACKGROUND

In order to justify certain design assumptions questioned by CYGNA (Reference 1), documentation was provided indicating that Gibbs & Hill I

had assumed that the cable tray and supports act as a systes (Reference 2).

As part of this " systems" approach, the following behavior was assumed:

A.

The moments introduced by the eccentricities between the load application points (i.e., tray centroid) and the member resistant centroid were balanced by the load couples between adjacent I.

supports. More specifically, for longitudinal supports (e.g., SP-7 with brace, Detail 8, Drawing 2323-S-0903, etc.), the development of torsion in the beam due to longitudinal loading eccentricity is I

prevented due to the development of flexure in the cable tray. This tray moment is subsequently balanced by a vertical load, coupled between adjacent supports.

Similarly, torsion in the beam and the weak axis bending in the hanger due to the vertical load placement eccentricities ss well as

, the bending moment in the beam due to the transverse load placement

,I eccentricities are all balanced by either vertical or transverse load couples between adjacent supports.

Such moment transfers as described above are only possible if full I

rotational and translational compatibility exists between the cable tray and support beam. The relative stiffness between the trays and their supports can also effect the percentage of the moment to be I

balanced by the load couples between supports. Gibbs & Hill assumes

~

that the compatibility is provided by the heavy duty and friction types of tray clamps. See Review Issue 18 for a discussion of CYGNA's concerns regarding clamp behavior.

B.

In the design of trapeze support hanger members for compression loads, the trays provide lateral bracing at points along the length of the hanger. Similarly, for cantilever type supports, the. : ray provides lateral bracing to the beam (see Review Issue 4).

,E C.

For trapeze type supports, the longitudinal and transverse support 5

systems act independently. Therefore, the longitudinal supports are designed for longitudinal loads only, i.e., no transverse or vertical load contribution is considered (also see Review Issue 5).

D.

Additional tensile forces introduced by rotation of the base angles about the bolt pattern axis is minimized by the hanger attachment to the tray (also see Review Issue 3).

'I A10.2 I

1682R I

I APPENDIX 10 ISSUE NO. 10: SYSTEM CONCEPT (Cont'd) 1.0

_ BACKGROUND (Cont'd)

E.

For trapeze type supports, out-of plane seismic inertial loads from two-way support frames (self-weight excitation) are resisted by the I

longitudinal supports. However, as discussed in Review Issue 6, these inertial loads have not been considered in Gibbs & Hill's zdesign of longitudinal' supports.

F.

The cable tray supports use channel sections for the beam and hanger members. The typical connection between the beam and hanger is a lap joint, with the channels attached back-to-back. This type of I

connection will introduce bending moments and torsion in the members due to the eccentricity between the section neutral axes (Reference 1, Question 2.2).

Gibbs & Hill addressed this issue in Reference 2, it.dicating that a portion of the effect is resisted as additional loads in the cable tray, and the net effect on the stress level in the s1pport is less than a 3 percent increase.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYST A.

The " system concept," used to justify support design, assised the I

zoments and torsion in tray and hanger, due to tray load r eacement eccentricities, are balanced by load couples between ad jacent supports.

This is dependent on tray clamp behavior aucug other factors.

B.

In the " system concept," to justify design of trapeze support I

members and cantilver support membert for compression loads, trays were assumed to provide lateral bracing. This assumption requires justification.

I C.

For the design of longitudinal trapeze supports, transverse and vertical loads were not considered. Only longitudinal loads were considered.

D.

In support design it was assumed that tray attachment minimizes additional tensile force in anchor bolts due to rotation of base angles about bolt pattern axis.

E.

Magitudinal support design did not consider out,-of plane seismic inertial loads transmitted from self-weight excitation of two-way trapeze support frames.

I 1682R I

l APPENDIX 10 ISSUE NO.10: SYSTEM CONCEPT (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd) i F.

De support design did not adequately address bending and torsion due to eccentricity between section neutral axes in lap joint connections between beam and hanger members. The tray was assumed to resist a portion of additional loads caused by the eccentricity.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

The HVAC support design captures the duct on all sides of the duct.

He syssetry of the suppcrt connection to the duct reduces the bending moments in the duct due to longitudinal or transverse I

loading and eccentricities between the duct and its capturing steel members. Induced bending moments in the duct and support were not i

considered in the design.

B.

The HVAC design calculations did not establish a criterion for evaluating compression member effective length based upon lateral bracing of the support by the duct.

C.

The HVAC design calculations included the effects of transverse, longitudinal, and vertical loads in the evaluation of longitudinal supports. Thus, this issue does not impact the HVAC system design.

D.

The HVAC support design assumed sufficient load-bearing capacity in the connection between transverse supports and the duct to prevent significant out-of plane deflections for these supports. The effects of possible increases in bolt tension loads due to weak axis rotation of these suppcres was not considered.-

E.

De HVAC longitudinal support design included the effects of the out-of plane inertial loads from adjacent transverse supports.

Thus, this issue does not impact the HVAC system design.

F.

The HVAC support design calculations did not include the effects of eccentricitien between seaber neutral axes in modelirg connections I

between membere.

3.0 ACTION PLAN TO RESOLVE THE ISSUE l

A.

Design verification procedures conservatively consider the local l

compatibility and global distribution of loads throughoat the HVAC system.

I A10.4 1682R

I APPENDIX 10 ISSUE NO. 10: SYSTDi CONCEPT (Cont'd) 3.0 ACTION PLAN TO RESOLVE ThE ISSUE (Cont'd)

B.

In general, transverse supports will be positively connected to the supported ductwork in three directions such that compression in individual frame members may be evaluated at a K value of 1.0.

C.

Transverse, vertical, and longitudinal loads are considered in the design of longitedinal supports.

D.

See Items 3.0A and B above.

E.

See Item 3.0A and B above.

F.

The effects of eccentricities at lap joints are considered in the support design verification (See also Item 3.0A above).

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

1.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.031 dated August 31, 1984, Attachment A, question 2.

2.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA), dated September 28, 1984 with attached calculations.

5.0 IMPLDIENTATION OF THE RESOLUTION A.

The procedure for applying duct loads to the HVAC supports is specified in Attachments B1, B2, and B3 of Reference 3.

B.

In general, the positive connection between the transverse support and ductwork is provided by physical means per Section 3.1.2.B.10, I

3.3.1. A.3, and Attachment 15 of Reference 4.

Transverse supports not so connected will be evaluated on a case-by-case basis.

I.

K1/r, the effective slenderness ratio, will be limited to 200 for compression members and 1/r will be limited to 300 for tension members. A member is a tension member if it is normally in tension g

and the addition of a transitory seismic load creates a compressive E

stress of r.o more than 50 percent of the allowable compressive stress. Attachment El of Reference 3 specifies slenderness ratio procedures used in design verification.

' I A10.5 l

1682R I

I APPENDIX 10 ISSbE NO.10: SYSTEM CONCEPT (Cont'd) 5.0 IMPLEMENTATION OF THE RESOLUTION (Cont'd)

C.

Attachments B1, B2, and B3 of Reference 3 specify consideration of three directional loads for longitudinal support design verification.

D.

HVAC support anchorage components are evaluated for all applied loads, whether induced by'self-weight of the support or by restraint of duct loads. Procedures are referenced in Attachment G of Reference 3.

E.

Attachments B1, B2, and B3 of Reference 3 specifies the procedure I

for the consideration of the self-weight excitation of transverse supports.

F.

Section IV.1.a of Reference 2 and Sections I.b, I.d, IV Attachments E2, H, I, and J of Reference 3 specify that all eccentricities must be considered for the design verification.

I I

I I

I I

I I

lI

~

I g

.10. e 1682R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION I

EVALUATION t.ND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS) t APPENDIX 11 ISSUE NO. 11: VALIDITY OF NASTRAN MODELS I

I E

5 I

I I

I

~

l5

'E E

, I g

All.1 l-t 1682R I

I APPENDIX 11 ISSUE NO. 11: VALIDITY OF NASTRAN MODELS

1.0 BACKGROUND

CYGNA has questioned the validity of the NASTRAN models used in the Gibbs & Hill generic studies, such as the Working Point Deviation Study (Reference 1), the qualification of Detail Di References 2 and 3), and the Dynamic Amplification Factor Study (Reference 4). The analysis models consist of identical supports, separated by equal spans. This modeling will influence the system frequencies and seismic response and may not be representative of an actual installation, where a mixture of support types, non-uniform spans and tees or elbows in the tray are used.

2.0 UNDERSTANDING OF THE ISSUE I.

2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS NASTRAN models used in generic studies (Working Point Deviation Study, qualification of Detail D1, and dynamic amplification factor study) assume a row of identical supports, not representative of actual mixed g

supports and spans. System frequencies and seismic response may be 5

incorrect.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTl!MS Use of the NASTRAN model in support of generic studies is not germane to the design of HVAC supports. These generic studies have not been I

performed for HVAC supports. Thus, this issue does not impact the HVAC system design.

I 3.0 ACTION PLAN TO RESOLVE THE ISSUE t

This issue is specific to the generic studies performed by Gibbs & Hill and is not relevant to the present HVAC support requalification effort.

"As-built" support configurations and span lengths are used for the design verification. Appropriate configurations representing actual plant conditions are used for generic studies.

I 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Sets 2-6.

2.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 3, Sheets l

234-243, Revision 9.

3.

Gibbs & Hill Calculation Binder DMI-13C, Set 1.

4.

Gibbs & Hill Report, " Justification of the Equivalent Static Load I

Method Using a Factor of 1.0 Times Peak Spectrum Acceleration for the Design of Cable Tray Supports; Comanche Peak Units 1 and 2."

I l

All.2 l

1682R lI 1

APPENDIX 11 ISSUE NO. 11: VALIDITY OF NASTRAN !!0DELS (Cont'd)

E 5.0 IMPLEMENTATION OF THE RESOLUTION As specified in Sections III.1 and III.2 of Reference 2, "as-built" support configurations and span lengths are used for design s

verification. Appropriate configurations representing actual plant conditions are used for generic studies.

L a

r F5 L

E

[

A11.3 1602R r

c k

f TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 12 I

ISSUE FO. 12: WORKING POINT DEVIATION STUDY I

E E

[

r L,

e E

A12.1 1682R

I APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY

1.0 BACKGROUND

Cable tray supports employ angle sections as braces in the following configurations: in-plane for trapeze type supports, out-of plane for longitudinal trapeze supports, and in various other orientations for other support types. The original designs for supports assumed that neutral axes of all members at a connection intersected at a common I

point, thus no connection eccentricities were considered. The connection details shown on the design drawings (e.g., Details 4 and 5 on Reference 6) provided a brace working point location which was not consistent with the design assumptions.

Based on the discussion with TUGC0 personnel (Reference 5), CYGNA learned that the QC inspectors had difficulty in determining the design requirements for the working point locations, and Gibbs & Hill had been requested to provide clarification on the requirements and an allowable tolerance on the working point locations. DCA 20278 and DCA 20418 were I

issued in response, and the Working Point Deviation Study (References 1 and 2) was performed to consider the fact that the member neutral axes did not intersect at a common point and to provide the requested tolerances.

The following are comments on the analyses performed as part of this study.

A.

Gibbs & Hill study (References 1 and 2) does not fully consider the effects of previously approved design change documentation.

The analyses of the generic support types did not consider the effects of all generic design change documents which allow I

deviations from the original support designs (also see Review Issue 21.A.).

I Due to the overstress of certain components of several support types, a limiting spectral acceleration was calculated, and cut-off elevations were established using the individual floor response g

spectra.

Frames below the cut-off elevations were not checked for g

compliance with the study parameters. Frames above the cut-off elevation were analyzed on a case-by-case basis, but the analyses did not consider the effects of design change documents associated with the individual support.

B.

The effects of vertical and transverse loads on longitudinal support frames were not considered in the Working Point Deviation Study (also see Review Issues 5 and 10).

C.

The portion of the study that evaluated longitudinal trapeze I

supports only checked member stress interaction as specified in Section 1.6.1 5

1682R I

cL APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd)

1.0 BACKGROUND

(Cont'd) of Reference 3.

No evaluation was made to enraure that the 5

connections, base angles, and anchor bolts are also adequate.

y D.

Modeling Assumptions L'

1. Instead of modeling a longitudinal support in the tray run, one end of the tray was assumed to be fixed. The effect of this tray

{

boundary condition on the system response was not justified.

Based upon the review of the NASTRAN models used in the Dynamic Analysis Program (Reference 4), CYGNA learned that Gibbs & Hill's p

modeling of these fixed ends did not account for the response spectrum input at those points, but instead fixed them to an absolute rigid ground. If the same modeling technique was applied in the Working Point Deviation Study, the results of g

those response spectrue analyses may be incorrect.

2. The analysis assumed a single 24-in. tray per support level and P

did not assess the impact of more realistic multiple tray 5

loadings or other tray widths.

3. Eccentricities were not properly modeled (also see Review Issue 10).

4. The cable trays were modeled as translationally and rotationally l

fixed to support beams. This assumption of tray attachment fixity was not justified (also see Review Issue 18).

5. The run configurations selected may not be representative of actual installations. Parameters include systems of identical I

supports, uniform 8 ft.-6 in. support spacing, and the assumed worst case frame dimensions (also see Review Issues 11 and 28).

hs B

6. The base angle modeling assumed a simply supported beam for two-bolt base connections. In reality, the concrete reactions l

(prying actions) provide flexural restraint to the base angle (See also Issue 26).

I

7. Excitation in the longitudinal tray direction was not considered.

8. The out-of-plane translational degrees of freedom were restrained on trapeze type supports, resulting in an unrealistically

{

restrained system.

A12.3 1682R

I APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd)

1.0 BACKGROUND

(Cont'd)

E.

Gibbs & Hill did not check all support components when determining the controlling support element.

For example, support type E4 was I

assumed to be limited by the load capacity of the HILTI expansion anchors. CYGNA's review indicated that the actual governing component was the Richmond Inserts which were not checked by Gibbs &

Hill.

F.

Working Point Location for Two-Bolt Brace Connections on Iongitudinal Supports I

The working point location shown on the design drawing does not coincide with the actual line of action of the brace load for two-I bolt brace connections, e.g., Details "F" and "C" on Gibbs & Hill Drawing 2323-S-0903, and the brace concrete attachments for support types L-Al through L-A4, L-B, L-B, L-B, L-C, L-C,

1 2

4 1

2 and L-C4 on Gibbs & Hill Drawing 2323-S-0902. Rese offsets may 8

induce larger tensile loads in the anchorages than originally considered in the designs. Rese connections were not evaluated as part of the Working Point Deviation Study.

G.

Arbitrary Allowed Working Point Deviations I

Several support types within CYGNA's review scope have specified allowable working point deviations without any supporting calculations.

1. Detail N (Gibbs & Hill Drawing 2323-El-0601-01-S) Gibbs & Hill Calculation Binder 2323-SCS-216C, Set 3, Sheet 5 indicates an allowable deviation of 9" + 3" for brace connection to beam.

Calculations are not incluIed.

~

2. Detail V (Gibbs & Hill Drawing 2323-El-0601-01-S) Gibbs & Hill Calculation Binder 2323-SCS-216C, Set 3, Sheet 5 states " Low Stress, Brace Working Point Deviation of 6-in. is acceptable."

Calculations to support thic statement are not included.

H.

Working Point Deviations by Similarity Several support types within CYGNA's review scope have specified I

allowable working point deviations based on similarity to standard support types.

1. Detail J (Gibbs & Hill Drawing 2323-El-0601-01-3) is qualified by I

similarity to Case B.

3 I

A12.4 I

l l

1682R I

l APPENDIX 12 ISSUE No. 12: WORKING POINT DEVIATION STUDY (Cont'd)

A

1.0 BACKGROUND

(Cont'd) r

2. Detail 11 (Gibbs & Hill Drawing 2323-S-0905) is qualified by similarity to Detail 8 (Gibbs & Hill Drawing 2323-S-0903).

The calculations for Case B3 and Detail 8 (Gibbs & Hill Calculation Binders 2323-SCS-215C, Sets 2 and 4) indicate that these support types will be overstressed for the allowed working point deviation. Case-by-case evaluations of Case B3 and Detail 8 p

supports were performed to determine if all "as-designed" supports were acceptable. The support types which had been qualified by similarity were not included in these case-by-case reviews; hence, there is no assurance that they are not overstressed also.

k I.

Use of Enveloping Cases

{

The Working Point Deviation Study evaluates several support types by grouping them with an enveloping support of similar configuration.

Reference 1, Set 2 evaluates two groups. Group 1 includes Cases A, B, and C, considering Case C3 to envelope the other 3

3 3

two.

Group 2 includes Cases A4 B, and C, considering Case 4

4 C3 to envelope the others. For each analysis, the enveloping case is found to be overstressed, and a case-by-case "as-designed" review of supports of that type is conducted. The enveloped cases are not all included in the case-by-case reviews, and a separate evaluation is not performed to show design adequacy of the other support types on a generic basis.

J.

Compressive Load Capacity of Members As discussed in the status for Review Issue 4.A, Gibbs & Hill considered the effect of multiple, discrete axial loads on the buckling capacity of the hangers la response to CYGNA's concerns.

The same effect was considered in the member evaluations for this study. Gibbs & Hill did not properly apply the effect, since the factor is a function of the applied loading, and Gibbs & Hill did not calculate it for each load case (Reference 7).

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CART.R TRAY SY3TEMS A.

The Working Point Deviation Study for brace connection

[

eccentricities developed allowable working point deviations for L

generic acceptance of installed supports.

A12.5 1682R r

I APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

Ihe study did not incorporate effects of all design change notices for individual supports.

. E Alao. cut-off *1 vation= " ra t bli=h d 6==a on contro111=a a

spectral accelerations, and frames below cut-off elevations were not checked for effects of design changes.

B.

The Working Point Deviation Study for brace connection eccentricities did not consider the effects of vertical and transverse loads on longitudinal support frames.

C.

The Working Point Deviation Study for brace connection eccentricities in longitudinal supports only checked member stress

(

interaction; adequacy of connections, base angles, and anchor bolts was not evaluated.

D.

The Working Point Deviation Study for brace connection I

eccentricities incorporated the following questionable modeling assumptions:

E

1. Modelins trar ends as fixed instead of =odelins ionsitudinal

W supports.

2. Multiple tray ands locdings ignored.

3. Eccentricities improperly modeled.

f

4. Tray modeled as fixed to support by tray attachment.

5. Run configuration (systems of identical supports, uniform support I

spacing, worst case frame dimensions) not based on actual installations.

6. Base angle for two-bolt connection modeled as simply supported I

beam.

7. Excitation in longitudinal tray direction ignored.

8. On trapeze supports, out-of-plane translational degrees of f reedom were improperly restrained.

E.

In the Working Point Deviation Study for brace connection eccentricities, all support components were not checked to determine the governing component.

A12.6 1682R I

APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd) h 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) r 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS (Cont'd) 5 F.

The Working Point Deviation Study for brace connection eccentricities did not evaluate two-bolt brace connections on r

longitudinal supports, for which the working point location shown on D

design drawing does not coincide with line of action of brace load.

[

G.

As a result of the Working Point Deviation Study for brace connection eccentricities, allowable deviations were specified for some support types without supporting calculations.

F L

H.

In the Working Point Deviation Study for brace connection eccentricities, some supports were qualified by similarity to supports which were later found overstressed. Overstressed supports f

were included in case-by-case review, supports qualified by similarity were not.

I.

In the Working Point Deviation Study for brace connection eccentricities, enveloping support types were found overstressed and were evaluated case-by-case; other support types were not subsequently qualified.

J.

In the Working Point Deviation Study for brace connection eccentricities, the effect of multiple, discrete axial loads on buckling capacity of hangers was not properly considered as a function of applied loading and load case.

i 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS I

I A.

The HVAC design calculations were based on "a built" information and no generic acceptance criteria were used. Thus, this issue does l

not impact the HVAC system design.

B.

The HVAC design calculations included the effects of vertical and l

transverse loads on longitudinal supports. Thus, this issue does not impact the HVAC system design.

C.

'Ihe HVAC design calculations checked member stress (including base angle), weld stress, and anchor bolts for each support. Thus, this issue does not impact the HVAC system design.

D.

The Working Point Deviation Study for brace connection eccentricities need to consider the following sodeling assumptions:

A12.7 1682R

l I

APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS (Cont'd)

I

1. The HVAC ducts were modeled between supports for HVAC design calculations. Ducts were not assumed rigidly fixed to ground.

Thus, this issue does not impact the HVAC system design.

2. "As-built" duct spans were used for design. No assumptions were made concerning lengths or number of ducts attached to the support. Thus, this issue does not impact the HVAC system design.

I.

3. HVAC support member eccentricities were not properly modeled.

J

4. The HVAC ducts were not modeled as rotationally fixed to 3

transverse supports.

5. Run configurations were based on "as-built" duct and support information. Thus, this issue does not impact the HVAC system design.

6. HVAC support base angle was modeled as simply supported beam.

7. The excitation in the longitudinal direction of the duct was I

determined and included in the design calculations. Thus, this issue does not impact the HVAC system design.

8. Transverse HVAC supports were assumed to be two dimensional frames restrained in the out-of plane direction.

E.

Each HVAC support component was evaluated as part of the design I

calculations. Thus, this issue does not impact the HVAC system design.

F.

The HVAC design calculations did not include eccentricities for I

bolted connections.

G.

The HVAC design is based on "as-built" information. Allowable i

deviations were not necessary; however, eccentricities were not modeled.

E The HVAC design permitted a support to be qualified by similarity to H.

another support.

I 8

1682R I

APPENDIX 12

]

ISSUE NO.12: WORKING POINT DEVIATION STUDY (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYST2MS (Cont'd)

I.

'!he HVAC design calculations did not group support types for generic qualification of supports.

J.

The HVAC design calculations evaluated member buckling using the total combined applied load for the particular member.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

"As-built" information is used for HVAC design verification.

B.

Iongitudinal supports are design verified for the application of longitudinal, transverse, and vertical loads.

lI All members, connections, base angles, and anchor bolts are design C.

verified for adequacy.

l D.

Modeling assumptions listed as Items 2.2D 1-8 above are specifically I

addressed as the resolution of other issues.

l E.

All members, connections, base angles, and anchor bolts are individually design-verified for adequacy.

l F.

"As-built" support configurations and end connection details are used for support requalification. Offsets and eccentricities are I

appropriately considered.

1 I

G.

See Item 3.0F above.

H.

Each HVAC support is individually design verified and similarity is

}

not used.

I.

See Item 3.0H above.

I J.

The effect of multiple, discrete axial loads on buckling capacity of support members is evaluated in the design verification.

i l

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Sets 2-6.

l 2.

Gibbs & Hill Calculation Binder 2323-SCS-216C, Sets 1-5.

I A12.9 1682R I

I APPENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd) 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

(Cont'd)

3. AISC Specification for the Design, Fabrication and Erection of Structural Steel for Buildings, 7th Edition.

4. " Cable Tray Raceway System Dynamic Analysis Program," Gibbs &

Hill, March 19, 1985.

5. Communications Report between M. Warner (B&R/ TUG 00 QC) and W. Horstaan, J. Russ (CYGNA) dated November 16, 1985.

6. Gibbs & Hill Drawing 2323-S-0903.

I

7. Communications Report between B.K. Bhujang et al. (Gibbs & Hill),

R.M. Kissinger (TUGCO) and W. Horstaan et al. (CYGNA) dated September 14, 1984.

5.0 IMPLEMENTATION OF THE RESOLUTION A.

Sections III.1 and III.2 of Reference 2 specifies the use of "as-built" information for HVAC design verification.

B.

Attachments B1, B2, and B3 of Reference 3 specify the consideration I

of three directional loads for longitudinal support design verification.

C.

Section III.2.3 of Reference 2 specifies that all members, I

connections, base angles, and anchor bolts are design verified for adequacy.

I D.

Resolutions for Issues 3, 7, 10, 11, 18, and 28 address the modeling assumptions raised in this issue. Appendices 3, 7, 10, 11, 18, and 28 of this document specify the implementation of the resolution of these issues for the HVAC design verification.

E.

See Item 5.0C above.

I F.

Section IV.1.a of Reference 2 and Attachment E2 of Reference 3 requires that offsets and eccentricities be modeled and taken into account in the design verification.

G.

See Item 5.0F above.

H.

Section III.2 of Reference 2 specifies that all HVAC supports and their components be design verified.

I A12.10 1682R

E AP'PENDIX 12 ISSUE NO. 12: WORKING POINT DEVIATION STUDY (Cont'd) 5.0 IMPLEMENTATION OF THE RESOLUTION (Cont'd)

I.. See Item 5.0H above.

I J.

Support member lengths and effective length K values used to evaluate buckling capacity of support members with appropriate loads and load cases are specified in Attachment El of Reference 3.

I I-I I

I I

g I

I I

I I

I i

A12.11 1682R I

j l

0 I

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 13 1SSeE NO. 13,.EoecEo S,EcIEAL ACCEL,EiT10NS g

I I

I I

I I

I I

I I

I I

I m3.1 1684R I

I APPENDIX 13 ISSUE No. 13: REDUCED SPECIRAL ACCELERATIONS

1.0 BACKGROUND

For the qualification of the supports discussed below, Gibbs & Hill used reduced spectral accelerations based on a calculated support-tray systen I

frequency. These analyses assumed that all supports on a tray run are of the same type and have equal s' pacing (also see Review Issue 11). T hese studies are not representative of the cable tray installations at CPSES.

A.

A reduced acceleration was used for the analysis of transverse supports, such as type A4 which was used in analysis of Alternate Detail 1 (Reference 1). This acceleration corresponds to a I

calculated frequency which is higher than that corresponding to the spectral peak. This frequency was calculated using a system model of identical supports equally spaced at 8 ft.-6 in. and a tray weight of I

35 psf. The results of this study may not be valid for all installations as discussed in Review Issue 11.

B.

For longitudinal supports (e.g., type SP-7 with brace (Reference 3),

I L-Al (Reference 2, etc.), the frequency calculations did not include the effect of the axial frequency of the tray and the eccentricities between the tray and support.

C.

The flexural stiffness of the base angle supporting the brace of the longitudinal supports was not considered in frequency calculation (References 3, 4).

Flexural deformation of the base angle can result I

in significant reduction in support frequency.

2.0 UNDERSf ANDING OF THE ISSUE I

2.1 UNDERSIANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS I

A.

In calculation of reduced spectral accelerations for qualification of transverse supports, tray weight and tray span used did not represent, or envelope, all support installations.

j B.

In calculation of reduced spectral accelerations for qualification of longitudinal supports, frequency calculations did not include effects of axial frequency of tray and eccentricities between tray and support.

l C.

In calculation of reduced spectral accelerations for qualification of I

longitudinal supports, flexural stiffness of base angle supporting the brace of the support was not considered in frequency calculations.

I I

l ll3 A13.2 l

1684R I

I APPENDIX 13 ISSUE NO. 13: REDUCE.D SPECIRAL ACCELERATIONS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'o) 2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSTEMS I

A.

For HVAC components the hanger frequencies and corresponding system frequencies are based on "as-built" conditions. No reduction in spectral acceleration was taken.

i B.

The frequencies in each of the three directions for each longitudi-l nal support and associated duct span are based on "as-built" conditions which in turn are used to establish the spectral accelerations for design verification.

C.

The flexural stiffness of the base angles has been considered in the frequency calculations for longitudinal HVAC supports.

3.0 ACTION PLAN TO RESOLVE THE ISSUE I

A.

"As-built" conditions are used in design verification along with the consideration of load eccentricities and the flexural stiffness of hanger base angles for the calculation of support and system frequencies in each of the three directions.

B.

Same as Item 3.0A above.

C.

Base angle assemblies are modeled at each bolt location. Assemblies consist of a short section of base angle with a given anctor bolt size and edge distance. Spring rates and allowable loads for each I

assembly are analytically determined. Base angle spanning between anchor bolts is included as part of the support model.

I 1

4.0 LISr OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

l 1.

Gibbs & Hill Calculations, " Analysis of Alternate Detail 1."

I 2.

Gibbs & Hill Calculation Binder SCS-101C, Set 3, Sheet 247 Revision 9.

3.

Gibbs & Hill Calculation Binder SCS-215C, Set 4.

l 4.

Gibbs & Hill Calculation Binder SCS-101C, Set 2, Sheets 131 & 132, Revision 5.

I I

I A' % 3 1684R I

L APPENDIX 13 ISSUE No. 13: REDUCED SPECTRAL ACCELERATIONS (Cont'd) 5 5.0 IMPLEMENTATION OF THE RESOLUIION r

A.

The use of "as-built" conditions along with the consideration of load eccentricities and the flexural stiffness of hanger base angles for the calculation of support and system frequencies is described in Sections IV.2.c, III.1, and III.2 of Reference 2.

l" B.

Same as 5.0A above. Attachments B1, B2, and B3 of Reference 3 describe longitudinal support loading.

C.

Section III of Reference 3 specifies the boundary condition modeling criteria for design verification. Anchorage assembly stiffness and allowable load values are specified in Attachment G9 of Reference 3.

g Prying action effects are an integral part of the anchorage assembly stiffnesses and allowables.

EI t

1 I

1 A13.4 1684R

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WITH AISC SPECIFICATIONS I

I I-

\\

I I

I I

I lI I

I I

I I

A14.1

' I 1684R I

I APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WITH AISC SPECIFICATIONS

1.0 BACKGROUND

Reference 2 commits to designing the cable tray supports in accordance with Reference 1.

Gibbs & Hill has not properly considered the requirements of Reference 1, as discussed below.

A.

Unbraced Length for Axial Buckling Section 1.8.4 (Reference 1) requires that KL/R be less than 200 for compression members. Depending on the approach selected for the resolution of Review Issue 4, this requirement may not be met.

For I

example, if the friction type clamp cannot provide adequate restraint in the longitudinal direction, the K value should be taken as 2.0 for trapeze type and cantilever type supports. Consequently, KL/R=257 for a 5 ft.-9 in. C6x8.2 hanger or beam.

B.

Unbraced Length for Latera1 Torsional Buckling I

Section 1.5.1.4.6a (Reference 1) requires that Equation 1.5-7 be used to calculate the allowable bending stress for channels. In the denominator, "L" is the unbraced length of the compression flange.

I CYGNA found the following instances where the AISC Specifications were not considered or were improperly applied:

1.

Gibbs & Hill's Working Point Deviation Study (see Review Issue I

12) uses 22 kai for the allowable flexural stress without checking Equation 1.5-7.

Since the frame heights are on the order of 144 in., an allowable flexural stress of 15 kai is calculated by Equation 1.5.7.

2.

Detail SP-7 and similar supports consider "L" to be the distance from the base attachment to the tray centerline and not to the I

outside tray rail where the load is applied. Use of the larEer distance will result in lower allowable bending stresnes.

C.

Bolt Holes in Member Flanges Reductions in the section properties of beams due to bolt holes in I

their flanges per Section 1.10.1 (Reference 1), were not considered (see Review Issue 9).

l l

D.

Lacing of Double Angles Double angle braces are designed as composite members, without providing lacing per Section 1.18.2.4 (Reference 1), (also see Review Issue 7).

l

^"'

I 1 e4R

'I

E APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WIIH AISC SPI CIFICATIONS (Cont'd)

1.0 BACKGROUND

(Cont'd)

E.

Eccentric Connections Section 1.15.2 (Reference 1) discusses eccentric connections. This section requires that any axial members not meeting at a single l

working point be designed for the eccentricities. For example, this I

section of the specification applies to supports with single angle braces (SP-7 with brace, L-A, etc.). The gusset plates connected 1

to these braces must also be-designed for the eccentricities.

F.

Oversize Bolt Holes Section 1.23.4 (Reference 1) specifies bolt holes to be 1/16-in.

I larger than the nominal bolt diameter. The bolt holes for anchor bolts in base plates / angles (per Gibbs & Hill Drawing 2323-S-0903) and for tray clamps (per DCA 17838, Revision 8) are specified as I

1/8-in. larger than the nominal bolt diameter. Therefore, the bolt holes in Gibbs & Hill's designs should be considered oversized and should be treated as such in bearing connection calculations.

G.

Use of the Allowable Compressive Stress for Secondary Members For the design of the longitudinal brace for support type SP-7 with I

brace, the brace was assumed to be a secondary member, and allowable compressive stresses were calculated per Section 1.5.1.3.3 (Reference 1).

Since this is the sole member which provides longitudinal load I

carrying capability, it should be considered a primary member, and Sections 1.5.1.3.1 and 1.5.1.3.2 are applicable.

2.0 UNDERSIANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

AISC requirement on unbraced length for axial buckling (KL/R less than or enual to 200) may be violated for support compression members, due to inadequate restraint by tray clamps.

B.

AISC equation for bending stress in channels was improperly used or not considered.

Incorrect unbraced lengths resulted in overestimated allowables, I

i C.

Reductions in section properties of beams due to bolt holes in flanges were not considered as required by AISC.

t lI j

A14.3 i

1684R

i I

APPENDIX 14

)

ISSUE NG. 14: NON-CONFORMANCE WITH AISC SPECIFICATIONS (Cont'd) 2.0 UNDERSIANDING OF THE ISSUE (Cont'd) 2.1 UNDERSIANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSIEMS (Cont'd)

D.

Double angle braces were designed as composite members, but no lacing was provided as required by AISC.

E.

Design of supports using single angle braces did not consider eccentric connections as required by AISC (impacts axial members, gusset plates).

I F.

Bolt holes for anchor bolts and tray clamps were specified oversize with respect to AISC. Calculations did not consider oversize holes.

G.

AISC specification of allowable compressive stress for secondary members was improperly applied to the design of a longitudinal brace which is a primary member.

I 2.2 UNDERSIANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSIEMS I

A.

The HVAC design cr.lculations did not check angle members for conformance with AISC unbraced length recommendations on axial buckling.

B.

The HVAC design calculations did not check angle members for conformance with AISC unbraced len8th recommendations on lateral torsional buckling.

I C.

The HVAC design calculations did not coneider reductions in member section properties due to bolt holes. Fabrication and installation I

drawings specified neither bolt hole size nor tolerances for bolt holes.

l D.

The HVAC design did not use double angles as composite members.

I Thus, this issue does not impact the HVAC system design.

I j

E.

The HVAC design calculations did not consider the effects of l

eccentric connections for angle members.

F.

The HVAC design did not specify bolt hole sizes. Design calculations did not consider possible effects of oversized bolt holes.

G.

The HVAC design calculations considered all members as primary members. Thus, this issue does not impact the HVAC system design.

I I

A14.4 lI 1684R I

I APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WIIH AISC SPECIFICATIONS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

Slenderness ratios (K1/r) for compression members are limited to 200 I

in accordance with AISC Specification Section 1.8.4.

Member lengths are determined based on centroidal intersection points of connecting members. K and r values depend on duct attachment.

HVAC supports which have positive duct attachment are prevented fron sidesway by axial and rotational stiffness of the duct. The K value for this type of support is taken as unity in accordance with AISC I

Specification Section 1.8.2.

The r values for post members can be based on the orthogonal properties of the post angle since the shear strength of the duct wall prevents the post from buckling in the plane of the duct wall.

Appropriate K values for HVAC supports unattached to the duct are developed on a case-by-case basis if required. The r values for unattached posts are based on the principal properties of the post angle unless the capturing steel of the support touches the duct at three corners of the duct.

In this case, the torsional resistance of the duct allows for r values based on orthogonal properties.

Under certain conditions, HVAC vertical posts can be classified as tension members. A maximum slenderness ratio (1/r) limit of 300 is I

applied to these members.

If a vertical post member is subject to static tension and if the combined static plus dynamic load is less than 50 percent of the design compression strength, the member is I

classified as a tension member. Regardless of the member classification, a full compressive stress check is performed in accordance with the AISC Specification for any member subject to a I

compressive load regardless of the amplitude of the load and regardless of whether it is a static or dynamic load.

B.

AISC Equation 1.5-7 is applicable to channels. Since HVAC supports I

use structural angles, Section 1.5.1.4.6b of the AISC Specification is applicable. Justification of this applicability is addressed in a study.

C.

The reduction in beam section properties due to used and unused bolt holes is considered in the HVAC design verification.

D.

None o

A14.5 1684R I

I APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WIrH AISC SPECIFICATIONS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cout'd)

E.

All significant eccentricities are considered in the HVAC design I

verification of support members including haces.

F.

The AISC Specification is silent on the requirements for oversized anchor bolt holes. For bearing joints, oversize holes for 3/4-in.

I and 1-in. diameter bolts are commonly taken as 15/16 in and 1-1/4 in., respectively. The HVAC anchorage bolt holes are maintained at these diameters or less.

Bolts are used in four design details:

1.

Anchorage, 2.

Attachments between duct and supports, 3.

Flanged joints between duct sections, and 4.

Connections between support frame members.

The degree of oversizing of holes for anchor bolts is within commonly accepted practice for concrete anchors. AISC Specifications do not I

govern this particular joint detail. For the most part, holes for 3/4-in. and 1-in. diameter HILTI Kwik-bolts and HILTI Super Kwik-bolts and for 1-in. Richmond Inserts are 1/8-in. larger than the bolt diameter. The concrete anchors installed with these hole sizes are still capable of developing the required strength.

' Bolting designed to carry loads between the ductwork and supports is I

required only to handle shear forces. As such,' bearing and shear evaluation is handled in a ctraightforward manner, even for oversized holes.

Ductwork flanged joints incorporate closely spaced S/16-in. diameter bolts. In order to accommodate minor misaligt.nent at fit up, some holes were enlarged. The ' conservative number of bolts assures that I

occasional hole modifications have negligible effect on the structural integrity of the joints.

I The vast majority of support frame members are joined by welded connections. Only a few joints are bolted designs. These joints are evaluated in their "as-built" condition.

I I

A14.6 1684R I

APPENDIX 14 ISSUE NO. 14: NON-CONFORMANCE WITH I

AISC SPECIFICATIONS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd)

G.

All bracing members are design verified using AISC Specification primary member stress allowables.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

AISC Specifications for the Design, Fabrication and Erection of I

Structural Steel for Buildings, 7th Edition.

2.

CPSES, FSAR, Sections 3.8.3.2 and 3.8.4.2.

5.0 IMPLEMENTATION OF THE RESOLUTION A.

Design verification of support members is performed in accordance I

with AISC Specification 1.8.4.

K values for supports attached to ducts are taken as unity in accordance with AISC Specification 1.8.2.

Effective K values used to calculate member slenderness i

I ratios of supports unattached to duct will be developed on a case-by-case basis if required.

B.

I Allowable compressive bending stress for structural angles is taken as 0.60Fy in accordance with Section 1.5.1.4.6b of the AISC Specification.

Lateral bracing requirements of Section 1.5.1.4.6b are specified in Attachment V of Reference 3.

A study to justify I

this approach is documented in Reference 6 Book 6.

C.

Bolt holes identified in "as-built" data as specified in Section 3.1.2.B.ll, 3.1.2.B.12, and Attachment 12 of Reference 4 are considered according to the AISC Specification as required by Reference 3, Attachment E2.

D.

None E.

Section IV.l.a of Reference 2 and Sections I.b, I.d, IV, and I

Attachments E2, H, I, and J of Reference 3 provide instructions for incorporating eccentricities for support members, including braces.

I F.

The actual anchor bolt hole sizes are confirmed to be within the acceptable limits. These acceptable limits, if different than AISC, shall be justified by engineering analysis and/or test.

For ductwork to support connections, oversize holes for such bearing connections are evaluated in accordance with Reference 17.

I I

A14.7 I

1684R

L APPENDIX 14 ISSI'E NO.14: NON-CONFORMANCE WITH g

AISC SPECIFICATIONS (Cbut'd)

L 5.0 IMPLEMENTATION OF THE RESOLUTION ((but'd) t F

Bolted connections of support f rame members are evaluated in L

accordance with Section VIII of Reference 3.

p The effects of cversize bolt holes for bolted connections between duct sections (companion angle joints) is being evaluated and will be documented in Reference 6, Book 13.

G.

Section IV.1.f of Reference 2 specifies the allowable stress criteria used for all support members. All support members are evaluated as primary members in accordance with the AISC Specification.

L FL FL

[

A14.8 1684R

TEXAS UTILITIES GENERATZNG COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 15 ISSUE NO. 15: MEMBER SUBSTITUTION

.I I

I I

I I

I I

I I

A15.1 I

1684R I

APPENDIX 15 ISSUE NO. 15: MEMBER SUBSIITUIION 1.0 BACKGROJND Note 9 on Gibbs & Hill Drawing 2332-S-0901, Revision 4, stateF:

" Structural members shown on drawing numbers 2323-S-900 series cay be substituted by one step heavier chape of the same size."

This note allows c aft to substitute a member from one series with a I

member from another serius, e.g., an American Standard Channel (C) for a Miscellaneous Channel (MC) or vice versa, as long as the substituted shape is heavier than, but of the same depth as the original member.

CYGNA is concernr* that this note allows the use of subi.titute sections which are heavier, bet have lower section moduli.

At a later date, Reference 2 was issued, providing the following I

a clarification:

" Structural members shown on drawing numbers 2323-S-900 series may be I

substituted by a member of the same size and next heavier shape determined by the material on site. The next step heavier shape will be governed by section:: as shown in AISC Manual of Steel Construction. Examples are shown on sheet 2 of 2."

The examples shown on Sheet 2 of Reference 2 include the substitution of a C4x7.25 for a C4x5.4, a C6x10.5 for a C6x8.2, etc. This clearly I

indicates that the substitution should be of the same aeries as the specified member.

I CYGNA's concern is what types of substitutions were performed by the craft and accepted by the QC inspectors during the time between the issuance of Reference 3 and Reference 2.

CYGNA was unable to locate any requirements for documenting member substitutions.

Within CYGNA's walkdown scope, such a substitutio1 was identified for support number 5654 (see Review Issue 20). The design required an I

MC6x12, and the installed member was a C6x13, which has a smaller section modulus (S=5.80 in3 for a C6x13 compared to S=6.24 in3 for an MC6x12). For the other supports listed in Review Issue 20, the required MC6x12s were substituted with C6x8.2s, a substitution not permitted by Reference 2.

I I

1684R I

l l

APPENDIX 15 ISSUE NO. 15: MEMBER SUBSIITUTION (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS IT APPLIES TO COLE TRAY SYSIEMS Design specification allowed substitution of structural members with members having potentially lower section moduli. Documentation of I

substitutions was inadequate.

2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSTEMS l

Support qualification was based on "as-built" member configurations.

Thus, this issue does not impact HVAC system design.

3.0 ACTION PLAN TO RESOLVE THE ISSUE HVAC supports are design verified based on "as-built" member configurations.

4.0 LISI 0F RELEVANT DOCUMENIS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS L

1.

Communications Reports between R.M. Kissinger (TUGCO) and J. Russ (CYGNA), dated January 17, 1985, 8:15 a.m. and 3:45 p.m.

2.

CMC 69335, Revision 1, dated 9/21/82.

3.

Gibbs & Hill Drawing 2323-S-0901, Revision 4.

5.O IMPLEMENTATION OF THE RESOLUrION Design verification of HVAC supports is based on "as-built" member configurations as specified in Section III.2 of Reference 2.

A15.3 1684R

TEXAS UTILITIES, GENERATING COMPANY COMANCHE PEAK STEAM ELECIRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 16 ISSUE NO. 16: WELD DESIGN AND SPECIFICATIONS l

=

A16.1

[

1684R

[

APPENDIX 16 ISSUE NO. 16: WELD DESIGN AND SPECIFICAIIONS l

\\

1.0 BACKGROUND

r CYGNA has noted the following discrepancies in the weld designs for cable tray supports.

A.

The design drawings are missing the weld details for several support l

types as described in Reference 1, Attachment C.

B.

Per discussions with Gibbs & Hill /IUGC0 (References 2, 3, 4, and 5),

CYGNA has noted that the weld sizes shown on the assembly drawings

(

differ from those shown on the design drawings and those that were assumed in Gibbs & Hill calculations.

F L

C.

Eccentricities were not considered in weld connections.

1.

Detail SP-7 with brace and similar connections requires a partial

(

penetration groove weld at the gusset plate / beam connection. The design calculations did not consider the eccentric load transfer i

fron the brace member. The eccentricity of the brace loads results in a weld stress in excess of the allowable.

2.

Weld designs for the lap joints between channels and between the base angle and attached channel did not consider the eccentricity between the applied loads from the connecting members and the plane of the weld.

D.

The weld designs did not consider the thickness of the connected parts. This issue was identified by DCA 2365, Revision 2, but was never considered in the design calculations. Gibbs & Hill's weld designs assumed that the full weld throat would be developed without considering the thickness of the connected member. For example, the weld size for support designs employing C6x8.2 channels with a fillet weld crossing the web of the channel is limited to the 0.2 inch web thickness. Gibbs & Hill designs specified a 5/16-in. fillet weld size and did not reduce the throat to account for the minimum l

material thickness. Cases where this may be a problem include:

Details E, F, G, H, J, and K on Gibbs & Hill Drawing 2323-El-0601-01-S; SP-7 using an L6x4x3/4 base angle; and the Detail 2/2A on Gibbs & Hill Drawing 2323-S-0903 as modified per CMC 58338.

E.

Gibbs & Hill assumed an incorrect minimum weld length for the beam / hanger base angle connection. Gibbs & Hill assumed a weld length of 1-k, where 1= angle leg width and k= distance from back of angle leg to end of fillet. However, because of the existe:ca of the curve with radius, r, (approximately equal to one-half the ~eg thickness), at the angle toe, the actual weld length is 1-k-r.

A16.2 1684R

t APPENDIX 16 g

ISSUE NO. 16: WELD DESIGN AND SPECIFICATIONS (Cont'd) l 2.0 UNDTRSIANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSIEMS A.

Weld details were not provided on design drawings for several supports.

B.

Different veld sizes were shown on assembly drawings, design drawings, and in calculations.

Ih C.

Eccentric loads were not considered in design of welds for brace /

gusset plate / base connections, lap joint connections between P

channels, base angle connections.

b l

D.

Thicknesses of connected members were not considered in weld designs. Specified designs may have excessive weld throat.

H E.

Design calculations assumed an incorrect minimum length for beam /

hanger base angle connection, due to radius on angle leg.

i~

2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSIEMS A.

Weld details were missing from HVAC support design drawings.

B.

Welding discrepancies between "as-built" and "as-designed" were found for HVAC supports.

L C.

The HVAC design calculations did not completely and consistently account for eccentric loads on welded connections.

D.

The HVAC design calculations correctly considered thickness in weld design for structural members. However, the thickness for weldment of sheet steel to structural steel was not correctly considered.

E.

The HVAC design calculations assumed a conservative weld length based i

on 75 percent of the "as-built" weld length.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

This issue is not applicable. Design verification is based on weld details from "as-built" support drawings. An approach to evaluate inaccessible welds is being developed.

B.

Same as Item 3.0A above.

A16.3 1684R

APPENDIX 16 ISSUE NO. 16: WELD DESIGN AED SPECIFICATIONS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd)

C.

Eccentric loads are considered in the design verification of all

[

welded connections.

L D.

Design sferification is based on weld details from "as-built" support m'

drawings. Weld design verification is performed in accordance with i

the requirementa of AISC for structural members and AWS D1.1 for sheet steel.

[

E.

See Item 3.0A above.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS l

1.

N.H. Williams (CYGNA) letter to V. Noonan (USNRC) " Response to NRC Questions," 83090.023, dated March 8,1985.

L 2.

Communications Report between Chang and Huang (Gibbs & Hill) and Horstman, Russ and Williams (CYGNA) dated October 27, 1984.

p 3.

Communications Report between Chang and anang (Gibbs & Hill) and Horstman, Russ and Williams (CYGNA) dated November 13, 1984.

p 4.

Communications Report between Chang and Huang (Gibbs & Hill) and Russ 1

(CYGNA), dated November 17, 1984.

5.

Communications Report between R.M. Kissinger (IUGCO) and J. Russ l

(CYGNA), dated November 30, 1984.

6.

N.H. Williams (CYGNA) letter to J.B George (TUGCO), " Cable Tray

{

Support Review Questions," 84056.041, dated February 13, 1985.

5.0 IMPLEMENTATION OF THE RESOLUTION l

A.

This issue is not applicable. Design verification is based on "as-I built" information as specified in Section III.2 of Reference 2 and on worst case assumptions for inaccessible welds as specified in

)

Attachment X of Reference 3.

For inaccessible welds, a procedure

'g based on statistical projections will be developed and documented in g

Reference 6 Book 20.

B.

Same as Item 5.0A above.

C.

Section IV.1.a of Reference 2 and Sections I.b and I.d of Reference 3 l

specify that eccentric loads are considered in the design verification of all welded connections.

D.

Design verificacion is based on "as-built" inforar cion.Section IX of Reference 3 specifies that both weld and base metal thickness must be appropriately considered in weld qualification.

A16.4 1684R

APPENDIX 16 ISSUE NO. 16: WELD DESIGN AND SPECIFICATI0HS (Cont'd)

I h.'

' 5.0 IMPLEMENTATION 07 THE RESOLUTION (Cont'd)

G E.

Same as Item 5.0A above.

E t

I A

F l'

u6.5 1684R i

m

+

COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCIS AND DUCT SUPPORTS)

APPENDIX 17 ISSUE NO. 17: EMBEDDED PLATE DESIGN I

4 1

- 8 I

e 8

I I

I I

I

~

8 A17.1 8

1684R I

APPENDIX 17 i

ISSUE NO. 17: EMBEDDED PLATE DESIGN

1.0 BACKGROUND

A.

Gibbs & Hill performed capacity calculations for cable tray support attachments to embedded strip plates. CYGNA's review of these I

calculations indicates that the calculated capacities may not have considered the effect of prying action on the tension in the Nelson Studs.

B.

Questions from CYGNA's pipe support reviewers and cable tray reviewers on the stiffening requirements for embedded plate moment connections elicited conflicting responses from TUGC0 personnel. The I

pipe support response indicated that attac1ments to embedded plates act as stiffeners for moment connections (Reference 2), while the cable tray support response indicated that any moment attachment must be stiffened or sufficiently analyzed (Reference 3).

C.

UYGNA has noted that calculations for cable tray supports attached to embedded plates did not consider the capacity reductions for attachment locations given in Gibbs & Hill Specification 2323-SS-30,

" Structural Embedments" (Reference 1).

CYGNA has requested any documents which address the corrective action associated with the issuance of Specification 2323-SS-30 (Reference 9).

~

D.

A review of Brown & Root Procedure CCP-45 (Reference 7) indicates that any two adjacent attachments to an embedded strip plate must be saparated by a minimum of 12 in.

Based on a discussion between CYGNA and TUGC0 (Reference 4), it was determined that even though the installation procedure requires this separation, the inspection procedures for cable tray supports do not require an inspection of

=

this attribute.

CYGNA walkdowns noted several instances where the separation between attachments to embedded plates were less than 12 in.

(Also see Pipe l

Support Review Issue 9.)

CYGNA is concerned that the lack of control l

of attachment spacing may have an impact on the design adequacy of the attachments.

E.

Installation of Details E, F, C, and H on Embedded Plates Reference 5 is the design calculation for the installation of Support Details E, F, G, and H (Gibbs & Hill Drawing 2323-El-0601-01-S) on i

embedded strip plates. A maximum tributary tray span of 7 ft.-6 in.

is used in these calculations. Note 9 on Reference 6 states:

a l

A17.2 I

I 1684R l

lI l

APPENDIX 17 ISSUE NO. 17: EMBEDDED PLATE DESIGN (Cont'd)

1.0 BACKGROUND

(Cont'd)

I

The supports will have e location tolerance of +12 in. in the direction parallel to the tray and +2 in. perpendicular to the tray.

However, spacing between any two adjacent supports shall not exceed 9 ft.-0 in. for l

Unit 1 and Common Areas...unless otherwise I

noted on the drawing."

Supports installed in accordance with this drawing note may have to resist loads due to a 9 ft.-0 in. tributary span, 1 ft.-6 in.

greater than the design tributary span.

'h F.

Gibbs & Hill Specification 2323-SS-30 (Reference 8) provides spacing W

requirements between embedded plates and HILTI arpansion anchors.

During CYGNA's cable tray support walkdowns, an instance tras noted where an embedded plate was located near an opening in a concrete wall. Several HILTI expansion anchors were installed within the opening, on the concrete surface perpendicular to the surface with the embedded plate, potentially violating the requirements of 2323-SS-30.

CYGNA was unable to determine how the minimum spacing requirements would be applied to situations where the expansion anchor is installed in a surface perpendicular to the embedded plate.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS If APPLIES TO CABLE TRAY SYSTEMS

=

A.

Previous design calculations for capacity of support attachments to 4

embedded strip plates may have ignored effect of prying action on tension in Nelson studs.

B.

Pipe support designers and cable tray support designers used inconsistent design practice on stiffening of moment attachments to I

embedded plates. Pipe support design assumed attachments act as stiffeners; cable tray support design indicated moment attachments must be stiffened or analyzed.

C.

Design calculations for cable tray supports attached to embedded plates did not consider capacity reductions given in design specifications for specific locations.

I

^ " '

E 1,d, I

APPENDIX 17 ISSUE NO. 17: EMBEDDED PLATE DESIGN (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) t 2.1 UNDERSIANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

I D.

Installation procedures specified minimum separation for attachments to embedded plates; inspection procedures for supports did not require a check of attachment separation.

Walkdowns found supports attached in violation of specification.

E.

Due to design specifications on some support details for embedded plate attachments, some supports may resist loads from larger tributary tray spans than were assumed in design calculations.

g F.

A case was found where HILTI expansion anchors were installed on 5

concrete surface perpendicular to embedded plate; specifications on

ninimum spacing do not specifically sddress this case.

2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSIEMS A.

The HVAC design calculations used the Gibbs & Hill load capacities 1

found in Specification 2323-SS-30 for evaluating embedded plate attachments.

B.

The HVAC design did not include stiffeners at embedded plate attachment points.

Connections to plates were modeled as translational restraints only.

I C.

The HVAC design considered the load capacity reductions for various attachment locations as specified in 2323-SS-30.

(

~

l g D.

The HVAC installation procedures and inspection procedures each W

require conformance to the separation criteria given in 2323-SS-30.

l E.

The design calculations for the HVAC support attachment to embedded plates were based on "as-built" information. Actual tributary duct span lengths were used in developing loads on those plates.

I F.

The design calculations for HVAC support attachments to embedded plates did not anticipate or address design requirements beyond those l

l specified in 2323-SS-30.

I i

1684R I

APPENDIX 17 ISSUE NO. 17: EMBEDDED PLATE DESIGN (Cont'd) 3.0 ACTION PIAN TO RESOLVE THE ISSUE I

A.

Support attachments to embedded strip plates are treated as six degrees of freedom (moment resisting) restraints in design verification. Attachment loads are evaluated in accordance with Reference 11.

For cases found to be unacceptable per Reference 11, i

l the embedded plate will be evaluated by Stone & Webster.

B.

The embedded plate qualification procedure specifies acceptable methods of analysis for embedded plates.

C.

These capacity reductions are considered. Supports affected by reduced anchorage capacity are identified after embedded plate evaluation.

D.

Reference 4 requires identification of instances where support 8

anchorage is provided by an embedded plate. Violations of the specified minimum separation for attachments to embedded plates are shown on "as-built" support drawings, and their effects are evaluated during embedded plate evaluation.

E.

"As-built" duct spans are used for support design verification and for determination of embedded plate loads.

F.

Configurations in which anchor bolts are installed on a surface perpendicular to an embedded plate are analyzed on a case-by-case basis.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.041, dated February 12, 1985, Attachment A, question 1.

2.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA) dated April 19, 1984, page 11.

!'w 3.

Communications Report between Williams, Russ and Horstman (CYGNA),

Kissinger and Keiss (TUGCO) and Bhujang, Huang, and Chang (Gibbs &

Hill) dated September 15, 1984.

4.

Communications Report between M. Warner (TUGCC) and N. Williams, J.

Minichiello and J. Russ (CYGNA) dated February 27, 1985.

5.

Gibbs & Hill Calculation Binder 2323-SCS-146C Set 4, Sheet 3-9, 21.

g e.

e1bbs e e111 era,1.g 2323-S.e91e, R. vision 3.

I A17.5 I

1684R I

i i

APPENDIX 17 ISSUE NO 17: EMBEDDED PLATE DESIGN (Cont'd) 4 4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE 'IRAY SYSTEMS (Cont'd) 7.

Brown & Root Installation Proceduro CCP-45, " Permanent and Temporary Attachments to Weld Plates," Revision 1, 8/18/80.

8.

Gibbs & Hill Specification 2323-SS-30, Appendix 4, " Design Criteria for Embedded Plate Strips," Revision 1.

9.

N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 21, 1985.

5.0 IMPLEMENTATION OF THE RESOLUTION A.

The embedded plate qualification procedure (Referent:e 11, Appendices 4, 4W, 5, and SW) has been developed and used in design verification, as incorporated in Appendix 2 of Reference 11.

For cases not acceptable per Reference 11, embedded plate evaluation will be performed by Stone & Webster as described in Appendix 2 of Reference 2 and Section VII of Reference 3.

B.

See Item 5.0A above.

C.

See Item 5.0A above.

D.

Use of Section 3.1.2.B.6 and Attachments 5 and 13 of Reference 4 provides the "as-built" information required to evaluate the effects of support attachments to embedded plates.

E.

Procedures to determine appropriate duct span lengths used in design verification are specified in Section III.1 of Reference 2 and I

Section II of Reference 3.

F.

In-plane spacing requirements between HILTI bolts and embedded plates t

l are presented in Appendix 2 of Reference 11 and Appendix 2 of w

Reference 2.

Evaluation of other configurations are analyzed on a case-by-case basis.

I I

I I

A17.6

, I l

1684R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION

~

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS) l APPENDIX 18 ISSUE NO. 18: SYSTEM TO SUPPORT CONNECIIONS r<

L r

L

[

[

[

A18.1

(

1684R r

8 APPENDIX 18 ISSUE NO. 16.

SYSTEM TO SUPPORI CONNECTIONS 1.0 BACEGROUlO E

Two general categories of cable tray clamps are used at CPSES.

" Friction" type clamps are installed on transverse type supports (e.g.,

~I A, B, SP-7, etc.). These clamps are assumed to provide vertical 1

1 and horizontal transverse load transfer.

" Heavy duty" clamps are i

installed on longitudinal trapeze supports (e.g., L-A :

L-B, etc.),

1 1

three-way supports (e.g., SP-7 with brace, Detail 8 on drawing I

2323-S-0903, etc.), and transverse supports, where interferences (e.g.,

tray splice plates, fittings, etc.) prevent the installation of friction type clamps. Heavy duty clamps are designed to transfer vertical, I

horizontal transverse, and longitudinal tray loads to the cable tray support beam. References 1 and 2; DCA 3464, Revision 23; DCA 6299, Revision 7; and DCA 20331, Revision 0 provide clamp configuration details.

In addition to the indicated load transfers between trays and supports, Gibbs & Hill has assumed other load transfer mechanisms in order to justify behavioral assumptions made in the support designs. For I

" friction" type clamps, the following assumptions have been made in order to justify the system concept (also see Review Issue 10).

I The trays will provide out-of plane bracing to trapeze supports to reduce the buckling length of the vertical hanger members (also see Review Issue 4).

The trays will provide lateral bracing to the compression flanges of the horizontal beams (also see Review Issue 24).

I The trays will provide out-of plane bracing to supports to prevent frame translation which would result in increased anchor bolt tensile loads (also see Review Issue 3).

The cable trays will transfer out-of-plane inertial loads from transverse supports to longitudinal supports on the same tray run (also see Review Issue 6).

I the development of minor axis bending moment in the beams due to the l

horizontal eccentricity between the beam neutral axis and the clamp bolt is minimized by a bending moment in the cable tray (also see Review Issue 24).

For vertical loading, the development of torsion in the beam due to I

the eccentricity between the clamp location and the beam shear center is prevented by flexure of the cable tray. This assumes a full moment fixity between the tray and the support beam (also see Review Issue 24).

I lI 1684R lI l

I APPENDIX 18 I

ISSUE NO. 18:

SYSTEM TO SUPPORT CONNECTIONS (Cont'd)

1.0 BACKGROUND

(Cont'd)

For heavy duty clamps, all of the above assumptions are also applicable, and an additional assumption is made by Gibbs & Hill.

The development of torsion due to longitudinal loads on three-way I

supports using composite beam sections (e.g., SP-7 with brace, Detail 8 on Drawing 2323-S-0903, etc.) is prevented by flexure of the cable tray. This assumes a full moment fixity between trsy and support beam (Review Issue 24).

The assumptions described above are valid only if the clamps can provide suitable displacement and rotation compatibility between the tray and support beam. Based on a discussion with TUGC0 (Reference 3), CYGNA lI determined that installation tolerances (Reference 2; DCA 6299, Revision l

7; DCA 20331, Revision 0; and CMC 93450, Revision 4) have been adopted l

which allow gaps between the tray side rails, the support beain, and the l

tray clamps.

In order to provide the assumed compatibility, " friction" type clamps must be cinched sufficiently.to develop friction between the I

tray / beam and tray / clamp interfaces. The existence of gaps will preclude the development of the normal contact force required for frictional I

resistance.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS If APPLIES TO CABLE TRAY SYSTEMS Assumptions made on the load transfer from cable trays to hangers may be I

invalid due to existence of installation gaps and inadequate rotational and displacement capability of both " friction" and " heavy duty" clamps.

These assumptions are:

Cable trays provide the out-of-plane bracing which can reduce buckling length on posts, reduce longitudinal hanger displacement and provide transfer of hanger's out-of plane inertial load (self weight I

excitation) to longitudinal hangers.

Cable trays provide lateral bracing for the compression flange of the horizontal tiers (beams).

Cable trays provide moment resistance capability between trays and horizontal tiers.

I 1

1684R I

I APPENDIX 18 ISSUE NO. 18: SYSTEM TO SUPPORT CONNECTIONS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS Longitudinal HVAC duct supports are welded (or bolted) directly to the HVAC duct skin.

his provides a positive attachment and a well-defined load transfer mechanism. he HVAC design calculations assumed this connection would constrain the rotation of the support in the vicinity of I

the duct attachment. Rotation about the vertical and transverse support axes were assumed to be resisted by the shear stiffness of the duct wall and force couples on adjac.ent supports.

Transverse HVAC duct supports are designed to capture the duct.

he transverse and vertical load is transferred through direct bearing on the support. Rotations of transverse supports were assumed to be constrained I

in the same way they were constrained for longitudinal supports as described above.

Additionally, it was assumed that sufficient friction existed in the I

longitudinal direction causing the self-weight of the transverse support to be carried by adjacent longitudinal supports.

3.0 ACTICN PLAN TO RESOLVE THE ISSUE Longitudinal HVAC duct supports are welded (or bolted) directly to the I

HVAC duct skin.

Bis provides a positive attachment and a well-defined load transfer mechanism.

Each transverse support is evaluated as attached to the duct in all three directions.

Transverse supports are modified per Section 3.1.2.B.10, 3.3.1.A.3, and Attachment 15 of Reference 4 to assure this positive load transfe r.

Transverse supports not so connected will be evaluated on a case-by-case basis.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Drawing 2323-S-0902, Revision 5.

2.

TUGC0 Drawing TNE-SI-0902-02, Revision CP-2.

3.

Communication Report between T. Keiss ('IUGCO) and W. Horstaan (CYGNA) dated November 15, 1984.

I I

I A18.4 1684R I

J APPENDIX 18 ISSUE NO. 18: SYSTEM TO SUPPORT CONNECTIONS (Cont'd)

F 5

5.0 IMPLEMENTATION OF THE RESOLUTION Sections 3.1.2.B.10, 3.3.1. A.3, and Attachment 15 of Reference 4 identify the "as-built" support configurations, including the description and

]

modification of any gaps between the support and duct.

Attachments B1, B2, and B3 of Reference 3 describe the criteria for load transfer between supports and ductwork and Section II of Reference 3 describes the methods to be used for distributing loads among the supports on a duct run. Attachments B1, B2, and B3 of Reference 3 specify the analysis procedures to be used for design verification.

I I

I lI l

I l

I I

1I lI 1

I I

I A1B. 5 16B4R I

iuu mu8 m

i i u

H TEXAS UTILITIES GENERATING COMPANY f

COMANCHE PEAK STEAM ELECTRIC STATION s,-

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES F '.

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

L, '

APPENDIX 19 ISSUE NO. 19: FSAR LOAD COMBINATIONS F

l E

t.

L

[

I E

A19.1

{

(

1684R r

APPENDIX 19 ISSUE NO. 19: FSAR LOAD COMBINATIONS I

Reference 1 defines the loads and load combinations applicable to the design of cable tray supports. CYGNA's review of the cable tray support designs indicates that only dead weight and seismic inertial loads are considered.

I For supports installed in the Reactor Buildings, the loads associated with a LOCA may be appliccble, including pipe whip, jet impingement, and I

thermal loads. Two support types within CYGNA's review were designed for installation in the Reactor Building, Detail A (Gibbs & Hill Drawing 2323-El-0500-04-S) and Detail C (Gibbs & Hill Drawing 2323-El-0500-01-S).

The design calculations for these supports, I

References 2 and 3, respectively, did not consider these additional loads.

2.0 LTDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS I

CYGNA interprets the load combinations of Section 3.8.4.3 of FSAR as being applicable to cable trays in Reactor Building.

Thus, LOCA associated loads, such as pipe whip, jet impingement, and thermal loads may have to be used for the design verification of cable trays.

I 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS I

The HVAC design calculations did not include the effects of LOCA related loads such as pipe whip, jet impingement, and thermal loads for HVAC ducts and supports in the Reactor Building.

3.0 ACTION PLAN TO RESOLVE THE ISSUE The effects of LOCA pipe whip and jet impingement loads are addressed by I

the CPSES damage study. None of the ductwork in the Reactor Building is classified as safety-related. The LOCA loads are listed under the section " Load Combinations for Factored Load Condition" of Section

. 3 3.8.4.3.3 of the FSAR. Furthermore, the FSAR specifies that thermal 1

l 3 loads may be neglected when they are secondary and self-limiting in nature and the material is ductile. This is the case for the CPSES HVAC g

Systems. Thermal effects on anchorages shall be addressed in a generic g

study.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

CPSES FSAR, Section 3.8.4.3.

2.

Gibbs & Hill Calculation Binder SCS-103C, Set 1, Sheets 14-19.

3.

Gibbs & Hill Calculation Binder SCS-103C, Set 2, Sheet 32.

1 A19.2 1684R I

l

APP 2NDIX 19 ISSUE NO. 19: FSAR LOAD COMBINATIONS (Cont'd) 5.0 IMPLEMENTATION OF THE RESOLUTION The CPSES damage study group has summarized the results of their investigation of ductwork in the Reactor Building and is documented in Reference 9.

For the reasons stated in Section 3.0 above, thermal loads I

are not explicitly considered in design verification.

I I

I I

I lI l

lI

~

I I

I lI lI A19. 3 1684R l

l

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 20 ISSUE NO. 20: DIFFERENCES BETWEEN INSTALLATION AND DESIGN 7 CONSTRUCTION DRAWINGS WITHOUT APPROPRIATE DOCUMENTATION I

I I-I I

I I

I I

I I

I l

I I

A20.1 I

1684R I

I APPENDIX 20 ISSUE NO. 20:

DIFFERENCES BErWEEN INSIALLATION AND DESIGN / CONSTRUCTION DRAWINGS WITHOUT APPROPRIATE DCCUhENTATION

1.0 BACKGROUND

CYGNA performed walkdown inspections on 49 of the 92 supports within the review scope. Certain discrepancies between the "as-built" support I

configurations and the design requirements were noted.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS II APPLIP.S TO CABLE TRAY SYSTEMS I

Walkdowns revealed numerous discrepancies between "as-builts" (installations) and designs (design / construction drawings) without documentation. Change notices have been issued for most.

2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSTEMS Limited walkdown inspections of HVAC duct supports have revealed numerous I

discrepancies between the "as-installed" support configurations and the "as-designed" configurations.

3.0 ACTION PLAN TO RESOLVE THE ISSUE All concerns in this issue are addressed by the fact that "as-built" configurations are used in HVAC support design verification. A position I

will be developed to address clearance requirements between HVAC supports and other components.

4.0 LIST OF RELEVANI DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEM 9 1.

Gibbs & Hill, Inc., Support Layout Drawing 2323-El-0713-01S.

2.

Brown & Root, Inc., Fabrication Drawing FSE-00159.

3.

American Institute of Steel Construction, Inc., Manual of Steel Construction, 7th Edition.

4.

Gibbs & Hill Support Layout Dra. ring 2323-El-0601-01-S.

5.

Gibbs & Hill Support Layout Drawing 2323-El-0700-01-S.

6.

Gibbs & Hill Cable Tray Support Design Drawings 2323-S-0900 series.

I 7.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Walkdown Questions," 84056.026, dated August 23, 1984.

I I

^' '

I 1 e4, I

l APPENDIX 20 ISSUE NO. 20: DIFFERENCES BETWEEN INSIALLATION AND DESIGN / CONSTRUCTION DRAWINGS WITHOUI APPROPRIATE DOCUMENT ATION (Cont'd) 4.0 LISr OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont'd)

I 8.

Communication Report between M. Warner, J. van Amerongen (TUGCO) and W. Horstman (CYGNA) dated October 25, 1984.

9.

Communication Report between T. Webb, M. Hamburg (TUGCO) and W.

I Horstman (CYGNA) dated October 10, 1984.

10. Communication Report between M. Warner, C. Biggs (TUGCO) and W.

Horstman (CYGNA) dated October 10, 1984.

11. Brown & Root Procedure No. CEI-20, Revision 9, " Installation of HILTI Drilled-In Bolts."

12. L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA), " Comanche Peak Steam Electric Station CYGNA Review Questions," dated September I

6, 1984.

13. N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Walkdown Questions," 84056.021, dated August 16, 1984.

14. N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Reviev Questions," 84056.089, dated October 21, I

1985.

15. Brown & Root Instruction QI-QAP-11.2-28, " Fabrication, Installation I

Inspections of ASME Component Supports, Classes 1, 2, and 3,"

Revision 29.

I 5.0 IMPLDENI ATION OF THE RESOLUTION All concerns in this issue, except the one noted below, are addressed by the fact that "as-built" configurations are used in hanger design I

verification, as specified in Section III.2 of Reference 2.

Criteria specified in the position on clearance requirements, when finalized, will be followed in HVAC support design verification.

I I

I 1684R I

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 21 l

ISSUE NO. 21: DESIGN CONTROL I

I I

I I

I I

lI I

I I

I I

A21.1 I

1684R

I APPENDIX 21 ISSUE NO. 21: DESIGN CONIROL

1.0 BACKGROUND

A.

During the course of the design and construction of cable tray supports, a large number of design change documents (DCAs and CMCs)

I have been issued that affect the support designs. These design changes can be grouped into two categories. Generic design changes are issued against Gibbs & Hill support design drawings (e.g., 2323-S-0901) and may affect all installations of one or more generic I

support designs. Individual design changes are issued against a support layout plant (e.g., 2323-El-0601-01-S) and affect one or more individual support installations.

CYGNA's review has identified several areas where oversights or errors may occur in the handling of these design changes. These may I

been incorporated in the design drawings.

be due in part to the large number of design changes which have not 1.

In the process of performing generic evaluations of support design adequacy (e.g., the inclusion of base plate flexibility in response to IE Bulletin 79-02, the Working Point Deviation Study, the evaluation of the effects of weld undercut / underrun, etc.),

I Gibbs & Hill based their calculations on the original support designs without considering the effects oi~ all applicable generic design changes (Reference 27).

2.

In some cases, as a result of the generic studies discussed above, the design limits for a support type were made more restrictive than those of the original design. In order to I

qualify existing supports which had been specified based on the original design limits, a case-by-case design adequacy review was performed for all individual supports which exceeded the revised design limits. These reviews were based on the "as-designed" I

configurations for the individual supports, and did not include the effects of applicable individual design changes (Reference 27).

I 3.

The design changes for individual supports are tracked by the cable tray support plan drawing number rather than by the support I

number.

In order to locate all design changes affecting a given support, one must manually search through all design changes affecting all supports on the applicable support plan. CYGNA has observed that some support plans have over 200 design changes I

outstanding.

In order to expedite this effort, the TUGC0 Field Structural Engineeering Group (FSEG) maintains a list of design changes sorted by individual support number. -This list is not a I

controlled document, and CYGNA's review noted several discrepancies between the design changes listed I

A21.2 I

1684R

E APPENDIX 21 ISSUE NO. 21: DESIGN CONIROL (Cont'd) i

1.0 BACKGROUND

(Cont'd)

I for individual supports and those located by CYGNA through a search of design change documents at the Document Control I

Center.

It is CYGNA's understanding, however, that this informal log is relied upon by the field engineer to determine which design changes should be considered in their evaluation.

4.

A discussion with TUGCO cable tray support installation QC personnel (Reference 23) indicated that the method of locating design changes for support inspection purposes was very I

cumbersome and placed an undue burde.n on the inspectors in assembling inspection packagec. TUGC0 QC indicated that the inspectors typically relied on the list of design changes I

included in the Brown & Root construction package as a basis for inspection without independently verifying the completeness of the package.

I 5.

CYGNA has noted instances where the design review for the verification of design changes may have been inadequate. The design changes allowed deviations from the original design that I

invalidated certain assunptions on which the original design was based. However, the design revie1. did not note this and did not assess the impact of the change on the design basis. In other cases, the design review did not assess tha impact of the change on all components of a support that would be affected. Examples of this include:

I Base angles are designed assuming a minimum distance of 3 in.

from the bolt hole to the end of the angle. This distance is used in the calculation of the resisting moment arm when a bending moment is applied to the base angle. CMC 1970 I

reduced this distance to a minimum of 1-1/4 in. The design review for this CMC did not consider the impact of this reduction on the anchor bolt designs.

Cable tray supports are designed for a frame width based on a minimum distance of 3 in. from the outside tray rail to the I

inside of the flange of the hanger (see Review Issue 28.A).

CMC 2646 allows the hanger to be notched so that the tray rail actually overlaps the inside flange of the hanger. T his can result in cable tray supports which do not meet the I

minimum width required by the design. The design review for this CMC only addressed the reduced section properties at the notch without considering the effect on the support width.

I

^

I 1,g, I

I APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd)

1.0 BACKGROUND

(Cont'd)

Cable tray supports are designed to act as a system, with the cable tray acting as a link between supports (see Review Issue 10). CMC 93450 allows gaps between the cable trays and the clamps attaching them to the supports. The frictional force between the clamps and the trays, which is required to prevent relative axial displacement between the trays and the I

supports, is eliminated by the gap. The design review for this CMC does not address the effect on the system behavior of the cable trays (see Review Issue 18).

I.

Cable trays are qualified for an 8 ft.-0 in. maximum span (see Review Issue 25.B).

DCA 1594 provides an installation location tolerance for the supports, resulting in a maximum I

spacing of 9 ft.-0 in, between supports. The design review for this CMC does not consider the effect of the increased span on the cable tray qualification.

B.

Criteria Violations in Individual Support Specifications on Support Plans In the generic design of cable tray i.upports, support dimension and loading limitations are determined for each support type. These limitations are typically stated in the design calculations, but are I

not shown on the generic support design drawings (Reference 4). T he dimensions for each support are specified in a descriptive block on the support plans (Reference 1), and the loading is indicated by the supported tray width shown.

The tray supports listed below were identified as having loadings or support geometries which exceeded the design limitations. Prior to I

the CYGNA review, justifying documentation did not exist for the following individual support designs.

1.

Support Nos. 3025, 3028, 2861, Type D1 Drawing 2323-El-0713-01-S specifies these supports as Type U1 (except beam to be MC6x16.3), L-11 ft.-9 in., h=4 ft.-2 in., and I

shows a tray width of 78 in. The design calculations for Type Di supporta limit L to less than or equal to 8 ft.-0 in.

and tray width to 48 in.

I 1 I A21.4 1684R I

I

_ APPENDIX 21 ISSUE NO. 21: DESIGN CONTRCL (Cont'd)

1.0 BACKGROUND

(Cont'd)

I 2.

Support No. 2607, Type Al I

Drawing 2323-El-0601-01-S specifies dimensions of L=2 ft.-9 in, and h=4 ft.-6 in. for this support. The design calculation for this support type liniits h to less than or equal to 2 ft.-4 in.

3.

Support No. 657, Type A1 Drawing 2323-El-0601-01 specifies this support as Type A, L=7 1

I ft.-0 in., h=2 ft.-0 in. The design calculation for this support type limits L to less than or equal to 6 ft.-0 in.

4.

Support No. 734, Detail "H", Drawing 2323-El-0601-01-S This drawing specifies that one beam is to be an MC6x15.1, rotated 90* from its normal orientation. The support design I

requires the use of C6x8.2 beam sectio The section modulus of MC6x15.1aboutitsweakaxis,1.75in.gs., is smaller than that of C6x8.2 about its strong axis, 4.38 in.3 Therefore, this support should be reevaluated for vertical loads.

Rotating the MC6x15.1 90* from its normal orientation significantly increases the longitudinal stiffness of the I

support. This rotation, together with CMC 00164, which requires l

the use of a " heavy duty clamp," can introduce significant longitudinal loads to the support. The support design requires I

the addition of a longitudinal brace if longitudinal loads are to l

be resisted.

5.

Support No. 3011, Type SP-6 Drawing 2323-El-0713-01-S specifies dimensions of L-8 ft.-9 in.

and h=4 ft.-6 in. The design calculation for this support type limits L to less than or equal to 6 ft.-0 in.

6.

Support Nos. 2992, 2994, 3005, 3017, 3021, 3111, 6654, Type A2 Drawing 2323-El-0713-01-S specifies dimensions of L-8 ft.-3 in.

and h=4 ft.-2 in., and shows a tray width of 78 in. The design calculation for this support type limits L to less than or equal I

to 6 ft.-0 in. and the tray width to 48 in.

I I

1684R I

I APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd)

1.0 BACKGROUND

(Cont'd) 7.

Support Nos. 95 and 112, Type SP-7 I

Drawing 2323-El-0700-01-S specifies these supports as Type SP-7, L=5 ft.-1 in. and shows a tray width of 48 in. The design calculations for Type SP-7 limits the tray width to 30 in.

8.

Support No. 758, Detail "V", Drawing 2323-El-0601-S Drawing 2323-El-0601-01-S specifies this support as Detail "V",

.I h =8 ft.-4 in., h =7 ft.-3 in., h =4 ft.-0 in., 1 =5 1

2 3

1 ft.-9 in., 1 =2 ft.-3 in., a=2 ft.-6 in., and shows a tray 2

width of 66 in. The design for the support detail limits the tray width to 60 in.

9.

Support Nos. 765 and 767, Detail "J", Drawing 2323-El-0601-01-S Drawing-2323-El-0601-01-S specifies these supports as Detail "J",

L-8 ft.-6 in., h =10 ft.-10 in., h =9 ft.-6 in., and h "3 2

3 ft.-6 in., and a ows a tray width of 66 in. The design for the support detail limits the tray width to 48 in.

Additionally, Gibbs & Hill was not consistent in establishing controlling criteria (i.e., support dimensions, tray width, etc.) in I

support designs. As an example, in several support designs, the support frame was designed for a particular height and width while the anchorages were designed using reactions from a frame with a I

different height and width. The lack of a single limiting configuration may affect the support dimensions as shown on the cable tray support plans. Within CYGNA's scope, support types E, SP-6, 4

and SP-8 are affected.

C.

Consideration of "As-Built" Support Conditions in Generic Reviews Which Require a Case-By-Case Review 1.

The SP-7 weld underrun analysis considered 5/16-in. fillet welds which are specified on the design drawings. However, the FSE-00159 fabrication drawings specify smaller weld sizes. In I

addition, the underrun analysis did not consider the effects of any design changes to the supports which were reported in CMCs and DCAs (see Review Issue 21.A).

I 2.

The Working Point Deviation Study did not include the effects of all applicable design changes (see Review Issue 12).

I 1684R I

I APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd)

1.0 BACKGROUND

(Cont'd)

D.

Inconsistencies in the Evaluation of Cable Tray Supports for Thermo-Lag Application 1.

Tray cover weights were not included in the development of the allowable span length tables (References 19 and 20) for fire-Protected cable trays.

2.

CYGNA believes that longitudinal supports are not evaluated for the added weight of fire protection. CYGNA noted evidence of the above in the fire protection reviews for cable tray segment T120SBD07. A longitudinal support (Type L-A ) was assumed to 1

provide transverse restraint in the fire protection calculation.

The calculated transverse load was compared to an assumed design capacity, but no longitudinal load was calculated. The original design for this support type assumes that only longitudinal restraint is provided. Note that the calculations (Reference 21) reviewed by CYGNA had not been design-reviewed at the time they were received from TUGCO.

3.

Gibbs & Hill performed calculations to determine the design capacity for supports to use as a comparison to the tray loads including fire protection (Reference 21). A tributary span of 9 ft.-0 in, was assumed. The actual design span was 8 ft.-6 in.; therefore, the Reference 21 calculations i

i overestimated the support design capacity.

4.

For several tray segments within CYGNA's review scope, the tray I

weight, including fire protection, exceeded the design limit of 35 psf by up to 6 percent, but engineering evaluations were not performed as required by Reference 20. See Reference 27, question 3, for a listing of the affected tray segments.

l S.

For tray segment No. T130SCA46, side rail extensions were I

installed, but a special evaluation was not provided as required by Reference 20 (see Review Issue 25.C.1).

CYGNA has requested additional information on the fire protection evaluation process in Reference 27.

E.

Tray Spans Between Supports Used in the Original Support Layout 1.

Reference 13 indicates that cable trays are to be designed and qualified for 8 ft.-0 in. transverse and vertical spans.

A21.7 1684R 1

_ _ _A

APPENDIX 21 ISSUE NO. 21: DESIGN CONIROL (Cont'd)

1.0 BACKGROUND

(Cont'd)

Reference 10, Note 13, allows a location tolerance for supports of 11/2 of the Richmond Insert spacing parallel to the tray, and limits the maximum spacing between supports to 9 ft.-0 in.

Gibbs

& Hill cable tray support design calculations assume a maximum tributary span of 8 ft.-6 in. to account for a support spacing of 8 ft.-0 in.

on center and an erection tolerance of 16 in.

CYGNA reviewed the tray support plans for segments within the review scope (Reference 12) and noted 15 locations where the "as-designed" tray span exceeded 8 ft.-0 in.

CYGNA's walkdown of these tray segments identified 5 locations where the "as-built" tray spans exceeded 9 ft.-0 in. (see Reference 11). This indicates that the design and installation limitations for support spacings may not have been complied with in the preparation of support layout drawings and in the field.

2.

Reference 13 indicates that cable trays are to be designed and qualified for 40 ft.-0 in. longitudinal spans. Longitudinal support design calculations assume a maximum longitudinal tributary span of 40 ft.-0 in.

For several supports within CYGNA's review, the support plan drawings (Reference 12) showed these supports to have tributary spans greater than 40 ft.-0 in.

(see Reference 11).

In addition, several horizontal tray segments were not provided with any longitudinal supports (see Reference 11). This indicates that the design limitations for the location of longitudinal supports may not have been complied with in the preparation of support layout drawings.

F.

Lack of Calculations for Change Notices CYGNA has noted several design reviews of change notices where the CVC was marked to indicate that new or revised calculations were not required.

CYGNA considers some of the design changes to be significant, such that calculations should have been provided to justify their acceptability.

In some cases, calculations marked "for reference only" are attached to the CMC which the reviewer had accepted without new or revised calculations.

G.

Design Calculation Retrievability and Completeness During the course of the Phase 2 and 4 reviews, CYGNA experienced difficulty in assemblying complete support design calculation sets.

CYGNA noted that Gibbs & Hill has similar difficulty. The following examples illustrate CYGNA's concerns.

A21.8 1684R I

APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd)

1.0 BACKGROUND

(Cont'd) 1.

In Phase 2 of CYGNA's IAP, CYGNA requested an evaluation of the effect of torsion in the C4x7.25 beams on the support design adequacy. Gibbs & Hill provided calculations (Reference 14, I

Sheets 28-33) which evaluate torsion in the beams. These calculations were performed in 1982, but were not included in the indicated calculation binder (the cover sheet for Reference 14 indicated that the total number of sheets was 6).

Subsequent to CYGNA's review of these calculations, they were added to form Revision 1 of Reference 14.

2.

CYGNA requested a list of all calculations relevant to several generic support designs (Reference 15). Gibbs & Hill provided a list of calculation binder and sheet numbers for each support I

type. The review of these calculations by CYGNA indicated that there were edditional calculations relevant to the support designs which had not been included on the list.

For example, I

the Working Point Deviation Study involved several supports listed in Reference 15, but was not referenced in Gibbs & Hill's response.

The difficulties in identifying and locating all calculations pertinent to a support design may be in part attributable to Gibbs &

Hill's methods of controlling structural design calculations. CYGNA I

observed that, as a general rule, Gibbs & Hill did not revise or supersede older calculations. In performing generic studies (e.g.,

Working Point Deviation Study, weld undersize / undercut, evaluation of I

torsional stresses in members, etc.) or performing design reviews for generic design changes, the new calculations evaluate only the effects of the changes. These new calculations may reference the previous calculations as a source of data, but the previous calculations are not superseded by the new calculations, nor are they revised to reflect the results of the design change or generic study. Hence, it is extremely difficult, from reviewing an original design calculation, to determine if it is still applicable to the support design.

It is also difficult to identify and locate generic study or design change review calculations that are applicable to the i

support design.

H.

Lack of Controlled Design Criteria At the initiation of this review, the cable tray support design criteria used by Gibbs & Hill consisted cf a calculation set in a structural calculation binder (Reference 9).

CYGNA's review of this I

document indicated that insufficient detail was given to assure that cable tray support designs were performed in a c:asistent manner and i

i 1684R I

I

1 I

APPENDIX 21 ISSUE NO. 21: DESIGN CONIROL (Cont'd)

1.0 BACKGROUND

(Cont'd) that the designs satisfied the requirements of the CPSES FSAR.

Examples of the impact of an incomplete design criteria include:

1.

CYGNA has noted instances where the field design review group did not utilize the proper criteria to evaluate support adequacy.

The evaluations for fire protection compared the "as-built" I

support load to a design load consisting of the allowable distributed load over a 9 ft.-0 in. tributary tray span. Since the maximum tributary span assumed in the current design I,

calculation is 8 ft.-6 in., the use of a 9 ft.-0 in, opan overestimates the allowable load.

2.

CYGNA has asked what supplements to the 7th Edition of AISC I

Specifications were committed to in the FSAR. No evidence was found to indicate that proper direction was given to design engineers to utilize the requirements of any supplements to which CPSES was committed.

I.

Differences Between Design Drawings and Assembly Drawings CYGNA performed a review of the cable tray support assembly drawings (Reference 25), which are used for construction purposes, and evaluated the accuracy of these drawings via a comparison with the I

applicable design drawings (References 1 and 4). Numerous drawing discrepancies were noted, these included:

I Incorrect weld sizes specified for fillet welds (also See Review' Issue 16.A).

Incorrect weld patterns.

Incorrect member sizes specified in the " Bill of Haterial."

Incorrect anchor bolt connection details.

Incorrect support dimensions.

Members that are not required by the design.

For a detailed listing of the individual discrepancies, see Reference 24.

I l

A21.10 I

l 1684R I

(

APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS

[

A.

This issue raises the following concerns related to proper implementation of design change documentst Both generic and individual design changes may be overlooked in previous calculations, inspections, and reviews.

Generic design adequacy evaluations and case-by-case reviews failed to address all changes.

(

Design change log not a controlled document.

Changes difficult to track for inspection purposes.

Design changes invalidated some assumptions and were not checked in design review process.

(

A related concern identified by the Project is as follows

The design documents (Drawing TNE-S2-0902-02 and ECP 10A Rev 3)

{

covering the installation of shin plates underneath C-clamps until early 1986 inadvertently permitted the shin plate to be inside the web of the tray by as much as 1/2 in.

If the shin plate is installed

(

in the field in this extreme condition, performance of the tray could L

.potentially be impaired.

4

, P,. For some supports, limits on support dimensions and loadings as

(

determined from generic supports were exceeded without providing appropriate documentation.

C.

"As-built" conditions were not considered in generic design reviews

[

which led to case-by-case reviews, e.g., SP-7 weld underrun ' analysis and working point deviation study.

(

D.

In the evaluation of supports with Thorno-Las fire protectiont Tray cover weights were not addressed in determining span

{

allowables.

Longitudinal supports were not evaluated for fire protection

[-

weight.- Incorrect tray span was used in comparison for support capacity.

(

A21.11 1684R

[

f

APPENDIX 21 ISSUE No. 21: DESIGN CONTROL (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.1 UNDERSTANDING OF THE I' SUE AS II APPLIES TO CABLE TRAY SYSTEMS (Cont'd)

I Tray spans with excessive tray weight were not evaluated as required.

Special evaluation for side rail extensions was not performed.

E.

Design and installation limits on tray spans may have been exceeded in drawings and in the field for support spacing in general and longitudinal support spacings in particular.

F.

Calculations for some significant design change notices were lacking or uncontrolled.

G.

CYGNA review encountered miscellaneous problems in assembling complete support design calculation sets. Old calculations were not I

revised or superseded when new studies were performed, resulting in confusion over governing documents.

I H.

Support design criteria used by designers were insufficient to ensure consistency and compliance with FSAR. Improper tributary span criteria were used for field design review. There was no control on use of AISC supplements by designers.

I.

Numerous discrepancies were found between support design and assembly drawings used in construction such as incorrect veld sizes, weld I

patterns, member sizes, anchur bolt connection details, and support dimensions.

In addition, members were shown which were not required by design.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

Design changes for the HVAC system were reviewed. Proper documentation of design reviews was not provided.

B.

Generic supports were not used for HVAC support design. Thus, this issue does not impact the HVAC system design.

C.

See Item 2.2B above.

I D.

The HVAC support design was based on "as-built" configuration.

Actual span lengths and duct accessory weights were included, Thus, this issue does not impact the HVAC system design.

E.

See Item 2.2D above.

I t

A21.12 1684R I

L

~

(

APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSIANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS (Cont'd)

F.

Calculations for some HVAC design change notices were lacking or uncontrolled.

[

G.

The HVAC design calculation sets may not have been properly controlled.

H.

HVAC design criteria used by designers were insufficient to ensure consistency and compliance with FSAR. There was no control on use of AISC supplements by designers.

[-

I.

The HVAC design was based on "as-built" information.

{

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

All accessible portions of the supports are design verified based on "as-built" conditions. For inaccessible portions of the HVAC h

systems, hidden attributes are appropriately considered.

B.

See Item 3.0A above.

[

C.

See Item 3.0A above.

D.

HVAC duct dead weight including sheet steel, joining angles, and reinforcing angles, insulation weight, and accessory weight are considered in support design verification.

E.

"As-built" duct and support information is used in the design verification of ducts and duct supports.

[

F.

HVAC support design verification calculation packages are controlled by the use of appropriate quality assurance procedures. See Item 3.0A above.

h G.

See Item 3.0F above.

H.

Support design verifica?. ion is performed in compliance with the CPSES

{

FSAR (Reference 12) anC the AISC Specification 7th Edition (Reference

17) including Supplements 1, 2, and 3.

Applicable codes and standards are specified in project procedures.

I.

See Item 3.0A above.

[

[-

A21.13 1684R V

APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd) 4.0 LISI 0F RELEVANI DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Drawings 2323-El-0601-01-S, 2323-El-0700-01-S, 2323-El-0713-01-S.

2.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray

{

Support Design Review Questions," 84056.022, dated August 17, 1984, questions 1, 2, and 6.

p 3.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray L

Support Design Review Questions," 84056.025, dated August 21, 1984, question 1.

~

4.

Gibbs & Hill Cable Tray Support Design Drawings 2323-S-0900 Series.

5.

Gibbs & Hill Calculations for Support Numbers 3025, 3028, 2861, CYGNA Technical File 84056.11.1.225.

6.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA), " Responses to CYGNA Review Questions," dated September 4, 1984, with attached l

calculations.

7.

Gibbs & Hill Calculation Binder 2323-SCS-1010, Set 3, Sheets 206, l

Revision 6.

8.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA), " Response to l

CYGNA Design Review Questions," dated September 11, 1984, with I

attached calculations.

9.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 5.

l

10. Gibbs & Hill Drawing 2323-S-0901, Revision 4.

l

11. N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support and Electrical Review Questions," 84056.019, dated August 10, 1984, questions 2.1 and 2.2.

l

12. Gibbs & Hill Drawings 2323-El-0601-01-S, 2323-El-0700-01-S, and 2323-El-0713-01-S.

l

13. Gibbs & Hill Specifications 2323-ES-19, Revision 1 " Cable Tray Specification."

14. Gibbs & Hill Calculation Binder 2323-SCS-1110, Set 8.

15. Communications Report between P. Huang (Gibbs & Hill) and J. Russ (CYGNA) dated June 13, 1984.

A21.14 1684R b

APPENDIX 21 ISSUE NO. 21: DESIGN CONTROL (Cont'd) 4.0 List OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

(Cont'd)

16. L.M. Popplewell (TUGCO) letter to N. Williams (CYGNA), " Comanche Peak Steam Electric Station CYGNA Review Questions," dated August 27, 1984 I

with attachments.

17. R.E. Ballard (Gibbs & Hill) letter to J.B. George (TUGCO), " Cable I

Tray Supports CYGNA Phase 4 Audit Activities," GIN-69377, dated August 24, 1984, with attachments.

18. L.M. Popplewell (TUGCO) letter to N. Williams (CYGNA), " Comanche Peak I

Steam Electric Station CYGNA Review Questions," dated September 11, 1984, with attachments.

19. Gibbs & Hill Calculations Binder 2323-SCS-111C, Set 7.

20. TUGC0 Instruction CP-EI-4.0-49, Revision 1.

21. Cable Tray Thermo-Lag Evaluation, Saf eguards Building, Elevation 790 f t.-6 in., CYGNA Technical File 84056.11.1.1.315.

22. N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.041, dated February 12, 1985.

I

23. Communications Report between M. Warner (TUGCO) and N. Williams et al. (CYGNA), dated February 27, 1985.

I

24. N.H. Williams (CYGNA) letter to V. Noonan (USNRC), " Response to NRC Questions," 83090.023, dated March 8, 1985.

25. Brown & Root Cable Tray Hanger Assembly Dr. awing -FSE-00159.

I

26. Gibbs & Hill Design Procedure DP-1, " Seismic Category I Electrical Cable Tray Supports," Revision 0, dated 6/11/84.

27. N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 21, 1985.

28. N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.027, dated August 27, 1984.

5.O IMPLEMENTATION OF THE RESOLUrION A.

Design verification of supports is performed using "as-built" I

configurations, as specified in Section III.2 of Reference 2.

Hidden i

attributes are appropriately considered per Reference 3, Attachment X.

A21.15 l

1684R lI

I APPENDIX 21 ISSUE No. 21: DESIGN CONIROL (Cont'd) 1 5.0 IMPLEMENI ATION OF THE RESOLUfION (Cont'd)

B.

See Item 5.0A above.

C.

See Item 5.0A above.

D.

Attachment C1 of Reference 3, and Section III.2 of Reference 2 specifies procedures for incorporating duct and duct accessory I

weights in support design verification.

E.

See Item 5.0A above.

I.

F.

Requirements of Appendix K of Reference 18 are followed for the preparation and control of calculation documentation.

G.

See Item 5.0F above.

H.

The CPSES FSAR and the AISC Specification 7th Edition including I

Supplements 1, 2, and 3 are referenced in Sections II.la and II.2 of Reference 2 and are complied with in the preparation of Ebasco project documents, and in support design verification.

I.

See Item 5.0A above.

I I

I I

I I

^ '

I

_R I

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION I

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 22 ISSUE NO. 22 DESIGN OF SUPPORT No. 3136, DETAIL "5", CTH DRAWING 2323-S-0905 I

I I

I I

I I

I I

l l

I I

A22.1 1684R I

E APPENDIX 22 I

ISSUE NO. 22: DESIGN OF SUPPORT NO. 3136, DETAIL "5", CrH DRAWING 2323-S-0905

1.0 BACKGROUND

Support No. 3136, located at elevation 790 ft.-6 in, at the Auxiliary Building / Safeguards Building boundary, is embedded in a fire wall. In I

reviewing the design calculations for this support (Reference 1), CYGNA noted several concerns. A list of CYGNA's questions was provided (Reference 2, Attachment A) to Gibbs & Hill for their review. These concerns included:

Justification for not considering tornado depressurization loads was not provided.

The original cable tray support is Seismic Category I, while the fire wall is Seismic Category II.

Justification for this conflict in I

design classification was not provided.

Several errors were found in the finite element model and in the calculations.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERRT ANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS The following discrepancies were found:

Tornado depressurization loads were not considered.

Support is Seismic Category I while fire wall is Seismic Category II.

Errors were found in finite element model and calculations.

2.2 UNDERSIANDING OF THE ISSUE AS Ir MAY APPLY TO HVAC SYSTEMS Tornado loads were not considered as design loads for HVAC duct and duct I

supports. No duct supports were designed to be attached to Seismic Category II structures.

3.0 ACTION PLAN TO RESOLVE THE ISSUE The design verification walkdown will identify any HVAC supports attached to Seismic Category II structure. Any supports identified are evaluated on a case-by-case basis.

I I

A22.2 1684R

I APPENDIX 22 ISSUE NO. 22: DESIGN OF SUPPORI No. 3136, DERAIL "5", CrH DRAWING 2323-S-0905 (Cont'd) 4.0 LIST OF RELEVANr DOCUMENIS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder SAB-1341, Set 3, Revision 0.

2.

Communication Report between B.K. Bhujang (Gibbs & Hill) and N.

Williams, et al. (CYGNA) dated October 20, 1984.

3.

Gibbs & Hill Calculation Binder SAB-1341, Set 3, Revision 1.

I 4.

N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 21, 1985.

5.0 IMPLEMENTATION OF THE RESOLUTION In design verification Seismic Category II structures are not assumed to I

provide support for HVAC systems. The potential for interaction between Seismic Category II structures and HVAC systems is evaluated on a case-by-case basis.

I I

!I

'I I

A22.3 1684R l

)

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECIRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPIORTS)

APPENDIX 23 ISSUE NO. 23: LOADING IN STRESS MODELS I

I I

I I

I I

I I

I A23.1 1684R lI

APPkNDIX23 ISSUE NO. 23: LOADING IN SIRESS MODELS

1.0 BACKGROUND

For the design of standard support Cases Ag, B, C, and D,

i 1

i where i=1 to 4, finite element analyses were performed (Reference 1) using the program STRESS. Single beam elements were used to model the horizontal members (beams). The analytical results may be inaccurate due to the following concerns:

A.

Tray loads were applied at the beam / hanger intersection, rather than

{-

within the span of the beam where the tray is physically located.

Modeling the load placements in this fashion eliminates the effects of bending and torsion due to vertical loads on the beams, and for Cases D, will totally remove the load applied at the wall i

connection from the support.

(See CYGNA's Phase 2 observation CI-00-03 in Reference 4).

B.

The applied loads are calculated based on an 8 f t.-0 in. tributary tray span. The actual design span is 8 f t.-6 in. if installation tolerances are considered.

C.

The support design drawings _ (Reference 3) specify the support frame heights as the distance from the bottom of the concrete to the top of the C4x7.25 beam. The models considered this distance to be from the concrete to the centerline of the beaz, thus underestimating the support height by 2 in. This error is also found in the related design calculations for the trapeze supports.

2.0 UNDERSIANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS A.

Finite element analyses of standard supports, performed using SIRESS program, applied tray loads at beam / hanger intersections rather than within beam span where tray is actually located, eliminating certain loading effects. Tids situation was corrected for braced frames, but has not been addressed for unbraced frames.

B.

Finite element analyses of standard supports, performed using SIRESS program, calculated applied loads based on a tributary tray span which did not consider installation tolerances. This situation was corrected for braced but not for unbraced frames.

C.

Frame heights input to the analysis models were incorrectly specified, resulting in analysis of frames shorter than actual lengths.

A23.2 1684R

APPENDIX 23 ISSUE NO. 23: IDADING IN STRESS MODELS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS l

A.

Duct hanger verification has been performed using the FEASA-2D, 3D, and HANGER programs in which duct loads applied at the duct / hanger intersection properly represent the load distributions.

B.

Applied loads from HVAC systems were based on "as-built" duct span l

lengths. Therefore, the tributary span lengths used for analysis include installation tolerances. Thus, this issue does not impact the HVAC system design.

C.

The HVAC support qualification modeled the entire height of the support frame. Thus, this issue does not impact the HVAC system design.

3.0 ACTION PLAN TO RESOLVE THE ISSUE

.A.

Duct loads are realistically applied to HVAC supports in a manner consistent with "as-built" connection details.

B.

Duct span lengths used in design verification are taken from "as-built" layout drawings.

C.

Supports are design verified using analysis models in which centroidal and shear center axes are considered per standard engineering practice to define spatial relationship between support members.

[

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Computer Output Binder 2323-DMI-5P.

{

2.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Set 2.

(

3.

Gibbs & Hill Drawing 2323-S-0901, Revision 4.

4.

CYGNA Energy Services, " Independent Assessment Program Final Report - Volume 1, for Texas Utilities Services, Inc., Comanche Peak Steam Electric Station," Report No. TR-83090, Revision 0.

A23.3

f APPENDIX 23 ISSUE NO. 23: LOADING IN STRESS MODELS (Cont'd)

{

l 5.0 IMPLEMENTATION OF THE RESOLUTION k

A.

Attachment B3 of Reference 3 specifies the modeling techniques used to attach the duct to the support.

B.

The use of "as-built" duct lengths is specified in Section III.1 of Reference 2, and Section II of Reference 3.

(

C.

"As-built" support dimensions are used in design verification as specified in Section III of Reference 2.

Modeling techniques used for design verification are specified in Section I of Reference 3.

[

A23.4 1684R

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 24 ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERS

[

l

[

(

y A24.1 1684R

APPENDIX 24 I

ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERS

1.0 BACKGROUND

In the design of cable tray support flexural members (i.e., beams and hangers), Gibbs & Hill did not consider several important items as discussed below.

A.

Additional major axis bending stresses due to transverse loads are introduced by the vertical eccentricity between the cable tray

(

centerlines and the been neutral axis (Reference 1). Gibbs & Hill I

provided calculations (Reference 2) indicating that the increase in bending stress did not exceed 2.5 percent of the allowable stress 1evel. However, the analysis incorrectly assumed that the beam was a

{

fixed-fixed member, effectively isolating it from the remainder of the support structure.

In addition, the load transfer mechanism that was assumed to be provided by the tray clamps may not be applicable to all clamp configurations (also see Review Issue 18).

B.

Minor axis bending of the beams due to transverse loading is introduced by the horizontal eccentricity between the beam neutral axis and the location of the tray clamp bolt holes in the beam's top flange (Reference 1).

Gibbs & Hill response (Reference 2) did not consider the allowed tolerance in bolt hole gage per DCA 17838, Revision 8.

A load transfer mechanism was assumed to be provided by the clamp, allowing the trays and supports to act as a system. This assumption results in increased transverse loads on adjacent supports and no minor axis flexure in the beams. The validity of this assumption depends on the resolution of Review Issues 10 and 18.

C.

Vertical loading introduces torsion into the beam due to the horizontal offset between the tray clamp location and the shear center of the beam. In Gibbs & Hill's response (Reference 2), the torsional moment was completely eliminated, based on an assumed moment resistance provided by the tray clamps and the tray / support system concept (also see Review Issue 10 for the acceptability of this concept).

D.

Torsion is introduced into the beam by longitudinal loading due to:

1.

The vertical offset between the tray centerline and the beam shear center (for longitudinal trapeze type supports, e.g.,

L-A ' L-B }*

l l

2.

The vertical offset between the tray centerline and the shear center of the composite beam (for longitudinal supports similar to SP-7 with brace, Detail 8, Drawing 2323-S-0903, etc.).

A24.2 1684R

_ _ = _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

APPENDIX 24 ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERy (Cont'd)

1.0 BACKGROUND

(Cont'd)

L Gibbs & Hill's evaluation of the torsional effects are included in Reference 2.

The evaluation of torsion due to loading type 1 only considers the eccentricity between the shear center and the top of

(

the tray rungs for ladder type trays or the tray bottom for trough type trays. The centroid of the tray fill is a more appropriate location from which to calculate the eccentricity. For loading type 2, the longitudinal load is applied at the bottom of the tray side rails, rather than the centroid of the tray fill. The tray clamps are assumed to provide rotational restraint to the top flange of the I,E r

composite beam, and all torsional moments are assumed to be resisted L

by a couple formed betwen adjacent vertical supports through flexure of the cable tray. All these assumptions must be justified per Review Issues 10 and 18.

E.

Gibbs & Hill has not consistently considered the reduction in the beam section properties due to bolt holes through the flanges (also see Review Issue 9) and weld undercut effects. Based on CMC 58338, H

Revision 0, the welded connection between the beam and hanger can a'

include vertical fillet welds crossing the web of the beam, thus weld undercut would affect the beam capacity at this critical location.

Weld undercut could also affect the beam capacity at beam-to-base angle / plate connection for the cantilever type of supports.

In addition, based on the tray installation tolerances provided in Gibbs & Hill Specification 2323-ES-100, Section 2.28, and the effcet of CMC 2646, Revision 5, the tray clamp can be located such that the bolt hole is in the same cross-sectional plane as the effect of weld undercut. Thus, it is possible that both reductions may occur simultaneously.

F.

Gibbs & Hill has not evaluated the effects of shear stresses on beam acceptability.

Shear stresses will be introduced by two loadings:

1.

Direct shear stresses due to the applied forces 2.

St. Venant shear stresses associated with torsional loads (see Items 1.0C and 1.0D above)

CYGNA's review indicates that direct shear stresses are minor and generally do not govern the design of flexural members. When these stresses are considered in combination with the potentially large St.

Venant shear stresses, the effect can be a significant factor in the member design (Reference 3).

A24.3 1684R

APPENDIX 24 ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERS (Cont'd)

1.0 BACKGROUND

(Cont'd) l G.

Gibbs & Hill generally assumes an allowable major axis bending stress of 22 kai for member designs. The capacity reduction based on the unsupported length of the beam's comprescion flange (AISC Equation h

1.5-7) is either not considered at all or not properly considered (also see Review Issue 14). Justification is provided, based on the assumption that the tray and tray clamp will provide lateral bracing

{

to the beam's compression flange. This assumption is dependent on 1

the tray clamp's ability to provide bracing (also see Review Issue

18) and neglects compression of the bottom flange due to support frame sidesway and seismic uplift. For the cantilever type of supports, the "1" value in Equation 1.5-7 is improperly selected as discussed in Review Issue 14.

2.0 UNDERSEANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS A.

In design of flexural members under transverse loadings, additional major axis bending stresses result from vertical eccentricity between cable tray centerlines and beam neutral axis. Analyses incorrectly modeled the beam as a fixed-fixed member and may have assumed unrealistic tray clamp behavior.

B.

In the design of flexural members under transverse loadings, minor axis bending may be induced by the horizontal eccentricity between the beam neutral axis and the location of tray clamp bolt holes in the beam top flange. Calculations did not consider hole gage tolerance, and may have assumed unrealistic behavior for the tray clamps.

C.

In flexural members, vertical loads induce torsion due to the horizontal offset between tray c1 amp location and supporting beam shear center. This torsional moment should be appropriately

{

considered.

D.

The methods and assumptions used to account for the effects of torsion induced in hanger beam members by longitudinal tray load vertical eccentricities contain assumptions which require justification.

E.

Design calculations for flexural members inconsistently considered reduction in section properties due to flange bolt holes and weld undercut. Specifically, weld undercut at vertical fillet welds across web, and at beam-to-base angle / plate weld for cantilevers was not addressed. Also bolt holes coplanar with weld undercut was not considered.

A24.4 1684R

APPENDIX 24 ISSUE NO. 24: DEFIGN OF FLEXURAL MEMBERS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd)

.l

)

2.1 UNDERSTANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS (Cont'd) e F.

Design calculations for hanger flexural members did not include shear L

stress effects due to direct shear, St. Venant shear from torsional loads, and the combination of the two.

I G.

In design of flexural members, capacity reduction due to the unsupported length of the compression flange, per AISC Equation 1.5-7, was not properly considered. Calculations used unrealistic tray r

clamp behavior assumptions to justify the above. Compression of L

bottom flange, due to frame sidesway and seismic uplift, was not considered.

2.2 UNDERSTANDING OF THE ISS"E AS If MAY APPLY TO HVAC SYSTEMS A.

The HVAC supports are designed to capture the duct completely on four sides. This load transfer minimizes the effect of the vertical eccentricity between duct centroid and capturing steel.

B.

The HVAC support duct capturing arrangement (as explained in Item

?.2A above) minimizes the effect of horizontal eccentricity between t.he duct attachment point and the neutral axis of the horizontal capturing member.

Calculations did not consider hole gage tolerance.

C.

The HVAC transverse supports are designed to carry vertical as well as transverse loads through bearing against the capturing steel.

Therefore, there is no offset between load application and shear center and no torsional moment developed. For longitudinal support design, load is transferred through welds and may create an offset which results in a torsional moment. This moment was not considered in HVAC design calculations.

D.

The design of flexural members for longitudinal HVAC supports did not consider the torsional moment. See Item 2.2C above.

E.

The HVAC design calculations did not address reductions in member section properties due to bolt holes or weld undercut.

F.

The HVAC design calculations considered the combined effects of direct and torsional shear stresses (i.e., St. Venant shear stresses). The direct shear stresses were combined by the SRSS method and added absolutely to the torsional shear stress.

i A24.5 1684R

APPENDIX 24 ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERS (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS If MAY APPLY TO HVAC SYSTEMS (Cont'd) i l

G.

The HVAC design calculations did not apply AISC Equation 1.5-7 in the l

design of flerural members. Unsupported member lengths were not l

checked for lateral torsional buckling.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

In support design verification, support member bending due to vertical eccentricity effects of transverse duct loads is considered.

B.

In support design verification, support member bending due to horizontal eccentricity effects of transverse tray loads is considered. The effect of bolt hole gage tolerance is insignificant.

C.

In support design verification, the torsion induced onto support members from horizontally offset duct loads is considered.

D.

In support design verification, the torsion induced onto support members from the vertical offset of longitudinal duct loads is considered.

E.

The reduction in beam section properties due to bolt holes and weld undercuts is considered in design verification.

F.

Shear stresses due to direct shear, St. Venant torsional shear, and the combination of the two are included in support design verification.

G.

AISC Equation 1.5-7 is applicable to channels. Since HVAC supports use structural angles, Section 1.5.1.4.6b of the AISC Specifications is applicable. Justification of this applicability is addressed in a I

study.

4.0 LIsr 0F RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS I

1.

N.H. Williams (CYGNA) letter to J.B. George (TUGCO), " Cable Tray Support Review Questions," 84056.031, dated August 31, 1984.

2.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA) " Comanche Peak Steam Electric Station Cygna Review Questions," dated September 28, 1984.

I A24.6 1684R

APPENDIX 24 ISSUE NO. 24: DESIGN OF FLEXURAL MEMBERS (Cont'd).

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont 'd) 3.

Communications Report between E. Berkor et al. (Gibbs & Hill) and M. Engleman et al. (CYGNA) dated April 11, 1985.

4.

Gibbs & Hill Drawing 2323-S-0903.

5.0 IMPLEMENTATION OF THE RESOIIITION l

A.

h e effects of transverse duct load eccentricities are considered in support design verification as specified in Attachments B1, B2, and

(

B3 of Reference 3.

1hese attachments also describe modeling techniques and load application procedures used in support design verification.

B.

The effects of transverse duct load eccentricities are considered in support design verification as specified in Attachments B1, B2, and B3 or Reference 3.

Tnese attachments present a detailed description of modeling techniques and load application procedures used in support design verification.

C.

1he effects of vertical duct load eccentricities are considered in support design verification as specified in Attachments B1, B2, and B3 of Reference 3.

1hese attachments present a detailed description of modeling techniques and load application procedures used in support design verificatien.

D.

Torsion due to vertical offset of longitudinal loads is considered in support design verification as specified in Attachments B1, B2, and B3 of Reference 3.

1hese attachments present a detailed description of modeling techniques and load application procedures used in support design verification.

E.

Attachment E2 of Reference 3 specifies requirements regarding

(

reduction in support member section properties due to the presence of both known and unidentified bolt holes. Welds not meeting the requirements of Reference 13 (VWAC) on weld undercut are identified per References 14 and 15 and subsequently dispositioned.

F.

Shear stresses are considered in support design verification as specified in Reference 3 Section VI.

G.

Allowable compressive bending stress for structural angles is taken as 0.60Fy in accordance with Section 1.5.1.4.6b of the AISC Specification. Lateral bracing requirements of Section 1.5.1.4.6b are specified in Attachment V2 of Reference 3.

A study to justify this approach is documented in Reference 6 Book 6.

A24.7 1684R

B

{

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 25 ISSUE NO. 25: SYSTEM STRUCTURAL QUALIFICATION IL F

Fl l

l ll 1l ll l

h A2 5.1 1684R

I APPENDIX 25 ISSUE NO. 25: SYSTEM SrRUCIURAL QUALIFICATION

1.0 BACKGROUND

The qualification requirements for cable trays are outlined in References I

1 and 4.

In reviewing related specifications, calculations, and installations of cable trays, CYGNA has noted several areas of concern.

j A.

Qualification of cable trays is performed through static load testing W

ar.d calculation of loading interactions for dead load plus three components of seismic load (Reference 1, Section 3.9 and Reference 3).

Seismic loads are calculated by the equivalent static load 1

method, using total tray dead weight times the peak spectral acceleration. No apparent dynamic amplification factor (DAF) is used. Reference 6, Section 5.3, and Reference 7 recommend the use of a DAF=1.5 unless justification is provided (see also Review Issue 8).

B.

The interaction equation specified for checking cable tray capacity (Reference 1, Section 3.9.4) is limited in its application and may I

have been used incorrectly.

The testing and qualification of cable trays is based on an 8 ft.-0 I

in. simply supported tray span (References 1 and 3); yet, Reference 8, Note 13, allows a support installation tolerance resulting in a maximum tray span of 9 ft.-0 in, for Unit 1.

The capacity values derived in the tray testing are total loads (in 1bs) uniformly distributed over an 8 ft.-0 in. section of cable tray i

(Reference 3). These values, F, F and F1 as used with the I

interactionequation,areonlyappike,abletotraysectionswith8 n

l ft.-0 in. spans. However, for the fire protection evaluation calculations (Reference 2) and tray span violation calculations (Reference 9), total loads for various tray spans were calculated as f'n w*L, where w is the tray unit weight and L is the tray span.

This load was compared with the rated tray capacity using the interaction equation.

For evaluation of trays with spans other than 8 ft.-0 in., a capacity comparison must be made in terms of tray bending moment which is I

proportional to (w*L2), rather than the total load on the tray section.

For example, if an 8 ft.-0 in. tray span will support a total distributed load of 1600 lbs (200 lb/ft), by increasing the I

span to 10 ft.-0 in., a uniform load of 128 lb/ft (1280 lbs) would result in the same bending moment at mid span. Therefore, the capacity for the 10 ft.-0 in, span would be 1280 lbs and not the 1600 lbs assumed.

I A25.2 1684R

. - ~... -

1 APPENDIX 25 ISSUE No. 25: SYSTEM STRUCTURAL QUALIFICATION (Cont'd)

1.0 BACKGROUND

(Cont'd)

.J t ( r j.T.l C.

CYGNA has noted several instances of modifications to cable tray SJ " '. ? O hardware without adequate justification or documentation.

$f s[

W.,

1.

Tray Segment No. T130SCA46 is assumed to be a 23-in. x 6-in.

$MOj ladder-type tray in the fire protection evaluation calculations y-f-6 1 for Safeguards Building Elevation 790 ft.-6 in. (Reference 10).

Jp1q CYGNA's walkdown indicates that this tray is actually a 24-in. x p,P.f 4-in. ladder-type tray with 6-in. side rail extensions added to p, g.g increase the tray depth. The tray qualification test report f f:),;K (Reference 3) does not provide qualification for trays using side W.,. A-6 g ;[*

([:'.b rail extensions. The procedure governing fire protection evaluation (Reference 11), Subsection 3.2.2.2 states:

'.:t

.5

.... =

61.

"Evaluatico process described in 3.2.2 is pl.3.# n not applicable to the cable trays (and

$y j [~.

their supports). For such cases, actual M,:,/h E

"as-built" configuration of the tray F-} '(

system with actual cable weight shall be

$i ;...

taken into account and proper engineering evaluation performed. No standard 3, i,.f,

methodology is recommended, but shall be (w.l<?

based on acceptable engineering practice."

W 9.2

' f).f:.

The referenced calculations do not perform an evaluation of this

'(:). iM tray segment. These calculations (Reference 10) were obtained from TUGC0 prior to their design review; therefore, this possible 01 L_.

omission may be corrected through the design review process.

/ h...q w

2.

Tray Segment T120SBC35 is joined to a tray reducer with side rail hh

-:.' v/ ' :.

splice connector plates. These plates have been modified by s

removing portions of their bottom flanges such that only the web x '.. : - [ '

area remains. This connector will not satisfy the requirements

.f. 7.9 of Reference 1, Section 3.7, Paragraph f, which states that

,.[-9:(

'(M:;d!

connectors "shall have mem nt and shear strengths at least equal

.;g to those of the continuous uncut side rail." CYGNA was unable to p

locate documentation justifying this modification of g83:;y vendor-supplied hardware, ff.g;c

,9 77

.l.Q':..y

+.p@.Dm f'.

.c.

.cx *,.%

lw. n '1 ' 1.'

t f.;.a-J n

f -t

$y, -

W&'

~

_=_

A25.3

.% A p:

1684R hh,

.. hW5

ic,.

r

l APPENDIX 25 ISSUE NO. 25:

SYSTEM STRUCTURAL QUALIFICATION (Cont'd)

L I

1.0 BACKGROUND

(Cont'd)

J D.

Cable tray section properties are calculated using the static test results (Reference 3). The moment of inertia is calculated based on the flexural displacement formula for a simply supported beam. For i

horizontal transverse loading (i.e., in the plane of the rungs),

ladder-type cable trays show a truss-like behavior, and the deflection will be due to both flexure and shear deformations.

F H

This will affect the calculated moment of inertia as used in any Gibbs & Hill analyses which consider the tray properties for r/

frequency or displacement calculations.

L 200 UNDERSTANDING OF THE ISSUE

(

2.1 UNDERSTANDING OF THE ISSUE AS II APPLIES TO CABLE TRAY SYSTEMS No dynamic amplification factor (DAF) was used in qualification of g

A.

cable trays, as recommended in FSAR and IEEE standard.

1 B.

The interaction equation was improperly used for spans greater than 8 ft.

C.

Several instances were found of modifications to vendor supplied hardware for cable trays, without adequate justification or documentation. This particularly involves side rail extensions to trays and tray splice connector plate modifications.

/

Splices / connections used for given tray types do not always conform to the acceptable splice patterns defined by the cable tray manufacturer or by approved field design change authorizations.

CYGNA had identified at least one instance in which tray splices

[

different than approved by the. manufacturer or justified by other design documents have been used. During the "as-built" program implementation, TUGC0 has noted that this is not an isolated

{

occurrence, but is reasonably frequent. The ability of predicting tray performance is predicated on knowing that tray splices behave as intended by the manufacturer.

Inadequate justification exists for many of the modifications of the splices encountered in the field.

Lacking the knowledge on the behavior of the modified splices prevents confirmation of adequate behavior of the tray. A reportable deficiency report has been issued.

D.

Cable tray moment of inertia calculations do not consider shear deformations under horizontal transverse loading of ladder-type trays. This affects tray properties used in support frequency or displacement calculations.

A25.4 1684R

B APPENDIX 25 ISSUE NO. 25: SYSTEM STRUCIURAL QUALIFICATION (Cont'd) 2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSf EMS A.

HVAC duct design used a 1.0 DAF for static analysis in frequency ranges above 20 Hz.

Dynamic analysis was used in calculations for

^

frequency ranges under 20 Hz.

B.

The HVAC duct load capacity was evaluated using "as-built" duct span lengths with equivalent duct beam properties generated from dynamic I

and static test data (Reference 7). Thus, this issue does not impact the HVAC system design.

I x

C.

No instances of modifications which may affect duct qualification have been found. Thus, this issue does not impact the HVAC system design.

D.

The H7AC duct beam properties are based on dynamic modal testing.

jL 1-i Shear deformations in the duct beams are calculated based on a reduced shear modulus of elasticity derived from test data.

3.0 ACTION PLAN TO RESOLVE THE ISSUE h

A.

For design verification performed using equivalent static analysis, a 3

factor of 1.5 is applied to the spectral acceleration for frequencies l

below the cut-off frequency. A factor of 1.0 is applied to accelerations for frequencies at or above the cut-off frequency. See I

also Section 3.0 of Appendix 8.

B.

The HVAC duct design verification is based on "as-built" duct span information. Ultimate moment capacities are properly considered.

l C.

The HVAC duct design verification is based on "as-built" data. Any I

modifications to the duct construction will be evaluated on a case-by-case basis.

D.

Duct span modeling considers shear flexibility of the duct based on a I

reduced shear modulus of elasticity.

4.0 LIsr OF RELEVANI DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Specification 2323-ES-19, Revision 1.

2.

Gibbs & Hill Structural Calculation 2323-SCS-111C, Set 7, Revision 1.

3.

T.J. Cope, Test Report and Calculations for the Qualification of Cable Trays.

I A25.5 1684R

(

l B

l APPENDIX 25

)

ISSUE NO. 25: SYSTEM STRUCTURAL QUALIFICATION (Cont'd) l B

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS (Cont 'd) 4.

CPSES FSAR, Section 3.10B.3, Amendment 44.

l S.

Gibbs & Hill Specification 2323-ES-100, Revision 2.

l l

6.

IEEE " Recommended Practices for Seismic Qualification of Class 1E Equipment for Nuclear Power Generating Stations," STD 344-1975.

7.

CPSES FSAR Section 3.7B.3.5.

j 8.

Gibbs & Hill Drawin8 2323-S-0901, Revision 4.

9.

L.M. Popplewell (TUGCO) letter to N.H. Williams (CYGNA), " Response to CYGNA Review Question 2.1 of Letter 84056.019," dated August 27, 1984 l

with attached calculations.

10. Cable Tray Thermo-Lag Evaluation Safeguardo Building, Elevation 790

}

ft.-6 in., CYGNA Technical File 84056.11.1.1.315.

11. TUGC0 Instruction CP-EI-4.0-49, Revision 1.

12. N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 21, 1985.

I 5.0 IMPLEMENTATION OF THE RESOLUTION l

A.

Procedures are specified in Sections IV.1.d, IV.2.d and V of Reference 2 regarding use of the 1.5 factor in equivalent static analysis.

B.

Design verification of the duct is based on "as-built" information as specified in Section III of Reference 2.

l C.

"As-built" duct information is used as specified in Attachment Cl of Reference 3 and Section III.1 of Reference 2.

D.

Duct member property calculations are based on Reference 7 and will be documented in Reference 6 Book 15.

B I

A25.6 1684R 1

TEXAS UTILITIES GENERATING COMPANY

' J COMANCHE PEAK STEAM ELECTRIC. STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES i

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 26 ISSUE NO. 26: BASE ANGLE DESIGN

(

[

[

[

[

t A26.1 1684R

~.

I APPENDIX 26 ISSUE NO. 26: BASE ANGLE DESIGN

1.0 BACKGROUND

A.

In References 1 and 2, the base angles were modeled as simply I

supported beams. This modeling technique does not include the stiffening effects of concrete bearing at the angle ends.

B.

The principal axes were not considered in the analyses of the base I

angles subjected to the various loadings.

C.

The base angle lengths due to the maximum spacing of the Richmond Inserts were not considered in the working point analyses.

For support types D, D, L-A, L-A, SP-4, SP-6, SP-8, and D.

i 2

i 4

Detail 11 (Drawing 2323-S-6905) the design calculations do not I

include an evaluation of the base angles.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSIANDING OF THE ISSUE AS If APPLIES TO CABLE TRAY SYSTEMS I

A.

In base angle design, angles are modeled as simply supported beams, ignoring stiffening effects of concrete bearing at angle ends.

B.

In design of bace angles for various loadings, the principal axes I

were not considered.

C.

The " Working Point Deviation Study" for brace connection I

eccentricities did not address the most critical spacing of Richmond Inserts in determining base angle lengths.

D.

For some cable tray support types, design calculations did not I

include an evaluation of base angles.

2.2 UNDERSTANDING OF THE ISSUE AS II MAY APPLY TO HVAC SYSIEMS I

A.

For HVAC supports, base angles are modeled as simply supported beams without consideration of the concrete bearing surface in the model.

B.

In the design of HVAC support base angles, the principal axes were not modeled in the two dimensional models. The principal axes were modeled for three dimensional models.

C.

No Working Point Deviation Study has been performed for HVAC supports. The HVAC design was performed using "as-built" support conditions. Thus, this issue does not impact the HVAC system design.

I g

.2e.2 1684R I

APPENDIX 26 ISSUE NO. 26: BASE ANGLE DESIGN

,2.0 UNDERSTANDING OF THE ISSUE (Cont'd) 2.2 UNDERSTANDING OF THE ISSUE AS If MAY APPLY TO HVAC SYSTEMS (Cont'd)

D.

Base angle evaluation was included.for each HVAC support. Thus, this

(

issue does not impact the HVAC system design.

3.0 ACIION PLAN TO RESOLVE THE ISSUE

[

l A.

The stiffening effects reduce the moments (stresses) along the base angles. Therefore, ignoring the stiffening effects of concrete bearing at the ends of base angles and modeling them as simply

[-

supported beams is conservative.

B.

The design verification of HVAC supports considers the principal axes

(

in the analyses of base angles.

C.

The design verification of base angles for HVAC supports is based upon "as-built" dimensions.

D.

All base angles for HVAC supports are design verified.

4.0 LIsr OF RELEVANE DOCUMENIS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Sets 2 through 6.

2.

Gibbs & Hill Calculation Binder 2323-SCS-101C, Set 1.

5.0 IMPLEMENTATION OF THE RESOLUEION A.

No action is required.

B.

The principal axes are considered in the design verification of base angles for HVAC supports as specified in Attachment E2 of Reference 3.

{

C.

The design verification of HVAC supports considers "as-built" dimensions as specified in Section III.2 of Reference 2 including member lengths and Richmond Insert spacing.

D.

All base angles for HVAC supports are design verified as specified in Section III.2 of Reference 2.

A26.3

(

1684R J

lI f

TEXAS UTILITIES GENERATING COMPANY l

COMANCHE PEAK STEAM ELECTRIC STATION l

l EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 27 l

ISSUE NO. 27: SUPPORT QUALIFICATION BY SIMILARITY I

1 lI I

I I

g I

I I

I

'I I

I I

^"

I g

1ee4R

I APPENDIX 27 ISSUE NO. 27: SUPPORI QUALIFICATION BY SIMILARIIY

1.0 BACKGROUND

A.

In Gibbs & Hill design calculations, several support types were qualified by similarity to another support type without showing similarity.

CYGNA's review of the geometry, loading, connection details, etc. indicated that the designs were not obviously similar, and that calculations should have been provided. Supports in this I

category are:

1.

Detail A, 2323-El-0700-01-S Reference 2 states that Detail A is similar to Case SP-7.

CYGNA noted that the cantilever length for Detail A is greater than for I

SP-7 and that the anchor bolt attachment is unlike the attachment for SP-7.

2.

Detail N, Drawing 2323-El-0601-01-S Reference 1 states that Detail N is similar to Details V and R on the same drawing. CYGNA noted that the frame geometry and tray locations for Detail N was unlike either of the cited details.

3.

Detail J, Drawing 2323-El-0601-01-S Reference 1 states that Detail J is similar to Case B. CYGNA 3

noted that the member sizes used are different than those for Case B, and the frame dimensions exceed the design limits for 3

iI Case B3-4.

Detail V, Drawing 2323-El-0601-01-S Reference 1 states that Detail V is similar to Detail B, drawing 2323-El-0713-01-S. CYGNA noted that Detail B is a three bay frame with braces in all bays and was designed as a pinned I

truss. Detail V does not have braces in all bays, and if the same design technique is applied, the frame would be statically unstable.

!I B.

Allowed working point deviations for individually designed supports were established based on similiarity to standard support types j

without justification. See Review Issue 12.H for a discussion of W

this topic.

l I

A27.2 1

1684R I

h i

p APPENDIX 27 ISSUE NO. 27: SUPPORT QUALIFICATION BY SIMILARITY (Cont'd) 2.0 UNDERSTANDING'0F THE ISSUE 2.1. UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS A.

In design calculatione, some supports were qualified by similarity to supports not obviously similar, and no justification was provided.

In the Working Point Deviation Study for brace connection B.

eccentricities, allowable working point deviations were established for some supports based on similarity without justification.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS h

A.

Qualification of HVAC supports by their similarity to other supports lacks proper justification.

B.

The Working Point Deviation Study does not apply to HVAC supports.

1his study was conducted specifically for the cable tray hanger program.

3.0 ACTION PLAN TO RESOLVE THE ISSUE Each individual HVAC support is design verified. If grouping is used,

{.

the grouping shall follow a generic grouping procedure.

4.0 LIST OF RELEVANT D6CUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-104C, Set 1.

2.

Gibbs & Hill Calculation Binder 2323-SCS-104C, Set 5.

3.

R.E. Ballard (Gibbs & Hill) letter to J.B. George (TUGCO),

(

GTN-69361, dated August 21, 1984, with attachments.

4.

R.E. Ballard (Gibbs & Hill) letter to J.B. George (TUGCO),

GTN-3-69377, dated August 29, 1984, with attachments.

5.0 IMPLEMENTATION OF THE RESOLUTION Each HVAC support is design verified as specified in Section III.2 of

[.,

Reference 2.

~

m

(

A27.3

-1684R

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 28 ISSUE NO. 28:

CRITICAL SUPPORT CONFIGURATIONS AND LOADINGS E

E d

E E

E

+

A28.1 1686R

I I

APPENDIX 28 l

ISSUE NO. 28: CRITICAL SUPPORT s FIGURATIONS AND LOADINGS

1.0 BACKGROUND

)u A.

Gibbs & Hill design calculations (References 1, 2, and 3) for trapeze type supports considered only a limited number of support aspect F

ratios. Justification was not provided to show that the chosen aspect ratios would provide the critical configuration to evaluate I

all components of the support design. The determination of aspect l

ratios was baced on an assumed frame width based on supported tray width and the maximum frame height. The frame width determination l

assumed that:

(a) trays were installed with a minimum 6-in.

l horizontal spacing, (b) the distance betwent trays, and 0 in, between l

tray side rails and the hanger flange. Reference 4 indicates that cable tray installations at CPSES allow a maximum tray width of 36 in.

l l

B.

In the design of the frame members for trapeze supports, Gibbs & Hill typically applied the loadings to the frame in a symmetric pattern.

I In reviewing the support layout plans, CYGNA has noted that the cable

}

trays are of ten located in an asymmetric fashion on the supports.

This could result in higher stresses in the support members and I

higher loads on the anchorages than considered in the design.

I 2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS I

A.

Aspect ratios used in calculations for trapeze supports may not I

represent critical configurations adequately to evaluate all support components. Frame width assumptions were unrealistic.

,I B.

Design calculations for frame members in trapeze supports applied loads symmetrically; actual tray loads are of ten asymmetric.

l 2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS l

A.

HVAC design calculations did not rely on critical support

'a configurations or loadings. Each HVAC support design was based on g

"as-built" configuration. Thus, this issue does not impact the HVAC system design.

I B.

HVAC design calculations did not rely on critical support configurations or loadings. Each HVAC support design was based on l

"as-built" configuration and actual duct load locations were used.

Tsus, this issue does not impact the HVAC system design.

A28.2 1686R

[

APPENDIX 28

. ISSUE NO. 28: CRITICAL SUPPORT j

CONFIGURATIONS AND LOADINGS (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE "As-built" configurations and duct locations are used in the design verification.

[

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTEMS 1.

Gibbs & Hill Calculation Binder 2323-SCS-1010, Set 1.

2.

Gibbs & Hill Calculation Binder 2323-DMI-SP.

(

3.

Gibbs & Hill Calculation Binder 2323-SCS-215C, Sets 2-5.

4.

Gibbs & Hill Specification 2323-ES-19, " Cable Trays," Revision 1.

5.

N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO) " Cable Tray / Conduit Support Review Questions," 84056.089, dated October 28, 1985.

5.0 IMPLEMENTATION OF THE RESOLUTION Section III.2 of Reference 2 specifies that "as-built" configurations are used in the design verification of HVAC supports.

[

[

[

[

[

A28.3 1686R

,f

H I

TEXAS UTILITIES GENERATING COMPANY COMANCHE PLAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 29 l"

ISSUE NO. 29: CUMULATIVE EFFECT OF REVIEW ISSUES u

Eu

[

E E

E E

[

[

A29.1 1

1686R C

APPENDIX 29 ISSUE NO. 29: CUMULATIVE EFFECT OF REVIEW ISSUES

1.0 BACKGROUND

In this Review Issues List, a number of the cited issues may lead to small unconservatisms when occurring singly in a support design. Such unconservatisms may usually be neglected. However, since several of I

these issues pertain to all cable tray support designs on a generic basis, their effect can be cumulative, such that many small unconservatisms may be significant. Therefore, any reavaluation of I

support designs should consider the cumulative effect of all pertinent Review Issues.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS Small unconservatisms resulting from separate issues may have significant cumulative effect for HVAC supports impacted by more than one issue.

2.2 UNDERSTANDING OF THE ISSUE AT IT MAY APPLY TO HVAC SYSTEMS Small unconservatisms resulting from separate issues may have significant cumulative effect for HVAC supports impacted by more than one issue.

3.0 ACTION PLAN TO RESOLVE THE ISSUE This issue is inherently addressed by the comprehensive engineering approach to the design verific. tion of the HVAC ducts and duct supports and by the implementation of extensive "as-built" analysis, qualification, and test activities.

As discussed in the introduction, all the generic technical HVAC issues fall into four categories: deviations between the "as-designed" and "as-I built" HVAC systems, control of design documents, analysis assumptions and methods, and design assumptions and methods.

I The "as-built" vs. "as-designed" issues are addressed cumulatively via the comprehensive "as-built" program. 100 percent of all accessible HVAC system components are "as-built."

Inaccessible components critical to I

the design verification effort are rendered accessible or are classified as " hidden attributes." Hidden attributes will be conservatively qualified via statistical studies and evaluated for worst effect in design verification. In this manner, actual "as-built" conditions will I

be properly taken into account.

In addition, this program resolves instances of improper installation and poor construction quality.

I The issues related to control of design documents are cumulatively addressed by virtue of the design verification program which will generate "as-built" design documentation, support drawings, and qualification calculations on 100 percent of the HVAC ducts and duct I

supports.

A29.2 1686R

L APPENDIX 29 ISSUE NO. 29: CUMULATIVE EFFECT OF REVIEW ISSUES r

L 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd) 3 All analytical issues (analysis assumptions and methods) and design r

L issues (design criteria and assumptions) are simultaneously addressed by the development of procedures and instructions and supported by studies and tests which systematically consider each issue. By virtue of the j

I overall approach which is implemented, the cumulative effect of these issues is addressed directly.

p In summary, the overall design verification approach fully addresses and L

resolves each of the generic technical issues both individually and collectively, provides 100 percent "as-built" documentation of the HVAC system designs including resolution of improper installation or 7

construction. This ensures that the margin of safety in the HVAC system L

is acceptable.

{

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CABLE TRAY SYSTDIS None 5.0 IMPLEMENTATION OF THE RESOLUTION No further action required.

y L

r L

7 L

E E

[

[

{

A29.3

{

1686R

('-

TEXAS UTILITIES GENERATING COMPANY C0lnNCHE PEAK STEAM ELECTRIC STATION

{:

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 30 r

L ISSUE No. 30: SYSTEM DAMPING VALUES g

[

[

[

[.

[

[

[

[

mm W

[

[

A30.1 s

[

1686R

{

APPENDIX 30 ISSUE No. 30: SYSTEM DAMPING VALUES 1.0' BACKGROUND Dsaping values of 4 percent and 7 percent have been used for the

(.-

evaluation OBE and SSE seismic inertial conditions respectively in the design serification of CPSES cable tray systems. The validity of-these values has been questioned by CASE (Hearing Transcript 13196, 13303-

{

13307, 13318, 13454-13461).

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS See Section 1.0 above.

E.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS

]

The evaluation of OBE and SSE seismic loads in the design of CPSES HVAC l

{

supports used damping values of 2 percent and 3 percent for the OBE and SSE seismic events, respectively.

b 3.0 ACTION PLAN TO RESOLVE THE ISSUE The HVAC design verification uses 2 percent and 4 percent damping for the

(-

OBE and SSE seismic load cases, respectively, as specified for welded steel structures in NRC Regulatory Guide 1.61.

If 4 percent damping SSE response spectra are not readily available, 3 percent SSE spectra any conservatively be used.

{

4.0 LIST OF RELEVANT DOCUMENTS FOR CABLE TRAY SYSTEMS

(

None 5.0 IMPLEMENTATION OF THE RESOLUTION Damping values of 2 percent for OBE and 4 percent for SSE are used for design verification of HVAC ducts and duct supports as specified in p

Section IV.1.c of Reference 2.

If 4 percent damping SSE response spectra L

are not readily available, 3 percent SSE spectra may conservatively be used.

[

[

[

[

A30.2 1686R

{

r

I-TEXAS UTILITIES GENERATING COMPANY j

COMANCHE PEAK STEAM ELECTRIC STATION h

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES L

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 31 ISSUE No. 31: MODELING OF BOUNDARY CONDITIONS rL b

rL E

E

[

[

[

[

[

[

[

A31.1 1686R hiae ii n i i i i

L

{

APPENDIX 31 ISSUE NO. 31: MODELING OF BOUNDARY CONDITIONS

1.0 BACKGROUND

CASE has questioned in testimony the behavior of bolted hanger anchorages and the modeling technique used to represent their boundary conditions.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CABLE TRAY SYSTEMS See Section 1.0 above.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS f

The HVAC design calculctions did not address the effects of oversized I

bolt holes on suppott aciffness, bolt capacity, or overall dynamic-response of the support.

l 3.0 ACTION PLAN TO RESOLVE THE ISSUr.,

Base angle assemblies are modeled at each bolt location. Assemblies consist of a short section of base angle with a given anchor bolt size I

and edge distance. Spring rates and allowable loads for each assembly are analytically determined. Base angle spanning between anchor bolts is J

included as part of the support model.

4.0 LIST OF RELEVANT DOCUMENTS l

I None I

5.0 IMPLEMENTATION OF THE RESOLUTION

!I Sections III of Reference 3 specify the boundary condition modeling criteria for design verification.

Anchorage assembly stiffness and l

allowable load values will be documented in Reference 6 Book 3.

Prying action effects are an integral part of the anchorage essemblies.

l l

A31.2 1686R

L

{

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES p

L FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 32 ISSUE NO. 32: CONCRETE VOIDS

[-

[

[

[

['

[

[

[

[

[

[

[

[

[

A32.1

[

1686R

[

I APPENDIX 32 ISSUE NO. 32: CONCRETE VOIDS 1

1.0 BACKGROUND

The IRC inspected a sample of HVAC duct supports in February,1986.

During this inspection it was noted that several supports had voids in I

excess of 1/16 in. in the concrete beneath the base angle of the support.

This gap between the base angle and the concrete was not I

included on the "as-built" drawings of the supports.

2.0 UNDERSTANDING OF THE ISSUE Ibe HVAC support construction procedures did not specify a tolerance for I

flush mounting of the base angle to the concrete surface.

The HVAC design calculations did not account for the effects of an uneven bearing surface on bolt tension loads.

3.0 ACTION PLAN TO RESOLVE THE ISSUE I

All accessible span steel with a void in excess of 1/16 in. will be grouted.

For inaccessible base angles, the existence of uneven bearing surfaces between the base angles and concrete will be conservatively addressed analytically.

If modifications are required due to I

conservatisms taken, design verification may be opplemented by test data.

4.0 LIST OF RELEVANT DOCUMENTS I

i None

5. 0 IMPLEMENTATION OF THE RESOLUTION Grouting of existing gaps between base angles and concrete shall be l

g performed in accordance with Section 3.1.2.B.9 and 3.3.1.A.10 of B

Reference 4.

The analysis procedures for evaluating base angles and anchorages are specified in Attachment G and Section III of Reference 3.

I I

I I

I A32.2 1686R I

e

TEXAS UTILITIES GENERATING COMPANY

~

COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 33 ISSUE NO. 33: GAPS BETWEEN DUCT AND DUCT SUPPORT E

E E

[

l b

E

[

[

[

E

[

[

A33.1 1686R

.sPPENDIX 33 ISSUE NO. 33: GAPS BETWEEN DUCT AND DUCT SUPPORT

1.0 BACKGROUND

The transverse duct supports rely on bearing between the duct and capturing support member to restrain vertical and transverse loads.

The existence of a gap between the duct and support could render a transverse

{

support inactive if the duct travel is insufficient to close the gap.

Duct support fabrication and inspection procedures did not specify a tolerance for fit-up between the duct and duct support.

This condition r

(

was not represented on duct support "as-built" drawings.

1 2.0 UNDERSTANDING OF THE ISSUE

[:

i The HVAC design calculations did not address the effects of the possible l

loss of load restraining capability of transverse supports. No design

{

fit-up tolerance was specified.

3.0 ACTION PLAN TO RESOLVE THE ISSUE

(

Transverse supports are modified per Reference 4 such that the total transverse gap between duct and support does not exceed 1/8 inch.

In addition, transverse supports are modified per Reference 4 to assure zero

{

vertical gap between duct and support.

This assures positive transverse and vertical load transfer between duct and support.

4.0 LIST OF RELEVANT DOCUMENTS None

(

5. 0 IMPLEMENTATION OF RESOLUTION Sections 3.1.2.B.10, 3.3.1. A.3, and Attachment 15 of Reference 4

{

identifies the "as-built" support configurations, including the description and modification of saps between the transverse support and duct. Attachments B1, B2, and B3 of Reference 3 describe the load application requirements for design verification of transverse supports.

[

[

{

{

A33.2 h

1686R

(

TEXAS UTILITIES GENERATING COMPANY

{-

COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES

[-

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

L APPENDIX 34 ISSUE NO. 34: ATTACHMENT OF TRANSVERSE r

SUPPORTS TO DUCTS L

[

[

[

[

[

[

[

[

[

[

[

A34.1

[

1686R

L-APPENDIX 34

{

ISSUE NO. 34: ATTACHMENT OF TRANSVERSE SUPPORTS TO DUCTS

1.0 BACKGROUND

(

Transverse duct supports are designed to carry only transverse and vertical duct loading through bearing of the duct against the capturing support member. Some transverse supports have been constructed such that l

{

the capturing steel is welded to the duct forming a positive attachment between the duct and the support. The concern is that this connection may add significant longitudinal load to the support from the duct.

(

2.0 UNDERSTANDING OF THE ISSUE The transverse HVAC supports have been designed for transverse and

(

vertical loading only. The positive attachment of a transverse support to the duct may generate longitudinal loading which the support has not been designed to carry. The HVAC design calculations did not address the effects of longitudinal loading on transverse supports.

{

3.0 ACTION PLAN TO RESOLVE THE ISSUE

(

Design verification is performed based on "as-built" duct to support connection details per Reference 4.

{

4.0 LIST OF RELEVANT DOCUMENTS None b

5.0 IMPLEMENTATION OF THE RESOLUTION Procedures for load transfer are given in Attachments B1, B2, and B3 of

(

Reference 3.

[

[

[

[

[

[

A34.2 1686R E

g TEXAS UTILITIES GENERATING COMPANY

[.

COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES

(

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 35 ISSUE NO. 35: INTEGRITY OF DUCT JOINTS

[

[

l 1

[

[

[

[

[

[

[

[

[

{

(

A35.1

[

1686R

[

APPENDIX 35 ISSUE NO. 35: INTEGRITY OF DUCT JOINTS

1.0 BACKGROUND

A series of duct tests were performed to establish duct structural properties and load capacities for HVAC ducts (Reference 7).

Compression axial loads were selected for testing to determine the duct load capacity I

since these loads are critical for measurements of duct buckling and bending moments. There is concern that this approach may have inadequately determined duct load capacities since the transverse duct I

joints were not tested in tension for leakage and structural integrity.

2.0 UNDERSTANDING OF THE ISSUE Duct sections are bolted together at flange connections creating a joint that is sealed with a neoprene gasket. The load capacity and failure point for this connection in tension was not established by test. The I

design calculations for HVAC duct were based on the compressive load capacity of the duct. If the tension load capacity is less than the compressive load-capacity, the structural integrity and leak tightness of the duct sections may be compromised.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A test will be performed to demonstrate the effects of a tension load (longitudinally through the duct) on the structural integrity and leak tightness of the duct and transverse duct joints. The test will determine the duct load capacity in the transverse joints.

4.0 LIST OF RELEVANT DOCUMENTS 1.

" Seismic Qualification Report of Seismic Category I Ductwork and ihngers," CCL Report No. A-424-81-10, January 18, 1985.

2.

" Evaluation of Non-Conforming Welds," CCL Report No. A-579-83, July 25,1983.

3.

" Duct Test Evaluation," CCL Report No. A-414-81, February 19, 1982.

4.

" Cable Tray Supports - Review Issues List," CYGNA - Rev. 12, November 20, 1985.

5.

" Addition of Thermo-Iag to Ten HVAC Supports...," CCL Letter to Bahnson, June 15, 1984.

I I

A35.2 1686R i

APIENDIX 35 ISSUE No. 35: INTEGRITY OF DUCT JOINTS 5.0 IMPLEMENIATION OF THE RESOLUTION Development of a test procedure is in progreso to address this issue.

This test will use ductwork identical to the ductwork installed at CPS ES.

Duct load ccpacities for longitudinal applied tension loads will be determined. Particular emphasis will be placed on the duct joints for both leakage and structural integrity. Ultimate load levels for leak tightness and structural integrity will be established. The results of I

this test will be incorporated into design verification.

A35.3 1686R

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION I.

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 36 ISSUE NO. 36: EFFECTS OF OPENINGS IN DUCTS I

I I

I I

I I

I I

I I

A36.1 1686R l

I APPENDIX 36 ISSUE NO. 36: EFFECTS OF OPENINGS IN DUCTS

1.0 BACKGROUND

In the duct testing program used to establish duct structural properties and load capacities (Reference 7), a concern was raised that the effects of openings in ducts were not addressed.

2.0 UNDERSTANDING OF THE ISSUE I

The test program conducted to determine duct structural properties was limited to HVAC duct test samples with no openings in the duct walls.

Actual duct construction requires that sections of the duct wall be I

removed for installation of branch connections, grills, registers, and access doors. The HVAC design program did not address the effects of these openings on duct section properties or duct load capacity.

3.0 ACTION PLAN TO RESOLVE THE ISSUE A test will be performed to determine the effects of duct openings on the duct section properties and load capacities.

4.0 LIST OF RELEVANT DOCUMENTS 1.

" Seismic Qualification Report of Seismic Category I Ductwork and Hangers," CCL Report No. A-424-81-10, January 18, 1985.

2.

" Evaluation of Non-Conforming Welds," CCL Report No. A-579-83, July 25, 1983.

3.

" Duct Test Evaluation," CCL Report No. A-414-81, February 19, 1982.

4.

" Cable Tray Supports - Review Issues List," CYGNA - Rev.12, November 20, 1985.

5.

" Addition of Thermo-Lag to Ten HVAC Supports...," CCL Letter to Bahnson, June 15, 1984.

I I

5.0 IMPLEMENTATION OF THE RESOLUTION I

Development of a test procedure is in progress to address this issue.

This test will use ductwork identical to the ductwork installed at CPSES. The results of this test will be incorporated into design verification.

I I

I A36.2 1686R

I TEXAS UT1L1 TIES eEsEEATI e Co eisT COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPO,TS)

APPENDIX 37 1SSUE so. m me,-m EA ES,e,SE e,,I E.Ax,E,SsE,,E g

I I

I I

I I

I I

I I

I I

I A3 7.1 1

L APPENDIX 37 ISSUE NO. 37: NON-LINEAR RESPONSE OF FIRE DAMPER SLEEVE

1.0 BACKGROUND

1 Allowable loads for transfer air sleeves (in which fire dampers are

~

g L

installed) were based on 1.5 x peak acceleration value, because the sleeves rest in the wall opening with a 1-in. top gap and 1/2-in. side gaps. For cases where the vertical response of duct at the sleeves is

[

expected to be more than 13, non-linear seismic response of the sleeve cannot be properly addressed by this approach.

e 2.0 UNDERSTANDING OF THE ISSUE L

See Section 1.0 above.

l

[

3.0 ACTION PLAN TO RESOLVE THE ISSUE No action is required since these transfer air sleeves and fire dampers

{'

are not Seismic Category I.

In addition, although gaps exist, the design details preclude separation of the equipment from the opening in which it is installed because the sleeves are trapped by supporting members.

4.0 LIST OF RELEVANT DOCUMENTS 1.

" Seismic Qualification Report of Seismic Category I Ductwork and Hangers," CCL Report No. A-424-81-10, January 18, 1985.

2.

" Evaluation of Non-conforming Welds," CCL Report No. A-579-83, July 25,1983.

3.

" Duet Test Evaluation," CCL Report No. A-414-81, February 19, 1982.

(

4.

" Cable Tray Supports - Review Issues List," CYGNA - Rev.12, November 20, 1985.

{

5.

" Addition of Thermo-Iag to Ten HVAC Supports...," CCL Letter to Bahnson, June 15, 1984.

5.0 IMPLEMENTATION OF THE RESOLUTION None required.

C

[

[

[

A37.2 1686R C

u TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION J

EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES L

FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

AFPENDIX 38 ISSUE No. 38: BUCKLING OF CANTILEVER LEG OF BASE ANGLE F

L L

r u

[

[

l'

[

[

{

A38.1 1686R

l

{

APPENDIX 38 ISSUE NO. 38 BUCKLING OF CANTILEVER LEG OF BASE ANGLE F

L

1.0 BACKGROUND

Concern has been raised that the duct support design methodology for g

L evaluating the cantilever portion of the support's base angle had not been checked for worse case loading to assure that no buckling of the vertically projected angle leg occurs.

2.0 UNDERSTANDING THE ISSUE p

Local stresses and localized buckling of the base angle vertical leg were L

not evaluated as part of the HVAC design calculations.

3.0 ACTION PLAN TO RESOLVE THE ISSUE i

The buckling behavior of the base angle vertical leg is addressed in design verification.

4.0 LIST OF RELEVANT DOCUMENTS 1.

" Seismic Qualification Report of Seismic Category I Ductwork and

[

Hangers," CCL Report No. A-424-81-10, January 18, 1985.

2.

" Evaluation of Non-conforming Welds," CCL Report No. A-579-83,

{

July 25, 1983.

3.

" Duct Test Evaluation," CCL Report No. A-414-81, February 19, 1982.

4.

" Cable Tray Supports - Review Issues List," CYGNA - Rev.12, November 20, 1985.

(

5.

" Addition of Thermo-Iag to Ten HVAC Supports...," CCL Letter to Bahnson, June 15, 1984.

{

5.0 IMPLEMENTATION OF THE RESOLUTION Design verification of base angles includes consideration of buckling of angle legs as specified in Attachment E2 of Reference 3.

c A38.2 1686R f

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 39 ISSUE No. 39: ANCHOR BOLT PERPENDICULARITY REQUIREMENTS

I

I f

'I I

' I I

lI I

I

I lI lI I

I

^

I g

1eeeR

I APPENDIX 39 ISSUE NO. 39: ANCHOR BOLT PERPENDICUIARITY REQUIREMENTS

1. 0 BACKGROUND Any existing angle between the actual "as-built" center line of drilled-in expansion anchors and a line perpendicular to the concrete surface was I

not included on the "as-built" design drawings for HVAC supports. 'lhis condition could significantly alter the overall support load capacity by reducing anchor bolt load capacity and must be addressed.

I 2.0 UNDERSTANDING OF THE ISSUE

'Ihe HVAC design calculations did not address anchor bolt angularity or I

establish an acceptable tolerance of anchor bolt angularity for design purposes. Ihe HVAC support design must consider any reduction in the rated capacity of anchor bolts due to installation angle.

3.0 ACTION PLAN TO RESOLVE THE ISSUE I

The "as-builting" and design verification programs include identification and disposition of out-of-tolerance HILTI bolt installations.

4.0 LIST OF RELEVANT DOCUMENTS None

5. 0 _ IMPLEMENTATION OF THE RESOLUTION Per Section 3.1.2.B.6 of Reference 4, anchor bolts with angularity greater than six degrees are identified and modified.

I I

I I

I I

A39.2 1686R I

L

{'

TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC AND TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 40 ISSUE NO. 40: MEASUREMENT OF EMBEDMENT FROM TOP OF CONCRETE TOPPING I-u I

L

[

E E

[

[

[

[

[

A40.1 1686R f

APPENDIX 40 ISSUE NO. 40: MEASUREMENT OF EMBEDMENT FROM TOP OF CONCRETE TOPPING

1.0 BACKGROUND

Note Sa on Gibbs & Hill Drawing 2323-S-0910, Sheet G-4a allows reduced expansion anchor embedment for certain supports at lower building elevations. SDAR CP-80-05 states that the integrity of the architectural I

topping cannot be assured, thus evaluation of all affected designs must be made to satisfy the corrective action requirements of the SDAR. CYGNA has not reviewed any design calculations resolving the above-mentioned I

note with the implications of the SDAR. The generic design calculations do not address the note.

Such a reduction in anchor embedment is not acceptable for 1/4-in. and I

3/8-in. HILTI Kwik-bolts with 2-in. embedment requirement, since these j

bolts are embedded in topping only. Additionally, HILTI does not manufacture a 1/4-in. Kwik-bolt of sufficient length to accommodate the I

2-in. topping and the 2-in. minimum embedment in structural concrete.

l The anchor embedment reduction may not be acceptable for other sizes of l

HILTI Kwik-bolts, depending on the actual accelerations applicable to the floor-mounted supports versus the design accelerations. The affected support types within CYGNA's scope are the CSM-18 series and CST-17.

2.0 UNDERSTANDING OF THE ISSUE 2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CONDUIT SYSTEMS Note 5a on Sheet G-4a of Drawing 2323-S-0910 allows the concrete topping thickness to be considered in determining embedment length of anchors at I

building elevations 832 ft.-6 in. and below. The integrity of the concrete topping cannot be assured and generic design calculations do not address the above note.

Although this is applicable to Unit No.1, this issue does not apply to Unit No. 2 since Note 3 on Sheet G-3a of Drawing 2323-S2-0910 specifies that the minimum embedment length of HILTI Fulk-bolt is the embedment I

length in structural concrete. 'Ihe 2-in. topping on the concrete floor is not considered as a part of structural concrete. Paragraph 2.4.5 of Brown & Root Procedure CEI-20 Revision 9 ensures that the embedment length of HILTI Kwik-bolt is consistent with the design.

2.2 UNDERSTANDING OF THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS The HVAC design did not specifically address reduction of embedment length of concrete anchors due to concrete topping for floor-mounted supports.

I A40.2 i

l

(

1686R I I

APPENDIX 40 ISSUE NO. 40: MEASUREMENT OF EMBEDMENT FROM TCP OF CONCRETE TOPPING (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE HVAC supports mounted on floors with 2-in. topping are identified and design verified with a 2-in. reduction of embedment length. The 1/4-in.

and 3/8-in. diameter concrete anchors used in floors with 2-in. topping, if any, will be replaced.

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CONDUIT SYSTEMS 1.

SDAR CP-80-05, "Use of Architectural Concrete in Floor Slabs," dated August 8,1980.

2.

N.H. Williams (CYGNA) letter to W.G. Counsil (TUGCO), " Cable Tray / Conduit Support Review Questions," 84056.094, dated October 30, 1985.

3.

G.L. Madsen (NRC) letter to R.J. Gray (TUGCO), dated January 25, 1985.

{

5.0 IMPLEMENTATION OF THE RESOLUTION Attachment G7 of Reference 3 identifies floors with concrete topping and I

requires a reduction of 2 in. from the embedment length of concrete enchors to be included in the design verification of supports mounted on

(

these floors.

ll 1

ll l

ll A40.3 1686R

I TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC AND TECHNICAL ISSUES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 41 ISSUE NO. 41: BOLT HOLE TOLERANCE AND EDGE DISTANCE VIOLATION I

1I 1

I lI l

I l

I 1

I 1

I l

l 1I A41.1 1686R

APPENDIX 41 ISSUE NO. 41: BOLT HOLE TOLERANCE AND EDGE DISTANCE VIOLATION j

[

1.0 BACKGROUND

A.

Reference 1 allows bolt hole tolerances which vary with the bolt

(

size, whereas the AISC Specifications provide zero bolt hole tolerances. Therefore, the bolt holes in Gibbs & Hill designs should be considered oversized and should be treated as such in bearing connection calculations.

{

B.

Reference 2 requires that a minimum clear distance be maintained for oversize holes. Some Gibbs & Hill designa do not provide the minimum

(

edge distances required in the AISC Specifications. For example, support types CA-Sa and CSM-42 provide edge distances of 3/4 in.

Per Reference 2, 25/32-in. is required.

2.0 UNDERSTANDING OF THE ISSUE l

2.1 UNDERSTANDING OF THE ISSUE AS IT APPLIES TO CONDUIT SYSTEMS A.

AISC Specification allows bolt hole in steel-to-steel connections to i

be 1/16-in. larger than bolt diameter. Since bolt hole tolerances

(

specified in Note 15 on Sheet G-lb of Drawing 2323-S-0910 and in Note 7 on Sheet G-la of Drawing 2323-S2-0910 are more than 1/16 in.,

bolt holes are considered to be oversized.

B.

Edge distance in structural angle with oversized hole in support types CA-Sa and CSM-42 should be increased to comply with requirements of the AISC Specification.

2.2 UNDERSTANDING THE ISSUE AS IT MAY APPLY TO HVAC SYSTEMS A.

The HVAC design did not specify bolt hole sizes.

Design calculations did not consider possible effects of oversized bolt holes. See Appendix 14, Item 2.2F for a discussion of this issue.

B.

The HVAC design did not check bolt hole edge distances for compliance with AISC Specifications.

(

3.0 ACTION PLAN TO RESOLVE THE ISSUE A.

See Appendix 14, Item 3.0F.

D.

Bolt hole edge distances are shown on "as-built" support drawings and checked for AISC violations in the design verification on a case-by-case basis. A statistical study shall be developed for

[.

inaccessible bolt holes as the need arises.

[

A41.2

[.

1686R

APPENDIX 41 ISSUE NO. 41: BOLT HOLE TOLERANCE AND EDGE DISTANCE VIDIATION (Cont'd)

(

4.0 LIST OF RELEVANT DOCUMENTS REVIEWED BY CYGNA FOR CONDUIT SYST1MS 1.

Gibbs & Hill Drawing 2323-S-0910, Sheet G-lb, Note 15.

2.

AISC Specification, 7th Edition, Section 1.16.5, Minimum Edge Distance.

3.

AISC Specification, 7th Edition, Section 1.23.4, Riveted and Bolted Construction - Holes.

(

5.0 IMPLEMENTATION OF THE RESOLUTION A.

See Appendix 14, Item 5.0F.

B.

Appendix 2 of Reference 2 specifies that the "as-built" bolt hole l

edge distances are checked for compliance with the AISC F

Specifications. AISC bolt hole violations, if any, will be evaluated L

as required.

[

[

[

[

E E

A41.3

{

1686R

[

[

ll TEXAS UTILITIES GENERATING COMPANY COMANCHE PEAK STEAM ELECTRIC STATION EVALUATION AND RESOLUTION OF GENERIC TECHNICAL ISSITES FOR HVAC SYSTEMS (INCLUDING DUCTS AND DUCT SUPPORTS)

APPENDIX 42 l l TERA IRR NO. DAP-E-M-504 1

5 DETERMINATION OF HEAT LOADS FOR EQUIPMENT SIZING I

I I

I I

I I

I I

I I

I A42.1 I

1686R

I APPENDIX 42 TERA IRR NO. DAP-E-M-504 DETERMINATION OF HEAT LOADS FOR EQUIIMENT SIZING

1.0 BACKGROUND

The areas of concern regarding the calculated HVAC heat loads, and the I

deviations and discrepancies uncovered in the process of the DAP review by the Third Party are documented in Discrepancy / Issue Resolution Reports (DIRs) listed in Reference 1.

The main concerns identified are:

1.

Calculation approaches were taken which were not consistent with established engineering methods in the applicable fields of concern.

2.

Incorrect assumptions were utilized.

3.

Improper and/or incorrect inputs were utilized.

4.

An incomplete inventory of heat load sources in areas under consideration was utilized.

5.

An inaccurate implementation of calculation results into equipment specifications was utilized.

I 6.

Additional minor deviations and discrepancies were identified during DAP as documented in the DIRs listed in Reference 1.

I 7.

TERA believes that the DAP reviews provide adequate confidence that all significant technical issues have been identified and that similar issues would be found in the remaining design documentation.

2.0 UNDERSTANDING OF THE ISSUE In general, this issue may impact the environmental qualification of safety related components located in the various plant areas.

3.0 ACTION PLAN TO RESOLVE THE ISSUE 1.

The calculation approaches utilized by Ebasco are consistent with established engineering methods.

2.

Assumptions in the calculations are clearly identified and follow I

appropriate engineering practice.

3.

Verifiable and appropriate design inputs are utilized.

, I A42.2 1686R i

I APPENDIX 42 TERA IRR NO. DAP-E-H-504 DETERMINATION OF HEAT LOADS FOR EQUIPMENT SIZING (Cont'd) 3.0 ACTION PLAN TO RESOLVE THE ISSUE (Cont'd) 4.

Complete inventory of heat load sources (see Paragraph II, Background) are utilized in all areas based on as-built configuration.

5.

Consistency reviews are performed to ensure the calculation results are implemented in the equipment specifications.

6.

Additional deviations and discrepancies identified during DAP will be resolved by closure of all DIR's.

I 7.

This Action Plan is generic to resolve similar issues in the design documentation of all HVAC Systems.

4.0 LIST OF RELEVANT DOCUMENTS l g 1.

Texas Utilities Generating Company, Comanche Peak Steam Generating l

3 Station, TERA IRR No. DAP-E-M-504, " Determination of Heat loads for l

HVAC Equipment Sizing," October 3, 1986.

5.0 IMPLEMENTATION OF THE RESOLUTION All HVAC Systems will be design verified in accordance with the Action i

Plan described herein and the program description in Reference 20.

1 I

I A42.3 1686R t

.