ML062490259

Enclosure 13, E. I. Hatch Request to Implement an Alternative Source Term - EPRI Technical Report 1007896, Seismic Evaluation Guidelines for HVAC Duct and Damper Systems.

ML062490259
Person / Time
Site:	Hatch
Issue date:	04/30/2003
From:	Electric Power Research Institute
To:	Office of Nuclear Reactor Regulation
References
1007896
Download: ML062490259 (160)

Text

Enclosure 13 Edwin I. Hatch Nuclear Plant Request to Implement an Alternative Source Term EPRI Technical Report 1007896, Seismic Evaluation Guidelines for HVAC Duct and Damper Systems

I--Frc2I Seismic Evaluation Guidelines for HVAC Duct and Damper Systems Technical Report

Seismic Evaluation Guidelines for*HVAC Duct and Damper Systems 1007896 Final Report, April 2003 EPRIProject Manager i R. Kassawara' EPRI -3412 Hillview Avenue, Palo Ato.California 94304

• PO Box 10412, Palo Alto. California 94303

• USA 800.313.3774

• 650.855.2121 - askeprd@eprl.com

• www.epri.com

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY. PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR 'A PARTICULAR PURPOSE, OR (11) THAT SUCH USE DOES NOT INFRINGE ON OR.

INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (111).THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER.

(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF:.THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT ABSG Consulting ORDERING INFORMATION . .

Requests for copies of this report.should be directed to ERIP Orders and Conferences, 1355 Willow Way, Suite 278, Concord,. CA94520, (800) 313-3774, press 2 or internally x5379, (925) 609-9169, 1(925) 609-1310 (fax). -. . .." .

l.,Eectric Power Research Institute and EPRI are 'registered service .marks of the Electric Power ResearchInstitute, Inc. '-.EPRI. ELECTRIFY THE WORLD is a service mark of the Electric.Power Research Institute, Inc.

Copyright © 2003 Electric Power ResearchInstitute, Inc.' All rights ieserved.

CITATIONS This report was prepared by ABSG Consulting 118 Portsmouth Avenue, Suite B202 Stratham, NH 03885 Principal Investigators P. Baughman F. Beighi J. Dizon J. White This report describes research sponsored by EPRI.

The report is a corporate document that should be cited in the literature in the following manner:

Seismic Evaluation Guidelinesfor HVAC Duct and DamperSystems, EPRI, Palo Alto, CA:

2003.1007896.

• -* *__:_. . . _

• i"ii

REPORT

SUMMARY

This report provides guidelines that can be used to perform an experience based seismic capability verification of HVAC duct and damper systems in nuclear power plants. The report

.,summarizes seismic experience data from strong-motion earthquakes for these systems and identifies the characteristics of systems that could lead to failure or unacceptable behavior in an earthquake. The seismic experience data show that HVAC duct and damper systems exhibit extremely good performance under strong-motion seismic loading, with the pressure boundary being retained in all but a handful of cases.

Background:

The Seismic Qualification Utility Group (SQUG) provides guidelines for seismic capability Verification of nuclear plant electrical and mechanical equipment; relays, tanks and heat exchangers; and electrical raceway systems using seismic experience and test data. As part of this effort, the performance of HVAC duct and damper systems in 100 power and industrial

-facilities in more than 20 sttrong-motion earthquakes has been compiled into a seismic experience database. This database has been used to establish guidelines to seismically verify as-installed HVAC duct and damper systems and screen out potential failure modes and undesirable conditions that could lead to seismic damage or failure.

Objective To provide guidelines that can be applied to as-installed HVAC duct and damper systems to demonstrate seismic ruggedness.

Approach The research team assembled data on the seismic performance of HVAC duct and damper data to systems from over 20 strong-motion earthquakes since 1971. The team'studied these experience determine failure modes capacities, and success parameters. They used the recorded data to develop guidelines for evaluation of ductwork and dampers.

Results The guidelines in this report can be6used to demonstrate the eismic caiability.of HVAC duct and damper systems.The recomrhended seismic adequacy review procedure includes doc'umentation re'view, in-planrt'screeninig Walkdowns, analytical review' of selected duct runs and supports; and identification and resolution of outliers that do not meet the screening or analysis*s, criteria. Documentation is reviewed to determine input parameters such a's system identification; function, -system boundaries, operating conditions, materials, and seismic input. Field walkdowns, which, shouldbe performed by qualified personnel who meet SQUG experience and training requirements for Seismic Capability-Engineers, are-used to screen the HVAC duct and damper systems for known seismic vulnerabilities and undesirable conditions that could lead to v

damage or failure in a seismic event. The walkdown team reviews the as-installed HVAC duct and damper system against a checklist of conditions to assess acceptability. As part of the field walkdowns, the review team selects and details representative, worst-case samples of duct runs and duct supports for analytical review. The guidelines include criteria for this analytical review.

.Appendix Aof this report summarizes the seismic experience database for HVAC duct systems.

'Appendix B summarizes the seismic experience database for dampers.

EPRI Perspective..

,The use of seismic experience data from actual strong-motion earthquakes has proven to be a reliable.and cost effective method for seismic capability verification. Accordingly, SQUG has developed the Generic Implementation Procedure, which structures the method and applies it to some twenty different classes of nuclear plant equipment, relays, tanks and heat exchangers, and electrical raceways.-Aspart of SQUG's ongoing effort to expand the method to new classes of equipment, this report extends the method to HVAC duct and damnper systems and provides guidelines to demonstrate seismic adequacy of existing HVAC duct and damper systems. The appendices to this report provide a summary of seismic experience database for HVAC duct and damper systems that has been assembled from power, and industrial facilities located in the'strong-motion-areas of over 20 earthquakes. This database provides valuable information on the performanceof HVAC duct and damper systems in earthquakes, and it will enhance the industry's overall database of seismic performance of equipment and systems.

" Keywords Damper Duct Earthquakes HVAC Seismic Effects Seismic Experience Data Seismic Verification Guidelines vi

CONTENTS 1 INTRODUCTION ........................................................................................................ 1-1 1.1 Background ....................................................................................................... 1-1 1.2 Overview of Guidelines .......................................... 1-1 2 APPLICABILITY AND QUALIFICATIONS ......................................................................... 2-1 2.1 Applicability................................................................. 2-1 2.2 Qualifications ................................................ 2-1 3 WALKDOWN SCREENING GUIDELINES ................................. 3-1 3.1 Overview of Walkdown Guidelines ........................................................................... 3-1 3.2 Structural Integrity Review ....................................... 3-2 3.2.1 D uct Span ....................................................... .............................................. 3-2 3.2.2 Duct Tie-dow ns ................................................................................................ 3-2 3.2.3 Duct Joints............................................................... 3-3 3.2.4 Riveted Lap Joints ............................................................................................ 3-3 3.2.5 Appurtenances ........................................... 3-3 3.2.6 Flexibility Mounted Heavy Equipment ................................................................ 3-3 3.2.7 Branch Flexibility ............................................................................................... 3-5 3.2.8 Cantilevered Duct..................  ; ..................................................................... 3-5 3.2.9 Ductu Corrosion . ............. ....... ..... ........ ,.............. ............ ............ 3-5 3.3 Support System Review ...... 3-5 33 Beam r. clampS ........

" ................ ........... ... ........ ...... ............ .......... 3-6 3.3.2 Channel Nuts .......... .... ... ............ .............. ....... ......... ..... .......... 3-63..-6...

33.3.3. Cast Iron Anchor Embedment r. ....- ........................ 3t...6........

6......

3,6 3.3.4 Broken Hatlrdware '..... ;.. .. ... ........... .... .... ..... ....... .............. ...................... 3-6 3.3.5: Su p. Corrosion ............ .......

Support ..... .... ............................ .................... 3-6 3.3.6 ' C oncrete*Q uality ..... .3-................................ . .................................... 3-7 3.3.7 Welded Attachments ....................................... 3-7 vii

3.3.8 Rod Hanger Fatigue ......................................................................................... 3-7 3.4 Seism ic Interaction Review ....................................................................................... 3-7 3.4.1 Proximity and Falling Hazards ......................................................................... 3-7 3.4.2 Flexibility of Attached Lines ....................... ............................... .................. 3-8 3.4.3 Differential Displacement Hazards ............... ;..................................................... 3-8 3.5 Pressure Boundary Integrity Review ................................. 3-8 3.5.1 'Duct Joints and Stiffener Spacing ............................... 3-8 3.5.2 Round Duct Supports ....................................... 3-9 3.5.3 Flexible Bellows....................................................................... .......................... 3-9 3.6 Selection of Bounding Configurations ........................................................................ 3-9 3.6.1 Selecting Bounding Duct Support Samples ..................................................... 3-10 3.6.2 Selection of Bounding Duct Configurations......................... 3-10 4 ANALYTICAL REVIEW CRITERIA ........................................ 4-1 4.1 Overview of Analysis Criteria.......................................4-1 4.2 Dead Load and Seismic Stress .................................... 4-2 4.2.1 Allowable Bending Stress for Rectangular Ducts ..................... 4-4 4.2.2 Allowable'Bending Stress for Circular Ducts .................................................... 4-4 4.3 Pressure Stress in Ducts .............. ........................................................................... 4-6 4.3.1 Pressure Stresses in Rectangular Ducts ........................................................... 4-6 4.3.2 Pressure Stresses in Round Ducts .............. o.................................................... 4-9 4.4 Pressure Stresses in Stiffeners ................................................................................ 4-10 4.4.1 Stiffener Evaluation for Rectangular Ducts ........................................................ 4-10 4.4.2 Stiffener Evaluation for Round Ducts ............................................................. 4-12 4.5 Duct Support Evaluation ..... .......... ........................................................................ 4-13 4.5.1 Metal Fram e ............ .... ,.. ............................................................................ 4-13 4.5.2 -'Rod Hanger Fatigue Evaluation................ ............ .................... 4-14 14...................

4.5.

.45.3 *! n hhorage Ealation r g Evalua ion .... . ... .*..... ..i ........................ . :.:...:...ii....4-i 4....5...

4.5.4 ý**Redundancy and Consequence Test *....... ............ ..... 415 5 DOCUMENTATION................................................ 5-1 6 OUTLIERS .................................... 61 6.1 Iden'tification of Outliers .................................... ....... 6-1 6.2 Outlier Re'solution ......................................... o............. 6-1

.viii

7REFERENCES .................................................... 7-1 A HVAC DUCT SYSTEM EARTHQUAKE EXPERIENCE DATA .................................... A-1 A .1 Introduction .............. .............................................................................................. A -1 A.2 Earthquake Experience Database .................................. A-1 A.2.1 Facilities Surveyed in Compiling the Database ................................................ A-2 A.2.1.1 1983 Coalinga, California Earthquake .................................................... A-34 A.2.1.2 1984 Morgan Hill, California Earthquake .................................................... A-35 A.2.1.3 1985 Mexico Earthquake .......................................................................... A-36 A.2.1.4 1987 New Zealand Earthquake .............................................................. A-37 A.2.1.5 1987 Whittier, California Earthquake ......................... A-39 A.2.1.6 1988 Alum Rock, California Earthquake .................................................. A-41 A.2.1.7 1989 Loma Prieta Earthquake ........................................................... A-42 A.2.1.8 1990 Philippines Earthquake ................................................................... A-46 A.3 Summary of Observed Damage.; ....................................................................... A-50 B HVAC DAMPER EARTHQUAKE EXPERIENCE DATA ........................ B-1 B.1 Definition of Equipment Class ..................................... B-1 B.1.1 Equipment Anchorage ................................................................................. B-1 B.1.2 Equipment Applications ............................................................................. B-2 B.1.3 Application in N uclear Plants ............................................................................ B-2 B.2 Database Representation for Dampers ................................... ............................... B-2 B.2.1 Basis for the Generic Bounding Spectrum ................................................. B-16 B.3 Instances of Seismic Effects and Damage ........................................................ B-18 B.4 Sources of Seismic Damage ............................................................................... B-18 B.5 Caveats for Dampers .,,.-.. ............  %.............................. ............................ B-18 B3.6 Reý13ferences B....

19...........

DEVELOPMENT EC OF ALLOWABLE. SPANS FOR SHEET METAL DUC.TS .................. C-i C.1. ,Rectangua Ducts ............................................ ...C-C.2 : Circular Du t . ................I............I....................... C-2 D SEISMIC AND PRESSURE TESTING OF HVAC DUCTS....`,.......................... D-1 D.1 Introduction............................................. D-i D.2 ? Duct Test Programs

/HVAC ............. . .... ... .................. ................ D-1 D.2.1 Summary of Tests Performed for TVA Ducts ....................... D-1 ix

D.2.2 Summary of Tests Performed for Limerick Ducts ............................................... D-2 D.2.3 Tests Performed at Other Plants ...................................................................... D-3 D.3 Conclusions from Test Programs .......................................................................... D-3 D .4 References ................................................................................................... ........... D -4 E ROD FATIGUE EVALUATION GUIDELINES ... .... . E-1

.E.1 Introduction ....... ............................................... ..... ........................................... E-1 FGUIDELINES FOR LIMITED ANALYTICAL REVIEW OF SUPPORTS ............................... F-1 F.1 Introduction ...................................................................................................... F-1 F.2 Dead Load Check F-3 F.3 Vertical Capacity Check ........................................................................................... F-3 F.4 Ductility Check . ............................... F.4...........................

F-4 F.5 Lateral and Longitudinal Load Check ............. ................... F-7 X

LIST OF FIGURES Figure 2-1 Seismic Motion Bounding Spectrum ...................................................................... 2-2 Figure 3-1 SMACNA Duct Joints ................................................. ......................................... 3-4 Figure 4-1 Rectangular Duct Configuration .................................. 4-6 Figure 4-2 Value of u for Rectangular Ducts [15] ..................................................................... 4-8 Figure 4-3 Load Going to Stiffener on a Rectangular Duct When L/S > 10.0 [6] ..................... 4-12 Figure A-1 Sylmar Converter Station, 1971 San Fernando Earthquake. Strap-Hung and Wall-Mounted Duct with Wall Penetrations ......................... A-29 Figure A-2 Glendale POwer Plant, 1971 San Fernando Earthquake. Cantilever Bracket Supported Rectangular Duct ....................................... A-29 Figure A-3 El Centro Steam Plant, 1979 Imperial Valley Earthquake. Trapeze Rod-Hung Rectangular Duct With Close Up of the Trapeze Detail ...................... A-30 Figure A-4 Bay Milk Products, 1987 New Zealand Earthquake. Long Vertical Cantilever Supported by the Roof at One End and Guy Wires at the Other .................................... A-31 Figure A-5 Califomia Federal Bank Facility, 1987 Whittier Earthquake. Typical Strap-Hung Rectangular Duct with Vertical Cantilevers and Diffusers ..................................... A-32 Figure A-6 Watkins-Johnson Instrument Plant, 1989 Loma Prieta Earthquake. Large, Insulated Round Duct with Branch Ducts and Cable Supports ................................. A-33 Figure A-7 Pacific Bell Watsonville, 1989 Loma Prieta Earthquake. Run of Trapeze Rod-hung Rectangular Duct ......................... ............. .......... ........ .......................................... A-33 Figure A-8 Gates Substation, 1983 Coalinga Earthquake. An HVAC Diffuser Fell from the Suspended Ceiling ................................................... ........................................... A-34 Figure A-9 Wiltron Facility, .1984 Morgan Hill Earthquake. A 4-Foot Long Vertical Cantilever Broke from its Supporting Heade~rand Fell ................................................ A-35 Figure A-10 Wiltron Facility, 1984 Morgan Hill Earthquake.-A Branch Line Tore at a Wall Penetration Due to Flexible HeaderMotior i.,.'....;.......... ....,.....;.....J................. A-36 Figure A-i 1 Fertimex Packaging Plant, 1985 Mexico Earthquake. A section of Duct Tore when the Duct Jumped off the Final Support in a Long Run ................................ *.... A-37 Figure A-12 Caxton Paper Mill, 1987 New Zealand Earthquake. A long, Unrestrained Run of Duct Constructed of Riveted Lap Joints,(Top) and a Taped Repair of a Sheared Joint (Bottom ) ................................................................................................ A-38 Figure A-13 Ticor Facility, .1987 Whittier Earthquake. A Flexible Bellows has Torn Due to the Motion of an Attached Fan on Vibration Isolation Mounts: ........................... A-40 Figure A-14 Sanwa Data Processing Center, 1987 Whittier Earthquake. A Duct above the Battery Racks Deformed at the Joints of an Angled Offset Section .......................... A-41 xi

Figure A-15 East Ridge Mall, .1988 Alum Rock Earthquake. A Flexible Bellows Tore Due to the Motion of Attached Air Handlers' on Vibration Isolation Mounts ......... A-42 Figure A-16 Seagate Technology, 1989 Loma Prieta Earthquake. A Strap Support Broke and the Attached Duct Fell to the Floor ......................................................... A-43 Figure A-17 Watkins-Johnson Instrument Plant, 1989 Loma Prieta Earthquake.

The Flexible Bellows 'Connecting HVAC Ducting to an In-Line Axial Fan Tore ............... A-44 Figure A-18 Watkins-Johnson Instrument Plant, 1989 Loma Prieta Earthquake.

The Support Anchorage for a Roof-Mounted Duct was Distressed.................................. A-45 Figure A-19 Pacific Bell, Watsonville,:1989 Loma Prieta Earthquake. A Vertical

• Cantilevered Section of Duct Fell to the Floor with its Attached Diffuser ........................ A-46 Figure A-20 Valley Steam Plant Forced Draft System, 1971 San Fernando Earthquake ....... A-47 Figure A-21 Drop IV Hydro Plant,*1*979 Imperial Valley Earthquake. Ceiling Mounted D ucting .................................. ................... ................................................................... A -47 Figure A22 SCE Rosemead Headquarters, 1987 Whittier Earthquake. HVAC Dented from Sway of Adjacent Fixtures..: ................................................... A-48 Figure A-23 Watkins-Johnson, 1989 Loma Prieta Earthquake. HVAC Ducting Atop Roof Level .................................................................................................................... A -48 Fig ure A-24 Magnolia Plant, Burbank, Ducting at Induced Draft Fan, 1971 San Fernando Earthquake ...................................................................................... .... ............ A-49 Figure A-25 El Centro Steam Plant, 1979 Imperial Valley Earthquake ................................. A-49 Figure B-1 Exploded View of a Typical Damper .................................................................... B..

-3 Figure B-2 Typical Damper Blades .or Louvers ........................................................................ B-3 Figure B-3 Typical Damper Actuators ...................................... B-4 Figure B-4 Pneumatic Damper at El Centro Steam Plant Subjected to the 1979 Imperial Valley Earthquake ...... ..... .................................................................................... B-4 Figure B-5 Louver Style Damper on the Boiler Structure at the El Centro Steam Plant which Experienced the 1979.Imperial Valley Earthquake ...................... B-5 Figure B-6 Pneumatic Actuator at the Puente Hills *Landfill Gas and Energy Recovery Plant .................... ............................................... B-5 Figure B-7 Radial Type Damper at the El Centro Steam Plant which was Subjected to the 1979 Imperial Valley and 1987 Superstition Hills Earthquakes ................................. B-6 Figure B-8LouverType DamPer at Humb*oldt B..y...Plant.... ower ......................... B-7 Figure B-9 Radial and Louver Typ..Dampers at the,Humboldt Bay Pdwer Plant which Experienced the 1975 Femdale Eart.hquake ;*L..........:..............B-7 Figure"B-i 0 Motor-operated ,Damp~er at Adak*Naval Station, which ,Experience,d th-e 1986-

-Adak AlIaska Earthiqu ake'. .... B-B Figure B-11 Damper at Adak Naval S-ation ;B-8 Figure b-12 Pneumatically Conitrolled da&nper at UC Santa Cruz Applied Science Building'Subjected to -1989 Loma Prieta Earthquake.......,...................... B-9 ein Electric Figure ExB-13 .dte*9 Motor4N .Fir6.Dampjerait AES ...Placerita forarhridge"Earhq-uake, ý *:.;.

• .

  .*. ..Cogeneration

..... ..... Plant

,:.................................. B-1 0

• ~ Eperie-nbed the .1994 NorthrdeErhuk

~~ -

Figure B-i4 Pneumatic Damper with Long*A6uator" _tValley Steam Plant Which Experienced the 1971 San Fernando and the 1994 Northridge Earthquakes ........... B-11 xii

Figure B-15 Pneumatic Louver Control Damper at Pasadena Power Plant which Experienced Several Database Earthquakes B-1 2 Figure B-16 Heavy Pneumatic Controller with Independent Support for a Large Damper at Pasadena Power Plant Located Very High in the Boiler Structure................ B-12 Figure B-17 Floor-mounted Air Operated Damper with Remove Actuator at Burbank

,'.Power Plant Experienced the 1971 San Fernando and the 1994 Northridge

-Earthquakes.................... ....................................................................... B-13 Figure B-1 8 Large Independently Supported Damper Controller at the Burbank Pow er Plant ................................................................................................................... B-14 Figure 8-19 Inventory of Dampers within Experience Database ........................................ B-15 Figure E-1 Bounding Rod Fatigue Spectra ............................................................................... E-1 Figure E-2 Fatigue Elevation Screening Chart for 1/4 inch Diameter Manufactured All-thread Rods.'Weight Corresponds to the Total Supported Load (i.e., on both Rods).

Length Corresponds to Clear Length ....................................... .... E!2...........................

E-2 Figure E-3 Fatigue Evaluation Screening Chart for 3/8 inch Diameter Manufactured All-thread Rods.-Weight Corresponds to the Total Supported Load (i.e., on both Rods).

Weight Corresponds to Clear Length............ ................................................................... E-3 Figure E-4 Fatigue Evaluation Screening Chart for 1/2 - inch Diameter Manufactured All-thread Rods. Weight Corriesponds tothe Total Supported Load (i.e., on both Rods).

Length Corresponds to Clear Length.............................................................................. E-4 Figure E-5 Fatigue Evaluation Screening Chart for-5/8-inch Diameter Manufactured All-thread Rods. Weight Corresponds to the Total Supported Load (i.e., on both Rods).

Length Corresponds to Clear Length ............ ................................................................... E-5 Figure E-6 Fatigue Evaluation'Screening Chart for 3/4-inch Diameter Manufactured All-thread Rods. Weight Corresponds to the Total Supported Load (i.e., on both Rods).

Length Corresponds to-Clear Length ............................................................................... E-6 Figure F-1 Vulnerable Duct Elbow Adjacent to Rigid Lateral Restraint .................................... F-2 Figure F-2 Examples of Potentially Non-Ductile Connection Details and Configurations ......... F-5 Figure F-3 System Frequency,Estimation using Beam-on-elastic-foundation Approxim ation ................... ................................ . ..................................................... F-8 Figure F-4 Dunkerley's Equation Frequency Estimation Methodology ................ F-9 Xiii

LIST OF TABLES Table 2-1 Temperature'Limitations for Duct Materials ........................... 2-1 Table 4-1 Value of K for Rectangular Ducts [15] ............................... 49 Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database ......... A-3 Table A-2 HVAC Duct Seismic Experience Database .................................................... A-1 3 xv

-INTRODUCTION 1.1 Background This report provides guidelines for seismic adequacy review of HVAC duct and damper systems.

The screening guidelines are primarily based on seismic experience data that show that most types of HVAC duct and damper systems exhibit extremely good performance under strong-motion seismic loading, with the pressure boundary being retained in all but a handful of cases.

The guidelines provide a method to screen and identify features ?seismic vulnerabilities and weaknesses.

The guidelines rely on the evaluation of seismic failure mechanisms for duct and damper systems from seismic experience data presented in Appendices A and B of this report. The data show that the damage to duct systems are generally limited to direct seismic damage of the duct or supports, and local damage due to seismic interaction with adjacent commodities. Seismic damage to HVAC d uct'systems documented in the seismic experience database can be attributed to the following categories:.

Broken andFallen CantileveredSections. Cantilevered sections of duct and duct diffusers have broken due to high inertia loading at weak joints, and due to inadequate flexibility of short duct segments to accommodate header movement.

• Opened and Sheared Seams. Light gage circular duct constructed with riveted lap joints have opened up and sheared in past strong-motion earthquakes. This damage has occurred at locations subject to high bending strain in very flexible duct systems.

• Duct Fallen off Support. .The database includes one example where the end of a cantilevered duct section jumped off of its end hanger support and was damaged. The duct was not tied

".:to the support, and was subject t6 high levels of seisnmic motion.-.

S Equipmenton Vibration Isolators. HVAC dudct-ha'sbeen damaged by excessive movement of'in-line equipment components supported oin vibration isolators.' **-.-*""ýi::;: *:*

The seismic experience database indicates hat dampers ossess haacteristics'that generally preclude &imiage in earthquakes Theexperience datab esecontains no instances of damage or

• • -:.1 ÷ significant seismic effects to dampers or their actuators.

1.2 Overview of Guidelines The guidelines for seismic adequacy review of HVAC duictand d'ampei systems include the following sections:

1-1

Introduction

" Applicability and Qualifications (Section 2)

• Walkdown Screening Guidelines (Section 3)

• Analytical Review Criteria (Section 4)

Section 2 provides general requirements the HVAC duct and damper system must meet to be able to use these guidelines for seismic verification. Section 2 also includes qualification requirements for individuals who perform the seismic adequacy review.

Section 3 presents guidelines for conducting in-plant seismiciadequiacy review of the HVAC duct and damper system'fi including supports. These walkdownguidelines are used to screen out potential failure modes indicated by seismic experience data, and to ensure database representation of the duct and damper system. As part of the walkdown, repreisentative worst-case examples of duct supiporIs are identified by the walkdown team and detailed for analytical review. In addition, representative worst-case examples of duct runs are identified by the walkdown team and detailed for analytical review for duct systems that require pressure boundary integrity to be maintained.

Section 4 includes criteria for performing analytical review of representative samples of duct systems and supports selected by the walkdown team. When 'these representative samples do not pass the analytical review, further evaluations should be conducted and the sample expanded as appropriate.

The results of the walkdown are documented in walkdown notes and forms included in Section 5.

Section 6 describes outliers and how they may be resolved.

References are included in Section 7.

A summary of the seismic experience database for HVAC duct systems is included in Appendix A. The seismic experience database for dampers is included in Appendix B. These appendices provide details on the performance of HVAC duct and damper systems at selected industrial and power plant facilities in actual strong-motion earthquakes.

4 . ' " i 4  : . " . . -

1-2

S 2' APPLICABILITY AND QUALIFICATIONS

.2.1. Applicability These guidelines apply to existing heating, ventilation and air-conditioning (HVAC) ducts, dampers and supports. Appurtenances such as registers, access doors, turning vanes, filters, louvers, air diffusers and similar components n6rmally attached to HVAC ducts are also included. These guidelines apply to duct fabricated of h6t-rolled and cold-rolled carbon steel, galvanized sheet steel, stainless steel and aluminum within the following maximum operating temperature limitations:-

Table 2-1 Temperature Limitations for Duct Materials Material Maximum Temperature Hot-Rolled Carbon Steel 400OF Cold-Rolled Carbon Steel 400OF Galvanized Sheet Steel 4006F Stainless Steel 400°F Aluminum 300OF The guidelines are applicable to any HVAC duct and damper system at any elevation in a plant where the nuclear plant free-field ground motion 5% damped seismic design spectrum does not exceed the Seismic Motion'Bounding Spectum of Reference [I]. The Bounding Spectrum is shown in Figure 2-1. '01 2.2 Qualificationis These guidelines are intended to be applied by qualified eigineers who meet the training and experience requirements defined in this section. -

2-1

Applicability and Qualifications 1.00 OO

• 0.9 0.60 0.40

-GroundAcceleration u 0.33 g 0.20

• 0.00 I I* I I I I ' T I I I p p 0 2 4 i' 10 12 14 1i 1 .20 22 24 26 2a 30 Frequency (Hz)

Figure 2-1 Seismic Motion Bounding Spectrum The Seismic Review Team (SRT) should consist of at least two engineers who meet the requirements for Seismic Capability Engineers (SCEs) as defined in Section 2 of Reference [1].

These individuals are required to be degreed engineers, or equivalent, who have completed a SQUG developed training course on seismic adequacy verification of nuclear power plant equipment. They are required to have at least five years experience in earthquake'engineering applicable to nuclear power plants and in structural or mechanical engineering. At least one engineer on each Seismic.Review, Team should be a licensed professional engineer.

In addition, qualified users of these guidelines must be familiar with the following topics:

" Content and intent of the-guidelines

• HVAC duct and support design requirements of the'Sheet Metal and Air Conditioning.

Contractor's National Association,-Inc.,,(SMACNA), inciuding References [4 through 7]

• Seismic experience data fr HACductand damper,systems .

2-2

wALKDOWN SCREENING GUIDELINES 3.1 Overview of Walkdown Guidelines This section presents requirements for performing the in-plant screening review of HVAC duct and damper systems for structural integrity, support review, seismic interaction, and pressure boundary integrity. Requiremerits are also provided for the selection of bounding/sample configurations for subsequent analytical evaluation. Analytical evaluation criteria are covered in Section 4. Screening -and evaluation Worksheets (SEWS) for recording information from the in-plant screening review are provided in Section 5.

The HVAC duct system seismic evaluation consists of two phases, (1) an in-plant screening review of field conditions to evaluate as-installed configurations for seismic deficiencies and (2) the analytical evaluation of selected duct and/or support configurations. The specific requirements for the evaluation are dependent upon the functional pressure boundary integrity

-requirements desired.

The in-plant screening review of HVAC duct systems encompasses the following:

" *Review duct system structural features that may lead to poor performance as illustrated by the seismic experience and test data (Section 3.2).

" Review support system for undesirable conditions that may lead to poor performance (Section 3.3).

• Review potential seismic interaction hazards (Section 3.4).

• Review duct system fea tu'ries toprovide a high confidence level that pressure boundary integrity is assured. These requirements are based on seismic experience and test data (Section 3.5).

Identify bounding configurations/samples for analytical evaluations (Section 3.6).

Items not meeting the in-plant screening review should be identified as outliers. Outliers require a more detailed review (see Section 6).'

An analytical evaluation should also be conducted for bounding configurations/samples of duct and/or supports selected during the in-plant review. Where pressure integrity is required following an earthquake, duct and support configurations should be selected to provide representative, worst-case, bounding samples.This will typically involve a careful review of available drawings and Collection of as-built informatin'. Analysis of bounding configurations for duct and supporitsneedinrg p'r'essuire boundairyintegrity canbe used to assure performance of 3-1

Walkdown Screening Guidelines a larger duct population. Where structural integrity (prevention of collapse and falling) is the only concern, analysis of a-random sampling of supportconfigurations is sufficient, along with the satisfaction of the in-plant screening review requirements. If the selected configurations do not pass the analytical review, the sample population should be expanded to identify the population of HVAC system configurations that meet the required seismic criteria.

Regardless of the pressure boundary integrity requirements, the HVAC duct evaluation includes an assessment of structural integrity and potential interaction, and analysis of support configurations. If pressure boundary integrity is required; the HVAC in-plant screening review' also includes requirements for.duct pressure boundary assessment and a selection of bounding configurations for analysis. Items not satisfying the analytical evaluations are outliers that may require more detailed analysis or modification.

3.2 Structural Integrity Review This section describes HVAC duct and support attributes for review'during the in-plant screening review walkdowhs. These attributes have led to poor seismic performance based on past earthquakes and testing (see Reference 3, Appendix A, 'Appendix B and Appendix D).

3.2.1 Duct Span Duct span governs the'seismic and dead load stress in the duct. Developing allowable duct spans and maximum cantileverlengths forvarious duct sizes prior to the in-plant screening review will facilitate the scre6ning of as-installed spans. An example of how allowable spans can be developed is given i~nAppendix C which is based on the analytical review requirements presented in Section 4.1. Lateral and vertical spans that appear to be beyond the developed requirements should be noted for further evaluation.

3.2.2 Duct Tie-downs Ducts should be secured to their supports to preclude the possibility of displacing, falling or sliding off during a seismic event. Systems do not have to be secured to every support unless' the supports are at the maximum spacing described in.Section 3.2.1. The HVAC duct should

.be securelyattached to the last hanger supportathe terminal end of the duct run. Similarly,

• 1. s-upports config-ued to limit the ]atei-a mov ement ofhe HVAC du*t system should also'be

..attached to the duct. Seismic experience data indicate tfiatA mde of failure for HVAC dtict systems subject t6'eirthquake loading isthe duct falling offof d supports.-An example of this occurred at the Fertimex plant during the61985 Mexico City erthquake (Figure A-l1 of AppendixýA).

The SRT should use experience and judgment when evaluating where duct tie-downs are required.For example, attachment to the last support is no required if the distance from the end of the:ductt0o the next t6 last support does niot exceed the maimirim allowable cantilever length.

In this case the duct would be seismically adequate without taking credit for the last support in the ýduct run. " - - - -

3-2

Walkdown Screening Guidelines 3.2.3 Duct Joints HVAC joints should be visually inspected to verify their structural integrity. Joints

" . (including connected tees and elbows) that are observed to be loose, incomplete, corroded, or

-otherwise suspect (such as those repaired with duct tape or fiberglass, or missing rivets, screws, etc.) should be reviewed in detail. Seismic experience data have shown that such joints are often the point of excessive leakage or failure of HVAC systems in an earthquake. A corroded riveted duct joint failed at the Caxton Paper Mill as a result of the 1987 New Zealand earthquake (see AppendixA,.Section A.2.1). In addition, HVAC without pressure boundary requirements and with runs consisting of slip joints without pocket locks, rivets or screws should be reviewed to assure that the differential displacement between the two adjoining ducts due to seismic loading wvill not causejoint separation. Figure 3-1 shows different SMACNA duct joints as described in Reference [4] .to aid in identifying slip-type joints.

3.2.4 -RivetedLap Joints Round HVAC duct with light gage riveted lap joint construction' should be considered outliers and subjected to more detailed investigation. The seismic expeiience database contains isolated cases of damage occurring to this kind of duct construction, such as the failure at the Wiltron Electronics Plant during the Morgan Hill earthquake (Figure A-9 of Appendix A).;More detailed investigation should be performed to assure the seismic adequacy of this type of duct.

3.2.5 Appurtenances Appurtenances attached to6HVAC ducts must be checked to assure they will not fall in the event of an earthquake. This equipment includes items such as dampers, turning vanes, registers, access doors, filters, louvers, and air diffusers. Earthquake experience data have shown that intake and discharge screens and vanies that are inadequately attached to the duct (i.e. only slipped into place and not fastened'with screws or rivets) have fallen during seismic events.

Figure A-8 of Appendix A shows this type of failure. Appurtenances not positively attached to the duct that appear to beat risk of falling during an earthquake should be evaluated to determine if failure will affectlthe functioning of the HVAC system and whether they will become an interaction hazard with"other neaiby safety related equipment.-Appurtenances projecting from the duct (cantilevered) should bereviewed to'assure connections are seismically adequate.

3.2.6 Flexibility.Mounted Heav Eqiment

H'vAc systems often have heavy, pieces of mechanical equipment mounted in-line with theduct.

. .Examples include fans, olers, dryers, dampers with motor operators; and bloivers. Earthquake experience datfi have shown that large pieces of equipment mounted in-line on flexible supports

• (e.g.,4without lateral and longit6dinal braci ig).can damage the duct from excessive displacement during an earthquake. Thisoccu'rred at the"Watkins-Johnson Plant during the 1989 Loma Prieta earthquake (Figure'A-17 of Appendix A)Mechanical equipmint should be investigated to determine if the joints connecting the equipment to the.ductare sufficiently, flexible to accommodateJany expected swvinging of the equipment'during a-seismic event. Potential interiactions betwween -sWvi nging mechanical 'equipmentf aiiid the HVAC duct or other safety related equipment should also be investigated (see Section 3.4).

3-3

Walkdown Screening Guidelines T DAFVE SLIP, STANDING DONE SLi:

• LAP T4--REINFORCED T-2 I-M S I F 1IP .

PLAIN "5"SUP TAHEMMED ""SUP. REINFORCED-S- SLIP, T48DOUBLE 5"S'SWIý

. T-84 REINFOWCEDW TA- rT4& REINFORCED) .T-7 H

INSIDE SUP JOINT 'STANDING S STA bNGS iALT.1 STANDING S 4ALT.)

-,T.IO

1.11i T-12 T.9

... ANINS- "-" STANDIINGS JL STANDING SEAM :ANGLE REJNFORCED."

(BAR REINFORCED) [ANGLE REINFORCED) T.15 -.STANDING SEAM T-13 T-4.

T,1 T.7POCKET LOCK BA)i-REINF ORCEO CAPPED FLANGE POCKET LOcK, :T-4 IT-118 REINFORCED POCKET LOCK) T-16 j ~On T1WELDED FLANGE - -_.,.COMPANION ANGLES' FLANGED - . LANGED:

It.. -,IORbr&CEDW .FJ -. CAULK OR GASKET) :-.CAUL ORTGASE -. I.AuL O S Fiure 3-1 i' SMACNA DuctJoints .

Heavy equipment with-connected HVAC du-tmay beflooi-mi ntedon vibration isolation pads. Earthquake experience]data have shown examplesof excessive leakage and failures of such HVAC systems due ,to'insufficient restraint of.this equipment-Excessive leakage and failures have been caused by flor-m-ounted equipment falling off their isolation pads and damaging attached du'cts in'the process. Figure A-13 in Appendix A-shows one such failure wvhere a flexible bellows was torn due to the motion of an attached fan on vibration isolation mounts.

3-4

Walkdown Screening Guidelines The SQUG GIP [1] provides guidelines for seismic verification of HVAC equipment such as fans (axial and centrifugal), air handlers and chillers; Heaiy equipment that is flexibly'supported or on vibration isolation pads should be evaluated separately using the SQUG GIP or identified as outliers for further evaluation.

132.7 Branch Flexibility Earthquake experience data have indicated that "hard points" are prone to seismic damage.

• Examples of hard points include locations such as wall penetrations and rigid supports on short stiff branches that are attached to flexibly supported duct. This type of seismic damage occurred at the Wiltron Electronics Plant during the Morgan Hill earthquake (Figure A-10 of Appendix A). Short, stiff branches on a flexibly supported header should be identified as outliers and checked for adequate flexibility to accommodate the expected header motion during a seismic event based on the guidelines in Section 4.1.

3.2.8 CantileveredDuct Earthquake experience data include isolated cases of cantilevered duct sections separating and falling from the main duct heade-r. An example of inadequate attachment occurred at the Pacific Bell Watsonville facility during-the Loma Prieta earthquake where a vertical cantilevered duct section separated and fell to the floor (Figure A-19 of Appendix A). Another example occurred

-at the Wiltron facility during the Morgan Hill earthquake where a vertical cantilever broke from its supporting header and fell (Figure A-9 of Appendix A). Cantilever duct sections should be adequately restrained to prevent excessive loads at the cantilever attachment point. The cantilever should be supporteddso that the maximum allowable cantilever length is not exceeded.

Unrestrained short cantilever ducts that meet the maximum allowable cantilever length should be reviewed to insure positive attachinent to the supporting headers.

3.2.9 Duct Corrosion Excessive corrosion of HVAC ducts should be evaluated for its effect on structural integrity.

Light surface corrosion is generally not a concern butt heavy flaking or pitting might be. Seismic experience dati haveshown thats"giificant c 0osion maylead to poor seismic performance for

- many plant items. Corrosion reviews are especially important in damp areas of a plant such as

.pump h6u ses.-Evaluations should c`nsideran estimatedstrengthk redutiondue to corrosion.

Significant corrosi6n sholl.d gene~aly.b~ identified for repair.* tI 3.'Support Sytem R'eview This section describes support attributes for review. uring the in-plant screening review waikdoWns. These attributes have led t6 poor seismicpefrformaihe in similar distributed type.

Ssystem, ssuch is pipingcable tay an conduit systems [l and 3]. Existing duct systems judged to have similar, potentialy poor seismic performance attributes, shall be documented as outliers.

3-5

Walkdown Screening Guidelines 3.3.1 Beam Clamps Beam clamps should not be oriented in such a way that gravity loads are resisted only by the frictional forces developed by the clamps. Beam clamps oriented this way might loosen and slip off in an earthquake and possibly cause a collapse of the system.

3.3.2 Channel Nuts Channel nuts used with light metal strut framing systems.should have teeth or ridges stamped into the nut where itbears0on the lip of the channel when slip resistance is relied upon to maintain structural integrity. Laboratory tests have shown that in a seismic environment, channel nuts without these teeth or ridges have significantly lower slip' resistance capacity than those with the teeth or ridges. Excessive galvanization or loose and flaking galvanization on the strut channel may also lead to reduced bolt resistance to slippage. Channel nuts should be visually reviewed on a random basis to provide reasonable assurance that teeth or ridges are present when required for structural integrity, and that the nuts are properly engaged on the frame sections.

3.3.3 Cast Iron Anchor Embedment Threaded rod hangeraanchorembedments constructed of cast iron should be evaluated because of potential brittle failure modes'.Plant documentation'should be consulted to determine whether anchor embedments are cast iron. Earthquake experience data includes examples where heavily loaded rod hangers threaded into cast-iron inserts have failed [8]. Failure modes include anchor pullout and anchor fracture where rods are only partially threaded into the anchor.

3.3.4 Broken Hardware Any observed missing or broken hardware for HVAC duct and supports should be noted so that repair or replacement may be provided. This includes examples such as missing nuts or bolts on connections, bent or damaged support members, dented duct seams, separated duct joints, torn expansion joints'and similar defects. HVAC related hardware that is missing or broken should be evaluated to determine thte consequences that this would have on the HVAC system. In particular, it should be determined if the integrity of the -VAC pressure boundary could be affected.

3.3.5 Support Corrsion Excessive'corrosion of ,ACiduc supports and support comiponents (iicludiuhg anchorage) should be evaluated frrits effect on'structural integrit. Light surface corrosion is generally not "a concern but heavy flaking or*pitting might b~e.Seismic experience data have shown that significanit corrosion may.lead to poor-seismic performance for many plant items. Evaluations should considerihe effects of -anestimated stiength reduction or loss of support due to corrosion.

Significant corrosion sshould generally be id*etified foir repair.. -. ,

3-6

Walkdown Screening Guidelines 3.3.6 Concrete Quality Gross defects or large cracks in the concrete to which the duct supports are attached should be evaluated for their potential effects on seismic performance.visibly large cracks, significant

'spalled concrete, and serious honeycombing in the vicinity of HVAC duct support anchors should beareas Iconcrete considered as gross defects. The walkdown team should consider grossly. defective as outliers and include supports anchored to marginally defective concrete in the sample selected for the limited analytical review.

3.3.7 Welded Attachments Support connections containing obviously undersized welds, incomplete welds, or welds of poor.quality (i.e., with significant burn-through) require analytical review incorporating reduced capacities. Seismic experience data' and shake table tests have shown that welds not capable of developing the strength of connected members may be subject to a brittle-type failure mode during seismic loading..

3.3.8 Rod HangerFatigue Although no specific instance of fatigue failure has been identified for HVAC duct rod hangers, raceway shake table tests have shown that short, fixed ended, heavily loaded rod hangers may be

-subject to low Cycle, high strain fatigue failures during seismic events [1 and 8]. Rod hangers that may be subject to high strain low cycle fatigue effects should be investigated in greater detail. The rod fatigue _evaluaiionrequirements outlined in Section 4.4.2 should be used to address rod fatigue concerns. Rods to be evaluated are characterized as follows:

o Rods double nutted toflanges of steel members o Rods threaded into shell-typeIconcrete expansion anchors o Rods connected by rod couplers to non-shell type concrete expansion anchors o Rods threaded into rod couplers which are welded to overhead steel embedments.

3.4. Seismic Interaction Review. .

Requirements for the evaluation of s,eismic interaction follow theYequire menis outlined in Appendix-D of Referene[.

3.4." -Proximityand Falling Hazardsi athThe walkdown team soi aware of igsues associated withliseismic interaction and be aleforpontial seismic interactionhazards. Duct systems attached to or in the vicinity of unanc nred or unrestrained bock wals should be noted and evaluated. Only cmpnents credible and significantifnteractioniSouirces should be considered as outliers. Damage that may occur to the duct itself as well asto any s-afety* related equipment that the duct may interact with should be considered.

3-7

Walkdown Screening Guidelines If an isolated support with questionable structural adequacy is found, the walkdown team should perform further evaluation of its adequacy or exercise judgment regarding the likely consequences of failure. If the adjacent spans are not excessive (see Appendix C for span length requirements), and if adjacent supports have a sufficiently high factor of safety (see Section 4.4),

failure of a single support can be acceptable if the walkdown team judges the adjacent supports have adequate margin for the increase in distributed load. The effect of the assumed failed support swinging or falling should be evaluated as a seismic spatial interaction hazard for fragile components in proximity.

3.4.2 Flexibility of Attached Lines Distribution lines such as small bore piping, tubing, conduit or cable that are connected to dampers can potentially fail if there is insufficient flexibility to accommodate relative motion between the damper and the adjacent equipment or structures. Straight, in-line connections in particular are prone to seismic damage or failure. The walkdown team should review distribution lines connected to dampers to insure there is adequate flexibility between the damper and the first support on the building or nearby structure.

3.4.3 DifferentialDisplacement Hazards Ducts spanning from one structure to another should be checked to assure that they can accommodate any relative movement of the structures. Experience data indicate there can be excessive leakage or failures for duct systems without sufficient flexibility at spans experiencing differential displacement [3]. If this condition is identified, stress criteria established in Section 4 of this report should be used.

3.5 Pressure Boundary Integrity Review This section applies to HVAC duct systems where a high confidence level of pressure boundary integrity is required for functional considerations. Examples where pressure boundary integrity may be required include the following:

" Systems with little or no margin for airflow

" Systems where leakage could significantly change system balance

• Systems that separate clean from potentially contaminated or hazardous material (such as battery room exhaust).

The following are in-plant screening requirements to achieve a high level of confidence of pressure boundary integrity.

3.5.1 Duct Joints and Stiffener Spacing Stiffeners prevent bulging of the duct panels due to internal pressure. Lateral joints such as companion angles, and lateral reinforcements, typically of steel angles, are considered as stiffeners. Earthquake experience and test data have demonstrated that duct systems that met the SMACNA guidelines performed well during earthquakes. Items to be checked for the given 3-8

Walkdown Screening Guidelines system operating pressure requirements include sheet metal gage, stiffener size and spacing, and panel dimensions. For bolted duct connections, it is also necessary to check minimum flange height, number of bolts, maximum hole spacing, and ring size where segments of round duct are bolted together. Applicable sections from the SMACNA standards include Section 7 of Reference [6] Chapters 4 and 12 of Reference [7]; and Chapters 1 and 3 of Reference [5].

3.5.2 Round Duct Supports

'Round HVAC duct runs supported such that the duct is point loaded should be considered outliers unless theduct is reinforced at the point of support. An example of this situation is a round duct supported by a rod hanger without a saddle.

3.5.3 Flexible Bellows Flexible bellows connecting HVAC duct to in-line equipment may become damaged if they do not have enough slack to accommodate differential motion between the equipment and the duct. Bellows are typically not designed to resist any large'differential motions imposed by the earthquake. If reasonable estimates of bellows flexibility cannot be determined by judging the available slack in the as-installed configuration, then manufacturer's data should be reviewed.

3.6 Selection of Bounding Configurations As part of the in-plant screening review, representative, worst-case HVAC duct and duct supports should be selected as bounding configur'ations.' The extent of the sample should be determined by the Seismic Capability Engineers based on the diversity, complexity and extent of the systems being reviewed. The samples should include representative samples of the. -

major different types of duct and duct supports for the HVAC duct and damper systems being reviewed. As a general guideline, 10 to'20 different sample supports'and 1 to 4 sample duct runs should be selected for facilities evaluating multiple HVAC systems.'These selected configurations should be evaluhated, using the analytical review criteria in Section 4. Detailed-evaluation of bounding, worst-case configurations assures the seismic adequacy of the entire population. When selected c6nfigurations do-n6t pass tihr*alyticali'e-vie-wthe selected. .. ' . i population should be expinded to idefitify the population of.HVAC system configurations that-

... :meetthe required seismic critera.For example, all supports and duct runsthat are represented by the'item that failed should be located and identified for modification or further (more

detailed) review.,-

Theprocedlure for the selectio~i~of boundary configurations for duct and sirt. system analytical review is dependent upon thefunctional requirementsof the system. For duct systems requiriing structural integrity or reasonable assurance for-pressure boundaryintegrity (where potential sm all tears orleaks are acceptable).,;the sample selection only needs to include worst-case bounding-duct supporis. For systems where full pre~su'e boundary integrity is required, the worst-case bounding sample should include theduct run itself as well as the

• SupportS. -- -

3-9

Walkdown Screening Guidelines The walkdown team needs to understand the analytical review requirements presented in Section 4 prior to performing in-plant screening reviews and selection of bounding configurations. The goal is to establish a biased, worst-case sampling, representative of and bounding the major different HVAC configurations in the plant. This bounding of worst-case samples will be subject to analytical review.

Notes'should be taken describing the basis for selection of eiach configuration. The location

of the selected configuration should be noted, and detailed sketches of the as-installed .

condition should be made. As-built sketches should include the duct and support configuration, dimensions, connection details, anchorage attributes, member sizes, and loading. Any additional information that may be considered relevant to the seismic adequacy of the selected configuration. should be noted in detail.

Building elevation should be taken into account when Choosing HVAC duct configurations as bounding samples. Identical systems at two different elevations in the plant experience different :seismic environments. The higher the building elevation, the greater the seismic demand. Therefore, it is possible that a system appearing to have few seismic vulnerabilities which is located at an upper elevation in a building may.actually, have a greater probability of failure than a system located at a lower elevation with a worse configuration. The walkdown team members should acquaint themselves with the differing seismic demand environments in the buildings being inspected by reviewing the floor response spectra before selecting the bounding sample.

3.6.1 Selecting Bounding Duct Support Samples The most heavily loaded support for each duct configuration should be selected as a bounding case. Long spans, insulated duct, supports carrying multiple ducts, top supports of vertical runs, heavy in-line components and isolated "stiff' supports on rod hung systems are indicators of heavy load. Duct support configurations to consider are long HVAC runs with few supports providing lateral or longitudinal restraint, long vertical runs, runs with seemingly weak curved sections, and runs with large, flexibly mounted in-line equipment. Of paiticular importance are duct supports that aplpearto have'more loading than originally designed for. Heavily loaded supports can be identified by the piesence of other plant components attached to the supports, such as supports for pipe, cable trays, and conduit.

Selection of a boundinj duct support should c6nsidericonditions where anchorage appears tobe the weak link'in theload path. Duct supports with anchorage'that appears marginal for the supported weight should be investigated. Anchorage with undersized weldsincomplete elds,.

or welds of poor quality should also be evaluated. Overhead support steel, such as st~ee angle, tsed Spcifically as an anchoi point to support the duct system should haveits anchorage to the

building structure evaluated. .

3.6.2 SelectionofBoundingDuctConigurations ,

--When appropriate; the selection should include duct systems with evidence of extreme or over-pressiue loads; anrd/or duct systems that appear to have unusual loading conditions. Examples include du~cttru-ns thaitsupport other equipment items'(such as race w-ays or piping), ducts that are

• shielded, heavily insulated or covered with fireproofing, and ducts with suspect flexible joints.

3-10

ANALYTICAL REVIEW CRITERIA 4.1 Overview of Analysis Criteria Analytical evaluations shall be performed on the selected bounding or sample HVAC duct and support configurationsi'equired to achieve duct system function following a seismic event.

The selection of duct and/or support configurations shall be consistent with the requirements of Section 3.6. The duct evaluation criteria-are based primarily on the design approach utilized in SMACNA's construction- standards for round and rectangular industrial .duct [6, 7]. Equations for computing pressure stresses in duct and stiffeners are taken directly from SMACNA standards. Use of this procedure resultsin 'a conservative estim ate of the true duct capacity and is compatible with data obtained from various test programs listed as References [9 throughl3].

The pressure boundary integrity review of HVAC duct considers the combined effects of pressure, dead weight and seismiic loads on the duct. The combined dead load and seismic

-stress is checked agaiins'talfactored allowable working stress for acceptance. The general stress combination equations are given below:

Horizontal Rectangular Duct fDL + [(EQ,) 2 + (EQh)2 ]0. < 1.7 Fb Eq. 4-1 Vertical Rectangular Duct 2

[(EQh,) + (EQh2)2 ]°'5< 1.7 Fb Eq. 4-2 Horizontal Circular Duct fDL +EQ <I13:Fb - Eq. 4-3

'EQh-<l1. 7 ,Fb Eq. 4-4 Vertical Circular Duct  :-. -

EQh < -. 7 - -Eq. 4-5 Pressure Stress-- -. -. -- _

4-1

Analytical Review Criteria Where:

fDL = Dead load bending stress p

Pressure stress Bending stress resulting from DBE seismic loads in the vertical direction EQh= Bending stress resulting from DBE seismic loads in the horizontal transverse direction. The additional, subscripts 1 and 2 refer to stress components from two orthogonal transverse seismic loading conditions Fb Bending stress allowable (normal working stress allowable)

Fp Pressure stress allowable (normal working stress allowable)

The 1.7 increase in allowable stress accounts for the short duration of seismic loading.

This increase is consistent with realistic allowable capacities for cable tray support components in Section 8.3.8 of Reference [1].

The effect of longitudinal seismic loading on the ducts is typically not significant since these forces are usually distributed over many support points..The-effects of longitudinal seismic loading should be combined with transverse and vertical seismic loading by the Square Root of the Sum of the Squares (SRSS) method in the stress calculations.

4.2 Dead Load and Seismic Stress Analysis for dead and seismic loads may be performed using either the equivalent static load method or the response spectrum method.

The equivalent static load method follows a tributary length approach using the spectral acceleration at the applicable frequiency (use peak floor spectral acceleration if frequency is unknown). An equivalent static coefficient of 1.0 times the spectral acceleration is used which is similar to the static coefficient used for equipment items addressed in Reference [1]. For this method, the bending moment is approximated by [6, 7]:

M =.w1 (Forducts.spanning over one or two spans) Eq. 4-7 8

w2 ~ -7

-M =-- (For ducts spanning over 3 or more supports) Eq. 4-8' Where:

w -appliedi linear load (lb/in) ->

I1 :--=tribuiary span(in) - .

M- = duct bending moment (in-*b) . -.

4-2

Analytical Review Criteria Other configuration anomalies, such as cantilevered duct sections, shall be considered on a

. .case-by-case basis.

Bending stresses due to axial response of a duct system may result if the axial run of duct is not

-braced in the longitudinal direction along the run of duct. If the axial restraint is provided by the first lateral restraint around a bend in the system, then the bending stress in theduct at the lateral.

restraint should be checked also for longitudinal motion of a tributary.span of the axial run.

Alternatively, longitudinal load resistance along an axial run may be provided by framing action between the-duct itself and the supports if the duct is adequately attached to the supports. In this case, the additional bending moment in the duct (about the transverse horizontal axis) must be checked.

Theresponse

• spectrum method requires modeling of sufficient ducting to analytically represent the expected dynamic response of the system. In general, this includes duct up to anchor points or equivalent restraint. Modal combinations are performed using the Square Root of the Sum of the Squares (SRSS) method. The analyses should consider all modes up to 33 Hz and include a minimum 90% mass participation.

For both methods, a critical damping ratio of 7% is appropriate for determining the seismic loads. This damping ratio is 'aconservative estimate of derived damping ratios from actual shake table tests [9 through 13].

Bending stresses for dead weight and seismic loads are derived using the duct section modulus as follows:

M fb= M Eq. 4-9 Z

where:

fb = Bending stress (psi)

• M = Applied bending moment (in-lb)

- 3 Z = Duct section modulus (in) .

..."tFor re ctangular du-cts,Refereince [6] limits the effective'are'aof sheetinietal for calculatioh of the

. *ductsection modulustoa 2-inch by 2-inch region at the four .orners of the duct A reduced sectioiiimodulus isthu's calculated by assuming onlythese corners are effective in resisting

. behding, For roound ducts, the full section is availablefor resisting the bending moment on the duct [7].

In addition, frequency correction factors of 0.59 and 0.87 for pocket lock and companion angle constructions, respectively, ijust betpplied to adjust he calculated rectangular duct frequency based on analytical corrI elation:f test results I(Appendix C). Duct joints that do not fit any of the Figure 3-1 duct joint types and can not be shown to behave ina manner.equivalent to one of them should be evaluated separately.

4-3

Analytical Review Criteria Allowable bending stresses differ for rectangular and round ducts, as detailed in the following sections.

4.2.1 Allowable Bending Stress for RectangularDucts The allowable bending stress for normal operating conditio~ns as specified by SMACNA[6],

is 8 ksi for carbon steel,:galvanized sheet and stainless steel -materials. This corresponds to 0.27 times the mini'mum yield strength of 30 ksi for typical materials used for industrial duct construction, within the specified temperature range.

The SMACNA standard for rectangular industrial duct construction [6] does not include design of duct fabricated of aluminum. A reasonable allowable'bending stress for normal operating conditions for aluminumi maylbe taken as 4.9 ksi. This corresponds to 0.27 times the minimum yield strength of 21.ksi for aluminum materials, times a yield strength reduction factor of 0.86 for temperatures up to 300 degrees Fahrenheit.

The normal allowable bending stress for rectangular ducts may be increased by a factor of 1.7 for DBE loads as detailed atthe beginning of Section 4. This increase may be taken for ducts having pocket lock and companion angle (or equivalent) joints. This applies to joint types.T-1 through T-3 and T-15 through T-24 of Figure 3-1, since Appendix D tests wereperformed on joints that are structurally similar to these types of duct joints. The normal allowable stress should not be increased by. 1.7 for DBE for ducts with potentially weakerjoints that rely on friction or crimping..Joints such as types T-4-through T-14 of Figure 3-1 are examples of potentially weaker joint.types..

Duct joint that-do not fit any of-the Figure 3-1 duct joint types and can not be shown to behave in a manner equivalent to one'of them should be evaluated separately.

4.2.2 Allowable Bending Stress for CircularDucts The allowable bending stress for circular ducts as specified by SMACNA [7] depends on the duct materials, operating temperature and diameter to thickness ratio.. ."

The normal allowable bending stress for hot rolled carbon ste (basedon a minimum yield stress of 33 ksi and maximum temn pe'ratuie of 400 degrees Fahrenh6eit) is as follows:,

Fb' 10.7 ksi 294(hot rolled carbon'steel)

Fb T 3140ksi,- foir Db/t 29(htrled M carbon steel)'ý

-o -(ot ' o 4-4

Analytical Review Criteria Where:

D = Diameter of circular duct (in) t = Duct thickness (in)

Fb = Bending stress allowable (normal working stress allowable)

The normal allowable bending stress for cold rolled carbon steel and galvanized sheet (based on a minimum yield stress of 32 ksi and maximum temperature of 400 degrees Fahrenheit is as follows:

'Fb = 11.0 ksi for D/t < 285 (cold rolled carbon steel, galvanized sheet)

_3140 Fb 1.

D/t ksi for D/t > 285 (cold rolled carbon steel, galvanized sheet)

The normal allowable bending stress for stainless steel is as follows. The following are minimum values that envelope parameters given iiithe SMACNA standard [7] for types of stainless steel typically used for duct. These values assume minimum yield strength of 30 ksi and maximum temperature of 400 degrees Fahrenheit. Higher allowable stress values may be obtained from Reference [7] for materials with a higher minimum yield strength and lower temperature, based on more detailed analysis.

Fb = 8.8ksi for D/t < 113 (stainless steel)'

Fb D/t ksi 993 for D/t > 113 (stainless steel)

The normal allowable bending stress for aluminum is as follows. The following are minimum values that envelope parameters given in the SMACNA standard [7] for types of aluminum typically used for duct. These values'assume a minimum yield strength of 21 ksi and maximum temperature of 300 degrees'Fahrenheit. Higher allowable stress values may' be obtained from Reference [7] for materials with'a higher minimum yield strength and lower temperature, based.,

on moredetailed analysis. -

Fb =6.0 ksi for D/t < O(lmnm

-662.

Fb- - ksi for D/t 10(aluminum)

- ~D/t.

,The normal ýhlowable beniding'stress for round ducts may be increased byafact6r'of1.7 for.

DBE loads as detailed at the beginning of.Seciion 4. This increase may be taken for ducts having pocket lock and companion afigle (or equivalent) joints.This applies to j6int types T-I through T-3 and T-15 through T-24 of Figure 3-1,;since Appendix D tests were pefformed on joints that are structuraliiy'similar to these types 4f duct joints. The normal' allowable stress should not be increased by 1.7 for DBE for ducts with potentially weaker joints such .as'types T4 through T- 14 of FigUre 3-1 -These jointsare potentially weaker because they rely on friction or crimping to transfer force across the joint.

4-5

Analytical Review Criteria

\

4.3 Pressure Stress in Ducts

• The effect of stress in HVAC duct material from internal pressure shall be accounted for in'the analytic evaluation of HVAC duct requiring pressure boundary integrity. These pressure stresses.

are checked against pressure stress allowables established in the SMACNA guidelines.

4.3.1 PressureStresses in RectangularDucts The SMACNA design of rectangular ducts is based on simplifying assumptions which permit the reduction of the analysis from a three-dimensional to a two-dimensional problem. Each of the four sides of the duct is assumed to act as an independent two-dimensional panel. Duct panel stresses are computed based on thin plate bending equations found in Reference [15].

For a given rectangular duct. the largest cross-sectional dimension (i.e. width or height) is used for stress analysis (see Figure 4-1). The applicable plate bending equations are dependent on the ratio of this maximum duct dimension, S, to the duct stiffener spacing, L.

- SwM9eX(W)

Figure 4-1 Rectangular Duct Configuration 4-6

Analytical Review Criteria Two simplified models are used to calculate duct pressure stresses. The following notations are used:

H Height of duct (in)

W = Width of duct (in)

.S = Max (H.W) (in)

L'= Stiffener spacing (in) t . Duct thickness (in)

.E = Young's Modulus of duct material (psi) adjusted for temperature. Use 9.5x10 6 psi for stainless steel, and 9.2x10 6 psi for aluminut m . Slightly higher

' values maybe obtained using more detailed analysis from Reference [7].

v = Poisson's ratio (dimensionless), taken as 0.30 for all duct materials p Applied pressure (psi)

IfL 5 S:

The duct panel is idealized as'one-way plate bending over a fixed-ended span, L, with axial in-plane tensile reactions resisting the increase in panel length.due'to bending curvature.

Let:

T = Axial tensile reactionresisting the increase in length due to bending curvature Db E t3/(12(1-v 2)) (plate bending stiffness coefficient) Eq. 4-10 u 0.5L(T/D)" 5 Eq. 4-11 To obtain u, use Figure 4-2 taken from Reference [15]. To use this chart,.the variable U, is first calculated as::

Uz (E2 tl)/((1_v2)2 p2 L0) Eq. 4-12 The quantity log 0(10 4 then O.U,"5) gives the ordinate of the curve in Figure 4-2, and the

,corresponding abscissa giv'es t erequired yalue of u. After determining U, the maximum

-stressesin the'plate are'calculated as follows:':

The maximu mitensile sfr'ess is [15:

1 q

=(E' U ) (t/L) /(3 (1 -V.)) - -13

The maximum bending stress is [5:

fj = (P12)(Lt)2 3(u tanh(u))/(u:, tanh(u)) Eq. 4-14 Maximum total priressure siress is: - E..4-1 fi- fl+f 2 Eq. 4-1 4-7

Analytical Review Criteria Figure 4-2 Value of u for Rectangular Ducts [15]

fL > S:"

• .Asthe stiffener'spacing exceeds the Width of the Critical duct section, the 'restrainingeffect of the panel sideiedges increasingly influenIces the stress distribution-within the panel, requiringthe use' of a second set of stress equations. -.

The panel is modeled as a u'niformly.loaded rectangular two-way ýiatefixed on the two opposite

• edges at the stiffeners and hinged onthe edges alongthe Sides. The maximum bending moment occurs at the mid-points of the fixed edges and is given by [15]: -

Mmax K pS. Eq.4-16 4-8

Analytical Review Criteria A list of K values for various UrS ratios, is given in Table 4-1.

Table 4-1 Value of K for Rectangular Ducts [15]

Values of Parameter K uS K US K 1.0 -0.0697 1.7 -0.1090 1.1 -0.0787 .1.8 -0.1122 1.3 -0.0868 1.9 -0.1152 1.4 -0.0938 2.0 -0.1174.

1.5- -0.0998 3.0 -0.1191

'1.6 -0.1049 .- 0.1250 The resulting stress is:

fp = Kp S2/t Eq. 4-117

.Through the use of equations Eq. 4-16 and Eq. 4-17, the panel pressure stresses can be calculated for any combination of system pressure and duct dimensions.

The allowable pressure.stress for rectangular carbon steel, galvanized sheet and stainless steel ducts is taken from Reference [6] as:

Fp .24 ksi (carbon steel, galvanized sheet and stainless steel)

The allowable pressure'stress for aluminum ducts may be taken as:

.Fp = 15 ksi (aluminum).

4.3.2 PressureStresses in'Round Ducts The pressurecapacity 6f circuar ducting is controlled byeither buckling of the'duct 'skin' or buckling (or yielding) of the'duct stiffeners asusumingciiegatiV-duct pressure. Duct skin buckling is influenced by.the duct end -onditions. The following notations are use d:

. . =Duct diameter.(in)," -

L:

• Stiffener spacinga(in) ...

t* Ductskin thickniess (in)

ýn Cnca duct presur (psi) c Critical stiffener spacing (in) 4-9

Analytical Review Criteria The critical duct pressure as determined in Reference [7] isdependent on the spacing of the-stiffeners. The critical spacing of the stiffness is defined as the spacing beyond which the duct is regarded as unstiffened, because the stiffener's are no longer contributing to the capacity of the duct to resist negative pressure. The critical spacing is as follows:

" , 1. 11.5 D -,r/-t Eq. 4-18 When the circumferential stiffener spacing is less than critical spacing, the allowable duct pressure is as follows:

P,= 18.1x10 6 (t/D) 2.5 (D/L) psi (carbon steel, galvanized sheet) Eq. 4-19 P= 16.1x10 6 (t/D)2"5 (D/L) psi (stainless steel) Eq. 4-20 Pn= 5.6x10 6 .(t/D) 2"5 (D/L) psi (aluminum) Eq. 4-21 When the duct is unstiffened or when the circumferential stiffener spacing is greater than thecritical spacing, the maximum duct pressure is as follows:

Pn= 16.2x! 06 (t/D)3 .psi (carbon steel, galvanized sheet) Eq. 4-22 P,= 14.5x 106 (t/D)3 psi (stainless steel) Eq. 4-23 P,=5.1xlWO (t/D)3 _psi (aluminum) Eq. 4-24

" These formula are valid for carbon steel, galvanized sheet steel and stainless steel up to 400 degrees Fahrenheit and for'aluminum up to 300 degrees Fahrenheit. They are based on temperature adjusted Young's M0duli of 29.5x106 psi for carbon steel and galvanized sheet steel, 26.3x10 6 psi for stainless steel, and 9.2x10 6 psi for aluminum. Slightly higher values for pressure may be obtained for specific stainless steel and aluminum materials at lower temperatures by using more detailed analysis from Reference [7].

The critical duct pressure should be used as the pr'essure stress allowable, Fp, and compared with the actual iiressure. -

4.4 PresSUre Stresses ii Stiffeners 4.4.1 Stiffener Evaluation for RectangularDucts`.'

Let:

q . Tribuaryl oad to stiffener(lb/in) - -

p = Duct piessure (psi) 1. - -

H = Height of duct (in) 4-10

Analytical Review Criteria W = Width of duct (in)

S = Max(H,W)

L = Stiffener spacing (in)

Fb(SFnm = Allowable bending stress in the stiffener (ksi)

Following analysis of the panels, the duct stiffeners are checked for two conditions:

• Maximum deflection < S/360 Maximum bending stress in the stiffener < Fb(STIFF)

The load transmitted to the stiffener from the duct panel is dependent on the ratio of L/S.

The tributary load to the stiffener, q, is calculated as follows:

For L/S < 2.0, q=pL Eq. 4-25 For 2.0 < LIS

• 10.0, q =p(1.25 - 0.125L/S) L Eq. 4-26 For L/S > 10.0, q = tributary load resulting from pressure p being applied on an area bounded by lines radiating at 450 from the ends of the stiffener (see Figure 4-3).

= p(S/2) Eq. 4-27 The stiffener stress evaluation forthe above loading conditions is dependent upon whether the stiffener ends are fixed or pinned.

Stiffeners welded at their ends to stiffeners from the adjacent side of the duct provide bending moment transition and are considered fixed. Such stiffeners should be analyzed as follows:

f=qS2 c/(l0I)<F - . . Eq. 4-28 A 3q S~(384 E 1):5 S/360:- *E.42 where:- -

I: . =iMoMment

=! of inertiaofthe&stiffener (in4 )-

c - Distancebetween neutral axis and extreme fiber of stiffener (in)

E = Young'siModulus' ofstiffen(er(psi) adjusied for-teiiiperature.,Use 29.5x 106 psi for stainless steeli and 9.2x 106 psi for alumijium. Slightly higher

.Values mayebbtainedUsingre detailed analysis from Reference [7].

d = maximum stiffener displacemenht(in) -

4-11

Analytical Review Criteria

.LOAD'GOING TO STIFFENER XS.

Figure 4-3

-Load Going to Stiffener on a Rectangular Duct When US , 10.0 [6]

Stiffeners are considered pinned regardless of whether they are bolted at their ends, tack welded, or.not connected at their ends. Such stiffeners should be evaluated as follows:

f =q S2 c/(8I) < Fb(STIFF) Eq. 4-30 d = 5q.S4/(384 E 1) < S/360 Eq. 4-31 The allowable bending stress in the stiffener is set as follows:

FsTim =- 24ksi (Carbon steel, galvanized sheet steel)

Fus-n -" 19.2 ksi (Stainless steel)

Ftsnm 13.1 ksi (Aluminum)

Inadequate stiffeners'will need to be supplemented. Stiffeners placed on only two opposite sides

"of a rectanguiar duct and meeting the above criteria are adeqiuate as long as the panel width is less than 72 iniche-s. For panels-of longer size,ý stress concentration becomes excessive and additional stiffeners are required.

ner Evaluation for,Rou'nd'Ducts 4.4.2 Stiffe".Z:s --

.The capacity of round duct stiffeners is controlled by buckling or yielding,ýwihere the theoretical buckling strengih'is proportional to the moment of inertia of the7Stiffener,-and the yield strength is proportional tothe area. Both of the following equations for moment ' f inertia and stiffener area for the res6ectiveý-mat-fial t'yp-(e'mu;tUt erfo*r-be Tatisfied [7]:.. .:

4-12

Analytical Review Criteria is> IMIN = 1.6x308 L D3 P, (Carbon steel, galvanized sheet steel) Eq.. 4-32 is > IMIN = 1.7xl0" L D3 Pl (Stainless steel) Eq. 4-33 is> IMIN = 5.0x10 8 L D3 P,(Aluminum) Eq. 4-34 As>AMIN = 6.3xiO 5 L D P, (Carbon steel, galvanized sheet steel) Eq. 4-35 As> AMIN = 7.6x10 5 L D P.(Stainless steel) Eq. 4-36 As> AMIN = 10.8x10-5 L D Pn(Aluminum) Eq. 4-37 where:

Is = Moment of inertia of stiffener (in4)

A8 = Area of stiffener (in2)

Pn = Applied pressure in duct (psi)

L = Stiffener spacing (in)

D Duct diameter (in)

Higher values may be obtained for specific materials and for lower temperatures by using more detailed analysis from Reference [7].

4.5 Duct Support Evaluation 4.5.1 Metal Frame The selection of support configurations for evaluation'shall be consistent with the requirements of Section 3.6. Simplified support evaluation requirements, consistent with those presented in' Section 8 of Reference [1] for limited analytical review of raceway supports, are applicable for the seismic adequacy verification of duct supports. These include the following checks:

1.. Dead load 2.'-Vertical_ capa'cit, (5 x Dead Load).

3.:3Ductility A

4. :Lateral loadcheck 5.- Longitudinal load check

6. Rod hangei fatigue evaluation '

4-13

Analytical Review Criteria The 5 times dead load check is used for overhead HVAC duct supports to account for the dynamic characteristic differences in terms of system damping between the HVVAC and raceway systems. That is, the 3 -times dead load check established for raceways is factored up,by'the difference in spectral acceleration demand due to the lower damping in HVAC systems (on the order of 7%) as compared to raceway systems damping (on the order of 15%).

A discussion of the requirements for the dead load check, vertical capacity check, ductility review, lateral and longitudinal loads checks is included in Appendix F. The rod hanger fatigue evaluation guidelinesare presented in Appendix E.

For systems in which detailed modal response spectrum analysis is performed, the duct support frame should be evaluated for the resulting seismic loads combined with dead loads.

Loads from other attached systems, such as conduit orpiping, should also be considered. All steel -components such as bracket members, support members, and internal support framing connections should be checked, using maximum allowables as defined in Part 2 of the AISC [2].

Exceedance of AISC allowable stresses is permitted as follows. If the high stress results in plastic hinge formation in support members; and support instability.(vertical-load-carrying capability) does not occur, and the additional support deflection can be accommodated by the duct system, the configuration should be adequate. The effecton pressure boundary integrity shall be considered in these cases of support overstress. The basis for the evaluation should be documented within the limited analytical review calculations.

The buckling analysis of vertic'al support members and lateral bracing should also follow the criteria of Part 2 of AISC [2]. It is recognized that many support configurations have structurally redundant members. If buckling is predicted to occurin a support member which does not affect the overall stability of the duct system, the support may still be acceptable. For example, if a lateral brace is found to buckle under imposed seismic loading, but vertical capacity is not jeopardized, the duct can be analyzed ignoring the presence of the brace. If the duct system stresses are acceptable without the lateral brace and spatial (proximity-related) interaction due to duct seismic displacement is not a problem, then the support is acceptable.

4.5.2 Rod HangerFatigue Evaluation Short, fixed ended, heavily loaded rod ha nger maybe subject to low cycle, high strain fatigue failures duringseismicevents [8]. Rod hangers of concern are typically of fixed end connections.

These rods aire characterized a's follows!:

e Rods double-nutted to flanges of steel members'

• Rods threaded into shell-type concrete expansion anchors

• Rods connected by rod couplers to non-shell type concrete expansion anchors

• Rods threaded into rod couplers welded to overhead steel embe`dments .

Rod hangers that may be'subject to high sirain 16w iycle fatiguei ffects should be'investigated in greater detail. The rod fatigue evaluation guidelines outlined in Attachment E should be used to address any rod fatigue concerns.

4-14

Analytical Review Criteria 4.5.3 Anchorage Evaluation Capacity values for anchors should be taken from Reference [1]. The provisions of these anchorage guidelines should be followed, including edge distance, bolt spacing; and inspection procedures. Tightness checks are not required for expansion anchor bolts that are normally subjected to tensile forces due to dead weight, since the adeqhacy of the anchorage set is effectively proof tested by the dead weight loading. This applies to expansion anchors for overhead and wall mounted supports.

4.5.4. Redundancy and Consequence Test Isolated cases of a suppbrt-not'meeting the'analytical review guidelines may be accepted" if the HVAC support system has redundancy so that postulated support failure would have no consequence to overall systemperformance. Adequate redundancy is demonstrated if the adjacent supports are capable of sustaining the additional weight resulting from the postulated support failure.

4-15

• 5 DOCUMENTATION A summary package should be assembled to document and track the Seismic Capability Engineer's evaluation activities. Documentation should include records of the HVAC duct and damper systemsevaluated, the 'dates of the walkdowns, .thenames of the engineers conducting the evaluations, and a summary of results. Recommend data sheets for the summary package are given in Exhibits 5-1 to 5-4 and are described below. The Outlier Seismic Verification Sheet (OSVS) given in Section 5, Exhibit 5-1 of the SQUG GIP [1] may be used to document outliers.

Exhibit 5-1 provides'Screening Evaluation Worksheet (SEWS) that can be used to document the walkdowns. The SEWS includes reminders,; as a checklistfor primary aspects of the evaluation guidelines; however, the walkd*wn engineers should be familiar with all aspects of the seismic eyaluation guidelines during in-plant screening reviews and not rely solely on the checklist.

The checklist items on the SEWS are worded so that all acceptable conditions are answered Y (for yes). Any condition'that is answered N (for no) or U (for unknown) is an outlier. The SEWS should be signed and dated by all members of the SRT.

Exhibit 5-2 provides a Duct Support Analytical Review Data Sheet for recording information on the supports selected as the worst case,. representative samples.

Exhibit 5-3 provides a Tracking Summary for the Duct Support Analytical Review Data Sheets.

As items are completed and resolved, the responsible engineers should initial the line item on the tracking sheets to confirm!final closure.

Exhibit 5-4 provides an HVAC Duct System Analytical Review Data-Sheet for recording information on the duct system selected as worst case, representative samples for systems required to maintain pressiiure boundary. ,

J.Exhibit 5-5 provide's a HVAC System Outlier S.eet (HSOS) foi documenting outliers.

-An outlier is an HVAC duct system or support feature that does` not meet the screening guidelines in Section 3,'or an HVAC.duct or support selection that fails aheinalytical review criteria in Section.4 The outlier sheet identifies the screening guidelines that are not met, the reasonsfor the outlier, and the proposed methedof resolvuingth oUtier. l Outliers are discussed iAn Section 6.0. ---

Photographs imay -beused to supplement documentation as required. When used as foi-mal documentation for the sumn&a-ypackages, phot6graphs sh6bUld be clearly labeled for identification. -

5-1

Documentation Exhibit 5-1 Sheet 1 of SCREENING AND EVALUATION WORKSHEETISEWS)

HVAC System I.D.

Damper Equipment I.D.

System Description and Boundaries HVAC System Locations and Reference Drawings Duct Materials and Sizes Linear Weight:

Duct Insulation Total References Concurrent Pressure and Temperature Applicability

1. Is the operating temperature less than the temperature limitations in Table 2-1? -a Y N U N/A

2. Plant ground spectrum is enveloped by the SQUG Bounding Spectrum (Figure 2-1)? Y N U N/A Does duct meet applicability criteria?,.. Y N U

...Pressure Boundar Integrity Review s' anresr boundary integnrity 6 *Y: N U N/A 1.Is anyspessrtohe required

- IFlthe aiswer to the above question is NO, SKIP THIS SECTION and proceed to the Structural Integrity Review.-.

~--2. 2~ Stiffener spacings are within the guidelines Y N N/A

- .,,-. flanged jointssatisfy SMACNA requequirements

.,Bolted Y !u N U N/A N/A

4. "N'opoint'supported round dc Y my. N U N/A

5. Flexible bellows ca-n accrmmodate motions":,: N uj N/A

6. No additional concavernsme?K .: Y Are the above caveats met?: - IY. 'N U 5-2

Documentation Exhibit 5-1 Sheet 2 of SCREENING AND EVALUATION WORKSHEET ýSEWS)

Duct System I.D.

Damper Equipment I.D.

• Structural Integrity Review

1. Support slans satisfy the criteria Y N U N/A

2. Ducts are properly tied-down to the supports Y N U N/A

3. Industry standard duct jointsare utilized Y N U . N/A

4. Slip joints can accommodate displacements Y N U N/A

5. Round duct joints exclude riveted lap joints Y N U N/A

6. Appurtenances are positively attached toduct Y N U N/A

7. Heavy in-line equipment is adequately restrained Y N U N/A

8. No stiff branch with flexible header Y N U N/A

9. Cantilevered duct section is attached to last support Y N U N/A

10. Ducts are free of corrosion detrimental to integrity Y N U N/A

11. System is free of obvious damage or defects Y N U N/A

12. No other concerns? Y N U N/A Are the above caveats met? Y N U Support Review

1. Beam clamps are oriented to preclude slipping off the support Y N U N/A

2. Channel nuts have teeth or ridges Y N U N/A

3. Cast iron inserts . . , Y N U N/A

4. No broken or obviously defective hardware Y N U N/A

5. Support is free of excessive corrosion Y N U N/A

6. Welded joints appear to'be of good quality Y N U N/A

7. Does the anchorage appear adequate? Y N U N/A

8. No stiff supports or hard spots In long flexible duct runs? Y N U N/A

9. No short, fixed ended heavily loaded rod hangers subject to potential fatigue failure?. Y N U N/A

10. No additional concerns Y N U N/A

Are the above caveats met? y N U Darnoer Review

1. Damper is similar to and bounded by the seismic experiencedata for

2. dampers'in Attachment B-:- Y 1N U N/A

'.'Damper'operator/actuator not of cast iron

2. N U N/A 3., Attached lines have sufficient slack and flexibility', Y. N U N/A

4. Damper controls mounted separately from the damper adequately anchored Y N U N/A S5. Motor or pneumatic operator mounted on the damperbhas adequate anchorage and load path . Y U N/A

6. Is duct at the damper location free from signs of distortion that could interfere N with damper operation? .. _. ' U N/A

7. No other adverseconcerns' Y N U N/A Are the above caveats met?. Y N U 5-3

Documentation Exhibit 5-1 Sheet 3 of SCREENING AND EVALUATION WORKSHEET (SEWS)

HVAC System I.D.

Damper Equipment I.D.

Seismic Interaction Review 1.. Free from impact by nearby equipment Y N U N/A

2. No collapse of overhead equipment, distribution systems or masonry walls Y N U N/A 3' Able to'accommodate differential displacements Y N.U N/A

4. No other adverse concerns Y N U N/A Y N U Are the above caveats met?

IS THE HVAC DUCT AND DAMPER SYSTEM SEISMICALLY ADEQUATE? Y N U Supports Selected for Analytical Review" Duct System Selected for Analytical Review Comments

'All aspects of the equipment s seismic adequacy have been addressed. - - -

'%Evaluaed by -Dt:_____

(All team members) 5-4

Documentation Exhibit 5-2 Sheet 1 of DUCT SUPPORT ANALYTICAL REVIEW DATA SHEET

.HVAC Duct System:- Selection No.:

Plant Location:__

Description and Sketch:

CERTIFICATION: (Signatures of at least two Seismic Capability Engineers are required; one of whom is a licensed professional engineer.)

Print or Type Name/Title Signature Date Print or Type Name/Title Signature Date 5-5

Documentation Exhibit 5-3 Sheet 1 of HVAC DUCT SUPPORT ANALYTICAL REVIEW TRACKING

SUMMARY

HVAC Duct Plant Selection Final Initials/

-System Location Number Resolution Date Designation 5-6

Documentation Exhibit 5-4 Sheet 1 of HVAC DUCT SYSTEM ANALYTICAL REVIEW DATA SHEET HVAC Duct System: Selection No.:

Plant Location:_

Description and Sketch:

SCERTIFICATION: (Signatures of at least two Seismic Capability Engineers are required; one of whom is a licensed professional engineer.)

Print or Type Name/Title ,'.. Signature: Date' Print or Type Name/Title,--. Signature Date 5-7

Documentation Exhibit 5-5 ý "

• Sheet 1 of HVAC SYSTEM OUTLIER SHEET (HSOS)

OUTLIER NO.

1. OUTLIER IDENTIFICATION AND LOCATION HVAC System I.D."

Location

2. OUTLIER ISSUE DEFINITION

a. Identify the screening guidelines that are not met, or indicate if the analytical review selection fails the analysis criteria.

Applicability Damper Review Pressure Boundary Integrity Interaction Effects Structural Integrity Review Support Analytical Review Support Review Duct Analytical Review

b. Describe all the reasons for the outlier:

3. PROPOSED METHOD OF OUTLIER RESOLUTION (OPTIONAL)

a. Define the proposed method(s) for resolving the outlier:

CERTIFCATIO: (Sigature f atlattoSimcCpblt.ninesaerqie;oeo hmi

b. Provide Information needed to Implement proposedimethod(s) for resolving the outlier:

!CERTIFIiCATION:, (Sign~atures of at leastitwto :SeismiecCapability Engin'eers'are required; one of whom is a licensed professional engineer.)

Printor Type Name/Title --- - Signature, Date Print or Type Name/Title *Signature Date 5-8

OUTLIERS 6.1. Identification of Outliers An outlier is defined as an HVAC duct, damper or support feature that does not meet the.

screening guidelines in Section 3,'or an HVAC duct or support selection that fails the analytical review criteria in Section 4. The guidelines and analysis criteria are intended to be used on a generic basis for seismic adeq~uacy review of HVAC systems (including supports). HVAC duct, dampers or supports that-do not pass the generic criteria may still be shown to be seismically adequate by obtaining additional information or by performing additional evaluations.

6.2. Outlier Resolution

• An outlier may be shown'to beadequiatefor seismic loadings by performing additional

• evaluations to demonstrate thereis adequate seismic margin. These additional evaluations and alternate methods shofiuld be thoroughly documented to permit independent review.

.Methods to determine the available seismic margin are contained in EPRI NP-6041-SL, RI [14].

• In some cases it may be necessary to exercise engineering judgment when resolving outliers, since strict adherence to the screening guidelines is not abýsolutely required for HVAC systems to be seismically adequate.-These judgments, however, should be based on a thorough understanding of the background and philosophy used to develop these screening guidelines as described in this report. The justification and reasoning for considering an outlier to be acceptable should be based onnmechanistic principles and sound engineering judgment.

The screening guidelines contained in'this report have been reviewed to ensure that they are appropriate for generic use; however,'.the altemative evaluation methods and engineering

- j.....Judgments used to1resolve outliers are not subject to the same level of pee review.7Therefore,

. ..the evaluations and judgmentsuused to resolve outliers should be thoroughl -documented so that independent reviews can be performed if necessary.

6-1

S7'

.. :REFERENCES

1. Seismic Qualification Utilities Group (SQUG), "Generic Implementation Procedure (GIP)

. for Seismic Verification of Nuclear Plant Equipment," Revision 3A, December 2001.

2. AISC, "Manual of Steel Construction," 8 "Edition, American Institute of Steel Construction, Chicago, IL, -1978.

3. Porter, K., G. S. Johnson, M. M. Zadeh, C. S. Scawthorn, and S. J. Eder, August 1993, "Seismic Vulnerability of Equipment in Critical Facilities: Life-Safety and Operational Consequences," Prepared for National Center for'Earthquake Engineering Research, November 1993.--

4. Sheet Metal and Air Conditioning Contractors National Association, Inc., "HVAC Duct Construction Standards, Metal and Flexible," Chantilly, Virginia, First Edition, 1985.

5' Sheet"Metal and Air Conditioning Contractors National Association, Inc., "HVAC Duct Construction Standards,. Metal and Flexible," Chantilly, Virginia, Second Edition, 1995.

6. Sheet Metal and Air Conditioning Contractors National Assdciation, Inc., "Rectangular Industrial Duct Construction Standards," Chantilly, Virginia, Copyright 1980.

7. Sheet Metal and Air Conditioning Contractors National Association, Inc.,

"Round Industrial Duct Construction Standards," Chantilly, Virginia, Second Edition, September 1999.

8. The PerformanceofRaceway Systems in Strong-motion Earthquakes,EPRI, and EQE Engineering: March'1991. Report EPRI NP-7150-D.

.. 9. Yow, Dr. J. Roland., "'Status Report on -theRecent'History of HVAC Ductwork Design for Nuclear Power Stations and Current Industry Activity for theCommittee on'Materials and

'Structural Design," Prepaied foir the American Society of Civil Enginfeers, September 1980.

10. ..McPherson, R.'Keith,l".Duct TestiReport to "iet ine Loiad Carryifi Capabilities and

-Cross Sectional Properties ofSafety.Related Duct forwashington Public Power Supply Steam Nucleair Project No. 2 April 1982 -

S11*. Desai,*S.C., et al., "Structural Testing of S6eismic Category i HVAC Duct Specimens,"

• Second ASCE Confereneeon CiVil Engineenng and Nuclear Power. Volume I. Knoxville, Tennessee, Sepiember 1980. V-7-1

References

12. Neely, B. B., and L.-Warrix, "A Procedure for Seismically Qualifying HVAC Ducts Used in Nuclear Power Plants," Second ASCE Conferenceon Civil Engineering and Nuclear Power, Volume I, Knoxville, Tennessee, September 1980.

13.. Kato, T. and T. Nakatogawa et al. "Limit Strength of Rectangular Air Ventilation Ducts Under Seismic Design Condition," Transactions of th.e .10th International Conference on Structural Mechanics in Reactor Technology, Volume K2, 1989.

14. A Methodology for Assessment of NuclearPower Plant Seismic Margin:Revision 1,EPRI, Palo Alto,*CA, NTS Engineering, Long Beach, CA, and RPK Consulting, Yorba Linda, CA: July 1991.Report NP-6041.

15. Timoshenko, S. and S. Woinowsky-Kreiger, "Theory of Plates and Shells," McGraw Hill, Second Edition, 1959.

7-2

HVAC DUCT SYSTEM EARTHQUAKE EXPERIENCE DATA A.1 Introduction.

• This.attachment documents the performance history of HVAC duct and duct support systems under seismic Jloading'. The bulk of data Was obtained from extensive field investigations of systems that have experienced strong motion earthquakes. Further information on the performance of HVAC duct systems was gained fr6m a literature search on earthquake damage in past earthquakes.

A summary of the known damage data for. the performance of HVAC duct systems when subject

• to seismic loading is presented. The seismic experience database includes many examples of ducts that have performed well in actual earthquakes. The presented data focuses on examples of ducts that have performed poorly in seismic excitations, with a discussion of the attributes of the

'installations that caused them to perform poorly.

• HVAC ducting is found at nearly all industrial sites. The seismic experience database therefore includes a vast amount of data' onthe survivability of ducting installed in many different ways, and experiencing mannydiffere -t seismic excitation levels. The large number of duct systems that have survived earthquakes indicate the inherent ruggedness of these systems. The limited,

• • -.smaller set of HYAC duct systems that have been found to have performed poorly in a seismic event point out key characteristics of HVAC installations that may contribute to seismic damage.

A.2 Earthquake Experience Database The seismic experience database is founded on studies of over 100 facilities located in the strong-motion areas of mori than twenty Strong-motion earthquakes that have occurred in the United States, Latin'Ameriea; New Zealand, and otherpartsof the World since 1971.

Th daetabaseA-) wa ofp dt - f The database was compiled through surveys of the following types of facilities: ,

" ' ýFssil-fueled power planis --

• Hydroelectric power pl ants-

• Electrical distribiution siubstations"-;* -

S*oil processin' and refiiing facilities ---- - -

water treatment and pumping stations W

A-I

HVA C Duct System EarthquakeExperience Data

• Natural gas processing and pumping stations

" Manufacturing facilities

" Large commercial facilities In general, data collection efforts focused on facilities located in the areas of strongest ground motion for each earthquake investigated. Facilities were sought that contained substantial inventories of mechanical and electrical equipment or control and instrumentation systems.

Because of the number of earthquake-affected areas and types of facilities investigated, there is a wide diversity in the types of installations included in the database. For the HVAC duct of focus in this study, there is a wide diversity in size, configuration, type of building, local soil conditions, and quality of construction.

The database currently includes in detail fourteen earthquakes from which duct data have been processed for this report. Each earthquake includes several different sites investigated within each epicentral area. The earthquakes investigated range in Richter magnitude (M) from 5.5 to 8.1. The Strong motion duration is as high as forty seconds. Local soil conditions range from deep alluvium to rock.

The buildings housing the ductwork have a wide range in size and type of construction. As a result, the database covers a wide diversity of seismic input to duct installations, in terms of seismic motion, amplitude, duration, and frequency content.

A.2.1 FacilitiesSurveyed in Compiling the Database Information on each database facility, its performance during the earthquake, and any damage or adverse effects caused by the earthquake were collected through the following sources:

" Interviews with the facility management and operating personnel usually provide the most reliable and detailed information on the effects of the earthquake on each facility. At most facilities, several individuals were consulted to confirm or enhance details. In most cases, interviews are recorded on audio tape.

• Observations by earthquake reconnaissance teams are documented and photographed.

Typical observations include descriptions and details of both damaged and undamaged installations or equipment and any indications of the cause of damage, such as substantial ground settlement or evidence of seismic interaction.

" The facility operating logs provide a written record of the conditions of the operating systems before and after the earthquake. Operating logs list problems in system operation associated with the earthquake and usually tabulate earthquake damage to the facility. Operating logs are useful in determining how long the facility may have been out of operation following the earthquake and any problems encountered in restarting the facility.

• The facility management often produces a report summarizing the effects of the earthquake following detailed inspections. These reports normally describe causes of any system malfunctions or damage.

" Earthquake damage can often be inspected prior to repairs if the facility can be surveyed immediately following the earthquake. This has been the case in most of the earthquakes included in the database.

A-2

HVA C Duct System EarthquakeExperience Data Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database Estimated Peak Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Large electrical substation San Fernando, CA Sylmar Station 0.65 Earthquake 1971 (M6.6)

Large electrical substation Rinaldi Receiving 0.50-0.75 Station Valley Steam Plant Four-unit gas-fired. 0.40 power plant Burbank Power Plant Six-unit gas-fired power plant 0.25-Glendale Power Plant Five-unit gas-fired 0.30 power.plant 0.30 Pasadena Power Plant Five-unit gas-fired power plant Point Mugu, CA .Ormond Beach Power Large two-unit oil fired. 0.10 Earthquake 1973 (M5.7) Plant. power plant Humboldt Bay Power Two gas-fired units, 0.30*

Ferndale CA Earthquake 1975 (M5.5) . Plant one nuclear unit Santa Barbara, CA Goleta Substation Electrical substation 0.26*

Earthquake 1978 (M5.7)

Four-unit gas-fired Imperial Valley, CA El Centro Steam Plant 0.42*

Earthquake 1979 (M6.6).

• power plant Drop IV Hydro. Plant Two-unit hydroelectric 0.30 Humboldt, CA .Humboldt Bay Power Two gas-fired units 0.25 Earthquake1 980 (M7.0) ",Plant.. one nuclear unit Ground acceleration measured by an Instrument at the site

-Average of two horizontal components A-3

HVA C Duct System EarthquakeExperience Data Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak Earthquake Magnitude Facility Type of Facility *Ground Acceleration (g)**

Coalinga, CA. . Main Oil Pumping Plant Pumping station feeding oil 050 Earthquake 1983 (M6.7) pipeline from Coalinga area Union Oil Butane Plant Petrochemical facility to 0.60 extract butane and propane from well waste gas Shell Water Treatment Petrochemical facility to 0.60 Plant demineralize water prior to steam injection into oil wells*

Coalinga Water Potable Water purification 0.52 Treatment Plant facility Coalinga Substation Electrical substation No. 2 Shell Tank Farm No. 29 Oil storage .0.38 Pleasant Valley. Pumping station to supply 0.56*

Pumping Plant water from the San Luis Canal to the Coalinga Canal San Luis Canal Agricultural pumping stations 0.20-0.60 Pumping Stations (29) taking water from the San Luis Canal Gates Substation Large electrical substation 0.25 Kettleman Compressor Natural gas pipeline booster 0.20 Station station Morgan Hill, CA United Tech Chemical Large research facility for *0.50 Earthquake 1984 (M6.2) Plant missile systems development IBM/SantaTeresa.,. Large computer facility for 0.37*

Facility .. . software development San Martin Winery Winery," .0.30

-Wiltron Electronics Plant Electronics manufaciuring 0.35 facility ,-

~Metcalf Suibstaition .'Large electrical substation . 0.40

• Ground acceleration measured by an Instrument at the site Average of two horizontal components A-4

HVA C Duct System Earthquake Experience Data Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Morgan Hill, CA' Evergreen Community Large college complex with 0.20 Earthquake 1984 (M6.2) College self-contained HVAC power

. (cont'd) plant Mirassou Winery Winery 0.20 Chile Earthquake 1985 .Bata Shoe Factory Four-building factory 0.64 (M7.8) and tannery San Isidro Substation *Electrical substation. 0.58*

Llolleo Water Pumping Water pumping station 0.78 Plant Oil/acetate/acid storage Terquim Tank Farm 0.55 tank farm Vicuna Hospital Four-story hospital 0.55 Rapel Hydroelectric Five-unit hydroelectric plant 0.40*

• Plant-San Sebastion Electrical substation 0.35 Substation Concon Petroleum Petrochemical facility 0.30 Refinery producing fuel oil, asphalt, gasoline and other petroleum products.

Oxiquim Chemical Plant Chemical facility producing 0.30 various chemicals, including feed stock for paint Ingredients Concon Water Pumping,. Water pumping station 0.30 Rena Powr Plant*:, Two-unit coal fired 0.30

- -, power plant -

LauaVerde Power- Two-unit coal-fie 0.30 Plant -power plant LasVetaasCoppe Cpper refinery/fou ndr 0.22 Refinery ,power pant

• Ground acceleration measured by an Instrument at the sit& ..

• • Average of two horizontal components.':,-.!

A-5

HVAC Duct System EarthquakeExperience Data Table A-i Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Chile Earthquake 1985 Las Ventanas Power Two-unit coal-fired peaking 0.25 I

1.Ivii.o) tuuIII U) Plant plant San Cristobal Electrical substation 0.25 Substation Las Condes Hospital Four-story hospital 0.20 Mexico Earthquake La Villita Power Plant Four-unit hydroelectric plant 0.14 1985 (M8.1)

SICARTSA'Steel Mill Large modern steel mill 0.25-0.50 Fertimex Fertilizer Plant Fertilizer plant 0.25-0.50 Adak, Alaska Adak Naval Base Diesel-electric power plants, 0.25 Earthquake 1986 (M7.5) electrical substations,.

sewage lift stations, water treatment plant, steam plants North Palm Springs, CA Devers"Substation Large electrical distribution 0.85*

Earthquake 1986 (M6.0)

Chalfant Valley, CA Control Gorge Two-unit hydroelectric plant 0.25 Earthquake 1986 (M6.0) Hydro Plant San Salvador Soyopango Substation Electrical substation 0.50 Earthquake 1986 (M5.4)

San Antonio Substation Electrical substation 0.40 Cerro Prieto, Mexico Power Plant 1 Geothermal power plant 0.20-0.30.

• Earthquake 1987 (M5.4)

Power Plant 2 Geothermal power plant 0.20-0.30 Bay of Plenty, New Edgecumbe Substation 230/115kV substation 0.5-1.0 Zealand Earthquake' 197(M6.25)

Ground acceleration measured by an Instrument at the site Average of two horizontal compon ents ý',.-

A-6

HVA C Duct System Earthquake Experience Data Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

• Estimated Peak Earthquake Magnitude Facility Type of Facility . Ground Acceleration (g)**

Bay of Plenty, New New Zealand Distillery Liquor distillery 0.50-1.0

-Zealand Earthquake 1987 (cont'd) (M6.25)

Bay Milk Dairy Products 0.50 Caxton Paper Mill Paper and pulp mill 0.40-0.55 Kawerau Substation 230/115kV substation 0.40-0.55 Whakatane Board Mill Paper mill producing 0.25 cardboard Matahina Dam Two-unit hydro-electric plant 0.26*

Whittier, CA Earthquake Olinda Substation Electrical substation 0.65*

1987 (M5.9)

SCE Central Dispatch Data Processing Center 0.56*

Headquarters SCE Headquarters Large office complex 0.42*

California Federal Bank Data processing facility 0.40 Facility Ticor Facility Data processing facility 0.40 Mesa Substation* Electrical substation 0.35 Sanwa Bank Facility Data processing facility 0.40 Alhambra Telephone Three-story concrete frame 0.40 Station building Rosemead Telephone Two-story steel-frame 0.40 Station building Central Telep'hone Three steel-frame high-rise 0.15 Station buildings Wells Fargo Bank Data processing facility 0.30 Facility Center Substation Electrical substation 0.26*

Lighthype Substation Electrical Substation 0.30 Ground acceleration measured by an Instrument at the site Average of two horizontal components A-7

HVA C Duct System EarthquakeExperience Data Table A-i Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

,Whittier, CA Earthquake Del Amo Substation Electricai Substation 0.20 1987 (M5.9) (cont'd)

Pasadena Power Plant Five-unit gas-fired 0.20

.power plant Glendale Power Plant five-unit gas-fired 0.25 power plant Commerce Refuse-to- One-unit gas-fired 0.40 Energy Plant power plant Puente Hills Landfill

• One-unit gas-fired 0.20 Gas and Energy power plant Recovery Plant Superstition Hills El Mesquite Lake 16 MW gas-fired power plant 0.20 Centro,'CA 1987 (M6.3) Resource Recovery Plant El Centro Steam Plant Four-unit gas-fired 0.25*

power plant Loma Prieta Earthquake Moss Landing Power Seven-unit gas-fired 0.34 1989 (M7.1) 'Plant. power plant Gilroy Energy Cogen One-unit combined gas 0.40 Plant. turbine and steam turbine plant Cardinal Cogen Plant One-unit combined gas 0.25 turbine and steam turbine plant

-iUCSC Cogen Plant One-unit diesel cogeneration .0.44

.Three-unit gas-fired

-Hunter's Point Plant. 0.15 power plant:.

Protreiro Plant One-unit gas fired plant 0.15 Meticalf Substation 500 kvsubstation- 0.30

" - San Mateo Substatioin 230 kV substation 0.20 Ground acceleration measured by an Instrument at the

• Average of two horizontal components A-8

HVAC Duct System EarthquakeExperience Data Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Loma Prieta Earthquake National Refractory Large brick & magnesia 0.30 1989 (M7.1) (cont'd) extraction plant Green Giant Foods Concrete tilt-up food 0.33 processing plant Watson Wastewater Sewage treatment plant 0.40 Treatment Santa Cruz Telephone Three-story concrete shear 0.50 Station wall switching station Watsonvilie Telephone Four-story concrete shear 0.33*

Station wall switching station Seagate Technology Concrete tilt-up 0.40 Watsonville manufacturing facility Santa Cruz Water Potable water purification 0.42 Treatment facility Soquel Water District One-story wood-frame office 0.50 Headquarters complex with small pumping station &storage tanks Lipton Foods Concrete tilt-up food 0.30 processing and packaging facility Lone Star Cement Large cement factory 0.25

• Watkins-Johnson One-, -two- and three-story 045

. Instruments."-",.,. concrete & steel-frame buildings for light manufacturing.

Riconada Water Potable water processing 0.30 Treatment Plant facility .

IBM/Santa Teresa Steel-frame high-rise 0.20 Facility. - complex for software development Ground acceleration measured by an Instrument at the site Average of two horizontal components A-9

HVAC Duct System EarthquakeExperienceData Table A-1 Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak

.Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Loma Prieta Earthquake EPRI Headquarters Two-and three-story 0.25 1989 (M7.1) (cont'd) concrete-frame office San Martin Winery Winery 0.30 Central Luzon Baguio Telephone Telephone switching station Phillipines Earthquake 1990 (M7.7)

Cabanatuan Substation 230 kV substation La Trinidad Substation 230 kV substation San Manuel Substation 230 kV substation Moog Manufacturing Manufacturing plant Plant Valle de Estrella, Costa Bomba Water Water treatment plant Rica Earthquake 1991 Treatment Plant (M7.4)

Cachi Dam 1,000 MW hydroelectric plant 0.12*

Changuinola Power Diesel power plant Plant Limon Telephone Telephone switching station Momn Power Plant 140 MW thermoelectric power plant RECOPE Refinery Oil refinery Sierra Madre, California Pasadena Power.Plant. Five-unit gas-fired power 0.20 Earthquake 1991 (M5.8) plant

- Goodrich Substation 230 kV substation' 0.30 Cape MendocinoV, PALCO Co-generation :Two-unit power plant,' 0.47 California Earthquake :Plant 1992 (M7.0)

• Ground acceleration measured by an Instrument at the site,- --.

Average of two horizontal components'.

A-1O

HVA C Duct System Earthquake Experience Data Table A-1

.Summary of Sites Reviewed in Compiling the Seismic Experience Database (Continued)

Estimated Peak

• Earthquake Magnitude Facility Type of Facility Ground Acceleration (g)**

Cape Mendocino, Humboldt Bay Power Two gas-fired units,. 0.24 California Earthquake Plant one nuclear unit 1992 (M7.0) (cont'd)

Centerville Beach Naval facility 0.40*

Station Landers and Big Bear, Cool Water Generation Four-unit power plant, 0.36*

California Earthquake Plant two gas/oil-fired and 1992 (M 7.6) two combined cycle units

..Mitsubishi Cement Plant Cement plant LUZ Projects Solar electric generating 0.35 station Northridge, California AES Placerita Two-unit electric gas turbine 0.60 Earthquake 1994 Cogeneration Plant generators (80MW), two heat (M 6.7) recovery steam generators and one 20 MW steam turbine generator ARCO Placerita Two-unit electric gas turbine 0.60 Cogeneration Plant generators and two heat recovery steam generators Pitchess Cogeneration 21 MW electric gas turbine 0.50 Plant and heat recovery steam.

generator Olive View 6MW power generation 0.72 Cogeneration Plant and heat recovery system Valley Steam Plant. '.fFo6ur-unit gas-fired power

->:i  :,0.40 plant' -

Burbank Power Plant -Six-unit'ga's-iired power plant' 0.30:

Glendale Power Plant Five-unit gas-fired power 0.25

-

• lant (148MW),

Ground acceleration measured by an Instrument at the site Average of two horizontal components>:-::

A-II

HVAC Duct System EarthquakeExperience Data Standard procedures used in surveying database facilities fou0rson collecting all information on damage or adverse effects of any kind caused by the earthquake 'Seismic damage to well-engineered facilities is normally limited to only a few items except at sites that experience very high seismic motion, that is, in excess of 0.50g Peak Ground Acceleration (PGA), or greater than thirty seconds of strong motion.;

Information on damaged and undamaged ductwork consists of photographs, measurements made

• .-:atthe site, visual observations, qualitative assessments of details and workmanship, and information supplied by personnel at the individual sites. This'information includes typical.

assemblies, unusual details or systems, and supports that appear to.be especially weak and prone to damage or failure.

An extensive search of the seismic experience database revealed thirty-nine sites in fourteen.

different earthquakes where ducting experienced PGAs of at least 0.25g. Eighteen of the thirty-nine experienced 0.40g or greater. The database sites represented a wide variety of duct sizes, shapes, configurations and suppiorttypes. Round and rectangular ducts were found at seventeen and thirty-five sites, respectiVely, with sizes ranging from six to seenty-two inches. The above data have been compiled and summarized according to database site, duct construction type and size, support type, building type, and noted damage. This information is shown in Table A-2.

The large number of duct systems that have survived earthquakes indicates the inherent ruggedness of these systems.'The light gauge. sheet metal ducts'were constructed with pocket locks, companion angles, andriveted-coinnections. In many cases the ducting had no stiffener angles and still survived the strong motion. Generally, the database HVAC ducts were supported with either' rod hangers or long sheet metal straps; however, there were also instances of frame-.

mounted ducts. Some HVAC ducts were hlung with rope, cables; or wire. Rod hafiger supports .

were typically trapezes which'were attached to concrete ceilings with expansion anchors, or.

either clamped or threaded and tapped into overhead steel structures. Sheet metal strap supports:'.

were usually spot welded to the duct sides and attached to overhead ceilingswith' expansion anchors. Figures A-I through A'7 illustrate some of the typical database duct configurations and supports that have suývived past strong-motion earthquakes.

It is important to note that nearly all of the HVAC duct installations in the database facilities.

were-designed and installed without specific consideration of seismiceloads.Also some facilities

-,were up to forty years old 'at the time of their.earthquakes. In addition to the effectsoof age, the initial instllation and any subSequent miodifications to database ductsSand theirSupports.

included all of the noimal oversights and deieciencies of industrial construction.

Ductwork ruggedness was demonstrated inmo instances, t ortmre atrbue -- a-,-"tiiincsbut'the're were some cases in which'.

one r reattu Iearthquake, ed to seismic damageA summary; organized by of the-configurati6nsand structural characteristics which cohtributed to the dafiiage is given below.

A-12

HVAC DuCt System Earthquake Experience.Data Table A-2 HVAC Duct Seismic Experifenice Database

" Duct Type _.,".,_ Support Type Damage .. Building Type

' D I:e;Rect Siz Rod Canti- F Concrete BlockI Framed Site Rnd uct e Strap Frame Falng Dented Tilt-Up SaFrame PGAangle . ..... ug Lever Shearwall Shearwall ADAK 0.25 X ', 1010 _ _....._*_" X ADAK 0.25 X XLX 100 X X '___X .,___

ADAK 0.25 X' .12X12 . X " _" . _,._. X ADAK 0.25 X-, 12X12' "X"__ ." _,_" X ADAK, 0.25 . X . "'_"-_"_,'X ADAK ': 0.25 X 12X12, ____ X .. ___ "'.._"._,

ADAK' 0.25 .,X '." 240 'r_"__ "___"

_____ X BATASHOE 064 " "!":::4. 0.6 XX2 BATA SHOE,. ..... . . . .. ,

FA CTO RY FACTO RY * . 0 64, . . . .. .,.X : : "3 24I : " ' ": Z_ _ _ _ __

• _ __ _ _ _ _ _ _ _ _ _ _ _ _ X ._ _

BAY MILK,. 0.50 - XI ý 24 X 24., . ._ X BAY MILK 0.5I X VERT... CABLE

_ _._-RUN BURBANK., 0.30 -X . 15"0 _.__,..____1_.X BURBANK.";: 0.30 X ., -LARGE . NV."

BURBANK 030 'ARGE" X' POW ER _._'. ,_'_,"._,_ __,R BURBANK- 030 POWER ,X0 3RG BURBANK., 03 k LARGEr POWER ....- 0.____ _._ ____" '-_"

CALFED 0.40 . X 60X48 ,"._ X CALFED 10.40 . '- X '30X8" '_" 'X X "_"_*

CALFED 0.40 _'-. X-. 30X8 _____ X .8 "XX CALFED . 0.40 X 30X8, X ____", X Legend:: NV Not Visible"': FR.ý Framed-CSW Concrete Shear Wall BR' Braced RC Reinforced Concrete NA Not Applicable A-13

HVA C Duct System Earthquake Exkperience Data Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct Type _:__ Support Type Damage Building Type Ron Rect" Se Rod Cantl- Concrete Block/ Framed Site,, PGA Round angle iDuct Size Hung Strap Lever Frame Faling Dented Tilt-uP Frame Shearwall CAXTON 0.40 ..... X 16 X 16 NV NV NV NV CAXTON, ,;,, 0.40 _,._ , X 18X12 NV NV NV NV -_* _, FRI/CBW CAXTON', 0.40 X _'___:"r " , 800 ROPE . " -"..... " v _ FRICBW CAXTON 0.40 "_:;; NV" 8'0 NV- NV NV NV ___ " __ 'FR/CBW CAXTON. 0.40 " _ X'!, NV X " ..

__._' "___ __"_."_... ... FRICBW CAXTON'- .4 'X . FLEX DUCT __ ____ __ __

CAXTON. 0.40 ""_. ,X 18X18 X "._" . ._- ,, .,_ _ __ _,_

CAXTON: 0.40 Xz .", . 18!0 . . ___ _ _" _.- SPLIT .. ... _FR/CBW .:

CAXTON:; 0.40 X _____ 18"0 __.... _'. 'SPLIT FRJCBW CAXTON 0.40: X X' .* BEAM FR/cIBW

________ ______ _____ CLAMP__ _ _ _ ___ _ __ _ _ _ _

CAXTON" 0.40 X "'i8"0 .... BEAM.

......... __"____:=. __...

. CLAMP _ __ _ ___ _CBW__ _ __ _ __F__ _ _ _

CAXTON: 0.40 X" 8 BEAM8' FRCBW CAXTON 6.40 X 18'0 B FR-CBW CLAMP CAXTON  ; 0.40 X* 18 X18 X __FR/CBW CAXTON-,' 0. X, 18'X8 X PROPPED " RC CAXTON --, 0'0.40 1 , X 30X30 _._ __*_X CAXTON,,,,.,,, 0.40 X_:_ 18 X 18, ._" PROPP __ED x CAXTON,:r 0.40 - X 18 X 18 .... PROPPED __ X CAXTON 0.40 .. X NV '_._"_. _X CAXTON 0.40 X NV CAXTON, 0.40 X 18 X18i X" X CAXTON 0.40 X X CAXTON 0.40 ' .X NV NV CSW*"

CAXTON 0.40 _-," X . X CS'._W ._

,CAXTON 0.40  : .120.

0).X X CAXTON 7 0.40 X NV ___ ___ __ X CAXTON.;' 0.40 X i, " 1 NV I" Legend: NV Not Visible FR Framed CSW Concrete Shear Wall BR Braced RC Reinforced Concrete- NA Not Applicable A-14

HVA C Duct System Earthquake Experience Data Table A-2D HVAC Duct SeimicExperience Database (Continued)

Duct Type' j  : . Support Type Damage Building Type site *":* P6A "i'oUnd "Rze-* :: '- Til-u FrFrame SiteL "' ;A Roun Re l S Rod Cantl- Frame Falling Dented Concrete BlockF Framed angle , Hung Lever Shearwall Shearwall COMMERCE, 0.40 X I 24'0... . X " _, __ x COMMERCE 0.40 X: X 24",12X12 ,,_ X .,_X

, ". , .I

-..,, *:-*X,,, 1 124

........ *:,12" 2 4/..!i"0 1:, " . .. ' " "";  : *'  : ' *. L *, .: , .

• i COMMER*C'E. 10.40  :"X.I' - . .. .. . . "

COMMERCE 0.40 r - 200 X .. . ___ _'_.. '_". . _ ... BRA ED COMMERCE' 0.40 X. _ _ 20"'0 X. _.. _ ". __ _ ____...'BRACED COMMERCE. 0.40 16024"0 OX X, "____ _.._'_ _ BRACED COMMERCE, 0.40 ___ _ X 60X60 X "". " . .. 1 X ".._"X COMMERCE- 0.40 X X . X 24"030X30 ____ _X... _' X___

COMMERCE, 0.40 X. '. 24"0 X,__,'_'___X _"_"_" X CONCON 3 x 18X 2,i PROPPED X PETROLEUM r_....

DEVERS'..., 0.85 x  :. NV.. X DEVERS ".*.; 0.85 "'"V "

DROP IV. 0.3 X 18" X 24" LIGHT CSW DROPiV . 0.30 X 24"X48 LIGHT CSW DROP IVi: 0.30 . X VARIOUS LIGHT CSW.

ELCENTRO .25 3XX36 X X x EL CENTRO

.42.,

.25X x x EL CENTRO'.:.42 Xi 60'X24' X X

.42. '24X24 ANGLE EL ,CENTRO .25

• , ... .. _ LEGS EL CENTRO ;42, 24X X ANGLE

.25

____. '__,424 4_- _ LEGS ____ ..

ELCENTRO .42,

.25 , " X 36X72 X X

EL CENTRO .42

,_.25

x. 36X2"__,

x x I_-

Legend: NV Not VIsMl I FR .Framed CSW Concrete Shear Wat BR Braced Reinforced Concrete NA Not Applicable A-15

HVA C DuctSystem EarthquakeExperience Data Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct Type _._-___Dctpe_ Support Type

• Damage Building Type P:, .n Reite Duct Size RatStrap Cant- Frame Falling Dented Concrete Tilt-Up Frame Framed gae gHung eLever, Shearwall Shearwall ELCENTRO

.___'__, "*..25 ': X V. 42*X42" m.,_ "____ __ _ _ __ _ _ _ _ _ __ X _ __ _ _ _ __ _ _ _

ELCENTRO .42 X: VARIES ANGLE.

.25 ____ RI___S ___ __ LEGS ___ _ _ _ ___ ___

ELCENTR..2 EdE . R L, ,L o.2 ` X,: ' VARIES ANGLE

______ .25 __ _ _LEGS'____ ___ _____

ECETOý 42.1 x x ELCNr .... 42, , ' . X N60GX80 EL CENTRO.

_ *25

-42, X .48VX 30 X ELCENTRO .42.X

.. 25 X_"_.. . "_

36 X98..

X EL CENTRO .42, ' " . 1 - X

________ -. 25-X N EL CENTRO .42,g X i -. X :83.25 X EL CENTRO .42, X EL CENTRP 6 . , *  :'  ::i~i :.,X EL CENTRO -25

.42, , {:I .,~ N NV NVX I POSTS".;i

POS S. ...._ :.25 ,_ __,.

ELCENTRO .42,x20X

__________ E42.,

.25 EL CENTRO . X , __

20 0" _________ _ ____

EL.NR 42,1 " '" ýX '2 XX EL CENTRO. .42,

.2, X, . I 6.". X . X P ELCENTRO .25 '-X 24*X60* X ELCENTRO .42. I X r, T 24X60 X6 * ..... X

.. .25 Legend: NV Not Visible,. . . FR Framed csw Concrete Shear Wall BR Braced RC Reinforced Concrete NA. Not Applicable A-16

HVAC Duct System EarthquakeExperience Data Table A-2 HVAC Duct Seismic Experience Database (Continued)

DuctType"{, Support Type Damage _ Building Type "Ret- " Rod Framed Site PGA Round Re Duct Size Strap Canti- Frame Falling Dented ConcreteTilt-Up Frame an.

'e , Hung Lever Shearwali TFShearwall EL CENTRO' .42,25 X

.. ... ., ,XX .X X ELCENTRO:' 42, X 48X24" ,

ELCENTRO .X'24*

"X' 4.425 ELCENTRO .42, . "x . VA.: x

.. 25 ,.. __." '4_8/_"" ___':

NA.___ _-_

ELCENTRO .42,. VAR. X' NA

ý.25 _____48X48 ____ _________

, ELCN-.425N

.25 .__ - ___ ___ __ ___ _ ___ NA

______.__.__,____, NA NA NA EL CENRO.42,

.25 X NA NA NA NA N

ELCENTRO

• ...25  ; ' :_... ., X.. . 48....X 24'i

___, _... _ _ __ _ __ X EL CENTRO .42,

...25 , ' x. . .

x x ELCENTRO. .42, , , X NV EL'NRO 42,.. .

ELCENTRO- .42 , X " I.LA:RGE ' X . .

EL CENTRO ,' .42,.E EL CENTRO . .. '.42,

.25 _ __ , . X .. 3

........ 2 __ "____.. _ RPE ELCN 42

'25XPRPE X 36XNV1201 PROPPED

.25' EL CENTRO .42,P RO

.25

'.25 X, . . 82 . . 1 ELCENTRO"

.42,1

'..25.

~:

Xl'" . '48 X x ELCENTRO. .42, - X. 24 1X24 X STEAM . .:

'25 , 24. .

Legend: NV Not Visible. " FR Framed csw Concrete Shear Wae BR Braced RC Reinforced Concrete NA Not Applicable A-17

HVAC Duct System EarthquakeEiperienceData Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct.Type.

...... ...... , . Support Type Damage Building Type

• ug strap Frame Flig Dented Tilt-Up Frame.

• ite ,.

Site Rou"d"Rect. -

GA'A"Rond.

PGA 'uItSie

)uct Size Rod Canti- Concrete BlockI Frame Framed angle Hung Lever F aShearwall " Shearwall____

EL CENTR1O .42.,..,*  :,.X.:: 24X 30'. .. ______

ELCENTRO: .42, xX .. 3___... X

".25 .

STEAM :-

X8 ELCENTRO .! .42, STEAM -'.25 "'24. X x FERTI'IMEX 0.25- : " X '

FERTIMEX . .X 24X* "X- X x x

FERTIME'X 0.25 X 12Xi6 I

, .. . 0.5 -,__..: _ . , .. ,.___ ,,, ... __

X 5,- 24 X24 30* X

• X FERtXMX X FERTIMEX 0.25- .. 4X, x 4 X 0.5 - . ._"_X4 X  ::iX ' '

FERTIMEX 0.25-*x ,60X 30,x X

FERTIMEX 0.25- ' " ýX 12X18 X FETME0.5

• 2 ;X, ~30X1 b:X X 2*_

X _ __x_

x FERTIMEX 0.25-x

_______ 0.5  : _

X _16' _ _

FERTIMEX. 0.25- X' NA NA - NA NA X 72X72 x FERTIMEX

'. , 52_' ,x... . X FERTIMEX

_ _ __ _ _ 0 .5.24 0.25 . X X2 3 0_ X__1 _8_

X X_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

FERTIMEX. 0.25- .'NV. NV NV N N X FERTIMEX 0.5 24x24 NV NV Legend: NV I NotVisOibe * . FR Framed; CSw

• Concrete ShearWatt Ban Braced " ,

RC Reinforced Concrete NA Not Apptiae.

A-I8

HVAC Duct System'EarthquakeExperience Data Table A-2 HVAC Duct Seismic Exdperience Database (Continued)

Duct Type. _ Support Type I, Damage '_ Building Type Site PGA Round Rect-, Duct Size Rod CnrtBlc/Framed Strap Canti- Frame Falling Dented Concrete Bloc iFrame angle' Hung Lever Shearwall Shearwall G LROY 040  : X" 24x 48 x NA NA NA NA COGEN ____.u______ ___......_....._...*

GILROY 00 . X 18 X 12 _NONE NONE NONE, NONE . CSW COGEN . .. _ .. _ _.,_ _ "_- __ __"_'

GLENDALE 0.30 X- ,LARGE:

POW ER ,', .. . .. . . . ._. _ _.. . _ _ _ _

GLENDALE 0.30 X'! LARGE POWER:" 0.30_ __ __"___ ___"_ _

GLENDALE 030 X'. LARGE X GL N R,*

ALpw 03 ' . .M.: X :':..: . LARGE',. :**i::: :::

POWER.. .....

_____ ___X GLENDALE POWER 0.3o X, - LARGE GLENDALE 0 0 X 'ARG' ,

POWER ;, "_-_03__

GLENDALE . 0.3" X- 3 V -36 ___.___,__..

POWER _. _____ _"____ _____ ___,_ ______X_"___ __24_X_4______

GLENDALE.

GLNAE0.30 . X~ X LARGE 24X2 N.

POWER~ _ _ _ _ __ _ _ _ _ _ _ _

GLNDALE POWER I 030.

03. ' . X.,:: 38" 18u '

.::NV X' GLENDALE' - NV, NV NV NV POWER .0X 1X GLENDALE;00.30 GLENDALE POWER,_

1X_ _ _

N _

PR_OPP,'. _ X GLENDALE r IV W GLENDALE POWER., .0'i

0.130. .

::i'X ` ." 18: XV

6. I' 'X " NVV GLENDALE ""- N .

.POWER - 'i, P0.30 X -25"0 .

Legend: NV. Not Vsie"le

• FR Framed csw Concrete Shear Waltl BR Braced PC Reinforced Concrete

• NA Not Applicable A-19

HVAC Duct System Earthquake'Exp-erienceData Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct Type , Support Type Damage Building Type

[: '.. ":; Rect-r . .. r '""Rod Canti- C nrt Boc/* Framed Site, PGA Round n ! Duct Size PGA.dangle R Hung Strap Lever Lever Frame Falling Dented Concrete Shearwall Block/ Tilt-Up Frame Shearwall GLENDALE 030 X 18X30 _

PROPPED POWER,, ,,:_0.30 .  ;,*,"__ _*

GLENDALE ! X POWER _,_ _ _ _ -"

• _"

GLENDALE 0.30 . . ..` x POWER ___ ___ ___ ______________

GLENDALE POWER~~:. 01 POW GLENDALE R[. , 0.30  ::.' ' "" X

• X::,  : ' 30 24X18'X 18*:" -X""

POWER 3 -.

GLENDALEpo E . 0.30 ,_".:..".:X':. 18"

.X"v-]. ,i GLENDALE.

POWER _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _

GLENDALE X NV .

POWER

• BAY HUMBOLDT .30 0.;"

.25, ' .36.0. X,.. w .... , ": x x x*"

HUMBOLDT',

BA _. .25',

-._ :" _ . ..30' . .... ,"_.. . .. ._

NV " _ _ _ _ _'_ _ __*_ __][

BAY .30 ....... .... .. .. .... _.._._ _.....

HUMBOLDT ,25.

HUMBOLDT .25, NV BAY .30 ,____'_'__,_.._______ ______,

BAY..-- !* ,0 : 7: : ..

HUMBOLDT% .25. *,

HUMBOLDT BAY .30__________ .2RE, .X HUMBOLDT .2,5,*X HUMBOLDT, .25; X B AY .30 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Legend: NV -  ; - NotVlsIbae"'-,, ' ý. "-,

• FR Framed CSW Concrete ShearWanI BR Braced RC Reinforced Concrete NA Not Ap~licae

: . q, ;

A-20

HVA C Duct System Ea'rthquakeExperience Data Table A-2 HVAC Duct Seismic Experience Database (Continued) r

____ , Duct Type. 7': Support Type Damage Building Type Site Sie PA Rou PGA oud ange m uctSz Duct. *i Rod ung Strap Lever-Canti-r Frame Falling Dented Concrete BlockI sheaaI Tilt-Up Frame Framed HUMBOLDT 3.25, BAY 3 18'0,X HUMBOLDT- .25, X BA..Y.-:L . 30 X'. VARIES ....

• ______ _______,__..........

HUMBOLDT,..25, BAY.0 _ _

x _

NV '.7 X ' . CSW HUMBOLDT. .25; x 300 X BAY .30: ... . .'"

.303-*0 HUMBOLDT' HUBLT .25, 2, x . xX, iV x*

BAY __ ,_ _ ____.30 HUMBOLDT, .25, BAYi BAY : :_ ..

• :3M"' .. :* . :. . .. __

HUMBOLDT-: .25, ... 1.. ' ..

HUMBOLDT: .25, " '

BAY'- .30 1_ 0: o . _ x _ __.

IBM SANTA 037.,;;' .... 24........

TERESA ______ __u___,_. .... ________, ___.__"_X..__._ _2X._12._ _._

IBM SANTA:' 0'37 ' ;W"X "24X1. X -X TERESA ____._ ___,___....___.______ __,____ ______24_X___

K E T TLE M AN 0 .2 0 i : : X : . . ' : ' .:.: . 16 "0 : ' .

LA.VILLITA rd 0.14 ': '"!' X  :..NV ... .'"X LA V ILLITA *' .'14 i:: ' ' :' X '18 X 18 ' X X HUMBOLDT .25 1 "O'ý BAY . ' ,. .3 , , ,;., . ,

HUMBOLDT '.25, ,'.24,, 24.8 18.. x* X .

BAY . 0 - - _. * ,

HUMBOLDT . *.25,  :.  ? , -! 24 X 18 X" BAY°.

HUMBOLDT .:

30: ,,_ ___.__ _ _.__ __._ _ __ __

.25 BAY .30 X . L E Legend: NV

• Not Visible FR Framed CSW Concrete Shear Wail BR Braced RC Reinforced Concrete NA Not Applicable A-21

HVA C Duct System EarthquakeExperience Data Table A-2 HVAC Duct Seismic'Experience Database (Continued)

.. r'DuctType .Support Type Damage --

_..___,._*_ Building Type Site PGA Round angle, Rect, Duct Size Rod . Canti- Concrete Block/ Tilt-Up Frame Framed Hung Lever Frame Failing Dented reBlo.Shearwall HUMBOLDT .25, . X x BAY. .30 - - - _.._ _._._

IBM SANTA 037 X 24XVARIES X

.. 0..7 . . . . 2_XVARIES _

TERESA "__..

IBM SANTA, 03 X ;24X12 X X TERESA *... .. -

IBMSANTA 037, . X 24X12 X X TE R ES A , ._._..24 12 ... . " ___

LAS . 0.22 0..... .  :' 36X36 X. x LAS- .18.X18. CSW VENTANAS':-ý VENTANAS 7. .. . X- . X LAS VENTANAS 0.22 X- 20 0 . X X COP._ _ _ __ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _

MEQUITE LAKE 0.207*-u;u* X 240,360 MEQUT 20.2 :;X, X 24'03204 MESQUITE- ,,

LAKE 0.20 . X 24X24 MESQUITE LAKE,' 0.20 X LARGE. x MESQUITE,,.

MESQUITE: .20 020 " Xm X: 2 V

'2"0"'x x NA NA. NA NA' LAKE LAKE . ,... - .. ..

. A... , ..

MESQUITE 0.20. .... X x.

LAKE" - . _L_ _ _ __:__

MESQUITE 0.20 X IV x x AKMESQUITE. 0.20  : X:x " 0..0isi, X*"X.

x" Legend: NV Not VISilOe FR Framed CSW Concrete Shear Wan BR Braced RC Reinforced Concrete NA Not A4picae A-22

HVAC Duct System EarthquakeExperience Data Table A-2: .

HVAC Duct Seismic Experience Database (Continued)

__ ... ____ " DuctType SupportType- Damage Building Type PGA Round Rect- Rod Canti- ConcreteBlock/ Framed Site PGA Round Duct Size Strap Lev Frame Falling Dented l Tilt-Up Frame angIe Hung er Shearwail Shearwall MT. UMANUM' 0.50 X.::,, ::,VARIES X X CSW.

MT. UMANUM 0.50 X 24X16 X X CSW MT. UMANUM" 0.50 X 1r 24X16 X . . CSW MT. UMANUM 0.50 . X .16X 12 X," % ,X CSW MT. UMANUM. 0.50 X :16X 18, :X X CSW PAC BELL-'

WATSON-. 0.33 ':X 12X6, CSW VILLE .. _ _ _ _ __ _ _ _...._""

MT. UMANUM 0.50 X' X 16X18. X X CSW MT. UMANUM' 0.50 ' X '.16 X 18 X X CSW MT. UMANUM 0.50 X 16"16 x CSW MT. UMANUM 0.50 X "'.16X.16 X CSW ORMOND' BEACH ' , 01. - I.' ---

PAC BELL . 0 .' , 2 ALHAMBRA 0.40 X . .....

PACBELL WATS..,.' ",:.,* 0.33 X 24X8 X o~. ... '-.. . '^ *. 24X-PAC BELL -" X CSW WATSONVILLE 0.33 X 12. 6 X PAC BELL WATSONVILLE ,.3 X 24 X 8 CSW PAC BELL WATSONVILLE 0 16 X.8....

Legend: NV NOt VisiW.O FR Framned cSW Concrete Shear Wall BR Braced RC Reinforced Concrete. NA Not Applicable A-23

HVAC Duct System EarihquakeExperience Data Table A-2 :. ,

HVAC Duct Seismic Experience Database (Continued)

IDuct Type __I____ ' ___...,Support Type " Damage Building Type dRet " Rod Cant- .Concrete Block" Framed Site, PGA Roundii Duct Size Strap Frame Failing Dented Cc Tilt-Up Frame

. angle Hung Lever, ShearwaUl Shearwall PAC BELL WATSONVILLE 03 X NV- .. sw PAC BELL 00..33 . *'

WATSONVILLE . ,___ .__":. , "."X.

PUENTE HILLS 0.20 " -*X' NV.. LEGS NA 'NA. NA NA RENCA . 0.30 X X X RINALDI 0.'50 X :NV:

SAN MARTIN , X NV X SAN MAR'riN: .3. ..... " NV....

.30, SANWA K: 0.40 K. . X- '30X18 ' NV NVW - NV X SANWA BANK 0.40 X X X X SCEMX '24X24 CSW ROSEMEAD. 0.4*2 SC' 0.42 X 3X15ý X X 0W SEAGATEI;:,* 0.410 *: X : i x

,::,i,*:

• ...:*,8"0 ii X , ,.. X " *. ... " "", ,

x SEAGATE"- 040 0:. '*i Xi""":?::*:

.. 12"016"0 l' Ix...X r. . .":X " " '"'

SEAGATE,-, 0.40 X 12"0 160 X X . X ,

SEAGATE", 0.40 'X , 12.16"0 . X SEAGATE:,!.; 0.40 X, 24"0 .. X SEAGATE.' 0.40 X',X 240 X X*

Legend: W 1, oNV t Visible FR Framed CSW , Concrete Shear Waln BR Braced RC Reinforced Concrete NA Not Applicable A-24

HVAC Duct System EarthquakeExperience Data Table A-2.

HVAC Duct Seismic Experience'Database (Contihued)

Duct Type Support Type Damage Building Type

"*'-  ;?" Rect- Rod Coti Cocrt Blc/ Framed Site PGA Round Duct Size Rod strap canti Frame Falling Dented Concrete Bock Tilt-Up Frame Framed angle Hung Lever Dt Shearwall T Shearwall SICARTSA 0.25 X NV. X X X SICARTSA 0.25 Xi NV' X X-SICARTSA 0.25 ' X 42 X 42 XX SICARTSA. 0.25 *ý' 12X212 NV NV NV NV X SICARTSA 0.25 ' X NVM ' X X SYLMAR .' 0.65 -X';7- 12X8 ,

SYLMAR ; 1'56' " BR 0.6 STEEL SYLMAR -- '.. 0.65.X 2

",BR

. .. .. _-_._ _.. _ __.. ST EEL SYLMAR ' 0.65 ,X X 28X18 "NV BR

__I__ . I. I 0.

. - I..____ 5.I _ _ BR STEEL '"

SYLMAR 7 0.65 X X 180 0, NV BR

__-._ _ _STEEL SYLMAR 0~i..065

+:: , ;'! ... X". 24X1RI8S .X:}V..

' STEELBR;... . .

SYLMAR 0.65 X VARIES.* NV STEEL.

sUA "I ,0.65I  :  :'X l2.X 'I BR, IYMA

" . "I. ... .""1 'VAR...ES. __ _ STEEL ": "

SYLMAR;:.'. 0.65"'" . SYLMR'

'6X1 "2X6 NVBR .65X BR:,

..... _ __ _ _ _ __ __ _ _ STEEL SYLMAR 0.65 3X1 4X N BR

__ -_ -STEEL SYLMAR'" 0.65 X ... . .*.._"BR_ BR.

....... _STEEL SYLMAR 0.65 X 20 X24 X BR

- ... ' STEEL Legend:

• NV Not VisibO . FR .Framed CSW Concrete Sheaf Walt BR Braced nC Reinforced Concrete NA Not Applicable A-25

HVA C Duct System Eai~hq.uake Experience Data Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct "Tye Support Type Damage ' Building Type

"*:ect Rod Cant:1' Concrete BIock/ I. p FramedSerwl Site-, PGA Round Duct Size SRod Canti- ame Falling Dented Ci.Fram Framed angle u Hung Lever Shearwall Shearwall SYLMAR 0,.65

• X VARIES BR SYLMAR 0.65,

':)!.,,. . X:'.:i".!.

(!.; 60 X 18 " -':X. : * " .'"

____________STEEL

. . ... "* "STEBR BR ..

SYLMAR' 065 :*  :.X 18X12: ,,

. ... ... . .. ... . "* , . . . " , 'S TEE L SYLMAR, 0'65 X' -30X 12 X BR SYLMAR 065 ... X  : ___________ _____STEEL BR

. . X. .. ., .. . ... ... STEEL SYLMAR &5 , .!. 20 Xa :':,. ,'x". .. .'.STEBR SYLMAR 065' 1X.2X1 X BR

_ __ _STEEL UCSC COGEN" 0.44 X;200. CABLES FR/CBW UNION OIL 0.60 X,, LARGE.,

UNION OIL-. 0.60 "'"X: LARGE. x x VALLEY STEAM .0.40.... , 36X36, VALLEY ' 040 ' X ARGE STEAM ____._ _...._._...__ ___.....__:___

VALLEY .

STEAM 0.40 X, 1, .

VALLEY 0.40 XX_'_

VLEY " 0.40 X NV SPRINGs X VALLEYAM 0.40 X NV sPRINGS X VALLEY 0.40 STEAM 0.40 . . 40 STEAM"", , .

VALLEY 0.40 X LARGE Legend: NV- Not Visiole . FR.* Framed CSW Concrete Shoear Wal BR Braced R0 Reinforced Concrete NA Not Appflcable A-26

HVAC Duct System Eadthquake Experience Data TableA.2:-2.!'1; ,...'

HVAC Duct Seismic Experience Database (Continued)

Duct Type _____:__ ___ Support Type Damage Buildin"g Type Site PGA Round Duct Size Rod ap Canti- Frame Falling Dented Concrete Block Tilt-Up F rame Framed angle Hung t Lever i Shearwall ' ' a Shearwall WATKIN uA5 X 30 X 30 UNISTRUT X WATKINS 0.45 ' 360,120 CABLES ...

J O HN S ON . .. ... *_.

WATKiNS-. W KIS 0.45 as X. ... . . UNISTRUT. *x .

J.OHNS,ON X

,."4 JOHNSON _r___ .__'UNST UT__ ____..

WA KNS5 0.:45 i*::i*: i:  ::::" 8X '8 ,:ý.:,.. UNISTRUT r X WATKINS JOHNSON . X N NV NV " x; NV WATKINS,, 04:._ ':':L: .: X.' N0X3 NiV-: "

WATKINSr WATKINS, 0 0

JOHNSON WATKINS: 5 . ... * ... 18X V.. UNISTRUT LINI.T ANC.."

T N ,-."..LW JOHNSON 04 '.":'ir.,'!x"?  : ,'"i V . " V " N V."

WAS :;-

JOHNSON _C.a0.48:!::r" 45 NVx WATKINSC 03 X 24'.UI' BELL" ;;.: .... -":"" 1__ ___X___",___NV _" _! .*x ' : ."_X_," ",___

WATS-PAC B E LH , -.:O' 0..

,-  ;.. X:'. , 1X:

. . . .. . . Nx V IV BELL ______: ______ _____-__ __-_,____,__. _;_____ :l "

WATS. PAC 033 ', XI X AX BELL .,.,.. ....-'33_ . .. _ _

Legend: NV csw NOtVsroAmeI. -,'. FR . Framed Concrete Shear Watl BR . Braced Re Reinforced Concrete NA .. Not Awpicabe A-27

HVA C Duct System EarthquakeExperience Data Table A-2 HVAC Duct Seismic Experience Database (Continued)

Duct Type Support Type Damage Building Type Site Rect- Rod Canti- Concrete Block/ Framed PGA Round. Duct Size Strap Frame Faiiing Dented Tilt.Up Frame angle Hung Lever Shearwall Shearwall WATS.

WASTE 0.40 X 30 X16 X X WATER _

WHAKATANE 0.25 X 18 X 10 X RC WHAKATANE 0.25 X X, RC WHAKATANE 0.25 X 24X 10 'X RC WHAKATANE 0.25 X 24X 12., X" RC WHAKATANE 0.25 X 20 X 12 X RC WHAKATANE 0.25 X 16" 0  :'NV' X WILTRON, 0.35 X 12"0' X X X X WILTRON., 0.35 X 120 X X X X WILTRON 0.35 X. 12" 0 X X X X WILTRON 0.351. X X 12"0,1, X X X X WILTRON 0.35 X 12"0 X "X X X A Legend: NV NotVisible FR Framed CSW Concrete Shear Wan BR Braced RC Reinforced Concrete NA Not Applicable A-28

HVA C Duct System EarthquakeExperienceData Figure A-1 Sylmar Converter Station,.1971 San Fernando Earthquake. Strap-Hung and Wall-Mounted Duct with Wall Penetrations.

FigureA-2 Glendale Power Plant, 1971 San Fernando Earthquake, Cantilever Bracket Supported Rectangular Duct A-29

t HVAC Duct System EarthquakeExperience Data 1i NIMMINIaht'.491:1

..-.Figure A-3 Steam Plant, 1979 Imperial Valley Earthqua e.,Trapeze Rod--Hung Recta'n'gular'-

Duct ivith Close Up'of the ap e ze Det il A-30

,a M HVA C Duct System EarthquakeExperience Data Figure A-4 Bay Milk Products, 1987 New Zealand Earthquake. Long Vertical Cantilever Supported by the Roof at One End and Guy Wires at the Other A-31

HVAC Duct System EarthquakeExperience Data Figure A-5 :s 3/4 -- i-Bi< --

C ifra FeIderal Bank iacility, .1987,Whitier Earthquake.Typical Strnp-Hung Rectangular "Duct with Vertical Cantilevers and Diffusers'. -

A-32

H *C Duct System Earthquake Experience Data Figure A-6 Watkins-Johnson Instrument Pla'nt, 1989 Loma Prieta Earthquake. Large, Insulated Round Duct with Branch Ducts'and Cable Supports

.Figure A-7'. ,-

P:-acific Bell Watsonville, 1989 Loma Prieta Earthquake Run of Trapeze Rod-hung Rectangular Duct A-33

HVAC DuctSystem EarthquakeExperience Data A.2.1.1 1983 Coalinga, California Earthquake The Coalinga, California earthquake occurred at about 4:43 P.M., local time, on May 2, 1983, and had a Richter magnitude of 6.7. It was centered near the town of Coalinga which is midway between San Francisco and Los Angeles. Coalinga is situated in a large oil field that includes numerous petrochemical and other industrial and power installations.

Gates Substationis located on the 500 kilovolt (kV) intertie that runs north to south through the Califomria Central Valley. The facility has two control buildings, several shops, and storage buildings.All of ihese structures are one-story structures of reinforced concrete block or precast concrete construction. All were designed to the'seismic standards of the concurrent Uniform Building Code, seismic zone IV, or more stringent requirements imposed by the operating facility.

Gates Substation is located about fourteen miles southeast of the main shock's epicenter, and about an equivalent distance south of the nearest strong motion record at Pleasant Valley Pumping Plant. Standard ground motion attenuation formulae indicate a PGA of approximately 0.25g for the site's distance from the epicenter.

During the earthquake, an HVAC diffuser fell from a suspended ceiling. The diffuser was slipped into place and supported from the ceiling, but was not attached to'the HVAC ducting (see Figure A-8).

• Figure'A-8z Gates Substation, 1983 Coalinga Earthquake. An HVAC Diffuser Fell from the Suspended Ceiling A-34

HVA C Duct System EarthquakeExperience Data A.2.1.2 1984 Morgan Hill, California Earthquake The Morgan Hill, California, earthquake occurred on the Calaveras fault at 1:15 P.M., PST on April 24, 1984. The Richter magnitude 6.2 earthquake was centered approximately ten miles due east of San Jose. Despite localized pockets of damage to residences and commercial facilities, the damage to structures was generally light.

i..Wiltron, located on Mast Street in Morgan Hill, manufactures microwave and communication

.equipment for telephone and'other companies. The facility is housed in a reinforced concrete tilt-up building which has aplywood diaphragm roof. Based upon the nearest recording instruments at Anderson Dam and the nineteen mile distance to the epicenter, the site experienced an estimated PGA of 0.35g.

'At the Wiltion Facility, a'four foot'long vertical cantilevered section of HVAC ductwork broke from its supporting header and fell to the floor (see Figure A-9). The round duct was constructed of riveted lap joints which failed uider thec antilever's inertial loads.

114i-2-64 Figure A 91' -

Wiltron Facility, 1984 Morgan Hill Earthiuake. A 4-Foot Long Vertical CantileverBr'oke;' "

from Its Supporting Header and Fell Another section of HVAC duct at the same facilitysplit a seam wh ere a branch line entered a awall penetration (see FigureA-0). Thedamaged section.was approximately ten inches in

'diameter, branchingoff of an estimated twenty inch diameier header'.The seam pulled apart near the will;,approximiatelyfour feet fromtlhe branch point. The branchfapparently was not flexible

• enough'to accommodate'the header motion,-and the se*amwas too 'weakto resist the imposed differential displacement.

A-35

HVAC DuctSystem Earthquake ExperienceData Figure A-10

• Wiltron Facility, 1984 Morgan Hill Earthquake. A Branch Line Tore at a Wall Penetration

• -Due to Flexible Header Motion A.2.1.3 1985 Mexico E6thquake The Richter magnitude 8.1 earthquake of September 19, 1985 was centered near a large industrial area at LazTaro Cardenas on the west coast of Mexico:The industrial area includes a large steel mill and a fertilizer plant, as well as several other manufacturing and service facilities.

The industrial area is served by two large hydroelectric plants located on the Rio Balsas. Both the power plants and the industrial facilities are relatively new, having been constructed primarily in the 1970s and 1980s...

The Fertimex facility is a'large fertilizer plant on an island at the mouth of the Rio Balsas.

Reconnaissance teams observed several sand boils and settlement as large as twelve inches on the island. The site's PGA is esir'mated at O.25g based upon the nearest ground motion records at Zacatula; however, thesection of theisland which supports Fertimex's Packaging Plant is thobight to have experienced at least o.50g.- -

HVAC ducting was damaged on the-second floor.of the packaging plant's switchgear building.-

The second floor slab is approximately fifteen feet aboveg rade. The tW0story concrete-frame structure-is about 120 feet long, fifty. feet wide, and has eccentric rigidity due to the asymmetric loc"ationof brick in-fill and partial concrete walls. The icenitricity created high torsional accelerations in somfe regions of the structure. In one of these areas, the lasi section in a long duct run jumiped off the final support.: The resulting cantileverfailed at:*' andjacent support

- (see Figure -A-I 1). The HVAC ductsection was of pockeIt lock orisiiicti6n and was not

• positively; attached to the riod hung trapeze Support. Had it beed attachedthe damage would

• likely have been -avoided. Also in the same area, one of the duct's rod supprrtspulled its" expansion anchor from the concrete ceiling. The concrete quality was questionable and the ribs on the non-drilling shell anchor's cone expander were flat rather than slanted.

A-36

HVAC Duct System Earthquake Experience Data

"~~ ~~~~

• '..r* 2:.*. ...........

,Figure A-li,_-

Fertimex Packaging Plant, 1985 Mexico Earthquake. A section of Duct Tore when the Duct Jumped off the Final Support in a Long Run A.2.1.4 1987 New Zealand Earthquake On March 2, 1987 at 1:43 P.M.,ý a Ri chter magnitude 6.2 earthquake'struck the eastern Bay of

. Plenty region of North Island,' New Zealand. The earthquake was preceded at 1:36 P.M. by a M5.2 foreshock and followed at 1:52 P.M. by a M5.2 aftersh6ck. The main event, centered about four miles northwest iof the small town of Edgecumbe, propagated along a previously unmapped fault that opened a large surface rupture and caused widespread soil failures. Strong.

ground motion also affected the nearby towns of Kawerau, Te Teko, and Whakatane. An average horizontal PGA of 0.26g was recorded approximately six miles from the rupture, and PGAs from 0.30gto 1.Og were estimated in the affected area.

th'e' Caxton Paper Mill is located on t'he outskirts of Kawrau, about five miles from 'the faulIt and along a line etendiig inthe direction of surface rupture.Based upon the ground motions recorded at the Matahina- Dam and a comparison of the Modified Mercalli ifitensities for the dam and the paper mill,* the mill's PGA is estimated to'be O.4Og.

The :facility's* paper machinebuildings (Nos. 2 and 3) are flexible high-bay steel frames and reportedly deflected excessively du'ing the earth'quak'e.Damaged HVAC duct was found in both buildings.'-.

At Paper.Machine Building No. 2, there were several instances of sheared ductwork joints; however nSoectionsfell to thie floor. The circular duct was mounted near the'ceiling and constructed of riveted lap joints (see Figure A-] 2).

A-37

HVAC Duct System EarthquakeExperience Data

'FigureAl12

-...Caxton Paper Mill;1987 New Zealand Earthquake. A long, Unrestrained Run Of Duct

' ,Constructedof Riveted Lap Joints'(Top) and a Taped Repair of a Sheared Joint (Bottom)

}Papei Machine Building No. 3 is taller aind ex1eienced more damage. he round ductwork was fastened by riveted laap joints and 'supported from; the roo ftruss with rod hangers and beam clamps. Large deflection of the ductwork pulled adjacent sectiois of ducting apart allowing a portionf topry itisielf away fr*ri the supports 'andfall to the operating floor. Inspection of the fallen ductwork noted heavy corrosion at the riveted joint.

A-38

.Hi C Duct System Earthquake Experience Data A.2.1.5 1987 Whittier, California Earthquake On Thursday, October 1, 1987, at 7:42 A.M., a Richter M5.9 earthquake occurred due east of Los Angeles near the city 6f Whittier, California. The shock caused damage over a large area o f. the Los Angeles Basin. The main shock was followed by numerous aftershocks, including a M5.5 aftershock at about 3:00 a.m. on Sunday, October 4, which further damaged structures already weakened by the initial shock.

The City of Commerce Refuse-to-Energy Plant is locatedapproximately Seven miles southwest of the epicentral area. The plant was constructed in 1985, and its buildings were designed according to the current Uniform Building Code for seismic zone IV. The 11i5.MW plant is housed.in a large steel-frAme structure, including'an enclosed high-bay refuse storage pit, with adjoining office complex; open turbine deck,'and open steel-frame boiler tower. The PGA is estimated as 0.40g, based upon the records at the Bulk Mail Center and a comparison of the Modified Mercalli Intensities at similar sites.*The Bulk Mail Center is less than a mile south of the plant and has similair soil conditions.

Damage to the Commerce Energy Plant was minimal but included an HVAC diffuser which fell in an office area. The diffuser was apparently not secured to the duct main run.

The main office of the:Southern California Edison (SCE) Headquarters is located within a mile of the epicefiter and has ground motion equipment located on site. The four-story concrete shear "wall structure endured a PGA of 0.42g and sustained the most significant structural damage of the three buildings'in the cormplex. An HVAC fan in this building dislodged from its spring isolators and displaced enough totear the flexible bellows coupling to the duct on its discharge side.

The Ticor Data Processing Center is a two-story concrete tilt-up building constructed around 1980. It is a somewhat complicated structure combining steel and reinforced concrete internal framing with a spanreiee second floor, a metal roof deck, and exterior concrete wall panels.

The building suffered substantial damage iJnicluding shear cracks in Wall panels, spalling and fracture of the second floor slab, separation of joints between wall panels and framing, and a torn expansion joint in the roof. .

Nonstructural damage was also extensive a*ndi included HVAC duct. Roof-mounted HVAC

..equipment at Ticor.Was severely damaged and the system was shut down. Most of the equipment was mounted'oh vibration isolators without lateral (seismic)restaint. TWO axial fans had

• shified off theirimounts, rupturing their duct attachments (see Figure A-] 3).".

A-39

HVAC Duct System EarthquakeExperience Data Figure A-13 Ticor Facility, 1987 Whittier Earthquake. A Flexible Bellows has Torn Due to the Motion of an Attached Fan on Vibration Isolation Mounts

.-The free-field record take adjacent to the SCE Headquarters is near enough to the Ticor facility

t. essentially be considered a site record. Both the Ticor and SCE sites are on soft'alluvial deposits laid down from the nearby Sanh Gabriel River.:- . -

. :The SCE free-field accelerogragh is likely representative of the e6t'fective fr*fieldground .

motion at Ticor. Although the peak acceleration exceeded 0.0g in b th horizonta directions, and the response spectra show'relatively broadband frequency content,:the motion was very short in duration,.with only ihreeto fivie ciiycles of significant amplitude.-'--:

--The.'Sanwa Data-Processing Center is housed in adjoining steel-frame concrete panel sided buildings of abou' 100,000 squar'e&feet.each, on four~stagge'red floorlevels. The center contains

.data processing equipmient mounted on raised llo6rsas well as orfice fa'ilities. The roof includes a penthouse for HVAC equipment.

A-40

HVAC Duct System Earthquake Experience Data The Sanwa facility is located in the Repito Hills, a shallowTormation of sedimentary rock that penetrates the surrounding alluvial valleys. The nearest record at Garvey Reservoir, with a peak horizontal acceleration of about 0.40g, is a reasonable representation of the effective free-field motion experienced by the site. The strong motion instrument is founded on compacted alluvium, less than a mile and a half from the Sanwa facility.

HVAC ducts in the space above the raised ceiling experienced movement and permanent distortion without excessive leakage, failure or loss of function. !In addition, a duct above the battery racks, approximately twelve inches by twelve inches, deformed but did not fall. The long run was supported at the ceiling by sheet metal straps and had no companion angles or stiffeners.

The duct deformed at the j4oints of an angled offset section which contained an HVAC register (see Figure A-14).

Figure A-14. ..

'Sanwa Data Processing Center,'1987.Whittier Earthqu'ake. A Duct a~bov'e the Battery Racks Deformed 'atthe Joints of an Angled Offset Section X2.1.6 1988 Alum Rock,-California Earthquake:*

The Alum -Rock'earthquake6 had a low.PGA (0..15g) and relatively mfinor djamag'e; however, there. was HVAC relate .ddamfag'e in the'third floor,mechaniica'l penithouse of the1 EastRigMal The daIm Iage -occu.rred when air handling units, mounted'ofi vibra ,tion isolationprn Ml.

Rsidgeu lateral supp.ort,-de'flec.ted an'd tore the attached flexible bellows' to -the "adjaicent ducting (see Fig'ure A-15):'.

A-41

HVAC Duct System Earthquake Experience Data.

ocurda 'gasgeto

").9 h a nra al nea Lma rita '.Pa rudsaiga Figure A-15 East he Ridge Mall, 1988 m copue dik Alum&Rock Ea\thquake. A Flexible Bellows Tore Due to the Motion' d

of Attached Air Handlers on Vibration*Isolation Mounts A.2.1.7 1989 Lorna Prieta Earthquake .. .

'At 5:04 P.M., Tuesday,'October -17, 1989,',a Ric'hter magnituide 7.1 earthquake struck "approxim~ately te~n mil es northeast of-.Sfiantaruz, -California. The'twenity second earthquake strong as 0.65g was recorded iin both't'hehorizontal and v'ertic'al dýiriections in thieep'i central -area.

The coputerdikrive manufacturing plant operated by Seagate Technology is housed -in aconcrete tilt4u building rmad of adjoining onead tosryecin;T e sitiloated

\approximatel tw mie orth-west of an infstrument in downtown, Watsonville.ýSoil cnitions in the vicinfity of Seagate are labieled "fl uvial facie6sK" a form of marintfe'tefira~ce -de'positis' hara istic o e Watlsonville area. Theth one A uilding where the strong-motion.

instruments are located isoembded in flood plai oitn unconsolidated snanAd sl The' Seagate site therefore appearsto'be on somewhat firmer soil., Based upon the observed cectswithineth building, -a-reasonableestimate of the peak horizo nitalkground shacceleration is 0.40g.

A-42

HVAC Duct System Earthquake ExperienceData The sections of Seagate's circular duct are lap jointed (without rivets or bolts) and hung from the ceiling with sheet metal straps. During the earthquake, a portion of the duct fell to the floor when a strap broke at the duct connection and the attached section pulled free of its joints (see Figure A-16).

,-1 intumn sites are. founde 66s-dmeyay.k Figure A-16 Seagate Technolog'y, 1989 Loma Prieta Earthquake. A Strap Support Broke and the

-Attached Dut Fell to the Floor l.s ,,

The Watkins-Johnson ct* i.r ~fiH Instrument ~

Plant is'an expansion of a small instrument assembly operation that was started in the 1950s. The site includes eight buildings of various 'Constructionand fvintage Oie built (see into the base17) -of-hillsides-wswithin a small val1ley.

Th'uct "+ieA-

• The nearest instruments are' At the Lick O-bservatory _(CDMG) and in the Earth Sciences Building on the University of California, Santa Cruz. Both instrumienft sites are just over five miles away

'from Watki ns-Johnson and 'each measured PG, s greater than 0.Ag Th9 CCcmu instumet stesre~uried n sdiieiiaryrock wereas the Watkins-Johnson plant is ia

.smal valley ~with alluvial deposits o'erlying "sedimen.t.aryIrock.;The site -conditIions at the plant 7are t'he'refore somewhat softer compared to those of the nearest, instrumients..Usinig the eod and acomparison of te Modifie'dMercalli IntensteteWtisJh~~ iePAi estimated as OA.45g.+

Buildinig'niumber six at the Watkin Is-Johns-on Instrument Pliant is' a -prefabricated steel structure.

tru i 967, the structure ha penhouse roughl 'thirty feet above grade.

Inside te pent cting cicular HVAC ductingk to an in-line axial*

an t' Fi d hung and the fan wassuported with a rod hanger/spring arrang~ement. The bellows were not designed to resist the differential motion imposed by the earthquake.

A-43

HVAC Duct System EarthquakeExperience Data

.Figure A-17 Watkins-Johnson Instrument Plant, 1989 Loma Prieta Earthquake. The Flexible Bellows Connecting HVAC Ducting to an In-Line Axial Fan Tore Also at Watkins-Johnson"'sbuilding number six, the supportframe anchorage for a large rectangular roof-mounted duct was distressed. The P-1000 unistrut frame and its clip angle anchorage were not designed to withstand the inertial loads. The duct was not damaged and, other than the minor anchorage distress, the support survived as well (see Figure A-I8).

Pacific Bell's Watsonville switching station is a four story concrete shear wall structure which endured a measured PGA of 0.33g. During the earthquake, a vertical cantilevered section of duct and its attached diffuse6rfell to the floor (see Figure A-] 9). Closer inspection revealed insufficient positive attachment between the cantilever and the header.

A-44

F/HVAC Duct System EarthquakeExperience Data Figure A-18 Watkins-Johnson Instrument Plant, 1989 Loma Prieta Earthquake. The Support Anchorage for a Roof-MountedDuct was Distressed A-45

HVAC Duct System EarthquakeExperience Data Figure A-19 Pacific Bell;Watsonville, 1989 Loma Prieta Earthquake. A Vertical Cantilevered.'

Section of Duct Fell to the Floor with its Attached Diffuser A*.2.1.8` 1990 P hilippines EarthqUake6. , . . -.

O.nMonday,July 16,.1990, at4:26 P.Mo.local time, the heavily populated island ofLueon,

..:,Republic of Philipines, was struck by an earthquake 6f magnitude 7.7. The ethquakeWas'.:

.caused bymajor rupture along the Philippine and Digdig faults, extendingapproximately seventy miles alofig the northern edge of th6 Central Plains and into the Cordillera Central. ',';.

The Texa'instrument- faeility in Baguiio City -,was-about forty miles norhwest f th-epicenter in a region of extensive landslides. No accurate estimate of theg round motioni exists. In orie region'6 of the facility, round, slip-jointed duct pulled apart at its seams and fell4t6 the floo"r._-The-rod hung duct had no positive connection between sections'and was attached to unanchored equiprnentand flexibly mounted fume hoods , creating the differential motion failure.The diffu'sers in the building's clean room also fell along with the room's suspended ceiling.

A-46

HVAC Duct System EarthquakeExperience.Data Figure A-20 Valley Steam Plant Forced Draft System, 1971 San Fernando Earthquake Figure A-21

'Drop IV HyrOln,17 merial Valley Earthquake.Ceilin'g' Mounted Ducting A-47

HVA C Duct System Earthquake Experience Data

- .'*>>'~---* -. *-

Figure A-22."

SCE Rosemead Headquarters, 1987 Whittier Earthquake. HVAC Dented from Sway of Adjacent Fixtures

" . '.: j*

Figure -23

-Watkins-Johnson, 1989 Loma Prieta Earthqtuake. HVAC Ducting Atop Roof Level A-48

HVAC Duct System Earthquake Experience Data

~L Figure A-24' Magnolia Plant, Burbank, Ducting at Induced Draft Fan, 1971 San Fernando Earthquake

-Figure A-25 El Centro Steam Plant, 1979 Imp~erial Valley Earthquake A-49

HVA C Duct System Earthquake Experience Data A.3 Summary of Observed Damage The cases of duct system damage listed above are generally limited to direct seismic damage of the ducting or supports. The database search also uncovered a number of instances in which HVAC ducting was dented or damaged by interaction with adjacent commodities. These cases include impact with flexibly supported piping, false ceilings, and equipment. HVAC diffusers have fallen from false ceilings on several occasions, typically when the ceiling is not properly restrained against lateral motion and the diffuser is not attached to the structural slab above.

In summary, seismic damage to HVAC duct systems from the seismic experience database can be characterized as follows:

" Broken and Fallen CantileveredSections. Cantilevered sections of duct and duct diffusers constructed of riveted lap joints and simple friction connections have broken or fallen in past strong motion earthquakes. The cases of damage appear to be the result or:

- High inertial loading of the cantilever sections causing high reaction forces at relatively weak joints

- Flexible headers developing high seismic stresses in short duct segments not flexible' enough to accommodate the motion

" Opened and ShearedSeams. Light gage circular duct constructed with riveted lap joints have opened up and sheared in past strong motion earthquakes. This damage has occurred at locations subject to high bending strain in very flexible duct systems.

" Duct Fallen off Support. The database includes one.example where the end of a cantilevered duct section jumped off of its end hanger support and was damaged. The duct was not tied to the support, and was subject to high levels of seismic motion.

" Equipment on Vibration Isolators.HVAC duct has been damaged by excessive movement of in-line equipment components supported on vibration isolators.

A-50

B SHVAC DAMPER EARTHQUAKE EXPERIENCE DATA Dampers are sheet metal fabricated devices that consist of a system of parallel vanes or louvers toeither per it or prevent air flow. The actuators controlling the position of these louvers can be operated manually, electrically or pneumatically.

B.1 Definition of Equipment Class Dampers are part of any heating, ventilating and air conditioning'(HVAC) system, and are found at nearly all industrial sites. The principal functions of this equipment are control of air flow and isolation of HVAC systems.:Some dampers'at nuclear plants are used in safety related applications and must function under extreme conditions of violent weather, radiation, temperature, seismic shock, and high pressure transients (due to loss of coolant accident or tornado transient). Dampers are self-supporting structures that do not require additional integral supports or bracing. These devices are typically used in the following applications:

• Inlet or outlet side of an air handler

" In-line in HVAC ductin'g

'Mounted in walls to allow Or prevent air flow between rooms Dampers may be operated passively, manually, or actively. The louvers of dampers are tied together by a common linkage which is externally controlled by an electric, pneumatic or manual actuator. Typical -c*rnponents mounted on an air operated actuator are air tubing, flexible conduit, solenoid operated valves and pressure gages. Air receiver tanks that supply air to the solenoid valves require separate evaluation.

BA.1. Equipment-Anchorage - -,

.Daiipers are an integralpart bf thefansiý airihandlers Hand VAC ducting and as such are-,

characterized as in- ine co ponents. Dampers in fans or airhandlei-s a-e par of the equipment and are evaliuaed With the "Rule of The Box S..6mel dampers such as fire dampers are mounted in walls or ceilings and therefore-are not co6nsidered as in-line components.:These devices are,.

normallyiattached to the si*pporiing equipment, ducting; rpeneratidns inowallsand ceilingsbby bolts,;rivetsor Welding along'their perimeter-.-Heavy motor-.operaied or pneumhatic dampers typically have their iwn-supporting system.

B-1

HVAC DamperEarthquakeExperience Data B.1.2 EquipmentApplications Dampers are typically operated pneumatically, electrically or manually. In the'.case of the pneumatically controlled and motor-operated dampers, such as flow/pressure control and isolation/shutoff dampers, a pneumatic or electrical signal is sent to the actuator to either open, *close or modulate the louver position. Some dampers, such as pressure relief and tornado protection dampers, are'self actuated when quick differential pressure changes are detected and use counterweights or counterbalances to return to normal position. Some fire dampers have fuseable link that would break in a fire and force the damper to close.

B.1.3 Application in Nuclear Plants Dampers. are used in all nuclear plants for control of air flow and isolation. of HVAC systems.

Dampers are utilized in the HVAC systems to perform one or more of the following functions:

o' Flow and PressureControl - Used to control a given flow rate or pressure within a system.

Actuators may be electrical,: pneumatic or manual.

o Balancing - Used to establish a flow. and. pressure relationship within a system. Actuation is through a manual adjustment hand-quadrant that is left at a pre-set level.

• Isolation/ShutoffControl,- Used to isolate or sealoff a portion of the system from selected flows. This type of damper is used only in an open/close application. Actuators could be electric, pneumatic or manual..

• Backdraft Control - Utilized where reverse flow of air is undesirable or could cause system inefficiencies. Actuation is by counterweight or counterbalance.

" PressureRelief- Used to0protect the system from excess pressure or damaging surges.

'The dampers are closed under normal conditions and open very quickly when positive pressures are detected. Actuation is by counterweight or counterbalance.

• Tornado Protection' Used at the intake or exhaust openings of the HVAC system.

During tornado conditions this damper closes automatically. Actuation is by counterweight or counterbalance.

• Isolation Shutoff- Used to prohibit any leakage passing through the damper and downstream.

• Actuators are typically eitherpneumatic or manual.

, FiireDampers_-Mounted in walls -orteilingsand is used for isolation of two separate but adjacent areas in case of fife. -

B.2Database Representation for Da'mpers

-Figures B-1 through B-3 show typical components of.dampers:

Figures B-4through B-I 8 present examples of dampters withinthe 'database. The database inventory of dampers includes at leastI 75 examples, representing 20 sites and 14 earthquakes studied. Of this inventory, there are no instances of seismic damage.

B-2

HVA C DamperEarthquake Experience Data EXTERNAL.

PosmQ*4N INDICATOR Figure B-1 Exploded View of a Typical Damper Hemme Bla~des p.

Figure B-2D Typical Damper Blades or Louvers N B-3

HVA C DamperEarthquake Experience Data

~i

. PNEUMATIC Van* type yinI -- o F 1e Figure B-3

%*hTR iu~b ;T" I Typical Damper Actuators.

188 Figure B-4 ':

Pneumatic Damper at El Cintr6 Steam Plant Subjected tothe 1979 Imperial Valley Earthquake.

B-4

". VA C Damper EarthquakeExperience Data Figure B-5 Louver Style Damper on the Boiler.Structure at the El Centro Steam Plant which Experienced the 1979 Imperial Valley Earthquake Figure B-6 Pneumatic Actuator at the Puente Hills Landfill Gas and Energy Recovery Plant B-5

HVAC DamperEarthquake ExperienceData 7 i- -" np5" i5fi Figure B Radial Type Damper at the El Centro Steam Plant which was Subjected to the 1979 Imperial Valley and 1987 Superstition Hills Earthquakes B-6

HVAC DamperEarthquakeExperience Data

'I

'4" Figure B-8 Louver Type Damper at Humboldt Bay Power Plant

.:i.-.* * .-. !. . .*

FigureB-9 Radial and LouverT ype Dampers at the'HumboIdt Bay Poweir Plantf whichExperienced the 1975 Ferndale Earthquake.

B-7

HVAC DamperEarthquakeExperienceData

., 77 X, V ..

Figure B-10 Motor-operated Damper at Adak Naval Station, which Experienced the 1986 Adak Alaska Earthquake

-'Figure Bt-1 Damper at Adak Naval Station

• B-8

HVA C DamperEarthquake Experience.Data Figure B-12 Pneumatically Controlled damper at UC Santa Cruz Applied Science Building Subjected to 1989 Loma Prieta Earthquake B-9

HVAC DamperEarthquakeExperienceData -

Figure B-13 Electric Motor for a Fire Damper at AES Placerita Cogeneration Plant Experienced the 1994 Northridge Earthquake B-IO

HVAC Damper EarthquakeExperience Data Figure B-14 Pneumatic Damper with Long Actuator at Valley Steam Plant which Experienced the 1971 San Fernando and the 1994 Northridge Earthquakes' k

B-II

HVAC DamperEarthquakeExperience Data Figure B-15, Pneumatic Louver Control Damper at Pasadena Power Plant which Experienced Several Database Earthquakes.-,

Figure B416 Heavy Pneumatic Controller with. Independent Support for a Large Damper at Pasadena Power Plant Located Very High in the Boiler Structure B-12

" " CDamperEarthquakeExperience Data HVA Figure B-17 Floor-mounted Air Operated Damper with Remove Actuator'at Burbank Power Plant Experienced the 1971 San Fernando and the 1994 Northridge Earthquakes B-13

HVAC DamperEarthquakeExperience Data Figure B-18 Large Independently Supported Damper Controller at the Burbank Power Plant Figure B-19 presents a bar chart that illustrates the inventory of dampers at various database sites as a function of estimated PGA.

The database represents a wide variety of damper configurations. Pneumatic, motor driven and manual dampers are well represented. Some damperis in the database are housed in steel boxes which ate anchored to0the ground or to the building's structural steel. Heavy pneumatically operated dampers in the database have their own independent supporting system, and their usually long actuators attach to the side of the duct for louver control within the duct.

B-14

n viiu .vanr rllpA...,wle*__

Peak Grun"Ac7 g V( IHJal iper Earthquake Experience Data Peak Gr'o~und Accelerati .on (AýVerage Horizontai) 50

' .C, 45 Ccu

.40 I - ' , ' --

35 ClD C ) cu CD a) E cu 30 o- : -o c n CL__ ... O n m l o 2 *-

4-.

!E ,. U) 0L. . L_ CD E Cu 0C m >%' cu >

25 ~~~  :'"

az "

'"C.o.,** 0 "L E 0 N- D a) EL"t-0LCun C'.'

0CD E 0)

~ C)=3 CD0 a) H-cou n 0 0C Cu CD 20 CUa'4C CuCD~

5.

n,,,C a) 0l N  : E

-3

. C Cu -.

15 -~ C 'u~ a) 'UU)C l CD CuuC Cu > ... - ' CD 2' cu 0 10 Cu CTO H 0 N (Da C

~ ua cuC Ia~C Cu .U 3 5 ~* CD' Cuo 0cCuCu Cl) 0

,020g . 30g 0.40g 0.50 g 0.60g Fgure B-19 Inveniotr of Darmpers w within Experience Database B-15

HVA'C DamperEarthquakeExperienceData B.2.1 Basis for the Generic Bounding Spectrum The seismic experience database includes a vast amount of data on the performance of dampers of various configurations and installations which experienced many different seismic excitation levels.: The Generic Bounding Spectrum developed by SSRAP [B-2] to represent the motion at typical data sites was based on'the average horizontal free field motion from each of the four reference database sites: Sylmar Converter Station (1971 San Fernando), El Centro SteamPlant (1979 Imperial Valley); Pleasant Valley Pumping Plant (1983 Coalinga), and Llolleo Pumping Plant (1985 Chile). The average of the four ground motion spectra is referred to as the Reference Spectrum. This spectrum is a conservative representation of the ground motion level to which the earthquake experience data demonstrate seismic ruggedness. In other words, the Reference Spectrum is used a's a measure of the equipment capacity which has been'demonstrated by experience. The GenericBounding Spectrum is obtained by dividing thisReference Spectrum by 1.5. This 1.5 factor is to acco0unt for the possibility that floor spectra within about 40 feet above garde in the nuclear power plant might be amplified over the ground spectra more than occurred in the database plants. Thus, the resultant Bounding'Spectrum is directly applicable for comparison with Ground Spectra.?The capacity as defined as either the Reference Spectrum or the Bounding Spectrum, coupled with caveats on equipment attributes and installation, is then compared to the demand as defined in the GIP Table 4-1.

El Centro Steam Plantexperienced a peak ground acieleration of 0.42g during the 1979 Imperial Valley Earthquake. Strong-motion at the site lasted about 15 seconds. The site ground motion is based on measurements from an instrument located within 1/2 mile of the plant.

This plant includes many pneumatic and manual controlled dampers. The positioners for these dampers are enclosed in steel boxes which are then anchored to the ground or the building structural steel. There were no instances of damage to the dampers or their operators in the earthquake.

The Sylmar ConverterStation located near the fault rupture of the 1971 San Fernando.

Earthquake, is estimated to haVe experienced at least 0.65g peak ground acceleration, with about 10 seconds of strong motion.

Eight instances of dampers are included in the database'at tiis facility. None of the dampers

,-...expeneieced any seismic 6--efet.

The Shdeli.WterTreatmentPlant is locatedabout two miles norh ofthe Main Oil Plant,.

,The peak ground acceleration experenced at this se during the 1983 Ccalinga Earthquake

.is conseratively estimated at 0.60g. - '

At this site only onedocument6ed case' of a butterfly damper exists in the database. This damper remained undamaged as a result of the earthquake.: - -

• The IBM/Sntia TeresadComputer-Facilityexperienced a PGA-of 0.37gwith strong motion occurring for about eight'seconds during the 1984 Morgan Hill Earthquake. This-facility included several motion monitors, one located in the free field 100 yardsTr6oim the main building.

B-16

. HVAC DamperEarthquakeExperience Data The database includes one pneumatic operated damper at this facility. This damper was not

• damaged in the earthquake.

" Valley Steam Plant experienced ground shaking during both the,1971 San Fernando earthquake and the 1994 Northridge earthquake. The peak ground acceleration at the site due to each of these earthquakes was approximately 0.40g. The plant,*which'includes four units with a total

...generating capacity of 513 MW, is located about 10 miles from the epicenter and three miles from the fault of the San Femando Earthquake.

Twenty four of the pneumatically operated dampers at this plant are represented in the seismic experience database. None of these dampers sustained any damage due to the above earthquakes.

Burbank Power Plant,located in the Burbank/Glendale area of the San Fernando Valley, is estimated to have experienced a peak ground acceleration of 0.25g, with about 10 seconds of strong motion, during the 1971 San Fernando Earthquake; This plant also experienced the 1994 Northridge earthquake with'an estimated peak free field acceleration of 0.30g. This plant consists of five steam generating units and two gas turbine units..

A total of 35 pneumatically operated dampers at this site are represented in the database.

No damage was reported to these dampers as a result of the above earthquakes.

PasadenaPower Planthas the unique distinction of being the onlysite included in the seismic experience database that has been shaken at comparable levels of intensity by four earthquakes, each producing'a level of moderate ground motion comparable to a design basis event for a nuclear plant in the eastern United States. The Pasadena Plant experienced the magnitude 6.6 San Fernando earthquake ini1971, the magnitude 5.9 Whittier earthquake in 1987, the.magnitude,5.8 Sierra Madre earthquake in 1991, and finally the magnitude 6.7 Northridge earthquake in 1994.

The peak ground acceleration experienced by this site'during these shakings is-estimated to be about 0.20g.

The database includes a total of 24 pneumatically operated dampers at this facility.

These dampers functioned properly during and after the above mentioned earthquakes with no damage. .. .

AES PlacentaCogenerationPlantexperienced a peak ground acceleration of at least o.60g during the 1994 Northridge Earthquake [B-3]. The estimated site' ground motion is based on'.-:,'::.-

measurements from-several instruments located a few'kilometers fromi he plant.

-Twentiysmall motor operated fire dampers f;o Halon system isolation are includedinthe database forithis plan't:No damage, as a result of the Northridge earthquake, was repoited for these-dampers.

B-17

HVA C DamperEarthquakeExperience Data B.3 Instances of Seismic Effects and Damage The experience database contains no instances of seismic effects to dampers. The database contains no evidence of the malfunction of dampers during or immediately after an earthquake.

In addition no instances of seismically induced damage to dampers were found in an extensive literature search. In general dampers can be classified as inherently rugged equipment.

B.4 Sources of Seismic Damage The seismic experience database indicates that dampers possess characteristics that generally preclude damage in earthquakes. The experience database contains no instances of damage or significant seismic effects to dampers or their actuators.

B.5 Caveats for Dampers The equipment class of Dampers described below has been determined to be seismically rugged based on earthquake experience data, provided the intent of each of the caveats listed below is met. This equipment class includes all components of dampers installed in HVAC systems (or other types of duct systems). Fire dampers which are installed in walls or ceilings are also within this equipment class. Damper components are louver blades, actuators (pneumatic, electrical, and manual, as well as automatic counterweight and counterbalance actuators),

attached air tubing and rigid or flexible electrical conduit, solenoid valves and pressure gages.

Dampers are sheet metal fabricated devices that consist of parallel flaps to either permit or prevent air flow. Dampers are an integral part of fans, air handlers and HVAC ducting and in case of fire dampers they are installed in walls or ceilings. The flaps or louvers of dampers are tied together by a common linkage which is externally controlled by an electric, pneumatic or manual actuator. Automatic dampers are operated by a pre-set counterweight or counterbalance.

Attachment of dampers to the HVAC ducting or equipment is through bolting, riveting or welding provided around the perimeter of the damper housing. The pneumatic or electric motors that control the actuation are typically attached to the damper housing; however, they also could be mounted on a nearby wall or floor with rack and pinion connection provided for the actuator.

Dampers with heavy motor-operated actuators (typically greater than about 200 pounds) that are installed in-line in HVAC ducting are also represented in the database. This type of damper, however, should have its own independent support system.

The Bounding Spectrum represents the seismic capacity (defined as free-field motion at effective grade) of dampers when the damper meets the intent of the following inclusion and exclusion rules. Note, however, that when the specific wording of a caveat is not met, then a reason for concluding that the intent has been met should be provided on the SEWS.

DMPR/BS Caveat I - EarthquakeExperience Equipment Class. The damper should be similar to and bounded by the DMPR class of equipment described above. The equipment class descriptions are general and the Seismic Capability Engineers should be aware that worst case B-18

HVA C DamperEarthquakeExperience Data combination of certain parameters may not be represented in the generic equipment class.

These worst case combinations may have reduced seismic'capacity and should be carefully evaluated on a case-by-case basis.

DMPRIBSCaveat 2 - Damper Operator/ActuatorNot of Cast Iron. The intent of this caveat is to avoid the brittle failure mode of cast iron as evidenced by poor performance of some cast iron components in the past earthquakes. Note that the database does not contain actuators with cast iron components; therefore, it is not necessary to determine the material of the damper control components unless it appears to the seismic capability enginfeers to be made of.cast iron.

DMPRS/BS Caveat 3 - Sufficient Slack and Flexibility of Attached Lines. Sufficient slack and flexibility should be present in attached lines (e.g., air tubing, electrical conduit) to preclude a

-linebreach due to differential seismic displacement of the equipment and the line's nearest

• .support. AlSo,fordamper positioners with independent supports (i.e., not mounted integrally on the duct) the effect of differenti*l displacement on the actuator.(with actuator defined as the rod connected atone end to the positioner and at the other end to the'duct louver controls) needs to be considered. The issue'here is to watch out for cases where the'actuator is connected to a rigidly mounted positioner at one end and to a rod hung duct system at the other.

DMPRS/BS Caveat44 Adequiate Anchorage.Damper controls when mounted on the ground or nearby structures should be properly anchored in accordance with-the guidelines of GIP section 4.4. When the motor- or pneumatic operator is mounted on the duct at the damper location the adequacy of the attachment point to the duct skin or its stiffeners should be ensured.

DMPR/BS Caveat 5 -'DuctDistortion.The duct at the damper location should be carefully investigated for any'signs of distortion as this would interfere with the damper operation.

B.6 References B-I. Quality Air Design, "Design of Nuclear Dampers," Division of ACDC Inc.

B-2. Senior Seismic Review Advisory Panel (SSRAP), "Use of Seismic Experience and Test Date to ShbwRuggedness of Equipment in Nuclear Power Plants,"

Sandia Report SAND92-0140, Part 1, 1992..

'B-3. The January17, 1994,NorthridgeEarthquake. Effects on Selected IndustrialFacilities and Lifelines, Prepared by EQE Intemra'tional, Electric PoWerResearch Institute, July, 1995...- .

BA4.,.,-.TheOctAberi:1987 WhittierEarthqake:'Effecis'on Selected Po6wer, lndustrial," -

and CommeiidalFacilities,Prepared by Electric Power Research Jnstitute, .

' 'andEQE EngineeringInc; December, 1990. EPRI NP-7.126 - .

,"~ ~ . een ., nc .ec...

B-19

C DEVELOPMENT OF ALLOWABLE SPANS FOR SHEET METAL DUCTS Allowable span length charts for horizontal ducts with seismic loads are developed to check for conformance with SMACNA standards. These charts may be used during the in-plant screening review or to guide'sample selection for analytical reviews. The screening charts consider seismic and dead weight loading; pressure loads are decoupled since they only influence duct thickness and spacing between duct stiffeners. Seismic loading consists of horizontal and vertical static approximations using peak spectral acceleration. Dead load stresses are summed absolutely with the Square Root Sum of the'Squares (SRSS) of the vertical and horizontal seismic stresses.

Allowable span length calculation ci-iteria are developed using analytical requirements presented in Section 4. HVAC duct systems not meeting these spans should be selected for analytical evaluation. The process for developing allowable span charts is described below.

C.1 Rectangular1Ducts The evaluation guidelines (see'Section 4) for rectangular ducts define the section property of a rectangular duct as being co6mprised of 2- by 2-inch angular sections at each comer.

The maximum expected bending moment is approximated by:

M= '1-(For ducts spanning over one or two spans) Eq. C-1 8

M =wl 10. (For ducts

• spanning

-: . over 3 or more supports) Eq. C-2 where: -

Sw =a:-appiedload (lbs/in)".

I '. span between vertical supports (in)

The rectangular duct allowable span length for the typicalcase"(for ductsspanning over 3 or more supports) is determined by:.-

2

• 1=F..I6~b -v1¶ * - Eq.GC-3

-,[7(2H +2wV)p ý C-I

Development ofAllowable Spansfor Sheet Metal Ducts where:

p = duct wall material density (lb/in3 ). Note: an equivalent weight density should be used to account for additional material weight on the duct wall, such as joints and stiffeners.

Fb = allowable material stress (psi)

H duct height(in)

W. = duct width (in)

Sa = horizontal peak spectral acceleration (g's)

Sv = vertical peak spectral acceleration (g's)

R = (horizontal restraint span length)/(vertical support span length)

K1 = a derived constant (in") based on a rectangular duct with a linear weight of 2p t (H + W), section moduli based on 2-inch by2-inch angular sections at each comer, such that the section modulus about the horizontal is 8t [H - 2 + (2/H)] and the section .

modulus about the vertical is 8t [W - 2 + (2/W)], and an SRSS summation of seismic stresses, resulting in Se 2 R4W .62 .(HS2H2. .O)2]

H2 /2-H+1.o0)

(H Note that the allowable span length equation is independent of duct thickness since duct section modulus and duct weight are. both linear with'respect to duct thickness.

For a given duct geometry, allowable'span length screening charts can be developed using Eq. C-3 for various span' ratios*and speciral'acceleration levels. Material allowable bending stress should be taken-as defined in Section 4.1.1.

C.2 Circular Ducts The evaluation guidelines (see Section 4) for circular duct support spacing define the duct section modulus tobe:, -

Z (O.25)(71$D.t -- Eq."-

- where:-

.D d*u*ict diameter (in)

.t ,- ,du tithickness (in)"_ . -

duct section m6dulus <in 3 ) ,

C-2

Development ofAllowable Spansfor Sheet Metal Ducts The maximum bending moment is approximated by:

M - (For ducts spanning over one or two spans) Eq. C-5 w2 M - (For ducts spanning over 3 or more supports) Eq. C-6 The design of duct support spacing for circular ducts is governed for small spans by duct bending.: As the span length increases, buckling controls. Allowable stresses reflecting these modes of failure are given in Section 4.1.2.

The circular duct allowable span length for the typical case (for ducts spanning over 3 or more supports) is determined by:

r 5Fb D 1112. = I Eq. C-7

• L2pK 2 where:

p = duct wall material density (lb/in 3 ). Note: an equivalent weight density should be used to account for additional material weight on the duct wall, su'ch as joints and stiffeners.

Fb = allowable material stress (psi)

Sa horizontal peak spectral acceleration (g's)

Sv= vertical peak spectral acceleration (g's)

R = (horizontal restraint span length)/(vertical support span length)

D = duct diameter (in).

K2= a derived co0nstant (lb/in 2) based on a circular duct with a linear weight of p7 D t, a section modulus of 7t D2t/4, and an SRSS summation of seismic. stresses, resulting in

.1,+(Sa2 R4 +S 2

)1/2

• -,For agivenduct geometry, alloWable span length screening charts can be developed using Eq. C-7...

C-3

SEISMIC AND PRESSURE TESTING OF HVAC DUCTS D.1 Introduction Several tests were conducted by testing facilities to demonstrate the inherent resistance of HVAC

.systems to seismic damage in combination with pressure loadings. The test pressure loading (both positive'and negative -pressure) was generally severil times the typical normal operating pressures in the ducts at nuclear power plants. Similarly, theseismic test loading, in the form of biaxial input motions or equivalent static loadings, was greater or equivalent to the maximum seismic demand at most nucleai-power plants. The tests confirmed that the HVAC ducts

.constructed to SMACNA standards have adequatestructural integrity and functional capability for the postulated DBEloads, as well as the normal operating pressure loads.

D.2 HVAC Duct Test Programs

• D.2. I Summary of Tests Performed for TVA Ducts Vibration testing of rectangular ducts, which included both pocket lock and companion angle duct constructions, was Conducted. Three different duct sizes (60"X24", 48"xl 8", and 36"x24")

with width-thickness ratios ranging between 602 and 1671; -andconstructed to SMACNA standards, were tested. Four of each duct size were available for testing for a total of twelve test specimen (six ducts with-pocket locks and six ducts with companion angles). Each duct size was tested, nonconcurrentlyin the two directions perpendicular to the longitudinal axis of the duct specimen. The sheet metal thickness ranged from 20 ga. to 22ga. and the duct span lengths varied from 14 to 28 feet. Inorder to tune the test setups to a first mode resonance of 8 to 1I Hz.,

which.was the frequency rangeof the dominant response-as~defin'dby the required response spectra (RRS) with a peak acceleiation valsue of 6.4g, a variable support was designed to alter the struictural response of theduct/support system -

S. loa ds of up to 6.4g with n r very littl6edamage. The compainion angle ducts experienced minor, h ighly focalized failures in the-duct skin'thati occurred as small separations in the duct skin corners or near. a stiffener. Theselocaized separations remained sufficiently closed that air delivery would not`be significantly impaired. Ducts with pocket lock construction demonstrated an unexpected capability forrsustaining higih dyamic loads. The more flexile joints and higher damping in this type of construcion arethe primary ireasons tthat no 1ocal filures, such as found with the companion angle ducts, were observed with the pocket lock constriction.

D-1

Seismic andPressureTesting ofHVAC Ducts The average damping values obtained from testings forcompanion angle and pocket lock ducts were about 7% and 10%, respectively. In addition, first mode natural frequency of each duct specimen was determined during testing. These tests revealed that the fundamental mode frequency of both pocket lock and companion angle ducts'wg less than what would be predicted based on beam theory and using the SMACNA four corner method to calculate the effective moment of inertia of the duct section. The resulting reduction factors used to adjust Calculated natural frequencies are 0.59 and 0.87 for pocket lock and companion angle constructions, respectively.

All duct specimen were subsequently iested to failure. The peak acceleration value of actualtest response spectra (TRS) at'failure'ranged from 10.2g to 14.0g for the companion angle ducts, and froml 1.Og to 16.2g for the pocket lock ducts. Analysis of the test results, ising the acceleration levels sustained at failure, indicated a bending stress at failure ranging from 25.2 to 51.7 ksi calculated by the SMACNA four corner effective section method. The ,general failure mode

-for companion angle ducts was a gradual, very ductile failure, with no complete separation of sections and with no gross opening of the pressure boundary. The general failure -modefor pocket lock ductswas usually a sudden openingof the crimped joint. A sudden,-catastrophic type of failure resulted and actual separation of duct sections caused the span to fall to the test table.

It is noted that two of the pocket lock ducts could not be failed due to force limitations of the shake table.

D.2.2 Summary of Tests Performedfor Limerick Ducts The test program for ducts at Limerick consisted of testing seventeen test groups. Each test group consisted of three identical specimen except for one test group which had one specimen. Fifteen test groups included rectangular ducts with sizes ranging from 24"x24" to 96"x48". All duct specimen were of welded construction with a minimum sheet thickness of 18 ga., and the actual width-thickness ratios varied from 502 to 1605.Stiffener angle sizes ranged from l"xl"x1/8" to 3"x3"xl/4", with spacings ranged from 24" to 48".

All specimen were tested for negative pressure with the exception of one specimen that was tested for positive pressure. The average negative test pressure ranged from 17.8 to 104.2 inches of water gage and the positive test pressure was 48.0 inches of water gage. Duct spans were from 8 to 12 feet long, andall"duct Specimen weresimply-supported on the ends along the bottom end stiffener widths with'the exception of two specimen that were supported along their end vertical stiffeners mounted on thie'h~ight of the ducts. -

-All ducts were-subjected to live load orseismic load simulation tests,' or both, follwed bythe pressur test to failure.or to a maXimumnegative pressre of 14.psi (-407.0" w.g.). Application.

-of live lbad and simulated seismic loadwas accomplished by predetermined steel weights

.eandbagged sand.Test internal pressures (negative or positive pressures)were appied to the specimen bby.an electrical pump -connected in series *with an accunmulator tank. Only ducts in two groups were subjected td simulated seismic loading. The test sequence began:with the live load tests followed by.the seismic load tests, if any. Thereafter, the "pressure tests to failure began.

None of the ducts failed during the live load or seismic load teSts. Aliducts failed during the pressure load testing,-with eXception of the 8" diameterduct that did not fail..'

D-2

Seismic and Pressure Testing ofHVAC Ducts In general, the test results demonstrated that failure modes of ducts were not catastrophic and there is a great reserve strength after failure. The negative pressure loading was the most important loading, since the failure mode under positive pressure was stiffener buckling whereas under negative pressure loading the duct failure mode was either the stiffener buckling or the corner crippling of the sheet metal. Dead load, live load, and seismic stresses in duct acting as a beam between supports were relatively low. The test results also supported using duct width-thickness and height-thickness ratios of up to 1500, as opposed to 500 and 200, respectively, per American Iron and Steel Institute (AISI) requirements.

D.2.3 Tests Performedat OtherPlants Similar duct tests were also conducted for CPSES and CP&L plants to verify the structural integrity of ductwork, particularly with longitudinal seam construction, under combined seismic and pressure (both positive and negative) loadings. The results of these tests are generally in agreement with the duct tests described in detail above.

D.3 Conclusions from Test Programs In general, the tested ducts were either constructed to the SMACNA standards or were of a less conservative construction. The tests, collectively, provided the following results:

• The duct beam properties established based on the test results are comparable to the method prescribed by SMACNA guidelines but are less conservative.

• The average damping values for companion angle and pocket lock construction were established to be about 7% and 10%, respectively.

• Long spans of ducts (14' to 28') performed adequately under seismic input motions with a peak acceleration value of up to 6.4g.

• When tested to failure, such as seismic input motions with peak acceleration value ranging from 10.2g to 16.2g, the failure of the duct specimen was very gradual and of ductile nature, except for the ducts with pocket lock construction in which the crimped joints would suddenly open and cause a catastrophic type failure.

The overall conclusion from these limited tests indicates that as long as brittle failure of duct section connections is precluded, duct deformation under increasing loads is very ductile.

Furthermore, for HVAC ducts with typical span lengths of about 15 feet and constructed to the SMACNA standards, duct capacity can be expected to significantly exceed typical demand under the combined normal operating and seismic loadings postulated for most nuclear power plants.

D-3

Seismic andPressureTesting of HVAC Ducts D.4 References D-1. Neely, B. B., Warrix, L, "A Qualification and Verification'lmprovement Test Program for HVAC Ducts'Used in Nuclear Power Plants," presented at Century 2 Pressure Vessels and Piping Conference, August 1980..

D-2. -Neely, B. B., Warrix, L., "A Procedure for Seismically-Qualifying HVAC Ducts Used in Nuclear Power Plants,, presented at Second ASCE Conference on Civil Engineering and Nuclear Power, September 1980.

D-3. Desai, S. C., K. P. Buchert,ýand E. A. Marcinkevich,; '"Structural Testing of Seismic Category I HVAC Duct Specimens," Second ASCE Conference on Civil Engineering and Nuclear Power,-Volume I., Knoxville, TN, September 1980.

D-4. Dizon, J. 0., E. J. Frevold, and P. D. Osborne, '"Seismic Qualification of Safety Related HVAC Duct Systems and Supports," 1993 ASME Pressure Vessel and Piping Division Conference, Denver, Colorado, July 1993.

D-4

ROD FATIGUE EVALUATION GUIDELINES E.A Introduction Shake table tests have shown that the seismic capacity of fixed-end rod hanger trapeze supports is limited by the'fatigue life of the hanger rods. Rod hanger trapeze supports should be evaluated for possible fatigue effects if they; are constructed with fixed-end connection details.

Fixed-end connection details include double-nutted rod ends at connections to flanges of steel members, rods threaded into0shell-type concrete expansion anchors and rods connected by rod coupler nuts to non-shell concrete expansion anchors. Fixed-end connection details also include rods with lock nuts at cast-in-place light metal strut channels and rod coupler nuts welded to overhead steel.

This attachment describes a screening method for evaluating rod hangers for fatigue based on

• theuse of rod fatigue bobundi ng spectra (shown in Figure E-1) and generic rod fatigue evaluation screening charts (shown:in Figure E-2 through E-6).

1.5 -

.2 1.0 J5g V.:.

0.5 .50g k

r I 10.1 0A10

- Frequency (Hz)

Figure E-1 Bounding Rod Fatigue Spectra E-l

Rod FatigueEvaluation Guidelines 1/4" THREADED RODS 450

Ii 1500

'250.

ao -

a 12 WrIkio Ac,60Pabla Rod Lon'gth IL; Ni Figure E-2 FatigueElevation Screening Chart for 'h"inch Diameter Manufactured All-thread Rods.

Weight Corresponds to the Total Supported Load (i.e., on both Rods). Length Corresponds to Clear Length The screening charts'are directly applicable to hangers constructed of manufactured all-thread rods in overhead suspended system runs with uniform length hangers. The charts may also be used for evaluation of supports constructed of field threaded rods and for short, isolated fixed-end rod hangers in more flexible systems with relatively much longerrod hangers; guidance is given later in this appendix on how to adjust the parameters when evaluating these special cases.

fatigue' evaluation should be conducted for rodianger *supports that have roids with fixed end

. A

onnection details. Forirod'hung HVAC duct systems with'rods of uniform length, c.

the fatigue ealuationi is conducted as flows (a) Obtaintthe 5%damped floor response spectrum for the l6cation of the support attachment-.

(b): Comparie the Bounding Rod Fatigue'Spectraof Fiigure E-l with the damped floor response spectra. oFra given ZPA, if a-Rod Fatigue Spectrumrentirely envelops the floor response spectrum, proceed to step (c).- f the Rod Fatigue&Spectrum does not entirely envelop the floor response siectrum,'then compare the Rod Fatigue spectrni with the floor response spectrum (unibroadened) at the frequency of the support. Supportfrequency may be estimated as follows:

E-2

Rod FatigueEvaluation Guidelines f 1 U Ks where:

= Wi 1 /g (lbs-sec/in)

= 24EI/L 3 + Wui/L (trapeze support, lbs/in)

S.. Wequtv'

= total dead weight on the pair of rod supports (lbs)

= gravitational constant (386.4 in/sec 2)

E Young's modulus of rod hanger material (psi)

-. 4 I moment of inertia of rod root section (in)

= length of rod above top tier,(in) 3/8ý 4TREADED RODS fO . 0.W9 0.7 .ZPAm IA'

'ii."*" n"' 6 "r' 38f6 /nn 0 aS 12- 16 20 34 U2 MknmAmptabI Rodwih(.L Figure E-3 Fatigue Evaluation screening Chart for 3/8 inch Diameter Manufactured All-thread Rods.

'Weight Corresponds to'the Total Supported Load (i.e., on both Rods). Weight Corresponds to Clear Length r' E-3

Rod FatigueEvaluation Guidelines 1/2. THREADED RODS (0.40-.

o5o and 0.75 ZP^s)

ý3

• 2.8

.2.4 12.8 120

• 10 I" . . .9Minimm Acceptble Rod Length (L In.)

Figure E-4 Fatigue Evaluation Screening Chart for.1/2 - inch Diameter Manufactured All-thread Rods.

Weight Corresponds to the Total Supported Load (i.e., on'both Rods). Length Corresponds to Clear Length If the bounding Rod Fatigue Spectrum does not envelop the floor response spectrum at the frequency of interest, then a more detailed evaluation should be conducted (by requirements other than the screening'evaluation requirements presented herein).-

(c) Enter one of the Fatigue EvaluationiScreeningCharts showninFigures E-2 through E-6

'corresponding to the diameter of the threaded rod.-Use the curve associatedwith the

-acceleration (0.33g, 050g or.0.75g) of the Rod Fatigue Bounding-Spectrumr of the'previous step. If hanger length is greater than minimum acceptable length, and suport dead weight is less than maximum acceptable weight, then the support is acceptable.*This chart is applicable for all:continuously threaded rods. For field threaded rods see (d) below.

(d) ,If field threaded rods are to be evaluated, then the screening chart rnay be used formodified

'* rod lengths and weight For1field threaded rds,rdouble the weight and decrease rod length by .1/3 befoie using thechart. -

E-4

Rod FatigueEvaluation Guidelines 5/8 ,.THREADED RODbS

• g ZPAS) iOna o0.7Th 4OA

.4A S3.0 12.0 i.5 0 .10 .20 .30 Mh*~ujmt Acmpt" Rod Lengt (L IM)

Figure E-5.

Fatigue'Evaluation Screeninig ,Chart for 5/8-inch Diameter Manufactured All-thread Rods.

Weight Corresponds to the Total Supported Load (Le.,, on both Rods). Length Corresponds

..to Clear Length "rqec~ he ma R`d b, I"'dnwAc~l i dt,

'If isolated, short fixed-end rod hangers are used in a system with predominantly longer, more flexible hangers, a special e1valuiation should 'be conducted that decouples the response effects

• of the short isolated rod. The pecial evaluation proceeds as flos

,(a) Estimate'the freq-uencyof the system, neglecting the isolate, shortrod support.,

isoee acf xe-n short rodhagrarusdnasytm o thpeoinlyogr, A ilated, using

-TFa eEvu sti tiornforimula givef above may be used, provid ngthat the ength, of the loniger rods isconsidered.-

(b) WAssue that the rod fatigueTounling spectrue Laieops the applicable floor, response spectrum at ~this frequency of iniekrsit.

(c)liBle -calcnlate an equival ent weight f onucth ationofailesht rnod using the fsriquenc rof the long rod's aas follows: -

24(rapze sport) u 42qItf 3 )2 0~ -gL?

E-5

Rod FatigueEvaluation Guidelines (d) Enter the appropriate Fatigue Evaluation Screening Chart (Figures E-2 to E-6) by-using the above calculated equivalent weight and length fof th1e isolated short rod hanger. If these parameters are in an acceptable region on the Fatigue Evaluation Screening Chart, then the isolated, short, fixed-end rod hanger is seismically, adequate.

-Whenusing the charts, the simpleequations given in this section for calculating response frequency should be used for consistency since these'are the same equations used to generate the screening charts (that is, the screening charts are based on the simplified results obtained from detailed fatigue analysis, considering capacities determined by component test results).

.3/4' THREADED RODS.

(OM~g U-909 arv 0.75g ZPAs)

LS

~40 I .0

~23-

.0 1.5 1.0

- h-1irmm rads~d Lar* ALkNO

'Faigure 1.Evauaio

'- enn hI-tfr

-cr ,/i' F n ameter. Manufactured All-thread Rods.

I egtCorresp6n'ds to the Total Supported Load (i.e.; on both Rods). Length Corresponds to Clea-r Length E-6

GUIDELINES FOR LIMITED'ANALYTICAL REVIEW OF SUPPORTS F.1 Introduction.

A Limited Analytical Review (LAR) should be performed to assess the structural integrity of HVAC. duct supports chosen as representative, worst-case boundingsamples of the evaluation

• scope of HVAC duct SYstems. The purpose f the LAR is not to estimate actual seismic response and system performarnce during a DBE. Rather, the LAR is intended to demonstrate that the HVAC duct supports are at least as rugged as supports that performed well as evidenced by past experience, using empirical methods, plastic design principles, and engineering judgment.

There are several steps in the LAR process that must be understood in their entirety in order to ensure that the intent of the evaluation guidelines are met. These steps include:

. Dead load check

• Vertical capacity vheck

. Ductility review Lateral and longitudinal load check

, Rod hanger fatigue evaluations The above checks are described in detail in the following sections except for the i6d hanger fatigue evaluation. Guidelines for rod hanger fatigue eValuation are contained in Appendix E.

..:The first check to be performed is a itandaid, dead load design check. Supports not passing this check are outliers This check serves the functions6 fan inclusionruile. Mos't f the earthquake experience database supports are conservatively assumed to'have been adequately designed for

..dead weight.,Adequate dead load design is thus the first important step for verification of seismic adequacy. This check is discussed in- Section F.2.

The second check is the vertical capacity tchecki. Thisý heck ensures high capacity.*" arnchorage and primary anchor connections for the.support, using simple calcukiaional methods Position.

retention is considered the most impoIrtant aspect"of ensuring structural integrity.,This check is described in Section F.3. ` - -

The third check* i-sdiictility review. This'requires an assessment of how the support responds to lateral and longitudinal seismic motion,and what are the"weak links in tlIe* upport load path.

The next two checks are the lateral and longitudinal load checks. These checks are static F-I

GuidelinesforLimitedAnalytical Review of Supports coefficient approaches for evaluating support capacity. If fiildire modes are ductile, then the lateral and longitutdinal checks may not be required. See section F.4 for a discussion of the ductility review, and section F.5 for the lateral and longitudinal checks.

It is important for the evaluator to understand the functional goals (following the DBE) for the HVAC duct system being reviewed. If the seismic evaluation is being perf6rmed solely to ensure structural integrity, then support flexibility and ductility principles may be used to their fullest extent. Conversely, if duct system pressure boundary integrity is of high concern, then the evaluator must use caution when applying the ductility guidelines contained herein.

When ductile,.plastic deformation of supports is allowed in either the lateral or longitudinal directions of motion, judgment must be passed on the potential consequences of this support behavior on the duct system. For example, co0nsider an axial iruii of duct withanelbow at the end to a transverse rni. If, in the longitudinal direction for the'axial rfun, the supports are allowed to go into ductile plastic deformation, then the evaluator must ensure that the first lateral support around the elbow to the transverse run will respond in a similar manner. If not, the support may act as a hard spot; and cause potentially detrimental consequences to the duct elbow or that first lateral support on the transverse run (see Figure F-I)..

RIGID CEIUNG PENETRATION *<

FLEXIBLE DUCT.,

.,,EISMJIC LOAD FROM"LONG'--.. """ " ..

Figure F-1 Vulnerable Duct Elbow Adjacent to Rigid Lateral Restiaint. r" F-2

Guidelinesfor Limited Analytical Review ofSupports F.2 Dead Load Check A detailed dead load design review of the representative worst-case bounding sample HVAC duct supports should be conducted using normal design working stress allowable loads.

The check should consider the as-installed configuration, connection detailing,and loading

..condition of the support. All components such as bracket and trapeze cross members, vertical support members, internal framing connections,-and support anchorage should be-checked.

All system eccentricitiesinelUding load to anchor point eccentricity, should be considered.

Evaluation of clip angle bending stresses may be excluded for trapeze supports suspended.

from the overhead. Loads from other attached systems, suich as piping or conduit, should be Srcosidered.

Consideration should also be given to the seismic adequacyiof the wall to which the HVAC duct supports are attached. Reinforced concrete structural walls are not a concern but masonry walls should bechecked to verify that they are seismically adequate.Anchorage into transite walls (asbestos fiber board) and gypsum board partitions should be considered outliers.

Reduced anchor bolt capacities should be used for expansion anchors in masonry block walls. The anchorage of partition walls and shielding walls should be checked.

F.3 Vertical Capacity Check The check concentrates on the support anchorage, focusing on the weak link in the support anchorage load path. High veftical eapacity is one of the primary design attributes that is given credit for good seismic performance. The Vertical Capacity Check evaluates whether the vertical capacity to dead load demanid ratio is as least as high as that of support systems in the earthquake experience database that performed well. The high vertical capacity provides considerable margin for horizontalearthlquake'loading.

The Vertical Capacity Check is an equivalent static load check, in which the support is subjected to 5.0 times Dead Load -in the downward direction. This check is limited to the HVAC duct support primary connections and anchorage. It is not necessary. to evaluate clip angle bending stress or secondary support members. The0lower support member of floor-to-ceiling configurations and base-mounted supports should be checked for buckling.

Eccentricities resulting in Airchor prying and eccentricities between vertical support members

'..'and anchor points should, in general, be igniored.Thisboncept is theresult of back-analyses.of-earthquake experience database supports and iscon'sisteint vith limit state condiitionsobserved lnteaboratories.

.-Forcantilever bracket support types; theeccentricity of the cantilevered dead load should

-eignored-

"For trapeze frame and rod-hunig supports, load distribution between the two vertical framing members should be considered only if thecenter.of the1loadis significantlyldistant from the centerline of tie support frame. The bending strength and stiffness§6f frame members should be checked-for transfer.of the 10ad beween anchor bolts when overhead support is provided by light..

-metal framing with anchor bolts spaced at relatively large intervals'and when multiple anchor bolts are needed to resist the vertical load.

F-3

Guidelinesfor Limited Analytical Review of Supporns For most HVAC duct support sý,siems, the anchorage will be found to be the weak link in the load path. For these cases of HVAC duct supports the Vertical Capacity Check is simply a comparison of anchor capacity to 5.0 times the supported load...

If the 5.0 times Dead Load guideline is not met, then the support should be classified as an outlier.

F.4 Ductility Check An evaluation should be conducted of the supports selected for review to characterize their response to lateral and longitudinal seismic motion as either ductile or potentially non-ductile.

The purpose of the ductility check is to identify support configurations that require a lateral and longitudinal load check (discussed in Section F-5).

Supports suspended only froim overhead may be.characterized as ductile if they can .respond to lateral seismic motion by swinging freely without degradation of primary vertical support connections and anchorage.' Ductile, inelastic performancelsuch as clip angle yielding or vertical support member yielding is acceptable so long as deformation does not lead to brittle or premature failure of overhead vertical support.

Review of typical HVAC duct support systems in the earthquake experience and shake table test databases indicates that many overhead mounted support types are inherently ductile for.

lateral seismic motion. Back-analysis of many database supports predicts yielding of members and connections. These database systems performed well;,with no visible signs of distress.

Ductile yielding of suspended supports results in a stable, damped swaying response mode.

This is considered to be acceptable seismic response and use of the support plastic moment is permitted.

• The ductility review of anchorage connection details is most important for rigid-type suspended supports. Supports with rigid, non-dUctile anchorage that do not have the capacity to develop the plastic strength of the vertical support members can possibly behave in a non-ductile fashion.

Examples include large tube steel supports welded to overhead steel with relatively light welds, or rigid supports welded to large base plates and outfitted with relatively light anchorage. These types of support systems are not well represented in the database and are not preferable since theyave a brittle failure mode.

The seismic design of certaini HVAC duct support members' my hav, been controlle'd byhigh:

frequency requirements ratherthandesign loads yet richors may havebeen sized b6y the design loads. These types of supports may have low seismic margin due to loads placed n' the support. *

"which were not considered by the original design.Supports with rigidnon-2ductile'anchorageare subject to further strength review.(see Section F.5). -

Examples of ductile and non-ductile 'support connection detaiisiand configurations are described below and illu si'ated infigureF-2..

F-4

Guidelinesfor Limited Analytical Review of Supports A.

I i = == . 01 ..

.. 1

• I WAL=*

I ~ i AKNMLLMff 0.D W'

I Figure F-2 . -

Examples of Potentially Non-Ductile Connection Details and Configurations"

--- -StdardCatalogLightMetal,;StrutFramingMembers, Clip*Angles, and Bolt withChannel' i Nuts.unbraced supports suspended from overhead,"c6nsti'u'cted of stanidard Cataioglight metal,

-,strut framing channels, cip angles, and bolts with channel nuts may be characterized as ductile.

-. . This includes supports constructed of standard ¢ctalog light'meiai strt fruramin gusseted, clip.

angle connections.,

Welded Steel Member. If an anchor point connection weldig stronger thant:heyertical member,.

then a plastic hinge will be-able .to"fofm intfhe y'eriical membei,:allowing dtictile response without weld failur6e.A support is seismically rugged so long as overhead support is maintained.

In this case, plastic hinge action in the'vertical member prevents transmission of loads capable of F-5

Guidelinesfor Limited Analytical Review ofSupports failing the welded anchorage point."For open channel struictiural sections, an all-around fillet weld whose combined throatthicknesses exceed the thickness of the part fastened, may be considered capable of developing the plastic hinge capac'ity of an open channel section vertical member. If the Plastic hinge capacity of the framing support member exceeds the capacity of the weld, then a brittle failure is possible, which is not acceptable seismic performance. For light metal, strut framing members, welded connections are likely to be non-ductile and thus'not capable of developing plastic moment capacity of the framing member.

Ceiling ConnectionPlate Secured with ExpansionAnchors'. Supports with overhead anchorage provided by a plate attached to concrete with expansion anchors should be evaluated for ductility as follows. The anchoiage may be characterized as -ductile if it is stronger than the plastic flexiiral strength of the vertical support member. A simple anchor moment -capacity estimate

may be used; by multiplying the bolt pullout capacity times the distance between the bolts or center of bolt groups. In some cases, it may be possible to demonstrate ductility if the ceiling

• connection'plate is the weak lirik in the anchorage load path. This is similar to the case of clip angle bending. The key to'characterizing a support as ductile ornon-ductile is reviewing the anchorage load path, and determining if the weak link responds in the ductile or.brittle manner.

Braced CantileverBrackeiand Trapeze Frame Supports. The presence of a diagonal brace in a support has'thevpotential of significantly increasing the pullout loads on anchorage when the support is subjected to horizontal motion. This is a.functi6n 6f the support geometric configuration,the realistic capacity of the brace, and the realistic capacity of the anchorage.

Non-ductile behavior is possible when the .brace reaction of horizontal load, plus dead load, has the capability of exceeding the primary support anchor capacity. If a brace buckles or has a connection failurebefore primary, Support anchor capacity is reached, then the support may be considered as ductile. Braced supports are subject to further horizontal load capability review with a focus on primary support anchorage.

UnbracedRigid Trapeze Frames. Trapieze frames constructed as moment-resisting frames, such as those with a'number of stiff cross-beam members welded to the two vertical supports have the potential of significantly increasing the pullout loads on anchor bolts when the frame is subject to horizontal motion. Non-ductile behavior is possible when the rigid frame anchor-.-:..

point reactions to horizontal load exceed the anchor capacity.u nbraced rigid trapezeTrames are subject to further horizontal load strength review with focus on anchora'ge.

Floor-mountedSu"ppor)-ts.Plastii havior of flosrmounted ipportsiay lead to structural instability. Ductility,-as def ned by these gUidelines; only~applies to suspen ed systems.

_6 or-moUnted supports are characterized as non-ductile, and are subject tifurther horizontal strength eiwwt focus on stability. '-

"RodHangerTrapeze Supports. Supports constructed ofi hre'ded steel rods~with fixed-enid connection detailsat the ends of the rods behave in aductile manner under horizontal motion;-

however, relatively -shorrods may U-ndergo very large strains due to bendingigmposed by horizonrtal !seismicmotion, at the fixed ends 6f the rods. Low',cycle fatigue~may govern response.

Rod hanger trapeze isiupporis with-short, ixed-endrods should be evaluated for low cyclefatigue effects. -X F-6

GuidelinesforLimited Analytical Review of Supports If a support is characterized as non-ductile or has questionable ductility, then its lateral load capacity should be verified, as discussed in Section F.5. If a support is characterized as ductile,

• and the duct system as a whole, depending on the functional goals, is judged capable of handling anticipated plastic deformation of supports, then it may be judged that no further lateral or

-longitudinal load check is necessary.

F.5 Lateral and Longitudinal Load Check A Lateral and Longitudinal Load Check should be performed for the bounding case HVAC duct supports that are-characterized as potentially non-ductile. The'Lateral and Longitudinal Load Check is in the form of an equivalent static lateral load coefficient.

S. If a support is non-ductile or has questionable ductility, then it should be analyzed for dead load plus a transverse acceleration of 1.0 times the Peak Spectral Acceleration (PSA) of the

-.in-structure response spectrum, at 5% damping, for the anchor point in the plant where the HVAC duct system is attached.

To evaluate a given support, transverse loads for the two horizontal axes should be applied, and capacities checked, non-concurrently. That is, two separate load cases should be checked.

' For example, one load case would be dead load plus loads due to north-south seismic motion, and the second would be dead load plus loads due to east-west seismic motion.

For'these loading conditions, judgment must load.be used to ascertain the tributary mass, or length

-of duct run, to consider foreach direciion of As a'general guideline, tributary length of duct for consideration for each direction of load should include one-half the length of duct to the next supports, on either side of the support being evaluated, that'resist load in that direction.

This general rule may not always apply, and tributary spans should be adjusted as judged necessary based on stiffness considerations for the duct systems.'

It is not required that the PSA always be used; this is intended to be*a "first screen" method.

If a reasonably accurate estimate of dominant mode response can be made, then the spectral acceleration associated with the frequency estimate may be made.,As appropriate, beam-on-elastic-foundation approximations or Dunkerley's equation approximations may be used

.(separately or togethier);Thes frequenicyestimati6n approaches are shown in Figures F-3 and F-4. When these methods are used, the basis for their applicability should be documented

wit the LARcalc'ulations.

-Thesiimple equivalent static load coefficieni method mayi be too conservative for very low wiih iong'drops from heceiling anchorage to theHVAC duct.The static ofrequencysupports coefflcient method predicts very high *connection bending moments in thesecases. In this case, thebending moment imposed onhtheceiling connection may be limited by peak seismic deflection and not seismic acceleration.: An alternative loading condition of dead load plus reaction forces due.to a realistic estimate for imposed Seisimic deflection may be used. Seismic deflectionimayybe calculated by using floor spectral ýdisplacementat a16ower bound frequiency estimate;i.considering'only single degree-of-freedom pendulum response of the support.

F-7

Guidelinesfor Limited Analytical Review of Supports w, IEID

.~III IKI LD Figure F-3 System Frequency Estimation using Beam-on-elastic-foundation Approximation For diagonally-braced supports.with ductile overhead anchorages, the load reaction imposed on the support anchorage during the Lateral Load Check does not need to exceed the buckling capacity of the brace or its connections. For diagonally-braced supports where the anchorage is not ductile, the portion of the lateral load that is not resisted by the brace should be redistributed as bending stress to the oveihead connection. The loads in the diagonal bracewill cause additional vertical and horizontal loads on the anchorage, and should be accounted for.

An upper and lower bound estimate should be used for buckling capacity of the brace, whichever is worse, for the overhead anchorage.There is considerable-variation in test data capacity for light metal strut framing connections. An upper boundtypes. estimate of 2.0 times the manufacturer's suggested capacities can be usedfor these connection

-Connections A and Bare-partially welded connection deiails. Partial Welds cannot develop the plastic moment capacity of the vertical member, and are considered.non-ductile.

• Connection C is the non-ductile rigid boot connection.

• Connection D is a rigid moment-resisting frame and should be checked for horizontal load.

S*Connections E and F are diagonally bracedand should be checked for horizontal load.

For IDuse SMACNA 2-inch corner section method-earetypi.ca -

qequv, f , =(0.87-.)(7t!2)(EI g/wLg

' : . s y s te m , . f , fl*D .. .: .

• 0.87 is forcompanion angle duct. Use 0.59 for other'joints in Figure 3-1.

F-8

Guidelinesfor Limited Analytical Review ofSupports

, wEID r-. ~~~~~~1 Figure F-4 Dunkerley's Equation Frequency Estimation Methodology For IDuse SMACNA 2-inch comer section method Ks, I are typical Wequiv W "

f, (1/27) (Ks gfWequiY) 4 12 fD = (0.871*)(c / 2)(EID g/w1w) /

2 2]"-12 fystem = [ +1/) D()if) 0.87 is for companion angle duct. Use 0.59 for other joints in Figure 3-1.

7 j.

F-9

Program: About EPRIR.

Nuclear Power, creates science and technology solutions for cEPRI the global energy and energy services industry. U.S.

electric'utilities established the Electric Power

• Research Institute in 1973 as a nonprofit research consortium for the benefit of utility members, their

• customers,. and society. Now known simply as EPRI.

the company provides a wide range of innovative products and services to more than 1000 energy-

- related organizations In 40 countries. EPRI's multidisciplinary team of scientists and engineers

--draws on a worldwide network of technical and business expertise to help solve today's toughest

-energy and environmental problems.

EPRI. Electrify the World

ýd C 2003 Electric Power Research Institute (EPRI).Inc.All rights q

.reserved. Electric Power Research Institute and EPRl are registered service marks of the Electric Power Research Institute. Inc.

"EPRI. ELECTRIFYTHEWORLD isa service mark of the Electric Power Reserch Institute.In. .-.

Printed on recycled paper in:h'e Linkeod States of1A0r 796 1007896 EPRI

• 3412 Hillview Avenue, Palo AltoCalifornia 94304 - P0 Box 10412. Palo Alto.California 94303

• UJSA 800.313.3774 *6S0.855.2121

• askeprl@eprl.com

• www.epri.com

Enclosure 14 Edwin I. Hatch Nuclear Plant Request to Implement an Alternative Source Term R. P. Kennedy Peer Review Comments on EPRI Seismic Evaluation Guidelines for HVAC Duct and Damper Systems

Peer Review Comments on EPRI Seismic Evaluation Guidelines for HVAC Duct and Damper Systems R.P. Kennedy February 7, 2004

1. Introduction The EPRI Technical Report.1007896 entitled Seismic Evaluation Guidelinesfor HVA C Duct and Damper Systems(Ref.1) provides an earthquake experience based approach for verifying the seismic adequacy'of HVAC duct and

• damper systems..,I.tis my understanding that Ref. 1 has not been subjected to a detailed review by an independent peer review panel in a manner-similar to that performed for other classes of equipment evaluated using an earthquake experience based seismic evaluation approach. Although not from an independent peer review panel, this report presents my individual independent peer review of the seismic evaluation -guidelines presented in Ref. 1.

The seismic: evaluation'approach recommended inRef. 1 consists of a two-step process. The first'steP-cvnsists of a detailed in-plant seismic Walkdown screening review of the HVAC duct systems to be evaluated. This review is to be conducted by'a Seismic Review Team (SRT) that consists of at least two qualified engineers that must mutually agree that the walkdown reviewedHVAC duct system hasPpassed the seismic screening so that it is eligible tohave its seismic adequacy verified by. the earthquake experience based approach:' Guidance for this seismic walkdown review is presented in Section 3 of Ref. 1.

For the second step, the SRT selects a bounding sample of HVAC duct systems and supports to be subjected to a-simplified analytical review. Details for .

.thisanalyticalreview are presented in Section 4,of Ref. 1. The simprlifed analytical approach presented in SectionA4 of Ref. I is very similar to*the Design-.

by .Rule ap~proach presented in Ref. 2 for -VAC duct systems and their -supports."

Ref. Z* as very.thoroughlyreviewed and accepted by an independent peer review.'

panel . .....

.The6 bove summarized two-step process is also Nery simiiarto the -

.earthquake xp*rience based approach developed by.SQUG andf presented in Section 8 of.Ref'.3;f r Cable and Condujit RiewaySystmsand t.hir supports.:'

Ref. 3 was also-very-thoroughly reviewed and accepted by an independent peer review panel.

RP'X" Structural Mechanics Consulting 28625 Mountain Meadow Road, Escondido, CA 92026

,(760)751-3510. 0 (760) 751-3537 JFax) email:' rpkstruct@earthlink.net

I served as chairman of the five member independent Senior Seismic Review and Advisory Panel (SSRAP) which provided considerable technical review and advice during the development of the SQUG (Ref. 3) approach for evaluating the seismic adequacy of 20 classes of equipment plus Cable and Conduit Raceway Systems and their supports. SSRAP (Ref. 4) unanimously endorsed the SQUG (Ref. 3) approach Tor use on existing components in existing nuclear power plants.

Furthermore, I served as a member of a four member independent panel established by the U.S. Nuclear Regulatory Commission to provide advice on the use of this earthquake experience based approach for the seismic qualification of new equipment, cable trays, and HVAC duct systems in new plants. In Chapter 5 of Ref. 5, this panel explicitly endorsed the earthquake experience based Design-by-Rule approach proposed in Ref. 2 for HVAC ducts and their supports. The independent panel stated:

"The Panel fully supports the idea of 'design-by-rule' for HVAC ducts. This requires simplified design procedures with minor computational needs. The Panel observed that, in the past, significant efforts were expended for nuclear power plants to analyze and design HVAC ducts. The lessons learned from past practice and experience, if incorporated in the new design rules, will significantly reduce cost without sacrificing confidence in performance.

Therefore, the Panel not only endorses a new design approach but also encourages it."

Therefore, even though the detailed material presented in Ref. 1 has not been reviewed by an independent peer review panel, the overall approach has been reviewed and endorsed by independent peer review panels.

My review of Ref. 1 has heavily concentrated upon whether important aspects of the SQUG approach (Ref. 3) for Cable Raceways and the Design-by-Rule approach (Ref. 2) for HVAC systems and their supports have not been incorporated into Ref. 1.

2. Overall Conclusions In general, I find the seismic evaluation guidelines for HVAC Duct and Damper Systems and their supports presented in Ref. 1 to be excellent. However, I believe that Ref. 1 is deficient in certain details that are included in either Ref. 2 or 3. These minor deficiencies are discussed in the remainder of this report. I recommend that these minor deficiencies be corrected. Each minor deficiency van be easily corrected and will have very little overall impact on the use of Ref. 1.

3.: Minor Deficiencies in Ref.: I1

. 3.1 Limits on"Applicability.

In Section 2.1 of Ref. 1, it is stated that the guidelines are applicable to any HVAC duct and dampingsystem at any elevation in a plant where the nuclear p lantfree-field ground motion 5% damped seismic design spectrum does not exceed the'Seismic Motion Bounding Spectrum of Ref. 1.

I do not consider this limit to be sufficient. HVAC duct systems can be supported at very high elevations in a variety of buildiigs where the in-structure-response-spectra (ISRS)rcanibe much higher than the free-field ground motion. I don't believe that the experience data adequately covers this situation.

Section 3.1 of Ref. 2 -restrits its proposed Design-by-Rule method to situations where the horizontal zero period accelerationf(ZPAh) at the HVAC support anchorage does not exceed 2.0g. J -doubt that very many situations exist where.ZPAh exceeds 2.Ogwhen. the free-field spectrum is less than'the Bounding Spectrum. Even so*I strongl, believe that the ZPAh" ess than 2.Og limit is an important additionalilimitation that: should be included in Ref. 1. I doubt that it can be demonstrated that anhyof the HVAC duct earthquake experience data base included situations where ZpAh exceeded 2.0g. Without a significant amount of

-suchdata, the 2.Og limitation is needed.

3.2 Duct Span Lengths'Bet'een Vertical Supports Section 3.2.1 of Ref. 1suggests 1 that tables of allowable. duct spans and maximum cantilever lengiti foýrvarious duct sizes be devloPed prior to the seismic walkdown screening of duct systems., Dexelopment of these allowable

..span tables should be a prewalkdrwn rquirementfand not just a suggestion.

"Section 3.2.1 refers to ppendiX, C as an example:of how a tabulaion of allowable spans can be develope. Here again, Appendix C Ashuld be a

.requirementandnit just anexample. iFurihe ore:itiw0uidbe helpful to have an exaiimple appli*ation of Apýendix C with an-examjlesei olef'screening"tables fo-r -

..do9me realistic situatn -

in additionh,-some ýupper limit on vertical support spans should be

.established. ;:This ilmit *should be based pi spahns 6bserved in the earthquake.

.experience database.Ref_. 2 whic*hi*aS based oiithe'experience data in Ref. 6 established thefollowing limits on support spans for the Design-by-Rule method:

1. Duct support to support spans should not exceed 15 feet.

2. Supports should be provided within 5 feet from fittings such as Ts and

• Ys in each branch of the fitting 3.. Duct cantilevered length (beyond end of last support) should not exceed

• 6 feet.

These limits are intended to place the duct spanswithin the limits of extensive earthquake experience data. Unless significant amounts of new earthquake experience data can be used to justify higher span limits, I believe that these limits from Ref. 2 'should be incorporated into Ref. 1.

3.3 Seismic Interaction Review Sections 3.4 of Ref. 1requires that the SRT conduct a'seismic interaction review. However, very little guidance isgiven in Ref.- I for the Proximity Interaction review. "Akey element of this review is to estimate the seismic induced displacement of both the duct systems and of any adjacent "itemthat might damage the duct upon impact. .'Some guidance on how to make these.displacement estimates for the duct system should be included in Ref. 1. At least some limited guidance is presented in Section 3.3 of Ref. 2. This guidance could at least serve as a start for guidance in Ref. 1.

3.4 Vertical Capacity Check Section 4.5.1 of Ref. I1requires a Vertical Capacity check of the vertical supports. Further guidance is-given in Appendix F. This check is to verify that the duct supports lie within the range of duct support capacities within the earthquake experience data base.

451.1section dealsolily with the metal frimine.Section 4.5.3;deals with anchorage. No Yertical Capacity check isdreuird inSection4.5.3. It needs to made very clear in Section 4.5 that the Vertical.Capacity -check appiies b6th to the metal frame and .the ancorage. Appendix F does properly include anchrage in

this'check. Even so, it should be made clearin' Sectioh 45.- .

The fourth pargraph'on Page+4 '-4f Section,4.51 -statesth'atit is .

"peritted t exceed AISC allowable stresses in certain siuations.:HoWever, it is myiunderstanding that essentially all6 f the -successful duet suppqrtsin the Ref. 6 earthquake-exipe nen~ce -data base .pa-ssed the Vertical Capacity check at AISC' allowablestress levels. For this reason,..Section 6.1.2 of Ref.2 requires that the Vertical Capacity check to be passed at AISC allowable stress levels. Unless it can be demonstrated that a significant number of the successful duct supports in

the earthquake experience data do not pass the Vertical Capacity check at AISC

. allowable 'stress levels, I -strongly recommend that the Vertical Capacity check be limited to the AISC allowable stresses.

Ref. 1 does not clearly delineate the limits of the Vertical check. For ductifalue nlyca failure modes, only loadsse primary stresses from verticalloads ne'dtb ed tO be included:, Stresses relieved by small displacements do not have to be included in the Vertical Capacity check. Some useful guidance on this topic is given in Section 6.1.2'of Ref. 2.

The Vertical Capacity check is made for a vertical load Pv defined by:

PV = Fv * (Dead Load) (1) where Fv is a vertical load increase factor defined in Ref. I by:

Fv5.Og . (2)

The independent peeri review panel which reviewed Ref. 2 did not consider FV from Eqn. (2) to be adequate f6r high'seismic lateral forces. As a result Ref. 2 uses:

Fv Greater[5.0g, 6.0(ZPAh)] (3) where ZPAh is the zero period acceleration at the support anchor. The net effect of this change isto increase Fvlwhen ZPAhexceeds 0.83g.

In my opinion, Ref. ,Ishoulduse Eqn. (3) to define Fv unless it can be demonstrated that a significant number of the successful duct supports in the data base will not pass'the 'Vertical Capacity check when Fv from Eqn. (2) is replaced

- byFv from-Eqni:(3).

S

• i3.5 Peer ReviewPRequire-ment The eahquake experience based sisemic .valuation approaches, presented

• .in:Refs.lthrough 3 rely heaviIy-on the judgment and experience of theSRT. This ju~dgme~nt.andexpeeneeis usedin lieu bfexteinsive analyses. 'Asa* result, both the'SSRAP'r*p6rt (Ref.;4) and the`SQUG approach;(Re f.3) require independent peer review.of te judg'ments andtconelusions made bythe SRT as well as a

'sampling review of the limited analytical evaluations.

However, Ref. 1 does not require this independent peer review. I consider, this to be a fatal deficiency in Ref. 1 that must-be correcied., Independent peer review is an integral part of an experience based appioach.:`i.-

4. Final Comments I fully concur with and support the use of theRef.l 1 seismic evaluation guidelines for HVAC duct and damper systems and their supports so long as the, minor deficiencies identified in Section 3 are corrected. In the meantime, I suggest that users of Ref. 1 should implement the changes -recommendedin' Section 3 for their plant.. specific use. I don't believe Secioni3ficatly . ev that th any an'f ofthese changes

• 'ech will swl significantly affect the usefulness of Ref. 1.

References 1.- EPRI, Seismic Evaluation Guidelinesfor HVA C Duct and DamperSystems, Technical Report 1007896,April 2003

• 2. ARC, Advanced Light Water Reactor (AL WR) First-of-a-Kind-Engineering (FOAKE)Project on Design Conceptsfor HVA C Ductingand Supports, April 1995

3. Seismic Qualification Utilities Group (SQUG), Generic Implementation Procedure(GIP)for-Seismi'c Verification ofNuclearPlantEquipment, Revision 3A, December 2001.

4. Senior Seismic Review Advisory Panel (SSRAP), Use of Seismic Experience and Test Data to Show Ruggedness ofEquipment in Nuclear PowerPlants, Sandia Report SAND92-0140, Part 1, 1992 5.Bandyopadhyay, K.K-iKana, D.D.,Kenfiedy,- R.P., ,and 'Schiff, A.J., An 5.

• • . EvaluationýofMethodology.foriSeishicQualfldtin of Equ ent Cable .

,Traysand Ducts in ALWR Plants' by Use of Experienet:Data, NURE G/CR-6464 ; ~prepared for U.S Nuclear - e-,ator Commission, July 1997 -,:

6. -ARC,*Advan~ed Light Watie-Reaco6tio(LWR*)Firsi-of-.*Kind-Engineering-. '.'

(FOAKE)Projecton the PerforianceofHVACiDucts and Supports in' Earthquakesaind.Tests, .April 19

Enclosure 15 Edwin I. Hatch Nuclear Plant Request to Implement an Alternative Source Term R. P. Kennedy Peer Review of the Hatch Unit 1 Seismic Verification of the Turbine Building Exhaust Ductwork

Peer Review of the Seismic Verification of the Turbine Building Exhaust Ductwork for Hatch Nuclear Plant Unit 1.

by R.P. Kennedy.

November24, 2004

1. Introduction The Hatch Unit I Turbine Building exhaust ductwork was not originally seismically designed. This ductwork is required to be seismically evaluated as part of a review for-control room habitability. The ductwork was evaluated using the 'earthquake experience based methods documented in EPRI Report 1007896 (Ref. 1) plus my previous peer review comments on Ref. 1 as documented in Ref. 2.

The experience based seismic evaluation'of the Turbine Building exhaust

• ductwork is presented in Refs-.3 through 5 which I have reviewed: Reference 3 presents the overall approach and resulisof the experience based seismic evaiuation. This' evaluation relies heavily on the professional experience and judgment of the two seismic capability engineers(P.'Baughmnand D.P. Moore) conducting the detailed seismic walkdown of the Turbine Building exhaust ductwork.

Ref. 4 developsHVAC duct allowable span tables for the various floor' levels and duct sizes for Hatch Unit 1 Turbine and Control Buildings. These'span tables were developed prior to the seismic walkdown documented in Ref. 3. The -seisnicwalkdown team assured themselves that alliscreened ductwork satisfied both the ailowable spans tabulated in Ref. 4 and the experience based limits that I previously re'6mmended in Ref.

2. Generally the Ref. 2 limits iontiolled. All ductwork satisfied the allowable-span limits of both Refs. 2 and 4.

During theseismic walkdownbounding eases for both the ductork and its "suports *wereselected for further'analytical review. These analytical rviews aire presented in-Ref. 5... -

2.-- Conduct'of Peer-Review.::-

The initial 'step of my peer review*was to review Refs 3 through 5 in order to -

-assess:

• :1,:"!;Compliance with the requirements ofRefs. Iand-2 2 Thooughness"of the documentation'ofwalkdown results

3. 'Searching for deficiencies in the bounding calculations rPX~ _______________Structural Mechanics Consulting 28625 Mountain Meadow'Road, Escondido, CA 92026 (760)751-3510 * (760) 751 -3537 (Fax)email:'rpkstruct@earthlink.net

My findings on the Refs. 3 through 5 documentation are presented in Section 3.

• Next,I spent one -full day at Hatch Unit Lin a peer r*view walkdown of the Turbine Building exhaust ductwork. I was accompanied on my walkdown by the original walkdown team of P. Baughman and D.P. Moore. During my walkdown I extensively discussed with the walkdown team their basis for decisions and I independently assessed these decisions'. The primary purpose of my walkdown was to search for any situations where:

-1. I might differ with the walkdown team's judgment

2. Potentially bounding cases for ductwork and supports were not evaluated*

inRefs. 5 My walkdown reviewed :100% of all Turbine Building exhaust ductwork not within high radiological areas i.e., condenser bay.. This sample constituted more than 50% of the ductwork reviewed by the walkdown team. I did not enter the high radiological areas. However, the walkdown team made extensive photographic documentation of the ductworkan'd sutpports in these high radiological areas. I reviewed this photographic d6oumentation.- I didn't see any ductwork or supports in these photographs that diffe red meaningfully from ductwork and supports which I did walkdown.

A summary of my pebrrieview walkdown and its findings are presented in Section

4. Lastly, my peer review conclusions are presented in Section 5.

3. Peer Review Findings oCnceming Refs. 3 through 5 The documentation presented in Refs. 3 through 5 fully complies with all requirements of Refs. I and 2. The documentation is both technically correct and complete. The Screening and Evaluti6n Worksheets very thoroughly document the

.ducting and supports. I fouind no6deficiencies-in the bounding calculations. I support all conclusionis'

*.

.4. Sumhmary and Findings of Peer Revie Walkdown - - .

atachedIThe table'summaizes the du&ting and -supports that I inspected. I found i-.th documentatio'noffScreening this ductworkin the and Evaluaiion'Worksheets of Ref. 3 to *e excellent. I didn't find any'situation where I differed from the walkdown team's judgment.

Under' ItemC (see'attached table), I noted that the fans on the reactor building roof are mounted on vibration isolators. I carefully inspected these'isolators.-My judgment is that the spring within these isolators is sufficiently entrapped laterally and longitudinally that it cannot laterally collapse. Therefore, t0oncur with the walkdown team's judgment that these isolators are seismically acceptable.:

The Item E duct passes through a non-seismicallydeisgned concrete block wall.

Section 6.1 of Ref. 3 identifies this situation as a potential seismic interaction outlier that needs to be resolved (outlier No. 4). 1 concur. D.P. Moore orally described a proposed resolution of removing concrete block around the duct sufficiently such that collapse of:

" .,theconcrete block wall would not impact the duct. Iconsider this approach to resolving

'"this potential outlier issue to be acceptable.

I f6und two support conditions that were potentially of concern to meand were hot covered by any of the Ref. 5 bounding analyses. These two support conditions are briefly'discussed.

-A long sectionofthe 44-inch by 40-inch *bu in Item A is vertically 21feet supported by six 2.5x25 inch angles acting as compression legs.'The longitudinal direction insupported length of these legs is 14.5 feet whilethe lateral unsupported

-lenigth is 7.25 feet.'The KL/r ratio for these legs is large *o that their compressive buckling capacity is small.-However, the duct weight is also1small. Based upon a quick calculation, I concluded that these supports are acceptable.

In a number of locations, smaller ducts are hung from I-inch by 3/32-inch vertical straps. At supports, in some cases, these straps are bent.90 degrees to make an angle and are bolted with one 1/4-inch bolt tothe support. This bolt is eccentric from the strap by as much as %-inch. I was concerned about the bending stress introduced into the strap at the bend by this eccentfricity.--From my walkdown, the bounding case was at the item D14-_

inch diameter circular ducts where supports were up to 8-feet apart. However, because this 8-feet section of duct weighs only about 40 lbs, I judged this strap hangar eccentricity to be acceptable.

Although, I probably would have included bounding calculations' for the above two cases in Ref. 5, I 0onsider the judgmefint bf the Walkd*wn team in selecting boinding:.

cases to be acceptable.. -

5.verllConclus'ion -

• The experience based seismic'evaluation of the Hatch Unit-1 Turbine Building.;:

exhaust ductwork presented in Refs. 3 through 5 iseekllent. This evaluation fully coiimplies with tihe evaluation iequirements of Refs. l*and2. The walkdow*n team perfo6med a very thorough and competent evalu ti6n i didn't identify any open issues. I fully concur with the ýcbnclusions of Ref. 3.

6. References

1. EPRI Report 1007896 "Seismic Evaluation Guidelines for HVAC Duct and Damper Systems," prepared by ABSG Consultingi.Final Report, April 2003.

2. R.P. Kennedy, ".Peer Review Comments on EPRI:Seismic Evaluation Guidelines for HVAC Duct and Damper Systems," February 7,2004.

3. ABS Consulting Report 1302241-R-001, "Hatch Nuclear Plant Unit 1: Seismic Verification of theTurbine Building Exhaust Ductwork,". October 6, 2004.

4. ABS Consulting Calculation 1294895-C-001, "Hatch Nuclear Plant Unit 1:

HVAC Duct Allowable Span Tables for Seismic Walkdown," February 20, 2004.

5. ABS Consulting Calculation 1302241-C-002, "Hatch Nuclear Plant Unit 1:

HVAC Duct and Support Analytical Review," October 6, 2004.

Hatch Unit I Turbine Building Exhaust Ductwork Inspected During Peer Review

- ~ - -- r Item- lev. I Building Ductwbrk Inspected - ý' .'

ItmEev uidn DutokIrnce A -164 Turbine Bldg. H-16050 44x40 Duct between T9 and TI0.5

" "___Control Bldg. H-16194 between TA and TD B 164 Control Bldg. H-16194 20x20 Duct between T10.5 and T13 at

___ TA C 185 Reactor Bldg. Roof H-16232 Ductwork from Control Bldg. to filters

-130 and from filters' to stack D 130 Turbine Bldg. H-16048 14-inch Circular Duct between TI and

__T7 between TA and TE E 112 Turbine Bldg.. H-16047 12x12 Duct between T2 and T7 along

___ _ ____ ____TA F 112 Turbine Bldg. H-16047 !8x20 Duct in'RBCCW. Heat Exchanger Room between T7 and T 10

_ ._ ._ between TB and TD, G"__* 112 Control Bldg. H-16053.* 16x14 Duct between T9 and T12 along TB.5

.-5-