ML20153D400

Forwards Response to NRC 851022 Staff Evaluation Rept on LaSalle Single Angle Member Design. Majority of Concerns Inapplicable to Specific Design of HVAC Hanger Frames. Methodologies Will Be Incorporated Into FSAR

ML20153D400
Person / Time
Site:	LaSalle
Issue date:	02/13/1986
From:	Farrar D COMMONWEALTH EDISON CO.
To:	Harold Denton Office of Nuclear Reactor Regulation
References
1293K, NUDOCS 8602240136
Download: ML20153D400 (27)

Text

,

t..

1 N CommonweaHh Edison g

) One First N:tional Ptrts Chrctgo, Ilknois -

O' Address Reply to: Post Office '3ox 767 -

J

.- Chicago Illinois 60690 t

, February 13,~1986 3

j l

\\

Mr. Harold R. Denton, Director i

4 Office of Nuclear Reactor Regulation-U.S. Nuclear Regulatory Cosmission j

Washington, DC.20555 3

Subject:

LaSalle County Station Units 1land 2

}

Design of Single-Angle Members

}

License NpF-11 and NpF-18 i

NRC Docket Mos. 50-373 and 50-374 i

j

Reference:

Letter dated October 22, 1985 from W. F.

Butler to D. L. Farrar regarding LaSalle County Station Design of Single Angle Members.

I f

f

Dear Mr. Denton:

1 In October 1985, the NRC transmitted to Commonwealth Edison a 3

i report entitled " Staff Evaluation Repoi-L on LaSalle Single Angle Member

.,j Design." The report had four concerns regarding:

1) allowable flexural stress i

2) useof0.95f(ricaxesforresolutionofmomentsandforflexural as ths allowable flexural stress for SSg loads i

3) use of geome i

stress calculaticns 1

4) interpretation of AISC interaction equations 1.6.1 and 1.6.2 Conunonwealth Edison has thoroughly reviewed the design procedures 3

]

used for HVAC hangers by our architect engineer, Sargent & Lundy. We ha're concluded that the majority of the concerns raised by the NRC are not applicable to the specific design of the HVAC hanger frames constructed of i

single angle members.

To. gain further assurance of the validity of the design methodology, commonwealth Edison has obtained independent reviews by two prominent steel design engineers, professor T. V. Galambos and'Dr.

]

Georhard Hasijer.. The NRC Staff conclusions were based, in part, on correspondence from these two experts. 'However,'the correspondence between these individuals and the NRC regarding Sections 1-5 and 1-6 of the AISC

.i specification was general in nature and did not pertain'to the specific j

features of the Cosmonwealth Edison HVAC hanger frame design. More recent coinnents of both reviewers contained in the attachment,' confirm that these

]

~

general observations did not pertain to the special case of single angle hanger frame design.

\\

l gangg %

gg

[

P

I H. R. Denton February 13, 1986 In conclusion, cosunonwealth Edison has reaffirmed that the' design procedures used for HVAC hanger frame single angle members is based on sound engineering principles, employs the latest research results on the rubject, and meets all licensing connaitments.

As a result of this determination, the unbraced length, the slenderness ratio and the single angle member design methodologies will'be incorporated into the LaSalle Final Safety Analysis Report at the earliest opportunity.

Please direct any questions you may have concerning this matter to this office.

Very truly yours, D. L. Farrar Director of Nuclear Licensing im Enclosure cc: Region III Inspector - LSCS A. Bournia - NRR M. Parker - State of Ill j

1293K i

.m

.u.

,1 tc c'; '

(; /[.

s t v,

.1

' k;l 5

• s t;

's

RESP HSE T0:

NRC STAFF EVALUATION REPORT r

DN i

2,.

i LA, SAP.LE SINGLE-ANGLE MEMBER DESIGN

.g

+

i.,

3-.

'S Commonwealth Edison _ Company January 30, 1986 u

f

,Y" s

~

l LS:

a s

l CONTENTS I.

Summ e ry a nd Co ncl u s i o n s................................... Pa ge 1 II.

All owa bl e Bendi n g Stresse s................................ Pa ge 3 A.

Aus tral i an Standa rd AS125 0-81........................ Pa ge 3 B.

Allowa bl e Stre s s for SSE Load s....................... Pa ge 4 III.

Design Calculations....................................... Page 5 A.

Reference Axes.......................................

Page 5 1.

Han ger Frame Members............................ Pa ge 5 2.

Longi tudinal Diagonal Brace Members............. Page 6 B.

Interpretation of AISC Interaction Equations......... Page 6 1.

General.........................................

Page 6 2.

Unbraced Hanger Frames.......................... Page 7 -

3.

Braced Hanger Frames............................ Page 7 4.

Longi tudinal Diagonal Braces.................... Page 8 C.

C V a l u e............................................. Pa g e 8 b

D.

Dus t Pl a te Re s trai nt................................. Pa ge 8 IV.

S E I SH ANG Ve r i fi ca t i o n..................................... Pa ge 8 Attachments 1.

Letter from Dr. Geerhard Haaijer - 1/15/86 2.

Letter from Professor T. V. Galambos - 2/20/85 3.

Letter from Professor T. V. Galambos - 1/9/86 1

l i

f

'~ 1 g

t

.o I.

Summary and Conclusions i

i In February 1985, Commonwealth Edison Company (CECO) met with' the NRC staff to discuss the design procedures used for HVAC hanger frames I

constructed of single angle members..

The basic item covered at. the meeting dealt with the use of 1/t ratios in excess of 270.

At the conclusion of the meeting, it was agreed that 1/t ratios of up to 900 could be used, based upon the CECO. design methodology..Following the meeting, the NRC requested that sample calculations for the design of hanger frames be submitted ~ for review.

These calculations were transmitted to the NRC in late February 1985.- In October 1985,. the ' NRC transmitted to CECO a report entitled, " Staff Evaluation Report on La Salle Single Angle Member Design."

The report addressed the following concerns:

1.

Allowable Flexural Stress the NRC staff has indicated disagreement with the method of determining the allowable. flexural stresses for single angles based on the Australian Steel Code, l

AS 1250-1981, and the research work of Leigh and Lay. - The staff l

report further indicated that the Australian procedure is not.

currently used in the United States, is not' contained in any.

l American code, and has not been accepted by the NRC or the AISC..

i l

2.

Use of 0.95F as the Allowable Flexural Stress for.SSE Loads - the j

y staff indicated that 0.95F ' may exceed _. the ~ critical buckling-y stress for short and. intermediate length members.

I Page'11 1.

m s'

3.

Use of Geometric Axes for Resolution of Moments and for Flexural Stress Calculations the report indicate'd that the use of principal axes is required in all cases for making the required calculations.

4.

I_nterpretation of AISC -Interaction Equations 1.6.1 and 1.6.2 - the NRC implied that the AISC interaction equations may not have been interpreted correctly.

\\

l l

CECO has thoroughly reviewed the design procedure used for HVAC hangers and has concluded that the majority of the concerns raised-by the NRC are not applicable to the specific case of the design of the HVAC hanger frames constructed of single-angle members.

Items such as l

eccentricity effects, shear stresses due to torsion, and the use of principal axes for the design of longitudinal diagonals were considered l

applicable.

These items were reviewed and - their effects; were determined to be insignificant; however, CECO has elected to revise the computer program, SEISHANG, to incorporate these. -items for future designs.

l-To gain further assurance of the validity of the design methodology, CECO has obtained independent reviews by two eminent steel design-l engineers, Prof. T. V. Galambos of the University of L Minnesota and l

l Dr. Geerhard Haaijer Vice President and Director of Engineering of the AISC. The NRC staff conclusions were based, in part, on correspondence-from Prof. Galambos and Dr. Haaijer which was attached to the staff l.

re po rt..

It should be noted -that the correspondence. between these l-Page 2 l

1 i

q,y

, o individuals. and tihe NRC : regarding Sections f 1-5 'and[1-6 of. the AISC -

' Specification was" general. observations and 'did - not pertain to ' the unique features of the : CECO single-angle.HVAC hanger f.rame :. design.

P Comments ~ of both : reviewers, contained in.: Attachments 'l through: 3. -

~

confirm that these general observationsf do not pertain ~ to thet special i

case of single angle hanger frame design.

In conclusion, CECO has reaffirme'd.that L the ' design' procedure of the HVAC angle members is based on sound: engineering principles', employs the latest research results.on,the subject, and. meets-all licensing commitments.-

II. Allowable Bending Stress A.

Australian Standard AS1250-81-'

As indicated in AISC's attached' letter, the AISC Specification is i

\\,

silent on the design of single' angle-members subject to axial and l

t bending loads.

In particular, the Specification does not; provide an allowable bending. stress for.such members.-

However, the

+

(

Australian Steel Code- 'does contain equations. that: provide _-- an -

-i allowable flexural stress.

At this time, the' Australian Code'and

~

l research work are the -. definitive o documents oni the subject'. of:

angles loaded in bending.

TheJattached letter from ;the-AISC I

affirms. that - the.' Australian, methodology will be ' incorporated :in :

the forthcoming: AISC. LRFD Manual... In addition,' the AISC Lpublished :

r 1

]

. the results in11ts " Engineering Journal". in 1983; furthermore, Lthe.

i

. 'c c

Page;3

.*^- -,

y

@l y

Institute presented the results:in a series of seminars on steel

~

design held across the country beginning in late - 1983.

The results were contained in the seminar notes and were submitted.to-the NRC on February 22, 1985.

The fact that the AISC has chosen to publish the results in its " Engineering Journal" and in seminar -

notes is a definite' indication of the ~ approval of the Institute.

Finally, Prof. Galambos has also endorsed the = Australian Steel Code and the research of Leigh and Lay dealing with the critical lateral-torsional buckling of single angles in bending.

In the CECO design procedure an allowable stress of 0'.6F has b'een' y

conservatively used for angles: for OBE loads.-

The maximum allowable for normal loads permitted by the Australian Code. is 0.66F. Prof. Galambos concurs with a maximum allowable of 0.66F y

y for b/t ratios in the compact range, and with a reduction below 0.6F only when the b/t ratio exceeds 13.

All CECO HVAC. hanger y

frames utilize angles with _b/t ratios less than'12._ cConsequently, the allowable of 0.6F is - below the ' allowable' specified in. the y

~

Australian Code, and -is equal to the, reduced : value suggested by -

Prof. Galambos for b/t ratios between 11 and 13.

o B.

Allowable Stress for SSE Loads The design allowables -available Lin the Australian Steel' C' ode 'are,

-as is' the case in most design codes, intended for use with normal:1 9f 3,.

loading. situations.

Under the Extreme Environmental; load;:J2f

' /;n my m

Page 4 n;

.i conditions, tne SRP permits an allowable stress increase of 1.6 times lthe normal.

In ' addition, CECO has _ chosen to limit the allowable stresses-in the. Extreme Environmental case (SSE) to 0.95F, as documented in the FSAR.

y Both the AISC and Prof. Galambos agree that this stress limit will not exceed the critical buckling stress of single-angle members in the short and intermediate 1/t range.

In fact,.the shape factor that ' defines the ratios of. the ultimate moment capacity.to yielding capacity of angles is as high as -1.75 to 1.80.

Consequently, there is sufficient safety margin against angle bending capacity at 0.95F under SSE leads.

y III. Design Calculations A.

Reference Axes i

1.

Hanger Frame Members In single-angle hanger frames, the' bending moments arise from seismic displacements; consequently, the loads are applied to the angle members at the joints.

The joints, as wel_1 as'the i

{

support points, provide restraint to the angles normal to the-

,i

-j direction of the applied load.

Both. -the AISC and i

Prof. Galambos agree 'that when - such -lateral' restraint is present, stress calculations in single-angle members. can be-made-on the ba s i s

of-- the 1 simple flexural formula.

.t Page 5

^

Consequently, 'the use - of the geometric axes for' stress

. calculations rather than the' use of-principal axes is appropriate for this particular ~ case of single angles loaded as described previously.

For members subject.to axial loads and bending, the interaction calculations may also be based on the geometric axes rather than principal axes.

2.

Longitudinal Diagonal Brace Members In the case of the longitudinal diagonal brace members, the maximum moment, which is due solely to the member self-weight excitation, occurs at the mid-span of the member.'-

CECO concurs that for this situation the use of the principal axes is appropriate.-

CECO has verified that the - longitudinal diagonal members, using the principal axes, meet: the design requirements.

In addition, the SEISHANG computer program Lis being revised to utilize.the principal axes for the design of these mer.bers.

B.

Interpretation of AISC Interaction. Equations.

1.

General i

CECO has correctly interpreted and. applied ' the interaction H

equations of-Section 1~.6 of the ' AISC Specification.

The 1

purpose of equation ~ 1.6-la is to insure overall' stability of

~

the member in question, while ~ equation 1.6-lb is for a check' Page 6

of strength against _ yielding.

Where-axial' load -is small-(i.e.,- fa/Fa < 0.15) AISC equation 1.6.2 may be applied as an i

alternate but conservative criterion for checking member interaction.

2.

Unbraced Hanger Frames The members of longitudinally, unbraced hangers are subject 4

to axial load and uni-axial bending..

Consequently, when evaluating the capacity of the member according to -the-interaction equations of Section 1.6, calculations are made' for the point of maximum moment in the Jspan of each member.

For these hangers, the angle memb'ers have slenderness ratios such that the yielding criterion (1.6-lb), and not: the stability criterion (1.6-la),' governs-the design.

J 3.

Braced Hanger Frames Angles in the longitudinally braced ' hangers are subject to axial loads and bi-axial bending.

Since - the ' bending moments -

are induced by seismic displacements,Jand since the maximum moments in two coordinate directions result from seismic

]

- 1 motions in the respective two orthogonal directions,. the

- l maximum' moment in each coordinate direction -will ' not occur-simultaneously.

Therefore, consistent with i the ' ~ SRP,: the :

interaction effects may be calculated ion _ af probabilistic

'l basis through an application of combination by. square ~ root of-

-Page.7.

i H

y' q

r

+wg p

g w.

.p to 4'

v-m g

g p

e-+:-

the - sum of the' squares: (SRSS).

CECO has verifled that interaction values calculated on this~ basis are enveloped by the' interaction values computed in the initial design for the~

l HVAC hanger members.

l l

l 4.

Longitudinal Diagonal Braces l

braces s' tisfy the Ceco has verified that the longitudinal a

interaction equations of Section.1.6 when the principal axes l

l are used as the basis of stress calculations, l

l I

j-C.

C Values b

l l

Values of Cb greater than 1.0 were not employed in the design of HVAC hangers.

The value of 1.1 was only used in the. sample calculation submitted to the NRC in February 1985.

D.

Duct Plate Restraint l

1 l

~

l The restraint effect of the duct plate was not assumed in the 1

design of HVAC angle frames.

l 1

IV. SEISHANG Verification l

All designs of HVAC hanger members are performed by Sargent' & Lundy's l

proprietary _ program, SEISHANG.

Four specific items regarding the l

program were cited in the NRC report as requiring further. verification:

l c

Page-8 l

1.

' Verification that closely spaced modes are combined in accordance with R.G. 19.2 SEISHANG is a general purpose ' program.specifically designed for analysis and design of hangers.

Three methods to combine modal responses are available.

These methods are:

a) square root of the sum of the squares method, b) the - double sum method, and c) the absolute double sum method, which combines closely spaced modes in accordance with R.G. 1.92.

2.

Verification that member global stiffness matrices include - the transformation from local principal axes to global axes, including the eccentricity effects due to end shear and axial forces (C.G.-

to shear center torques).

}

SEISHANG is a beam finite element program.

The member global stiffness matrices are obtained by transformation of the member stiffness matrices in the local-member-coordinate system to the global coordinate system. The local member. coordinate axes may be defined to be aligned with the geometric or the principal axes of the member cross section. -When the member local axes are aligned with -the geomitric axes, the cross-sectional properties for-those

.i members are specified about the geometric axes.. Similarly, when i

the member local axes are -aligned with the principal axes, the cross-sectional properties for that member are specified about the principal axes.

A. new option will be incorporated ' to permit page 9

e member force and stress calculation in the member ' principal axes l

even when the member local coordinates are aligned with the-member.-

geometric axes.

1 j-

~

l The program currently does not. compute the ~-eccentricity effects

^

c'ue to : end shear and a'xial. force (C.G. to shear center effects);

because they are insignificant.

This - is because of the small member size used to fabricate hangers coupled witn the; small; axial.-

i l'

and shear forces which ars applied.

A survey of typical 1 hangers,

)

incorporated, in which the moments due to eccentricity were 1

confirmed this judgement. The program, however, is being modified; i

to add these rather insignificant effects.

4 1

3 3.

Verification that shear stresses. due to torsion and 1 warping:

effects in WF and channel beams are evaluated.

t l-b Shear stress due to torsion and warping effects.in WF and channel beams were not being evaluated in the program because.for the q

j hanger geometry and the system configuration being ~usedito support j

2-

~

cable tray and HVAC duct.

The individual hanger members ~are not subjected to significant torsional loads.

The hanger; behavior 1s 1

j planar and the member forces are predominant 1y' moment and axial'

~

load, t

!~

A general purpose stress checking. option has:been added that gives a

the user the option to include the torsionalsand 4 restraint. of; warping effects in stress evaluation.when desirable.:

Page50

'.~,,,

<,y of s

1 13

-4.

Verification of

1 program -

1 All -versions of ~ the SEISHANG program 'have been vElidated in:

c accordance with the S&L General' Quality. Assurance Manual'.and the-

[

S&L General Office Procedures that :are in effect' at thectimeiof l

-validation.

Validation of the: program is performed each time the progra'm is modified for-'.any ; reason.. This validation consists? of one or both of.the following:

4 r

E a.

A set. of standard validation ' problems are-run on._.the ? new I

version. - The new results are chedked against corresponding j'

results from the existing program version to validate all the T

existing options of the program.

t b.

- For new options added, new validation problems a're set upland:

i

?

the' program results for' these problems are validated against -

~

r hand calculations and/or results obtained - from 'another l

irdependent and validated computer'. program 1with.. 'similarf J

capabilities.;

i j

The ' validation for each program version-is'_ documented.in calculation books and filed with the Computer Software L'ibrary.

i-l-

p Page 11'-

4 r

4

s ATTACHMENT 1 Letter from Dr. Geerhard Haaijer 1/15/86

AMERICAN INSTITUTE OF STEEL CONSTRUCTION, INC.

The Wng!ey Budding 400 North 1Achigon Avenue Chicago, lihnois 60611 4185 Q12) 670-2400 GEERHARD HAAUER, PhD.

w. %

Orector ce tyre **m January 15, 1965 Mr. Thomas G. Longlais Head, Structural Engineering Dir.

Sargent & Lundy Engineers 55 East Monroe Street Chicago, IL 60603

Dear Tom,

Your letter of December 23, 1985 addresses several ques-tions related to the design of single angle supports, which were raised by the Nuclear Regulatory Commission (NRC).

Apparently, our previous comments-in correspon-dence with you and Mr. Hans Ashar of NRC did not satis-fy the latter.

Before answering your questions, it is important to point out that the American Institute of Steel Construction, Inc. is not a regulatory' body that prescribes design rules and practices.

The Specifica-i i

tion and Code of Standard Practice are voluntary con-census documents.

Specifically, the Committee-on Spe-cifications consists of structural engineers with wide experience and high professional standing, represent-ing a wide geographical distribution ' t.hroughout the United States.

The membership of the committee is made up of approximately equal number representing design en-gineers in private practice, engineers involved in re-search and teaching, and engineers emp&oyedchy steel fabricating companies and auppliers.

The Specification developed by this committee, as approved by the AISC Board of Directors, is.widely used by designers and code.

authorities.

However, AISC recognizes the authority and responsibility of the licensed professional for.the design of structures within his'or her scope of expertise.

To aid design professionals.in exercising their authority and responsibility, AISC publishes design aids, manuals-and the Engineering Journal.

Reference is often made j

to the worldwide literature cn1 steel design.

l

Mr. Thomas G. Longlais January 15, 1986 Page 2.

1.

The AISC Specification is indeed silent on the design rules for laterally unbraced angles subject to bend-ing.

AISC, however, has frequently referenced the Australian research on this topic (1,2) in a nation-al series of seminars entitled " Steel Design Cur-rent Practice" beginning in late 1983.

In addition, the Australian article " Safe Load Tables for Lateral-ly Unsupported Angles" has been reprinted in the First Quarter, 1984 AISC Engineering Journal with the per-mission of the Australian ISstitute of Steel Construc-tion.

This design methodology is based on an earlier version"of the Australian Standard AS 1250-81(3) in the absence of angle-beam criteria in the present AISC Specification.

The Australian methodology for the design of single angles subject to bending is being used for an example in the 1986 First Edition of the AISC Load and Resistance Fac'.or Design (LRFD)

Manual.

2.

In the proposed LRFD Specification, the flexural de-sign capacity of a double angle beam is constant at the yield moment value until the onset of elastic buckling.

There appears to be no need for an in-elastic transition done because of the absence of major internal residual stresses in angle sections.

A minimum inelastic rotation capacity of 3 is a com-mon criterion for. limiting slenderness ratios in the AISC Specifications (see pg. 34-on new LRFD Specifi-cation).

This is approximately equivalent to the requirement that the computed elastic critical stress be at least 3 times as large as the yield stress, to ensure that the plastic moment will be reached.

Thus, the Australian recommendation that single angle flex-ural yielding will control in the slenderness zone wherein the computed elastic buckling moment is at least 3 times the yield moment is valid.

Furthermore, compact angles have a high shape factor to justify use of a 0.66F allowable bending stress.

y 3.

AISC interaction equations 1.6.la and 1.6.lb are in-tended for the design of beam-columns subject to concurrent axial loads and biaxial' moments.

.In the design of beam-columns, it is recognized that either yielding or stability may. control member behavior.

i

Mr. Thomas G. Longlais January 15, 1986 Page 3.

Equation 1.6.la checks. stability with maximum moments along the length of a beam.

Equation 1.6.lb is in-tended to be a stress strength check against yielding.

In this latter. situation, stresses should be combined at the critical member support cross-section and-should not be based upon the maximum moments along the member length as in equation 1.6.la.

Also, equations 1.6.la and 1.6.lb are intended for the simultaneous application of concurrent static loads.

If a member is subject to time dependent dynamic loads, i.e.,

seismic loads, then a rational method which ac-counts for the nonconcurrence of maximum axial bending moments about the two axes and axial loads may be utilized when employing equations 1.6.la and 1.6.lb.

4. We agree that in cases where lateral support of un-symmetric sections is provided at the location of con-centrated loads, simple bending theory will apply, i.e.,

the horizontal and vertical axes'(geometric) of the angle may be used to compute the bending stresses.

Otherwise, the principal axes of the member must be used in the interaction equations.

Sincerely yours, L

Geerhard Haaijer GH/cd cc:

N.

Iwankiw i

l l

m

REFERENCES

1. Leigh J.M. and Lay M.G.,

"The Design of Laterally Unsupported Angles", BHP Technical Bulletin 13(3),

Nov. 1969, pp. 24-29

2. Thomas, B.F. and Leigh, J.M.,

"The Behavior of Laterally Unsupported Angles", BHP Melbourne-Research Laboratory Report MRL 22/4, Dec., 1970

3. Australian Standard AS 1250-81, Australian Insti-tute of Steel Construction j

ATTACHMENT 2 Letter from Professor T. V. Galambos 2/20/85-

F i

• c...

UNIVEFGTY OF MINNESOTA cecenmn: e C.x aac ?.5ce a Eng.nec :nq

s..

o 122 c.o and f.t rcra' Erg neenng 0icmg

. ;. ; mw m.cs

500 Pc stvv Dove S E s

l PAnneapo3s; P.'innesma 55455-0220 j

. (612) 373-2968 20 February 1905 -

SARGENT & LUNDY FEB 25 bo:

Mr. Kennc th T. Eostal Assistant Manager, Structural Department

%[j,YG Snrgent and Lundy, Engineers 55 East Ifonroe St.

Chi c t.go, IL 60603 Deer I!r. Kostal:

This is a reply to your 1ctter of February 14, 1935, in which you deal with questions relating to the ficxure of laterally unsupported angle-beams.~ The design of laterally unsupported angle-beams.is not covered by the AISC Specification, for the reasons nentioned by Mr. Fdinger in his letter to you.

Most emphatically, Sec. 1.5.1.4.1 doe s not apply to angic-beams.

I have been aware of the Australian research sincc 1970, when I was a

~

visiting engineer at the Melbourne Research Laboratory of BHP.. The research was then in progress.

I witnessed some of the tests, and I reviewed the analytical derivat2ons with the authors.

I subsequently reviewed this work when I gave a lecture for ATSC on this topic in February or March of 1984. This work is correct and applicable to laterally unsuppseted angle-beams.

I also agree with the treatment of this topic in the Australiac Standard 1250-1981 ~. The work of the elastic buckling strcss F and Lay and Leigh sFould be used to determinethen tne allowable stress should be c applicable.

Sincerely yours, hA

/

T. V. Galammos Professor of Civil Engineering i

T/G:shh 1

l Nu.--.

a

ATTACHMENT 3 Letter from Professor T. V. Galambos 1/9/86 1

I A

l k

t.

t

{ TWIN CmES UNIVERSITY OF MINNESOTA Department of Civil and Mineral Engineering i 122 Civil and Mineral Engineering Building l 500 Pinsbury Drive S.E.

l 4

{

! Minneapolis. Minnesota 55455-0220 d

f (612) 373-2968 1

January 09,1986 Mr. T.G. Longlais, Head Structural Engineering Division j

Sargent and Lundy, Engineers 55 East Monroe St.

l Chicago, 111. 60603 j

Dear Mr. Longlais,

j.

This letter is to sum up our discussions and to put down my conclusions on.

4 the subject of the Sargent and Lundy design method for single-angle hangers for HVAC ducts in several nuclear power plants. This investigation started with your letter of Nov. 8,1985, in which you outlined the basis of your method and -

4 with which you enclosed copies of previous correspondence on.the subject. This' correspondence included an exchange of letters between Dr. Hartzman of the.U.S.

l Nuclear Regulatory Commision and myself, in which I made some general--

observations without specifically addressing the complete details of-your single-angle design method. Subsequent to your initial letter we had a further-exchange of letters to provide additional clarification, and.we had two meetings in my office at the University of Minnesota between you, Dr. S. Fang, and myself (on Nov. 26,1985 and on Jan. 3,1986). As a consequence of these letters and -

meetings, and after considerable study and analysis, I believe that I am quite familiar with both the philosophy and the method used by Sargent and Lundy to I

design these angles. My comments in this letter are intended to assist Sargent I

and Lundy in responding to NRC questions and to expand on my previous correspondence with the NRC in June 1985.

4 Before addressing specific issues I want to state that my comme.nts apply to the particular HVAC angle hangers used by Sargent and Lundy, and they are not t-meant to be a general design methodology for the design of single-angic beam-columns used in any kind of structural design. There S and L angles are primarily intended for the tension hanger support of HVAC ducts in several nuclear pow:r plants. Their end-framing is such that the ends are constrained about one of the geometric axes of the angle. By dynamic analysis it has been determined that under an extreme seismic event these hangers are subject to end moments about one, or sometimes both, geometric axes, and to small axial compressive forces which are usually (but not always)less than 15 percent of the allowable axial compressive force.

The check for combined bending and compression is covered in Sec. l.6.1 of--

the AISC Specification. Two criteria are required to be checked: 1) stability (Eq.1.6.la) and 2) cross section strength (Eq.1.6.lb). If the ratio f /F, is less than 0.15, then only one conservative interaction equation needs to be checked (i.e., Eq.1.6.2).

~

V' y9r

-.e-y

,s.

,-mi.

+w-i y

g, w

ap-+.-

e

-ee sqe

-p g-,

g.

g p.w'-

g

,r

-=

g y

4 e.-

+

wr+-,.

.9; e;

+

f 1

Letter to: T.G. Longlais; Sargent and Lundy, Engineers -

Page: 2 Date: January 09.1986 m

I i

. My first comments relate to the strenath check. There are two questions to L!

I be considered:

i

1) Is it appropriate to compute the flexural stress on the basis of the 1 ;

geometric rather than the principal axes?

1 F

2) What is the allowable ficxural stress F 7 b

j'-

It is my opinion that because the applied force and the restraint at ihn member ends is directed along one of the angle legs, i.e., along one of the-I geometric axes, the flexural stress should be comouted for the acomeltje axes The applied force will produce forces in.both the hanger angle and in the restraining members. The background arguments for this are presented in Sec.

l 11.4 of the book " Basic Steel Design" by Johnston, Lin and Galambos (3rd ed.,

j Prentice-Hall,1985).-

The value of F to be used for angles is not specifically defined in the-j AISC Specification.b A value of F = 0.66F is recommended for compact wide-b y

fhnge shapes bent about the major axis, and F = 0.75F for minor axis flexure.

5 h

These two values contain allowance for the different pl,astic shape factors for l

major and minor axis bending, i.e. l.10 and 1.55_ respect _ively. -It can be shown that the shape factor for a compact angle is about 1.8 when bending is about a-geometric axis,;.nd 1.75 when bending is about the minor principal axis. These values are larger than those for wide-flange shapes, so the use of F = 0.66F 4

b i

is conservative for angles. The use of F = 0.66F, for compact angles is thus b

t entirely within the intent of the AISC Specification.

1 The next comments refer to the questions related to the stability check.

The stability interaction equation in the AISC Specification is in a stress-format for reasons of convenience and convention. The original formulation of i

f the interaction equation is in terms of forces (design and ultimate compression and flexural forces; see, for example, Chap. II, in " Structural Steel Design",

editor L. Tall, Ronald Press,1974) which are translated into stresses. If, for example, the allowable flexural moment for the limit state of lateral-torsional j

buckling is formulated for flexure about a geometric axis, as was done by Leigh

~

and Lay in their paper, then the actual and the allowable stress may be normalized from moments about the geometric axis also. Thus it is permissible, for reasons of convenience, to use the stresses f and F in Eq.1.6.la with b

b j

reference to the geometric axes. A similar interpretation is used by the Steel Joist Institute for the design of cccentric single-angle webs in prefabricated

[

trusses. Several dozen angle columns were tested under three different end conditions about 15 years ago at Washington University to substantiate this method of design (see paper by Usami and Galambos, IABSE Memoirs,1971). This 4

I design method has been successfully used for over ten years, and tests on full I

scale trusses performed by John Leigh for his Master's thesis at Washington i

University show'ed it to result in a safe structure.

i The next question in connection with the stability check has to do with the :

r

}

value of F to be used in Eq. I.6-la. The AISC Specification is silent on this b

I I

m m-=-

r N-

~, v*e -v

• er v _-~ w e-~,.+-

wr v

-e = w w v' e---

r < v tv-c----

, ve - + w *v " r+ v

• * * *vW v w

. -* v

= - ++ *

~

f Letter to: 'T.G. Lon'glais; Sargent and Lundy,' Engineers Page: 3 '

Date: January 09c1986 subject, but a recent AISC lecture series acquainted American engineers _with the j

work of Leigh and Lay on the subject. This work was performed in 1970 at the BHP Steel Company Laboratory in Melbourne, Australia. Both Leigh and Lay are my former students, and I reviewed their work when I ' visited Australia in 1970. I.

~

{

have the highest regard for their research. I.have again reviewed their study and rederived their equation in the past weeks. It is my opinion that F can be

~

b safely determined by the criteria given in the Australian SAA Steel Structures Code AS1250-1981 Sec. 5. For the case of wide-flange beams I have compared the Australian design criteria with the new AISC Load and Resistance Factor Design rules and I found the Australian method to be more conservative in all cases I 4

j '

checked. The Australian buckling criteria assume that the cross-section pla'stifies when the clastic buckling stress equals or exceeds three times the yield stress if a wide-flange beam is bent about the major axis. This is a '

typical approach used on all types of buckling criteria in the Australian steel design specification, and it is justified in a number of papers by Max Lay and i

1 Nick Trahair (see also the figure on p.109 of Vol. II of the book by Atsuta and Wilfred Chen). Modern U.S. Specifications do not explicitly formulate the limits of plastic behavior in quite the same way, but the result is just about the same. For the single-angle beam the assumption of plastification under factored loads when F = 0.66F and the clastic buckling stress is 3F,1.7 to 1.8.

is, of b

y course, quite conservative because of the high shape factor of about l

Thus even the criterion of F = 0.95F,F = 0.66F, has a safety factor of about carries with it a factor of safety of

]

b about 1.9 against plastification, while i

j b

j 2.7. These remarks concern the behavior of compact angles (i.e., b/t < 65/ F ).

for non-compact angles we must use F = 0.6F,, or if b/t > 76/ F,, F = 0.60DF,.~

i b

b i

For the cases the Australian design equation must be topped out at 0.6F, or 0.6QF,, as appropriate.

4 i

It is my opinion that the use of the Australian criteria for determining f

F in Eq.1.6-la by Sargent and Lundy for the single-angle HVAC hangers is b

}

appropriate and conservative, resulting in safe designs.

I The use of AISC Eq.1.6-lb assures adequate strength at the ends of the hangers where the moments are maximum, while compliance with Eq.1.6-la check's the lateral-torsional stability of the member as a whole. For most of the practical angles for steel with F = 36 ksi, and for the compact angles for steel with F = 50 ksi only the s[rength equation needs to be checked if the ratio of lendth to thickness is less than 400.

The AISC interaction equations apply when the forces acting on them are acting concuirently. They would result in unduly conservative designs if the.

7 I

maxima of several load sets, which do not act simultaneously, were to be assumed to be applied.

The moment caused by the eccentricity of the axial force which is applied

- through one leg should be considered in design. This moment, however, is shared 1-by the restraining memb rs in proportion of the relative stiffnesses..The Usami experiments on restrained-end single angles, and the Leigh truss tests J

...a...

.u.

m.

T.G. Longlais; Sargent and Lundy, Engineers Letter to:

3 Page: 4 Date: January 09.1986 demonstrated the effectiveness of this assumption. With the relatively low axial forces in the Sargent and Lundy hanger angles and the participation of the restraining members the effect of the eccentricity is expected to be quite small.

In summary, whil'c the AISC Specification' does not contain explicit rules for the design of single-angle beam-columns, the Sargent and Lundy design metho'd is rational, it is conservative, and it meets the intent of the AISC Specification. I believe it to be a safe approach for the design of these hanger angles. In conclusion I would add that the safety of these hangers is further enhanced by the fact that compression' occurs only during a transient seismic event, and not always as in a gravity-type building column.

1 am satisfied that the Sargent and Lundy design method is safe and

~

adequate for the single-angle HVAC hanger.

Sincerely yours, 6

T. V. Galambos Professor of Civil Engineering TVG:sbh I

i

+

4 1

1 1

'I I

s

.