Dynamic Wind Loads and Cladding Design

Surcharge dynarnique due au vent et calcul du parement Dynarnische Windbelastung und Bemessung von Gebaudeverkleidungen

D.E. ALLEN

W.A. DALGLIESH

Division of Building Research National Research Council of Canada Ottawa, Canada

Windstorm damage t o cladding ( e x t e r i o r s u r f a c e s ) of buildings often r e sults in considerable economic l o s s and s o m e t i m e s p e r s o n a l injury. The s a m e c a r e should t h e r e f o r e be given to cladding design and r e s e a r c h a s is given t o t h e s t r u c t u r e . T h i s d i s c u s s i o n looks a t two a s p e c t s of t h e design of cladding t o r e s i s t wind loads: ( a ) how behaviour u n d e r dynamic wind l o a d s affects r e s i s t a n c e of both m e t a l (ductile) and g l a s s (brittle) p a n e l s ; (b) what r i s k s of cladding f a i l u r e a r e implicit i n North A m e r i c a n d e s i g n rules. The problem of d e t e r m i n i n g a c t u a l wind l o a d s on cladding, the g r e a t e s t unknown in t h e design p r o b l e m , i s not d i s c u s s e d . Desien of Claddine t o R e s i s t Wind Loads B e c a u s e turbulent wind i s a dynamic loading, r e c e n t design approaches f o r s t r u c t u r e s a r e b a s e d on a dynamic component that t a k e s into account wind turbulence, s i z e of building (averaging effect on t u r b u l e n c e ) , a n d dynamic amplification. A s i m i l a r approach i s suitable f o r cladding design ( F i g . 1)

w h e r e R i s the design r e s i s t a n c e , FS t h e design safety f a c t o r , wo t h e design wind p r e s s u r e , iG the m e a n wind p r e s s u r e equal t o the m e a n velocity p r e s s u r e 4 tirnes the shape f a c t o r Cp, Cg the g u s t f a c t o r , IT t h e r m s ( r o o t m e a n s q u a r e ) intensity of wind p r e s s u r e t u r b u l e n c e and g a p e a k f a c t o r relating peak wind p r e s s u r e t o r m s intensity of t u r b u l e n c e . F o r a stationary G a u s s i a n l o a d ing, the expected peak f a c t o r g i s given i n Table I, b a s e d on the t h e o r y given in Ref. (1) and assuming t h e r e a r e 0 . 3 p e a k s p e r s e c o n d , a figure obtained from full-scale p r e s s u r e measurements2. When IT i s z e r o t h e loading i s s t a t i c and the r e s i s t a n c e R c o r r e s p o n d s t o t h a t f o r a s t a t i c a l l y loaded panel. Since t h e effect of dynamic l o a d s on r e s i s t a n c e i s r e l a t e d only t o t h e turbulent component of t h e wind, i t i s appropriate t o consider t h i s and o t h e r s t r u c t u r a l effects, such a s dynamic amplification,

280

V - DYNAMIC WIND LOADS AND CLADDING DESIGN

by a modification of the peak factor g in Eq. ( I ) , i. e. by multiplying g by a "gust m a t e r i a l " factor, km, C g

,vlND
wm
Wo

= 1 + km. g o I T

(2)

The "gust m a t e r i a l " f a c t o r , b, depends both on the m a t e r i a l and s t r u c t u r a l behaviour and on the nature of dynarnic loading; i t will be shown that f o r cladding t h i s factor i s nearly equal t o 1. The n a t u r a l frequency of metal and g l a s s panels v a r i e s f r o m about F i g u r e 1: Wind P r e s s u r e Representa5 t o 50 Hz. In m o s t c a s e s t h i s f r e tion f o r Cladding Design quency i s considerably higher than significant wind turbulence frequencies. Thus during e l a s t i c deformations a panel can be represented a s a statically loaded s t r u c t u r e and dynamic amplif ication can be neglected. Only for unusual, high-f requency wind turbulence c r e a t e d locally by building c o r n e r s 4 o r by surface i r r e g u l a r i t i e s , such a s mullions, i s dynarnic magnification likely t o become significant.
I
I

Metal Cladding TABLE I Metal panels fail by yielding, by buckling, o r by f r a c t u r e of the connections. Two a s p e c t s of m e t a l behaviour a r e discussed in t h i s paper: r a t e effect, i. e. change of yield strength with r a t e of loading, and the effect of p l a s t i c deformation. Metal fatigue is neglected in t h i s study.

PEAK FACTOR AND GUST FACTORS FOR GLASS

Duration Expected Peak Glass (Fig. 3) of Mini F a c t o r for km. g mum, Load Loading

t
one h r .
3.9

4.0

2.0

According t o information on 10 min. 3. 4 3. 2 1.7 r a t e effect f o r steel, a ten fold i n c r e a s e in r a t e of loading inc r e a s e s the yield s t r e s s , uy, by approximately 2 ksi. Since the standard ASTM corresponds approximately t o 10 k s i p e r second, t h e r a t e effect testing r a t e can be approximated by

Y O

= 2 log

(j;)

(3)

where uo i s the yield s t r e s s according t o the standard t e s t . P l a s t i c deformation can a l s o absorb an exceptional t e m p o r a r y overload; although the m a t e r i a l m a y yield t h e r e m a y not be sufficient permanent deformation t o r e q u i r e replacement of the panel. Vickery looked into t h i s problem f o r a steel building s t r u c t u r e and found that the mean wind speed n e c e s s a r y to p r o duce s e v e r e damage i s about 20 p e r cent g r e a t e r than that n e c e s s a r y t o just produce yielding. Consider a simply-supported panel, undergoing plastic bending a t midspan. The panel i s subjected t o a half sinusoidal cycle of wind p r e s s u r e with

D.E. ALLEN

W.A. DALGLIESH

28 1

maximum amplitude wm G and forcing frequency n Hz (Fig. 1). T h e amount of plastic deformation, b p , during an excursion of wind p r e s s u r e beyond yield p r e s s u r e wy can be calculated numerically f r o m the equation of motion. Rate effect i s taken into account in accordance with Eq. ( 3 ) by expressing wy in t e r m s of wo, corresponding t o the standard test yield r e s i s t a n c e . The following damage c r i t e r i a a r e assumed: 6 1 = 5, whichever occurs permanent deflection:$ = - and section failure: 6,, L 50 ' f i r s t . The f i r s t c r i t e r i o n c o r r e s p o n d s t o the need forYpanelreplacement and the second t o reaching of maximum bending r e s i s t a n c e . Ln c o r r u g a t e d panels, local buckling o c c u r s before v e r y l a r g e plastic deformation. Calculations w e r e c a r r i e d out f o r 3 types of cladding spanning 10 ft: ( 1 ) sheet m e t a l (weight 1 psf, no = 20 Hz); ( 2 ) n o r m a l roof decking o r wall cladding (weight 8 psf, no = 10 Hz); ( 3 ) reinforced c o n c r e t e o r m a s o n r y (weight 50 - psf, no = 20 Hz). In the calculations it was a s s u m e d that wy = 50 psf and w/wy = 0.5. The r e s u l t s , given in F i g . 2 in t e r m s of &, show that for wind frequencies l e s s than about 1 Hz, v e r y little can be gained by taking into account plastic deformation and r a t e effect. Recalculation for n o r m a l cladding, C a s e (Z), assuming s t r a i n hardening f a c t o r s ( r a t i o of inelastic stiffness t o e l a s t i c stiffness) of 0.01 and 0.1 indicate that s t r a i n hardening i s no m o r e significant f o r dynamic r e s i s t a n c e than it i s f o r s t a t i c r e s i s t a n c e . Thus design wind loads f o r m e t a l cladding failing p r i m a r i l y by yield can be determined by assuming that the panel i s a static s t r u c t u r e which f a i l s when the wind p r e s s u r e exceeds the standard p l a s t i c r e s i s t a n c e . Some e x t r a r e s i s t a n c e i s available f o r r a r e local high-frequency wind turbulence. Windows G l a s s i s a brittle m a t e r i a l in which failure s t a r t s at an invisible surface o r 0.1 0.2 0.5 I 2 5 10 edge flaw. I t s strength i s dependent on F O R C I N G FREQUENCY n HZ window size (in accordance with the theory of weakest flaws), r a t e ( o r F i g u r e 2: Gust M a t e r i a l F a c t o r duration) of loading, and, t o a l e s s e r k for One Cycle of , extent, on t e m p e r a t u r e and relative Loading humidity. Rate of loading, the only effect of i n t e r e s t to t h i s study, can be approximated by the following criterion:=

'm

where w i s the pressgure and tf the t i m e a t failure. On the b a s i s of manufact u r e r s ' loading t e s t s a gradually applied p r e s s u r e in which f a i l u r e occurs at about one minute, the constant CM i s determined f r o m Eq. (4) t o be 4.62 wol ' , w h e r e wo i s the failure p r e s s u r e f r o m the standard t e s t . Based on

282

V - DYNAMIC WIND LOADS AND CLADDING DESIGN

one peak of sine wave loading (Figure I ) , Fig. 2 shows that g l a s s i s considerably m o r e sensitive to r a t e effect than m e t a l , although f o r a load of v e r y short duration (high n) m e t a l gains on glass because of its ductility. Because of this sensitivity to r a t e effect, one peak of loading does not p r o vide useful information f o r g l a s s . F i g u r e 3 shows the r e s u l t s of applying the damage c r i t e r i o n to a sustained random wind p r e s s u r e :

where x i s assumed to be a stationary Gaussian random process with a mean of zero and a standard deviation of one. The damage will be different f o r each random p r o c e s s of duration T but the expected damage can be determined f r o m Eq. ( 5 ) by expanding into powers of x and replacing

value based on the n o r m a l distribution, 1. 3 .(n-1)T. F r o m Eqs. ( 1 ) and (3), wo can be eliminated and the r e s u l t s can be expressed in t e r m s of h. or C g g' F i g u r e 3 shows the r e s u l t s a s a function of r m s intensity of t u r b u lence, IT, f o r 2 load durations: 10 minutes and 1 hour. The r e s u l t s indicate constant values of k,.g for high intensity of turbulence and constant values of C f o r low intensity g of turbulence; these values a r e given in Table I. The f o r m e r 0.02 0.04 0.060.080.1 D. 2 0.4 0.6 0.a1.0 applies to m o s t windows n e a r the R M S I N T E N S I T Y OF T U R B U L E N C E IT ground and the l a t t e r to windows in tall unobstructed buildings. F i g u r e 3: Gust F a c t o r s F o r G l a s s The values of km. g compare Stationary Gaus s ian Loading closely t o the expected peak f f a c t o r s g, which do not consider dynamic s t r u c t u r a l effect. The r e s u l t s o t h i s study therefore indicate that, except f o r windows subject to s m a l l turbulence, the value of k, f o r g l a s s can a l s o be taken a s 1.

..

s',

dt by its expected

Risk of F a i l u r e The consequences of cladding failure due to wind a r e not a s s e r i o u s as f o r s t r u c t u r a l collapse. North American building codes allow for t h i s , e i t h e r by a one-third i n c r e a s e in allowable s t r e s s f o r m e t a l cladding,'' o r by a reduced r e t u r n period f o r design wind load.'' Annual failure r i s k s implied by North American de s i g i rules a r e compared in Table I1 based on the following assumptions: (1) m e t a l yield s t r e n g t h is distributed normally with a m e a n value 1 . 1 5 times the specified r e s i s t a n c e (guaranteed minimum) and a coefficient of variation 0. 10; ( 2 ) g l a s s window strength i s distributed normally with a coefficient of variation 0. 25; (3) maximum annual hourly wind loads follow the e x t r e m e value Type 1 distribution with a coefficient of variation 0.30. Annual failure r i s k s of 0.01 to 0 . 3 p e r cent in Table I1 appear t o be considerably higher than indicated by actual damage, probably due t o conservative design assumptions. Comparative f i g u r e s in Table 11, however, indicate that

D.E. ALLEN - W.A. DALGLIESH

TABLE I1

ANNUAL FAILURE RISKS FOR CLADDING

Code

Design Wind Code Gust Return Factor, Period, Years Cg 10 30 50 2. 5 2. 0 2. 0

~eta'l Annual Safety Failure Factor Risk, O/o 1. 67 1. 25 1. 25 0. 007 0. 26 0. 13

Glass Annual Safety Failure Factor Risk, O/o 2. 5 2. 5 2. 5

4.

N B C ~ 1970

0 . 17 0 . 21 0 . 16

N B C ~ 1965 A N S I ~ 1972
a b

National Building Code of Canada - Sections 4. 1, 4. 6 and A m e r i c a n National Standards A58. 1

1972. S e e a l s o Ref. (10)

recent Canadian design r u l e s f o r m e t a l cladding a r e too conservative; a safety f a c t o r of 1.25 would give a n implied f a i l u r e r a t e of 0.16 p e r cent p e r year. Recent window f a i l u r e s in t a l l buildings indicate that falling g l a s s f r o m a broken window initiates f a i l u r e of a n u m b e r of other windows ( a kind of prog r e s s i v e collapse). F o r t h i s and psychological r e a s o n s it i s recommended t h a t design f a i l u r e r i s k s be reduced f o r t a l l buildings. Conclusions A study of m e t a l f o r r a t e effect and ductility, and of glass f o r r a t e effect, shows t h a t f o r turbulent wind loads, both t y p e s of cladding can be considered a s statically loaded s t r u c t u r e s in which f a i l u r e o c c u r s when wind p r e s s u r e exceeds the s t r u c t u r a l capacity a s d e t e r m i n e d by standard t e s t s . The f a c t o r k,in Eq. ( 2 ) , which t a k e s into account dynamic behaviour under turbulent wind loads, c a n be g e n e r a l l y taken a s 1. F o r g l a s s , m i n i m u m v a l u e s of the gust factor, C in Table I a r e recommended t o avoid f a i l u r e of windows subject t o steady g' winds. Annual f a i l u r e r i s k s f o r cladding indicate that r e c e n t Canadian r u l e s f o r m e t a l cladding a r e too conservative. Existing safety f a c t o r s f o r windows in t a l l buildings a p p e a r t o be too small. The a r e a of g r e a t e s t uncertainty in t h e design of cladding t o r e s i s t wind loads i s in t h e wind loading itself, in p a r t i c u l a r t h e intensity of turbulence, T h i s information should be obtained f r o m f u l l I T , and shape f a c t o r , scale m e a s u r e m e n t s an boundary - l a y e r wind-tunnel t e s t s .

"$

References (1) Davenport, A.G. "Note on the Distribution of the L a r g e s t Values of a Random Function with Application t o Gust Loading". P r o c . Inst. of Civil E n g i n e e r s , Vol. 28, 1964, p. 187. (2) Dalgliesh, W.A. "Statistical T r e a t m e n t of P e a k G u s t s on Cladding" J o u r n a l of the S t r u c t u r a l Division, ASCE ST9, September 1971.

V - DYNAMIC WIND LOADS AND CLADDING DESIGN

(3) Standen, N.M., Dalgliesh, W.A. and Templin, R. J . "A Wind Tunnel and Full-Scale Study of Turbulent Wind P r e s s u r e s on a T a l l Building". DME/NAE Q u a r t e r l y Bulletin No. 1971(4). National R e s e a r c h Council of Canada, J a n . 1972. ( 4 ) Ostrowski, J . S . , M a r s h a l l , R. D. and C e r m a k , J . E. "Vortex Formulation and P r e s s u r e Fluctuations on Buildings1' in Wind Effects on Buildings and S t r u c t u r e s . University of Toronto P r e s s , 1968. ( 5 ) Voorhees, H.R. "A Survey of Effects on Lower-Than-Usual R a t e s of Strain in the Yield and Tensile Strengths of Metals" ASTM Data S e r i e s DS44, M a y 1969. (6) ASTM Standard A370 "Standard Methods and Definitions for Mechanical Testing of Steel Products". ASTM Standards P a r t 4, J a n u a r y 1970. ( 7 ) Vickery, B. J. "Wind Action on Simple Yielding Structures'' J o u r n a l of Engineering Mechanics Division, ASCE, Vol. 96, EM2, A p r i l 1970. (8) Brown, W.G. "A Load Duration T h e o r y f o r G l a s s Design". Division of Building R e s e a r c h , R e s e a r c h P a p e r No. 508. National R e s e a r c h Council of Canada. J a n u a r y 1972. (9) P . P . G . Industries, Technical Service Report No. 101, G l a s s P r o d u c t Recommendations Structural.

(10) A m e r i c a n I r o n and Steel Institute, AISI Specification f o r the Design of Cold-Formed Steel S t r u c t u r a l M e m b e r s 1968. (11) National Building Code of Canada 1970 Section 4.1. Structural Loads and P r o c e d u r e s . Associate Committee on the National Building Code, National R e s e a r c h Council of Canada. (12) Engineering News-Record. F e b r u a r y 15, 1973, p. 14.

SUMMARY A study is made of the behaviour of ductile (metal) and brittle (glass) panels under dynamic wind loading. The results show that the effect of such factors as dynamic amplification, rate of loading and ductility can generally be neglected. Failure risks implicit in North American codes are compared; some changes in design rules a r e suggested, including an increase in safety factor for glass windows in tall buildings.

RESUME On a fait une Ctude sur le comportement des panneaux ductiles (metal) et O cassants (verre) dans des conditions de chargement dynamique d au vent. D'aprss

D.E. ALLEN

W.A. DALGLIESH

l e s r e s u l t a t s obtenus, l'effet d e f a c t e u r s c o m m e l'amplification dynamique, l a v i t e s s e d e c h a r g e m e n t e t l a ductilite e s t generalement negligeable. L e s r i s q u e s de r u p t u r e que sous-entendent l e s codes n o r d - a m e r i c a i n s s o n t m i s en r e g a r d ; q u e l q u e s modifications dans l e s r e g l e s d e c a l c u l sont s u g g e r e e s , y c o m p r i s une augmentation du facteur d e s e c u r i t e d e s f e n e t r e s vitrees dans l e s b2timents t r & s eleves.

ZUSAMMENFASSUNG D a s Verhalten von zahen (Metall) und sproden (Glas) Materialien in G e baudeverkleidungen unter dynamischer Windbelastung w i r d untersucht. Die E r g e b n i s s e deuten an, d a s s d e r Effekt von Faktoren w i e dynamisches Aufschaukeln, Belastungsgeschwindigkeit und Zahigkeit i m allgemeinen vernachlassigt w e r d e n kann. Die in nordamerikanischen Normen enthaltenen Bruchwahrscheinlichkeiten werden verglichen. Gewisse Verbesserungen in den Berechnungsmethoden w e r d e n vorgeschlagen, u. a. eine Erhohung d e s Sicherheitsfaktors fiir G l a s in hohen G e bauden.