Professional Documents
Culture Documents
THEODORE V. GALAMBOS
Load and Resistance Factor Design, abbreviated as LRFD, ample, beam theory is more accurate than column theory
is a scheme of designing steel structures and structural (e.g., in Ref. 1,0 = 0.85 for beams and 0 = 0.75 for col-
components which is different from the traditionally used umns), or that the uncertainties of the dead load are smaller
allowable stress format, as can be seen by comparing the than those of the live load (e.g., in Ref. 2, yD = 1.2 and 7^
following two inequalities: = 1.6). LRFD thus has the potential of providing more
consistency, simply because it uses more than one factor.
Rn/F.S. > Qm (1) The purpose of this paper is to describe the development
1 of an LRFD specification for steel structures.
4>Rn > t yiQni (2) SIMPLIFIED PROBABILISTIC MODEL
1
The strength R of a structural member and the load effect
The first of these inequalities represents the allowable stress
Q are both random parameters, since their actual values
case, while the second represents the LRFD design crite-
cannot be determined with certainty (Fig. 1). The strength
rion. The left side in each case is the design strength, and
of a structure, often referred to also as its resistance, is de-
the right is the required strength* The term Rn defines the
fined in a popular sense as the maximum force that it can
normal strength as given by an equation in a specification,
sustain before it fails. Since failure is a term tfrat is associ-
and Qni is the load effect (i.e., a computed stress or a force
ated with collapse, it is more useful, in the context of
such as bending moment, shear force axial force, etc.) de-
structural behavior, to define strength as the force under
termined by structural analysis for the loads acting on the
which a clearly defined limit state is attained. Such limit
structure (e.g., live load, dead load, wind load, etc.). The
states are, for example, the plastic mechanism, the plastic
term F.S. represents the Factor of Safety, 0 is termed the
moment, the overall or component buckling load, fracture,
resistance factor, and the 7 ? 's are the load factors associated
fatigue, deflection, vibration, etc. Not all of these limit states
with each load effect Q z .
cause "collapse" in the popular sense, and so it is appro-
The coefficients F.S. > 1.0, 0 < 1.0, and 7, > 1.0 all priate to define strength as "the limit state which deter-
serve the same purpose; they account for the uncertainties mines the boundary of structural usefulness."
inherent in the determination of the nominal strength and
Structural behavior is thus satisfactory if Q < R, while
the load effects due to natural variation in the loads, the
on the contrary, Q > R is unacceptable. Since Q and R are
material properties, the accuracy of the theory, the precision
random, it is theoretically not possible to state with certainty
of the analysis, etc.
The fundamental difference between LRFD and the
allowable stress design method is, then, that the latter
employs one factor (i.e., the Factor of Safety), while the Ptobabi I ity
Density
former uses one factor with the resistance and one factor
each for the different load effect types. LRFD, by
employing more factors, recognizes the fact that, for ex-
74
75
THIRD QUARTER / 1981
Design according to the 1978 AISC Specification,
1 I i 1 1 i 1 1 I 1 1 1 i 1 1 |
rmtrr ffffl'" Parti
t 4- 1 wD = 57.5X32 = 1840 plf
Fig. 4. Two-span beam example Live load reduction for AT = 896 ft2 and D/L =
57.5/50: 0.49
wL = 50(1 - 0.49)(32) = 809 plf
Obviously there is a certain element of judgment involved Mmax = wL2/8 = 3115 kip-in. (at center support)
in these values, but they represent the best that a number 0.9X3115 _ . ,
of experienced people could come up with on the basis of Srea = = 118 in. 3
q
the available information. Substitution of these values into 0.66 X 36
Eqs. (4) and (5) gives R = \AlZxFy and VR =0.13. Use W21x62, Sx = 127 in.3; Zx = 144 in. 3
Plastic analysis of the beam in Fig. 4 gives:11 Thus R = 1.17 X 144 X 36 = 6065 kip-in.
O P + wL)L2 With this value of R, and with VR = 0.13, Q = 2333
MP = (J)
11.66 kip-in. and VQ = 0.11, the reliability index [Eq. (3)]
is: (3 = 5.6.
The right side of Eq. (7) is the load effect Q. It consists of
the dead and the live load effect, and it can be written as: Design according to the 1978 AISC Specification,
Part 2
Q = Q.D + QL (8) wD = 1.7 X 1840 = 3128 plf
6
The mean and the standard deviation are then: wL = 1.7X809 = 1387 plf
Q=QB + QL (9) __ (4.515 X28 2 )(12) . 3
Z = 1 l m
VQ = (VD2QD2+VL2QL2)V2/(QD + QL) (10) ^ ~ 11.66X36 -
A study of the load effect statistics (Ref. 2) gives the Use Wl8x50, Z^. = 101 in. 3
following values: The resulting (3 is 3.5.
Qj> = 1.05QnZ) , ^ = 0 . 1 0 (11) Design according to the proposed AISC LRFD
criteria (Ref. 1)
QL = QUL, VL =0.25 (12)
The design criterion is Eq. (2).
where the subscript n relates to the nominal code-specified <t>Rn > \.2QnD+l.6QnL (14)
dead and live loads given in Ref. 2. The statistics include
the uncertainties due to the loads themselves, as well as the 0 = 0.85 for beams
effects of translating the loads into the idealized load effects WD = \.2X 57.5 X 32 = 2208 plf
and the idealization inherent in the uniform distribution.
WL = 1 . 6 X 3 0 X 3 2 = 1536 plf
The occupancy live load QnL is equal to:2
3.744 X 2 8 2 X 12 _. .
QnL = QoL [ 0 . 2 5 + 1 5/y/Tj] (13) Zrea = = 99 in. 3
req
where QoL is the load effect due to the basic code specified 11.66X36X0.85
uniformly distributed live load intensity, and the bracket UseW18x50: Zx = 101 in.3
is a live load reduction factor dependent on the influence
area Ai (Aj = 2AT and Ai = \AT for beams and columns, The resulting /3 is 3.5.
respectively, where AT is the tributary area). This simple exercise showed how the relative reliability
The following data are specified for the given of several designs can be compared and that the design
problem: of this particular beam according to Part 1 of the AISC
Dead load intensity: 57.5 psf Specification is conservative in comparison to the designs
Basic live load intensity: 50 psf according to Part 2 and the LRFD method. Similar
Beam length between centers of supports: 28 ft studies of thousands of different structure elements form
Beam spacing: 32 ft the basis of the LRFD criteria.
RESEARCH ON THE DEVELOPMENT OF LRFD
For the tributary area of 28 X 32 = 896 ft2, the reduced
live load is 30 psf, and The ideas of multiple factors and of probabilistic ap-
2 proaches to design are not new. No precisely documented
n _28 (1.05X 57.5X32)02) _ _ . .
QD = 1559 kip m historical account is intended here (see Refs. 4, 5, and 12
~ 11.66X1000 - - for such a treatise), but a brief general discussion will be
QL = 755 kip-in.; Q = 2333 kip-in.; VQ = 0.11 provided to put the present status of LRFD into focus.
76
LEVEL I I I . 1
Full Probabilistic Analysis - Research A n a l y s i s for
Special Structures
Past Experience;
Design Office
Current
Fig. 5. Hierarchies of probabilistic methods Specifications Studies
77
Compact
Compact W-beam:
W-beam:
Uniform
Uniform moment
moment Mp 1.02 0.06
Table 1. Load Statistics Continuous
Continuous Mechanism
Mechanism 1.06 0.07
78
Notation:
LTB = Lateral-torsional buckling
RESISTANCE FACTORS
the two-span beam of Fig. 4, but using a more advanced The loads and the load factors in the LRFD design crite-
method.4 rion [right side of Eq. (2)] are now set by ANSI, and the
The resulting /3's were then examined, and on the basis task of the materials specification groups is then to develop
of this across-materials survey, it was determined that, on (j)Rn values which are consistent with the target reliabilities
the average, structural design in the U.S. inherently had inherent in the load factors. Reference 4 gives the proba-
the following reliability index values: bility-based method, as well as charts and a computer
Gravity loads only: /3 = 3.0 program to facilitate this task. The resistance factors <>/ are
Gravity plus wind loads: /3 = 2.5 to be developed for the load combination \.2Dn + 1.6Ln
and for a target reliability index of (3 = 3.0.
While there was considerable fluctuation, these values The following simple example, based on the simplified
turned up most frequently,4 and the decision was made by approach of Eq. (3), will illustrate the method. For a simply
the ANSI A58 subgroup working on the problem to base supported beam of compact shape, Ref. 9 provides the
the new load factors on these values of /? as target relia- following statistical data:
bilities. Accordingly a new cycle of probabilistic analyses
was made (see Fig. 7), and the load factors listed in Table ? = 1.02, F P = 0.06, M = 1.05,
8 were finally developed; these were recommended for VM = 0 . 1 , F= 1.00, ^ = 0.05
adoption in the new ANSI A58.1 standard.2 Therefore, using Eqs. (4) and_(5), R = 1.07i? and VR =
The load combinations are of the following general 0.13. The load effect data is: Q = l.05Dn + Ln , where Ln
form: = Lo(0.25 + 1 5 / V 2 ^ r ) [ E q s ^ ) and (13)], and VQ is
given by Eq. (10). Noting that D = 1.05Z>, VD=0A.L
1.2Dn + 7 I -Q I - n + yQjn 05) = Ln, VL = 0.25 (Table 1), and letting
Mr=SK(f\j-l)
I I I I I I I I I I
400 800 1200 1600 2000
^KMr
^v. 2looo
O^
\-.^(^ V-i^|Vfy^
--o
LIMIT STATE: Web Local Buckling
O 1978 AISC Specification
-L L /D _L rh'f
Fig. 9. Reliability Index for simply supporte d compact wide- Fig. 10. Nominal strength of wide-flange beams under
flange beam, tributary area = 400ft2 uniform moment
80
81
THIRD QUARTER / 1981
2. American National Standards Institute Building Code 9. Yura, J. A., T. V. Galambos, and M. K. Ravindra The
Requirements for Minimum Design Loads in Building and Bending Resistance of Steel Beams Journal of the Struc-
Other Structures ANSI A58.1-81 draft. tural Division, ASCE, Vol. 104, No. ST9, Sept. 1978.
3. Cornell, C A. A Probability-Based Structural Code 10. Galambos T. V. and M. K. Ravindra Properties of Steel
Journal, American Concrete Institute, Vol. 66, No. 12, Dec. for Use in LRFD Journal of the Structural Division,
1969. ASCE, Vol. 104, No. ST9, Sept. 1978.
4. Ellingwood, B., T. V. Galambos, J. G. Mac Gregor, and C 11. Beedle, L. S. Plastic Design of Steel Frames John Wiley
A. Cornell Development of a Probability-Based Load and Sons, 1958.
Criterion for American National Standard A58 National 12. Ang, A.H.-S. Structural SafetyA Literature Rev-
Bureau ojStandards Special Publication 577, June 1980. iew Journal of the Structural Division, ASCE, Vol. 98,
5. Ravindra, M. K. and T. V. Galambos Load and Resistance ST4: April 1972.
Factor Design for Steel Journal of the Structural Division, 13. Freudenthal, A. M. Safety, Reliability and Structural
ASCE, Vol. 104, No. ST9, Sept. 1978. Design Journal of the Structural Division, ASCE, Vol. 87,
6. Benjamin, J. R. andC. A. Cornell Probability, Statistics, ST3, March 1961.
and Decision for Civil Engineers McGraw-Hill Book Co., 14. Puglsey, A. The Safety of Structures E. Arnold Pub-
1970 lishers, London, 1966.
7. The Nordic Committee on Building Regulations Recom- 15. Vincent, G. S. Tentative Criteria for Load Factor Design
mendations for Loading and Safety Regulations for Struc- of Steel Highway Bridges AISI Bulletin No. 15, 1969.
tural Design NKB Report No. 36, Nov. 1978. 16. Galambos, T. V. Proposed Criteria for L R F D of Steel
8. Canadian Standards Association Steel Structures for Building Structures AISI Bulletin No. 27, Jan. 1978.
BuildingsLimit States Design CSA Standard S16.1- 17. Wiesner, K. B. LRFD Design Office Study Engineering
1974. Journal, AISC, First Quarter, 1978.
82