Você está na página 1de 9

ACI JOURNAL TECHNICAL PAPER

Iitle no. 80-27

Load and Resistance Factors for Concrete Design

by James G. MacGregor

An overview of the methodology used and the assumptions made in


deriving load factors and resistance factors ¢ proposed for future
editions of the ACI Building Code are presented. The load factors are
based on those incorporated in ANSI A58.1-82.

"0
«<
Keywords: building codes; load factors; loads (forces); prestressed concrete; 0
....J
reinforced concrete; safety; safety factor; structural design.
II
0

In 1963 the ACI Building Code introduced ultimate


strength Design in which design was based on ultimate
limit states using loads increased by load factors and R =Strength
strengths reduced by strength reduction factors ¢. Fig. I - Strengths and maximum lifetime loads for a
These factors were based on a rudimentary statistical family of beams
analysis that was based on rather incomplete knowl-
edge. Derivation of the ACI Code safety provisions is
reviewed in detail in the appendix of Reference 1.
Since 1963 the understanding of safety theory, struc- PROBABILISTIC APPROACH TO STRUCTURAL
SAFETY
tural loadings, and reinforced concrete have expanded
Consider a large family of similar floor beams each
greatly. During the 1970s extensive work was done to
subjected to the same type of occupancy for a 50 year
derive load and resistance cJ> factors for steel struc-
lifetime. Each of the beams has been designed to sup-
tures.2 The load factors differed from those now in use
port an unfactored or service load moment Q". At one
in the design of concrete structures. Recognizing the
time during its lifetime each beam experiences a ''life-
potential confusion in design offices if steel structures
time maximum" moment. Since the dead load and the
were designed using one set of load factors and con-
live load are both random, each of the family of beams
crete structures using another, the ACI Building Code
will experience a different lifetime maximum moment.
Committee unanimously passed a motion in 1976 en-
The resulting family of lifetime maximum moments has
dorsing the concept of similar load factors for all ma-
a statistical distribution similar to that plotted along the
terials.
vertical axis in Fig. 1.
In 1979 the Center for Building Technology at the
Each of the beams concerned has a different mo-
National Bureau of Standards brought Drs. Cornell,
ment capacity due to random variations in the concrete
Ellingwood, Galambos, and the author together to de-
and steel strengths, effective depths, etc., as well as
velop a universal set of load factors for use in the de-
random and systematic errors due to simplifications in
sign of buildings. Results of this study have been pub-
the design equation (rectangular stress block, limiting
lished in References 3 through 5. It is envisioned that
strain, etc.). As a result, the family of moment capaci-
these load factors and compatible resistance factors
would be incorporated in the ACI, AISC, and AISI
Recc:ived Apr. 30, 1982, and reviewed under Institute publication policies.
codes and specifications. The basic set of load factors Copyng~t © 1983! American Co~cr.ete !nstitute. All rights reserved, including
has been incorporated in the 1982 ANSI A58.1 build- the makmg of COJ;>1es unless perm1ss1on 1s obtamed from the copyright propri·
etors: Pert1_nent dtscuss10n w1ll be published in the May-June 1984 ACI JouR-
ing design load standard. 6 NAL 1f rece1ved by Feb. I, 1984.

ACI JOURNAL I July-August 1983 279


James G. MacGregor, FACI, is a professor of civil engineering, University of
Alberta, Edmonton, Canada. He is a member of ACI Committee 318, Stan-
dard Building Code; and joint ACI-ASCE Committees 441, Reinforced Con-
crete Columns; and 445, Shear and Torsion. Dr. MacGregor is chairman of the
Canadian reinforced concrete code committee.

ties R has the distribution plotted along the horizontal


axis in Fig. 1. =Safety Margin
The 45 deg line in Fig. 1 represents the case where the
load effect Q equals the strength R. Combinations of Q Fig. 2 - Definition of failure
and R that fall above this line such as loading Q 1 acting
on structure R 1 result in failure (Q > R).
If we know the means and standard deviations of R to govern, in this case, reinforced and prestressed con-
and Q f/[, aR, Q, and aQ), we can define a new function crete structures for buildings. Safety provisions have
been developed to be compatible with the loads in the
Y = R - Q = safety margin (1) 1982 version of the ANSI A58-1 Standard6 and the de-
sign procedures in the 1983 ACI Code. 7
with mean
Definition of limit states
Y=R-Q (2) The limit states (design conditions) to be considered
should be specified in the code and defined in func-
and standard deviation tional and/or numerical terms. The load and resistance
factors presented apply only to the ultimate limit states
of failure due to flexure, shear, etc.
(3)

This is plotted in Fig. 2. The shaded portion of this Selection of design life
figure represents cases where R - Q < 0; it represents Although building codes do not specify a design life,
cases where failure occurs. The probability of failure is it is necessary to assume a value to define the extreme
the probability that R - Q < 0. If the type of distri- values of the loads. A 50-year life has been chosen.
bution is known, the probability of failure can be com-
puted from the number of standard deviations that Y
Definition of loads and load effects
exceeds zero. This is shown in Fig. 2 as {3a-' where {3 is
referred to as the "safety index" Loads
Because the load factors were derived for inclusion in
the 1982 edition of the ANSI A58.1 building design
{3 = Y!ay (4) load standard, 6 they are, of course, compatible with the
loads specified in that standard. The nominal or speci-
If {3 is increased by increasing the value of Y, the fied dead load corresponds to the value computed from
shaded area in Fig. 2 is reduced and the probability of the dimensions shown on the drawings and nominal
failure is reduced. Thus {3 is a measure of the reliability densities. The ANSI floor, wind, and snow loadings are
of a structural member. A considerably more complete chosen to represent the mean maximum value during a
version of this theory has been used to compute the 50-year life, and earthquake effects are given in terms
safety indexes presented in this paper. of the 50 year maximum peak ground acceleration. The
live load reduction factors in ANSI A58.1-82 were in-
cluded as part of the specified live loads.
STEPS IN DEVELOPMENT OF LOAD AND
RESISTANCE FACTORS FOR CONCRETE
STRUCTURES Load effects
A series of 11 fundamental decisions must be made The moments, shears, axial forces, or other effects
during the development of a limit states design code for produced by a load acting on a structural member are
structural design. The balance of this paper will illus- referred to as load effects. In safety calculations the
trate the application of this procedure to the calcula- load effects caused by a given set of loads are com-
tion of load and resistance factors for reinforced con- pared to the corresponding strengths or resistances of
crete structures. the members in question. Variability of the load effects
includes the variability of the load itself, its distribu-
Scope of proposed code tion on the structure, and the structural analysis. 2
The first step in the derivation of a limit states de- Probability distributions used to represent the various
sign code is to define the class of structures the code is load effects are given in Reference 3.

280 ACI JOURNAL I July-August 1983


Variability of resistances of reinforced and of a load is a moderately high fractile of the distribu-
prestressed concrete members tion of 50 year maximum values of that load. In com-
Actual strength of a reinforced concrete member may puting the maximum factored load effect for a prob-
differ from the calculated or nominal strength due to lem involving several variable loads, several combina-
variations in material strengths and dimensions of the tions are considered, with each of the loads being con-
member as well as variabilities inherent in the equa- sidered as the principal variable load in turn.
tions used to compute member strengths. Resistance Turkstra and Madsen" have shown that Eq. (7) ex-
distribution properties used in this paper are docu- presses the true lifetime maximum combination of
mented in a companion paper8 and in Appendix B of loads with much more accuracy than Eq. (6). The ACI
Reference 3. load combination rule in which the dead load is multi-
plied by the probability factor a was considerably less
Selection of code format accurate than either Eq. (6) or (7). The set of load fac-
Probability based limit states design is based on tors presented in this paper are based on Eq. (7). To
nominal loads or load effects multiplied by load fac- simplify design, however, a few fundamental load
tors that are generally greater than one and on nominal combinations are explicitly stated rather than using the
resistances multiplied by resistance factors, less than general form of Eq. (7).
unity, according to the equation

Factored resistance ~ Effect of factored loads (5) Resistance factors


The left-hand side of Eq. (5), factored resistance, can
A number of different formats for this equation are also be expressed in several ways. The most familiar of
reviewed in the following paragraphs. The final choice these to North American designers is the use of
must balance theoretical appeal, computational ease, ''strength reduction factors'' or ''resistance factors''.
accuracy, and user acceptance. In this format, the left-hand side of Eq. (5) is expressed
as ct>R. where cJ> is a resistance factor that applies to a
Load factors particular structural action such as flexure, shear,
The National Building Code of Canada9 uses the bond, axial compression, etc.
probability factor format given in Eq. (6) to specify the The other important method of specifying resistance
basic loading cases factors uses material partial safety factors. In the CEB
Factored load effects Model Code, 10 the strength of a cross section is com-
puted using design material strengths equal to f: l"fc
= U{yJ). + a('yLLn + 'YwW.)} (6)
and fyi'Ys where 'Yc and 'Ys are material partial safety
where U refers to the load effects due to loads in the factors or material understrength factors for concrete
brackets and D., L., etc., are the nominal or specified and steel, respectively. These partial safety factors are
load 'Yv• 'Yu etc., are load factors; and a is a load com- the same for all ultimate limit states. For average con-
bination probability factor equal to 1.0, 0. 7, and 0.6 if struction quality these terms have values of 1.5 and
one, two, or three loads are included in the inner 1.15, roughly corresponding to the 1 in 1000 under-
bracket. strength values of concrete and steel.
In this format and au·others discussed, the terms 'YD• Major factors to consider when deriving resistance
'YL• etc., account for variations in the dead or other factors include:
loads themselves plus variations in the load effects due 1. Variability in member strength due to variability
to uncertainties in the load models and the structural of the material properties in the structure. In the case
analysis. The factor a in this format accounts for the of a composite material like reinforced concrete, two or
reduced probability that maximum dead, live, wind, more material variabilities may have to be considered.
etc., loads act simultaneously. The ACI Code 7 uses a The material understrength factor approach is better in
similar combination for dead, live, and wind loads ex- this case since the variabilities of concrete and steel can
cept that the factored dead load is also multiplied by a. be considered independently.
The Comite Euro-International du Beton (CEB) 2. Variability in member strength due to variability
Model Code 10 uses Eq. (7) to define the basic factored of dimensions. Neither approach is adequate here.
load effects 3. Variability in member strength due to simplifying
Factored load effects assumptions in the resistance equations (e.g., use of a
(7) rectangular stress block in concrete design). The resis-
U ['YvlJ + 'YQ[Q,k +
i
E
>
l
(ao;Q;k)]J tance factor approach is better in this case since each
type of structural action has its own resistance factor.
where Q,k and Q;k are the characteristic values of the 4. Increased risk to building occupants if failure oc-
principal variable load Q, and some other less impor- curs without warning and the post-failure strength is
tant variable load Q;; a is the ratio of the arbitrary
0; less than original strength. Both approaches can allow
point-in-time value of the ith load to the characteristic for this in an empirical manner.
value of that load; and 'YQ is the load factor on the 5. Importance of member in structure. Neither pro-
combination of variable loads. The characteristic value cedure is entirely adequate. Resistance factors can be
ACI JOURNAL I July-August 1983 281
Typical ranre for 2.0
I• reinforced concrete I
4 r I· Typical ranre
5 3 ~ for steel
~-~4
14~-~3 1.5
3,..- -,..._
(3 f-2 ----L: r___
-- ---6

-
5
Ay = 400 11 1
:::!
Q
2 1- Curve Description R/Rn VR u
1 RC-Grade 60 O+L 1.05 0.11
~
2 RC - Grado 40 D+L 1.15 0.14 1.0
3 RC • Grade 60 D+S 1.05 0.11 :'§
1r 4
5
RC • Grede 40 D+S 1.15
Still O+L 1.07
0.14
0.13 .,.
t:
A.
6 Stul O+S 1.07 0.13
0~----L-----L-----L-----~----~----~
0 0.5 1.0 1.5 2.0 2.5 3.0 0.5 R/C, Grade 60 · Flexure
Lo/Dn or Sn/Dn
D+L

Fig. 3 - Safety indexes /3 for gravity loads on beams /3 = _._


o"--_ __,___ 3.0
__. . . .___.___ __.__
(from Ref. 3)
0 2 3 4 5
Ln/Dn
derived to reflect the relative importance of the average
beam or average column. Fig. 4 - Theoretical load and resistance factors for
6. Designers' familiarity with method used. The re- flexure in reinforced concrete beams - Dead plus live
load (from Reference 3)
sistance factor approach is currently used in the ACI
Code and is widely understood.
Based on a review of these items, the structural ac-
tion resistance factor or cJ> factor has been recom- ues are slightly more conservative than the values in
mended for use in material standards in the United current practice. The lower values for wind loading and
States. 3 particularly for seismic loading can be rationalized in
part by the fact that under wind and seismic loads a
number of cross sections, rather than one isolated
Selection of the method used to compute load member, must fail before the structure fails. In the case
and resistance factors of seismic loadings the load modeling is more vague,
The calculations of load and resistance factors pre- and possibly more important, a much higher degree of
sented have been carried out using an ''advanced first damage is allowed than under gravity or wind loading.
order, second moment procedure". This procedure, In the calculation of individual factors, the target (3
which gives a good approximation of the true reliabil- values given previously will be increased by 0.5 for
ity, originally presented by Rackwitz and Fiessler,l 2 is members displaying sudden, brittle failures (compres-
described in References 3 and 4. sion failures in columns, etc.) to reflect the increased
danger to occupants due to the lack of warning of fail-
Selection of target reliability ure.
Safety indexes inherent in current American practice
for various load combinations for typical members are Derivation of load factors
summarized in References 3 and 4. Representative val- Using the first order, second moment calculation
ues of (3 for gravity loads are in the range of 2.8 to 3.2 procedure with the target values of (3, it is possible to
for reinforced concrete beams, as shown in Fig. 3, and compute resistance factors ct> and load factors 'Y for
range from 3.5 to 4 for plant pretensioned beams with various loading cases. One particular set is presented in
low to moderate steel percentages, dropping to 3.1 for Fig. 4 for (3 = 3 and various ratios of live and dead
very high steel percentages. For tied columns failing in load. The resulting factors are not constant. Thus for a
compression the (3 values ranged from 3 to 3.5; while live/dead ratio of 0.5 the values of ¢, 'Yv• and 'YL are
for tied columns failing in tension (3 values of 2. 75 to 0.78, 1.15, and 1.5, respectively, while for live/dead =
3.5 were found. For shear failures the (3 values ap- 1.5 they have changed to 0.85, 1.08, and 1.88. As a
peared to be much lower, ranging from about 2.25 to particular load becomes dominant in the combination,
2.4. For load combinations involving wind or earth- its theoretical load factor increases.
quake and for counteracting loads D- W, the (3 values Studies of this type suggested that the values of the
appeared to be lower as discussed in References 3 and load factors at given live/dead or snow/dead ratios
4. were very similar and not especially sensitive to the type
Based on these studies, the target (3 values chosen for of material being considered. Furthermore, the magni-
the purpose of deriving load factors were 3.0 for (dead tude of the dead load factor 'Yv was relatively constant
+ live load) and (dead + snow load); 2.5 for (dead + in all the combinations, suggesting that the dead load
live + wind load); and 2.0 for (dead - wind load) and factor and resistance factor could be set independently
(dead + live + earthquake load). In general, these val- of the factors for variable loads.
282 ACI JOURNAL I July-August 1983
Dead load factor U = 1.2D + 1.3W + 0.5L (13)
The load factor for dead load in Fig. 4 ('Yv = 1.05 to
1.21) is much lower than in any existing or proposed U = 1.2D + 1.5£ + (0.5L or 0.2S) (14)
standard because the variability in the dead load itself
was small compared to other load variabilities. The U = 0.9D - (1.3 W or 1.5£) (15)
value of 'Yv was set at 1.20, believed to be the lowest
value that the profession would accept.
Other loadings
When only dead load was present, studies based on {3
Although the studies reported in References 3 and 5
= 3.0 gave 'Yv = 1.28 to 1.33 for ¢ = 0.80 and 'Yv =
considered only dead, live, snow, wind, and earth-
1. 35 to 1.41 for ¢ = 0. 85. On this basis a separate
quake loads, the ACI Code includes load factors for
loading case U = 1.4D was selected.
fluid loads F, soil loads H, and imposed deformations
T. Similarly, the AISC Specification includes ponding
Live and snow load factors load P. Finally, prestressing forces, classified as dead
From the reliability calculations, the values of 'Yv· 'YL• load in ANSI A58.1-1972, are present in many concrete
and¢ were found to be variable, as shown in Fig. 4 and members. Due to the absence of data on the variability
discussed earlier. As a result, if constant values are of these loading cases, load factors were arbitrarily as-
chosen for design, the resulting reliability will vary. signed based on the following arguments.
Linear programming was used to calculate values of 'Yu For F = weight and pressure of fluids with well-de-
'Ys• and ¢ to minimize the difference between the nom- fined maximum level, the variability of the maximum
inal resistance corresponding to the target {3 value ob- load is small, approaching that of dead load. The cur-
tained from the reliability calculation and that given by rent ACI load factor is the same as that for dead load.
the proposed set of load and resistance factors. 3 Although difficulties have been reported arising from
Results of this calculation suggested that for rein- this low value, these have stemmed from filling the tank
forced concrete beams the optimum values of¢ and 'YL with a heavier fluid, failure to consider differential set-
or 'Ys based on {3 = 3.0 and 'Yv = 1.2 were 0.82 to 0.93 tlements T, or inadequate attention to the limit state of
watertightness. Later in this paper a value of ¢ = 0.85
and 1.56 to 1.93, respectively. If 'Yv and 'YL or 'Ys were
set equal to 1.2 and 1.6, the optimum values of ¢, for is proposed for flexure. Based on this value, the value
reinforced concrete beams and {3 = 3.0, ranged from of 'YF corresponding to current practice would be 1.4 x
0.85/0.9 = 1.32. A value of 1.3 has been suggested.
0.81 to 0.86. On this basis the basic gravity load com-
binations were chosen to be For H = weight and pressure of soil and water in
soil, the variability of the maximum load can be high,
as for example, when ornamental landscaping leads to
(8)
unexpected mounds of earth where the design called for
a uniform fill. The current load factor on H in the ACI
(9) Code is 1. 7, the same as live load. It is proposed that
this correspondence be maintained and that a value of
and as chosen earlier 1.6 be used in the future.
For T = effects of imposed deformations such as
U = 1.4 D (10) differential settlements, shrinkage, temperature change,
etc., there is again no statistical data on which to base
a selection. The effect of imposed deformations on a
For members supporting both snow and live load, the
structure depends on the limit state in question. For
probability of the simultaneous occurrence of maxi-
ductile limit states such as flexure, the internal forces
mum snow load and maximum live load is very low.
due to imposed deformations are dissipated by inelastic
Probabilistic studies indicated this was best handled by
deformations leaving the ultimate strength unaffected.
introducing two new loading combinations
For brittle limit states such as shear or for serviceabil-
ity limit states such as cracking, stresses due to im-
(11) posed deformations will tend to proportionately reduce
the failure load. The ACI Code currently suggests con-
(12) sideration of U = 1.05D + 1.05 T + 1.28L or U =
1.4(D + n. In the proposed loading criteria the load
For live loads with a large sustained fraction the 0.5 factors for Tare taken equal to those for D.
factor in Eq. (12) becomes 1.0. Prestressing force has not been considered as a sepa-
rate loading case in the proposed load factor relation-
Wind and earthquake load factors ships, in effect, setting the load factor on prestress
Similar calculations were carried out for combina- equal to 1.0. This is in agreement with current Ameri-
tions of D, L, W, S, and E as outlined in References 3 can practice but is in conflict with the CEB Model
and 5. Calculations involving wind assumed a target {3 Code 10 which suggests that prestress should be multi-
of 2.5 and those involving earthquake a target {3 of 2.0. plied by 0. 9 if the prestress is beneficial and 1.2 if it
The resulting load combinations were unfavorably affects the limit state in question.
ACI JOURNAL I July-August 1983 283
10

0.8

04 Tension Failures

f3 ~ 3.0
0.2

05 o~--~O~l--~0~2~--0~3~--~0~4--~0~5--~0~6~--0~7~

p/pb Fig. 6 - Resistance factors for tied columns failing in


compression
Fig. 5 - Resistance factors for reinforced concrete
beams points represent weighted values for the most common
loading cases for various types of flexural members. It
can be seen that the level of cJ> drops significantly when
When a beam is prestressed, two general types of plpb exceeds 0.5. This is because asp approaches, Pb the
self-equilibrating force systems are introduced into the chance of compression failures in beams increases. 8
member. The first of these systems is the isostatic or This occurs, for example, if the concrete strength is
primary force system in which a compressive force in lower than specified and the steel strength is higher
the concrete equilibrates the tension in the tendons. At than specified.
ultimate load, the internal resisting couple consists of a It is proposed that the maximum value of pi Pb be
compressive force in the concrete and a tensile force in limited to 0.60 and that ¢1 (where f stands for flexure)
the steel separated by a lever arm. For a given cross be set equal to 0.85 for reinforced concrete beams in
section, the magnitude of these forces at ultimate is es- flexure. Similar studies for prestressed concrete beams
sentially independent of the level of prestress since the suggested that ¢ 1 should be 0.85, and except for beams
prestress force has, in effect, disappeared. Clearly, with produced in a certified plant, the value could be raised
respect to the ultimate limit state in flexure in a simply to 0.90.
supported beam, the primary or isostatic prestressing
forces should have a load factor of 1.0. Columns
In a statically indeterminate beam the deflections in- Reinforced concrete columns derive their strength
duced by eccentric prestressing may cause changes in partly from concrete and partly from the reinforce-
the beam reactions leading to a second set of self-equi- ment. In the compression failure range the failures are
librating forces and moments known as the hyperstatic brittle with little warning, while in the tension failure
or secondary forces. These forces are important at the region the failures are initiated by yielding of the rein-
serviceability limit state but tend to disappear near ul- forcement and hence are more ductile. For this reason
timate load if plastic hinging and moment redistribu- a target (3 = 3.5 was selected for computing cJ> factors
tion can occur. In checks of serviceability limit states for compression failures of tied columns, a target (3 =
and of ultimate limit states for heavily reinforced 3.0 for pure moment failures, and values between 3.0
beams, it may be desirable to factor the secondary ef- and 3.5 in the tension failure range. Fig. 6 is a plot of
fects of the prestress. the computed values of cJ> for compression failures. The
The prestressing force affects the shear strength of width of the band of cJ> values results from the different
concrete beams. Here also it may be desirable to factor variabilities of concrete and steel. In this instance the
the prestressing force using factors less than 1.0 when CEB system of material understrength factors would
the force has a beneficial effect on the capacity. reduce the scatter. A value of ct>c = 0.65 has been cho-
In these recommendations, the load factor for pre- sen for this case.
stressing force has been set equal to I. 0 for the ulti- Spiral columns have considerable ductility after
mate limit state because this is the correct value for the reaching their maximum capacity. A value of ct>c = 0. 70
primary or isostatic prestress force system and because has been selected for such columns based on a target (3
the effects of the secondary force system tend to dis- = 3.0 throughout the entire range of compression and
appear as the ultimate limit state is approached. tension failures.
In the tension failure region it is proposed that ct>c for
reinforced concrete tied columns vary linearly with ax-
Derivation of resistance factors ial load from 0.65 at the balanced load P.b to ¢ 1 = 0.85
Reinforced concrete beams in flexure when P. = 0. If P.b is less than zero, the value of ct>c
The range of ct> values for various loading cases is would jump from 0.65 to 0.85 at P. = 0. A similar
shown in Fig. 5 as a function of the ratio pi Pb of the variation is proposed for spiral columns starting from
longitudinal steel ratio to that at balanced failure. The 0.70.
284 ACI JOURNAL I July-August 1983
1.0 10·
f3 = 3.5 [For all e/h :S eb/h]

0.8 08

0. 6 h(7':":Jc::::::::::=:::::::::::=:::::==- ~ __<..-....:--..:.....~----
........ ~
f = 5 kSI
¢ 06 ¢=0.65
¢
O ) _ J o i s t s , one way slabs and shallow beams designed
• We1ghted Values 4 for vc and no stirrups

02

QL-----~~----~~----~------~----
0 05 1.0 1.5 20 0 50 100 150
L/D
Required rfy (psi)

Fig. 7- Resistance factors for tied columns Fig. 8 - Resistance factors for shear in beams

In the case of beams subjected to small axial com- Testing of final results
pressive loads, studies have shown that it is satisfactory The final stage in developing a set of load and resis-
to base the design flexural resistance on ¢ 1 times the tance factors for design is the comparison of designs
nominal moment resistance for flexure alone, ignoring carried out under the old and new systems to see if any
the increase in capacity due to axial loads. Thus, it is major changes occur. This stage is currently underway.
not necessary to consider the effect of small axial loads
on ¢ 1. DISCUSSION OF PROPOSED LOAD FACTORS
The moment magnifier equation for slender columns There are four major reasons for adopting new load
includes a resistance factor ct>s to account for variability and strength reduction factors.
in the stability calculation. Based on Monte Carlo 1. Common load factors for concrete, steel, and
studies of slender columns, it is proposed that a con- wooden structures will simplify the design of structures
stant value of ct>s = 0.65 be used for all slender col- of more than one material, such as a steel frame braced
umns regardless of whether they are tied or spiral or by concrete shearwalls, because the same factored
have P. less than P.b· loads will apply throughout the structure. The pro-
posed load factors are incorporated in the 1982 ANSI
A58.1 building design load specification6 and are slated
Shear in beams to be included in the upcoming AISC "Specification
The words "shear failure" refer to a family of fail- for the Load and Resistance Factor Design, Fabrica-
ure modes; some of which are related only in that shear tion and Erection of Structural Steel for Buildings."
forces are present. No completely satisfactory mechan- 2. Studies of present ACI load factors and strength
ical model exists for any of these modes, and to com- reduction factors indicate that reliability decreases as
plicate matters further, the ACI Code design equations the live to dead load ratio increases. 3 •4 To partially
for shear in beams and punching shear in slabs have overcome this problem, the ratio of the load factors on
been shown to be poor representations of the true live and dead load has been increased by 10 percent.
strengths. Similarly, for the present load factors the reliability also
Fig. 8 shows values of ct>v calculated for beams with decreases for very low live to dead load ratios. A spe-
varying amounts of web reinforcement for (3 = 3.0. cial load combination, Eq. (10), has been proposed for
The calculated values range from 0.42 to 0.49 for this case.
members satisfying ACI Code Sections 11.5.5.1(a) to 3. Structures with low live to dead load ratios de-
(c) which have been designed for v" = ct>vc. No account signed for wind using the current ACI Code Eq. (9-2)
has been taken of possible load sharing as implied in have a much lower reliability than for gravity loads. 3
the Commentary to ACI Code Section 8.11.8 since this This results from the use of 0. 75 x 1.4D = 1.05D as
is offset by a higher value of vc. For beams governed by the load factor on dead load. Eq. (11) provides a gen-
the minimum web reinforcement provisions, the calcu- erally higher and more consistant reliability than cur-
lated values of ct>v range from about 0.9 for beams with rent Eq. (9-2).
minimum stirrups and v" = ct>vv/2 down to a low of 4. Snow loads and live loads due to use and occu-
0.52 for a beam with minimum stirrups and v" = vc + pancy are statistically independent. As a result, the
50 psi. Lines representing ct>v = 0.65 and 0. 70 are shown probability that both will reach a maximum at the same
in Fig. 8. A value of ct>v = 0.70 is proposed as a com- time is very low. This is recognized in Eq. (11) and (12)
promise between these results and existing practice. which respectively consider the combination of maxi-
At present there is little or no definitive data on the mum live load and frequent snow load, and the com-
variability of punching shear in slabs. A rudimentary bination of maximum snow load and frequent live load.
study in Reference 8 suggests that this variability is This allows savings, particularly in the design of struc-
high, and on this basis the same ¢, factors are pro- tures with relatively light, long-span prestressed con-
posed for punching shear as for beam shear. crete roof systems.
ACI JOURNAL I July-August 1983 285
DISCUSSION OF PROPOSED RESISTANCE tawa), V. 3, No.4. Dec. 1976, pp. 484-513.
FACTORS 2. Ravindra, Mayasandra K., and Galambos, Theodore V., "Load
and Resistance Factor Design for Steel," Proceedings, ASCE, V. 104,
In general, the proposed resistance factors are ap- ST9, Sept. 1978, pp. 1337-1353.
proximately (1.2D + 1.6L)/(1.4D + 1.7L) times those 3. Ellingwood, Bruce; Galambos, Theodore V.; MacGregor, James
currently given in the ACI Building Code. The major G.; and Cornell, C. Allin, "Development of a Probability Based
exception is the value for shear which is considerably Load Criterion for American National Standard A58," NBS Special
smaller than would be expected from the ¢ values cur- Publication No. 577, National Bureau of Standards, Washington,
D. C., 1980, 222 pp.
rently used in the ACI Code. This results from the high
4. Galambos, Theodore V.; Ellingwood, Bruce; MacGregor, James
variability of shear strength which is a function of the G.; and Cornell, C. Allin, "Probability-Based Load Criteria: Assess-
variability of the tensile strength of concrete and also ment of Current Design Practice," Proceedings, ASCE, V. 108, ST5,
results from the poorly understood mechanism of shear May 1982, pp. 959-977.
failure. 5. Ellingwood, Bruce; MacGregor, James G.; Galambos, Theo-
dore V .; and Cornell, C. Allin, "Probability-Based Load Criteria:
The need for a reduction in the resistance factor for Load Factors and Load Combinations," Proceedings, ASCE, V. 108,
shear, while debatable, appears to be borne out by the ST5, May 1982, pp. 978-997.
incidence of shear failures observed in practice. Chap- 6. "Building Code Requirements for Minimum Design Loads in
ter 12 of Reference 13, which deals with causes of fail- Buildings and Other Structures," (ANSI A58.1-1982), American Na-
ures of buildings in service states: tional Standards Institute, New York, 1982, 100 pp.
7. ACI Committee 318, "Building Code Requirements for Rein-
"Shear is, however, the most common mode of fail-
forced Concrete (ACI 318-83)," American Concrete Institute, De-
ure in reinforced concrete buildings. In cast-in-place troit (to be published).
concrete buildings, two types of failure are most prev- 8. MacGregor, J. G.; Mirza, S. A.; and Ellingwood, B. R., "Sta-
alent: punching shear and diagonal tension. Punching tistical Analysis of Resistance of Reinforced and Prestressed Con-
shear usually is associated with flat-slab or flat-plate crete Members," ACI JouRNAL, Proceedings V. 80, No.3, May-June
1983, pp. 167-176.
construction, and is in fact practically the only way
9. National Building Code of Canada, National Research Council
such structures will fail. Diagonal tension failures, on of Canada, Ottawa, 1980, 546 pp.
the other hand, are associated with beam-slab elements 10. CEB-FIP Model Code for Concrete Structures, 3rd Edition,
and have been responsible for many problems in rein- Comite Euro-lnternational du Seton/Federation Internationale de Ia
forced concrete." Precontrainte, Paris, 1978, 346 pp.
II. Turkstra, Carl J., and Madsen, Henrik 0., "Load Combina-
tions in Codified Structural Design," Proceedings, ASCE, V. 106,
STI2, Dec. 1980, pp. 2527-2543.
PROPOSED IMPLEMENTATION OF NEW LOAD 12. Rackwitz, R., and Fiessler, B., "Note on Discrete Safety
AND RESISTANCE FACTORS Checking when Using Non-Normal Stochastic Models for Basic Var-
Due to the lack of design aids and computer pro- iables," Loads Project Working Session, Massachusetts Institute of
Technology, Cambridge, June 1976, 20 pp.
grams based on the new load and ¢ factors, it seems
13. "Short Course on Investigation of Structural Failures, Course
desirable that a transition period be provided during Content Notes," Notes prepared for ASCE, Wiss, Janney, Elstner
which both the current values or the new values are ac- and Associates, Inc., Northbrook, May 1976, 93 pp.
ceptable. For this reason it is proposed that the new
safety provisions be incorporated in an Appendix to the APPENDIX- SUGGESTED CODE STATEMENT
ACI Code and moved into the body of the code in a It should be noted that this proposal has no legal or
future edition. A proposed text is included as an Ap- other status until incorporated in the ACI Code.
pendix to this paper. APPENDIX X- ALTERNATIVE REQUIRED AND DESIGN
STRENGTH FACTORS
X.O-Notation
ACKNOWLEDGMENTS D= dead loads or related internal moments and forces
E = load effects of earthquake or related internal moments and
The major portion of this work was carried out and financed by the
forces
Center for Building Technology, National Bureau of Standards,
F = loads due to weight and lateral pressure of fluids with well de-
Gaithersburg, Maryland. The work at NBS was organized by B. Ell-
fined densities and controllable maximum heights or related in-
ingwood and involved Dr. Ellingwood, Professors C.A. Cornell and
ternal moments and forces
T.V. Galambos, and the author. Additional work was carried out at
H= loads due to weight and lateral pressure of soil and water in soil
the University of Alberta, Edmonton, Canada, by Dr. S.A. Mirza
or related internal moments and forces
and the author, and was financed by the National Research Council
L = live loads due to use and occupancy or related internal moments
of Canada. The support of these bodies is gratefully acknowledged.
and forces
The computing suport ofT. Reinhold and C. Mullin of the NBS staff
L,= roof live loads or related internal moments and forces
and S.E. Evison at the University of Alberta is particularly appreci-
P = loads, forces, and effects due to ponding
ated.
R = rain loads or related internal moments and forces
Thanks are due to the members of the load factor subcommittee of
S = snow loads or related internal moments and forces
ANSI A58 and Subcommittee 4 of ACI 318 who reviewed this work.
T = deformation induced forces and effects arising from contrac-
In particular, the help, guidance, and criticism of J .E. Breen, E.O.
tion or expansion resulting from temperature changes, shrink-
Pfrang, and C.P. Siess are appreciated.
age, and moisture changes; creep in component materials; and
movement due to differential settlement or combinations thereof
W= wind loads or related internal moments and forces
REFERENCES </>,= strength reduction factor for bearing
I. MacGregor, James G., "Safety and Limit States Design for cf>.= strength reduction factor for axial compression and flexure with
Reinforced Concrete," Canadian Journal of Civil Engineering (Ot- axial compression

286 ACI JOURNAL I July-August 1983


tf>1 =strength reduction factor for flexure X.2.6 - Where ponding P is significant in design, structures, struc-
tf>, =strength reduction factor for precast concrete connections tural members, and connections shall be designed for the load com-
tf>/X=strength reduction factor for plain concrete binations in Section X.2.2 with the factored force 1.2P added to Eq.
tf>,= strength reduction factor for slenderness effects (9-3) as follows
tf>,= strength reduction factor for axial tension and flexure with ax-
ial tension U = 1.2D + 1.6 (L., S, orR)
tf>,= strength reduction factor for shear and torsion + (0.5L or 0.8W)+ 1.2 P (X-13)

X.l- General X.2.7 - The load factor on L in Eq. (X-3), (X-4), (X-5), (X-8),
(X-9), (X-II), and (X-13) shall be taken as 1.0 for parking structures,
X.l.l- Structures, structural members, and connections shall be areas occupied as places of public assembly, and all areas where
proportioned to have design strengths at least equal to required specified live load exceeds 100 psf before live load reductions are
strengths calculated in accordance with provisions of this code using considered.
load factors and strength reduction factors </>.

X.1.2 - In lieu of the load factors and strength reduction factors X.2.8 - If resistance to impact effects is taken into account in de-
given in Sections 9.2 and 9.3, structures, structural members, and sign, such effects shall be included with live load L.
connections may be proportioned for adequate strength using the load
factors and strength reduction factors given in Sections X.2 and X.3. X.3- Design strength

X.1.3 - Strength design shall be based entirely on the set of load and X.3.1 -Design strength provided by a member, its connections to
strength reduction factors in Sections 9.2 and 9.3 or on the set of load other members, and its cross sections in terms of flexure, axial load,
and strength reduction factors in Appendix X. shear, and torsion, shall be taken as the nominal strength calculated
in accordance with requirements and assumptions of this Code, mul-
X.1.4- Members shall also meet all other requirements of this code tiplied by a strength reduction factor cf> stipulated in Section 9.3.2 and
to insure adequate performance at service load levels. 9.3.3.

X.2 - Required Strength


X.3.2 - Strength reduction factor tf> shall be as follows:
X.2.1 -Strength required by a member, its connections to other
members, and its cross sections, in terms of flexure, axial load, shear, X.3.2.1-Fiexure or flexure with axial tension:
and torsion shall be based on factored loads and forces calculated in (a) Non-prestressed members:
accordance with the factored load combinations stipulated in Sec- (i) with p or (p - p') not greater than 0.6p, ........... </>1 = 0.85
tions X.2.2 to X.2.2.5. The most unfavorable effects on the struc- (ii) with p or (p - p') greater than 0.6p, .. ............. cf>1 = 0. 75
ture, structural member, or connection shall be considered. The most (b) Prestressed members
unfavorable effects may occur when one or more of the variable con- (i) with w; (w + w, - w') or (w. + w,. - w:) not greater
tributing loads are not acting. In those load combinations which in- than 0.25 ....................................................... </>, = 0.85
clude L,, S, or R, the load which produces the most unfavorable ef- (ii) with w,, (w + w, - w') or (w. + w,. - w:) greater than
fect shall be used. 0.25 .............................................................. </>, = 0.75
(c) Precast prestressed members manufactured under plant control
X.2.2 - Required strength U to resist factored loads and forces shall conditions:
be as follows </>1 may be taken 0.05 greater than the values in (b).

u 1.4D (X-1)
u 1.2D + l.6L + 0.5 (L., S, orR) (X-2) X.3.2.2-Axial tension and flexure with axial tension ....... </>, = 0.85
u 1.2D + 1.6 (L,, S, orR) + (0.5L or 0.8W) (X-3)
u 1.2D + l.3W + 0.5L + 0.5 (L, S, orR) (X-4) X.3.2.3-Axial compression and flexure with axial compression:
u 1.2D + 1.5£ + (0.5L or 0.2S) (X-5) (a) Members with spiral reinforcement conforming to Section
u 0.9D - (1.3 W or 1.5£) (X-6) 10.9.2 .................................................................. </>, = 0.70
(b) Other members ................................................. </>, = 0.65
X.2.3 - Where deformation induced forces T may be significant in (c) For compression members with axia~ load capacities P. less than
design, structural members and connections shall be designed for the the nominal balanced axial load capacity P.,, </>, may vary linearly
load combinations in Section X.2.2 with the factored force 1.2T with P. from the applicable value in (a) or (b) to the applicable
added to Eq. (X-I) and (X-2) as follows value in X.3.2.1 when P. = 0; except that when PN ~ 0.051: A,,
the design flexural strength may be taken as </>1 times the nominal
u l.4D + 1.2T (X-7) moment resistance for flexure alone.
u 1.2 (D + T) + l.6L + 0.5 (L, S, orR) (X-8)
X.3.2.4-Shear and torsion ......................................... </>, = 0. 70
X.2.4 - Where loads due to fluid Fare significant in design, struc-
tural members and connections shall be designed for the load com-
X.3.2.5-Bearing on concrete ...................................... </>, = 0.60
binations in Section X.2.2 with the factored forces l.3F included in
Eq. (X-2) and (X-6) as follows
X.3.2.6-Stability (in Section 10.11.5) ........................... </>, = 0.65
u 1.2D + 1.3F + 1.6L + 0.5 (L, S, orR) (X-9)
u 0.9D + (1.3 W or 1.5£) - 1.3F (X-10) X.3.2. 7-Piain concrete ............................................ cf>p< = 0.55

X.2.5 - Where soil pressures are significant in design, structures, X.3.3 - Design strength provided by connections between precast
structural members, and connections shall be designed for the load members or between precast and cast-in-place members shall be taken
combinations in Section X.2.2 with the factored force 1.6H included as the nominal strength for the structural effects considered, multi-
in Eq. (X-2) and (X-6) as follows plied by a modified strength reduction factor. The modified strength
reduction factor shall be taken as the product of the appropriate
u 1.2D + 1.6 (L + H) and 0.5 (L, S, or R) (X-II) strength reduction factor from Section 9.3.2 and a strength reduction
u 0.9D - (1.3 W or 1.5£) - 1.3F (X-12) factor for connections cf>, taken equal to 0.85.

ACI JOURNAL I July-August 1983 287

Você também pode gostar