Design of 8-bolt

Stiffened Moment End Plates

THOMAS M. MURRAY and ANANT R. KUKRETI

This paper presents two design procedures for the back of the end-plate was not permitted; nodes showing
8-tension boh, stiffened, extended, moment end plate tension forces were released and new analyses con-
shown in Fig. 1. The resuhing end-plate design will be sat- ducted.
isfactory for use in Type 1 construction (rigid-frame) as The accuracy of the finite element model was then veri-
defined in Ref. 1 or, with appropriate modifications, as a fied with two series of tests. First, six stiffened tee-hanger
FR type (fully restrained) connection as defined in the tests were conducted. Each tee-hanger specimen consisted
LRFD design notation.^ With bolt size limited to a maxi- of two A36 steel tee-stubs connected by four rows of two
mum of IVi in. dia., the configuration is capable of devel- Vs-in. dia. A325 bolts. The specimens were loaded using
oping the full moment capacity of most available hot- a 200-kip capacity universal testing machine by applying
rolled beam sections. The two design procedures are load to the tee stems. Measurements were made to deter-
limited to use with A36 steel and A325 bolts. mine strains on and separation of the tee-stub flanges and
bolt shank strains. Experimental strains and displace-
ments were compared to predictions from the finite ele-
OVERVIEW OF ANALYTICAL AND ment model and were found to be in close agree-
EXPERIMENTAL STUDIES ment.^'^'^
The design procedures presented here are the result of ex-
tensive analytical and experimental studies on the behav-
ior of 8-bolt, stiffened, moment end plates.^ First, a hy-
brid 2D/3D finite element model was developed."^'^'^ It
was assumed the beam tension flange and a symmetrical
portion of the end plate acts as a stiffened tee hanger, as
shown in Fig. 2. One quarter of this section was then ana-
lyzed using the finite element model in Fig. 3. The end
plate and the beam flange-to-end7plate welds were mod-
eled using three dimensional (3D) subparametric ele-
ments. The stiffener and beam flange were modeled using
two dimensional (2D) elements. Both the bolt heads and
shanks were modeled with 3D elements. The nodes for the
shank and center portion of the bolt head were kept sepa-
rate to more accurately model actual behavior. Nonlinear
behavior of both the plate material (assumed to be elastic-
perfectly plastic) and bolt material (bi-linear) were in-
M
cluded in the analysis. Tension at support nodes on the

Thomas M. Murray is the Montague-Betts Professor of Civil

Engineering, Virginia Polytechnic Institute and State Univer-
sity, Blacksburg, Virginia.
Anant R. Kukreti is an Associate Professor, School of Civil Engi-
neering and Environmental Science, University of Oklahoma,
Norman, Oklahoma.
Fig. L 8-tension bolt stiffened moment end-plate

SECOND QUARTER / 1988 45

 Tee-stub Stiffener
Analogy

Beam Flange

Tee-Stub Flange
(End-Plate)

(a) End-Plate Connection (b) Stiffened Tee-Stub

Fig. 2. Stiffened tee-stub analogy

Pretension Force
Far Bolt Shank
Bolt Head

Near Bolt Shank -^^

-7
X

(a) Model of a Tee-Hangar (b) Details of Model

Fig. 3. Configuration of stiffened tee-hanger
finite element model

To further evaluate the finite element model, eight end- and vertical deflections were measured. The connections
plate connection tests were conducted. Each test specimen were also analyzed using the finite element model. Figure
consisted of two beam sections with end plates at each 4 shows selected results for one of these tests. These
end. The sections were bolted together and tested under curves show that the finite element tee-hanger model gives
pure moment, developed by a symmetric two-point load- results that compare well with full connection test results.
ing applied using a spreader beam. Separation of the end A similar conclusion was reached for each of the other full
plates near the beam tension flanges, tension bolt strains connection tests.

46 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

 70 Table 1. Practical Ranges for
Various Geometric Parameters
60 Parameter Low Intermediate High
tp (in.) 1/2 P/4 3
50 Pf (in.) IVs 13/4 21/2

db (in.) Vs 1 11/2
o
cc
o 40 .^'- 1
ts (in.) 5/16 1/2

8 (in.) 31/2 51/2 71/2

S 30 1- bp (in.) 6 10 16

20 h

F.E.M. (Near Bolt) Table 2. Practical Ranges for End-plate Thickness

10 . TEST (Near Bolt In)
TEST (Near Bolt Out) Corresponding to Various Bolt Diameters
End-plate Thickness
100 200 300 400 500 600 Bolt Diameter Minimum Maximum
MOMENT (in.) in. in.
(a) Bolt Force versus Moment Results
Vs Vi 11/4

3/4 Vi 1^2

700 - Vs 5/8 13/4

^,''
1 Vs 2
600 " •.-''' IVs 3/4 21/4
500 - VA 1 21/2

400 11/2 1 3

300 - /
200 eters. Only A36 steel end-plate material and A325 bolts
F.E.M.
.TEST
were considered. The six independent geometric variables
100 were then reduced to five dimensionless parameters. The
ij_ ^ normalizing variable was chosen as the end plate width bp.
0 —J 1 , „ 1 1

0.02 0.04 0.06 0.08 0.1 0.12 Once the five independent dimensionless parameters
PLATE SEPERATION were established, ranges for each of the six geometric pa-
(b) Moment versus Plate Separation Results rameters were selected based on usual detailing practice,
F/g. 4. Typical experimental and analytical results Table 1. Also, limitations were placed on the combina-
tions of bolt diameter and end-plate thickness as in Table
2. Based on these ranges, "low," "intermediate" and
"high" values of each of the dimensionless parameters
DEVELOPMENT OF THE were calculated. Using combinations of the three ranges of
BASIC DESIGN PROCEDURE the dimensionless parameters, 21 end-plate configurations
To develop equations which predict end-plate deflection, were selected. In addition, four special and extreme cases
end-plate strains and bolt force response to loading, sensi- were formulated using the smallest and largest values of
tivity and parametric studies were first conducted. The flange width. Thus, the resulting 25 cases bracketed all
sensitivity study was used to determine the most signifi- reasonable end-plate design configurations.
cant geometric and force related variables that govern Next, the 25 cases were analyzed using the previously
connection behavior. The following six geometric and two described finite element model. The following criteria was
force variables were found to be significant: tp = end-plate used to complete the detailing: (1) All edge distances were
thickness pf = distance from the face of the beam flange to 1.75 db. (2) The end-plate width equaled the beam-flange
the centerline of the nearer bolts (i^= nominal bolt diame- width. (3) The distance between bolt rows on the same
ter ts = end-plate stiffener thickness g= gage of vertical side of the flange was set at IVs d^. And (4), the length of
bolt lines bp = end-plate width F^ = factored beam flange the 45° profile stiffener was made to extend beyond the far
force and P^ = bolt pretension force. All other dimen- bolt centerline by a distance equal to the bolt diameter. In
sions necessary to define an end-plate configuration were the finite element analyses, the flange force was apphed
determined from these six independent geometric param- (after pretensioning of the bolts) in increments of y2oth of

S E C O N D QUARTER / 1988 47
the ultimate capacity of the eight bolts until failure oc-
curred. Failure was defined to occur when the ratio of the
secant modulus and the elastic modulus of the plate mate-
rial became equal to or less than 0.1 or when a bolt strain
reached a value of 0.00693 in./in.
Regression analyses, using results from the finite ele-
ment analyses together with the five non-dimensional pa-
rameter terms for each of the 25 cases, were then con-
ducted to generate prediction equations for maximum
plate separation, maximum end-plate strain and maximum
bolt force. The three best fit equations (with least square
fit values of 0.961, 0.979 and 0.988) were then rearranged
for design use as follows:

^ 0.00553 pf^-^^'g^''' F,
p\ 0.682 (1)
d,'
0.257 „0.148 17 1.017
_ 0.00371 p}>-^^' g"'^^ f,
(2)
d^ p 0.319
_ 2.305 X 10"^5 p„ /0.591 77 2.583
0.885 J 1.909 ^0.327 L. 0.965 + Pi (3)
d.
Fig. 5. End-plate tension flange area geometry
where tpi and tp2 are required end-plate thicknesses for a
maximum (factored) flange force F^, T^ is the correspond-
ing bolt force, and other variables are as previously de- BASIC DESIGN PROCEDURE AND EXAMPLE
fined. Equation 1 was developed from the end-plate dis- This allowable stress design procedure is only valid for A36
placement prediction equation using a limit of 0.02 in. to steel end-plate material and A325 bolts. Figure 5 defines
ensure sufficient stiffness for use in Type 1 or FR construc- the geometry.
tion."^ Equation 2 was developed from the prediction equa-
tion for maximum end-plate strain and Eq. 3 from the pre- Allowable Stress Design Procedure:
diction equation for maximum bolt force. The maximum 1. Select beam size.
bolt force T^ includes prying action effects and bolt pre-
tension Pr- Obviously, the larger of tpi and tp2 would be 2. Compute beam flange force F\
used to select a final end-plate thickness. The specified F = M/(d-tf),
minimum tensile strength of the bolt material (88 ksi for where
A325 bolts based on nominal bolt area) and r„ are used to M = beam end moment
determine the required bolt diameter. d = beam depth
For allowable stress design use, a factor of safety of 1.67 tf = beam flange thickness
is introduced into Eqs. 1 and 2 by substituting 1.67Ffor
F^, where F = beam tension flange force (unfactored). 3. Determine single bolt force T assuming 6.8 bolts are
Similarly, a factor of safety of 2.0 is introduced into Eq. 3. effective:
The resulting allowable stress design equations are: T = F/6.8

0.873 ^0.577 17 0.917 4. Determine the required A325 bolt diameter and select
_ 0.00885 PI'-^^^ g"-'^^ F bolt size:
^p\ (4)
d^ = V(4r/(Tr F,) with F, = 44 ksi
0.00625 p^Q-^^^g^-^^^F^Q^^
tp2 = (5) Or select bolt diameter from Table 1-A, p. 4-3, of the
AISC Manual.^
1.381 X 10-4^^591^2.583
T = + Pi (6) Select gage g, pitch pf, end-plate width bp and stiff-
ener plate thickness t^. The actual end-plate can be of
any width, but the width bp must not exceed the beam
In the application of Eqs. 4 and 5, a preliminary bolt di- flange width plus 1 in. or the actual width. The gage g
ameter is selected assuming that 6.8 of the 8 tension bolts must not exceed the beam flange width. The stiffener
are effective. This ratio must often be decreased depend- plate thickness should be approximately the same
ing on the results of Eq. 6. thickness as the beam web.

48 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

 6. Determine tpi from stiffness criterion (Eq. 4): Calculate tp2.
. 0.873 ^0.577 ^0.917
0.00885 pf"'"' g" ^ 0.00625 (1.5)^-2^^ (5.5)01^^ {221 Af''^'
tpi = ^^0.924 ^^0.112 ^^0.682 ^^' " {l.^f™ (0.5)^1^2 (9.0)^-^1^
7. Determine tp2 from strength criterion (Eq. 5): = 1.24 in.
0.00625 p ^0.257
" ^ ^ ^_0.148 c-l.Oiy
g'""''F Select end-plate thickness:
fp2 = ^^0.719 ,^0.162 ^ ^ 0 1.18 in.
tp = IV4 in. > max.
8. Select end-plate thickness, tp-. 1.24 in.
^pi
tp > max. Calculate ultimate bolt force:
-P2 1.381 X 10"^ (1.5)Q-^^i (227.4)2-^^'
9. Calculate ultimate bolt force (Eq. 6): " " (1.25)^-««^ (1.0)1-90^ (0.5)^-^2^ (9.0)0-96^ ^
^ 1.381 X 1 0 - 4 p 0.591 ^2.583 = 26.53 + 51 = 77.53 kips
^u 0.885 J 1.909 ^0.327 L, 0.965 ^ -* T
^p ^b Check bolt size:
10. Check bolt size T^ = 77.53 kips > 2 x 34.6 = 69.2 kips. n.g.
r . < 2 X Tallow Redesign is necessary.

where T^^xo^ = tension allowable load from Table Redesign

1-A, p . 4-3, of the AISC Manual.^ Increase bolt size to IVs-in. dia.
11. Redesign if necessary T^aiiow = 43.7 kips
Two possibihties exist. Either the end-plate thickness Pf = IVs + 1/2 = IVs in. pt = 3 X IVs = 33/8 in.
can be increased which reduces T^ or a larger bolt di- As previous tpi = 1.13 in. ^^2 = 1-16 in.
ameter can be selected. Use tp = 1^4 in.
Check bolt capacity i
^ _ 1.381 X 10-4 (1.625)0-^^1 (227.4)2-^^3
Example: \0.965 + 56
u ( 1 25)0-885 (1.125)1-909 (0.5)0.327 (9 Q)
Design an 8-tension bolt moment end plate to develop the
allowable stress design moment capacity of a W24 x 94, = 22.21 + 56 = 78.21 kips < 2 x 43.7
A36 steel, beam under gravity loading = 87.40 kips
W24 X 94 A36 steel bf = 9.065 in. d = 24.31 in. Therefore, IVs in. dia. bolt is satisfactory
tf = 0.875 in. t^ = 0.515 in.
M = M, = 444 ft /kips (AISC Manual,^ p . 2-7) Final design

Flange force: F = {AAA x 12)/(24.31 - 0.875) W24 X 94 A36 M = 444 ft/kips

= 227.4 kips End plate: VA x 9 in.
Stiff ener: V2 in.
Estimate single-bolt force: T = 227.4/6.8 = 33.4 kips Bolts: 8-11/8 in. dia. A325
Trial bolt size: Select 1-in. diameter A325 bolts g = 5^2 in. Pf = IVs in.
Pb = 3 % in.
fallow = 34.6 kips (AISC Manual,^ p . 4-3) > 33.4 kips

Select end-plate geometry:

diy = I'm. pf = I + Vi = IV2 in.
g = 5^2 in. bp = 9 in. DEVELOPMENT OF
Pb = 3 in. ts = V2 in. SIMPLIFIED DESIGN PROCEDURE

(Note: stiffener thickness is approximately the beam web Because of the difficulty of using Eqs. 4, 5 and 6, except
thickness, 0.515 in.) for completely computerized designs, an additional effort
was made to develop a simplified allowable stress design
Calculate tp^. procedure. First, end-plate connection designs were gen-
^ 0.00885 (1.5)^-^^' (5.5)Q-'^^ (227.4)'\0.917 erated using Eqs. 1, 2 and 3 for all hot-rolled A36 steel
^Pi - (1.0)0-924 (0.5)0-112 (9.0)0-^«2 beam sections at 100%, 7 5 % and 50% of full moment ca-
pacity (M = 0.66 F^ Sx). The effective number of tension
= 1.18 in. bolts was then computed from

SECOND QUARTER/1988 49
 2F/8 3. The pitch Pfirom the face of the beam tension flange to
Neff = N (7) the first row of bolts must not exceed IVi in. The rec-
ommended pitch is bolt diameter plus V2 in.
where N^ff = the effective number of tension bolts, F = 4. The spacing p^ must not exceed 3J^.
unfactored beam flange force, 7^ = maximum boh force 5. The gage g must not be less than 3V2 in. nor greater
calculated using Equation 3 with F^ = 2 F a n d N = 8. From than IV2 in.
this procedure, it was found a conservative value for the 6. Stiffener thickness t^ must be approximately equal to
effective number of bolts is six. Note a factor of safety of the beam web thickness.
2.0 was used to determine N^ff. 7. Bolt diameter must not be less than VA in. nor greater
Next, an equivalent pitch p ^ ^ was found so classical tee- than Wi in.
stub type calculations could be used for determining re-
quired end-plate thickness rather than Eqs. 1 and 2. From Simplified Allowable Stress Design Procedure:
Fig. 6, assuming an inflection point at p^Jl and two bolts
1. Select beam size.
per row, the tee-stub flange moment is
2. Compute beam flange force F:
Mpr^ = 2T{pfl2) = Tpf (8)
F = Ml{d-tf),
where T — force per bolt based on six effective bolts. The
where
required section modulus and tee-stub flange thickness
M= beam end moment
can then be calculated. To determine the corresponding
d = beam depth
required end-plate thickness p^ff is substituted for pf in
tf = beam flange thickness
Eq. 8.
To determine p^ff, numerous expressions were devel- 3. Determine single bolt force T:
oped and evaluated as follows. First, the required end-
plate thickness was determined for all hot-rolled beam T= F/6
section capacities (subject to hmitations included in the 4. Determine the required A325 bolt diameter and select
foflowing design procedure) using the tee-stub analogy bolt size:
(Fig. 5) and p^ff- The required end-plate thickness was
then rounded to the next highest Vsth in. and the capacity d^ = V(4r/(7T Fr) with F, = 44 ksi
of the connection determined from the minimum F„ calcu- Or select bolt diameter from Table 1-A, p. 4-3, of the
lated by rearranging Eqs. 1, 2 and 3 and using factors of AISC Manual.^
safety of 1.67 for Eqs. 1 and 2 and 2.0 for Eq. 3. Bolt size
was determined using N^ff = 6. The resultant allowable 5. Select gage g (not to exceed the beam flange width),
stress design capacity F^ap was then compared to the re- pitch Pf and compute effective pitch, p^ff.
quired unfactored flange force F. The following expression
for/?g^y^ resulted in values of FIF^ap less than 1.05 for all but Peff- W~^' Pf
19 of the 726 cases examined. The maximum FIF^ap ratio
was 1.10. Note that a value of FIF^ap greater than 1.0 is 6. Compute effective tee-stub analogy moment M^.
unconservative.
M^ = effective moment in plate caused by two bolts
with inflection point at Pef/2 (see Fig. 6).
reff - ^ Pf (6)
2r(peff/2) = Tpeff
The final result is the following simplified procedure for 7. Select effective end-plate width bp (not to exceed beam
determining bolt diameter and end-plate thickness of flange width plus 1 in.) and determine required end-
8-tension bolt, extended, stiffened moment end-plates. plate thickness tp'.
Sj^ = MJ(0J5Fy) with Fy = 36 ksi
SIMPLIFIED DESIGN PROCEDURE AND EXAMPLE
This allowable stress design procedure is only valid for A36
steel end-plate material and A325 bolts and is subject to the Example:
following limitations. Figure 5 defines the geometry. Using the simplified procedure, design an 8-tension bolt
moment end-plate to develop the allowable stress design
1. The connected beam section must be hot-rolled and in-
moment capacity of a W24 x 94, A36 steel, beam under
cluded in the Allowable Stress Design Selection Table
gravity loading. (Same as previous example).
in the 8th ed. AISC Manual.^
2. The effective end-plate width, e.g. the end-plate width W24 X 94 A36 steel bf = 9.065 in. d = 24.31 in.
used in the design calculations, must not be greater tf = 0.875 in. t^ = 0.515 in.
than the beam flange width plus 1 inch. M = M, = 444 ft./kips (AISC Manual,^ p. 2-7)

50 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

 Final design:
F
W24 X 94 A36 M = 444 ft/kips
End plate: P/s in. x 9 in.
Assumed Tee Stem Stiffener: Vz in.
Inflection Bolts: 8-1V8 in. dia. A325
(Beam Flange) g == SVi in. Pf = P/s in.
Point
Pb = ^Vs in.
Tee Flange The selected stiffener thickness is approximately equal
T -(End-Plate) to the beam web thickness (0.515 in.). The simplified de-
sign procedure resulted in a slightly thicker end plate
(Vs in.), but the same bolt diameter.

' Pf ACKNOWLEDGEMENTS
The research described herein was sponsored by the
American Institute of Steel Construction and performed
M PL at the Fears Structural Engineering Laboratory, Univer-
sity of Oklahoma. Test specimens were provided by Capi-
tol Steel Corporation, Oklahoma City, Okla. and Flint
MpL= (2T)(Pf/2) Steel, Tulsa, Okla. Valuable insight and encouragement
was given by John D. Griffiths, chairman of the AISC
Fig. 6. Tee-stub analogy moments oversight committee, and Gerald B. Emerson, Jim
Wooten and Nestor Iwankiw, members of the committee.
The authors also wish to acknowledge University of Okla-
homa graduate students Vipul Ahuja, Larry Curtis,
Mehdi Ghassemieh and Ali Mazroi, who performed much
Flange force: F = (444 x 12)/(24.31 - 0.875) of the analytical and experimental work.
= 227.4 kips Finally, the authors wish to acknowledge the generous
assistance of Prof. N. Krishnamurthy, who provided the
Single-bolt force: T = 227.4/6.0 = 37.9 kips basic finite element code from which the model described
in the paper was developed.
Bolt size: Select IVs in. dia. A325 bolts
T^aiiow = 43.7 kips (AISC Manual,^ p.4-3) > 37.9 kips REFERENCES
Select end-plate geometry: 1. American Institute of Steel Construction, Inc. Specifi-
cation for the Design, Fabrication and Erection of
di, = IVs in. pf = IVs + Vi = IVs in. Structural Steel for Buildings 1978, New York,
g = 5^2 in. bp = 9 in. N.Y.
Ply = 3diy = SVs in.
2. American Institute of Steel Construction, Inc. Load
and Resistance Factor Design Specification for Struc-
Effective pitch:
tural Steel Buildings 1986, Chicago, III.
Vg2 + p/ V5.5^ + 1.625^ ,, ^^^, 3. Ghassemieh, M., A. R. Kukreti and T. M. Murray
Peff = 3 '^ Pf = ^ (1.625) Inelastic Finite Element Analysis of Stiffened End-
= 1.86 in. Plate Moment Connections Research Report FSELI
AISC 83-02, Fears Structural Engineering Laboratory,
Effective plate moment: June 1983, University of Oklahoma, Norman, Okla.
4. Ghassemieh, M. Inelastic Finite Element Analysis of
Me= Tp,ff= (37.9) (1.86) Stiffened End-Plate Moment Connections Master of
= 70.49 in./kips Science Thesis, School of Civil Engineering and Envi-
ronmental Science, 1983, University of Oklahoma, Nor-
Required end-plate thickness:
man, Okla.
SR = MJ(0J5F) = 70.49/(0.75 x 36.0) 5. Ahuja, V. Analysis of Stiffened End-Plate Connec-
= 2.61 in.^ tions Using the Finite Element Method Master of
Science Thesis, School of Civil Engineering and Envi-
tp = \/6SR/bp = V6(2.61)/9
ronmental Science, 1982, University of Oklahoma, Nor-
= 1.32 in. Use V/s in. plate man, Okla.

SECOND QUARTER / 1988 51

6. Kukreti, A. R., T. M. Murray and M. Ghassemieh Fi- N =
Number of tension bolts, 8.
nite Element Analysis of Stiffened End-Plate Moment ^eff =
Effective number of tension bolts.
Connections with Multiple Bolt Rows Considering Ine- M =
Allowable stress design beam end moment.
lastic Material Behavior Proceedings, SECTAM XII, M, =
Effective moment in plate caused by two bolts
The Southern Conference on Theoretical and Applied with inflection point at /J///2.
Mechanics, Vol I, June 10-11, 1984, Auburn University = Tee-stub flange plate moment.
(pp. 521-526). Pb = Spacing of bolt rows on same side of flange.
7. American Institute of Steel Construction, Inc. Man- Peff = Effective pitch.
ual of Steel Construction 1980, Chicago, III. Pf = Distance from the face of the beam flange to the
centerline of the nearer bolts.
Pr = Bolt pretension force.
NOMENCLATURE = Required section modulus of end-plate.
SR
bp = End-plate width. = Strong axis beam section modulus.
d = Beam depth. = Beam-flange thickness.
di, = Nominal A325 bolt diameter. = End-plate thickness.
F = Unfactored beam-flange force. p\
= Required end-plate thickness to satisfy maximum
Fcap = Allowable stress end-plate connection design ca- allowed end-plate separation criterion.
pacity. = Required end-plate thickness to satisfy maximum
h2
Ft = Allowable tension stress of A325 bolt material, 44 aflowed end-plate strain criterion.
ksi. = End-plate stiffener thickness.
Fj, = Factored beam-flange force. T = Force per tension bolt.
Fy = Yield stress of A36 end-plate material, 36 ksi. T = Bolt force at factored load including prying
g = Gage of vertical bolt lines on end-plate. force.

52 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION