WT2017 Compresion
WT2017 Compresion
Email address
[email protected]
Citation
Yuwen Li. A Closer Examination of Design Approaches for Eccentrically Loaded WT Shapes in
Compression. American Journal of Civil and Environmental Engineering.
Keywords Vol. 2, No. 5, 2017, pp. 37-56.
WT Shapes, Abstract
Eccentrically Loaded, Design strengths of Structural Tees (WT-shapes, Tees) in axial compression are provided
Eccentric Loading, in Table 4-5 of the 14th edition of AISC Manual of Steel Construction (the Manual) [1],
Compression, with an assumption that the compression force is applied at the center of gravity (c.g.) of
Design Tables the cross section. However, in engineering practice, Tees are rarely loaded at the section’s
c.g.; rather, they are eccentrically loaded at the ends through the gusset plates that welded
or bolted to their flanges. Therefore, these Tees are subject to the axial compressive force
and the bending moment induced by the eccentric end connections. Although the design
Received: July 25, 2017 strengths of eccentrically loaded WT-shapes are not provided in the Manual, they can be
Accepted: August 14, 2017 calculated based on AISC Specifications for Structural Steel Constructions (the
Published: September 20, 2017 Specifications) [2], albeit the design process is tedious and time consuming. However, if
the Specifications are indiscriminately followed (called Approach 1 in this paper), the
calculated strengths are too conservative. This paper discusses an alternate approach
(Approach 2) allowed for by Commentary H2 of the Specifications, and the calculated
strengths based on Approach 2 are larger and more reasonable than those based on
Approach 1. Extensive Finite Element Nonlinear Analyses have been employed to
validate the proposed approach. Finally, design tables for eccentrically loaded WT-shapes
are provided to help engineers quickly determine the proper size of a WT-shape for their
project.
1. Introduction
WT-shapes are commonly used in the bracing system to resist large axial loads and/or to
satisfy the KL/r requirement for large unbraced lengths. When used as braces, Tees are
often connected to a gusset plate through their flanges which create an eccentric
connection due to an offset between the neutral axis of the Tee and the gusset plate.
Although braces are treated as truss members in the structural analysis, the induced
bending moment on a Tee due to the eccentric connection must be considered in the
member design. This additional bending moment is neither considered by design tables in
the 14th edition of AISC Manual of Steel Construction [1], nor by Design Examples,
Version 14 [3].
Gordon [4] presented Tables for Eccentrically Loaded WT Shapes in Compression,
based on the 13th edition of AISC Manual of Steel Construction (AISC 2005), but it is the
author’s opinion that while the available strengths from Gordon’s tables are conservative
(called Approach 1 in this paper), more capacity is available if a different design approach
is used. This paper discusses an alternative approach (Approach 2) permitted by the
Specifications [2] that yields larger and much needed capacities.
38 Yuwen Li: A Closer Examination of Design Approaches for Eccentrically Loaded WT Shapes in Compression
Section Wt. (lb/ft) Connected element Outstanding Element φcPn (kips) Reference
WT6x17.5 17.5 6.56x0.52 5.48x0.30 48.8 Example
76.2 Sec. E5 of the Specifications
L6x6x7/16 17.3 6.00x0.44 5.56x0.44
62.4 Table 4-12 of the Manual
Design Approach 2 in this article. Sxc, section modulus of the flange Sxc = Iy / y
4.1. Available Flexural Stresses Fcbx - Design The nominal flexural strength of the Tee, Mnx, under the
Approach 1 limit state of lateral torsional buckling should be calculated
according to Equation 9-4 of the Specifications [2].
According to Section E9 of the Specifications [2], for Tees The nominal flexural strength of the Tee member, Mnx,
loaded in the plane of symmetry, the nominal flexural strength under the limit state of flange local buckling in flexural
Mnx shall be the lowest value obtained according to the limit compression should be determined according to Section F9.3
states of yielding, lateral torsional buckling, and flange local of the Specifications [2], based on the slenderness of the
buckling, and the available flexural stress Fcbx can be flange. The limit state of flange local buckling does not apply
calculated as follows: for Fy = 36 ksi, since all 273 WT-shapes that included in the
ϕb M nx AISC Shapes Database V14.0 have compact flanges. There
Fcbx _ stem = (5) are only 10 WT-shapes that have non-compact flanges in
Sx
flexural compression for Fy = 50 ksi, they are listed in Table 2.
ϕb M nx These WT-shapes are excluded from this paper. Therefore, the
Fcbx _ flange = (6)
S xc limit state of flange local buckling is not a concern of the
discussion.
where, Sx, section modulus of the stem, Sx = Iy /(d-y)
Table 2. WT with non-compact Flanges (Fy = 50 ksi) (λrf = 24.1 < λ < λpf = 9.15).
(Figure 6). The singly symmetric I-Shaped members bent about used by Approach 2 is incorporated in the Microsoft Excel
their major axis are covered in Sections F4 and F5 of the spreadsheet to generate the design tables.
Specifications [2] depending upon the compactness of the web.
However as stated in Section F4, I-shaped members that are
applicable to Section F4 may be designed conservatively using
Section F5, therefore Section F5 will be used in this paper to aid
the evaluation of the available critical stresses in the flange.
5.3.2. The Critical Compressive Stress, Fcr, The available flexural stress, Fcbx_flange based on the limit
Based on the Limit States of Torsional state of yielding in flange that in compression
and Flexural-Torsional Buckling
Fcbx _ flange = ϕb Fy = 0.9(36.00) = 32.40 ksi
(Section E4 of the Specifications [2])
The controlling critical compressive stress, Fcr The available flexural stresses, Fcbx_stem and Fcbx_flange
Fcr = min(Fcrx, Fcry, Fcr) = 25.08 ksi ϕb M nx 0.9(1905.03)
Fcbx _ stem = = = 530.8 ksi, not control;
The available axial compressive stress, Fra Sx 3.23
ϕb M nx 0.9(1905.03)
Fca = φc Fcr = 0.9(25.08) = 22.57 ksi Fcbx _ stem = = = 138.2 ksi, not control
S xc 12.31
5.4. The available Flexural Stress, Fcbx_stem,
Fcbx_flange 5.4.3. The Nominal Flexural Strength, Mnx
Based on the Limit State of Flange
5.4.1. The Nominal Flexural Strength, Mnx Local Buckling (Section F9.3 of the
Based on the Limit State of Yielding for Specifications [2])
Stem in Tension (Section F9.1 of the
Specifications [2]) Check the compactness of flange for flexure,
Approach 1:
bf 6.56 E 29, 000
= = 6.3 < λ r = 1.0 = 1.0 = 28.4
2t f 2(0.52) Fy 36
American Journal of Civil and Environmental Engineering 2017; 2(5): 37-56 43
5.5. The Available Flexural Stress, Fcby Actual flexural stress fbx, due to Mrx
6.2. Available Flexural Stresses in Flanges 7.1. Finite Element Analysis of Centrically
Loaded Members
As discussed in the Design Considerations, the nominal
flexural strength Mnx shall be the lowest value obtained according AISC [1] provides axial capacities for centrically loaded
to the limit states of yielding, lateral torsional buckling, and W-shapes and WT-shapes in compression. In order to test the
flange local buckling, and the available flexural stress in flange FE modeling techniques and serve as a benchmark to analyze
Fcbx_flange can be calculated based on Equation (5). eccentrically loaded WT-sections, FE analyses were first
Excluding ten (10) WT-shapes, flange local buckling is not performed to analyze centrically loaded doubly symmetric
applicable to WT-shapes. The available stresses in the flange I-shape W14x82 and centrically loaded singly symmetric
based on the remaining two limit states – yielding and lateral WT6x17.5. Two element types were utilized in these analyses:
torsional buckling are presented in Figure 11. It can be concluded three Dimensional Semiloof Cross Section Beam (BXL4) and
that for WT6x17.5 with Grade 36 steel and unbraced lengths as quadrilateral Thick Shell (QTS8), elements used in the study
shown, the yielding, rather than lateral torsional buckling, are sufficiently refined and not to be discussed in this paper [5].
controls the available stresses in the flanges. Therefore, in order Figure 12 shows axial capacities of W14x82 obtained from FE
to design a Tee member more efficiently, it is critical to analysis and from Table 4-1 of the Manual [1]; Figure 13 shows
distinguish the section modulus for the flange and stem when axial capacities of WT6x17.5 obtained from FE analysis and
determining the yielding strength of the Tee under combined from Table 4-7 of the Manual [1]. Both Figure 12 and Figure
axial compression and flexural compression on the flanges. 13 indicate that finite element analysis using both beam element
46 Yuwen Li: A Closer Examination of Design Approaches for Eccentrically Loaded WT Shapes in Compression
and shell element yields very similar capacities for both 7.2. Finite Element Analysis of Eccentrically
methods for doubly symmetric I-shape and singly symmetric Loaded Tees
WT-shape. It can also be concluded that finite element analysis
can very accurately predict the axial capacity of both I-shapes Since finite element nonlinear analysis using either beam
and WT-shapes with large unbraced length, and that when the element or shell element yields similar capacities of
unbraced length is small, the axial capacities obtained from centrically loaded Tees, shell element (QTS8) was used to
Table 4-1 and Table 4-7 of the Manual [1] are conservative; the investigate eccentrically loaded WT - shapes, flange elements
conservativeness is greater with WT-shapes than with I-shapes. are taken at mid-thickness of the flange. The following
discussion is for: WT6x17.5 with unbraced length of 10 feet,
Gr. 36 steel, eccentrically loaded with an eccentricity of
1.6125”, and the initial imperfection is 0.125”.
Nominal Axial Capacity, Stress Distributions and Load vs. 40% of stem reach yield stress of 36 ksi (Fy), while the
Displacement Curve: The nominal axial capacity based on compressive stresses in flange are slightly lower than Fy.
nonlinear finite analysis is Pr = 87.23 kips, therefore, the Figure 17 shows the stress distribution before and after the
factored available strength φc Pr = 0.90*87.23 = 78.51 kips, load reaches the maximum load of Pr, the maximum tensile
which is slightly larger than 70.56 kips from the calculation stress in the stem reaches Fy when the load is at 0.80Pr, while
based on Approach 2. the compressive stresses in the flange is only 24 ksi. With the
Stress distribution along the stem (local y-y axis) at increase of applied load, more area of the stem is yielded, and
mid-height of the tee member is shown in Figure 16 and 17. stress in flange is increased, the compressive stress in flange
Figure 16 shows the stress distribution at the maximum load eventually reaches the Fy.
of Pr = 87.23 kips. The averaged nodal tensile stresses within
Load vs. mid-height displacement curve is shown in Figure 18. Linear variation is observed when the applied load is less than
0.80Pr, and the stresses in the WT6x17.5 are less than Fy (also see Figure 17). The load vs. displacement curve indicates
continual strength gain when more area of the stem reaches Fy. A 1.0-inch displacement is noted when WT6x17.5 reaches its
ultimate strength.
7.3. Axial Capacity Comparison Between WT-shapes based on Approach 2 are consistently larger than
Approach 2 and Nonlinear Finite Element those based on Approach 1, they are still less than the results
Analysis from this finite element analysis and are therefore
conservative. The conservativeness of the axial capacity of
eccentrically loaded WT-shapes based on Approach 2 is in
line with the axial capacity of centrically loaded WT-shapes
per Table 4-7 of the Manual [1], as shown in Figure 13.
8. Conclusions
This paper proposed an alternate approach (Approach 2)
allowed for by Section Comm. H2. of the Specifications [2] to
design the eccentrically loaded WT-shapes in Compression,
and the calculated strengths based on Approach 2 are larger
and more reasonable than those based on Section H2 of the
Specifications [2] (Approach 1). Extensive Finite Element
Nonlinear analyses have been employed to validate the
proposed approach. Design tables to facilitate implementation
of the procedure of Approach 2 have also been developed.
Only a few experimental tests on eccentrically loaded
WT-shapes in compression are available. The capacities of
two 15 ft long WT15x11 braces tested by Tito [9] agree very
well with the calculated capacities based on the proposed
approach; the percent difference between the experimental
Figure 19. Axial Capacities of WT6x17.5. factored average capacity and the factored capacity predicted
by Approach 2 is 94.85%.
The results of the finite element analysis and calculated The results from finite element analyses and experimental
axial capacity of WT6x17.5 based on both approaches tests clearly indicate that the proposed approach is quite
discussed in the paper are presented in Figure 19. As indicated accurate in predicting the ultimate capacity of eccentrically
previously, while the axial capacities of eccentrically loaded loaded WT-shapes in compression.
American Journal of Civil and Environmental Engineering 2017; 2(5): 37-56 49
References [6] LUSAS Bridge Software, Version 15.2 (2016), FEA Ltd,
Kingston-Upon-Thames, UK.
[1] AISC (2010), Manual of Steel Construction, 14th Edition,
American Institute of Steel Construction, Inc., Chicago, [7] Galambos, T. V. (2001), “Strength of Singly Symmetric
Illinois. I-Shaped Beam-Columns,” AISC Engineering Journal, 2nd
Quarter, 2001, pp. 65-77.
[2] AISC (2010a), Specification for Structural Steel Buildings,
June 2010, American Institute of Steel Construction, Inc., [8] Ellifritt, D. S., et al. (1992), “Flexural Strength of WT Sections,”
Chicago, Illinois. AISC Engineering Journal, 2nd Quarter, 1992, pp. 67-74.
[3] AISC (2011), Design Examples, Version 14, 2011, American [9] Tito, J. A. (2013), “Eccentric Compression Test of WT Shape
Institute of Steel Construction, Inc., Chicago, Illinois. Steel Braces,” 11th LACCEI Latin American and Caribbean
Conference for Engineering and Technology, 2013, Cancun,
[4] Gordon, M. E. (2010), “Tables for Eccentrically-Loaded WT Mexico.
Shapes in Compression,” AISC Engineering Journal, 2nd
Quarter, 2010, pp. 91-100. [10] Li, Y. W. (2012), “Axial Capacities of Eccentrically Loaded
Equal-Leg Single Angles: Comparisons of Various Design
[5] Rhodes, S. and Cakebread, T. (2013), “Understanding Methods,” AISC Engineering Journal, 4th Quarter, 2012, pp.
Buckling Behavior and Using FE in Design of Steel Bridges,” 131-167.
International Bridge Conference, Pittsburg, 2013.