Structural Engineering Department, Ain Shams University, Cairo, Egypt Email: Sherif - Ibraim@eng - Asu.edu - Eg
Structural Engineering Department, Ain Shams University, Cairo, Egypt Email: Sherif - Ibraim@eng - Asu.edu - Eg
Structural Engineering Department, Ain Shams University, Cairo, Egypt Email: Sherif - Ibraim@eng - Asu.edu - Eg
2 Sherif M. Ibrahim
3 Structural Engineering Department, Ain Shams University, Cairo, Egypt
4 Email: [email protected]
5
6 Abstract
7 This paper presents stability analysis of steel frames composed of linearly tapered columns
8 and partially tapered restraining beams. In the first part of the stability analysis, closed-form
9 equations for the bending stiffness of the symmetrical and unsymmetrical partially tapered
10 restraining beam are derived and presented in form of design charts. Subsequently, slope deflection
11 equations of linearly tapered column subjected to axial compressive load are then assembled along
12 with the bending stiffness equations of partially tapered beams to determine the elastic critical load
13 and the effective buckling length factor for non-sway and sway steel frames. Numerical examples
14 and efficient design charts are presented for several braced and unbraced steel frames. Moreover, the
15 effect of gravity loads acting on leaning columns connected to sway frames is investigated. Based on
16 this investigation it is concluded that correction of effective buckling length factor based on story
17 buckling concept can be extended to cover steel frames with tapered column as well.
18 Keywords: stability, effective buckling length, steel frames, tapered columns, partially tapered
19 beams
20 1. Introduction
21 Effective buckling length of steel frames is still one of the main design parameters for
22 column design of steel frames in several design specifications such as AISC [1]. Single-story steel
23 frames composed of fully and partially tapered members result in weight saving for many structural
24 and loading situations. Lee et al. [2] conducted one of the earliest elastic buckling analysis studies on
25 individual tapered columns and frames with tapered columns and prismatic beams. For individual
26 linearly web-tapered column with length (L), they introduced a length factor (g) which enabled the
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
27 calculations of elastic critical load (Pcr) of tapered column based on an equivalent prismatic column
28 with the smallest moment of inertia (I1) and a buckling length (gL) as follows:
𝜋 2 𝐸𝐼1
29 𝑃𝑐𝑟 = (𝑔𝐿)2
(1)
ℎ2
31 In which 𝛾 = −1 (3)
ℎ1
32 Where h2 and h1 are the cross-section depth at the largest and the smallest ends, respectively. For
33 steel frames, Lee et al. [2], provided design charts for the effective buckling length factor of tapered
34 columns restrained by prismatic girders using slope deflection equations. Their research was the base
35 for the provisions of tapered members design included in AISC supplement No. 3 [3]. Lee and
36 Morrell [4] showed that the use of AISC supplement No. 3 can be extended to determine the
37 effective buckling length of frames with linearly fully tapered columns and linearly fully tapered
38 beams as well. Hirt and Crisinel [5] presented an expression for the elastic critical load for
39 symmetrical web-tapered pin-ended column about its major axis based on a concept of prismatic
40 column with the same length but with equivalent moment of inertia (Ieq) as follows:
𝜋 2 𝐸𝐼𝑒𝑞
41 𝑃𝑐𝑟 = (4)
𝐿2
42 where,
𝐼
43 𝐼𝑒𝑞 = (0.08 + 0.92 √𝐼1 ) 𝐼2 (5)
2
44 In which I1 and I2 are the smallest and largest moment of inertia of the linearly tapered column,
45 respectively.
46 Marques et al. [6] derived an Ayrton–Perry formulation for columns with non-uniform cross sections
47 and proposed new design rules for tapered columns. They proposed an amendment to column design
48 equations in Eurocode [7] by introducing additional factor that accounts for the second order effects
49 of tapered columns that was later verified by experimental work of Tankova et al. [8]. Nuguyen et al.
50 [9] used artificial neural network (ANN) to predict the critical buckling load of web-tapered
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
51 columns. They used finite element method to generate a wide database set of 269 specimens to
52 develop ANN model. An ANN-based formula was proposed to estimate the critical buckling load of
54 Effective buckling length of columns in steel frames can be determined by performing stability
55 analysis. Formulation of the stability functions based on slope deflection equations for linearly web
56 tapered I-member subjected to axial compressive loading is well documented in the literature [10-
57 13]. Ermopoulos [14] used these equations to develop design charts for the effective buckling length
58 of frames with tapered columns and prismatic beams. El-Sadder [15] derived exact expression for the
59 stability functions of non-prismatic members with solid rectangular, solid circular and I-cross
60 sections. He used power series approach to solve the fourth-order ordinary differential equation with
61 variable coefficients. Raftoyiannis and Ermopoulos [16] studied the elastic stability of web tapered
62 and stepped columns with initial imperfections of parabolic shape according to EC3 [7]. Their
63 formulation was based on the exact solution of the governing differential equations for columns with
64 web-tapered cross sections. King et al. [17] derived a stiffness modification factors for linearly or
65 symmetrical partially tapered solid rectangular girders restraining prismatic columns in braced or
66 unbraced frames. They used these factors along with AISC alignment charts to calculate the effective
67 buckling length for frames restrained by partially tapered girders with solid rectangular cross section.
68 Bazeos and Karabalis [18] introduced an approximate method to calculate the elastic buckling load
69 for tapered columns with different end conditions and restrained by prismatic beams using simple
70 design charts. Their design charts were based on best fitting curves of wide range of web-tapered
71 columns solved by exact formulation that was presented by Karabalis and Bazeos [19]. From these
72 design charts, the location of a critical cross section of the web-tapered column was determined
73 based on the tapering ratio and the column’s end conditions. The critical buckling load was
74 determined based on the moment of inertia of this critical cross section and the well-known buckling
75 length factors. Saffari et al. [20] presented design-oriented charts to calculate the effective buckling
76 length of columns of gabled frames in which both columns and rafters are linearly web-tapered
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
77 members. Their charts were based on stability analysis and utilizing stability functions introduced by
78 Ermopoulos [10]. Serna et al. [21] proposed an approximate procedure to transform tapered column
79 subjected to non-uniform compressive loading into equivalent prismatic column with non-uniform
80 loading. Their approximate procedure has shown its effectiveness for different tapering ratios with
81 errors in the critical buckling load not exceeding 11% of the exact solution. Lu and Meng [22]
82 considered the effect of shear deformations on the stability analysis of steel frames with tapered
83 columns. They derived the exact elements of the stiffness matrix of tapered member from the
84 governing differential equations considering both axial force and shear deformations. They
85 concluded that shear deformation has minimal effect on columns with slenderness ratios greater than
86 50. Salem [23], used the lateral stiffness of unbraced frames with tapered columns to map them to
87 equivalent prismatic columns through closed-form equations. However, his method was limited to
89 Many researchers [19] and [24-27] developed finite element (FE) formulations to determine the
90 elastic buckling loads for tapered columns with general non-prismatic cross sections. These studies
91 cover various aspects of tapered columns boundary conditions, type of cross section and flexibility of
92 connections. Kucukler and Gardner [28-29] developed a stiffness reduction function for axial and
93 bending strength which can be incorporated into finite element analysis. This stiffness reduction
94 approach is able to consider the spread of plasticity, initial imperfections, and residual stresses.
95 Although these FE formulations are accurate, they are not quick tool for design oriented engineers.
96 Single story steel frames are usually constructed from welded I-sections that have constant flanges
97 and variable web depth. To optimize the frame design, columns are linearly tapered while beams are
98 partially tapered as shown in Fig. 1. A partially tapered beam is a beam that has web-tapered
99 segments connected to another prismatic segments. The available literatures [17] and [30] only
100 address the modification of bending stiffness of partially tapered girders with solid rectangular cross
101 section. Therefore, in this paper, stability analysis of braced and unbraced steel frames composed of
102 linearly tapered columns and partially tapered beams made from I-section is presented. Calculations
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
103 of the bending stiffness of partially tapered symmetrical and unsymmetrical beams are presented in
104 form of closed-form equations. Afterward, this bending stiffness is incorporated into stability
105 analysis of steel frames with linearly tapered columns to develop efficient and simple design charts
106 for the effective buckling length factor of these columns. Moreover, these design charts can be
107 applied to unbraced frames with leaning columns by applying the story buckling approach. The
108 effective buckling length is an important factor that is used to evaluate column available flexural
109 buckling strength using AISC [1] effective length method provided that all the limitations of this
111 Lb Lb
I2c I2c I2c
112 I2b I2b I2b
I1b I1b
113 a Lb a Lb a Lb
114
Lc
Lc
115
I1c I1c
I1c
116
117 a- Frame with symmetrical partially tapered beam b- Frame with unsymmetrical partially tapered beam
118 Fig.1. Single story frames with partially tapered beams and tapered columns
120 This section presents the derivation of the bending stiffness of partially tapered I-beam that provide
121 restraining effect to columns in frames permitted to sway or prevented from sway. The ratio between
122 the bending stiffness of partially tapered I-beam and that of prismatic I-beam with the smallest
123 moment of inertia represents a modification factor that can be utilized in the stability analysis of steel
126 Symmetrical partially tapered I-beam is generally used in girders with rigid ends. Thus, for sway
127 frames, the beam is bent in double curvature while for no sway frames the beam is bent in single
128 curvature [31]. The variation of the moment of inertia of linearly fully tapered I-beam, shown in Fig.
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
129 2 with respect to x-axis origin at virtual point of zero cross section is provided in [10,14 and 32] as
130 follows:
𝑥 2 𝑥 2
131 𝐼𝑥 = 𝐼1𝑏 [𝐴] = 𝐼1𝑏 [(𝑏−1)𝐿 ] (6)
𝑏
𝐼
132 Where, A is the distance from the smallest end to x-axis origin, 𝑏 = √ 2𝑏⁄𝐼 and 𝐼1𝑏 & 𝐼2𝑏 are
1𝑏
133 moment of inertia of the tapered beam at the smallest and the largest ends of the beam, respectively.
134 Similar equations to Eq. (6) can be used for the tapered parts of a symmetrical partially tapered beam
135 but considering the shift in the partially tapered beam x-axis origin shown in Fig. 3 as follows:
𝑥 2
136 𝐹𝑜𝑟 0 ≤ 𝑥 ≤ 𝛼𝐿𝑏 ∶ 𝐼𝑥 = 𝐼1𝑏 [𝑏 + (1 − 𝑏) 𝛼𝐿 ] (7)
𝑏
(1−𝑏) 𝑥 2
138 𝐹𝑜𝑟 (1 − 𝛼)𝐿𝑏 ≤ 𝑥 ≤ 𝐿𝑏 ∶ 𝐼𝑥 = 𝐼1𝑏 [1 − (𝐿 − (1 − 𝛼))] (9)
𝛼 𝑏 y
x
139
x Ix o
140
Lb I1b A
I2b
141
143
y
Lb
144
x
I1b o
145 I2b I2b
a Lb a Lb
146
149 In sway frames both ends of the restraining beam are subjected to equal end moments in the same
150 direction [31] as shown in Fig. 4. The slope moment-rotation equation in this case can be written as
151 follows:
1 𝐿 𝑀 𝐿 𝑀
152 𝜃= (∫0 𝑏 𝐸𝐼𝑥1 𝑥 𝑑𝑥 − ∫0 𝑏 𝐸𝐼𝑥2 𝑥 𝑑𝑥) (10)
𝐿𝑏 𝑥 𝑥
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
153 where, 𝜃 is the beam end rotation.
154 The variation of the bending moments along the beam span can be expressed as:
𝑀 𝑀
155 𝑀𝑥1 = 𝐿 𝑥 𝑎𝑛𝑑 𝑀𝑥2 = (𝐿𝑏 − 𝑥) (11)
𝑏 𝐿𝑏
156 Using Eq. (11) and considering the three segments of the symmetrical partially tapered beam, Eq.
160 Substituting Eq. (7) and (9) into Eq. (12) and performing the integration, the moment-rotation
6𝛼2 𝛼+(1−𝛼)(1−𝑏)
164 (1−𝑏)
[1 − ] + 1 − 6𝛼 + 12𝛼 2 − 8 𝛼 3 } (14)
𝑏𝛼
165 Lb
166 M
167 M
x y
168
M
169 x Mx2
Mx1
170 M
171 Fig. 4. Symmetrical partially tapered beam moment-rotation diagram in sway frames.
172 The term 𝛽1𝑠 , represents a modification factor that can be multiplied by the bending stiffness of
173 similar prismatic beam having the smallest moment of inertia (I1b) to obtain the actual bending
174 stiffness of symmetrical partially tapered I-beam which is required for stability analysis of frames
175 permitted to sway. The value of this modification factor is plotted in Fig. 5 for different tapered
176 length ratios ( 𝛼) and different (I2b / I1b) values . The obtained bending stiffness of symmetrical
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
177 partially tapered restraining beam can be used along with the ASIC [1] alignment charts to determine
178 the effective buckling length of prismatic columns in sway permitted frames.
6.5
M Lb
6.0
I2b I1b I2b
5.5
a Lb a Lb
5.0 𝟔𝑬𝑰𝟏𝒃
M= 𝜷𝟏𝒔 𝜽 M
𝑳𝒃
4.5
4.0
β1s
3.5
3.0
2.5
2.0
1.5
1.0
0 0.1 0.2 0.3 0.4 0.5
179
a
180 Fig. 5. Bending stiffness modification factor for symmetrical partially tapered I-beam restraining sway frames.
181
182 2.1.2 Symmetrical partially tapered beams in non-sway frames
183 For frames prevented from sway, the restraining beam is subjected to uniform and constant bending
184 moment (M) along its span as it is bent in single curvature [31]. The moment-rotation equation in
1 𝐿 𝑀 𝑀 𝛼𝐿𝑏 𝑥 (1−𝛼)𝐿𝑏 𝑥 𝐿 𝑥
186 𝜃= (∫0 𝑏 𝐸𝐼 𝑥 𝑑𝑥 ) = 𝐿 (∫0 𝑑𝑥 + ∫𝛼𝐿 𝑏
𝑑𝑥 + ∫(1−𝛼)𝐿 𝑑𝑥 ) (15)
𝐿𝑏 𝑥 𝑏 𝐸𝐼𝑥 𝑏 𝐸𝐼1𝑏 𝑏 𝐸𝐼𝑥
187 The integration in Eq. (15) can be solved considering that (M) is constant and substituting with
188 moment of inertia of the beam from Eqs. (7) to (9) for the three parts of the beam. The moment-
1 2𝛼2 𝛼+(1−𝛼)(1−𝑏)
191 where, = {1 − 2𝛼 + (1−𝑏) [ − 1]} (17)
𝛽1𝑛𝑠 𝑏𝛼
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
192 Fig. 6 illustrates the variation of the bending stiffness modification factor for symmetrical partially
193 tapered restraining beam for frames prevented from sway in terms of the tapered length ratio (𝛼) and
194 the ratio of the largest to smallest moment of inertia (I2b / I1b) of the beam.
3.5
Lb
2.0
1.5
1.0
0 0.1 0.2 0.3 0.4 0.5
a
195
196 Fig. 6. Bending stiffness modification factor for symmetrical partially tapered I-beam restraining no sway frames.
197
198 2.2 Unsymmetrical partially tapered beams with hinged far end
199 Unsymmetrical partially tapered I-beam is generally used in girders with rigid connection at the
200 largest end and hinged connection at the smallest end. Whether used as restraining beam in sway
201 permitted or sway prevented frames, unsymmetrical partially tapered I-beam is bent with single
202 curvature [31]. The variation of the moment of inertia of unsymmetrical partially tapered I-beam
(1−𝑏) 𝑥 2
205 𝐹𝑜𝑟 (1 − 𝛼)𝐿𝑏 ≤ 𝑥 ≤ 𝐿𝑏 ∶ 𝐼𝑥 = 𝐼1𝑏 [1 − (𝐿 − (1 − 𝛼))] (19)
𝛼 𝑏
206 The variation of the bending moment along the beam span with respect to x-axis is given by:
𝑀
207 𝑀𝑥 = 𝐿 𝑥 (20)
𝑏
208 The rotation at the rigid beam end can be determined from the following relationship:
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
1 (1−𝛼)𝐿𝑏 𝑀 𝐿 𝑀
209 𝜃= (∫0 𝑥 2 𝑑𝑥 + ∫(1−𝛼)𝐿
𝑏
𝑥 2 𝑑𝑥 ) (21)
𝐿𝑏 𝐸𝐼1𝑏 𝑏 𝐸𝐼𝑥
210 The above integration can be carried out by substituting the moment of inertia for the tapered
211 segment from Eq. (19) and considering that M and I1b are constants. The moment-rotation
212 unsymmetrical partially tapered I-beam with hinged far end at the smallest end can be written as
213 follows:
6 𝐸𝐼1𝑏 2 𝐸𝐼1𝑏
214 𝑀= 𝛽2𝑠 𝜃 = 𝛽2𝑛𝑠 𝜃 (22)
𝐿𝑏 𝐿𝑏
217 y
Lb
218
x
219 I1b o
I2b
Hinged end
220 M a Lb
221
223 𝛽2𝑠 𝑎𝑛𝑑 𝛽2𝑛𝑠 represent modification factors that can be multiplied by the bending stiffness of
224 similar prismatic beam having the smallest moment of inertia (I1b) to obtain the actual bending
225 stiffness of unsymmetrical partially tapered I-beam with hinged end which is required for stability
226 analysis of frames permitted to sway and prevented from sway, respectively. Fig. 8, shows these
227 modification factors for various tapered length ratio (𝛼) and various values of (I2b / I1b).
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
3.5 Lb
Hinged End
I2b I1b
3.0
a Lb 𝟔𝑬𝑰𝟏𝒃
M= 𝜷𝟐𝒔 𝜽
M 𝑳𝒃
𝟐𝑬𝑰𝟏𝒃
2.5 M= 𝜷𝟐𝒏𝒔 𝜽
b2s = b2ns /3
𝑳𝒃
2.0
1.5
1.0
0.5
0 0.2 0.4 a 0.6 0.8 1
228
229 Fig. 8. Bending stiffness modification factors for unsymmetrical partially tapered I-beam with hinged smaller end
230 restraining sway and no sway frames.
231 2.3 Applications on frames with partially tapered beams and prismatic columns
232 To verify and illustrate the use of the derived bending stiffness modification factors for partially
233 tapered beams, the effective buckling lengths of columns in three steel portal frames shown in Fig. 9
234 are calculated. The cross-section dimensions of beams and columns are listed in Table 1. The
235 calculations are performed using the basic equations of alignment charts instead of reading values
236 from the charts for accuracy reasons. The basic equations to determine the effective buckling length
237 of column in frames permitted to sway or prevented from sway are given in [31] and [33]. These
238 equations can be simplified in case of columns with hinged and fixed bases by substituting 𝐺𝐴 equals
239 to infinity and zero, respectively. Thus, the alignment charts equations can be rewritten for these
𝜋 𝜋 6
241 For sway frames with hinged base: (𝑘 ) tan (𝑘 ) − =0 (24)
𝐺𝐵
𝜋 𝜋
242 For sway frames with fixed base: (𝑘 ) 𝐺𝐵 + 6 tan (𝑘 ) = 0 (25)
𝜋 2 𝜋 𝜋
243 For no sway frames with hinged base: 0.5 (𝑘 ) 𝐺𝐵 + 1 − (𝑘 ) / tan (𝑘 ) = 0 (26)
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
𝐼𝑐 /𝐿𝑐
244 where, k is the effective buckling length factor of column and 𝐺𝐵 = in which Ic is the
𝛽𝑠 𝐼1𝑏 /𝐿𝑏
245 column moment of inertia, Lc is the column length, 𝛽𝑠 represents the relevant bending stiffness
246 modification factor obtained from Eqs. (14), (17) & (23) for partially tapered beam, I1b is the
247 smallest moment of inertia of the beam and Lb is the beam length. Eqs. (24) to (26) can be solved to
248 determine the value of k using iteration technique for any given value of 𝐺𝐵 . The three frames shown
249 in Fig. 9 are solved for different tapered length ratio, 𝛼, for the restraining beam and results are
250 compared with those obtained from linear buckling analysis (LBA) using STAAD.Pro software [34].
251 A numerical model for each of the three portal frames shown in Fig. 9 was constructed using
252 prismatic beam element for columns and the uniform part of the beam while a web-tapered beam
253 element is used for the tapered portion of the beam. An initial gravity load is applied on each
254 column. STAAD.Pro software [34] performs eigenvalue calculations to get buckling factor (BF)
255 which is the amount by which the initially applied loads need to be multiplied by to just cause
256 buckling (i.e. elastic buckling load = Pcr = BF × initial load). Eigenvalue procedure can be
258 1- First, the primary deflections are calculated by linear static analysis based on the provided
260 2- Primary deflections are used to calculate member axial forces. These forces are used to
261 calculate geometric stiffness matrix [Kg]. Both the large deformations effects and the small
265 In LBA, the output of the analysis is the elastic critical load 𝑃𝑐𝑟 , which can be used to determine the
𝑃
267 𝑘 = √𝑃 𝑒 (27)
𝑐𝑟
𝜋 2 𝐸𝐼𝑐
268 In which 𝑃𝑒 = 𝐸𝑢𝑙𝑒𝑟 𝑙𝑜𝑎𝑑 = , where E is the modulus of elasticity.
𝐿2𝑐
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
269 Lb = 20 m Lb = 20 m
Ic
Ic Ic
I2b I2b I2b
270
I1b I1b
a Lb a Lb a Lb
271
Lc = 8 m
Lc = 8 m
272
273 Ic Ic
Ic
274
276 Lb = 20 m
Ic Ic
277 I2b I2b
I1b
a Lb a Lb
278
Lc = 8 m
279
280
Ic Ic
281
282 c- Portal frame 3
283 Fig. 9. Dimensions of steel portal frames (1 to 3) with prismatic columns and partially tapered beams.
284 The results from alignment charts equations based on the modified bending stiffness of partially
285 tapered beam and those from LBA for the column effective length factor of the three steel portal
286 frames are provided in Tables 2 to 4. It is evident from these tables that the proposed modification
287 factors for partially tapered beam along with alignment charts equations provide accurate estimate
288 for the effective length factor compared to the results from LBA as the error percent is less than 1.7
289 % for case of frames permitted to sway and 2.4 % for frames prevented from sway. In order to
290 investigate the effect of partial tapering of beams on the effective length factor, the ratio between k-
291 factor at any given tapered length ratio (𝛼) to k-factor at tapered length ratio = 0, is indicated in
292 Tables 2 to 4. Frames with prismatic beams having the smallest cross section properties (i.e., tapered
293 length ratio = 𝛼 = 0) represents the ideal case where effective buckling length factors can be obtained
294 directly from AISC alignment charts. It is evident from Tables 2 to 4 that partial tapering in beam
295 has significant effect on reducing the effective buckling length factor for case of frames permitted to
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
296 sway (e.g. frames 1 and 2 for which maximum reduction in k-factor can reach 33% compared to the
297 similar frame with prismatic beam having the smallest cross section property, I1b) while this effect is
298 minimal for braced frames (e.g. frame 3 for which maximum reduction in k-factor is only 4%
300 Table 1
301 Dimensions and properties of portal frames 1 to 3 cross sections
Cross section Web dim. Flanges dim. Moment of inertia E (MPa)
Beam cross section 1 300 × 6 150 × 10 𝐼1𝑏 = 85.6 × 106 2.05 × 105
Beam cross section 2 700 × 6 150 × 10 𝐼2𝑏 = 549.6 × 106 2.05 × 105
302
303 Table 2
304 k-factor for frame 1 (symmetrical partially tapered beam)- sway permitted-hinged base
𝛼 𝛽1𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
305
306 Table 3
307 k-factor for frame 2 (unsymmetrical partially tapered beam) – sway permitted-fixed base
𝛼 𝛽2𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
0.4 1.04 5.62 1.53 2639 1.55 -1.07 0.89
308
309
310 Table 4
311 k-factor for frame 3 (symmetrical partially tapered beam)- sway prevented-hinged base
𝛼 𝛽1𝑛𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
312
314 The variation of the moment of inertia for a linearly tapered column can be written using equation
315 analogous to Eq. (6), but with the column notations as follows:
𝑥 2 𝑥 2
316 𝐼𝑥 = 𝐼1𝑐 [𝐴] = 𝐼1𝑐 [(𝑐−1)𝐿 ] (28)
𝑐
𝐼
317 Where, A is the distance from the smallest end to x-axis origin, 𝑐 = √ 2𝑐⁄𝐼 and 𝐼1𝑐 & 𝐼2𝑐 are
1𝑐
318 moment of inertia of the tapered column at the smallest and the largest ends, respectively.
319 The slope deflection equations for linearly tapered member shown in Fig. 10 in the presence of axil
320 compressive force are well documented in [13-14] and can be written in terms of the smallest
2𝐸𝐼1𝑐 𝛿
322 𝑀𝑎 = [𝛼1 𝜃𝑎 + 𝛼2 𝜃𝑏 − (𝛼1 + 𝛼2 ) 𝐿 ] (29)
𝐿𝑐 𝑐
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
2𝐸𝐼1𝑐 𝛿
323 𝑀𝑏 = [𝛼3 𝜃𝑎 + 𝛼4 𝜃𝑏 − (𝛼3 + 𝛼4 ) 𝐿 ] (30)
𝐿𝑐 𝑐
2𝐸𝐼1𝑐 𝛿
324 𝑉 𝐿𝑐 = [−(𝛼1 + 𝛼3 ) 𝜃𝑎 − (𝛼2 + 𝛼4 ) 𝜃𝑏 + (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 ) 𝐿 ] − 𝑃 𝛿 (31)
𝐿𝑐 𝑐
325 The last term in Eq. (31) represents the (𝑃 − 𝛿) effect on columns in sway frames. For completeness,
326 the other coefficient in Eqs. (29) to (31) are given by the following equations [14]:
𝑓4 (𝜇 2 +0.25) 2
328 𝛼2 = 𝛼3 = ℓ (33)
2𝑓0
330 where
ℓ
331 𝑓1 = {0.5 cos[𝜇 ln(1 + ℓ)] − 𝜇 sin[𝜇 ln(1 + ℓ)]} − 0.5 ℓ (35)
√1+ℓ
ℓ
332 𝑓2 = {𝜇 cos[𝜇 ln(1 + ℓ)] − 0.5 sin[𝜇 ln(1 + ℓ)]} − 𝜇 ℓ (36)
√1+ℓ
335 𝑓0 = 𝑓1 𝑓4 − 𝑓2 𝑓3 (39)
𝐿𝑐
336 ℓ= =𝑐−1 (40)
𝐴
𝑃𝐴2 𝑃 𝐿2
337 𝜇 = √𝐸 𝐼 − 0.25 = √(𝑐−1)2𝑐𝐸 𝐼 − 0.25 > 0 (41)
1𝑐 1𝑐
(𝜇 2 +0.25)(𝑐−1)2 𝐸𝐼1𝑐
339 𝑃= (42)
𝐿2𝑐
340 Substituting Eq. (42) into Eq. (31) results in the following equation:
2𝐸𝐼1𝑐 (𝜇 2 +0.25)(𝑐−1)2 𝛿
341 𝑉 𝐿𝑐 = [−(𝛼1 + 𝛼3 ) 𝜃𝑎 − (𝛼2 + 𝛼4 ) 𝜃𝑏 + (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − )𝐿 ] (43)
𝐿𝑐 2 𝑐
342
343
344
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
345
346 x y
347 b Ix a
x o
I2c I1c
348 Lc A
359 A number of assumptions, similar to some of those used to develop AISC [1] alignment charts, are
360 adopted in the current frame stability analysis and they are listed as follows:
𝑃
362 2- The tapered columns stiffness parameter 𝐿𝑐 √𝐸𝐼 is equal for all columns.
1𝑐
363 3- In frames prevented from sway, rotations at opposite ends of restraining beams are equal in
365 4- In frames permitted to sway, rotations at opposite ends of restraining beams are equal in
370 The other assumptions adopted in AISC alignment charts of constant members (for beams and
371 columns) cross sections and rigid end connections are not applicable since the current analysis takes
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
372 into considerations: a) variation in both columns and beams cross sections, and b) the effect of
374 Based on the above assumptions, a sub-assemblage consists of the tapered column and the partially
375 tapered beams at both ends of the column can be analyzed and the equilibrium equations can be
378 In frames prevented from sway, columns are restrained at both ends by restraining beams or by the
379 base support. In case that the restraining beam is partially tapered, the bending moment at the beam
380 end connected to the column can be obtained from Eqs. (16) and (22) according to the beam type.
381 For column prevented from sway, only Eqs. (29) and (30) are used to calculate the sum of bending
𝛿
382 moment at each end of the column and the term (𝐿 ) equals zero. Thus, for a column prevented from
𝑐
383 sway the sum of bending moments at both ends, (a and b), of the column can be written as follows:
387 the sum of the bending stiffness of partially tapered beams connected to column ends (a) and (b),
388 respectively. 𝛽𝑛𝑠 depends on the type of the partially tapered beam and is determined from Eqs. (17)
1
(𝛼1 + 𝐺 ) 𝛼2
𝐴
391 | 1
| =0 (46)
𝛼3 (𝛼4 + 𝐺 )
𝐵
392 Eq. (46) can be furtherly simplified in case that the column has hinged or fixed base at point (a). For
1
393 a column with hinged base the term 𝐺 = 0 and Eq. (46) can be rewritten as:
𝐴
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
𝛼1 𝛼2
394 |𝛼 (𝛼4 +
1
)| = 0 (47)
3 𝐺𝐵
395 For a column with fixed base, 𝜃𝑎 = 0, and only Eq. (45) is used to determine the buckling conditions
396 as follows:
1
397 (𝛼4 + 𝐺 ) = 0 (48)
𝐵
398 Different values of parameter (𝜇) are iterated till the results of Eqs. (46) to (48) are equal to zero for
399 given values of parameters (c) and (𝐺𝐵 ) which account for the tapering in the column and partial
400 tapering in restraining beams, respectively. Once the critical value of parameter (𝜇) is obtained, the
403 where 𝜇𝑐𝑟 is the value of parameter (𝜇) which satisfies Eqs. (46) to (48).
404 Therefore, effective buckling length factor, k, which can be used along with the smallest moment of
408 For linearly tapered column permitted to sway, Eqs. (29), (30) and (43) are used along with the Eqs.
409 (13) and (22) of the partially tapered restraining beams to write the equilibrium equations. The
410 general equation for the critical buckling load of column in sway frame can be written as follows:
3
(𝛼1 + 𝐺 ) 𝛼2 −(𝛼1 + 𝛼2 )
𝐴
| 3 |
411 𝛼3 (𝛼4 + 𝐺 ) −(𝛼3 + 𝛼4 ) =0 (51)
| 𝐵 |
(𝜇 2 +0.25)(𝑐−1)2
−(𝛼1 + 𝛼2 ) −(𝛼3 + 𝛼4 ) (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − )
2
413 of the sum of the bending stiffness of partially tapered beams connected to column ends (a) and (b)
414 respectively. 𝛽𝑠 depends on the type of the partially tapered beam and is determined from Eqs. (14)
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
1
415 and (23). For case of a column hinged at end (a), the term 𝐺 = 0 and Eq. (51) can be simplified as
𝐴
416 follows:
𝛼1 𝛼2 −(𝛼1 + 𝛼2 )
3
417 || 𝛼3 (𝛼4 + 𝐺 ) −(𝛼3 + 𝛼4 ) || = 0 (52)
𝐵
(𝜇 2 +0.25)(𝑐−1)2
−(𝛼1 + 𝛼2 ) −(𝛼3 + 𝛼4 ) (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − )
2
418 For case of a column fixed at end (a), 𝜃𝑎 = 0 , Eq. (51) can be simplified as follows:
3
(𝛼4 + 𝐺 ) −(𝛼3 + 𝛼4 )
𝐵
419 | (𝜇 2 +0.25)(𝑐−1)2
|=0 (53)
−(𝛼3 + 𝛼4 ) (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − )
2
420 Depending on the boundary conditions and the stiffness of the restraining beams at both ends of
421 column (GA & GB) and the tapering column parameter (c) Eqs. (51) to (53) can be solved using
422 iteration technique to obtain the critical buckling load and the effective buckling length factor using
426 Three portal frames with linearly tapered columns and partially tapered beams as shown in Fig. 11
427 are analyzed by stability analysis presented in the previous sections and results are compared with
428 LBA using STAAD.Pro software [34]. A numerical model for each of the three portal frames shown
429 in Fig. 11 was constructed using prismatic beam element for the constant part of the beam while a
430 web-tapered beam element is used for the columns and the tapered portion of the beam. Eigenvalue
431 analysis previously described in Section 2.3 is carried out to determine the elastic critical load for
432 each frame. The cross-section dimensions of beams and columns are listed in Table 5. It is evident
433 Tables 6 to 8 that the proposed modification factor for partially tapered beam along with stability
434 analysis of linearly tapered columns provide accurate estimate for the effective length factor
435 compared to the results from LBA with error percent less than 2.7 % for case of frames permitted to
436 sway and 0.34 % for frames prevented from sway. The partial tapering of beam has substantial effect
437 on the effective length factor for the case of frames permitted to sway with tapered columns (e.g.,
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
438 frames 4 and 5) as the k-factor can be as low as 60 % of the k-factor of similar frames with prismatic
439 beam having the smallest cross section properties as indicated in Tables 6 and 7. On the other hand,
440 as evident from Table 8 the effect of partial tapering of beam on the k-factor for braced frame 6 has a
441 marginal reduction on k-factor by 2.3 % at most compared to similar frame with prismatic beam
443
444 Lb = 20 m Lb = 20 m
I2c
I2c I2c
445 I2b I2b
I2b
I1b
a Lb I1b
446 a Lb
a Lb
Lc = 8 m
Lc = 8 m
447
448
I1c I1c
I1c
449
451 Lb = 20 m
I2c
I2c
452 I2b I2b
I1b
453 a Lb a Lb
Lc = 8 m
454
455
I1c
I1c
456
458 Fig. 11. Dimensions and properties of steel portal frames with partially tapered beams and tapered columns.
459 Table 5
460 Dimensions and properties of portal frames 4 to 6 cross sections
Cross section Web dim. Flanges dim. Moment of inertia E (MPa)
Beam cross section 1 300 × 6 150 × 10 𝐼1𝑏 = 85.6 × 106 2.05 × 105
Beam cross section 2 700 × 6 150 × 10 𝐼2𝑏 = 549.6 × 106 2.05 × 105
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
Column cross section 1 400 × 6 200 × 10 𝐼1𝑐 = 200.13 × 106 2.05 × 105
Column cross section 2 600 × 6 200 × 10 𝐼2𝑐 = 480.13 × 106 2.05 × 105
461
462
463
464 Table 6
465 k-factor for frame 4 (symmetrical partially tapered beam)- sway permitted-hinged base
𝛼 𝛽1𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
466
467 Table 7
468 k-factor for frame 5 (unsymmetrical partially tapered beam) – sway permitted-fixed base
𝛼 𝛽2𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
469
470
471
472
473
474
475
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
476 Table 8
477 k-factor for frame 6 (symmetrical partially tapered beam)- sway prevented-hinged base
𝛼 𝛽1𝑛𝑠 𝐺𝐵 K LBA Error 𝑘𝛼 /𝑘𝛼=0
Pcr (kN) k
478
480 The gabled steel frame shown in Fig. 12 was analyzed by Saffari et al. [20]. They calculated the
481 column buckling length factor (km) based on the moment of inertia of a cross section located at the
482 middle of the column height (Im) for special case of symmetrical partially tapered beam when 𝛼 =
483 0.5. This value of (km) can be transferred to a buckling length factor (k) based on the moment of
484 inertia of the smallest cross section from the following equation:
𝐼 𝑘
486 Therefore, 𝑘 = √ 𝐼1𝑐 𝑘𝑚 = [1+(𝑐−1)/2]
𝑚
(55)
𝑚
487 The length of the restraining beam of the gabled frame is taken as the actual total inclined length of
488 the rafter. The bending stiffness modification factor is calculated from Eq. (14) and considering the
489 tapered length ratio 𝛼 = 0.5. Table 9 shows comparison between the buckling length factor obtained
490 by the current analysis and results from Saffari et al. [20]. It is concluded that the current proposed
491 analysis result for the buckling length factor is almost identical to those of Saffari et al. [20] results
493
494
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
495
H = 2.31 m
I1b
496 I2c
I2b I2b I2c
497
Relative values of moment of inertia
498
Lc = 4 m
501 L=8m
502 Fig. 12. Gabled frame 7: Dimensions and relative cross section properties [20].
503 Table 9
504 k-factor for frame 7 (symmetrical partially tapered beam)- sway permitted-hinged base
𝛼 𝛽1𝑠 𝐺𝐵 Column K Saffari et al. [20] Error
[20] (Eq.55)
505
507 To provide efficient and simple design tool to evaluate the buckling length factor of steel frames
508 composed of linearly tapered columns and partially tapered beams a set of design charts are
509 presented in Fig. 13 to 16. The design charts show the buckling length factor in terms of two main
510 parameters which are the ratio of the largest to smallest moment of inertia of the column and the ratio
511 GB. The effect of the partial tapering in the restraining beams is accounted for by using the proposed
512 bending stiffness modification factors (𝛽1𝑠 , 𝛽2𝑠 , 𝛽1𝑛𝑠 𝑎𝑛𝑑 𝛽2𝑛𝑠 ) which depends on the shape and
513 end conditions of the partially tapered beam. It is evident from these charts that the buckling length
514 factor decreases by increasing the ratio of moment of inertia of the largest to the smallest end of the
515 column.
516
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
517
518
519
520
521
522
523
524
525
526
527
528
529 Fig. 13. Buckling factor for frames prevented from sway with hinged base.
530
531
532
533
534
535
536
537
538
539
540
541
542 Fig. 14. Buckling factor for frames prevented from sway with Fixed base.
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559 Fig. 15. Buckling factor for frames permitted to sway with hinged base.
560
561
562
563
564
565
566
567
568
569
570
571
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
572 Fig. 16. Buckling factor for frames permitted to sway with Fixed base.
574 Leaning columns in sway frames have no resistance to base shear and P-𝛿 effect of the portion of
575 loads on these columns must be resisted by other rigidly connected columns. Fig. 17 shows a steel
576 portal frame with two rigidly connected columns at the middle and two leaning columns at the two
577 edges. In this figure, the two leaning columns can be thought as a representation of all leaning
578 columns attached to the frame. The load ratio (n) is defined as the ratio between the sum of load
579 acting on all columns (leaning and rigidly connected) to the sum of load acting on the rigidly
580 connected columns only and it can be expressed by the following equation:
∑ 𝑃+𝑃𝑙𝑒𝑎𝑛𝑖𝑛𝑔
581 𝑛= ∑𝑃
(56)
582 The minimum value of (n) is unity which represents a frame without leaning columns. Frames with
583 (n) greater than unity always have leaning columns. The P-𝛿 effect resulting from leaning columns
584 must be considered in shear equilibrium Eqs. (31) or (43) by adjusting them as follows:
2𝐸𝐼1𝑐 𝛿
585 𝑉 𝐿𝑐 = [−(𝛼1 + 𝛼3 ) 𝜃𝑎 − (𝛼2 + 𝛼4 ) 𝜃𝑏 + (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 ) ]−𝑛𝑃𝛿 (57)
𝐿𝑐 𝐿𝑐
2𝐸𝐼1𝑐 (𝜇 2 +0.25)(𝑐−1)2 𝛿
586 𝑉 𝐿𝑐 = [−(𝛼1 + 𝛼3 ) 𝜃𝑎 − (𝛼2 + 𝛼4 ) 𝜃𝑏 + (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − 𝑛 ) 𝐿 ] (58)
𝐿𝑐 2 𝑐
587 Therefore, the buckling conditions for sway frames with leaning columns can be written as follows:
3
(𝛼1 + 𝐺 ) 𝛼2 −(𝛼1 + 𝛼2 )
𝐴
| 3 |
588 𝛼3 (𝛼4 + 𝐺 ) −(𝛼3 + 𝛼4 ) =0 (59)
| 𝐵 |
(𝜇 2 +0.25)(𝑐−1)2
−(𝛼1 + 𝛼2 ) −(𝛼3 + 𝛼4 ) (𝛼1 + 𝛼2 + 𝛼3 + 𝛼4 − 𝑛 )
2
589
P P
(n-1) P (n-1) P
590 I2c I2c
593
I1c I1c
Lb Lb Lb
594
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
595 Fig. 17. Steel portal frame with leaning columns.
596 Eq. (59) is solved by iteration technique for different values of n, GB and I1c/I2c. The effect of the
597 load ratio (n) on the buckling length factor of tapered columns in sway frames with hinged and fixed
598 bases is shown on Figs. 18 and 19, respectively. The curves in these figures are plotted for certain
599 values of GB. It is obvious from these figures that the buckling length factor, for any given value of
600 GB, increases with the increase of load ratio (n) resulting from additional P-𝛿 effect on leaning
601 columns. Moreover, for a specific value of GB , the buckling length factor curves for different ratios
602 (n) are parallel. This implies that the buckling length factor (kn>1) at specific value of GB, and at
603 specific value of n > 1 for all I2c / I1c , is just a constant multiplier of the buckling length factor (kn=1)
604 at n=1. To determine this multiplier, the relationship between the ratio (kn>1)/ (kn=1) and the ratio (I2c
605 / I1c ) is plotted in Figs. 20 and 21 for hinged and fixed base columns, respectively. For the case of
606 sway frames with hinged bases the value GB is varied from zero to 10 while for the case of fixed
607 bases GB is varied from zero to infinity. As expected, Figs. 20 and 21 show that the ratio (kn>1)/ (kn=1)
608 is almost constant for similar values of GB, irrespective of the value of I2c / I1c. Moreover, it is clear
609 from these figures that there is an upper bound value for the ratio (kn>1)/ (kn=1) for each value of n.
610 The value of this upper bound can be taken conservatively equals to √𝑛 which is indicated by the
612 Therefore, the buckling length factor (kn>1) in case of sway frames with leaning columns can be
613 determined by multiplying the buckling length factor (kn=1) in case of no leaning columns by the
616 Eq. (60) is similar to the story buckling approach discussed by Yura [35] and included in AISC [1],
617 which means that the story buckling approach is valid to sway frames composed of rigidly connected
619
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
620
621
622
623
624
625
626
627
628
629
630
631
632
633 Fig. 18. Effect of loading on leaning columns on the effective buckling length of hinged base sway frames.
634
635
636
637
638
639
640
641
642
643
644
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
645
646 Fig. 19. Effect of loading on leaning columns on the effective buckling length of fixed base sway frames.
647
648
649
650
651
652
653
654
655
656
657
658
659 Fig. 20. Ratio of k-factor for sway frames with leaning columns to k-factor without leaning columns-hinged base case.
660
661
662
663
664
665
666
667
668
669
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
670
671
672 Fig. 21. Ratio of k-factor for sway frames with leaning columns to k-factor without leaning columns-fixed base case.
673 5. Conclusions
674 The current paper developed closed form equations and design charts for bending stiffness
675 modification factors for two common types of partially tapered beams that can be used in stability
676 analysis of steel frames with prismatic or linearly tapered columns. For steel frames with prismatic
677 columns, these modifications factors can be used along with AISC alignment charts and extend the
678 applicability of this alignment charts to include partially tapered restraining beams. For steel frames
679 with linearly tapered columns, the proposed modifications factors are incorporated in the stability
680 analysis and simple design charts are presented for case of sway or no sway frames with different
681 boundary conditions. Results from the proposed approach agrees well with numerical results from
682 LBA and published results in previous literature. The results of the investigated cases showed that
683 partial tapering of restraining beam significantly reduces effective buckling length of frames
684 permitted to sway compared to that of similar frames with prismatic beams having the smallest cross
685 section while that effect is marginal in case of braced frames. Moreover, the additional P-𝛿 effect,
686 resulting from leaning columns in sway frames, on the effective buckling length was investigated. It
687 was proved that story buckling approach commonly used in prismatic steel frames can be extended
688 to steel frames composed of tapered columns and partially tapered beams.
689 6. References
690 [1] AMERICAN INSTITUTE OF STEEL CONSTRUCTION. Specification for Structural Steel Buildings. 130 East
691 Randolph Street, Suite 2000, Chicago, Illinois 60601, 2016. ANSI/AISC 360-16.
692 [2] Lee, G. C., Morrell, M. L. and Ketter, R. L. Design of tapered members. 1972, Welding Research Bulliten
693 173, pp. 1-32.
694 [3] AMERICAN INSTITUTE of STEEL CONSTRUCTION. Supplement No. 3 to the Specificationfor the Design,
695 Fabrication & Erection ofStrucutral Steel for Building. New York : AISC, 1974.
696 [4] Lee, G.; Morrell, M. Application of AISC Design Provisions for Tapered Members. 1975, Engineering
697 Journal, American Institute of Steel Structures, Vol. 12, issue 1, pp. 1-13.
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
698 [5] Hirt, M.A. and Crisinel, M. Charpentes Métaliques–Conception et Dimensionnement desHalles et
699 Bâtiments.: Traité de Génie Civil, Vol. 11. Press Polytechniques et Universitaires Romandes, Lausanne., 2001.
700 [6] Marques, L.; Taras, A.; Simos da Silva,L.; Grenier, R.; Rebelo,C. Development of a consistent buckling
701 design procedure for tapered columns. 2012, Journal of Constructional Steel Research, Vol. 72, pp. 61-74.
702 [7] Eurocode 3. Design of Steel Strucutres, Part 1-1: General Rules and Rules for Buildings. : BS-EN 1993-1-1,
703 2005, Incorporating Corrigenda February 2006 and April 2009.
704 [8] Tankova, T.; Martins, J.P.; Simões da Silva, L.; Simões, R.; Craveiro, H.D. Experimental buckling
705 behaviour of web tapered I-section steel columns. 2018, Journal of Constructional Steel Research, Vol. 147,
706 pp. 293-312.
707 [9] Nguyen, T.; Tran, N.N; Nguyen, D.D. Prediction of Critical Buckling Load of Web Tapered I-Section Steel
708 Columns Using Artificial Neural Networks. 2021, International Journal of steel structures,
709 https://doi.org/10.1007/s13296-021-00498-7
710 [10] Ermopoulos, J. Ch. Elastic Stability Analysis of Plane Rectangular Frames with Varying Stiffness
711 Members. National Technical University of Athens. Athens: 1984. Ph.D. thesis.
712 [11] Ermopoulos, J. Ch and Kounadis, A. N. Stability of frames with tapered built-up members. 1985, Journal
713 of Strucutral Engineering, ASCE, Vol. 111, issue 9, pp. 1979-1992.
714 [12] Ermopoulos, J. Ch. Buckling of tapered bars under stepped axial loads., 1986, Journal of Structural
715 Engineering, ASCE, Vol. 112, issue 6, pp. 1346-1354.
716 [13] Ermopoulos, J. Ch. Slope-deflection method and bending of tapered bars under stepped loads., 1988,
717 Journal of Contructional Steel Research, Vol. 11, issue 2, pp. 121-141.
718 [14] Ermopoulos, J. Ch. Equivalent Buckling Length of Non-uniform Members. 1997, Journal of
719 Constructional Steel Research, Vol. 42, issue 2, pp. 141-158.
720 [15] Al-Sadder, S.Z. Exact expressions for stability functions of a general non-prismatic beam-column
721 member. 2004, Journal of Constructional Steel Research, Vol. 60, issue 11, pp.1561-1584.
722 [16] Raftoyiannis, I. G.; Ermopoulos, J. Ch. Stability of tapered and stepped steel columns with initial
723 imperfections. 2005, Engineering Structures, Vol. 27, issue 8, pp. 1248-1257.
724 [17] King, W.; Duan, L.; Zhou, R.; Hu, Y.; Chen, W. K-factors of framed columns restrained by tapered girders
725 in US codes. 1993, Engineering Structures, Vol. 15, issue 5, pp. 369-378.
726 [18] Bazeos, N.; Karabalis, D. L. Efficient computation of buckling loads for plane steel frames with tapered
727 members. 2006, Engineering Structures, Vol. 28, issue 5, pp. 771-775.
728 [19] Karabalis, D.L. and Beskos, D.E. Static, dynamic and stability analysis of structures composed of tapered
729 beams. 1983, Computer and Strucutres, Vol. 16, issue 6, pp. 731-748.
730 [20] Saffari, H.; Rahgozar, R.; Jahanshahi, R. An efficient method for computation of effective length factor
731 of columns in a steel gabled frame with tapered members. 2008, Journal of Constructional Steel Research,
732 Vol. 64, issue 4, pp. 400-406.
733 [21] Serna, M. A.; Ibáñez, J. R.; López, A. Elastic flexural buckling of non-uniform members: Closed-form
734 expression and equivalent load approach. 2011, Journal of Constructional Steel Research, Vol. 67, issue 7,
735 pp. 1078-1085.
736 [22] Lu, N.; Meng, L. The element stiffness matrix of a tapered beam with effects of shear deformation and
737 its stability application. 2011, Advanced Materials Research, Vols. 308-310, pp. 1383-1388.
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993
738 [23] Salem, E. Mapping web-tapered member to a prismatic member for buckling analysis of sway frames-
739 closed form equation. 2013, Structural Stability Research Council Annual Stability Conference, St. Louis,
740 Missouri, April 16-20.
741 [24] Chan, S.L. Buckling analysis of structures composed of tapered members., 1990, Journal of Strucutral
742 Engineering, ASCE, Vol. 116, issue 7, pp. 1893-1906.
743 [25] Kim, M.C., Chang, K.C. and Lee, G.C. Elastic and Inelastic Buckling Analysis of Thin-Walled Tapered
744 members. 1997, Journal of Engineering Mechanics, ASCE, Vol. 123, issue 7,pp. 727-735.
745 [26] Valipour, Hamid R. and Bradford, Mark A. A new shape function for tapered three-dimensional beams
746 with flexible connections. 2012, Journal of Constructional Steel Research, Vol. 70, pp. 43-50.
747 [27] Chiorean, C. G.; Marchis, I. V. A second-order flexibility-based model for steel frames of tapered
748 members. 2017, Journal of Constructional Steel Research, Vol. 132, pp. 43-71.
749 [28] Kucukler, M., Gardner, L. Design of laterally restrained web-tapered steel structures through a
750 stiffness reduction method. 2018, Journal of Constructional Steel Research, Vol. 141, pp. 63-76
751 [29] Kucukler, M., Gardner, L. Design of web-tapered steel beams against lateral-torsional buckling through
752 a stiffness reduction method. 2019, Journal of Constructional Steel Research, Vol. 190, pp. 246-291
753 [30] Duan, L.; Lei, H.; Chen, W. Effective Length Factors of Compression Members, 2014, Bridge Engineering
754 Handbook: Fundamentals, Second Edition, pp. 427-450.
755 [31] Ziemian, Ronald D. Guide to Stability Design Criteria for Metal Structures: Sixth Edition: John Wiley and
756 Sons, 2010.
757 [32] Timoshenko, S. P. and Gere, J. M. Theory of Elastic Stability,. second edition. New York : McGraw- Hill,
758 1961.
759 [33] DUMONTEIL, P. Historical Note on K-factor Equations. 1999, Engineering Journal, AISC, Vol. 36, issue 2,
760 pp. 102-103.
761 [34] STAAD.Pro . SELECT Series 6, 2015, Bently Systems Inc., Exton ., PA, USA.
762 [35] Yura, Joseph. The Effective Length of Columns in Unbraced Frames. April 1971, Engineering Journal, AISC
763 Vol. 8, issue 2, pp. 37-42.
764
Preprint copy of: Effective buckling length of frames with tapered columns and partially tapered
beams, Journal of Constructional Steel Research 187(12)- DOI: 10.1016/j.jcsr.2021.106993