Stability Design of Steel Buildings

Finding K Can Be Complicated!

STABILITY DESIGN OF STEEL BUILDINGS:

Highlights of a New AISC Design Guide

Donald W. White,
Georgia Institute of Technology
&
Lawrence G. Griffis,
Walter P. Moore and Assoc. Inc.

Physical Behavior of Frames 2005 AISC Specification

• Always a load-deflection problem • Major update in stability provisions

• Never a bifurcation problem • Modifications to traditional approach –
Effective Length Method (ELM), Section C2.2a
P Py
• Creation of new method –
Direct Analysis Method (DM), Appendix 7
• Creation of a streamlined version of DM –
o = L/500
P First-Order Analysis Method (FOM), Section C2.2b
Mp
M
P
M

NASCC 07 1
Stability Design of Steel Buildings

Major Advantages of the DM Section C1 of the 2005 Specification

• All analysis and design methods must consider :

• No K factors are required! – Flexural, axial and shear deformations
• Internal forces are more accurate (in all components)
• Applies to all frame types – moment frames, braced – All component limit states
frames, column bracing systems & combined systems – P-Δ and P-δ effects
• More economical beam-column proportions in certain – Reduction in stiffness (and corresponding increases in
cases deformations) due to residual stresses and material
• The DM will become the preferred method in the next yielding
Specification – P-Δo and P-δo effects (i.e., member out-of-plumbness
& out-of-straightness)
• How does one address all of these requirements?

Flexural, Axial & Shear Deformations All Component Limit States

• Included where important by modeling the • Member axial resistance (Pc )

corresponding component flexibilities in the analysis, – Flexural buckling about the x-axis
e.g., – Flexural buckling about the y-axis (sym. members)
– Member or panel zone shear deformations – Torsional buckling (sym. members, in some cases)
– Connection deformations – Torsional-flexural buckling (y-axis flexure + torsion,
singly-sym. members)
– Diaphragm deformations
– Constrained-axis torsional buckling
– Column or brace axial extension & shortening • Member flexural resistance (Mc )
– Lateral-torsional buckling
– Flange local buckling
– Tension flange yielding
• Connection resistances …
• Panel zone resistance … etc.

NASCC 07 2
Stability Design of Steel Buildings

P-Δ and P-δ effects P-Δ effects

Captured by a 1st-order elastic analysis with

appropriate amplifiers or by a properly
configured explicit 2nd-order analysis

P-Δ and P-δ effects P-Δ and P-δ effects

B2

Recommendation: Thoroughly benchmark & verify

selected software packages providing
general purpose second-order analysis capabilities …
Then use these capabilities for final design of anything
beyond basic 2D rectangular frames
B2

NASCC 07 3
Stability Design of Steel Buildings

2nd-Order Analysis with ASD Loads 2nd-Order Analysis with ASD Loads

Applied Applied
Load Load
Analyze for 1.6x
W2 1.6WASD
Ratio α

W1 WASD
Divide
by 1.6x

R1 R2 Component force RASD R1.6ASD Component force

Ratio >α
• When using an explicit 2nd-order analysis: factor loads x 1.6; run
Due to 2nd-order effects, calculated internal forces are analysis; divide results by 1.6
NOT proportional to applied loads. • When using amplifiers (e.g., B1 & B2), simply include 1.6 factor in the
load term of the amplifier, i.e., αΣPr, where α = 1.6 in ASD

Reduction in Stiffness P-Δo and P-δo effects

due to residual stresses & material yielding • The ELM handles these effects on the overall resistance
implicitly, via the use of:
• The ELM handles these effects on the overall resistance – The AISC column strength curve +
implicitly, via the use of: – Calculated column effective lengths (KL )
– The AISC column strength curve + or overall flexural buckling loads (Fe )
– Calculated column effective lengths (KL ) with the exception of a new minimum lateral load
or overall flexural buckling loads (Fe )
requirement for gravity-only load combinations
• The DM handles these effects on the overall resistance
explicitly, by: • The DM handles these effects on the overall resistance
– Factoring all stiffness contributions nominally by 0.8 explicitly, by:
– Reducing flexural stiffnesses by an additional – Including a nominal out-of-plumbness (base value
τb = 4 (1 – p ) p for p > 0.5 (p = α Pr /Py ) Δo = L / 500)
in the structural analysis in the structural analysis

NASCC 07 4
Stability Design of Steel Buildings

P-Δo and P-δo effects Base Analysis & Design Methods in AISC (2005)

• Δo = L /500 = maximum tolerance from the AISC Attribute Effective Length Direct Analysis Method
Method (ELM) (DM)
Code of Standard Practice
• Recommendations for handling the P-Δo effect: Analysis Type Second-order elastic(1) Second-order elastic(1)
– Explicitly cant the frame geometry if facilitated by the
analysis software Notional Load None(2) 0.002Yi (or nominal out-of-
• Easier to understand and automate for general cases plumbness of Δo/L = 0.002)

– Apply notional loads Ni = 0.002Yi if explicit canting Effective Stiffness Nominal 0.8 * Nominal, except
of the frame geometry is not facilitated by the EIeff = 0.8τbEI if αPr > 0.5Py
analysis software
Axial Strength Pn Pn based on KL(3) Pn based on L (no K)
(Pn = QPy in some cases)

(1) Includes first-order elastic analysis with amplifiers

(2) Minimum lateral load of 0.002Yi required for gravity-only combinations
(3) K = 1 allowed when sidesway amplification < 1.1

Required Forces & Beam-Column Interaction Checks Modified AISC DP-13 Example

Effective Length Method Direct Analysis Method

Exterior columns: W12x65 Fy = 50 ksi

Interior columns: W8x31 E = 29,000 ksi
Beams: W24x62
Beam span lengths: 40 ft
Story height: 20 ft

Unfactored Loads: D = 1.0 kip/ft LRFD Load Combinations:

S = 1.2 kip/ft LC1: 1.2D + 1.6S
W = 10 kips LC2: 1.2D + 0.5S + 1.6W

Pr/Pc + 8/9 (Mr/Mc) = 1

Pr/2Pc + Mr/Mc = 1

NASCC 07 5
Stability Design of Steel Buildings

Summary of Results, Modified DP-13 Example Summary of Results, Modified DP-13 Example

Req’d Forces & Str. Chks, Critical Leeward Beam-Column, LC2

DM
ELM
DM
ELM

Fraction of design load corresponding

to U.C. = 1.0, leeward beam-column
17 % more
LC1 LC2 strength
Direct Analysis 1.18 0.95
Effective Length 1.05 0.81
118 % greater
moment

DM vs. ELM DM vs. ELM

• Both methods are legitimate • The DM simplifies the resistance calculations by:
• The DM gives a more accurate answer in general – Directly addressing overall stability effects in the
• The DM is more sensitive to the 2nd-order analysis analysis calculation of Pr & Mr
accuracy – Eliminating the need for K factors
– Both P-Δ and P-δ effects must be captured accurately • The DM applies in the same logical and consistent way
• The ELM requires some restrictions to control its lack of for all structure types
accuracy for certain structure types:
– Minimum lateral load of 0.002Yi for gravity-only load
combinations
– Δ2nd / Δ1st < 1.5

NASCC 07 6
Stability Design of Steel Buildings

First-Order Analysis Method (FOM) Stiffness Reduction for the DM

• Section C2b of AISC (2005) gives the requirements for

the First-Order Analysis Method (FOM) • Use 0.8E to reduce all elastic contributions to the
vvstiffness (do not reduce E in component resistance eqs.)
– design based on a first-order analysis, no 2nd-order
analysis required • Reduce EI of members with αPr / Py > 0.5 by an
vvadditional τb = 4 (αPr /Py ) (1 – αPr /Py )
• The FOM is really just the DM IN DISGUISE, COMBINED
WITH AN APPROXIMATE 2nd-ORDER ANALYSIS
– Applicable only to rectangular frame geometries
(unless you really know what you are doing!) Note:
– Based on an assumed B2 = Δ2nd / Δ1st = 1.5 (1.71 For frames in which the second-order effects are
using the DM reduced stiffness) small, the 0.8 stiffness reduction has a negligible
– Based on the use of the maximum Δ = Δ1st from all effect on the calculated internal forces
the stories as a uniform Δ over the full height of the
structure, to account for the overall P-Δ effects

Nominal Out-of-Plumbness for the DM Notional Load = Out-of-Plumbness Effect

• Use the base out-of-plumb tolerance from the AISC = ΣPr(i+1) x 0.002L / L
Code of Standard Practice Δo/L = 0.002
• Larger values of Δo/L should be used where a more
liberal tolerance is permitted
• Smaller values of Δo/L may be used where appropriate
controls are placed on the constructed geometry

NASCC 07 7
Stability Design of Steel Buildings

Out-of-Plumbness Æ Notional Load: Out-of-Plumbness Æ Notional Load:

Nonrectangular Frames Nonrectangular Frames

Example 12 of
Design Guide

Relaxation of Out-of-Plumbness Requirements in the DM Why Pn(L) in the DM?

• If Δ2nd / Δ1st < 1.5 (Δ2nd / Δ1st < 1.71 based on reduced • To capture any potential “non-sway” member failure under
stiffness): large axial compression
– Δo/L = 0.002 can be neglected • Alternative approach (extension to Specification):
in lateral load combinations – The axial resistance in the plane of bending Pni may be
taken equal to QPy when
• αPr < 0.1Pe(Li) or
• An appropriate δo = 0.001Li is included in the DM analysis
– Torsional, torsional-flexural and/or lateral-torsional
buckling limit states still must be addressed based on
the actual unbraced lengths in calculating Pn & Mn
• Use Pn(L) = min (Pn(Lx), Pn(Ly)) for biaxially-loaded beam-cols.

NASCC 07 8
Stability Design of Steel Buildings

BASIC CONCEPT of the DM More information may be found in …

Let the 2nd-order analysis do the work for you!

and in the NASCC 07 session on

“Stability Design of Steel Buildings”
(TH 8:00-9:30 & F 3:30-5:00)

NASCC 07 9