Basic Introduction to The function of a structural engineer is

Nonlinear Analysis to design — not to analyze

Norris and Wilbur
Ronald D. Ziemian 1960
Bucknell University

Analysis is a means to an end

rather than the end itself.

Role of the analysis: Impetus:

• forces, moments and deflections ⇒ design equations AISC Ch. C: P-∆, P-δ (App. 7)
Limit States Design App. 1: Inelastic Design
• insight into the behavior of a structure Seismic: Pushover Analysis
⇒ better the understanding, better the design Other: Progressive Collapse

Limit States Design: Nonlinear Analysis

Prior to limit of resistance, significant nonlinear
response, including
- geometrical effects (P-∆, P-δ)
- material effects (yielding, cracking, crushing) Available Software Education
- combined effects
Research

Nonlinear Analysis
™ Hand methods Lecture Overview
• Second-order effects
™ Brief Introduction (done!)
– i.e. Moment Amplification Factors (B1 and B2 factors)
• Material nonlinear effects ™ Computer Structural Analysis (Review?)
– i.e. plastic analysis (upper and lower bound theories) ™ Basis for Material Nonlinear Models
™ Computer Methods (focus of today’s lecture) ™ Incorporating Geometric Nonlinear Behavior
• Lots of variations ™ Critical Load Analysis
– all use same basic concepts (most important to today)
™ Overview of MASTAN2 software
– one approach will be presented (basis for MASTAN2)
™ Summary and Concluding Remarks
™ Please keep in mind
• All methods are approximate
• Not a substitute, but a complement to good engineering

1
 How does the computer get Equilibrium Equations
these results? D 40 kips Free Body
AD
f xD
Diagram BD
f xD D 40 kips
™ State-of-the-Art Crystal Ball? Not quite. CD
f xD
™ By applying 2 requirements and 1 translator
AD CD
• Two Requirements: AD
f yD f yD BD f yD
f yD
– Equilibrium (equations in terms of F’s and M’s, 1 per d.o.f.) A B C AD
f xD
– Compatibility (equations in terms of ∆’s and θ’s, 1 per d.o.f.)
CD
D BD f yD
• Translator “apples to oranges” r A f yD
x-d.o.f. u D : ∑ Fx = 0 be BD
em CD

BD
– Constitutive Relationship (i.e. Hooke’s Law, σ = E Є ) f xD f xD

Member CD
M
40 = AD
+ BD
+ CD

er
– Generalized to Force-to-Displacement (i.e. F=k∆) f xD f xD f xD

mb
– Re-write equilibrium eqs. in terms of unknown displacements

Me
y-d.o.f. vD : ∑ Fy = 0
™# of Equil. Eqs. = # of Unknown Displacements 0 = f yD
AD
+ f yD
BD
+ f yD
CD

Translator: Forces Æ Displacements

F Æ ∆ for all members vD
fyj j’ D
uD
fxj vj
j f xi = k11ui + k12vi + k13u j + k14v j Member AD:
(Undeformed) uj f xAAD = k11ADu AAD + k12AD v AAD + k13ADu DAD + k14AD vDAD
(Deformed) f yi = k21ui + k22vi + k 23u j + k24v j AD
f yA = k21 u A + k22
AD AD
v A + k23
AD AD
u D + k24
AD AD AD AD
vD
fyi f xj = k31ui + k32vi + k33u j + k34v j
AD
f xD = k31 u A + k32
AD AD
v A + k33
AD AD
u D + k34
AD AD AD AD
vD vA vB vC
fxi vi i’ AD
f yD = k41 u A + k42
AD AD
v A + k43
AD AD
u D + k44
AD AD AD AD
vD A uA B uB C uC
i ui f yj = k 41ui + k42vi + k43u j + k44v j
Big Question: Member BD: Member CD:
Where do these known stiffness coefficients k’s BD
f xB = k11BDu BBD + k12BD vBBD + k13BDu DBD + k14BD vDBD CD
f xC = k11 uC + k12
CD CD
vC + k13
CD CD
u D + k14
CD CD CD CD
vD
come from? Little Answer: BD
f yB = k21 u B + k22
BD BD
vB + k23
BD BD
u D + k24
BD BD BD BD
vD CD
f yC = k21 uC + k22
CD CD
vC + k23
CD CD
u D + k24
CD CD CD CD
vD
Function of member’s material and geometric BD
f xD = k31 u B + k32
BD BD
vB + k33
BD BD
u D + k34
BD BD BD BD
vD CD
f xD = k31 uC + k32
CD CD
vC + k33
CD CD
u D + k34
CD CD CD CD
vD

properties, including its orientation. BD

f yD = k41 u B + k42
BD BD
vB + k43
BD BD
u D + k44
BD BD BD BD
vD CD
f yD = k41 uC + k42
CD CD
vC + k43
CD CD
u D + k44
CD CD CD CD
vD

Substituting into Equil. Eqs. Substituting into Equil. Eqs. (cont.)

D D
40 kN Member AD: 40 kN Member AD:
AD
f xD = k31 u A + k32
AD AD
v A + k33
AD AD
u D + k34
AD AD AD AD
vD AD
f yD = k41 u A + k42
AD AD
v A + k43
AD AD
u D + k44
AD AD AD AD
vD

Member BD: Member BD:

A B C A B C
BD
f xD = k31 u B + k32
BD BD
vB + k33
BD BD
u D + k34
BD BD BD BD
vD BD
f yD = k 41 u B + k 42
BD BD
vB + k 43
BD BD
u D + k44
BD BD BD BD
vD
x-d.o.f. u D : ∑ Fx = 0 Member CD:
y-d.o.f. v D : ∑ Fy = 0 Member CD:
40 = f xDAD + f xD
BD
+ f xD
CD CD
f xD = k31 uC + k32
CD CD
vC + k33
CD CD
u D + k34
CD CD CD CD
vD 0 = f yDAD + f yD
BD
+ f yD
CD CD
f yD = k 41 uC + k 42
CD CD
vC + k 43
CD CD
u D + k44
CD CD CD CD
vD

d.o.f. u D : d.o.f. vD :
40 = (k31ADu AAD + k32ADv AAD + k33ADu DAD + k34ADvDAD ) + (
0 = k 41AD u AAD + k 42AD v AAD + k 43AD u DAD + k 44AD vDAD + )
(k BD BD
31 B u +k v BD BD
32 B
+k uBD BD
33 D
+k v BD BD
34 D
)+ (k BD BD
u
41 B +k vBD BD
42 B +k u BD BD
43 D +k v BD BD
44 D )+
(k CD CD
31 C u + k32 vC + k33
CD CD
u D + k34
CD CD
vD )
CD CD
(k CD CD
u
41 C +k v CD CD
42 C +k u CD CD
43 D + k 44
CD CD
vD )

2
 So, where are we at? Compatibility Eqs. (consistent deflections)
™ We have two equilibrium equations (1 per
D Member-to-Member
vD
d.o.f.) in terms of a lot of displacements: vDAD D
uD
uDAD
(
u D : 40 = k31AD u AAD + k32AD v AAD + k33AD u DAD + k34AD vDAD + ) v BD
vCD

D
(k )+ A B C D D

rA
BD BD
u + k32BD vBBD + k33BD u DBD + k34BD vDBD

be
uDBD uDCD
31 B

(k )

em

BD
CD CD
u + k32CD vCCD + k33CD u DCD + k34CD vDCD

Member CD
M
Member-to-Support
31 C

(k )+

er
vD : 0 = AD
u AAD + k 42AD v AAD + k 43AD u DAD + k 44AD vDAD

mb
41

(k )+ u A = u AAD = 0 u C = u CCD = 0

Me
BD BD
41 B u +k v
BD BD
42 B +k u BD BD
43 D +k v BD BD
44 D

(k CD CD
41 C u +k v
CD CD
42 C +k u CD CD
43 D +k v CD CD
44 D ) =
vA vA AD
= 0 vC = vCCD = 0
uB = uB = 0
BD u D = u DAD = u DBD = u DCD
What card haven’t we played yet? vB = vBBD = 0 vD = vDAD = vDBD = vDCD

Time for some serious simplifying

D
40 kN

After simplifying…
™ Applying Compatibility to Equil. Eqs.:
All = uD A B C

u D : 40 = k u ( AD
31
AD
A +k v AD AD
32 A +k u AD AD
33 D +k v AD AD
34 D )+ u D : 40 = k ( AD
+k BD
+k CD
) u? + (k AD + k BD + k CD ) v?
(k )+
33 33 33 D 34 34 34 D

0 = (k ) ( )
BD BD
u + k32BD vBBD + k33BD u DBD + k34BD vDBD AD
+k BD
+ k 43
CD
u?D + k 44AD + k 44BD + k 44
CD
v?D
31 B
vD :
(k CD CD
u
31 C +k vCD CD
32 C +k u CD CD
33 D +k v CD CD
34 D ) 43 43

vD : 0 = (kAD
u AAD + k 42AD v AAD + k 43AD u DAD + k 44AD vDAD )+
Since k’s are known, we have
41

(kBD BD
u + k 42BD vBBD + k 43BD u DBD + k 44BD vDBD )+
41 B
2 Equations and 2 Unknowns
(kCD CD
41 C u + k 42 vC + k 43
CD CD
u D + k 44
CD CD CD CD
vD )
All = 0 All = vD Solve for Unknown Displacements
Which simplifies to… uD = # and vD = ##

With all displacements, solve for Summary of Computer Approach

member forces… Member
f
AD:
=k u +k v +k u +k AD AD AD AD AD AD AD AD AD ™ For each d.o.f., write an equilibrium equation:
xA 11 A 12 A 13 D 14 vD

u A = u AAD = 0 AD
f yA = k 21 u A + k 22
AD AD
v A + k23
AD AD
u D + k24
AD AD AD AD
vD Fexternal = Σ fmember
AD
= k31 u A + k32
AD AD
v A + k33ADu DAD + k34
AD AD AD AD
Re-write (translate) each member force in terms of
f xD vD
v A = v AAD = 0 AD
= AD AD
+ AD AD
+ AD AD
+ AD AD
™
its end displacements (Stiffness Eqs.)
f yD k 41 uA k 42 vA k43 uD k44 vD

u B = u BBD = 0 Member BD:

BD
f xB = k11BDu BBD + k12BD vBBD + k13BDu DBD + k14BD vDBD fmember = Σ k member ∆ member end
vB = vBBD = 0 = k21 u B + k22 vB + k23 u D + k24
Substitute Stiffness Eqs. into above Equil. Eqs.
BD BD BD BD BD BD BD BD BD
f yB vD
™
uC = uCCD = 0 = k31 u B + k32 vB + k33 u D + k34
BD BD BD BD BD BD BD BD BD
f xD vD
™ Simplify Equil. Eqs. by applying member-to-member
BD
f yD = k 41 u B + k42
BD BD
vB + k 43
BD BD
u D + k44
BD BD BD BD
vD
vC = vCCD = 0 Member CD:
and member-to-support compatibility conditions
= k11 uC + k12 vC + k13 u D + k14 ™ Solve n Equil. Eqs. for the n unknown displacements
CD CD CD CD CD CD CD CD CD

u D = u DAD = u DBD = u DCD = #

f xC vD

™ Use Stiffness Eqs. to calculate member forces

CD
f yC = k21 uC + k22
CD CD
vC + k23
CD CD
u D + k24
CD CD CD CD
vD
vD = vDAD = vDBD = vDCD =# # CD
f xD = k31 uC + k32
CD CD
vC + k33
CD CD
u D + k34
CD CD CD CD
vD
™ Apply Equil. Eqs. to solve for reactions
CD
f yD = k41 uC + k42
CD CD
vC + k43
CD CD
u D + k44
CD CD CD CD
vD

3
 Lot’s of Questions Two Big Questions
™ So,this is how most commercial programs such ™ Where do those stiffness coefficients
as SAP2000, RISA, STAAD, etc. get the come from?
answer? • You mean the ones that relate member end
forces to member end displacements?
• Yes! Known as “Direct Stiffness Method”
• Yeah, those k’s ! <More to come on this>
™ So, all such programs will give the same answer?
™ What happens when we go static nonlinear
• Yes, as long as it is a static 1st-order elastic analysis.
or even dynamic?
™ Waita minute…Is this the basic analysis
• Same basic procedure, but apply loads in
procedure for the “finite element method”? increments and perform a series of analyses.
• Yes! Bit more tricky to get k’s, σ‘s, and Є’s Then, sum incremental results.
< Much more to come on this! >

Important Points Stiffness Coefficients, k’s

™ The only opportunity for most computer
analysis software to model the actual ™ Let’s start with high school physics
behavior of the structure is through the • Extension Spring Experiment
member stiffness terms.
Force
™ So, to include Before:
F
• first-order effects
• second-order effects ∆
After:
• material nonlinear behavior k
F
Must modify member stiffness!!! 1
™ Let’s review member stiffness F=k∆
∆ Displacement

Stiffness Coefficients, k’s (cont.) How about real structural members?

™ Axial force member
™ More “advanced” high school physics lab Before:
• Rotational Spring Experiment F

Before: ∆
Moment k After:
M 1
F
M ∆
After: θ ™ Stiffness k function of:
k • Geometry: Area and Length (A↑,k↑ & L↑, k↓)
1 • Material: Elastic Modulus (E↑,k↑)

θ Rotation
M=kθ F = k(A,L,E) ∆

4
 How about real members? (cont.) Other factor impacting stiffness
™ Flexural members ™ Orientation of member
Before: Before:
• consider axial force member:

M kv=EA/L kv=0
After: θ After: kh=0
∆

Vertical Horizontal
™ Stiffness k function of: F
Member Member
• Geometry: Moment of Inertia & Length (I↑,k↑& L↑, k↓) kh=EA/L
• Material: Elastic Modulus (E↑,k↑)

M = k(I,L,E) θ F = k(I,L,E) ∆

Orientation of axial force member Summary: Three Perspectives

kh=(cos2φ)EA/L kv=(sin2φ)EA/L

F1

φ φ

Important Point: Less vertical a member,

the less stiffness to resist vertical loads. ™ Reality: What you see…

Three Perspectives
Three Perspectives (cont.)
™ What your computer actually sees:
™ What you see on your computer screen:
DL
DL WL
WL

Collection of elements
connected by sharing Σk
common nodes

Assemblage of
equivalent springs {F} = [K]{∆}

5
 Analysis Review: Key Points
Material Nonlinear (Inelastic)
™ Reviewed the “Direct Stiffness Method”
• Equilibrium Æ Translator F(∆) Æ Compatibility ™ Best place to start is with a tensile test
™ Response of structure controlled by
stiffness of members (a.k.a. springs) σ = P/A
™ First-order elastic stiffness of member
σyield
function of: Perfectly Plastic (E = 0)

ict
• Material Property (E)

Elas
• Geometric Properties (A, I, L, and orientation) E

™ Time to go nonlinear… 1
let’s begin with material nonlinear
Є = ∆/L

Normal Stress: Structural Members Inelastic Behavior: Axial Force

™ For typical structural steel members P Originally:
(L/d>10), elastic/inelastic behavior E=0 Æ k=0 σ =P/A=0
Py=Aσy
controlled by normal stresses σ’s acting
∆
along the length axis of the member. k=EA/L Elastic:
™ Normal stress produced by: 1
P
∆ σ < σy
• Axial force (P/A) ∆
Yield:
• Major and/or minor axis flexure (Mc/I)
• Combination of above effects (i.e. P/A + Mc/I) Plastic Hinge σ = σy
Py

• Warping (not today!) at P = Py or Post-Yield:

∆

™ We will assume elastic-perfectly-plastic when P/Py = 1.0

σ = σy
material (often done for steel)

Inelastic Behavior: Inelastic Behavior: A

M
θ
Flexure M
Flexure (cont.)
A θ A
M M
A
Mp=Zσy Mp=Zσy
Section A-A -σy +σy Section A-A -σy +σy
My=Sσy My=Sσy

k=4E I/L k=4E I/L

E Ielastic M < My E Ielastic M=Sσy=My
1 1
θ θ

6
Inelastic Behavior: A
M
θ
Inelastic Behavior: A
M
θ
Flexure (cont.) Flexure (cont.)
A A
M M

Mp=Zσy Mp=Zσy
Section A-A -σy +σy Section A-A -σy +σy
My=Sσy My=Sσy
E=0 E=0
k=4E I/L k=4E I/L
EI<EIelastic My<M<Mp EI<<EIelastic My<<M<Mp
1 1
θ θ

Inelastic Behavior: A
M
θ Inelastic Behavior: Flexure
Flexure (cont.) M < Mp
A M Elastic: θ
M Plastic
Mp=Zσy
E I=0 Æ k=0
EI=0 Æ k=0
c
ti

Mp=Zσy M = Mp
as

k=4E I/L Plastic: θ

El

Section A-A -σy +σy 1

My=Sσy
θ
E=0
™ Plastic Hinge Model – Assume section as fully
k=4E I/L
EI = 0 M=Zσy=Mp elastic or fully plastic (neglect partial yielding)
1 ™ Plastic Hinge at M = Mp or when M/Mp = 1.0
θ

Types of inelastic models Simple Example:

™ We will employ a plastic hinge model
• A.K.A. “Concentrated Plasticity” 100 kips
• Section is fully elastic or fully yielded W12×65
• Plastic hinges only at element ends
™ Distributed plasticity (still line elements)
• A.K.A. “Plastic Zone”
12’ - 0” 12’ - 0” 12’ - 0”
• Captures gradual yielding through depth and along
length
• More accurate, but computationally more $$$ E = 29,000 ksi
™ Finite element with continuum elements ($$$) σy = 50 ksi

7
 Inelastic Behavior: A
M
θ

Combination P & M P
A
M
-σy
Mp=Zσy
+σy
Plastic
M < Mp
-σy
c
ti σ = Mc/I + P/A
as
El k=4E I/L

1
+σy
θ

Inelastic Behavior: M Plastic Hinge Criterion: A

M
θ
A θ
Combination P & M P/Py -σy P=Py P
for Plastic Hinge
P M=0 A
A 1.0
Yi -σy P<Py
eld
Su +σyM<Mp
r fa
-σy -σy ce
= + -σy
-σy P<Py
+σy +σy +σy
Elastic +σy M<Mp

Fully yielded M < Mp P < Py -σy P=0

+σy M=Mp
section when: M/Mp < 1 P/Py < 1
1.0 M/Mp

Material Nonlinear Analysis Simple Example (with axial force):

™ Employ “Direct Stiffness Method” applying loads
in increments: [K]{d∆} = {dF}
100 kips
™ During the load increment, check to see if
plastic hinge(s) form. If so, scale back load W12×65 P = 400 kips
increment accordingly.
™ Reduce stiffness of yielded members and
continue load increments 12’ - 0” 12’ - 0” 12’ - 0”
• k = kelastic + kplastic with kplastic = plastic reduction
™ Continue
to accumulate results of load
increments until all of load is applied or a plastic E = 29,000 ksi
mechanism forms. σy = 50 ksi

8
 No
Axial Force

Second-Order Effects Equilibrium Equations

™ Formulated on ™ Formulated on
™ A.K.A. “Geometric Nonlinear Behavior” Undeformed Shape Deformed Shape
™ Equilibrium Equations ∆
P ∆
P
• Reality: Should be formulated on deformed H H
shape
Truss is susceptible to
• Difficulty: Deformed shape (deformations) is
2nd-Order effects,
a function of the member forces, which are in
turn a function of the deformations luckily ∆ is often
(Chicken ‘n Egg) quite small.
• Remedy: Perform a series of analyses with
loads applied in small increments and update
geometry after each load increment.
Different reactions and member forces.

Equilibrium Equations Focus on Lateral Stiffness

™ Formulated on ™ Formulated on ™ Formulated on Undeformed Shape: Linear Response
Undeformed Shape Deformed Shape Before: After: Lateral Stiffness
H H k=3E I/L3 P ∆ is slope of H-∆
P ∆ kspring
∆ P H response curve
H k=3E I/L3
H 1

keff < k H
1
1
∆
∆
Effective klateral = kspring
lateral 1
stiffness
is reduced!
∆
M=HL M=HL+P∆

9
 Focus on Lateral Stiffness (cont.) Focus on Lateral Stiffness (cont.)
™ Formulated on Deformed Shape: Nonlinear Response ™ Equilibrium Formulated on Deformed Shape

Before: After: Let’s start by assuming L’ ≈ L,

∆ P Effective lateral
kspring stiffness is reduced ∆ P ΣMo = 0 RL = HL + P∆
H H kspring R = H + P∆/L
H kspring
R=kspring∆ kspring∆ = H + P∆/L
1
H = kspring∆ – P∆/L
klateral < kspring L’ L H = (kspring – P/L) ∆
1
™ Lateral Stiffness (slope of response curve)
∆
Pt. o
H = klateral∆ with klateral = kspring – P/L

Some thoughts here… Geometric Stiffness

™ This simple analysis becomes less “accurate” as
∆/L becomes large (i.e. ∆/L >> 1/5) ™ Effective lateral stiffness of a member:
• Remedy: Perform an incremental analysis and update • decreases as a member is compressed
geometry after each load increment…hence, limit – kg is negative for compressive P
∆/L in each step to some small amount – backpacker example
• Keep in mind serviceability limits are often • increases when subjected to tension
something like ∆/L < 1/400 – kg is positive for tensile P
– guitar string example
™ Most importantly, klateral = kspring – P/L takes on
the form: ™ Employing geometric stiffness approach
Geometric Stiffness • Other methods exist (i.e. stability functions)
k2nd-Order El. = k1st-Order El. + kg

How about real members? (recall…) Closer look at stiffness terms…

™ Flexural members subjected to axial force
™ Flexural members subjected to axial force
M
θ M
θ
P ∆
P P ∆
™ Stiffness k function of: F P
F
• Geometry: Moment of Inertia & Length (I↑,k↑& L↑, k↓) M = k(I,L,E,P) θ with
• Material: Elastic Modulus (E ↑,k↑) k = 4EI/L – 2PL/15 F = k(I,L,E,P)∆ with
• Axial Force: Compressive (P↑, k↓) k = 12EI/L3 – 6P/5L
Again, basic form:
M = k(I,L,E,P) θ F = k(I,L,E,P) ∆ k2nd-Order El. = k1st-Order El. + kg

10
 Geometric Nonlinear Analysis
™ Employ “Direct Stiffness Method” applying loads
in increments: Solve Equil. Eqs. {dF} = [K]{d∆}
™ At start of increment, modify member stiffness
to account for presence of member forces (such
as axial force):
• k = kelastic + kg with kg = geometric stiffness
™ At end of increment, update model of structural
geometry to include displacements
™ Continue to accumulate results of load increments
(∆i = ∆i-1 + d∆ and fi = fi-1 + df) until all of load is
applied or elastic instability is detected.

Comparison: 1st- and 2nd-Order Analysis Results 2nd-Order Inelastic Analysis

™ Employ “Direct Stiffness Method” applying loads
in increments: Solve Equil. Eqs. {dF} = [K]{d∆}
™ At start of increment, modify member stiffness
to account for presence of member forces and
any yielding:
Moments increase by ~10%
k = kelastic + kgeometric + kplastic
™ At end of increment, update model of structural
geometry to include displacements
™ Continue to accumulate results of load increments
(∆i = ∆i-1 + d∆ and fi = fi-1 + df) until all of load is
applied or inelastic instability is detected.

Critical Load Analysis (Basics)

™ Definition: Critical or buckling load is the
load at which equilibrium may be satisfied
by more than one deformed shape.

P Solution Solution P
#1 P #2

Big Q: How does computer software calculate this?

11
 Critical Load Analysis (Background) Example
™ Elastic stiffness of a member k = kel + kg
• kel is f(A or I, L, and E)
Demonstrate computational
• kg is f(P,L), also note directly proportional to P
method for calculating the
™ Elastic stiffness of structure [K] = Σk
elastic critical load (buckling P
• [K] = [Kel] + [Kg]
load) for the structural Rigid Beam
• [Kg] directly proportional to applied force
– i.e. Double applied forces, hence, double internal force
system shown.
distribution and double [Kg]
A, I
™ To the computer, “buckling” will occur when L, E
our equilibrium equations {F} = [K]{∆} permit
non-unique solutions, e.g. det[K] = 0.

Example: Key Stiffness Terms Example: Solution

P
P ∆lat P Rigid Beam
1. Apply reference load, and use 1st-
H Rigid Beam H ∆vert order elastic analysis to obtain
internal force distribution.
klateral A, I 2. Determine load factor λ at which
A, I L, E system stiffness degrades to
L, E Pcr
kvertical permit buckling.

klateral = 12EI/L3 – 6λP/5L

Vertical Stiffness: Lateral Stiffness:
klateral = 0 when λP = 10EI/L2
P = kvertical ∆vert H = klateral ∆lat
klateral = 12EI/L3 – 6P/5L Pcr = λP = 10EI/L2 (Ptheory=9.87EI/L2)

100 kips
Thoughts on Critical Load Analysis
™ Computer analysis for a large system:
W12×65 P = 400 kips
• First, apply reference and perform analysis
– Solve equilibrium eqs. {Fref} = [K]{∆}
– With displacements solve for member forces
• Second, assemble [Kel] and [Kg] based on {Fref}
λ=9.4
• Finally, determine load factor λ causing instability;
computationally this means find load factor λ at which
[K]=[Kel]+λ[Kg] becomes singular
– Determine λ at which det([Kel]+λ[Kg) = 0
– “Eigenvalue” problem: Eigenvalues = Critical Load Factors, λ‘s
Eigenvectors = Buckling modes
™ Accuracy increases with more elements per
λ=1.7 compression members (2 often adequate)

12
 MASTAN2:
Basic Introduction Complete
- Educational software
™ Where do I go from here? (Learning to drive) - GUI ⇔ commercial programs
• Review the slides (Read the driver’s manual) - Limited # of pre- and post-processing options
• Acquire nonlinear software (Borrow a friend’s car) to reduce learning curve
• Work lots of examples (Go for a drive, scary at first…) - Suite of linear and nonlinear 2D and 3D
• Apply nonlinear analysis in design (NASCAR? not quite) analysis routines
- Available with textbook
Acquire nonlinear analysis software or online at no cost
• Commercial programs
www.mastan2.com or
• Educational software (i.e. MASTAN2)
www.aisc.org [Steel Tools]

Lateral
Levels of Analysis: MASTAN2 MASTAN2
1st-Order
load,
Elastic
H
Elastic Critical Load
1st-Order Elastic: [Ke]{∆}={F}
Hec
Elastic Stability Limit
Hes
2nd-Order Elastic: [Ke + Kg]{d∆}={dF} 2nd-Order
Elastic
1st-Order Inelastic: [Ke + Kp]{d∆}={dF} Hic Inelastic Critical Load
1st-Order
2nd-Order Inelastic: [Ke + Kg + Kp]{d∆}={dF} Hp Inelastic Inelastic Limit Load
∆
Critical Load: [Ke + λKg]{d∆}={0} Inelastic Stability H = αP P
His
Limit
Actual
2nd-Order
Yield Surface:
P
H = αP
Inelastic

Function of P, Mmajor, and Mminor

Lateral displacement, ∆

Planar Frame: 1st-Order 2nd-Order

Summary and Conclusions
Load Elastic Elastic
Factor
1st-Order
Inelastic
2nd-Order ™ Provided an introduction to nonlinear analysis
Inelastic
• Review of direct stiffness method
E = 29,000 ksi
σy = 36 ksi
• Material nonlinear analysis (Inelastic hinge)
• Geometric nonlinear analysis (2nd-Order)
ud (in.) • 2nd-Order inelastic analysis (combine above)
14
1831
1665
1720
• Critical load analysis (“eigenvalue analysis”)
First
Second hinge hinge ™ Nonlinear…think modifying member stiffness!
8783 (1st-Order) ™ Overview and availability of MASTAN2
™ Now, its your turn to take it for a spin…
8783 (2nd-Order)

Mechanism Moments at Limit (in. kips)

13
 Gravity Load

Appendix Multi-story 25 kips 50 kips 25 kips

0.5P
Frame: P
™ Several examples to try out P

Lateral Load
™ Solutions by MASTAN2 P
P
™ Need a reference text with many P
examples? see Matrix Structural Analysis, P
2nd Ed., by McGuire, Gallagher, and Ziemian P
(Wiley, 2000) P

™ See tutorial that comes with MASTAN2

™ OK, time to jump in and start driving… Demonstrate:
<See Final Exam…> 2nd-Order Inelastic Analysis
E = 29,000 ksi
Non-proportional loading σy = 36 ksi

2nd-Order
MASTAN2
Inelastic
Model:
Analysis:
Gravity Load P = 30 kips

2nd-Order 2nd-Order
Inelastic Inelastic
Analysis: Analysis:
Limit State:
P = 37.7 kips
Plim = 40.3 kips

14
2nd-Order 2nd-Order Strength Limit State (P=40.3 kips)

Inelastic Inelastic
Analysis:
Analysis: Post-limit State: 1st Plastic Hinge (P=30 kips)

P = 40 kips ∆

Truss (Hoff et al.): MASTAN2

3P
Model:
All members:
20” A = 9.348×10-2 in2
I = 6.954 ×10-4 in4
E = 29,000 ksi
3P
3 @ 20”
P = 100 lbs

Demonstrate:
Elastic Critical Load
2nd-Order Elastic
Experimental (Plimit=220 lbs)

Elastic Critical Load: Pcr = 210.7 lbs 2nd-Order Elastic: Plim = 210 lbs

15
 P
Response Elastic Critical Load
M = αPL
Curves: P = 210 lbs Beam-Column:
2nd-Order Elastic
1st-Order
Elastic
W24×76
E = 29,000 ksi
L = 24’
Demonstrate:
Elastic Critical Load Analysis
1. Flexural Buckling (α=0.0)
2. Torsional Flexural Buckling
(α=0.04)

M = αPL

Elastic Critical Load (α = 0.0)

MASTAN2
Model:

Isometric View Elevation View Plan View

Elastic Critical Load Analysis (α = 0.04) Suspension System:

A = 5.40 in2
σy = 150 ksi

A = 50 in2
I = 20,000 in4
Z = 1,000 in3
σy = 50 ksi
Load
Factor
2
2nd-Order 1st-Order Inelastic
1.5 Inelastic
1st-Order
1 Inelastic

0.5 2nd-Order Inelastic

0
0 50 100 150 200 250
vmid (in.) Hinge Formation
Isometric View Elevation View Plan View

16