CE 809 - Structural Dynamics

Lecture 6: Response of SDF Systems to General Dynamic Loading

Semester – Fall 2020

Dr. Fawad A. Najam Prof. Dr. Pennung Warnitchai

Department of Structural Engineering
NUST Institute of Civil Engineering (NICE)
National University of Sciences and Technology (NUST)
H-12 Islamabad, Pakistan
Cell: 92-334-5192533, Email: fawad@nice.nust.edu.pk
Head, Department of Civil and Infrastructure Engineering
School of Engineering and Technology (SET)
Asian Institute of Technology (AIT)
Bangkok, Thailand
 Response to

General Dynamic Loading

Duhamel’s Integral Step-by-step Direct Integration

(Convolution Integral) Methods

• Based on the principle of • Based on the direct numerical

superposition (It considers the integration of the governing
general dynamic loading as a series equation of motion in incremental
of short impulses). form.

• Applicable only to linear systems • Applicable to linear and nonlinear

systems

2
Duhamel’s Integral

3
 Duhamel’s Integral
𝑝(𝑡)
 A general dynamic loading =
A series of short Impulses

 Each impulse produce its

𝑝(𝜏) own (impulse) response

0 𝑡
𝜏  The sum of these impulse
𝑝(𝜏) 𝑑𝜏
𝑑𝜏 𝑒 −𝜉 𝜔 (𝑡−𝜏) sin 𝜔𝐷 𝑡 − 𝜏 responses = the response to
𝑑𝑢(𝑡) 𝑚 𝜔𝐷 the dynamic loading

0 𝑡

4
 Duhamel’s Integral
Let 𝑑𝑢 (𝑡; 𝜏) is the response of a linear dynamic system at time 𝑡 due to impulse 𝑝(𝜏) 𝑑𝜏 at
time 𝜏.

𝑑𝑢 (𝑡; 𝜏) = 𝑝(𝜏) 𝑑𝜏 . ℎ(𝑡 − 𝜏) ………. (1)

Where

𝑒 −𝜉𝜔 𝑡−𝜏
ℎ (𝑡 − 𝜏) = sin 𝜔𝐷 𝑡 − 𝜏 , 𝑡>𝜏 ………. (2)
𝑚 𝜔𝐷
0 , 𝑡≤𝜏

ℎ (𝑡 − 𝜏) = unit impulse response (or response to unit impulse applied at 𝑡 = 𝜏).

5
 𝑑𝑢(𝑡; 𝜏) = 𝑝(𝜏)𝑑𝜏 . ℎ (𝑡 − 𝜏)

𝑑𝑢(𝑡; 𝜏 = 0. Δ𝑡) = 𝑝 (0. Δ𝑡) 𝑑Δ𝑡 . ℎ (𝑡 − 0. Δ𝑡)

𝑑𝑢(𝑡; 𝜏 = 1. Δ𝑡) = 𝑝 (1. Δ𝑡) 𝑑Δ𝑡 . ℎ (𝑡 − 1. Δ𝑡)

𝑑𝑢(𝑡; 𝜏 = 2. Δ𝑡) = 𝑝 (2. Δ𝑡) 𝑑Δ𝑡 . ℎ (𝑡 − 2. Δ𝑡)

𝑑𝑢(𝑡; 𝜏 = 𝑖. Δ𝑡) = 𝑝 (𝑖. Δ𝑡) 𝑑Δ𝑡 . ℎ (𝑡 − 𝑖. Δ𝑡)

𝜏= 𝑡

𝑢(𝑡) = 𝑝 𝜏 . ℎ (𝑡 − 𝜏) 𝑑𝜏
𝜏= 0

AIT Share 6
 Response to General Dynamic Loading - Duhamel’s Integral

By means of superposition the total responsive 𝑢(𝑡) can be obtained by summing all
impulse responses developed during the loading history.

𝜏= 𝑡

𝑢(𝑡) = 𝑝 𝜏 . ℎ (𝑡 − 𝜏) 𝑑𝜏 ………. (3)

𝜏= 0

The integration is called “Convolution Integral” in general theory of mathematics and

“Duhamel’s Integral” in structural dynamics.

7
 Response to General Dynamic Loading - Duhamel’s Integral

In Equation (3), it is assumed that the structure is initially at-rest condition

That is 𝑢 (0) = 0, 𝑢 (0) = 0.

For other cases, additional free vibration response must be added to the solution:

𝑡
𝑢 (0) + 𝑢 (0) 𝜉 𝜔
𝑢 (𝑡) = 𝑒 −𝜉 𝜔 𝑡 sin(𝜔𝐷 𝑡) + 𝑢 (0) cos(𝜔𝐷 𝑡) + 𝑝 𝜏 ℎ (𝑡 − 𝜏) 𝑑𝜏 ….. (4)
𝜔𝐷
0

In the following investigation, the initial at-rest condition is assumed.

8
 Response to General Dynamic Loading - Duhamel’s Integral

𝑡
1
𝑢 (𝑡) = 𝑝 𝜏 . 𝑒 −𝜉 𝜔 𝑡−𝜏
sin 𝜔𝐷 𝑡 − 𝜏 𝑑𝜏
𝑚 𝜔𝐷
0
𝑒 −𝜉 𝜔 𝑡 . 𝑒 𝜉 𝜔 𝜏 sin 𝜔𝐷 𝑡 cos(𝜔𝐷 𝜏) − cos(𝜔𝐷 𝑡) sin(𝜔𝐷 𝜏)

Therefore,

𝑡
1
𝑢 (𝑡) = 𝑝 𝜏 . 𝑒 −𝜉𝜔𝑡 𝑒 𝜉𝜔𝜏 sin 𝜔𝐷 𝑡 cos 𝜔𝐷 𝜏) − cos 𝜔𝐷 𝑡 sin 𝜔𝐷 𝜏 𝑑𝜏
𝑚 𝜔𝐷
0

9
 Response to General Dynamic Loading - Duhamel’s Integral
By rearranging the terms, we get

𝑡 𝑡
−𝜉𝜔𝑡 −𝜉𝜔𝑡
𝑒 𝜉𝜔𝜏
𝑒
𝑢 𝑡 = 𝑝 𝜏 .𝑒 cos 𝜔𝐷 𝜏 𝑑𝜏 sin 𝜔𝐷 𝑡 − 𝑝 𝜏 . 𝑒 𝜉𝜔𝜏 sin 𝜔𝐷 𝜏 𝑑𝜏 cos 𝜔𝐷 𝑡
𝑚 𝜔𝐷 𝑚 𝜔𝐷
0 0

So we can write 𝑢 𝑡 = 𝐴 𝑡 sin 𝜔𝐷 𝑡 − 𝐵 𝑡 cos 𝜔𝐷 𝑡

Where 𝑡
𝑒 −𝜉𝜔𝑡
𝐴𝑡 = 𝑝 𝜏 . 𝑒 𝜉𝜔𝜏 . cos 𝜔𝐷 𝜏 𝑑𝜏
𝑚 𝜔𝐷 0
𝑡
𝑒 −𝜉𝜔𝑡
𝐵𝑡 = 𝑝 𝜏 . 𝑒 𝜉𝜔𝜏 . sin 𝜔𝐷 𝜏 𝑑𝜏
𝑚 𝜔𝐷 0

10
 Response to General Dynamic Loading - Duhamel’s Integral

For undamped case,

𝑡
1
𝐴𝑡 = 𝑝 𝜏 . cos(𝜔 𝜏) 𝑑𝜏
𝑚𝜔 0

𝑡
1
𝐵𝑡 = 𝑝 𝜏 . sin(𝜔 𝜏) 𝑑𝜏
𝑚𝜔 0

The terms in parenthesis for (both damped and undamped cases) need “numerical integration”.

11
 Numerical Integration

Simple Summation:
𝑡
𝑓 𝜏 𝑑 𝜏 ≅ Δ𝜏 𝑓0 + 𝑓1 + 𝑓2 + 𝑓3 + ⋯ 𝑓𝑁−1
0

𝑓𝜏
Where 𝑓𝑖 = 𝑓(𝜏 = 𝑖. Δ𝜏) , and Δ𝜏 = 𝑡/𝑁

𝜏
𝑡
Δ𝜏
12
 Numerical Integration

Trapezoidal Rule:
𝑡
Δ𝜏
𝑓 𝜏 𝑑𝜏 ≅ 𝑓0 + 2𝑓1 + 2𝑓2 + 2𝑓3 + ⋯ 2𝑓𝑁−1 + 𝑓𝑁
0 2

Where 𝑓𝑖 = 𝑓(𝑖. Δ𝜏) , and Δ𝜏 = 𝑡/𝑁 𝑓𝜏 𝑓𝑖 + 𝑓𝑖+1

. Δ𝜏
2

13
 Solving Duhamel’s Integral using Numerical Integration

• For example, consider the numerical integration of

a function 𝑦 𝜏 = 𝑝 𝜏 cos 𝜔𝜏 as required to find
𝐴 𝑡 in Duhamel’s Integral.

• For convenience of numerical calculation, the

function 𝑦 𝜏 is evaluated at equal time increments
Δ𝜏 as shown in Figure.

• The integral 𝐴𝑁 can now be obtained

approximately by summing the ordinates, after
multiplying by weighting actors that depend on the
numerical integration scheme being used.

Source: Clough and Penzien (2003)

14
 Solving Duhamel’s Integral using Numerical Integration
Undamped Systems
Simple summation:

Δ𝜏
𝐴𝑁 = 𝑦0 + 𝑦1 + 𝑦2 + ⋯ + 𝑦𝑁−1
𝑚𝜔

Trapezoidal rule:

Δ𝜏
𝐴𝑁 = [𝑦0 + 2𝑦1 + 2𝑦2 + ⋯ + 2𝑦𝑁−1 + 𝑦𝑁 ]
2𝑚𝜔

Simpson’s rule:

Δ𝜏
𝐴𝑁 = [𝑦0 + 4𝑦1 + 2𝑦2 + ⋯ + 4𝑦𝑁−1 + 𝑦𝑁 ]
3𝑚𝜔
Source: Clough and Penzien (2003) 15
It is more efficient to write these equations in their recursive forms:

Simple summation:

Δ𝜏
𝐴𝑁 = 𝐴𝑁−1 + 𝑦 𝑁 = 1, 2, 3, …
𝑚 𝜔 𝑁−1

Trapezoidal rule:

Δ𝜏 𝑁 = 1, 2, 3, …
𝐴𝑁 = 𝐴𝑁−1 + 𝑦𝑁−1 + 𝑦𝑁
2𝑚𝜔

Simpson’s rule:

Δ𝜏
𝐴𝑁 = 𝐴𝑁−1 + 𝑦𝑁−2 + 4𝑦𝑁−1 + 𝑦𝑁 𝑁 = 2, 4, 6, …
3𝑚𝜔
Such that 𝐴0 =0
16
Source: Clough and Penzien (2003)
 Solving Duhamel’s Integral using Numerical Integration
Damped Systems
Simple summation:

Δ𝜏
𝐴𝑁 = 𝑒 −𝜉 𝜔 Δ𝜏 𝐴𝑁−1 + 𝑦𝑁−1 𝑒 −𝜉 𝜔 Δ𝜏 𝑁 = 1, 2, 3, …
𝑚 𝜔𝐷
Trapezoidal rule:

Δ𝜏
𝐴𝑁 = 𝑒 −𝜉 𝜔 Δ𝜏 𝐴𝑁−1 + 𝑦𝑁−1 𝑒 −𝜉 𝜔 Δ𝜏 + 𝑦𝑁 𝑁 = 1, 2, 3, …
2 𝑚 𝜔𝐷
Simpson’s rule:

Δ𝜏
𝐴𝑁 = 𝑒 −2 𝜉 𝜔 Δ𝜏 𝐴𝑁−1 + 𝑦𝑁−2 𝑒 −2 𝜉 𝜔 Δ𝜏 + 4 𝑦𝑁−1 𝑒 −𝜉 𝜔 Δ𝜏 + 𝑦𝑁 𝑁 = 2, 4, 6, …
3 𝑚 𝜔𝐷

17
 Solving Duhamel’s Integral using Numerical Integration

• The evaluation of 𝐵 𝑡 can be carried out in the same manner, however, in doing so, the
definition of 𝑦 𝜏 must be changed to 𝑦 𝜏 = 𝑝 𝜏 sin(𝜔𝜏).

• Having calculated the values of 𝐴𝑁 and 𝐵𝑁 for successive values of 𝑁 , the

corresponding values of response 𝑢𝑁 are obtained using

𝑢𝑁 = 𝐴𝑁 sin 𝜔𝑡𝑁 − 𝐵𝑁 cos 𝜔𝑡𝑁

Source: Clough and Penzien (2003)

18
 Numerical Example
Taken from Clough and Penzien (2003)
The unit of force is kips.
1.0 kip = 1000 lbs

The unit of displacement is ft.

The unit of velocity is ft/s.
The unit of acceleration is ft/s2.

1.0 g = 32.2 ft/s2

The unit of mass is the unit of force

divided by the unit of acceleration.
Kip- s2 /ft

A water tower subjected to blast load The unit of stiffness is the unit of force
divided by the unit of displacement.
Kip/ft
19
 Numerical Example
Taken from Clough and Penzien (2003)

20
21
 Numerical Example
Taken from Clough and Penzien (2003)

22
 Duhamel’s Integral

• The concept of convolution integral will be used again later when we study the response
of structures to random loadings from statistical view point (random vibration theory).

• The Convolution Integral is derived based on the principle of superposition. So, it is

applicable only for the response analysis of “linear systems”.

23
 Applications
Evaluation of Structural Response to Earthquake Ground Motions

𝑑 2 𝑢(𝑡)
𝑡 𝑑𝑢 (𝑡)
𝑡 𝑑 2 𝑢𝑔(𝑡)
𝑡
𝑚 +𝑐 + 𝑘 𝑢 𝑡 = −𝑚
(𝑡)
𝑑𝑡 2 𝑑𝑡 𝑑𝑡 2
𝑚
𝑢𝑔(𝑡)
(𝑡)

𝑢𝑔(𝑡)
(𝑡)

24
 Applications

Evaluation of Response Spectrum of

Earthquake Ground Motions

Computation of deformation (or

displacement) response spectrum
25
Step-by-step Direct Integration Method
Or Time-stepping Method

26
 Step-by-step Direct Integration Method

 General Dynamic loadings

 Linear & Nonlinear Structures

 In some important structural dynamic problems, the responses of structures are in

nonlinear range.

For example, the response of a structure subjected to a major earthquake.

27
 𝑢
𝑓𝑠 𝑓𝑠
𝑓𝑠 Linear
Elastic
Column
𝑓𝑠 = 𝑘 𝑢

𝑢 𝑢

𝑓𝑠 Steel Column
RC Column Elasto-plastic

Hysteretic Concrete Cracking

Response Rebar Yielding
𝑢 Nonlinear System Duhamel’s Integral is
Nonlinear Equation → not applicable

28
 Step-by-step Integration Procedure
Consider the dynamic equilibrium (in scalar form) of a nonlinear structure at time 𝑡 :

𝑓𝐼 𝑡 + 𝑓𝐷 𝑡 + 𝑓𝑠 𝑡 = 𝑝 𝑡 ………. (1)

Where
𝑓𝐼 𝑡 = 𝑚 𝑢 𝑡 𝑝 𝑡 is an arbitrary/general
𝑓𝐷 𝑡 ≠ 𝑐𝑢 𝑡 loading.

𝑝𝑡
Damping force may not be a linear function of velocity 𝑢 𝑡 .

𝑓𝑠 𝑡 ≠ 𝑘 𝑢 𝑡
𝑡
Restoring force is a nonlinear function of displacement 𝑢 𝑡 .

29
 Step-by-step Integration Procedure
𝑓𝐼 𝑡 + 𝑓𝐷 𝑡 + 𝑓𝑠 𝑡 = 𝑝𝑡 ………. (1)

At a small time Δ𝑡 later:

𝑓𝐼 𝑡 + ∆𝑡 + 𝑓𝐷 𝑡 + ∆𝑡 + 𝑓𝑠 𝑡 + ∆𝑡 = 𝑝 𝑡 + ∆𝑡 ………. (2)

Subtract Equation (2) by Equation (1), we get

Δ𝑓𝐼 𝑡 + Δ𝑓𝐷 𝑡 + Δ𝑓𝑠 𝑡 = Δ 𝑝 𝑡 ………. (3)

Where Δ𝑓𝐼 𝑡 = 𝑓𝐼 𝑡 + ∆𝑡 − 𝑓𝐼 𝑡 = 𝑚 Δ𝑢 𝑡

𝑑𝑓𝐷
Δ𝑓𝐷 𝑡 = 𝑓𝐷 𝑡 + ∆𝑡 − 𝑓𝐷 𝑡 ≅ . Δ𝑢 𝑡 = 𝑐 𝑡 . Δ𝑢 𝑡
𝑑𝑢 𝑡

𝑑𝑓𝑠
Δ𝑓𝑠 𝑡 = 𝑓𝑠 𝑡 + ∆𝑡 − 𝑓𝑠 𝑡 ≅ . Δ𝑢 𝑡 = 𝑘 𝑡 . Δ𝑢 𝑡
𝑑𝑢 𝑡 30
 𝑝 𝑡 + ∆𝑡
𝑝𝑡 Δ𝑝 𝑡 = 𝑝 (𝑡 + Δ𝑡) − 𝑝 𝑡

Time
𝑓𝑠 𝑡 𝑡 + Δ𝑡
Slope
𝑑𝑓𝑠
𝑓𝑠 𝑡 + ∆𝑡 Δ𝑓𝑠 𝑡 =
𝑑𝑢 𝑡
Δ𝑓𝑠 𝑡 = 𝑓𝑠 𝑡 + ∆𝑡 − 𝑓𝑠 𝑡
=𝑘𝑡
𝑓𝑠 𝑡 𝑑𝑓𝑠
Δ𝑢 𝑡 ≅ . Δ𝑢 𝑡
𝑑𝑢 𝑡

≅ 𝑘 𝑡 . Δ𝑢
𝑢
𝑢𝑡 𝑢 𝑡 + ∆𝑡
31
 Step-by-step Integration Procedure

We have introduced the following two approximations:

Δ𝑓𝐷 𝑡 = 𝑐 𝑡 . Δ𝑢 𝑡
𝑑𝑓𝑠
=
𝑑𝑢 𝑡

Δ𝑓𝑠 𝑡 = 𝑘 𝑡 . Δ𝑢 𝑡

They are equivalent to the assumption that the damping force

and restoring forces are linear within 𝑡 and 𝑡 + Δ𝑡.

“Piecewise Linear Approximation of Structural System”

32
 Step-by-step Integration Procedure

The incremental equation of motion (Equation 3) becomes,

𝑚 Δ𝑢 𝑡 + 𝑐 𝑡 Δ𝑢 𝑡 + 𝑘 𝑡 Δ𝑢 𝑡 = Δ𝑝 𝑡 ………. (4)

Introducing an assumption that “the acceleration response varies linearly during each
time increment”.

This yields quadratic and cubic variations of velocity and displacement, respectively.

33
 𝑢 𝑡 + Δ𝑢 𝑡
𝑢𝑡 “the acceleration response
Δ𝑢 𝑡
∆𝑡 𝑢 𝜏 =𝑢 𝑡 + .𝜏 varies linearly during each ………. (5)
∆𝑡 time increment”.

𝑡 𝑡 + ∆𝑡 𝜏
𝜏 𝑢 𝜏 =𝑢 𝑡 + 𝑢 𝜏 𝑑𝜏
𝑢 𝑡 + ∆𝑢 𝑡 0
………. (6)
𝑢𝑡 Δ𝑢 𝑡 𝜏 2
𝑢 𝜏 = 𝑢 𝑡 + 𝑢 𝑡 .𝜏 + .
∆𝑡 2

∆𝑢 𝜏
𝜏
𝑢𝜏 =𝑢 𝑡 + 𝑢 𝜏 𝑑𝜏
𝑢 𝑡 + ∆𝑢 𝑡 0

𝑢 𝑡 𝜏 2 Δ𝑢 𝑡 𝜏 3 (7)
𝑢 𝜏 = 𝑢 𝑡 + 𝑢 𝑡 .𝜏 + 𝑢 𝑡 . + .
2 ∆𝑡 6
34
∆𝑢 𝜏
 Step-by-step Integration Procedure
At 𝜏 = ∆𝑡, the above equations for velocity and displacement becomes,
Δ𝑢 𝑡 ∆𝑡 2 ………. (8)
∆𝑢 𝑡 = 𝑢 𝑡 . ∆𝑡 + .
∆𝑡 2
∆𝑡 2 Δ𝑢 𝑡 ∆𝑡 3
∆𝑢 𝑡 = 𝑢 𝑡 . ∆𝑡 + 𝑢 𝑡 . + . ………. (9)
2 ∆𝑡 6

Re-writing the above two equations in terms of Δ𝑢 𝑡 :

6 6 ………. (10)
∆𝑢 𝑡 = 2 . ∆𝑢 𝑡 − . 𝑢 𝑡 − 3 𝑢 𝑡
∆𝑡 ∆𝑡
3 ∆𝑡
Δ𝑢 𝑡 = . ∆𝑢 𝑡 − 3 𝑢 𝑡 − . 𝑢 𝑡 ………. (11)
∆𝑡 2
Equations (10) and (11) are derived from the “linear acceleration assumption”.
35
 Step-by-step Integration Procedure
Introducing Equations (10) and (11) into the incremental form of governing equation of
motion (Equation (4)), we obtain

6 6 3 ∆𝑡
𝑚 . ∆𝑢 𝑡 − . 𝑢 𝑡 − 3 𝑢 𝑡 + 𝑐 𝑡 . ∆𝑢 𝑡 − 3 𝑢 𝑡 − . 𝑢 𝑡 + 𝑘 𝑡 . ∆𝑢 𝑡 = ∆𝑝 𝑡
∆𝑡 2 ∆𝑡 ∆𝑡 2

Re-writing the above equation, we get,

𝑘 𝑡 . ∆𝑢 𝑡 = ∆ 𝑝 𝑡 ………. (12)
Where
6 3
𝑘 𝑡 = 𝑘 𝑡 + 2.𝑚 + .𝑐 𝑡
∆𝑡 ∆𝑡
6 ∆𝑡
∆𝑝 𝑡 = ∆𝑝 𝑡 + 𝑚 .𝑢 𝑡 + 3 𝑢 𝑡 +𝑐 𝑡 3 𝑢 𝑡 + .𝑢 𝑡
∆𝑡 2
36
 Step-by-step Integration Procedure
𝑘 𝑡 . ∆𝑢 𝑡 = ∆ 𝑝 𝑡 ………. (12)

Where
6 3
𝑘 𝑡 = 𝑘 𝑡 + 2.𝑚 + .𝑐 𝑡
∆𝑡 ∆𝑡
6 ∆𝑡
∆𝑝 𝑡 = ∆𝑝 𝑡 + 𝑚 .𝑢 𝑡 + 3 𝑢 𝑡 + 𝑐 𝑡 3 𝑢 𝑡 + .𝑢 𝑡
∆𝑡 2

Let’s assume that the calculation is made up to Time = 𝑡 and we are going to proceed to the
next time stop, 𝑡 + ∆𝑡.

Hence, 𝑢 𝑡 , 𝑢 𝑡 , 𝑢 𝑡 are known, and 𝑘 𝑡 , 𝑐 𝑡 , 𝑚 and ∆𝑝 𝑡 are also known.

∆𝑢 𝑡 can be determined. ∆𝑢 𝑡 and ∆𝑢 𝑡 can be derived from ∆𝑢 𝑡 by Eqs (11) and (10).
37
 Step-by-step Integration Procedure
Note:
Two assumptions are used in this step-by-step calculation.

1) Within 𝑡, 𝑡 + Δ𝑡 , Δ𝑓𝐷 𝑡 = 𝑐 𝑡 . Δ𝑢 𝑡 and Δ𝑓𝑠 𝑡 = 𝑘 𝑡 . Δ𝑢 𝑡

2) Within {𝑡, 𝑡 + Δ𝑡}, acceleration varies linearly

These assumptions are justified only when Δ𝑡 is sufficiently small, small Δ𝑡  small error.

Although the error in each step is small, the error can accumulate and becomes significant
when the number of steps is large.

The accumulation should be avoided by imposing the dynamic equilibrium condition at each
time step.

38
Time = 𝒕 𝑢(𝑡) and 𝑢(𝑡) are known Calculation
flow chart
Evaluate 𝑑𝑓𝑠 𝑑𝑓𝐷 𝑓𝑠 𝑡 = 𝑓𝑠 𝑢(𝑡)
𝑘 𝑡 = , 𝑐 𝑡 = ,
𝑑𝑢 𝑡
𝑑𝑢 𝑡 𝑓𝐷 𝑡 = 𝑓𝐷 𝑢(𝑡)

Impose dynamic 1
𝑢 𝑡 = . 𝑝 𝑡 − 𝑓𝐷 𝑡 − 𝑓𝑠 𝑡 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (1)
equilibrium condition 𝑚

6 3
𝑘 𝑡 =𝑘 𝑡 + . 𝑚 + .𝑐 𝑡
∆𝑡 2 ∆𝑡
6 ∆𝑡
∆𝑝 𝑡 = ∆𝑝 𝑡 + 𝑚 .𝑢 𝑡 + 3𝑢 𝑡 + 𝑐 𝑡 3 𝑢 𝑡 + .𝑢 𝑡
∆𝑡 2

∆𝑢 𝑡 = ∆𝑝 𝑡 /𝑘 𝑡 Equation (12)

3 ∆𝑡
∆𝑢 𝑡 = ∆𝑢 𝑡 − 3𝑢 𝑡 − 𝑢 𝑡 Equation (11)
∆𝑡 2

𝑢 𝑡 + ∆𝑡 = 𝑢 𝑡 + ∆𝑢 𝑡
Time = 𝒕 + ∆𝒕
𝑢 𝑡 + ∆𝑡 = 𝑢 𝑡 + ∆𝑢 𝑡 39
 Additional Notes
1. Response of any SDF system with any prescribed nonlinear properties can be evaluated
by “step-by-step integration”.

2. Response of any linear SDF system can also be evaluated by the step-by-step
integration.

3. To determine Δ𝑡 , we should consider :

 The rate of variation of the applied loading 𝑝 (𝑡)

 The nonlinearity of damping and stiffness properties.

 The natural period of structure (𝑇)

40
 Additional Notes
The choice of Δ𝑡 also depends on the nonlinear properties of damping and stiffness

Rule of thumb: 𝑓𝑠 Need a very

Δ𝑡/𝑇 ≤ 1/10 small ∆𝑡

My suggestion:
Δ𝑡/𝑇 ≤ 1/30 𝑢

4. The step-by-step integration technique will be extended for the calculation of responses
of nonlinear MDF systems later.
More attention will be paid on the accumulation of error – as it is a major factor in the
determination of Δ𝑡.

41
 Numerical Example
Taken from Clough and Penzien (2003)

𝑢(𝑡)

An elastoplastic frame
and dynamic loading

42
Numerical Example Taken from Clough and Penzien (2003)

43
 Numerical Example Taken from Clough and Penzien (2003)

Comparison of elastoplastic with elastic response 44

Thank you