Académique Documents
Professionnel Documents
Culture Documents
SD An Najah 2013 DS-alesandro PDF
SD An Najah 2013 DS-alesandro PDF
An-Najah National University Course “Fundamentals of Structural Dynamics” April 19 - April 23, 2013
April 19 - April 23, 2013
Lecturer: Dr. Alessandro Dazio, UME School
4 Schedule of classes
Date Time Topic
Fundamentals of Structural Dynamics 09:00 - 10:30 1. Introduction
2. SDoF systems: Equation of motion and modelling
Day 1
Fri. April 19 11:00 - 12:30 3. Free vibrations
1 Course description 2013 14:30 - 16:00 Assignment 1
Aim of the course is that students develop a “feeling for dynamic problems” and acquire the theoretical 16:30 - 18:00 Assignment 1
background and the tools to understand and to solve important problems relevant to the linear and, in 9:00 - 10:30 4. Harmonic excitation
part, to the nonlinear dynamic behaviour of structures, especially under seismic excitation. Day 2 11:00 - 12:30 5. Transfer functions
The course will start with the analysis of single-degree-of-freedom (SDoF) systems by discussing: (i) Sat. April 20
Modelling, (ii) equations of motion, (iii) free vibrations with and without damping, (iv) harmonic, pe- 2013 14:30 - 16:00 6. Forced vibrations (Part 1)
riodic and short excitations, (v) Fourier series, (vi) impacts, (vii) linear and nonlinear time history anal- 16:30 - 18:00 6. Forced vibrations (Part 2)
ysis, and (viii) elastic and inelastic response spectra.
09:00 - 10:30 7. Seismic excitation (Part 1)
Afterwards, multi-degree-of-freedom (MDoF) systems will be considered and the following topics will
Day 3 11:00 - 12:30 7. Seismic excitation (Part 2)
be discussed: (i) Equation of motion, (ii) free vibrations, (iii) modal analysis, (iv) damping, (v) Rayleigh’s Sun. April 21
quotient, and (vi) seismic behaviour through response spectrum method and time history analysis. 2013 14:30 - 16:00 Assignment 2
To supplement the suggested reading, handouts with class notes and calculation spreadsheets with se- 16:30 - 18:00 Assignment 2
lected analysis cases to self-training purposes will be distributed.
9:00 - 10:30 8. MDoF systems: Equation of motion
Lecturer: Dr. Alessandro Dazio, UME School
Day 4 11:00 - 12:30 9. Free vibrations
Mon. April 22 14:30 - 16:00 10. Damping
2 Suggested reading 2013 11. Forced vibrations
[Cho11] Chopra A., “Dynamics of Structures”, Prentice Hall, Fourth Edition, 2011. 16:30 - 18:00 11. Forced vibrations
[CP03] Clough R., Penzien J., “Dynamics of Structures”, Second Edition (revised), Computer and 09:00 - 10:30 12. Seismic excitation (Part 1)
Structures Inc., 2003. Day 5 11:00 - 12:30 12. Seismic excitation (Part 2)
[Hum12] Humar J.L., “Dynamics of Structures”. Third Edition. CRC Press, 2012. Tue. April 23
2013 14:30 - 16:00 Assignment 3
16:30 - 18:00 Assignment 3
3 Software
In the framework of the course the following software will be used by the lecturer to solve selected ex-
amples:
[Map10] Maplesoft: “Maple 14”. User Manual. 2010
[Mic07] Microsoft: “Excel 2007”. User Manual. 2007
[VN12] Visual Numerics: “PV Wave”. User Manual. 2012
As an alternative to [VN12] and [Map10] it is recommended that students make use of the following
software, or a previous version thereof, to deal with coursework:
[Mat12] MathWorks: “MATLAB 2012”. User Manual. 2012
6.1.2 Half-sine ........................................................................................................ 6-5 7.5 Elastic response spectra ................................................................... 7-42
6.1.3 Example: “Jumping on a reinforced concrete beam”............................... 6-7 7.5.1 Computation of response spectra ............................................................ 7-42
6.2 Short excitation .................................................................................. 6-12 7.5.2 Pseudo response quantities...................................................................... 7-45
6.2.1 Step force .................................................................................................... 6-12 7.5.3 Properties of linear response spectra ..................................................... 7-49
6.2.2 Rectangular pulse force excitation .......................................................... 6-14 7.5.4 Newmark’s elastic design spectra ([Cho11]) ........................................... 7-50
6.2.3 Example “blast action” .............................................................................. 6-21 7.5.5 Elastic design spectra in ADRS-format (e.g. [Faj99])
(Acceleration-Displacement-Response Spectra) .................................... 7-56
7.2 Time-history analysis of linear SDoF systems ................................. 7-3 7.6.3 Inelastic behaviour of a RC wall during an earthquake ........................ 7-63
7.2.1 Newmark’s method (see [New59]) .............................................................. 7-4 7.6.4 Static-cyclic behaviour of a RC wall ........................................................ 7-64
7.2.2 Implementation of Newmark’s integration scheme within 7.6.5 General definition of ductility ................................................................... 7-66
the Excel-Table “SDOF_TH.xls”.................................................................. 7-8 7.6.6 Types of ductilities .................................................................................... 7-67
7.2.3 Alternative formulation of Newmark’s Method. ....................................... 7-10
7.7 Inelastic response spectra ............................................................... 7-68
7.3 Time-history analysis of nonlinear SDoF systems ......................... 7-12 7.7.1 Inelastic design spectra............................................................................. 7-71
7.3.1 Equation of motion of nonlinear SDoF systems ..................................... 7-13 7.7.2 Determining the response of an inelastic SDOF system
7.3.2 Hysteretic rules........................................................................................... 7-14 by means of inelastic design spectra in ADRS-format ........................... 7-80
7.3.3 Newmark’s method for inelastic systems ................................................ 7-18 7.7.3 Inelastic design spectra: An important note............................................ 7-87
7.3.4 Example 1: One-storey, one-bay frame ................................................... 7-19 7.7.4 Behaviour factor q according to SIA 261 ................................................. 7-88
7.3.5 Example 2: A 3-storey RC wall .................................................................. 7-23 7.8 Linear equivalent SDOF system (SDOFe) ....................................... 7-89
7.4 Solution algorithms for nonlinear analysis problems .................... 7-26 7.8.1 Elastic design spectra for high damping values ..................................... 7-99
7.4.1 General equilibrium condition................................................................. 7-26 7.8.2 Determining the response of inelastic SDOF systems
by means of a linear equivalent SDOF system and
7.4.2 Nonlinear static analysis ........................................................................... 7-26 elastic design spectra with high damping ............................................. 7-103
7.4.3 The Newton-Raphson Algorithm............................................................... 7-28
7.9 References ........................................................................................ 7-108
7.4.4 Nonlinear dynamic analyses ..................................................................... 7-35
7.4.5 Comments on the solution algorithms for
nonlinear analysis problems ..................................................................... 7-38 8 Multi Degree of Freedom Systems
7.4.6 Simplified iteration procedure for SDoF systems with
idealised rule-based force-deformation relationships............................ 7-41 8.1 Formulation of the equation of motion............................................... 8-1
8.1.1 Equilibrium formulation ............................................................................... 8-1
8.1.2 Stiffness formulation ................................................................................... 8-2
8.1.3 Flexibility formulation ................................................................................. 8-3 10.3Classical damping matrices .............................................................. 10-5
8.1.4 Principle of virtual work............................................................................... 8-5 10.3.1 Mass proportional damping (MpD) ......................................................... 10-5
8.1.5 Energie formulation...................................................................................... 8-5 10.3.2 Stiffness proportional damping (SpD) .................................................... 10-5
8.1.6 “Direct Stiffness Method”............................................................................ 8-6 10.3.3 Rayleigh damping..................................................................................... 10-6
8.1.7 Change of degrees of freedom.................................................................. 8-11 10.3.4 Example....................................................................................................... 10-7
8.1.8 Systems incorporating rigid elements with distributed mass ............... 8-14
11 Forced Vibrations
9 Free Vibrations
11.1Forced vibrations without damping ................................................. 11-1
9.1 Natural vibrations ................................................................................. 9-1 11.1.1 Introduction ................................................................................................ 11-1
9.2 Example: 2-DoF system ...................................................................... 9-4 11.1.2 Example 1: 2-DoF system.......................................................................... 11-3
9.2.1 Eigenvalues ................................................................................................ 9-4 11.1.3 Example 2: RC beam with Tuned Mass Damper (TMD) without damping...
11-7
9.2.2 Fundamental mode of vibration .................................................................. 9-5
9.2.3 Higher modes of vibration ........................................................................... 9-7 11.2Forced vibrations with damping ..................................................... 11-13
9.2.4 Free vibrations of the 2-DoF system .......................................................... 9-8 11.2.1 Introduction .............................................................................................. 11-13
9.3 Modal matrix and Spectral matrix ..................................................... 9-12 11.3Modal analysis: A summary ............................................................ 11-15
12.4.3 Equation of motion in modal coordinates.............................................. 12-38 14 Pedestrian Footbridge with TMD
12.4.4 Response spectrum method ................................................................... 12-41
12.4.5 Response spectrum method vs. time-history analysis ........................ 12-50 14.1Test unit and instrumentation ........................................................... 14-1
14.2Parameters .......................................................................................... 14-4
13 Vibration Problems in Structures 14.2.1 Footbridge (Computed, without TMD) ...................................................... 14-4
14.2.2 Tuned Mass Damper (Computed) ............................................................. 14-4
13.1Introduction ........................................................................................ 13-1
13.1.1 Dynamic action ........................................................................................... 13-2
14.3Test programme ................................................................................. 14-5
13.1.2 References .................................................................................................. 13-3 14.4Free decay test with locked TMD ...................................................... 14-6
13.2Vibration limitation ............................................................................. 13-4 14.5Sandbag test ....................................................................................... 14-8
13.2.1 Verification strategies ................................................................................ 13-4 14.5.1 Locked TMD, Excitation at midspan ....................................................... 14-9
13.2.2 Countermeasures ....................................................................................... 13-5 14.5.2 Locked TMD, Excitation at quarter-point of the span .......................... 14-12
13.2.3 Calculation methods .................................................................................. 13-6 14.5.3 Free TMD: Excitation at midspan .......................................................... 14-15
13.3People induced vibrations................................................................. 13-8 14.6One person walking with 3 Hz......................................................... 14-17
13.3.1 Excitation forces......................................................................................... 13-8
14.7One person walking with 2 Hz......................................................... 14-20
13.3.2 Example: Jumping on an RC beam....................................................... 13-15
14.7.1 Locked TMD (Measured) .......................................................................... 14-20
13.3.3 Footbridges............................................................................................... 13-18
14.7.2 Locked TMD (ABAQUS-Simulation) ...................................................... 14-22
13.3.4 Floors in residential and office buildings .............................................. 13-26
14.7.3 Free TMD .................................................................................................. 14-24
13.3.5 Gyms and dance halls.............................................................................. 13-29
14.7.4 Remarks about “One person walking with 2 Hz” .................................. 14-25
13.3.6 Concert halls, stands and diving platforms........................................... 13-30
14.8Group walking with 2 Hz .................................................................. 14-26
13.4Machinery induced vibrations......................................................... 13-30
14.8.1 Locked TMD ............................................................................................ 14-29
13.5Wind induced vibrations.................................................................. 13-31 14.8.2 Free TMD ................................................................................................. 14-30
13.5.1 Possible effects ........................................................................................ 13-31
14.9One person jumping with 2 Hz ........................................................ 14-31
13.6Tuned Mass Dampers (TMD) ........................................................... 13-34 14.9.1 Locked TMD .............................................................................................. 14-31
13.6.1 Introduction............................................................................................... 13-34 14.9.2 Free TMD ................................................................................................. 14-33
13.6.2 2-DoF system ........................................................................................... 13-35 14.9.3 Remarks about “One person jumping with 2 Hz” ................................. 14-34
13.6.3 Optimum TMD parameters....................................................................... 13-39
13.6.4 Important remarks on TMD...................................................................... 13-39
• No soil-dynamics and no dynamic soil-structure interaction will - Seismic response through response spectrum method and
be treated (this is the topic of another course); time-history analysis
1.4 References
Blank page
Theory
[Bat96] Bathe KJ: “Finite Element Procedures”. Prentice Hall, Upper
Saddle River, 1996.
[CF06] Christopoulos C, Filiatrault A: "Principles of Passive Supple-
mental Damping and Seismic Isolation". ISBN 88-7358-037-
8. IUSSPress, 2006.
[Cho11] Chopra AK: “Dynamics of Structures”. Fourth Edition.
Prentice Hall, 2011.
[CP03] Clough R, Penzien J: “Dynamics of Structures”. Second Edi-
tion (Revised). Computer and Structures, 2003.
(http://www.csiberkeley.com)
[Den85] Den Hartog JP: “Mechanical Vibrations”. Reprint of the fourth
edition (1956). Dover Publications, 1985.
[Hum12] Humar JL: “Dynamics of Structures”. Third Edition. CRC
Press, 2012.
[Inm01] Inman D: “Engineering Vibration”. Prentice Hall, 2001.
[Prz85] Przemieniecki JS: “Theory of Matrix Structural Analysis”. Do-
ver Publications, New York 1985.
[WTY90] Weawer W, Timoshenko SP, Young DH: “Vibration problems
in Engineering”. Fifth Edition. John Wiley & Sons, 1990.
F+T = 0 (2.9)
– cu· – ku – mx·· – mu·· = 0 (2.10)
mu·· + cu· + ku = – mx·· (2.11)
– f k ( t ) – f c ( t ) + F ( t ) = mu··( t ) (2.2)
2 Single Degree of Freedom Systems Page 2-1 2 Single Degree of Freedom Systems Page 2-2
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
δA i = δA a (2.13)
k a sin(ϕ1)
• Thereby, both inertia forces and damping forces must be consid- Fp
ered Fk
( f m + f c + f k )δu = F ( t )δu (2.14) l
a cos(ϕ1)
ϕ1
l
2.1.3 Energy Formulation a
• Kinetic energy T (Work, that an external force needs to pro- sin(ϕ1) ~ ϕ1
vide to move a mass) cos(ϕ1) ~ 1
• Deformation energy U (is determined from the work that an ex- O
ternal force has to provide in order to generate a deformation) ϕ1)
l sin(ϕ
• Potential energy of the external forces V (is determined with Spring force: F k = a ⋅ sin ( ϕ 1 ) ⋅ k ≈ a ⋅ ϕ 1 ⋅ k (2.17)
respect to the potential energy at the position of equilibrium)
··
• Conservation of energy theorem (Conservative systems) Inertia force: Fm = ϕ1 ⋅ l ⋅ m (2.18)
dE Equilibrium
= 0 (2.16)
dt F k ⋅ a ⋅ cos ( ϕ 1 ) + F m ⋅ l – F p ⋅ l ⋅ sin ( ϕ 1 ) = 0 (2.20)
2 Single Degree of Freedom Systems Page 2-3 2 Single Degree of Freedom Systems Page 2-4
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
2 ·· 2
m ⋅ l ⋅ ϕ1 + ( a ⋅ k – m ⋅ g ⋅ l ) ⋅ ϕ1 = 0 (2.21) Spring force: F k ⋅ cos ( ϕ 1 ) ≈ a ⋅ ϕ 1 ⋅ k (2.24)
··
Circular frequency: Inertia force: Fm = ϕ1 ⋅ l ⋅ m (2.25)
K 2
a ⋅k–m⋅g⋅l
2
a ⋅ k- g External force: F p ⋅ sin ( ϕ 1 ) ≈ m ⋅ g ⋅ ϕ 1 (2.26)
ω = -------1 = ------------------------------------- = ------------ – ---
2 l
(2.22)
M1 2
m⋅l m⋅l Virtual displacement:
The system is stable if: δu k = δϕ 1 ⋅ a , δu m = δϕ 1 ⋅ l (2.27)
2
ω > 0: a ⋅k>m⋅g⋅l (2.23) Principle of virtual work:
( F k ⋅ cos ( ϕ 1 ) ) ⋅ δu k + ( F m – ( F p ⋅ sin ( ϕ 1 ) ) ) ⋅ δu m = 0 (2.28)
Principle of virtual work formulation
··
m ( a ⋅ ϕ 1 ⋅ k ) ⋅ δϕ 1 ⋅ a + ( ϕ 1 ⋅ l ⋅ m – m ⋅ g ⋅ ϕ 1 ) ⋅ δϕ 1 ⋅ l = 0 (2.29)
Fm
Fpsin(ϕ1) After cancelling out δϕ 1 the following equation of motion is ob-
tained:
Fkcos(ϕ1) δum 2 ·· 2
k m ⋅ l ⋅ ϕ1 + ( a ⋅ k – m ⋅ g ⋅ l ) ⋅ ϕ1 = 0 (2.30)
ϕ1
a
δϕ1
sin(ϕ1) ~ ϕ1
cos(ϕ1) ~ 1
O
2 Single Degree of Freedom Systems Page 2-5 2 Single Degree of Freedom Systems Page 2-6
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
l Energy conservation:
ϕ1 E tot = E def,k + E kin,m + E pot,p = constant (2.37)
a
1 2 ·2 1 2 2
sin(ϕ1) ~ ϕ1 E = --- ( m ⋅ l ) ⋅ ϕ 1 + --- ( k ⋅ a – m ⋅ g ⋅ l ) ⋅ ϕ 1 = constant (2.38)
2 2
cos(ϕ1) ~ 1
O Derivative of the energy with respect to time:
dE
1 2 1 2
Spring: E def,k = --- ⋅ k ⋅ [ a ⋅ sin ( ϕ 1 ) ] = --- ⋅ k ⋅ ( a ⋅ ϕ 1 ) (2.31) = 0 Derivation rule: ( g • f )' = ( g' • f ) ⋅ f' (2.39)
2 2 dt
2 · ·· 2 ·
Mass:
1 2 1 ·
E kin,m = --- ⋅ m ⋅ v m = --- ⋅ m ⋅ ( ϕ 1 ⋅ l )
2
(2.32) ( m ⋅ l ) ⋅ ϕ1 ⋅ ϕ1 + ( k ⋅ a – m ⋅ g ⋅ l ) ⋅ ϕ1 ⋅ ϕ1 = 0 (2.40)
2 2
·
After cancelling out the velocity ϕ 1 :
E pot,p = – ( m ⋅ g ) ⋅ ( 1 – cos ( ϕ 1 ) ) ⋅ l (2.33)
2 ·· 2
m ⋅ l ⋅ ϕ1 + ( a ⋅ k – m ⋅ g ⋅ l ) ⋅ ϕ1 = 0 (2.41)
by means of a series development, cos ( ϕ 1 ) can be
expressed as: The equation of motion given by Equation (2.41) corresponds to
2 4 Equations (2.21) and (2.30).
ϕ1 ϕ1 k x
2k
cos ( ϕ 1 ) = 1 – ------ + ------ – … + ( – 1 ) ⋅ ------------- + … (2.34)
2! 4! ( 2k )!
2 Single Degree of Freedom Systems Page 2-7 2 Single Degree of Freedom Systems Page 2-8
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
· § 2 ⋅ k – m ⋅ g ⋅ l·
ϕ 1,max = ¨ a------------------------------------
-¸ ⋅ ϕ 1 (2.44) Ground
© m⋅l
2 ¹
· al Tra
ϕ 1,max = ω ⋅ ϕ 1 (2.45) o
din
itu n
ng ctio
dir nsver
ect se
ion
L ire
d
3EI w
• ω is independent of the initial angle ϕ 1 k = … k = 2 -----------
3
-
Bridge in transverse direction H
• the greater the deflection, the greater the maximum velocity. F(t)
Frame with rigid beam
F(t)
3EI w
k = -----------
3
-
H
12EI s
k = 2 -------------
3
-
H
mu·· + ku = F ( t ) (2.46)
2 Single Degree of Freedom Systems Page 2-9 2 Single Degree of Freedom Systems Page 2-10
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• Transformations:
• Principle of virtual work • Using Equations (2.54) and (2.55) the work δA i produced by the
internal forces is:
δA i = δA a (2.49)
L L
L L δA i = ³0 ( EIψ''U ⋅ ψ''δU ) dx = δU U ³ ( EI ( ψ'' ) 2 ) dx (2.57)
δA a = ³0 ( t ⋅ δu ) dx + ³ ( f ⋅ δu ) dx
0
(2.50) 0
2 Single Degree of Freedom Systems Page 2-11 2 Single Degree of Freedom Systems Page 2-12
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
2 k
* ³0 ( EI ( ψ'' ) 2 ) dx
ωn = ------*- = ------------------------------------
L
- (2.60)
m 2
³ mψ dx
0
-> Rayleigh-Quotient
2 Single Degree of Freedom Systems Page 2-13 2 Single Degree of Freedom Systems Page 2-14
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
π 4 L πx 2
k = EI § -------· ³ § cos § -------· · dx
*
(2.63)
© 2L¹ 0 © © 2L¹ ¹
L
§ πx + 2 cos § ------ πx· § πx· ·
- sin ------- L
π 4 ¨ © 2L ¹ © 2L¹ ¸
§ · 1
= EI ------- ⋅ --- ¨ ---------------------------------------------------------------¸
© 2L¹ 2 ¨ π ¸
© ¹
0
4
π EI EI 3EI
= ------ ⋅ -----3- = 3.04 ⋅ -----3- ≈ --------
3
-
32 L L L
πx π 2 πx
ψ = 1 – cos § -------· , ψ'' = § -------· cos § -------· (2.65)
© 2L¹ © 2L¹ © 2L¹
3EI
ω = -----------------------------------------
- (2.64) *
3 • Calculation of the mass m
( 0.23mL + M )L
§ πx· · dx + ψ 2 § x = L
L 2
• Check of the boundary conditions of the deformation figure * § ---· M 1 + ψ 2 ( x = L )M 2
m = ³0 m © 1 – cos © ------
2L¹ ¹
-
© 2¹
(2.66)
πx
ψ ( 0 ) = 0 ? -> ψ ( x ) = 1 – cos §© -------·¹ : ψ ( 0 ) = 0 OK!
2L ( 3π – 8 ) π 2
m = -------------------- mL + § 1 – cos § ---· · ⋅ M 1 + 1 ⋅ M 2
* 2
© ©
(2.67)
π πx 2π 4¹ ¹
ψ' ( 0 ) = 0 ? -> ψ' ( x ) = ------- sin §© -------·¹ : ψ' ( 0 ) = 0 OK!
2L 2L
( 3π – 8 )
m = -------------------- mL + § -------------------· ⋅ M 1 + M 2
* 3–2 2
π 2 πx © 2 ¹
(2.68)
ψ'' ( L ) = 0 ? -> ψ'' ( x ) = §© -------·¹ cos §© -------·¹ : ψ'' ( L ) = 0 OK! 2π
2L 2L
*
m = 0.23mL + 0.086M 1 + M 2 (2.69)
2 Single Degree of Freedom Systems Page 2-15 2 Single Degree of Freedom Systems Page 2-16
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
* 6 13 2
• Calculation of the stiffness k EI = 200000 ⋅ 431.9 ×10 = 8.638 ×10 Nmm (2.75)
4 2
π 4 L πx 2 EI = 8.638 ×10 kNm
k = EI § -------· ³ § cos § -------· · dx (2.76)
*
(2.70)
© 2L¹ 0 © © 2L¹ ¹
By means of Equation (2.73) we obtain:
4
* π EI EI 3EI- 4
k = ------ ⋅ -----3- = 3.04 ⋅ -----3- ≈ -------- (2.71) 8.638 ×10 - = 4.9170
EI = 1.673 ------------------------
32 L L L
3 ω = 1.673 ----------- (2.77)
3 3
ML 10 ⋅ 10
• Calculation of the circular frequency ω
1 4.9170
f = ------ ⋅ ω = ---------------- = 0.783Hz (2.78)
2π 2π
3.04EI
ω = ------------------------------------------------------------------------- (2.72)
3 From Equation (2.74):
( 0.23mL + 0.086M 1 + M 2 )L
3.04EI EI
ω = ------------------------------ = 1.673 ----------- (2.73)
3 3 The first natural frequency of such a dynamic system can be cal-
( 1.086M )L ML
culated using a finite element program (e.g. SAP 2000), and it is
equal to:
The exact first natural circular frequency of a two-mass oscillator
with constant stiffness and mass is: T = 1.2946s , f = 0.772Hz (2.80)
2 Single Degree of Freedom Systems Page 2-17 2 Single Degree of Freedom Systems Page 2-18
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
2.3.3 Damping
M = 10t
• Types of damping
Damping
Internal External
Contact areas
Material
within the structure
2 Single Degree of Freedom Systems Page 2-19 2 Single Degree of Freedom Systems Page 2-20
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• Bearings • Dissipators
2 Single Degree of Freedom Systems Page 2-21 2 Single Degree of Freedom Systems Page 2-22
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
“A structure undergoes free vibrations when it is brought out of ωn = k ⁄ m [rad/s]: Angular velocity (3.7)
its static equilibrium, and can then oscillate without any external
dynamic excitation” ωn
f n = ------ [1/s], [Hz]: Number of revolutions per time (3.8)
2π
3.1 Undamped free vibrations 2π
T n = ------ [s]: Time required per revolution (3.9)
ωn
mu··( t ) + ku ( t ) = 0 (3.1)
• Transformation of the equation of motion
3.1.1 Formulation 1: Amplitude and phase angle
2
u··( t ) + ω n u ( t ) = 0 (3.10)
• Ansatz:
u ( t ) = A cos ( ω n t – φ ) (3.2) • Determination of the unknowns A and φ :
s
Damping constant: c N ⋅ ---- (3.31)
m
c 1 2
λ = – -------- ± -------- c – 4km (3.36)
2m 2m
u(t)/u0 [-]
• Transformation of the equation of motion
0
mu··( t ) + cu· ( t ) + ku ( t ) = 0 (3.39)
c k
u··( t ) + ---- u· ( t ) + ---- u ( t ) = 0 (3.40) -0.5
m m
2
u··( t ) + 2ζω n u· ( t ) + ω n u ( t ) = 0 (3.41)
-1
• Types of vibrations: 0 0.5 1 1.5 2 2.5 3 3.5 4
t/Tn [-]
c
ζ = ------ < 1 : Underdamped free vibrations
c cr
c
ζ = ------ = 1 : Critically damped free vibrations
c cr
c
ζ = ------ > 1 : Overdamped free vibrations
c cr
Underdamped free vibrations ζ < 1 The determination of the unknowns A 1 and A 2 is carried out as
usual by means of the initial conditions for displacement
By substituting: ( u ( 0 ) = u 0 ) and velocity ( u· ( 0 ) = v 0 ) obtaining:
c c c 2 k v 0 + ζω n u 0
ζ = ------ = --------------- = --------------- and ω n = ---- (3.42)
c cr 2 km 2ω n m m A 1 = u 0 , A 2 = --------------------------- (3.51)
ωd
in:
3.2.2 Formulation 1: Amplitude and phase angle
c 2 k
λ = – -------- ± -------- c – 4km = – -------- ± § --------· – ----
c 1 2 c
(3.43)
2m 2m 2m © 2m¹ m Equation (3.50) can be rewritten as “the amplitude and phase
angle”:
it is obtained:
– ζω n t
u ( t ) = Ae cos ( ω d t – φ ) (3.52)
2 2 2 2
λ = – ζω n ± ω n ζ – ω n = – ζω n ± ω n ζ – 1 (3.44)
with
2
λ = – ζω n ± iω n 1 – ζ (3.45) v 0 + ζω n u 0 2 v 0 + ζω n u 0
u 0 + § ---------------------------· , tan φ = ---------------------------
2
A = (3.53)
© ωd ¹ ωd u0
2
ω d = ω n 1 – ζ “damped circular frequency” (3.46)
λ = – ζω n ± iω d (3.47)
The motion is a sinusoidal vibration with
The complete solution of the ODE is: – ζω n t
circular frequency ω d and decreasing amplitude Ae
( – ζω n + iω d )t ( – ζω n – iω d )t
u ( t ) = C1 e + C2 e (3.48)
– ζω n t iω d t –i ωd t
u(t) = e ( C1 e + C2 e ) (3.49)
– ζω n t
u(t) = e ( A 1 cos ( ω d t ) + A 2 sin ( ω d t ) ) (3.50)
Displacement
0.8 5
0.7
0
0.6 2
ωd = ωn 1 – ζ
Tn/Td
-5
0.5
0.4 Tn -10
T d = ------------------
0.3
03 2
1–ζ -15
0.2
-20
0.1 0 1 2 3 4 5 6 7 8 9 10
0 Time (s)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Damping ratio ζ • Amplitude of two consecutive cycles
– ζω t
- The envelope of the vibration is represented by the follow- u0 Ae
n
cos ( ω d t – φ )
ing equation: ----- = --------------------------------------------------------------------------------
– ζω ( t + T )
(3.55)
u1 Ae n d
cos ( ω d ( t + T d ) – φ )
– ζω n t v 0 + ζω n u 0 2
u 0 + § ---------------------------·
2 with
u ( t ) = Ae with A = (3.54)
© ωd ¹
– ζω n ( t + T d ) – ζω n t – ζω n T d
e = e e (3.56)
- Visualization of the solution by means of the Excel file giv-
en on the web page of the course (SD_FV_viscous.xlsx) cos ( ω d ( t + T d ) – φ ) = cos ( ω d t + ω d T d – φ ) = cos ( ω d t – φ ) (3.57)
• Logarithmic decrement δ u0
δ = ---- ln § ------·
1
(3.62)
u0 N © u N¹
δ = ln § -----· = ζω n T d = ------------------ ≅ 2πζ (if ζ small)
2πζ
(3.59)
© u 1¹ 2 • Halving of the amplitude
1–ζ
The damping ratio becomes: u
---- ln § -----0-·
1 1
N © u N¹ ---- ln ( 2 )
δ δ N 1 - ---------
1-
ζ = ------------------------- ≅ ------ (if ζ small) (3.60) ζ = ---------------------- = ------------------ = ------ ≅ (3.63)
2 2 2π
2π 2π 9N 10N
4π + δ
10
Useful formula for quick evaluation
Exact equation
9
Approximation • Watch out: damping ratio vs. damping constant
Logarithmic Decrement δ
8
7 Empty Full
6
5
4
3
2
1 m 1, k 1, c 1 m2>m1, k1, c1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
c1 c1
ζ 1 = -------------------- ζ 2 = -------------------- < ζ 1
Damping ratio ζ 2 k1 m1 2 k1 m2
10
Displacement
5
a) b) 0
– f k ( t ) – f μ = mu··( t ) – f k ( t ) + f μ = mu··( t ) -5
• Solution of b) -15
-20
fμ
u ( t ) = A 1 cos ( ω n t ) + A 2 sin ( ω n t ) + u μ with u μ = ---- (3.64) 0 1 2 3 4 5 6 7 8 9 10
k Time (s)
Figure: f=0.5 Hz , u0=10 , v0 = 50, uf = 1
u· ( t ) = – ω n A 1 sin ( ω n t ) + ω n A 2 cos ( ω n t ) (3.65)
• Calculation example:
by means of the initial conditions u ( 0 ) = u 0 , u· ( 0 ) = v 0 we ob-
- Step 1:
tain the constants:
Initial conditions u ( 0 ) = u 0 , u· ( 0 ) = 0
A1 = u0 – uμ , A2 = v0 ⁄ ωn A1 = u0 – uμ , A2 = 0 (3.66)
• Solution of a): Similar, with – u μ instead of +u μ π
u ( t ) = [ u 0 – u μ ] cos ( ω n t ) + u μ 0 ≤ t < ------ (3.67)
ωn
π
End displacement: u § ------· = [ u 0 – u μ ] ( – 1 ) + u μ = – u 0 + 2u μ
© ω n¹
Displacement
5
the web page of the course (SD_FV_friction.xlsx)
0
• Characteristics of friction damping
- Linear decrease in amplitude by 4u μ at each cycle -5
• Particular solution: u p
2
u··p + 2ζω n u· p + ω n u p = f ( t ) (4.4)
2
u··h + 2ζω n u· h + ω n u h = 0 (4.5)
• Complete solution: u = u p + Cu h
An harmonic excitation can be described either by means of a 2
sine function (Equation 4.1) or by means of a cosine function u·· + 2ζω n u· + ω n u = f o cos ( ωt ) (4.6)
(Equation 4.2):
• Initial conditions
mu·· + cu· + ku = F o sin ( ωt ) (4.1)
u ( 0 ) = u0 , u· ( 0 ) = v 0 (4.7)
mu·· + cu· + ku = F o cos ( ωt ) (4.2)
4 Response to Harmonic Excitation Page 4-1 4 Response to Harmonic Excitation Page 4-2
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
2 fo
u·· + ω n u = f o cos ( ωt ) (4.8) - cos ( ωt )
u = A 1 cos ( ω n t ) + A 2 sin ( ω n t ) + ------------------ (4.16)
2 2
ωn – ω
• Ansatz for particular solution
By means of the initial conditions given in Equation (4.7), the
u p = A o cos ( ωt ) (4.9) constants A 1 and A 2 can be calculated as follows:
2 fo v0
u··p = – A o ω cos ( ωt ) (4.10)
A 1 = u 0 – ------------------
- , A 2 = ------ (4.17)
2
ωn – ω
2 ωn
By substituting (4.9) and (4.10) in (4.8):
2 2 • Denominations:
– A o ω cos ( ωt ) + A o ω n cos ( ωt ) = f o cos ( ωt ) (4.11)
- Homogeneous part of the solution: “transient”
2 2
Ao ( – ω + ωn ) = fo (4.12) - Particular part of the solution: “steady-state”
fo F 1 • Visualization of the solution by means of the Excel file given on
- = -----o ⋅ ------------------------------
A o = ------------------ - (4.13) the web page of the course (SD_HE_cosine_viscous.xlsx)
2
ωn – ω
2 k 1 – ( ω ⁄ ωn )2
fo
- cos ( ωt )
u p = ------------------ (4.14)
2 2 • Harmonic vibration with sine excitation
ωn – ω
fo
• Ansatz for the solution of the homogeneous ODE 2
- sin ( ωt )
u = A 1 cos ( ω n t ) + A 2 sin ( ω n t ) + ------------------
2
(see section on free vibrations) ωn – ω
u h = B 1 cos ( ω n t ) + B 2 sin ( ω n t ) (4.15) By means of the initial conditions given in Equation (3.7), the constants
A 1 and A 2 can be calculated as follows:
v0 fo ( ω ⁄ ωn )
A1 = u0 , A 2 = ------ – -----------------------
-
ωn ω2 – ω2
n
4 Response to Harmonic Excitation Page 4-3 4 Response to Harmonic Excitation Page 4-4
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
4.1.1 Interpretation as a beat • Case 1: Natural frequency SDoF 0.2 Hz, excitation frequency 0.4 Hz
2 500
u·· + ω n u = f o cos ( ωt ) with u ( 0 ) = u· ( 0 ) = 0 (4.18) Total response Envelope
400
300
The solution is:
200
Displacement
fo 100
2
- ⋅ [ cos ( ωt ) – cos ( ω n t ) ]
u ( t ) = ------------------
2
(4.19)
ωn – ω 0
-100
and using the trigonometric identity -200
-300
α–β α+β
cos ( α ) – cos ( β ) = – 2 sin § ------------- t· sin § ------------- t· (4.20) -400
© 2 ¹ © 2 ¹
-500
0 2 4 6 8 10 12 14 16 18 20
one gets the equation Time [s]
2f o ω–ω ω+ω
- ⋅ sin § ----------------n t· sin § ----------------n- t·
• Case 2: Natural frequency SDoF 2.0 Hz, excitation frequency 2.2 Hz
u ( t ) = ------------------ (4.21)
2 2 © 2 ¹ © 2 ¹
ω – ωn 80
Total response Envelope
60
that describes a beat with:
40
f + fn
Displacement
Fundamental vibration: f G = ------------ (4.22) 20
2
0
f – fn -20
Envelope: f U = ------------ (4.23)
2
-40
A beat is always present, but is only evident when the natural fre- -60
quency of the SDoF system and the excitation frequency are
-80
close (see figures on the next page) 0 2 4 6 8 10 12 14 16 18 20
Time [s]
4 Response to Harmonic Excitation Page 4-5 4 Response to Harmonic Excitation Page 4-6
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
2
u··p = 2A o ω n cos ( ω n t ) – A o ω n t sin ( ω n t ) (4.27)
---f- = 2.0250
----------------
fn 2.0000 By substituting Equations (4.25) and (4.27) in (4.24):
2 2
2A o ω n cos ( ω n t ) – A o ω n t sin ( ω n t ) + A o ω n t sin ( ω n t ) = f o cos ( ω n t )
(4.28)
2A o ω n = f o (4.29)
---f- = 2.0125
---------------- fo Fo ωn
fn 2.0000 A o = ---------- = ----- ⋅ ------ (4.30)
2ω n k 2
fo
u p = ---------- t sin ( ω n t ) (4.31)
2ω n
f 2.0000 • Ansatz for the solution of the homogeneous ODE
---- = ----------------
fn 2.0000 (see section on free vibrations)
Resonance!
u h = B 1 cos ( ω n t ) + B 2 sin ( ω n t ) (4.32)
4 Response to Harmonic Excitation Page 4-7 4 Response to Harmonic Excitation Page 4-8
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
4 Response to Harmonic Excitation Page 4-9 4 Response to Harmonic Excitation Page 4-10
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• Ansatz for the solution of the homogeneous ODE • Example 1: fn = 1Hz, f = 0.2Hz, ζ = 5%, fo = 1000, u0 = 0, v0 = fo/ωn
(see Section 3.2 on damped free vibrations) 60
Steady-state response
– ζω n t
uh = e ( B 1 cos ( ω d t ) + B 2 sin ( ω d t ) ) (4.45) 40 Total response
with:
Displacement
20
2 0
ω d = ω n 1 – ζ “damped circular frequency” (4.46)
-20
• Complete solution of the ODE:
– ζω n t -40
u = e ( A 1 cos ( ω d t ) + A 2 sin ( ω d t ) ) + A 3 cos ( ωt ) + A 4 sin ( ωt )
-60
(4.47) 0 2 4 6 8 10 12 14 16 18 20
Time [s]
By means of the initial conditions of Equation (4.7), the con-
stants A 1 and A 2 can be calculated. The calculation is labo- • Example 2: Like 1 but with F(t) = Fosin(ωt) instead of Focos(ωt)
rious and should be best carried out with a mathematics pro- 50
gram (e.g. Maple). 40
Steady-state response
Total response
30
• Denominations:
20
Displacement
- Homogeneous part of the solution: “transient” 10
0
- Particular part of the solution: “steady-state”
-10
• Visualization of the solution by means of the Excel file given on -20
20
the web page of the course (SD_HE_cosine_viscous.xlsx) -30
-40
-50
0 2 4 6 8 10 12 14 16 18 20
Time [s]
4 Response to Harmonic Excitation Page 4-11 4 Response to Harmonic Excitation Page 4-12
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
4.2.1 Resonant excitation (ω = ωn) - The amplitude is limited, i.e. the maximum displacement of
the SDoF system is:
By substituting ω = ω n in Equation (4.44) the constants A 3 and
A 4 becomes: fo Fo u st
u max = ------------2- = --------- = ------ (4.53)
2ζω n 2ζk 2ζ
fo
A 3 = 0 , A 4 = ------------2- (4.48)
2ζω n where u st = F o ⁄ k is the static displacement.
i.e. if damping is present, the resonant excitation is not a special
case any more, and the complete solution of the differential 2
equation is: - For small damping ratios ( ζ ≤ 0.2 ) ω d ≈ ω n and 1–ζ ≈1
hence Equations (4.51) becomes:
– ζω n t fo
u = e ( A 1 cos ( ω d t ) + A 2 sin ( ω d t ) ) + ------------2- sin ( ω n t ) (4.49) fo
u = ------------2- ( 1 – e
– ζω n t
) sin ( ω n t ) = u max ( 1 – e
– ζω n t
) sin ( ω n t ) (4.54)
2ζω n 2ζω n
fo fo – ζω n t
A = u max ( 1 – e ) (4.55)
A1 = 0 , A 2 = – -------------------------------- = – -------------------- (4.50)
2 2 2ζω n ω d
2ζω n 1 – ζ and the magnitude of the amplitude at each maxima j is
fo § sin ( ω d t ) – ζω n t· uj
----------
– ζω t
- = ( 1 – e n j ) sin ( ω n t j ) (4.56)
u = ------------2- ¨ sin ( ω n t ) – --------------------- e ¸ (4.51) u max
2ζω n © 1–ζ
2 ¹
Maxima occur when sin ( ω n t ) = – 1 , d.h. when
- After a certain time, the homogeneous part of the solution
Tn
subsides and what remains is a sinusoidal oscillation of the t j = ( 4j – 1 ) ⋅ ------ , j = 1…∞ (4.57)
4
amplitude:
Tn π
fo uj – ζω n ( 4j – 1 ) ⋅ ----- – ζ ( 4j – 1 ) ⋅ ---
4 2
A = ------------2- (4.52) ----------
- = 1–e = 1–e (4.58)
2ζω n u max
4 Response to Harmonic Excitation Page 4-13 4 Response to Harmonic Excitation Page 4-14
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
ζ = 0.01
40 40
abs(uj) / ust
umax / ust
30 30
0.02
20 20
0.05
10 10
0.10
0 0
0 0.05 0.1 0.15 0.2 0 5 10 15 20 25 30 35 40 45 0.20 50
Damping ratio ζ [-] Cycle
ζ = 0.01
abs(uj) / umax
0.6
0.4
0.2
0
0 5 10 15 20 25 30 35 40 45 50
Cycle
4 Response to Harmonic Excitation Page 4-15 4 Response to Harmonic Excitation Page 4-16
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
5 Transfer Functions The maximum dynamic amplitude u max given by Equation (5.5)
can be transformed to:
5.1 Force excitation 2
ωn – ω
2 2
2ζω n ω
§ · § ·2
u max = ¨ f o ---------------------------------------------------------
-¸ + ¨ f o ---------------------------------------------------------
-¸
The steady-state displacement of a system due to harmonic ex- © ( ω 2 – ω 2 ) 2 + ( 2ζω ω ) 2¹ © ( ω 2 – ω 2 ) 2 + ( 2ζω ω ) 2¹
n n n n
citation is (see Section 4.2 on harmonic excitation): (5.7)
u p = a 1 cos ( ωt ) + a 2 sin ( ωt ) (5.1)
2 2 2 2
( ω n – ω ) + ( 2ζω n ω )
with u max = f o ----------------------------------------------------------------
- (5.8)
2 2 2 2 2
[ ( ω n – ω ) + ( 2ζω n ω ) ]
2 2
ωn – ω 2ζω n ω
- , a 2 = f o ---------------------------------------------------------
a 1 = f o --------------------------------------------------------- - (5.2) 1
2 2 2 2 2 2 2 2 u max = f o -------------------------------------------------------------- (5.9)
( ω n – ω ) + ( 2ζω n ω ) ( ω n – ω ) + ( 2ζω n ω ) 2 2 2 2
( ω n – ω ) + ( 2ζω n ω )
By means of the trigonometric identity
fo 1
b
u max = -----2- ---------------------------------------------------------------------------------- (5.10)
2 2
a cos ( α ) + b sin ( α ) = a + b ⋅ cos ( α – φ ) where tan φ = --- (5.3) ω n [ 1 – ( ω ⁄ ω ) 2 ] 2 + [ 2ζ ( ω ⁄ ω ) ] 2
a n n
Equation (5.1) can be transformed as follows: Introducing the maximum static amplitude u o = F o ⁄ k = f o ⁄ ω 2n the
dynamic amplification factor V ( ω ) can be defined as:
u p = u max cos ( ωt – φ ) (5.4)
u max 1
It is a cosine vibration with the maximum dynamic amplitude V ( ω ) = ----------- = ----------------------------------------------------------------------------------- (5.11)
uo 2 2 2
u max : [ 1 – ( ω ⁄ ω n ) ] + [ 2ζ ( ω ⁄ ω n ) ]
2 2 The maximum amplification factor V ( ω ) occurs when its deriva-
u max = a1 + a2 (5.5)
tive, given by Equation (5.12), is equal to zero.
and the phase angle φ obtained from:
a2
tan φ = ----- (5.6)
a1
2 2
2ωω n [ ω – ω n ( 1 – 2ζ ) ]
2 2 5.1.1 Comments on the amplification factor V
dV
= --------------------------------------------------------------------------------
- (5.12)
dω 4 2 2 2 4 (3 ⁄ 2)
[ ω – 2 ( 1 – 2ζ )ω ω n + ω n ] 1
V ( ω ) = ----------------------------------------------------------------------------------- (5.20)
2 2 2
dV 2
[ 1 – ( ω ⁄ ω n ) ] + [ 2ζ ( ω ⁄ ω n ) ]
= 0 when: ω = 0 , ω = ± ω n 1 – 2ζ (5.13)
dω
The phase angle has the following interesting property: • ( ω ⁄ ω n ) ≈ 1 : (ζ very important)
2 1
dφ 2ζ [ 1 + ( ω ⁄ ω n ) ] • V ( ω ) ≈ ------
= ---------------------------------------------------------------------------------------------------
- (5.18) 2ζ
d ( ω ⁄ ωn ) 2
1 – 2 ( ω ⁄ ω n ) + ( ω ⁄ ω n ) + 4ζ ( ω ⁄ ω n )
4 2 2
• u max ≈ u o ⁄ ( 2ζ ) = F o ⁄ ( cω n ) : Damping controls the behaviour
dφ 1 1 180 • φ ≈ 90 : zero displacement when excitation force is maximum
at ω ⁄ ω n = 1 we have: = --- (= --- ⋅ --------- when φ in deg)
d ( ω ⁄ ωn ) ζ ζ π
(5.19)
8
7
F o cos ( ωt )
ζ = 0.05
u st = -------------------------- (5.21)
6 k
5
4
and its maximum is:
0.10
3
2
Fo
0.20
0.50
u o = ----- (5.22)
1 k
0.70
0
0 0.5 1 1.5 2 The steady-state dynamic response of the system is:
ω/ωn
, (5.24)
uo uo
90
In the next plots the time histories of u st ⁄ u o and u p ⁄ u o are
represented and compared.
Frequency of SDoF System fn = 1Hz (ωn = 6.28rad/s), Damping ζ = 0.1 5.1.2 Steady-state displacement quantities
5 ω ⁄ ω n = 0.9
4 V • Displacement: Corresponds to Equation (5.4)
3
Δt ω = 0.9 ⋅ 6.28
2
= 5.65 rad/s up
1 ------------ = V ( ω ) cos ( ωt – φ ) (5.25)
u / uo
0 V = 3.82 Fo ⁄ k
-1
-2 φ = 43.45°
-3 Excitation = 0.76 rad • Velocity: Obtained by derivating Equation (5.25)
-4 Steady-state response
φ 0.76
-5 Δt = ---- = ----------
ω 5.65 u· p
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 ------------ = – V ( ω )ω sin ( ωt – φ ) (5.26)
Time [s] = 0.14 s Fo ⁄ k
V ω ⁄ ω n = 1.0
5 u· p ω
4 ------------------------ = – V ( ω ) ------ sin ( ωt – φ ) (5.27)
3 ω = 1.0 ⋅ 6.28 ( F o ⁄ k )ω n ωn
2 Δt = 6.28 rad/s
1
u / uo
0 V = 5.00 u· p ω
-1 - = – V v ( ω ) sin ( ωt – φ ) with V v ( ω ) = ------ V ( ω ) (5.28)
---------------------
-2 φ = 90.00° F o ⁄ km ω n
-3 Excitation
= 1.57 rad
-4 Steady-state response φ 1.57
-5 Δt = ---- = ---------- • Acceleration: Obtained by derivating Equation (5.26)
ω 6.28
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Time [s]
= 0.25 s
u··p 2
------------ = – V ( ω )ω cos ( ωt – φ ) (5.29)
5 ω ⁄ ω n = 1.1 Fo ⁄ k
4 V
3 ω = 1.1 ⋅ 6.28
2 Δt = 6.91 rad/s u··p ω
2
1 ------------------------ = – V ( ω ) ------ cos ( ωt – φ ) (5.30)
u / uo
2 2
0 V = 3.28 ( F o ⁄ k )ω n ωn
-1
φ = 133.66°
-2 = 2.33 rad
-3 Excitation u··p ω
2
-4 φ 2.33
Δt = ---- = ----------
-------------- = – V a ( ω ) cos ( ωt – φ ) with V a ( ω ) = ------ V ( ω ) (5.31)
Fo ⁄ m
Steady-state response
2
-5 ω 6.91 ωn
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 = 0.34 s
Time [s]
8
• Half-power bandwidth
7
ζ = 0.05
6 10
5
V(Resonance)
2 9
4 ω = ω n 1 – 2ζ
Amp
cation factor Vv(ω)
7 2
6
ζ = 0.05 ζ = 0.05
1 2ζ
5 ω = ωn
0
ωa
Amplificati
4
0 10
0.10 ωb
3 0 0.5 1 1.5 2
1
2 0.20 V = ------ ω/ωn
1
0.50 2ζ
0
0.70 Condition:
0 0.5 1 1.5 2
ω/ωn 2
V ( ω ⁄ ω n = 1 – 2ζ ) 1 1
10
0.00 Resonant V ( ω ) = ----------------------------------------------------- = ------- ⋅ ---------------------- (5.32)
9 2 2 2ζ 1 – ζ
8
0.01
acceleration
Amplification factor Va(ω)
1 1
1- ----------------------
7
ζ = 0.05
---------------------------------------------------------------------------------- = ------ ⋅ (5.33)
6 2 2 2 2 2ζ 1 – ζ
5 ωn [ 1 – ( ω ⁄ ω n ) ] + [ 2ζ ( ω ⁄ ω n ) ]
4 ω = ----------------------
ω ·4
§ ----- ω 2
- – 2 ( 1 – 2ζ 2 ) § ------· + 1 – 8ζ 2 ( 1 – ζ 2 ) = 0
0.10 2
3 1 – 2ζ (5.34)
2 0.20
© ω n¹ © ω n¹
0.50
1 1
V = ---------------------- ω-· 2
0
0.70
2ζ 1 – ζ § ----- 2
= 1 – 2ζ ± 2ζ 1 – ζ
2
(5.35)
0 0.5 1 1.5 2 © ω n¹
ω/ωn
2
For small damping, the terms featuring ζ can be neglected: 5.2 Force transmission (vibration isolation)
ω-
----- ≈ 1 ± 2ζ ≈ 1 ± ζ (5.36)
ωn
This yield the solution for the half-power bandwidth: The mass-spring-damper system,
shown here on the right, is excited
ωb – ωa by the harmonic force
2ζ = ------------------- (5.37)
ωn
F ( t ) = F o cos ( ωt )
• Remarks on the frequency response curve
What is the reaction force F T ( t ) ,
• The natural frequency of the system can be derived from the res-
which is introduced in the founda-
onant response. However, it is sometimes problematic to build the
whole frequency response curve because at resonance the sys- tion?
tem could be damaged. For this reason it is often better to deter-
mine the properties of a system based on vibration decay tests
(see section on free vibration)
• The natural frequency ω n can be estimated by varying the Excita- The reaction force F T ( t ) results from the sum of the spring force
F s and the damper’s force F c
tion until a 90° phase shift in the response occurs.
• Damping can be calculated by means of Equation (5.15) as: F T ( t ) = F s ( t ) + F c ( t ) = ku ( t ) + cu· ( t ) (5.38)
with the trigonometric identity from Equation (5.3): • Representation of the transmissibility TR
Fo 2 2 2 10
F T ( t ) = ----- V ( ω ) [ k + c ω cos ( ωt – φ ) ] (5.41) 0.00
k
9
0.01
and by substituting the identity c = ( 2ζk ) ⁄ ω n : 8
ζ = 0.05
Trrasmissibility TR
7
ω 2
F T ( t ) = F o V ( ω ) 1 + § 2ζ ------· cos ( ωt – φ ) (5.42)
© ω n¹ 6
5
the maximum reaction force becomes: 0.10
4
F T,max
--------------- = TR ( ω ) (5.43) 3
0.20
Fo
2
0.50
where the quantity TR ( ω ) is called Transmissibility and it is 1
0.70 2
equal to: 0
0 0.5 1 1.5 2 2.5 3 3.5 4
ω 2 ω/ωn
TR ( ω ) = V ( ω ) 1 + § 2ζ ------· (5.44)
© ω n¹
- When ω ⁄ ω n > 2 then TR < 1 : Vibration isolation
2
1 + [ 2ζ ( ω ⁄ ω n ) ]
= ------------------------------------------------------------------------------ - When ω ⁄ ω n > 2 damping has a stiffening effect
2 2 2
[ 1 – ( ω ⁄ ω n ) ] + [ 2ζ ( ω ⁄ ω n ) ]
- High tuning (sub-critical excitation)
Special case: - Low tuning (super-critical excitation):
2 Pay attention to the starting phase!
ω
TR § ------ = 1· = ----------------------
1 + 4ζ
© ωn ¹
(5.45)
2ζ
5.3 Base excitation (vibration isolation) The right hand side of the ODE (5.48) can be interpreted as an
external excitation force F ( t ) = ky + cy· :
5.3.1 Displacement excitation
F ( t ) = ky go cos ( ωt ) – cy go ω sin ( ωt ) (5.49)
ω
= ky go cos ( ωt ) – 2ζ ------ sin ( ωt )
ωn
ω 2
= ky go 1 + § 2ζ ------· cos ( ωt + φ )
© ω n¹
ω 2
F o = ky go 1 + § 2ζ ------· (5.50)
© ω n¹
The mass-spring-damper system, shown here above is excited
According to Equations (5.10) and (5.11) the maximum displace-
by the harmonic vertical ground displacement
ment of the system due to such a force is equal to:
y g ( t ) = y go cos ( ωt ) (5.46)
Fo ω 2
u max = ----- V ( ω ) = y go 1 + § 2ζ ------· V ( ω ) (5.51)
What is the absolute vertical displacement u ( t ) of the system? k © ω n¹
The differential equation of the system is: By substituting Equation (5.44) we obtain:
mu·· + cu· + ku = ky + cy· (5.48) where again TR ( ω ) is the transmissibility given by Equation
(5.44).
5.3.2 Acceleration excitation The steady-state relative deformation u rel of the system due to
Pay attention: the harmonic ground acceleration y··g is given by Equation (5.1):
This base excitation, like u rel = a 1 cos ( ωt ) + a 2 sin ( ωt ) (5.59)
the excitation discussed
in the previous Section with the constants a 1 and a 2 according to Equation (5.2), and
5.3.1, is an harmonic ex- with:
citation and not an arbi-
trary excitation like e.g. Fo – my··go
an earthquake (see Sec- f o = ----- = ---------------- = – y··go (5.60)
m m
tion 7).
By double derivation of Equation (5.59), the relative acceleration
The mass-spring-damper system, shown above here, is excited u··rel can be calculated as:
by the harmonic vertical ground acceleration.
2 2
u··rel = – a 1 ω cos ( ωt ) – a 2 ω sin ( ωt ) (5.61)
y··g ( t ) = y··go cos ( ωt ) (5.53)
The desired absolute acceleration is:
What is the absolute vertical acceleration u··( t ) of the system?
2 2
u·· = u··rel + y··g = – a 1 ω cos ( ωt ) – a 2 ω sin ( ωt ) + y··go cos ( ωt ) (5.62)
The differential equation of the system is:
By substituting the constants a 1 , a 2 and f o given by Equations
mu·· + c ( u· – y· ) + k ( u – y ) = 0 (5.54)
(5.2) and (5.60), and after a long but simple rearrangement, the
after rearrangement: equations for the maximum absolute vertical acceleration of the
system is obtained as:
mu·· + c ( u· – y· ) + k ( u – y ) – my·· = – my·· (5.55)
m ( u·· – y··) + c ( u· – y· ) + k ( u – y ) = – my·· (5.56) u··max
----------- = TR ( ω ) (5.63)
y··go
mu··rel + cu· rel + ku rel = – my··g (5.57)
where again TR ( ω ) is the transmissibility given by Equation
mu··rel + cu· rel + ku rel = – my··go cos ( ωt ) (5.58) (5.44).
u rel, max
- = V(ω)
---------------------- (5.65)
2
( y··go ⁄ ω n )
• Natural frequency SDoF system: f n = 0.5Hz
• Excitation frequency: f 0 = 2.0Hz
• Case 1: Initial situation with ζ = 2% • Case 2: Increase of the damping rate from ζ = 2% to ζ = 20%
2 2
Excitation freq. [Hz]
f=fn=0.5Hz f=fn=0.5Hz
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Excitation amplitude [m/s2]
5 5
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
10 10
Excitation [m/s2]
Excitation [m/s2]
5 5
0 0
-5 -5
-10 -10
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
20 20
Aabs SDoF [m/s2]
0 0
-10 -10
-20 -20
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Time [s] Time [s]
• Case 3: Reduction of starting time from ta = 80s to ta = 20s (ζ = 2%) • Case 4: Change the start function for the amplitude (ζ = 2%)
2 2
Excitation freq. [Hz]
f=fn=0.5Hz f=fn=0.5Hz
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Excitation amplitude [m/s2]
5 5
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
10 10
Excitation [m/s2]
Excitation [m/s2]
5 5
0 0
-5 -5
-10 -10
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
20 20
Aabs SDoF [m/s2]
0 0
-10 -10
-20 -20
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Time [s] Time [s]
ω0 t u max
Ω ( t ) = ------- ⋅ t : ω ( t ) = ω 0 ⋅ --- ( 0 ≤ t ≤ t a ) (5.68) • Force excitation: ----------- = V ( ω )
2t a ta uo
• Parabolic variation of the excitation circular frequency
F T,max
ω0 2 t 2 • Force transmission: --------------- = TR ( ω )
Ω ( t ) = ------2- ⋅ t : ω ( t ) = ω 0 ⋅ § ---· ( 0 ≤ t ≤ t a ) (5.69) Fo
© t a¹
3t a
u max
• Sinusoidal variation of the excitation circular frequency • Displacement excitation: ----------- = TR ( ω )
y go
2ω 0 t a π t π t
Ω ( t ) = – -------------- cos § --- ⋅ ---· : ω ( t ) = ω 0 ⋅ sin § --- ⋅ ---· ( 0 ≤ t ≤ t a )
πt © 2 t a¹ © 2 t a¹ u··max
• Acceleration excitation ----------- = TR ( ω )
(5.70) y··go
• Double-sinusoidal variation of the excitation circular frequency u rel, max
- = V(ω)
----------------------
·· 2
sin § π ⋅ ---· t a
t ( y go ⁄ ω n )
ω0 © t a¹
Ω ( t ) = ------ 1 – ----------------------------- : ω ( t ) = ω 0 1 – cos § π ⋅ ---·
t
©
(5.71)
2 πt t a¹ • For further cases check the literature.
2.5
T0 0 for n ≠ j
³0 cos ( nω 0 t ) cos ( jω 0 t ) dt = ® (6.5)
2.0 ¯ T 0 ⁄ 2 for n = j
1.5 T0
³0 cos ( nω 0 t ) sin ( jω 0 t ) dt = 0 (6.6)
F
1.0
F ( t + nT 0 ) = F ( t ) for n = – ∞, …, – 1, 0, 1, …, ∞ (6.1) ∞ T0 T0
+ ¦ ³
n=1 0
a n cos ( nω 0 t ) cos ( jω 0 t ) dt + ³ b n sin ( nω 0 t ) cos ( jω 0 t ) dt
0
6.1.2 Half-sine 2
2A t p πt 4Aτ cos ( nπτ ) -
a n = ------- ⋅ ³ sin § -----· cos ( nω 0 t ) dt = ------------------------------------ (6.23)
T0 0 © tp ¹ 2 2
A series of half-sine functions is a good model for the force that π ( 1 – 4n τ )
is generated by a person jumping.
2A tp πt 4Aτ sin ( nπτ ) cos ( nπτ )-
4.0 b n = ------- ⋅ ³ sin § -----· sin ( nω 0 t ) dt = --------------------------------------------------------
T0 0 © tp ¹ 2 2
3.5
T0
Half-sine excitation
π ( 1 – 4n τ )
tp (6.24)
3.0
The approximation of the half-sine model for T 0 = 0.5s and
Force F(t) [kN]
2.5
t p = 0.16s by means of 6 Fourier terms is as follows:
2.0
4.0
Static term (n=0)
1.5 T0 First harmonic (n=1)
F
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.0
Time (s)
0.0
F
§ πt-· for 0 ≤ t < t
° A sin © ----
tp ¹ p -1.0
F(t) = ® (6.21)
° 0 for t p ≤ t < T 0 -2.0
¯ 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s)
The Fourier coefficients can be calculated at the best using a
mathematics program: • Note
The static term a 0 = 2Aτ ⁄ π = G corresponds to the weight of
A tp πt tp
a 0 = ------ ⋅ ³ sin § -----· dt = ----------
2Aτ
with τ = ------ (6.22) the person jumping.
T0 0 © tp ¹ π T0
2.5
2.0
• Contact time: - When the excitation frequency f 0 is twice as large as the
t p = 0.16s
1.5
natural frequency f n of the beam, the magnification factor
F
1.0
• Person’s weight:
0.5
G = 0.70kN V is small.
0.0 • Amplitude: - Taking into account the higher harmonics can be impor-
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
A = 3.44kN
Time (s) tant!
Displacement [m]
0.0020 Static term (n=0)
First harmonic (n=1) 0.0010
0.0015 Second harmonic (n=2)
Third harmonic (n=3)
D
0 0000
0.0000
-0.0005
-0.0005
-0.0010
-0.0015 -0.0010
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s) Time (s)
Displacement [m]
0.0200
0.0050
0.0100
0.0000 0.0000
-0.0100
-0.0050
D
D
0 0200
-0.0200
Static term (n=0)
First harmonic (n=1)
-0.0300
Second harmonic (n=2) -0.0100
Third harmonic (n=3)
-0.0400 Total (6 harmonics)
-0.0500 -0.0150
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s) Time (s)
0.0020 • The damped case can be solved in the exact same way. On the
0.0010
web page of the course there is an Excel file to illustrate this exci-
tation.
0.0000
• The maximum displacement of an undamped SDoF System under
-0.0010
a step force is twice the static deflection u st = F 0 ⁄ k .
D
-0.0020
• The deflection at the time t = ∞ of a damped SDoF System under
-0.0030
a step force is equal to the static deflection u st = F 0 ⁄ k .
-0.0040
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s)
• Step force: Tn=2s, Fo/k=2, ζ=0 6.2.2 Rectangular pulse force excitation
4.5
Dynamic response
4 Excitation
3.5
3
Displacement
2.5
1.5
2.5 k
2 F0
u· ( t 1 ) = ----- ω n sin ( ω n t 1 ) (6.31)
1.5 k
• Short duration of excitation ( t 1 ⁄ T n is small) The equation of an undamped free vibration is:
• Rectangular pulse: Tn=2s, t1=0.5s (t1/Tn=0.25), Fo/k=2, ζ=0% • Rectangular pulse: Tn=2s, t1=2s (t1/Tn=1.00), Fo/k=2, ζ=0%
4 4.5
Dynamic response Dynamic response
3 Excitation 4 Excitation
3.5
2
3
Displacement
Displacement
1 2.5
0 2
-1 1.5
1
-2
0.5
-3 0
-4 -0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
• Rectangular pulse: Tn=2s, t1=1s (t1/Tn=0.50), Fo/k=2, ζ=0% • Rectangular pulse: Tn=2s, t1=3s (t1/Tn=1.50), Fo/k=2, ζ=0%
5 5
Dynamic response Dynamic response
4 Excitation 4 Excitation
3 3
2 2
Displacement
Displacement
1 1
0 0
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
• Rectangular pulse: Tn=2s, t1=3.5s (t1/Tn=1.75), Fo/k=2, ζ=0% • Short rectangular pulse: Tn=2s, t1=0.05s, Fo/k=2, ζ=0%
5 2.5
Dynamic response Dynamic response
4 Excitation Excitation
2
3
2 1.5
Displacement
Displacement
1
1
0
-1 0.5
-2
0
-3
-4 -0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
• Rectangular pulse: Tn=2s, t1=4s (t1/Tn=2.00), Fo/k=2, ζ=0% • Short rectangular pulse: Tn=2s, t1=0.05s, Fo/k=2, ζ=5%
4.5 2.5
Dynamic response Dynamic response
4 Excitation Excitation
2
3.5
3
1.5
Displacement
Displacement
2.5
2 1
1.5
0.5
1
0.5
0
0
-0.5 -0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
E c I o = 171.9kNm 2
- Action
- Simplified system
Cross-Section
Boundary conditions: L 2 EI
k* = ³0 ( EI ⋅ ( ψ'' ) ⋅ dx ) = 104.37 ⋅ -----3-
L
(6.45)
ψ ( 0 ) = 0 , ψ ( L ) = 0 , ψ' ( 0 ) = 0 , ψ'' ( L ) = 0 (6.41)
L 2 =3.45m
By means of the mathematics program “Maple” Equation (6.40) P* = ³L =1.55m ( p ⋅ ψ ⋅ dx )
1
= 0.888 ⋅ P tot (6.46)
can be solved for the boundary conditions (6.41) and we get:
For this example, the modal properties characterizing the equiv-
[ sin ( βL ) + sin h ( βL ) ] ⋅ [ cos ( βx ) – cos h ( βx ) ]
1.508 ⋅ ψ = sin ( βx ) – sin h ( βx ) + ------------------------------------------------------------------------------------------------------------------ alent modal SDoF system are:
– cos ( βL ) – cos h ( βL )
(6.42)
m * = 0.439 ⋅ 2.06 ⋅ 5 = 4.52t (6.47)
with
52184-
k * = 104.37 ⋅ -------------- = 43571kN/m (6.48)
βL = 3.927 (6.43) 53
• Modelling option 2 1
0.8
0.6
Within a simplified modelling approach, it is assumed that the 0.4
slab remains elastic during loading. Sought is the maximum de- 0.2
ψ [-]
0
flection of the slab due to the explosion. -0.2
-0.4
-0.6
- Simplified system -0.8
-1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/L [-]
L 2 =8.45m
P* = ³L =6.55m ( p ⋅ ψ ⋅ dx )
1
= 0.941 ⋅ P tot (6.59)
The maximum elastic deformation of the SDoF system can be As a third option, the slab is modelled using the commercial finite
calculated using the modal pulse as follows: element software SAP 2000.
–3 - Numerical Model
I * = 0.5 ⋅ P * ⋅ t 0 = 0.5 ⋅ 180672 ⋅ 0.3 ×10 = 27.1kNs (6.65)
I* 27.1
v 0 = ------*- = ---------- = 2.63m/s (6.66)
m 10.3
T 1 = 0.100s (6.69)
0.06
c deformation [m]
0.04
0.02
-0.02
Elastic
-0.04
-0.06
-0.08
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Time [s]
- - - 0.100 0.064
7 Seismic Excitation From the previous figure it can be clearly seen that the time-his-
tory of an earthquake ground acceleration can not be described
by a simple mathematical formula. Time-histories are therefore
7.1 Introduction
usually expressed as sequence of discrete sample values and
The equation of motion for a base point excitation through an ac- hence Equation (7.1) must therefore be solved numerically.
celeration time-history u··g ( t ) can be derived from the equilibrium 2
of forces (see Section 2.1.1) as:
1
xg [m/s2]
mu·· + cu· + f s (u,t) = – mu··g (7.1) 0
Δüg
where u·· , u· and u are motion quantities relative to the base point
ü
-1
of the SDoF system, while f s (u,t) is the spring force of the system -2
that can be linear or nonlinear in function of time and space. The 9.5 9.6
2
time-history of the motion quantities u·· , u· and u for a given SDoF Zeit [s]
system are calculated by solving Equation (7.1). 1
x [m/s2]
Δü
4 0
Ground acceleration
ü
2 -1
üxg [m/s2]
0 -2
t t+Δt
9.5 9.6
-2
Zeit [s]
Time
-4
0 10 20 30 40 The sample values of the ground acceleration u··g ( t ) are known
Zeit [s]
Time [s] from beginning to end of the earthquake at each increment of
4
Response of a T=0.5s SDoF system time Δt (“time step”). The solution strategy assumes that the mo-
2 tion quantities of the SDoF system at the time t are known, and
x [m/s ]
2
-2
known initial conditions) and are carried out time step after time
-4
step until the entire time-history of the motion quantities is com-
0 10 20 30 40
Zeit [s]
Time [s] puted, like e.g. the acceleration shown in the figure on page 7-1.
7.2 Time-history analysis of linear SDoF systems 7.2.1 Newmark’s method (see [New59])
• Incremental formulation of the equation of motion
mΔu·· + cΔu· + kΔu = – mΔu··g (7.4)
t + Δt t t + Δt · t t + Δt·· t
u = u + Δu , u = u· + Δu· , u = u·· + Δu·· (7.5)
• Assumption of the acceleration variation over the time step:
2
Δu·· ( τ – t )
u ( τ ) = u + u· ( τ – t ) + § u·· + -------· -----------------
t t t
(7.9)
© 2¹ 2 Linear Acceleration:
Δt-
---- ≤ 0.551
1 1 T
The increments of acceleration, velocity and displacement dur- β = --- , γ = ---
ing the time step are: 6 2
• Solution of the differential equation: Option 2 7.2.2 Implementation of Newmark’s integration scheme
within the Excel-Table “SDOF_TH.xls”
Equation (7.14) can be transformed to:
t· t·· Equation (7.15) is implemented in the Excel-Table as follows:
Δu u u
Δu·· = -----------2 – --------- – ------ (7.17)
βΔt βΔt 2β
2
( m + cγΔt + kβΔt )
2 Δu·· – mΔu··g t··
u Δt § tu· Δt + tu··Δt
--------·
–k © 2 ¹
®
¯
= –c
®
¯
°
°
°
®
°
°
°
¯
®
¯
°
°
®
°
°
¯
and substituting Equation (7.17) into (7.13) we obtain: meq da ΔF ( t ) dv
dd
t·
γΔu γ u γ t
Δu· = --------- – ------- + Δt § 1 – ------· u·· (7.18) • In the columns C to E the so-called “predictors” dd, dv and da
βΔt β © 2β¹ are computed first:
Substituting Equations (7.17) and (7.18) into Equation (7.4) 2
t· t··Δt
gives Equation (7.19), which can be solved for the only remain- dd = u Δt + u -------- (“delta-displacement”)
2
ing variable Δu :
t
γc γc t γ dv = u··Δt (“delta-velocity”)
§ k + -----------
m
+ ---------· Δu = – mΔu··g + § --------- + -----· u· + § ------ – Δt § 1 – ------· c· u··
m m t
© 2 βΔt¹ © βΔt β ¹ © 2β © 2β ¹ ¹
βΔt – mΔu··g – c ⋅ dv – k ⋅ dd
(7.19) da = ------------------------------------------------------ = Δu·· (“delta-acceleration”)
meq
or in compact form:
• Afterwards, in the columns F to H the ground motion quantities
kΔu = Δp (7.20) at the time step t + Δt are computed by means of so-called
Substituting Δu into Equations (7.18) and (7.17) gives the incre- “correctors:
ments of the velocity Δu· and of the acceleration Δx··. In conjunc- t + Δt·· t
u = u·· + da
tion with Equation (7.5), these increments yield the dynamic re-
sponse of the SDoF system at the end of the time step t + Δt . t + Δt · t
u = u· + dv + ( da ⋅ γ ⋅ Δt )
°
°
®
°
°
¯
For linear systems we have: Δu
• m , c and k are constant throughout the whole time-history.
t + Δt t 2
• m̃ in Equation (7.15), as well as k in Equation (7.20), are also u = u + dd + ( da ⋅ β ⋅ Δt )
°
°
°
®
°
°
°
¯
constant and have to be computed only once. Δu
• Finally, in column I the absolute acceleration u··abs at the time 7.2.3 Alternative formulation of Newmark’s Method.
step t + Δt is computed as follows:
The formulation of the Newmark’s Method presented in Section
t + Δt·· t + Δt·· t + Δt·· 7.2.1 corresponds to an incremental formulation. It is possible to
uabs = u+ ug
rearrange the methodology to obtain a total formulation.
Observations about the use of the Excel-Table The equation of motion at the time t + Δt can be written as:
t + Δt·· t + Δt · t + Δt t + Δt··
• Only the yellow cells should be modified: m u+c u+k u = –m ug (7.21)
• The columns A and B contain the time vector and the ground ac-
celeration u··g ( t ) at intervals Δt ; for this ground motion the re- where
sponse of a Single-Degree-of-Freedom (SDoF) system will be t + Δt·· t
computed. To compute the response of the SDoF system for a dif- u = u·· + Δu·· (7.22)
ferent ground motion u··g ( t ) , the time and acceleration vector of the t + Δt · t
new ground motion have to be pasted into columns A and B. u = u· + Δu· (7.23)
• For a given ground motion u·· ( t ) , the response of a linear SDoF
g Using the expressions for Δu·· and Δu· given by Equations (7.17)
system is only dependent on its period T = 2π ⁄ ω n and its damp- and (7.18), the acceleration and the velocity at the time t + Δt
ing ζ . For this reason, the period T and the damping ζ can be cho-
can be written as:
sen freely in the Excel-Table.
t + Δt·· 1 t + Δt
u – u ) – --------- u· – § ------ – 1· u··
• The mass m is only used to define the actual stiffness of the SDoF t 1 t 1 t
2 u = -----------2 ( (7.24)
system k = m ⋅ ω n and to compute from it the correct spring force βΔt © 2β ¹
βΔt
f s = k ⋅ u . However, f s is not needed in any of the presented plots,
hence the defaults value m = 1 can be kept for all computations.
t + Δt · γ t + Δt γ t γ t
u – u ) + § 1 + ---· u· + Δt § 1 – ------· u··
t
u = --------- ( (7.25)
• In the field “Number of periods” (cell V19) one can enter the βΔt © β¹ © 2β¹
number of periods T i for which the response of the SDoF is to be
computed in order to draw the corresponding response spectra. Introducing Equations (7.24) and (7.25) into Equation (7.21) and
t + Δt
• The response spectra are computed if the button “compute re- solving for the only unknown u we obtain:
sponse spectra” is pressed. The macro pastes the different pe- t + Δt t + Δt
riods T i into cell S3, computes the response of the SDoF system, k⋅ u = p (7.26)
reads the maximum response quantities from the cells F6, G6, where:
H6 and I6 and writes these value into the columns L to P.
t + Δt t + Δt·· t t t f el um
p = –m ug + a 1 u + a 2 u· + a 3 u·· (7.28) f y = ------ , μ Δ = ------
Ry uy
m γc
a 1 = -----------2 + --------- (7.29)
βΔt βΔt
γ
a 2 = --------- + § --- – 1· c
m
(7.30)
βΔt © β ¹
γ
a 3 = § ------ – 1· m + Δt § ------ – 1· c
1
(7.31)
© 2β ¹ © 2β ¹
In the following the often used Takeda hysteretic model is dis- Unloading occurs along a straight line with stiffness k u . This un-
cussed in some detail. loading stiffness is computed by means of Equation (7.35) as a
function of the elastic stiffness k el and taking into softening ef-
The Takeda model was first described in [TSN70] and later mod-
fects in proportion of the previously reached maximum displace-
ified by various authors. The assumed force-deformation rela-
ment ductility μ Δ . The parameter α controls the unloading stiff-
tionship shown in the following figure was derived from the mo-
ness reduction and varies from structural element to structural
ment-curvature relationship described in [AP88].
element.
Large amplitude cycles Small amplitude cycles
–α –α
ξ+
p k u+ = k el ⋅ ( max { μ Δ+ } ) , k u- = k el ⋅ ( max { μ Δ- } ) (7.35)
+ + +
fs β⋅ξp β⋅ξp fs β⋅ξp
A+ B+ A+ A+ B+
Reloading follows a straight line which is defined by the force re-
k pl =r o ·kel fy k pl =r o ·kel
C+ versal point (u rev,0) and the point A. The location of point A is de-
β>0 β<0 termined according to the figure on the previous page as a func-
+ +
ku
-
ku
tion of the last reversal point B, the plastic strain ξ p and the dam-
k el
u rev
uy u rev
u
u rec u rev
+
u rec
u age influence parameter β . The parameter β allows taking into
account softening effects occurring during the reloading phase.
- -
ku ku
β<0
C- Again, in the case of RC walls meaningful parameters α and β
β>0
X- B- X- lay in the following ranges: α = 0.2…0.6 and β = 0.0… – 0.3 .
x- x-
-
β⋅ξp
-
β⋅ξp
These rules, which are valid for cycles with large amplitude, are
typically based on observations of physical phenomena made
The initial loading follows the bilinear force-deformation relation- during experiments.
ship for monotonic loading mentioned in the previous section.
On the other hand, rules for small amplitude cycles are based on
The exact definition of this so-called skeleton curve depends on
engineering considerations rather than on exact observations.
the structural element at hand. For example, in the case of Re-
They are designed to provide reasonable hysteresis lops during
inforced Concrete (RC) structural walls the elastic stiffness k el
an earthquake time-history, thus avoiding clearly incorrect be-
corresponds to 20 to 30% of the uncracked stiffness, while the
haviours like e.g. negative stiffnesses.
plastic stiffness k pl = r o ⋅ k el is approximated assuming an hard-
ening factor r o = 0.01…0.05 . The rules for small amplitude cycles are shown on the right of the
figure on the previous page.
If a reloading phase starts from a force reversal point (u rev,0) lay- 7.3.3 Newmark’s method for inelastic systems
+
ing between the two extreme force reversal point (u rec,0) and
- The Newmark’s numerical method discussed in Sections 7.2.1
(u rec,0) , then reloading does no longer occurs towards point A,
to 7.2.3 can be easily modified for application to nonlinear sys-
but towards a newly calculated point X, which lies between
tems. The following modifications are required:
points A and B. The position of point X is calculated using the
+ -
auxiliary variables x and x defined in Equation (7.36). • The mass m and the damping c are typically constant through-
+ -
out the whole time-history.
+ § u rev – u rec· + - § u rev – u rec· - • The stiffness k changes during the time-history, hence m̃ , re-
-¸ ⋅ βξ p ,
x = ¨ ------------------------ -¸ ⋅ βξ p
x = ¨ ------------------------ (7.36)
- + + -
© u rec – u rec¹ © u rec – u rec¹ spectively k , are no longer constant.
• If the stiffness changes within the time step iterations are need-
When a load reversal occurs before point X is reached, a new
ed (e.g. Newton-Raphson).
point C is defined as a temporary maximum and minimum. The
reloading in the subsequent cycles, which are smaller than the • For nonlinear systems the second solution strategy presented
temporary maximum and minimum is then always in the direc- in Section 7.2.1 (Option 2) has the advantage that the factors
tion of point C.
m γc γ
a = § --------- + -----· and b = ------ – Δt § 1 – ------· c
m
These rules for cycles with a small amplitude are a simplification © βΔt β ¹ 2β © 2β¹
of those described in [AP88], however they lead to very satisfac-
tory results and can be programmed very easily and efficiently. on the RHS of Equation (7.19) are constant throughout the
whole time-history and do not need to be recomputed at every
time step.
Remark
For the actual implementation of the nonlinear version of New-
mark’s time stepping method, it is suggested to use the formula-
tion presented in Section 7.4.4 in conjunction with the Newton-
Raphson algorithm described in Section 7.4.3.
7.3.4 Example 1: One-storey, one-bay frame • Frames with T = 3.75s , ζ = 5% subjected to “El Centro”
Δ=28.1cm 40
Displacement [cm]
20
0 0
−20
−20 Vel=39.5kN
−40
0 10 20 30 −20 0 20
40
Displacement [cm]
20
0 0
−20
12EI s My Vy −20
Δ=25.9cm Vy=19.8kN
- , T = 2π m
k = 2 ⋅ ------------- ---
- , V = 2 ⋅ ----------
- , Δ = ------ (7.37) −40
H
3 k y H⁄2 y k 0 10 20 30 −20 0 20
40
Displacement [cm]
• Parameters 20
Displacement [cm]
20
HEA 220 2181 0.95 272 123.4 0.50 5.7
• Frames with T = 0.95s , ζ = 5% subjected to “El Centro” • Frames with T = 0.27s , ζ = 5% subjected to “El Centro”
400
200 2
Displacement [cm]
Displacement [cm]
10
0 0 0 0
−1 −200
−10 Δ=11.3cm −200 Vel=246.7kN −2
Δ=1.5cm Vel=411.4kN
−400
0 10 20 30 −10 0 10 0 10 20 30 −2 −1 0 1 2
400
200 2
Displacement [cm]
Displacement [cm]
10
Δ=7.7cm Shear force [kN]
0 0 0 0
−1 −200
−10
Δ=1.2cm
−200 Vy=123.4kN −2 Vy=205.7kN
−400
0 10 20 30 −10 0 10 0 10 20 30 −2 −1 0 1 2
400
200 2
Displacement [cm]
Displacement [cm]
10
Δ=1.5cm
Shear force [kN]
0 0 0 0
−1 −200
−10 Δ=9.4cm −200 Vy=61.7kN Vy=102.9kN
−2
−400
0 10 20 30 −10 0 10 0 10 20 30 −2 −1 0 1 2
400
200 2
Displacement [cm]
Displacement [cm]
10
Shear force [kN]
0 0 0 0
Δ=2.2cm
−1 −200
−10
Δ=10.4cm −200 Vy=41.1kN −2 Vy=68.6kN
−400
0 10 20 30 −10 0 10 0 10 20 30 −2 −1 0 1 2
Time [s] Displacement [cm] Time [s] Displacement [cm]
yield force f y and the yield displacement u y of the Takeda SDoF -20
system are then estimated using the force reduction factor R y as -40
follows: -60
f el u el -80
0 3 8 11 14 17
R y = ----- = ------ (7.39) 1 2 4 5 6 7 9 10 12 13 15 16 18
fy uy Time [s]
Hysteretic behaviour of the nonlinear SDoF System 7.4 Solution algorithms for nonlinear analysis
Force-deformation relationship of the Takeda-SDoF System
problems
subjected to the same ground motion as Wall WDH4. In both di- In this section the more general case of system with multiple de-
agrams the same curve is plotted: On the left in absolute units grees of freedom is discussed. SDoF system can be seen as a
and on the right in normalised units. special case thereof.
80
40
0.5 The general equilibrium condition for elastic and inelastic static
20
and dynamic systems is:
0 0.0
-20 F(t) = R(t) (7.40)
-0.5
-40 In this equation F ( t ) is the time-dependent vector of the internal
-60 -1.0 forces of all DoFs of the structure and R ( t ) the time-dependent
-80 vector of the external forces.
-80 -60 -40 -20 0 20 40 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
Relative displacement x [mm] Displacement dutility mD=u/uy [mm] The vector R ( t ) depends on the problem analysed and is known.
In Equation (7.42) R is known and K is also known, therefore The actual solution of the equilibrium conditions of Equation
the unknown vector U can be computed by means of Equation (7.40) is often obtained by the numerical method for the iterative
(7.43): solution of nonlinear equations that was originally developed by
Newton.
–1
U = K R (7.43)
In the next section the so-called Newton-Raphson Algorithm for
The equilibrium condition of Equation (7.40) can only be solved the solution of Equation (7.40) will be discussed.
for linear-elastic systems by means of Equation (7.43).
For inelastic systems, due to successive yielding of the struc- 7.4.3 The Newton-Raphson Algorithm
ture, the stiffness matrix K is not constant over the course of the The Newton-Raphson Algorithm allows the solution of the load-
loading. ing of nonlinear springs with the following equilibrium condition:
For this reason Equation (7.40) must be solved in increments F(U(t) ) = R(t) (7.47)
(=small load steps) and iteratively. The approach is as follows:
t F ( U ( t ) ) = R ( t ) represents the internal spring force, which is a
• The nodal displacements U at the time t are known from the given nonlinear function of U ( t ) . The external force R ( t ) is a
previous load step; function of the time t .
t + Δt
• The nodal displacements U at the end of the load step Δt
For a system with 1 DOF the solution method of the Newton-
are determined by means of n iterations of Equations (7.44)
Raphson Algorithm can be illustrated by the figure on page 7-29.
and (7.45).
The algorithm consists of the following steps:
t + Δt i–1 i t + Δt i–1
KT ΔU = ΔR (7.44)
0) Up to time step t a solution was obtained and at time step t
t t
t + Δt i t + Δt i – 1 i the system is in equilibrium with ( U, R) ;
U = U + ΔU (7.45)
1) The initial conditions at the beginning of the iteration are de-
where: termined. The iteration commences with the nodal displace-
t + Δt i–1 t + Δt t + Δt i – 1 ment, the tangent stiffness and the internal force that have
ΔR = R– F (7.46) resulted at the end of the previous time step t . The external
t + Δt 0
and K T is the tangent stiffness matrix of the structure. initial loading increment ΔR within the time step is de-
termined by Equation (7.51).
7 Seismic Excitation
t+ΔtK 2
t+ΔtΔR1 T
t+ΔtK 1
t+ΔtΔR0 T
Course “Fundamentals of Structural Dynamics”
t+ΔtK 0
T
tR
Page 7-29
An-Najah 2013
R
t+ΔtΔU1 t+ΔtΔU2 t+ΔtΔU3
t+ΔtR
t+ΔtΔR2
7 Seismic Excitation
t+ΔtΔR1
t+ΔtK 0
T
t+ΔtK 0
t+ΔtΔR0 T
Course “Fundamentals of Structural Dynamics”
t+ΔtK 0
T
tR
Convergence criteria • For this reason the displacement criterion is typically used in con-
junction with other convergence criteria.
• A in-depth discussion of the convergence criteria can be found • A force criterion, which checks the residual forces, is given in
in [Bat96] and [AEM86]. This section provides only a short Equation (7.56).
overview.
t + Δt t + Δt i
• In Step 5) criteria are required in order to decide whether con- R– F-
---------------------------------------
t + Δt
≤ εF (7.56)
vergence of the iteration was obtained. Possible convergence R– F
t
criteria can be based on displacements, force or energy con-
siderations. ε F is the force convergence tolerance, which checks the mag-
nitude of the residual force after the ith-iteration against the first
• Since within the time step the unknown target displacement
t + Δt load increment of the time step.
U needs to be determined, it makes sense to prescribe that
the target displacement is reached within a certain tolerance • As for the displacement criterion, this force criterion should not be
interval. For this reason a possible displacement criterion for applied on its own because in some cases the target displacement
t + Δt
U may not have been reached. This may happen for systems
the convergence is:
with small post-yield stiffness.
i
ΔU - • The energy criterion in Equation (7.57) has the advantage
------------------
t + Δt
≤ εD (7.55)
U that it checks the convergence of the displacements and the
forces simultaneously.
where ε D is the displacement convergence tolerance.
t + Δt i T t + Δt t + Δt i – 1
• The vector U is actually unknown and must therefore be ap- ( ΔU ) ( R– F )
t + Δt i ----------------------------------------------------------------- ≤ ε E (7.57)
proximated. Typically U is used in conjunction with a suffi- 1 T t + Δt t
( ΔU ) ( R – F)
ciently small value of ε D .
• It is important to note that in some cases – although the criteria de- ε E is the energy convergence tolerance, which checks the
scribed by Equation (7.55) is satisfied – the wanted target dis- work of the residual forces of the ith-iteration against the work
t + Δt
placement U has not been reached. of the residual forces of the first load increment of the time
• This is the case when the computed displacements vary only step.
slightly during one iteration but these small increments are repeat- Choosing the tolerances ε D , ε F or ε E too large, can yield wrong
ed over many iterations. results, which can lead to the divergence of the solution in the
• Such a situation can result when the modified Newton-Raphson following load steps.
Algorithm is used (see page 7-30).
Choosing the tolerances ε D , ε F or ε E too small, results in an un- The numerical time integration of Equation (7.58) is often per-
necessary increase of the required iteration steps. The in- formed by means of the Newmark’s Algorithms ([New 59]). Ac-
creased accuracy is typically not useful. cording to these algorithms the displacement vector at the time
For numerical reasons it can also happen that too small con- t + Δt is estimated as follows:
vergence tolerances do not allow to reach convergence at all. t + Δt t t ·· t + Δt ·· Δt
2
U = U + U· Δt + [ ( 1 – 2β ) U
t
+ 2β U ] ------- (7.61)
For strongly inelastic systems it is recommended to check 2
the sensitivity of the results to the chosen convergence t + Δt t t ·· t + Δt ··
criteria and the chosen tolerances. U· = U· + [ ( 1 – γ ) U +γ U ] Δt (7.62)
From Equation (7.61):
7.4.4 Nonlinear dynamic analyses
t + Δt ·· 1 – 2 β t ··
U – U ) – -------- U· – § -------------· U
1 t + Δt t 1 t
U = ----------2- ( (7.63)
Similar to Equation (7.44) the equilibrium condition for nonlinear β Δ t © 2β ¹
β Δt
dynamic analyses is:
Substituting Equation (7.63) into (7.62):
t + Δt ··i t + Δt · i t + Δt i – 1 t + Δti–1 i t + Δt
M U +C U +( F + KT ΔU ) = R (7.58) t + Δt γ t + Δt γ t γ t ··
U· = -------- ( U – U ) + § 1 – ---· U· + § 1 – ------· U
t
Δt (7.64)
β Δt © β¹ © 2β¹
t + Δt i t + Δt i – 1 i
U = U + ΔU (7.59)
The expressions for the displacement, the velocity and the acceler-
For base excitation by means of ground accelerations the vector ation at the time t + Δt from Equations (7.59), (7.64) and (7.63) can
of the external forces is computed according to Equation (7.60). be substituted into the differential equation of motion (7.58) which
t + Δt t + Δt can then be solved for the only remaining unknown ΔU i :
R = –M 1 ag (7.60)
§ t + ΔtKi – 1 + ---------- γ
- M + --------- C· ΔU i =
1
where M is the mass matrix of the structure, 1 the norm vector © T 2 βΔt ¹
βΔt
with entries of unity for all DoFs in the direction of the excitation t + Δt t + Δt i – 1
t + Δt R– F –
and ag the ground acceleration at the time t + Δt .
– U ) – --------- U· – § ------ – 1· U
1 t + Δt i – 1 t 1 t 1 t ··
For this type of excitation the differential equation of motion M ----------2- ( U –
βΔt βΔt © 2β ¹
(7.58) has to be integrated numerically and – due to the inelas- γ t + Δt i – 1 t γ t γ t ··
C --------- ( U – U ) + § 1 – ---· U· + § 1 – ------· U Δt
tic behaviour of the system – the equation must be solved itera- βΔt © β¹ © 2β¹
tively and incrementally. (7.65)
or in a more compact format: 7.4.5 Comments on the solution algorithms for nonlinear
analysis problems
t + Δt i–1 i t + Δt i–1
K T ΔU = ΔR (7.66)
˜ ˜ • Contrary to the analysis of elastic systems, the analysis of in-
Equation (7.66) corresponds exactly to Equation (7.44) and is elastic systems is often interrupted before the targeted load or
also solved iteratively by the Newton-Raphson Algorithm. deformation state is reached.
When dynamic analyses are carried out, typical convergence • This situation arises if in one of the time steps convergence cannot
criteria also consider the inertia forces and, if present, damping be reached.
forces. Possible, often used convergence criteria are: • Typical causes for the failing convergence
t + Δt t + Δt i – 1 t + Δt ··i – 1 t + Δt · i – 1 • The convergence tolerances are too small or too large;
R– F –M U –C U
------------------------------------------------------------------------------------------------------------------- ≤ ε F (7.67) • The chosen values for the parameters of the solution algorithm are
RNORM
not appropriate;
i T t + Δt t + Δt i – 1 t + Δt i – 1 ·· t + Δt i – 1
( ΔU ) ( R– F –M U –C U· ) • The solution algorithm is not suitable.
------------------------------------------------------------------------------------------------------------------------------------- ≤ ε E (7.68)
1 T t + Δt t t ·· t· Typical example: If special measures are not taken, the algorithms
( ΔU ) ( R – F – M U – C U)
of the Newton family are not able to solve a system with a global
with negative post-yield stiffness;
• Apart from the algorithms by Newmark, many other algorithms
RNORM = ¦ m ij ⋅ g (7.69)
have been developed for solving the differential equation of
As alternative, depending at which point during the iteration proc- motion (7.58) (e.g. “Houbolt Method”, “Wilson θ Method”, “α-
ess convergence is checked, both criteria can be rewritten as: Method”). These methods are described in detail in [Bat96].
t + Δt t + Δt i t + Δt ··i t + Δt · i
The “α-Method” allows to introduce numerical damping, which
R– F –M U –C U- can be useful. The “α-Method” is described in detail in
---------------------------------------------------------------------------------------------- ≤ εF (7.70)
RNORM [HHT77].
i T t + Δt t + Δt i t + Δt i t + Δt i
( ΔU ) ( R– F –M U·· – C U· ) • Choice of the time step Δt for static analyses:
----------------------------------------------------------------------------------------------------------------- ≤ ε E (7.71)
1 T t + Δt t ·· t
– C U· ) • For static analyses the time t has no physical meaning. For this rea-
t
( ΔU ) ( R– F–MU
son the size of the time step Δt can be chosen almost arbitrarily.
As for static analyses, different convergence criteria exist also • As long as the algorithm converges and the variation of the exter-
for dynamic analyses and a discussion of these can also be nal loads is captured correctly, the size of the time step Δt has only
found in [Bat96]. a minor influence on the results.
• However, when geometric nonlinearities are considered or when • An in-depth discussion on the accuracy of time-integration meth-
the constitutive laws are a function of the strain history, care should ods and on so-called “conditionally stable” integration methods
be taken also for static analyses when choosing the size of the can be found in [Bat96].
time step. • Example:
• The time step size influences the convergence of the algorithm: The time-integration method by Newmark with γ = 1 ⁄ 2 and
For small Δt the algorithm converges more quickly; however, more β = 1 ⁄ 6 (linear variation of the acceleration over the length of the
steps are required. time step, see Section 7.2.1) is only stable if the criterion given in
• Time steps of a variable size can be advantageous. If the system Equation (7.73) is met for all natural periods T n of the system. For
is elastic or almost elastic, large times steps can be chosen. If the systems with many DoFs, higher modes can be especially prob-
system is close to its capacity, small time steps should be chosen. lematic and a very short time step is generally required when this
time-integration method is used.
• Certain analysis programs (see for example [HKS03]) determine
the time step size within chosen limits as a function of the conver- Δt-
----- ≤ 0.551 (7.73)
gence and the number of required iterations. Tn
• Choice of the time step Δt for dynamic analyses
• For this reason the “unconditionally stable” time-integration algo-
• The objective of the dynamic analyses is the solution of the differ- rithm by Newmark with γ = 1 ⁄ 2 and β = 1 ⁄ 4 is often used in
ential equation of motion (7.58) between the time t and the time seismic engineering.
t + Δt . For this reason the choice of the time step Δt plays always
an important role regarding the accuracy of the solution.
• The accuracy of the integration of the differential equation of mo-
tion (7.58) depends on the chosen time-integration algorithm and
on the ratio given by Equation (7.72), where T n are the natural pe-
riods of the system. To capture the motion components due to
higher modes the time step Δt has therefore to be reduced.
Δt
------ (7.72)
Tn
7.4.6 Simplified iteration procedure for SDoF systems with 7.5 Elastic response spectra
idealised rule-based force-deformation relationships
7.5.1 Computation of response spectra
In the case that the hysteretic behaviour is described by a se-
quence of straight lines, a so called “idealized rule-based force-
deformation relationships” like the Takeda model presented in
Section 7.3.2, it is possible to avoid implementing a Newton-
Raphson Iteration strategy. In this case adjustments are needed
in the case of:
• Stiffness change during loading
• Velocity reversal
• Transition between unloading an reloading.
In first case a secant stiffness can be iteratively computed until
the target point lays on the backbone curve, while in the second
and third case it is often enough to reduce the size of the time
step to limit error. These strategies are shown in the following fig- • Response spectra are used to represent the seismic demand
ure: on structures due to a ground motion record and design spec-
tra are used for the seismic design of structures.
• Response spectra shall be computed for all periods and
damping rates likely to be found in structures.
• Unless specified otherwise, the response spectra presented in
the following belongs to the north-south component of the May
18, 1940 “El Centro” Earthquake (see [Cho11]).
• Additional ground motion records can be downloaded for free
from:
1) http://db.cosmos-eq.org/scripts/default.plx
2) http://peer.berkeley.edu/nga/
ζ = 0%
Absolute acceleration [m/s2]
40
ζ = 2%
30 ζ = 5%
ζ = 10%
20 ζ = 20%
10
0
0.01 0.10 1.00 10.00 100.0 u··a + 2ζωu· + ω 2 u = 0 u··a + 2ζωu· + ω 2 u = 0
3
u = 0 u··a ( t ) = u··( t ) + u··g ( t ) = u··g ( t ) u··( t ) = – u··g ( t ) u ( t ) = – u g ( t )
ζ = 0%
ζ
Relative velocity [m/s]
= 2%
2 ζ = 5% 2.0
ζ
ag [m/s2]
= 10%
ζ = 20% 0.0
1 −2.0
ag,max = 3.13 m/s2
0 10 20 30
0.4
0
vg,max = 36.1cm/s
0.01 0.10 1.00 10.00 100.0 0.2
vg [m/s]
0.6 0.0
ζ = 0%
Relative displacement [m]
0.5 ζ = 2% −0.2
0.4 ζ = 5% 0 10 20 30
ζ = 10% dg,max = 21.1cm
0.3 0.2 Baseline
ζ = 20%
correction!
dg [m]
0.2 0.1
0.1 0.0
−0.1
0.01 0.10 1.00 10.00 100.0 0 10 20 30
Period [s] Time [s]
2 2 2
kS d k ( S pv ⁄ ω ) mS pv 0
E s = --------- = --------------------------- = ------------- (7.75) 0.01 0.10 1.00 10.00 100.00
2 2 2
1.0
Velocity
Pseudo-velocity
• Pseudo-acceleration
S pa
S pv = -------- log ( S pv ) = – log ( ω ) + log ( S pa )
0.0 ω
log ( S pv ) = – log ( f ) – log ( 2π ) + log ( S pa )
log ( S pv ) = log ( T ) – log ( 2π ) + log ( S pa )
-1 x Displacement
−1.0
0 5 10 15
1.0
Pseudo-acceleration ζ>0 100
Spv = 71 cm/s
Pseudo-velocity
10
00
0
... / ... max [−]
/s 2
Pseudo−velocity Spv [cm/s]
cm
0.0
10
0
Sd
44
10
00
=
=
11
pa
10
.
S
2
cm
-1 x Displacement
10
10
Ps
−1.0
0
eu
do
0 5 10 15
−a
]
cc
m
Time [s]
[c
.S
d
tS
pa
10
1
[c
en
u··a ( t ) = – ω 2 u ( t ) – 2ζωu· ( t )
m
(7.78)
em
/s
2
]
ac
1
pl
Time-history of the pseudo-acceleration A(t)
is
D
1
1
0.
• For ζ = 0 : u(t) = A(t)
• For ζ > 0 : At u max : u a = A however A < A max T=1s
Shift of the location of the maxima through damping 0.1 1.0 10.0
Period [s]
7.5.3 Properties of linear response spectra 7.5.4 Newmark’s elastic design spectra ([Cho11])
ζ = 0%
ζ = 2%
Spa / ag [-]
ζ = 5%
100 ζ = 10%
10
00
0
Pseudo−velocity Spv [cm/s]
vg = 36.1 cm/s
T [s]
10
0
10
dg
00
=
21
10
.1
cm
10
10
Ps
0
eu
/s 2
do
[-]
−a
cm
]
cc
g
/d
[c
Sp
.S
3
31
d
tS
pa
10
d
a
1
S
=
[c
/a
en
m
g
a
em
/s
g
2
[-]
]
ac
1
pl
Spv / vg [-]
is
D
1
1
0.
• Newmark’s elastic design spectra • Elastic design spectra according to SIA 261 (Art. 16.2.3)
6
ζ = 5% Zone 3b, ζ = 5%
5
αd
g
C
aa
0
dg
α
Pseudo−velocity Spv [cm/s]
vg 3
10
0
E
dg
10
00
g
a
10 B 2 Ground Class A
Ground Class B
10
10
Ground Class C
Ps
F 1
eu
Ground Class D
do
−a
]
cc
Ground Class E
m
[c
.S
A 0
d
tS
pa
10
1
[c
em
/s
Period [s]
2
]
ac
1
pl
is
1
0.
• Displacement elastic design spectra according to SIA 261 • Elastic design spectra according to SIA 261 (linear)
25 6
Ground Class C
15 Ground Class D 3
Ground Class E
2
10
1
Zone 3b, ζ = 5%
5
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Period [s]
0
25
0.01 0.10 1.00 10.00
Period [s] Ground Class A
Ground Class B
• The displacement spectra are computed from the acceleration 20
Ground Class C
spectra using equation (7.79)
Displacement Sd [cm]
Ground Class D
S pa 15 Ground Class E
S d = -------2- (7.79)
ω
10
• Displacement spectra are an important design tool (even within
force-based design procedures) because they allow a quick esti-
mate of the expected deformations, hence of the expected dam- 5
• Elastic design spectra: Newmark vs. SIA 261 7.5.5 Elastic design spectra in ADRS-format (e.g. [Faj99])
(Acceleration-Displacement-Response Spectra)
3.5
Elastic design spectrum according to
SIA 261, Ground Type A S pa = ω 2 S d
3.0
B C ζ = 5%
2.5 0.5 15
0.4
10
Spa / Ag [−]
Sd [cm]
2.0 0.3
Spa [g]
0.2
5
1.5
0.1
0.0 0
1.0 A 0 1 2 3 4 5 0 1 2 3 4 5
Period [s] Period [s]
Elastic design spectrum according to
SIA 261, Ground Type B D
0.5
Newmark’s elastic design spectrum 0.5
0.0 0.4
0.01 0.10 1.00 10.00 T = const.
Period [s] 0.3
Spa [g]
• The SIA 261 spectra, like the spectra of the majority of the stand- 0.2
ards worldwide, were defined using the same principles as New- 0.1
mark’s spectra.
0.0
• However, different ground motion were used: 0 5 10 15
Sd [cm]
- SIA 261 takes into account different ground classes;
- Different seismic sources were considered; Periods T correspond to lines running through the origin of the
- A larger number of ground motions was considered. axes, because:
• Note: in SIA 261 the corner period T A is not defined.
S pa = ω 2 S d and after reorganizing: T = 2π S d ⁄ S pa
0.2 ζ = 5%
0.1
Where:
0.0 f el
0 5 10 15 R y = ----- : Force reduction factor (7.80)
Sd [cm] fy
Maximum restoring force that the elastic SDOF system
• Design spectra are defined based on averaged response spectra. f el : (7.81)
reaches over the course of the seismic excitation u··g ( t )
For this reason, the spectral values of single response spectra
fy : Yield force of the inelastic SDOF system (7.82)
may differ significantly from the design spectra.
• This is a crucial property of design spectra and should be um
μ Δ = ------ : Displacement ductility (7.83)
kept in mind during design! uy
Maximum displacement that the inelastic SDOF system
um : (7.84)
reaches over the course of the seismic excitation u··g ( t )
uy : Yield displacement of the inelastic SDOF system (7.85)
• Results Quantity Elastic SDOF Inela. SDOF Ry=2 Inela. SDOF Ry=6
0.15
Elastic SDOF system T [s] 2.0 2.0 2.0
0.10 Inelastic SDOF system (Ry=2) Fmax [kN] 134.70 67.35 22.45
Inelastic SDOF system (Ry=6) Ry [-] – 2.0 6.0
0.05 uy [m]
Displacement [m]
– 0.068 0.023
um [m] 0.136 0.147 0.126
0.00
μΔ [-] – 2.16 5.54
-0.05
• Comments
-0.10
• Both inelastic SDOF systems show a stable seismic response.
-0.15
0 5 10 15 20 25 30 35 40
Time [s]
150
100
50
Force [kN]
-50
7.6.2 “Seismic behaviour equation” • More realistic representation of the decision possibilities
For seismic collapse prevention, the following approximate rela-
tionship applies
Δ
″quality″ of seismic behaviour ≈ strength × ductility (7.86) F
Δ
F
7.6.3 Inelastic behaviour of a RC wall during an earthquake 7.6.4 Static-cyclic behaviour of a RC wall
• Moment-curvature-relation-
Wall WDH4
ship at the base of the plas- 100
tic hinge zone.
50
Moment [kNm]
• Hysteretic behaviour of the RC wall under static-cyclic loading 7.6.5 General definition of ductility
500 …m
Ductility demand μ … = ---------
400
Wall WSH3 Fy …y
0.75 Fy
300
…u
Ductility capacity: μ … = --------
200 …y
Actuator force [kN]
100
0
μΔ=6 μΔ=5 μΔ=4 μΔ=3 μΔ=2 μΔ=1 μΔ=1 μΔ=2 μΔ=3 μΔ=4 μΔ=5 μΔ=6
−100
−200
−300 −0.75 Fy
−400 −Fy
−500
−100 −80 −60 −40 −20 0 20 40 60 80 100
Top displacement [mm]
• Comments
• The ductility capacity is a property of the structural member.
• The ductility demand is a result of the seismic excitation and also
a function of the dynamic properties of the structure.
• A structural member survives the earthquake if:
Plastic region of test unit WSH6 (left) and close-up of the left boundary region (right). Both pho-
tos were taken at displacement ductility 6.
f el um
R y = ----- , μ Δ = ------
fy uy
strain εu
μ ε = -----
ductility εy μΔ ( Ry ) = ?
10
ζ = 5% μΔ=1
φu μΔ=2
Sa [m/s2]
curvature
μ φ = ----- 5 μΔ=4
ductility φy μΔ=6
0
0.01 0.10 1.00 10.00 100.0
1.0
μΔ=1 ζ = 5%
μΔ=2
Sv [m/s]
rotation θu 0.5 μΔ=4
μ θ = -----
ductility θy μΔ=6
0.0
0.01 0.10 1.00 10.00 100.0
0.4
μΔ=1 ζ = 5%
0.3
μΔ=2
Sd [m]
displacement Δu 0.2 μΔ=4
μ Δ = ------
ductility Δy 0.1
μΔ=6
μΔ=6 Ry=6
Ry [−]
Ry=4
μΔ=4
μΔ=2
Ry=2
Ry=(2μΔ-1)0.5 Ry=μΔ
1 1
0.1 1.0 10.0
0.1 1.0 10.0
Period [s]
Period [s]
• In the small period range, already small reductions of the elastic
strength of the SDOF system yield very large ductility demands.
• If the ductility demand is very large, it can be difficult to provide the
Ry = μΔ Ry = 2μ Δ – 1 structure with a sufficiently large ductility capacity. This problem will be
further discussed during the design classes.
• Also in the large period range – where the “equal displacement princi-
ple” applies – large discrepancies between real and estimated ductility
demand can occur.
• The “equal displacement principle” and the “equal energy principle”
are “historical” Ry-μΔ-Tn relationships. In recent years a lot of research
has been done to come up with more accurate formulations (see e.g.
works by Krawinkler [KN92], Fajfar [VFF94], Miranda [Mir01], ...)
D
C
=α
aa
α
d
10
dg
A=
vg
00
0
g
a
0
E
dg
10
Vy [cm/s]
B C’ V/μΔ
D’
10
0
10
10
00
Vy [cm/s]
10
.5
1) 0
00
10
/μ Δ
10
F
Δ-
μ
(2
A/
E’
10
10
Ay
D
B’ Inelastic
0
=d
[c
10
m
g
design spectrum
0
A
/s
2
1
]
A’ for μΔ=4
]
1
m
10
a
1
A=
Ay
[c
F’
y
[c
10
D
m
1
/s
]
0.
2
m
]
[c
D
y
/μ Δ
D
1
1
0.
1
0.1 1.0 10.0 TF=33s
Period [s]
0.1 1.0 10.0
Period [s] • Maximum displacement of the SDOF system
um = μΔ ⋅ Dy
Note the new axes: D y = u y , V y = ω n u y , A y = ω n2 u y
• Yield strength of the SDOF system:
where: u y = yield displacement
fy = m ⋅ Ay
Construction of the spectra using Ry-μΔ-Tn relationships • Inelastic design spectra according to [NH82] (log. x-axis)
Ay / ag [−]
D = S d, inelastic = ------ ⋅ S d, elastic (7.88)
Ry
1
It should be noted that: Equal displacement principle
S pa, inelastic ≠ ω 2 ⋅ S d, inelastic (7.89)
• Ry-μΔ-Tn relationship according to [NH82]
0
0.01 0.10 1.00 10.00 100.00
° 1 Tn < Ta
° TA TB TC TD TE TF
° ( 2μ – 1 ) β ⁄ 2 T < T < T
° Δ a n b
° 2μ Δ – 1 T b < T n < T c' (EE principle) 2 μΔ=1
Ry = ® (7.90)
° T
μΔ=2
° -----n- μ Δ T c' < T n < T c μΔ=4
° Tc
° μΔ=6
Sd / dg [−]
° μΔ Tn > Tc (ED principle)
¯
1
Where: β = log ( T n ⁄ T a ) ⁄ log ( T b ⁄ T a ) (7.91)
T a = 1 ⁄ 33s , T b = 1 ⁄ 8s (7.92)
T c = Corner period between the constant Spa and the Equal displacement principle
constant Spv regions
T c' = Corner period between the constant Spa and the 0
0.01 0.10 1.00 10.00 100.00
constant Spv regions of the inelastic spectrum Period [s]
• Inelastic design spectra according to [NH82] (linear x-axis) • Ry-μΔ-Tn relationships according to [VFF94]
In [VFF94] Ry-μΔ-Tn relationships are defined as follows:
5.0 TB TC TD ζ = 5%
° cR T n
° c 1 ( μ Δ – 1 ) ⋅ ----- Tn ≤ T0
4.0 -+1
μΔ=1 Ry = ® T0 (7.93)
°
3.0
μΔ=2 c
° c1 ( μΔ – 1 ) R + 1 Tn > T0
Ay [m/s2]
μΔ=4 ¯
μΔ=6 Where:
2.0 T 0 = c 2 ⋅ μ ΔcT ⋅ T c ≤ T c (7.94)
T c = Corner period between the constant Spa and the
1.0
constant Spv regions
0.15 Model
TB TC TD ζ = 5%
Hysteresis Damping c1 cR c2 cT
Q Mass 1.0 1.0 0.65 0.30
For the Q-hysteretic model and mass-proportional damping, the • Inelastic design spectra according to [VFF94]
Ry-μΔ-Tn relationships by [VFF94] specialise as:
5.0 TB TC TD ζ = 5%
T
° ( μ Δ – 1 ) ⋅ -----n- + 1 Tn ≤ T0
Ry = ® T0 (7.95) 4.0
° μΔ=1
¯ μΔ Tn > T0 (ED principle)
μΔ=2
Where: T 0 = 0.65 ⋅ μ Δ0.3 ⋅ T c ≤ T c
3.0 μΔ=4
Ay [m/s2]
(7.96)
μΔ=6
T c = Corner period between the constant Spa and the
2.0
constant Spv regions
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.15
TB TC TD ζ = 5%
0.10
D [m]
μΔ=1
0.05
μΔ=2
μΔ=4
μΔ=6
0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Period [s]
• Inelastic design spectra in ADRS-format 7.7.2 Determining the response of an inelastic SDOF sys-
tem by means of inelastic design spectra in ADRS-
5.0
TB TC format
ζ = 5%
In this section the response of two example inelastic SDOF sys-
4.0
[NH82] tems is determined by means of inelastic design spectra in
μΔ=1 ADRS-format.
μΔ=2
Spa [m/s2]
3.0
• SDOF system 1 with Tn = 0.9 s
μΔ=4
μΔ=6 • SDOF system 2 with Tn = 0.3 s
2.0
• The spectra according to [VFF94] will be used (see Section 7.7.1)
TD • Example 1: SDOF system with Tn= 0.9 s
1.0
m fs
0.0 properties:
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Sd [m] C m = 100 t
5.0
TB TC fy = 80 kN
ζ = 5% k kel = 4874 kN/m
uy = 0.016 m
4.0 kpl = 0 kN/m
[VFF94] fy
μΔ=1 ζ = 5%
kpl
kel
μΔ=2
Spa [m/s2]
3.0
u
μΔ=4 uy um
2.0
μΔ=6
• Response of the elastic SDOF system 1:
TD
T n = 2π m 100- = 0.9s
1.0
---- = 2π -----------
k 4874
S pa = 2.62m/s 2
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
S d = 0.054m
Sd [m]
f el = 261.7kN
• Response of the inelastic SDOF system 1: • Example 2: SDOF system with Tn= 0.3 s
f el 261.7
R y = ----- = ------------- = 3.27 m fs
fy 80 properties:
μ Δ = R y = 3.27 (From Equation (7.95) since T n > T c = 0.5s ) fy,a = 120 kN
C m = 100 t uy,a = 0.0027 m
u m = u y ⋅ μ Δ = 0.016 ⋅ 3.27 = 0.054m = S d
Representation of the inelastic SDOF system 1 in the inelastic k kel = 43‘865 kN/m fy,b = 300 kN
design spectrum in ADRS-format: uy,b = 0.0068 m
kpl = 0 kN/m
fy
kpl
5.0
ζ = 5%
μΔ = 1 kel
u
uy um
4.0
• Response of the elastic SDOF system 2
T n = 2π m 100 - = 0.3s
3.0 ---- = 2π --------------
Spa [m/s2]
μΔ = 3.27 k 43865
Tn = 0.9s
Performance point S pa = 4.71m/s 2
2.0 S d = 0.011m
f el = 471kN
1.0 In this second example two different inelastic SDOF systems will
Capacity curve
um = 0.054m be considered: (a) A SDOF system with a rather low fy and (b) a
0.0
SDOF system with a rather high fy.
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Sd [m] • Response of the second inelastic SDOF system 2a
f el 471
• If the force-deformation relationship of the inelastic SDOF system R y = ----- = --------- = 3.93
fy 120
is divided by its mass m, the “capacity curve” is obtained, which
can be plotted on top of the spectrum in ADRS-format. In this case the resulting displacement ductility μ Δ is so large, that
• The capacity curve and the inelastic spectrum intersect in the “per- Equation (7.96) T 0 = T c = 0.5s results. After rearranging Equa-
tion (7.95), the displacement ductility μ Δ can be computed as:
formance point”.
The maximum displacement response is therefore: In this case the displacement ductility μ Δ will be such that Equation
(7.96) yields T 0 < T c = 0.5s . To compute μ Δ insert therefore Equa-
u m = u y ⋅ μ Δ = 0.0027 ⋅ 5.88 = 0.016m > S d tion (7.96) in Equation (7.95). This results in following expression:
Representation of the inelastic SDOF system 2a in the inelastic Tn
( μ Δ – 1 ) ⋅ ---------------------------------
- + 1 = Ry (7.97)
design spectrum in ADRS-format: 0.65 ⋅ μ Δ ⋅ T c
0.3
T0 0.383
The maximum displacement response is therefore:
2.0
u m = u y ⋅ μ Δ = 0.0068 ⋅ 1.73 = 0.012m > S d
Capacity curve μΔ = 1
1.0 um = 0.016m
μΔ = 5.88
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Sd [m]
• Note that the line (*) is no longer vertical as in Example 1, but in-
clined according to the equation μ Δ = ( R y – 1 ) ⋅ ( T c ⁄ T n ) + 1 .
7.7.3 Inelastic design spectra: An important note 7.7.4 Behaviour factor q according to SIA 261
3
• This characteristic of the design spectra shall be taken into ac- Elastic design
Sd / agd [-]
count during design by aiming at robust structures. q=1.5 spectrum
2
q=2
1
Φ0=1.5 q=3
q=4
0
0.01 0.10 1.00 10.00
Period [s]
7.8 Linear equivalent SDOF system (SDOFe) • Example: Inelastic SDOF system with Takeda-hyst. rule [TNS70]
u The properties of the inelastic SDOF system are:
..
fs -mug(t) fs • Damping: ζ = 5% (constant, proportional to kel)
..
mu Elastic • Mass: m = 100t
. (with elastic stiffness kel)
cu • Stiffness: k el = 4874kN/m
• Yield force: f y = 80kN
• Hysteresis: Takeda-hysteresis with r o = 0.05 , α = 0.5 , β = 0.0
kel Elastic
(with „effective stiffness“
ug(t) or „secant stiffness“ keff) + + –α
fm k u = k el ( max { μ Δ } )
.. fy - - –α
ug(t) kpl Inelastic k u = k el ( max { μ Δ } )
kel keff
u
uy um
Damping: ζe (7.100)
The differential equation of the SDOFe is: The maximum response of the SDOF system when subjected to the NS-
component of the 1940 El Centro Earthquake is:
2
u··( t ) + 2ζ e ω e u· ( t ) + ω e2 u ( t ) = – u··g ( t ) with ω e = k eff ⁄ m (7.101) x m = 0.073m , f m = 93.0kN
The question is how the viscous damping ζ e of the SDOFe can The properties of the corresponding SDOFe are:
be determined so that max ( u ( t ) ) = u m . fm 93.0 kN m- = 2π -----------
100- = 1.76s
k eff = ------ = ------------- = 1274 ------- , T e = 2π -------
um 0.073 m k eff 1274
ζ e = 22.89% , the viscous damping ζ e was determined iteratively!
Comparison between inelastic SDOF and SDOFe Comments regarding the example:
• The damping ζ e is in general larger than the damping ζ , since ζ e
100
of the SDOFe needs to compensate for the hysteretic energy ab-
ζ = 5%
sorption of the inelastic SDOF system.
ζe = 22.89%
• However, in rare cases it happens that ζ e < ζ . This shows again
50
the difficulties that are associated with the prediction of the seismic
response of inelastic SDOF systems.
Force [kN]
0.02 In particular in the sixties significant research has been dedicated to es-
timating the damping ζ e (see for example [Jac60], [Jen68] and [IG79]).
0.00 At that time the interest in linear equivalent systems was big because the
numerical computation of the response of inelastic systems was ex-
-0.02
tremely expensive. The basic idea behind estimating the damping ζ e
-0.04 was:
Inelastic SDOF The inelastic SDOF system dissipates energy due to ζ and due to the
-0.06
um Equivalent SDOF inelastic deformations, which are a function of its inelastic force-defor-
-0.08 mation relationship. The equivalent SDOF system, however, dissipates
0 5 10 15 20 25 30 35 40 energy solely due to its viscous damping. For this reason the following
Time [s]
relationship applies:
ζ e = ζ + ζ eq (7.102) fm = [ 1 + ro ( μΔ – 1 ) ] ⋅ fy
fo = ( 1 – ro ) ⋅ fy
where ζ eq is the viscous damping equivalent to the hysteretic energy ab-
sorption of the inelastic system. u1
f 1 = ----------------- ⋅ f 2
u1 + u3
The simplest method for estimating the equivalent viscous damping is to
f2 = [ 1 + ro ( μΔ – 1 ) ( 1 – β ) ] ⋅ fy
assume that the inelastic system and the linear equivalent system dissi-
pate the same energy within one displacement cycle. According to this u1 = μΔ uy – u2
assumption [Cho11] defines the equivalent elastic damping as: 1 + ro ( μΔ – 1 )
- ⋅ uy
u 2 = ---------------------------------
–α
μΔ
1 Ah
ζ eq = ------ ⋅ ------ (7.103) u3 = [ μΔ – β ( μΔ – 1 ) ] ⋅ uy
4π A e
Where: The equivalent viscous damping of other important hysteresis rules is:
A h : Energy dissipated by the inelastic SDOF system due to the ine-
lastic deformation of the system. The dissipated energy corre-
sponds to the area of the force-displacement hysteresis of the Elasto-plastic (EP) rule:
considered displacement cycle;
2 μΔ – 1
A e : Potential energy of the equivalent SDOF system at maximum dis- ζ eq, EP = --- ⋅ --------------- (7.105)
π μΔ
placement:
2
k eff ⋅ u m
A e = --------------------
2
Bilinear (BL) rule:
The inelastic force-deformation relationship of many structural RC ele-
ments can be described by the “Takeda”-hysteresis rule. According to 2 ( μΔ – 1 ) ( 1 – ro )
ζ eq, BL = --- ⋅ ------------------------------------------ (7.106)
Equation (7.103) the equivalent viscous damping of this hysteresis rule π μΔ ( 1 + ro μΔ – ro )
is:
1 ( f m + f o )μ Δ u y + f 1 u 1 – f m u 2 – ( f m + f o )u 3
ζ eq,Tak = ------ ⋅ ----------------------------------------------------------------------------------------------------- (7.104)
4π ( fm μΔ uy ) ⁄ 2
Rule according to Clough (Clo) [CP75]:
Where:
2 3 μΔ – 1
ζ eq, Clo = --- ⋅ ------ ⋅ --------------- (7.107)
π 2π μΔ
ζ eq - μ Δ -relationship for these important hysteresis rules: Comments regarding the comparison of the theoretical value with the
computed value of ζ e for the Takeda-SDOF system when excited by the
70
2/π El Centro earthquake:
60 Elasto-plastic • The computed value of ζ e was determined iteratively. Eight differ-
50 ent inelastic SDOF systems with different periods T n were consid-
Clough ered. The strength of each inelastic SDOF system was varied in
ζeq [-]
Tn=0.67s
30 7 Tn=0.50s • This shows again the difficulties associated with the prediction of
ζe [-]
The constants a to d are: • Takeda Thin (TT): Hysteretic rule that characterises RC struc-
tures which lateral stiffness is provided by walls and columns.
• Takeda Fat (TF): Hysteretic rule that characterises RC structures
which lateral stiffness is provided by frames.
• Ramberg-Osgood (RO): Hysteretic rule that characterises ductile
steel structures.
• “Flag-Shaped”, β=0.35 (FS): Hysteretic rule that characterises
prestressed structures with unbonded tendons.
ζ eq - μ Δ -relationships for the most important hysteresis rules according to
[GBP05]:
20
1) EPP
2) BL (ro=0.2)
Hysteresis rule a b c d 15 3) Takeda Thin
1) Elasto-Plastic (EPP) 0.224 0.336 -0.002 0.250 4) Takeda Fat
ζeq [-]
2) Bilinear, ro=0.2 (BI) 0.262 0.655 0.813 4.890 5) FS (beta=0.35)
10 6) RO
3) Takeda Thin (TT) 0.215 0.642 0.824 6.444
4) Takeda Fat (TF) 0.305 0.492 0.790 4.463
5) Ramberg-Osgood (RO) 0.289 0.622 0.856 6.460 5
6) “Flag-Shaped”, β=0.35 (FS) 0.251 0.148 3.015 0.511
Te=2.0s
from [GBP05] 0
0 1 2 3 4 5 6
The hysteresis rules 1) to 6) were chosen because they can be used to
μΔ [-]
represent the hysteretic behaviour of typical structural types:
• Elasto-Plastic (EPP): Hysteretic rule that characterises systems Important comments:
for the seismic isolation of structures (sliding systems that are • With these relationships an in a statistical sense improved esti-
based on friction). mate of the damping ζ e is obtained.
• Bilinear, ro=0.2 (BI): Hysteretic rule that also characterises sys- • For single systems subjected to a specific ground motion differenc-
tems for the seismic isolation of structures. The value of the post- es between the maximum response of the inelastic system and the
yield stiffness rokpl may vary significantly between different sys- maximum response of the equivalent SDOF with ζ e according to
tems. these improved ζ eq - μ Δ -relationships can still be significant!
5.0 TB TC TD 5.0 TB TC
[BE99]
ζ = 5%
4.0 10% 4.0
20%
Spa [m/s2]
3.0 3.0
Ay [m/s2]
30% ζ = 5%
2.0 2.0 10%
20% TD
1.0 1.0 30%
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12
Sd [m]
0.15
TB TC TD
5.0 TB TC
[TF99]
ζ = 5%
4.0
0.10 10%
Spa [m/s2]
3.0
20%
D [m]
ζ = 5%
30%
2.0
10%
0.05
20%
30 TD
40
1.0 50%
0.00 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12
Period [s] Sd [m]
7.8.2 Determining the response of inelastic SDOF systems • Response of the elastic SDOF system
by means of a linear equivalent SDOF system and
T n = 2π m 100- = 0.9s
elastic design spectra with high damping ---- = 2π -----------
k 4874
The computation of the seismic response of inelastic systems by S pa = 2.62m/s 2
means of linear equivalent systems was studied by Sozen and S d = 0.054m
his co-workers in the seventies (see for example [GS74], [SS76] • Response of the inelastic SDOF system
and [SS81]). The maximum response of the inelastic SDOF system will be com-
puted by means of the ADRS-spectra (page 7-106).
Today this approach gains new attention since the “Direct Dis-
placement-Based Design (DDBD)” approach, which was devel- Step 1: The capacity curve of the SDOF system is plotted on top
of the ADRS spectra.
oped by Priestley and his co-workers, is based on the idea of the
linear equivalent system ([PCK07]). Step 2: By means of Equ.s (7.102) and (7.104) the nonlinear
scale, which represents the damping ζ e as a function of the maxi-
This sections outlines the procedure for computing the response mum response of the SDOF system, is plotted along the capacity
of an inelastic SDOF system by means of an linear equivalent curve.
SDOF system and elastic design spectra with high damping. Step 3: Several spectra for different values of damping are plotted.
• Example: SDOF system with Tn= 0.9 s Step 4: The “Performance Point” is the point where the spectrum
with damping ζ e intersects the capacity curve at the same value of
m fs ζe .
Properties: Takeda hysteresis For the considered example the maxmimum response of the ine-
(α=0.5, β=0.0, ro=0.05) lastic SDOF system is:
C m = 100 t
S d = 0.065m
fy = 80 kN
k kel = 4874 kN/m Comments regarding the example:
uy = 0.016 m
kpl = 244 kN/m • To determine the “Performance Point” exactly, an iterativ approach
fy is typically required.
kpl
ζ = 5% kel • The linear equivalent SDOF system is fully defined by the period
u T e and the damping ζ e . The period T e results from the slope of the
uy um
line that connects the origin with the “Performance Point”.
For the example the spectra according to [BE99] will be used (Section 7.8.1).
• The damping values ζ eq used in the figures on page 7-106 were Determining the SDOF behaviour by means of elastic ADRS-spectra
determined according to Equation (7.104). In the figures on page 5.0 ζ = 5%
7-107 the damping ζ eq was determined using Equation (7.108).
The difference is, however, rather small.
• It should be noted that in both cases the computed maximum re- 4.0 10% ζ = 18.5%
sponse of the inelastic SDOF system does not comply with the 15%
Performance point
“equal displacement principle”.
Spa [m/s2]
3.0
20%
• The linear equivalent SDOF system leads often to results that do 25% Capacity curve
not agree with the “equal displacement principle”. This applies in
2.0
particular to SDOF systems with long periods or systems with
large ductility demands.
1.0 ζe = 5%
• A second example is presented on page 7-107. It is a SDOF sys-
20%
tem with a shorter period and a smaller ductility demand than in 10% 15%
ζe = 18.5%
Example 1. In this second example the “equal displacement prin- 0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
ciple” is approximately confirmed. Sd [m]
Alternative representation:
5.0 ζ = 5%
4.0
Spa [m/s2]
3.0
Tn = 0.9s
2.0
Te = 1.67s
1.0
3.0 [Faj99] Fajfar P.: “Capacity Spectrum Method Based on Inelastic De-
Te = 0.88s
mand Spectra”. Earthquake Engineering and Structural Dy-
namics, 28, 979-993, 1999.
2.0
[HHT77] Hilber H.M., Hughes T.J.R, Taylor R.L.: “Improved Numerical
Dissipation for Time Integration Algorithms in Structural Dy-
1.0
namics”. Journal of Earthquake Engineering and Structural
μΔ = 1.97 ζe= 12.8% Dynamics, Vol. 5, 1977.
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 [HKS03] Hibbit Karlsson & Sorensen: “ABAQUS Version 6.4 - User’s
uy = 0.02m um = 0.039m Sd [m] Manual”. Pawtucket, RI, 2003.
[KN92] Krawinkler H., Nassar A.A.: “Seismic design based on ductil- [PCK07] Priestley M.J.N., Calvi G.M., Kowalsky M.J.: “Displacement-
ity and cumulative damage demands and capacities”. Nonlin- Based Seismic Design of Structures”. IUSS Press, 2007.
ear Seismic Analysis and Design of Reinforced Concrete
[Saa91] Saatcioglu M.: “Modeling Hysteretic Force-Deformation Re-
Buildings, eds P. Fajfar and H. Krawinkler. New York: Elsevier
lationships for Reinforced Concrete Elements”. In Publication
Applied Science. 1992.
SP 127 “Earthquake-Resistant Concrete Structures, Inelastic
[GBP05] Grant D.N., Blandon C.A., Priestley M.J.N.: “Modelling Ine- Response and Design”, ACI, Detroit 1991.
lastic Response in Direct Displacement-Based Design”.
[SIA03] Swiss standard SIA 261: “Actions on structures”. SIA, Zürich
Rose School, Report 2005/03, IUSS Press, Pavia, 2005.
2003.
[GS74] Gulkan P., Sozen M.: “Inelastic Responses of Reinforced
[SS81] Saiidi A., Sozen M.: “Simple Nonlinear Seismic Analysis of
Concrete Structures to Earthquake Motions”. ACI Journal, Ti-
R/C Structures”. ASCE, Journal of the Structural Division,
tle No. 71-41. December 1974.
Vol. 107, No. 5, May 1981.
[IG79] Iwan W.D., Gates N.C.: “Estimating Earthquake Response of
[SS76] Shibata A., Sozen M.: “Substitute Structure Method for Seis-
Simple Hysteretic Structures”. ASCE, Journal of the Engi-
mic Design in R/C”. ASCE, Journal of the Structural Division,
neering Mechanics Division, Vol. 105, EM3, June 1979.
Vol. 102, No. 1, January 1976.
[Jac60] Jacobsen L.S.: “Damping in composite structures”. Proceed-
[TF99] Tolis S.V., Faccioli E.: “Displacement design spectra”. Jour-
ings of the 2nd World Conference in Earthquake Engineering.
nal of Earthquake Engineering vol. 3, No. 1, 1999.
Vol. 2. Tokio and Kyoto, Japan, 1960.
[TSN70] Takeda T., Sozen M.A., Nielsen N.N.: “Reinforced Concrete
[Jen68] Jennings P.C.: “Equivalent Viscous Damping for Yielding
Response to Simulated Earthquakes”. ASCE, Journal of
Structures”. ASCE, Journal of the Engineering Mechanics Di-
Structural Engineering, Vol. 96, No. 12, December 1970.
vision, Vol. 94, No. EM1, February 1968.
[VFF94] Vidic T., Fajfar P., Fischinger M.: “Consistent inelastic design
[Mir01] Miranda E.: “Estimation of inelastic defromation demands of
spectra: strength and displacement”. Earthquake Engineer-
SDOF Systems”. ASCE, Journal of structural engineering,
ing and Structural Dynamics 23(5) 507-521. 1994.
Vol. 127, No. 9, 2001.
[New59] Newmark N.M.: “A Method of Computation for Structural Dy-
namics”. ASCE, Journal of the Engineering Mechanics Divi-
sion, Vol. 85, No. 3, July 1959.
[NH82] Newmark N.M., Hall W.J.: “Earthquake Spectra and Design”.
Earthquake Engineering Research Institute (www.eeri.org),
1982.
• Stiffness matrix
k 11 k 12 ( k1 + k2 ) –k2
K = = (8.5)
k 21 k 22 –k2 k2
·· · ·
m 1 u 1 + ( c 1 + c 2 )u 1 – c 2 u 2 + ( k 1 + k 2 )u 1 – k 2 u 2 = f 1 ( t )
® (8.2)
¯ m 2 u··2 – c 2 u· 1 + c 2 u· 2 – k 2 u 1 + k 2 u 2 = f 2 ( t )
·· + Cu· + Ku = f ( t )
Mu (8.4)
8 Multi Degree of Freedom Systems Page 8-1 8 Multi Degree of Freedom Systems Page 8-2
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• Mass matrix M By means of the principle of virtual forces the vertical displace-
ment Δ at location d due to a unit force F = 1 acting at location
m1 0 a can be readily determined.
M = (8.6)
0 m2 FL 3 a d
Δ (α,δ) = – α δ ( α 2 + δ 2 – 1 ) ⋅ --------- with α = --- and δ = --- (8.9)
6EI L L
• Equation of motion
The flexibilty matrix consists of the following elemnts:
m1 0 u··1 ( k1 + k2 ) –k2 u1
+ = 0 (8.7) u = DF (8.10)
0 m2 ··
u2 –k2 k2 u2 0
u1 d 11 d 12 F1
·· + Ku = 0
Mu (8.8) = ⋅ (8.11)
u2 d 21 d 22 F2
8.1.3 Flexibility formulation The d ij factors can be computed by means of Equation (8.9) as
follows:
4 L3
d 11 = Δ (1 ⁄ 3,2 ⁄ 3) = --------- ⋅ ------ (8.12)
243 EI
7 L3
d 12 = Δ (2 ⁄ 3,2 ⁄ 3) = --------- ⋅ ------ (8.13)
486 EI
7 L3
d 21 = Δ (1 ⁄ 3,1 ⁄ 3) = --------- ⋅ ------ (8.14)
• Flexibility matrix D 486 EI
4 L3
d 22 = Δ (2 ⁄ 3,1 ⁄ 3) = --------- ⋅ ------ (8.15)
243 EI
L3
D = --------------- ⋅ 8 7 (8.16)
486EI 7 8
8 Multi Degree of Freedom Systems Page 8-3 8 Multi Degree of Freedom Systems Page 8-4
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• Mass matrix M
m1 0
M = (8.18)
0 m2
The stiffness matrix K of a beam element with constant flexural
• Equation of motion
and axial stiffness is well known:
m1 0 u··1 162 EI u F = Ku (8.21)
+ --------- ⋅ -----3- ⋅ 8 – 7 1 = 0 (8.19)
0 m2 u··2 5 L –7 8 u2 0
EA EA
·· + Ku = 0 -------- 0 0 – -------- 0 0
Mu (8.20) L L
12EI-
----------- 6EI
--------- 12EI- 6EI
F1 0 0 – ----------- --------- u1
8.1.4 Principle of virtual work L3 L2 L3 L2
F2 u2
6EI 4EI 6EI- 2EI
• See e.g. [Hum12] F3
0 --------- --------- 0 – -------- ---------
u3
= L2 L L2 L
⋅ (8.22)
F4 EA EA u4
8.1.5 Energie formulation – -------- 0 0 -------- 0 0
F5 L L u5
• See e.g. [Hum12] F6 0
12EI 6EI
– -----------
- – --------- 0
12EI 6EI
------------ – --------- u6
L3 L2 L3 L2
6EI 2EI 6EI 4EI
0 --------- --------- 0 – -------- - ---------
L2 L L2 L
8 Multi Degree of Freedom Systems Page 8-5 8 Multi Degree of Freedom Systems Page 8-6
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
Equation of motion:
mL 0 0 0 u··1 24 0 – 12 6L u1 0
0 0 0 0 u·· u2
⋅ 2 EI
+ -----3- ⋅ 08L 2 2L 2
– 6L ⋅ = 0 (8.25)
mL u··3 L u3 0
0 0 -------- 0 – 12 – 6L 12 – 6L
2
u··
4
2
6L 2L – 6L 4L 2 u4 0
12 6L – 12 6L 0 0 0 0
EI 2 2
K = -----3- ⋅ 6L 4L – 6L 2L (8.23)
L – 12 – 6L 12 – 6L mL 0 0 0 u··1 24 – 12 0 6L 0 u1
6L 2L 2 – 6L 4L 2 mL u··EI – 12 12 – 6L – 6L u
0 -------- 0
2
0 ⋅ 3 + -----3- ⋅ ⋅ 3 = 0 (8.26)
• Example: Cantilever ··
u2 L 0 – 6L 8L 2 2L 2 u2 0
0 0 0 0
u··4 6L – 6L 2L 2 4L 2 u4 0
0 0 0 0
Static condensation:
mL 0 0 0 u··1 24 – 12 0 6L 0 u1
mL u··EI u
0 -------- 0
2
0 ⋅ 3 + -----3- ⋅ – 12 12 – 6L – 6L ⋅ 3 = 0 (8.27)
u··2 L 0 – 6L 8L 2 2L 2 u2 0
Assemblage of the stiffness matrix 0 0 0 0
u··4 6L – 6L 2L 2 4L 2 u4 0
0 0 0 0
F1 12 + 12 – 6 L + 6L – 12 6L u1
F2 EI 2 2 2L 2 u2 ··
= -----3- ⋅ – 6 L + 6L 4L + 4L – 6L ⋅ (8.24) m tt 0
⋅
u t + k tt k t0 ⋅ u t = 0 (8.28)
F3 L – 12 – 6L 12 – 6L u3 ··
0 0 u0 k 0t k 00 u0 0
F4 6L 2L 2 – 6L 4L 2 u4
with L = L ⁄ 2
8 Multi Degree of Freedom Systems Page 8-7 8 Multi Degree of Freedom Systems Page 8-8
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
··
m tt u t + k tt u t + k t0 u 0 = 0 EI 6
® (8.29) k̂ tt = -----3- ⋅ --- ⋅ 16 – 5 (8.36)
¯ k 0t u t + k 00 u 0 = 0 L 7 –5 2
From the second row of Equation (8.29) the following expression after substituting L = L ⁄ 2 :
can be derived:
EI 48
–1 k̂ tt = -----3- ⋅ ------ ⋅ 16 – 5 (8.37)
u0 = – k 00 k 0t u t (8.30) L 7 –5 2
Substituting Equation (8.30) in the first line of Equation (8.29) we The final equation of motion of the cantilever is therefore:
obtain:
mL
·· + k u – k k k u = 0
m tt u
–1
(8.31) -------- 0 u·· u
2 EI 48
⋅ 1 + -----3- ⋅ ------ ⋅ 16 – 5 ⋅ 1 = 0
t tt t t0 00 0t t
(8.38)
·· + ( k – –1 mL u··3 L 7 – 5 2 u3 0
m tt u t tt k t0 k 00 k 0t )u t = 0 (8.32) 0 --------
4
T
and with k t0 = k 0t :
• Notes
T
·· + ( k – k k k )u = 0 –1
m tt u t tt 0t 00 0t t (8.33) • The “Direct Stiffness Method” is often used in the Finite Element
Method.
·· + k̂ u = 0 with k̂ = k – k k k T –1
m tt u t tt t tt tt 0t 00 0t (8.34) • The derivation of the stiffness matrix K for a beam element and in-
structions for assembling the stiffness matrix of entire structures
can be found e.g. in the following references:
Where k̂ tt is the condensed stiffness matrix, and in our case it is
equal to: [Prz85] Przemieniecki J.S.: “Theory of Matrix Structural Analy-
sis”. Dover Publications, New York 1985.
§ 1 1 ·
¨ ---------
2
– -----------2- ¸
EI ¨ 24 – 12 14L ⋅ 0 – 6 L ¸
k̂ tt = -----3- ⋅ – 0 6L ⋅ 7L [Bat96] Bathe K-J.: “Finite Element Procedures”. Prentice Hall,
L ¨ ¸ Upper Saddle River, 1996.
¨ – 12 12 –6 L –6 L 1
– -----------2-
2-
-------- 6L – 6 L ¸
© 14L 7L 2 ¹
(8.35)
8 Multi Degree of Freedom Systems Page 8-9 8 Multi Degree of Freedom Systems Page 8-10
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
u1 = u1
°
® L (8.41)
° u 2 = --- ⋅ θ 1
¯ 2
u1 1 0 u1
= ⋅ or u = Au (8.42)
θ1 0 2⁄L u2
°
®
°
¯
A
Using Equation (8.23), the equation of motion for free vibrations The matrix A is called coordinate transformation matrix and can
of the system becomes be used to transform the mass matrix, the stiffness matrix and
the load vector from one set of DoF to the other, i.e.
3m
-------- 0 T
2 u·· EI 12 – 6L u K = A KA (8.43)
⋅ 1 + -----3- ⋅ ⋅ 1 = 0 (8.39)
2 ·· 2 θ1
mL θ1 L – 6L 4L 0 M = A MA
T
(8.44)
0 -----------
8
T
F = A F (8.45)
8 Multi Degree of Freedom Systems Page 8-11 8 Multi Degree of Freedom Systems Page 8-12
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
For the example at hand, the stiffness matrix K expressed in the 8.1.8 Systems incorporating rigid elements with distribut-
set of DoFs u 1 and u 2 becomes: ed mass
The 2-DoF system depicted in the following incorporates a rigid
EI 12 – 6L
K = A KA = 1 0 ⋅ -----3- ⋅ ⋅ 1 0
T
(8.46) element with distributed mass μ .
0 2⁄L L – 6L 4L
2 0 2⁄L
EI
K = -----3- ⋅ 12 – 12 (8.47)
L – 12 16
3m
-------- 0
2
M = A MA = 1 0 ⋅ ⋅ 1 0
T
(8.48)
0 2⁄L mL
2 0 2⁄L
0 -----------
8
3m
-------- 0
M = 2 (8.49)
m
0 ----
2 The elements of the 2x2 mass matrix can be determined by
imparting a unit acceleration üa=1 to one degree of freedom
which yields the equation of motion of the 2-DoF systems ex-
while keeping the acceleration of the other degree of free-
pressed in terms of the DoFs u 1 and u 2
dom equal to zero (üb=0).
3m The resulting inertia forces are then applied as static forces
-------- 0 u··1 EI u
2 acting onto the system, and the elements of the mass matrix
⋅ + -----3- ⋅ 12 – 12 ⋅ 1 = 0 (8.50)
m u··2 L – 12 16 u2 0 are computed as the reactions to these static forces.
0 ----
2
8 Multi Degree of Freedom Systems Page 8-13 8 Multi Degree of Freedom Systems Page 8-14
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
μL
M = ------- ⋅ 2 1 (8.55)
6 1 2
EI
K = -----3- ⋅ 28 – 10 (8.56)
L – 10 4
and the equation of motion of the system for free vibration be-
comes:
8 Multi Degree of Freedom Systems Page 8-15 8 Multi Degree of Freedom Systems Page 8-16
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
9 Free Vibrations Also in the case of Equation (9.7), there is always the trivial so-
lution φ n = 0 , which corresponds to an absence of movement.
9.1 Natural vibrations If the matrix A has an inverse A –1 , then Equation (9.7) can be
rearranged as follows:
·· + Ku = 0
Mu (9.1)
–1 –1
Ansatz: A Aφ n = A 0 (9.8)
u ( t ) = q n ( t )φ
φ n where q n ( t ) = A n cos ( ω n t ) + B n sin ( ω n t ) (9.2) and therefore
Equation (9.5) is satisfied if q n ( t ) = 0 , which is a trivial solution If the determinant A is equal to zero, then the matrix is singular
meaning that there is no movement, because u ( t ) = q n ( t )φφ n = 0 . and has no inverse.
To obtain a nontrivial solution the term in brackets in Equation Therefore, Equation (9.6) has a nontrivial solution only if:
(9.5) must be equal to zero, i.e.:
2
2 – ωn M + K = 0 (9.11)
[– ωn M + K ]φ
φn = 0 (9.6)
2
The determinant yields a polynomial of order N in ω n which is
or:
called characteristic equation. The N roots of the characteris-
Aφ n = 0 with A = – ω n M + K
2
(9.7) tic equation are called eigenvalues and allow the calculation of
the N natural circular frequencies ω n of the system.
As soon as the natural circular frequencies ω n are computed, 9.2 Example: 2-DoF system
also the vectors φ n can be computed within a multiplicative con-
stant by means of Equation (9.6). There are N independent Vec-
tors which are called eigenvectors or natural modes of vibra-
tion of the system. We consider a regular 2-DoF
oscillator with
Summary m1 = m2 = m
• Natural circular frequencies and eigenvectors are properties of The equation of motion of the system corresponds to equation
the MDoF system and depends only from its mass and stiff- (8.7):
ness properties.
u·· u
• The index n refers to the numbering of the eigenvectors and m 1 0 1 + k 2 –1 1 = 0 (9.12)
0 1 u··2 –1 1 u2 0
the first mode of vibration ( n = 1 ) is commonly referred to as
the fundamental mode of vibration.
9.2.1 Eigenvalues
The eigenvalues are calculated from the determinant:
2
2 2k – ω n m –k
K – ωn M = = 0 (9.13)
2
–k k –ωn m
Fundamental mode:
2k – § ---------------- ⋅ ----· m
3– 5 k
–k φ 11
© 2 m¹
= 0
2
[ K – ω 1 M ]φ
φ1 = ⋅
φ 21 0 3---------------
– 5- --- k k
ω1 = ⋅ - = 0.618 ----
k – § ---------------- ⋅ ----· m
3– 5 k
–k 2 m m
© 2 m¹
(9.17)
“Mode” φ 11 2 -
---------------
we obtain two independent equations that can be used to deter- φ1 = = 1 + 5 = 0.618
φ 21 1
mine the elements of the first eigenvector φ 1 . The first row of the “Degree of freedom” 1
system yield the equation:
(------------------------
1 + 5 )k (1 + 5)
φ 11 – kφ 21 = 0 and φ 21 = --------------------- φ 11 (9.18)
2 2
9.2.3 Higher modes of vibration 9.2.4 Free vibrations of the 2-DoF system
Additionally to the fundamental mode of vibration, the consid- According to Equation (9.2), the free vibration of the 2-DoF sys-
ered 2-DoF system has a second mode of vibration. tem is:
The properties of this second mode of vibration can be comput- u = [ C 1 cos ( ω 1 t ) + C 2 sin ( ω 1 t ) ]φ
φ 1 + [ C 3 cos ( ω 2 t ) + C 4 sin ( ω 2 t ) ]φ
φ2
ed in analogy to the fundamental mode and the following results (9.20)
are obtained:
u1 φ 11
= [ C 1 cos ( ω 1 t ) + C 2 sin ( ω 1 t ) ] (9.21)
Second mode u2 φ 21
φ 12
3---------------
+ 5- --- k k- + [ C 3 cos ( ω 2 t ) + C 4 sin ( ω 2 t ) ]
ω2 = ⋅ - = 1.618 --- φ 22
2 m m
The still unknown constants C 1 to C 4 can be computed using the
“Mode” φ 12 1
initial conditions given by Equation (9.24) and become:
φ2 = = 1– 5 = 1
φ 22 ---------------- – 0.618
“Degree of freedom” 2 φ 22 u1 – φ 12 u2 φ 22 v1 – φ 12 v2
C 1 = -------------------------------------- , C 2 = ------------------------------------------------- (9.22)
φ 11 φ 22 – φ 21 φ 12 ( φ 11 φ 22 – φ 21 φ 12 )ω 1
φ 11 u2 – φ 21 u1 φ 11 v2 – φ 21 v1
C 3 = -------------------------------------- , C 4 = ------------------------------------------------- (9.23)
φ 11 φ 22 – φ 21 φ 12 ( φ 11 φ 22 – φ 21 φ 12 )ω 2
u1 ( 0 ) = u1
°
° u2 ( 0 ) = u2
Initial conditions: ® · (9.24)
° u1 ( 0 ) = v1
°·
¯u (0) 2 = v2
splacement u2 [m]
splacement u2 [m]
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
-0.2 -0.2
-0.4 -0.4
Dis
Dis
-0.6 -0.6
-0.8 -0.8
-1.0 -1.0
-1.2 -1.2
0 5 10 15 20 0 5 10 15 20
Time [s] Time [s]
1.2 1.2
First mode First mode
1.0 Second mode 1.0 Second mode
0.8 Total displacement 0.8 Total displacement
splacement u1 [m]
splacement u1 [m]
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
-0.2 -0.2
-0.4 -0.4
Dis
Dis
-0.6 -0.6
-0.8 -0.8
-1.0 -1.0
-1.2 -1.2
0 5 10 15 20 0 5 10 15 20
Time [s] Time [s]
-0.4 φ N1 φ N2 … φ NN
-0.6
• Spectral matrix
-0.8
0 5 10 15 20
ω 12 0 … 0
Time [s]
2 0 ω 22 … 0
Ω = (9.26)
0.8
First mode … … … …
0.6 Second mode
0 0 … ω N2
Total displacement
splacement u1 [m]
0.4
Equation (9.6) can therefore be rearranged as follows:
0.2
Kφ n = Mφ n ω n2 (9.27)
0.0
-0.2
and it is immediately apparent that the equation for all eigenval-
ues and all eigenvectors can be expressed in terms of modal
Dis
9.4 Properties of the eigenvectors vectors are linearly independent and can be chosen so that they
are orthogonal (proof complicated).
9.4.1 Orthogonality of eigenvectors So far we have shown that φ Tn Mφ r = 0 for n ≠ r . By means of
The orthogonality conditions of the eigenvectors are: Equation (9.32) we can prove also the φ Tn Kφ r = 0 for n ≠ r . We
have already seen that for n ≠ r the right hand side of Equation
T T
φ n Kφ r = 0 and φ n Mφ r = 0 for n ≠ r (9.29) (9.32) is equal to zero. For this reason also the left hand side of
Equation (9.32) must be equal to zero, which conclude the veri-
and can be proven by means of Equation (9.27). Equation (9.27) fication.
is first to be set up for the eigenvector vector n , and then pre-
T
multiplied with φ r on both sides: Example: 2-DoF system
T T
φ r Kφ n = ω n2 φ r Mφ n (9.30) In the following the orthogonality of the eigenvectors of the 2-
Afterwards, Equation (9.30) shall be transposed making use of DoF system presented in Section 9.2 is checked:
T T
the symmetry properties of the matrices K = K and M = M : • Relative to the mass matrix
T T
φ n Kφ r = ω n2 φ n Mφ r (9.31) 2
T 2 - m 0 ⋅ ---------------- = 2m (5 + 5)
φ 1 Mφ 1 = --------------- 1 ⋅ 1+ 5 ----------------------------- ≅ 1.382m
Now, Equation (9.27) shall be set up for the eigenvector vector 1+ 5 0 m (1 + 5)
2
T
r , and then pre-multiplied with φ n on both sides: 1
(9.34)
T T
φ n Kφ r = ω r2 φ n Mφ r (9.32)
1
Equation (9.32) can now be subtracted from Equation (9.31) T 2 -
φ 1 Mφ 2 = --------------- 1 ⋅
m 0 ⋅ (9.35)
yielding the following equation: 0 m 1– 5 = 0
1+ 5 ----------------
2
T
( ω n2 – ω r2 )φ
φ n Mφ r = 0 (9.33)
2
In the case that the eigenvalues are different, then for n ≠ r we T – 5- ⋅ m 0 ⋅ ---------------- = 0
φ 2 Mφ 1 = 1 1--------------- (9.36)
T 1+ 5
have ( ω n2 – ω r2 ) ≠ 0 and the expression φ n Mφ r must be zero. In 2 0 m
1
the case that an eigenvalue occurs more than once, the eigen-
1
T 1 – 5- ⋅ 2k – k ⋅ k
φ 2 Kφ 2 = 1 --------------- 1 – 5 = --- ( 5 + 5 ) ≈ 3.618k
2 –k k ---------------- 2
2
(9.41)
qi : modal coordinates N
u(t) = ¦ q i ( t )φ
φi (9.52)
The displacement vector u ( t ) and its double derivative i=1
··( t ) =
u ¦ q··i ( t )φφ i (9.46) in matricial form:
i
q1 ( t )
can be substituted into Equation (9.44), and the latter can be left-
T u ( t ) = Φ q ( t ) with q = … (9.53)
multiplied by φ n yielding the following equation:
qN ( t )
φ n M § ¦ q··i ( t )φ
φ i· + φ n K § ¦ q i ( t )φ
φ i· = 0
T T
© ¹ © ¹
(9.47) ·· + K * q = 0
i i
M*q (9.54)
The natural frequency corresponds to the result shown on page and making use of the possibility of decoupling of the equation
9-7. of motion, the total deformation u ( t ) under free vibration can be
computed as the sum of the contribution of all modes. The equa-
tion of motion of the nth decoupled SDoF system is:
m n* q··n ( t ) + k n* q n ( t ) = 0 (9.72)
where:
2
ω nd = ω n 1 – ζ “damped circular frequency of the nth mode” Blank Page
(9.85)
The the total deformation u ( t ) under free vibration is hence
N N
– ζω n t
u(t) = ¦ φi qi ( t ) = ¦ φi e [ A i cos ( ω id t ) + B i sin ( ω id t ) ] (9.86)
i=1 i=1
q· n ( 0 ) + ζω n q n ( 0 )
B n = -------------------------------------------- (9.88)
ω nd
Definition:
Eigenvectors: φ1 = 1 ⁄ 2 , φ2 = –1 (10.9)
1 1
• A system is classically damped if the matrix C * is diagonal
• A system is non-classically damped if the matrix C * is not di-
agonal
u·· u· u
m 2 0 1 + c 5 –4 1 + 3 –1 1 = 0 (10.11) u·· u· 1 u
0 1 u··2 – 4 4 u· 2 –1 1 u2 0 m 2 0 1 + c 6 –2 + 3 –1 1 = 0 (10.16)
0 1 u··2 –2 2 ·u
2 –1 1 u2 0
It is now attempted to decouple the equations by computing the
It is now attempted to decouple the equations by computing the
modal properties of the 2-DoF system:
modal properties of the 2-DoF system:
1--- 1--- 3---
1 –1 m 0 3--- 3k
M*
T
= Φ MΦ = 2 ⋅ 2m 0 ⋅ 2 = 2 (10.12) m 0 ------ 0
–1 1
0 m
1 1 0 3m M* = 2 , K* = 4 (10.17)
0 3m 0 6k
c n* = a 0 m n* (10.20) and using the results for MpD and SpD damping ratio ζ n be-
comes:
and the corresponding damping ratio ζ n becomes (see Section 3.2):
a0 a1
ζ n = --------- + ----- ω n (10.27)
c n* a 0 m n* a0 2ω n 2
ζ n = ----------------*- = ----------------*- = --------- (10.21)
2ω n m n 2ω n m n 2ω n
The coefficients a 0 and a 1 may be computed for vibration modes
i and j by means of equation (10.28):
10.3.2 Stiffness proportional damping (SpD)
C = a1 K (10.22) a0 1 a1
° ----- ⋅ ----- + ----- ⋅ ω i = ζ i
° 2 ωi 2
The damping constant of each mode of vibration is therefore: ® (10.28)
° ----
a0 1 a1
2 ° 2- ⋅ ω ----- + ----- ⋅ ω j = ζ j
2
c n* = a 1 k n* = a 1 ω n m n* (10.23) ¯ j
and the corresponding damping ratio ζ n becomes: In the case that ζ i = ζ j = ζ , coefficients a 0 and a 1 can be com-
2 puted as follows:
c n* a 1 ω n m n* a1
ζ n = ----------------*- = ------------------- = ----- ω n (10.24)
2ω n m n 2ω n m n * 2 2ω i ω j 2
a 0 = ζ ⋅ ----------------- a 1 = ζ ⋅ ----------------- (10.29)
ωi + ωj ωi + ωj
Remark
Both MpD and SpD, taken alone, are not a good approximation
of the behaviour of real structures. Studies have shown that dif-
ferent modes of vibration exhibit similar damping ratios.
k- 4
--- ⋅ 2m + 3k 0–k
Dam
4ζ m 4ζ mk
C = ------ ⋅ ------ ⋅ m = ------ ⋅ -------- ⋅ 5 – 1 3
3 2k k- 3 2 –1 2
0–k --- ⋅m+k 2
m
1
(10.32)
0
0 1 2 3 4 5 6 7 8 9 10
Check:
Circular frequency ω
1
--- 1 4ζ
1
--- – 1
Remarks
mk 5 – 1 mk
= ζ ⋅ -------- ⋅ 3 0
T
C * = Φ CΦ = 2 ⋅ -----
- ⋅ -------
- ⋅ ⋅ 2
3 2 –1 2 2 0 12 - If there are more than 2 modes of vibrations, then not all of
–1 1 1 1
them will have the same damping ratio.
(10.33)
- If more than 2 modes of vibrations should have the same
The damping matrix is indeed diagonal. damping, then a different damping modal shall be used. To
this purpose see e.g. “Caughey-Damping” in [Cho11].
with
F0 F0
f 1 = -------- and f 2 = – -------- (11.14)
3m 3m
The properties of the 2-DoF system are:
Each equation of the system (11.13) corresponds to the equation
m 1 = 2m , m2 = m (11.8) of motion of an undamped SDoF system under an harmonic sine
k 1 = 2k , k2 = k (11.9) excitation. The complete solution of these differential equations
has been discussed in Chapter 4 and it is:
c1 = 0 , c2 = 0 (11.10)
fn
2
- sin ( ωt )
q n = A 1 cos ( ω n t ) + A 2 sin ( ω n t ) + ------------------
2
(11.15)
The external excitation is: ωn – ω
or
sin ( ωt ) – ( ω ⁄ ω 1 ) sin ( ω 1 t )
°q = f -----------------------------------------------------------------
-
° 1 1 2
ω1 – ω
2
°
® (11.22)
° sin ( ωt ) – ( ω ⁄ ω 2 ) sin ( ω 2 t )
° q2 = f2 -----------------------------------------------------------------
2 2
-
° ω2 – ω
¯
11.1.3 Example 2: RC beam with Tuned Mass Damper (TMD) • RC beam with TMD
without damping
• RC Beam
• Damping ratio
ζ n = 0.0
• Modal mass
M n = 5.626t
• Modal stiffness
K n = 886kN ⁄ m
• Natural frequency
f n = 2Hz
• Excitation
• TMD (In this case damping is neglected)
As excitation a vertical harmonic sine force acting only on the
• Damping ratio
ζ T = 0.0 beam is assumed.
• Mass F ( t ) = F o sin ( ωt ) (11.26)
M T = 0.310t
with: ω : excitation frequency
• Stiffness
F o : static excitation force: F o = 0.8kN
K T = 44kN ⁄ m
• Natural frequency
• Solution
f T = 1.90Hz Both the transient and the steady-state part of the solution are
considered.
0.02 80
0.01
60
0
Amp
40
-0.01
Disp
20
-0.02
0
-0.03
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
Excitation frequency / Natural frequency of the beam [-]
-0.04
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time [s] • Remarks
0.02
where u st = F o ⁄ K n
0.01
- The solution was computed by means of the Excel file giv-
0 en on the web page of the course (SD_MDOF_TMD.xlsx)
-0.01
- The Tuned Mass Damper (TMD) is blocked
Dis
-0.02
- The natural frequency of the beam with TMD is: f n = 1.94Hz
-0.03
- At f = f n resonance occurs. In the diagram above the am-
-0.04 plification factor is limited, because the response of the
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time [s] system was only calculated during 60 seconds.
0.004 200
0.002
150
0
Amp
100
-0.002
Disp
50
-0.004
0
-0.006
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
Excitation frequency / Natural frequency of the beam [-]
-0.008
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time [s] • Remarks
0.02
where u st = F o ⁄ K n
0.01
- The solution was computed by means of the Excel file giv-
0 en on the web page of the course (SD_MDOF_TMD.xlsx)
-0.01
- The Tuned Mass Damper (TMD) is free to move
Dis
-0.02
- No resonance at f = f n occurs. Resonance occurs in corre-
-0.03 spondence of the first and of the second natural frequen-
-0.04
cies of the 2-DoF system.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 In the diagram above the factor A is limited, because the
Time [s]
response of the system was only calculated during 60s.
12 Seismic Excitation The system of equations governing the motion of the system is
F1 + T1 = 0
12.1 Equation of motion ® (12.1)
¯ F2 + T2 = 0
12.1.1 Introduction ·· ·· · ·
– m 1 ( x + u 1 ) – ( c 1 + c 2 )u 1 + c 2 u 2 – ( k 1 + k 2 )u 1 + k 2 u 2 = 0
In analogy to Section 2.1.1, the equation of motion of the system ® (12.2)
¯ – m 2 ( x·· + u··2 ) + c 2 u· 1 – c 2 u· 2 + k 2 u 1 – k 2 u 2 = 0
depicted here can be formulated by means of the d’Alembert
principle F + T = 0 applied to each one of the masses. and in matricial form:
y 1 = x ( t ) + l 1 + u s1 + u 1 ( t )
m1 0 x·· + u··1 c 1 + c 2 – c 2 u· 1 k 1 + k 2 – k 2 u 1
– – – = 0 (12.3)
y··1 = x··( t ) + u··1 ( t ) 0 m2 ··
x + u2 ·· –c2 c2 u2 · –k2 k2 u2 0
F 2 = – k 2 ( u s2 + u 2 – u 1 ) – c 2 ( u· 2 – u· 1 ) + m 1 g which is similar to Equation (8.3) meaning that the base point ex-
citation x ( t ) can be considered equivalent to two external forces
F 2 = k 2 u 1 – k 2 u 2 + c 2 u· 1 – c 2 u· 2 f 1 ( t ) = m 1 x··( t ) and f 2 ( t ) = m 2 x··( t ) acting on the masses m 1 and
m 2 . This is the same interpretation given in Section 2.1.1 for
SDoF systems.
12.1.2 Synchronous Ground motion Influence vector for some typical cases
As shown in the previous section, the equation of motion of a • Planar system with translational ground motion (Case 1)
system subjected to a base excitation is:
In this case all DoFs of the system undergo
·· + Cu· + Ku = 0
Mu (12.6) static displacements u s ( t ) which are equal
a
to the ground displacement u g ( t ) , hence:
where u·· is vector of the absolute accelerations of the DoFs of
a
the system while u· and u are the vectors of the relative veloci- 1
ties and of the relative displacements of the DoFs of the system,
respectively. ι = 1 = 1 (12.11)
…
The absolute displacement u a of the system can be expressed 1
as:
where 1 is a vector of order N , i.e. the
ua = us + u (12.7) number of DoFs, with all elements equal to 1.
where u s is displacement of the DoFs due to the static applica-
tion (i.e. very slow so that no inertia and damping forces are gen- • Planar system with translational ground motion (Case 2)
erated) of the ground motion, and u is again the vector of the rel- The axial flexibility of the elements
ative displacements of the DoFs of the system. of the depicted system can be ne-
glected, hence 3 DoFs are defined.
The “static displacements” u s ( t ) can now be expressed in func-
In this case DoFs 1 and 2 undergo
tion of the ground displacement u g ( t ) as follows:
static displacements which are
us ( t ) = ι ug ( t ) (12.8) equal to the ground displacement,
while the static displacement of
where ι is the so-called influence vector. Equation (12.6) can DoF 3 is equal to 0, i.e.:
now be rewritten as:
··) + Cu· + Ku = 0
M ( ιu··g + u (12.9) 1
ι = 1 (12.12)
·· + Cu· + Ku = – Mιu·· ( t )
Mu (12.10)
g 0
• Planar system with rotational ground motion. • Spatial system with multiple translational ground motion
Consider the spatial frame depicted here:
U6
U5 U4
U3
U2 U1
The depicted system is subjected to a rotational ground motion θ g
which generates the following static displacements of the DoFs:
Excitation in Excitation in
h1 h1 the y-direction
the x-direction
u s ( t ) = h θ g ( t ) hence ι = h (12.13)
2 2 ugx(t)
L L ugy(t)
Remark
Picture from: Chaudat T., Pilakoutas K., Papastergiou P., Ciupala M. A. (2006) “Shaking Table
If the planar system with rotational ground motion has more than Tests on Reinforced Concrete Retrofitted Frame With Carbon Fibre Reinforced Polymers
one support and every support is subjected only to the base ro- (CFRP),” Proceedings of the First European Conference on Earthquake Engineering and
Seismology, Geneva, Switzerland, 3-8 September 2006
tation θ g , then the static application of the base rotations typical-
ly create stresses within the system. Such a case must be con- The equation of motion of the frame structure for the ground mo-
sidered like a multiple support excitation (see Section 12.1.3). tions u gx ( t ) and u gy ( t ) neglecting damping is:
u··1 u1
§ 0 1 ·
u·· u2 ¨ ¸
2
¨ 1 0 ¸ Example of structure where often multiple support excitation is applied: Plan view of the
u·· u ¨ 0 u·· ( t ) + 0 u·· ( t )¸ dynamic model for the seismic analysis of a bridge in the transverse direction.
M 3 +K 3 = –M ¨ gx gy ¸ (12.15) The springs represent the piers.
u··4 u4 ¨ 0 1 ¸
¨ ¸ In this case it is distinguished between the DoFs of the structure
u·· u
5 5 ¨ 1 0 ¸
© 0 0 ¹ u a , which are free to move and whose displacements are ex-
u··6 u6
pressed in absolute coordinates, and those of the ground u g ,
and with which undergo the displacements imposed to the support. The
vector containing the displacements of all DoFs is hence:
m1 u··1 u1 – m 1 u··gy
m2 0 u·· 2 u2 – m u·· ua
2 gx u = (12.17)
I3 u·· u 0 ug
M = we obtain M 3 + K 3 =
m4 ··
u4 u4 – m 4 u··gy
The equation of motion of the system can hence be expressed
0 m5 u··5 u5 – m 5 u··gx
as (see [Cho11]):
I6 u·· u 06 6
(12.16) m mg ··
u c c g u· k kg u 0
a a a
+ + = (12.18)
Remarks T
m g m gg ··
ug T ·
u
c g c gg g
T u
k g k gg g p g ( t )
• For other cases see [Cho11] Sections 9.4 to 9.6.
where u s is the vector of the displacements of the DoFs of the If we now express the vector u s in function of the vector u g as
structure when the ground displacements u g ( t ) are applied stat-
u s = ιu g (12.24)
ically, and u is the vector of the relative displacements of the
DoFs of the structure. the so-called influence matrix ι can be computed, again mak-
The relationship between u s and u g ( t ) is given by the following ing use of the first line of Equation (12.20), i.e.:
system of equation: – k g u g = ku s = kιu g (12.25)
where p g, s is the vector of the support forces needed to impose The influence matrix ι is a N × N g matrix where N is the number
the displacements u g statically. If the system is statically deter- of DoFs of the structure and N g is the number of DoFs of the
minated, p g, s is equal to zero (See example of page 12-12). supports.
By introducing Equation (12.19) into Equation (12.18) we obtain By introducing Equation (12.26) into Equation (12.23) the final
the new system of equations: equation of motion of the system is obtained:
·· + cu· + ku = – ( mι + m )u
·· ·
g g – ( cι + c g )u g
mu (12.27) Example: 2-DoF system
Analogously, with the second line of Equation (12.21), an equa- The following 2-DoF system is subjected to multiple support
tion for the computation of the forces at the supports p g ( t ) can ground motion. Two different ground motions are applied to the
be setup and solved. degrees of freedom u 1 and u 5 . Sought is the equation of motion
of the system:
Remarks
In the case that the movement of the supports is the same at all u1
supports, u g becomes: ug = : displacements of the supports (excited, massless)
u5
u g = 1u g (12.30) (12.33)
and with ι = ι1 Equation (12.29) becomes
u2
·· + cu· + ku = – mιu·· u0 = : displacements of the supports (not excited, massless)
mu g (12.31) u6
By performig the needed calculations we obtain: The equation of motion of the system becomes:
1
--------- 0 M 0 0 0 u··3 6 0 –3 –3 u3 0
2 2 ·· 2
1 ˆ 1 0 ⋅ u 4 + ----- 0 6L 3L – 3L ⋅ u 4 0
= ----------- ⋅ k 00 = -----------4- ⋅ 4L 0 = 4L
–1 0 I 0 EI-
k 00 (12.41) ⋅ = (12.46)
k 00 16L 0 4L
2 1 0 0 0 0 u1 L3
·· – 3 3L 3 0 u1 p1 ( t )
0 --------2-
4L 0 0 0 0 u··5 – 3 – 3L 0 3 u5 p5 ( t )
– 6L 6L 1 We racall that:
--------- 0
2 2 2 2
T –1
k 0t k 00 k 0t = 2L 2L ⋅ 4L ⋅ – 6L 2L 6L 0 (12.42) ˆ EI k k g
6L 0 1
0 --------2-
2
6L 2L 0 – 6L k tt = -----3- ⋅ , ι = –k–1 kg (12.47)
T
L k g k gg
0 – 6L 4L
By performig the needed calculations we obtain:
18 0 –9 –9
T –1 2 1
k 0t k 00 k 0t = 0 2L 3L 3L (12.43) 3 --- 0
–1 1 ˆ L 6
– 9 3L 9 0 k = ------ ⋅ k = ------ ⋅ (12.48)
k EI 1
– 9 3L 0 9 0 --------2-
6L
§ 24 0 – 12 – 12 – 9 – 9 ·¸
18 0
¨
1---
ˆ EI ¨ 0 8L 2 6L – 6L – 0 2L 2 3L 3L ¸¸ 0 1
---
1
---
k tt = -----3- ⋅ ¨ (12.44) –1 L 6
3
EI – 3 – 3 2 2
L ¨ – 12 6L 12 0 – 9 3L 9 0 ¸ ι = – k k g = – ------ ⋅ ⋅ -----3- ⋅ = (12.49)
¨ ¸ EI 1 3L 3L 1 1
0 --------2- L – ------- -------
© – 12 – 6L 0 12 – 9 3L 0 9 ¹ 2L 2L
6L
6 0 –3 –3
2
ˆ EI
k tt = -----3- ⋅ 0 6L 3L – 3L (12.45)
L – 3 3L 3 0
– 3 – 3L 0 3
The vector of the effective forces becomes: 12.2 Time-history of the response of elastic systems
M
----- M
-----
As discussed in the previous sections, the equation of motion of
·· 2 2 u·· a MDoF system under base excitation is:
p eff ( t ) = – ( mι + m g )u g ( t ) = – ⋅ g1 (12.50)
I
– -------
I-
------ u··g5 ·· + Cu· + Ku = – Mιu·· ( t )
Mu g (12.52)
2L 2L
q n ( t ) : nth modal coordinate of the MDoF system ¦ m n*, eff = ¦ mn = m tot (12.60)
n=1 n=1
Further variables in Equation (12.53) are the modal mass m n*
and the modal stiffness k n* of the nth mode. These parameters where m tot is the total mass of the dynamic system.
are defined as follows: *
The effective modal height h n of the nth mode is:
T
m n* = φ n ⋅ M ⋅ φ n (12.56) θ N
* Ln θ T
k n* =
T
φn ⋅ K ⋅ φn =
2
ωn ⋅ m n* (12.57)
hn = ------ with L n =
Ln ¦ hj ⋅ mj ⋅ φjn and Ln = φ n ⋅ Mι (12.61)
j=1
ωn : nth modal circular frequency of the MDOF system • Significance of the effective modal mass m n*, eff
The modal participation factor Γ n is a measure for the contribu- The effective modal mass m n*, eff is the lumped mass of a single-
tion of the n-th mode to the total response of the system. It is de- storey substitute system which is subjected to a base shear
fined as follows: force V bn equal to the nth modal base shear force of a multi-sto-
T
rey system.
φ n Mι
Γ n = -----------------
T
- (12.58) If in addition the height of the single storey substitute system with
*
φ n Mφ n the lumped mass m n*, eff equals the modal height h n , the single-
storey system is subjected to a base moment M bn which is equal
In addition the so-called effective modal mass of the nth mode
to the nth modal base moment of the multi-storey system.
is defined as:
2
The following holds:
m n*, eff = Γ n ⋅ m n* (12.59)
N
V bn = m n*, eff ⋅ S pa, n = ¦ fjn (12.62)
j=1
s n = Γ n Mφ n (12.66) m3
m1*=2.180m
m1*Sa1
s n is independent on the normalization of the eigenvector φ n and hs
m2
we have that: m2*=0.646m
m3*=0.174m
h1*=2.50hs
h hs
N
m1 m2*Sa2
¦ sn (12.67)
h2*=0.72hs
h3*=0.48hs
= Mι m3*Sa3
n=1 hs
EI EI1* EI2* EI3*
Vb Vb1 Vb2 Vb3
Mb Mb1 Mb2 Mb3
where: This rule is often used as the standard combination method and
yields very good estimates of the total maximum response if the
Γn : modal participation factor of the n-th mode
modes of the system are well separated. If the system has sev-
S d (ω n,ζ n*) : Spectral displacement for the circular eigenfre- eral modes with similar frequencies the SRSS rule might yield
quency ω n and the modal damping rate ζ n* . estimates which are significantly lower than the actual total max-
S pa (ω n,ζ n*) : Spectral pseudo-acceleration for the circular ei- imum response.
genfrequency ω n and the modal damping rate ζ n* .
• “Complete Quadratic Combination (CQC)” Combination Rule
The contribution of the nth mode to the total displacement is:
N N
(j) (k)
u n, max = φ n ⋅ q n, max (12.69) u i, max ≈ ¦ ¦ ui, max ⋅ ρjk ⋅ ui, max (12.72)
j = 1k = 1
The maxima of different modes do not occur at the same instant.
An exact computation of the total maximum response on the ba- where
sis of the maximum modal responses is hence impossible. Dif- (j) (k)
u i, max and u i, max are the max. modal responses of modes j and k
ferent methods have been developed to estimate the total max-
imum response from the maximum modal responses. ρ jk is the correlation coefficient between nodes j and k :
12.3.2 Step-by-step procedure response spectrum for pseudo-accelerations. (The spectral dis-
placement S d (ω n,ζ n*) should be determined in the same manner)
The maximum response of a N-storey building can be estimated • Computation of the maximum displacements
according to the following procedure:
u n, max = φ n ⋅ Γ n ⋅ S d (ω n,ζ n*)
1) Determine the properties of the MDOF system
• Choose DOFs • Computation of the maximum equivalent static forces
• Determine circular eigenfrequencies ω n and eigenvectors φ n 4) Estimate the total response in terms of displacements and in-
ternal forces by means of suitable combination rules. Differ-
2
( K – ωn M ) ⋅ φn = 0 ent combination rules might be applied (ABSSUM, SRSS,
CQC).
• Compute the modal properties of the MDOF system ( M * , K * )
T T Comment
m n* = φ n Mφ n , k n* = φ n Kφ n
In order to consider the non-linear behaviour of the structure the
• Compute the modal participation factor Γ n
equivalent lateral static forces F n, max can be determined from
φ n Mι
T the spectral ordinate S pa ( ω n, ζ n, q ) of the design spectrum for
Γ n = -----------------
T
- pseudo-accelerations:
φ n Mφ n
F n, max = s n ⋅ S pa ( ω n, ζ n, q ) = Γ n Mφ n ⋅ S pa ( ω n, ζ n, q ) (12.76)
3) The maximum response of the n-th mode should be deter-
mined as described in the following. This should be done for
all modes n = 1, 2, …, N which require consideration.
• For all periods T n and for the corresponding damping ratios ζ n* ,
the spectral response S a ( ω n, ζ n ) should be determined from the
1. unit displacement u 1 = 1
2. unit displacement u 2 = 1
This 2-DoF system corresponds to the system presented in
Chapter 9 with the only difference that the 2 DoFs are swapped.
• Degrees of freedom (DoF)
Horizontal displacements u 1 and u 2 in correspondence of the
masses m 1 and m 2
• Masses
Both story masses have unit value, i.e. m 1 = m 2 = 1 , hence
• Damping
the mass matrix M is:
Damping is small and is neglected, hence the damping matrix C is:
m1 0
M = = 1 0 (12.77) C = 0 0 (12.79)
0 m2 0 1 0 0
or
3– 5
4 2 1 – ---------------- –1 φ 11
ω – 3ω + 1 = 0 (12.84) 2 ⋅ = 0 (12.88)
3– 5 φ 21 0
The solution of the quadratic equation yield the eigenvalues: –1 2 – ----------------
2
2 3− + 9 – 4- = 3 −
+ 5-
ω = ------------------------ --------------- (12.85) The first row yields following equation:
2 2
2 –( 3 – 5 )
--------------------------- φ 11 – 1φ 21 = 0 (12.89)
2
Normalizing the largest coordinate of the eigenvector to unity • Higher mode of vibration
( φ 11 = 1 ), φ 21 becomes: 2 3+ 5
The largest eigenvalue ω 2 = ---------------- leads to the
2
2 –( 3 – 5 )
--------------------------- – φ 21 = 0 (12.90) 3+ 5
2 2. circular natural frequency ω 2 = ---------------- = 1.62 (12.93)
2
or
In analogy to the fundamental mode, the second mode of vibra-
2
5–1 tion can be computed introducing the second eigenvalue ω 2 into
φ 21 = ---------------- = 0.62 (12.91)
2 the system of equations:
Hence the first natural mode of vibration is:
3+ 5
1 – ---------------- –1 φ 12
2 ⋅ = 0 (12.94)
3+ 5 φ 22 0
–1 2 – ----------------
2
2 –( 3 + 5 )
---------------------------- φ 12 – 1 = 0 (12.96)
2
or:
–2 1– 5
φ 12 = ---------------- = ---------------- = – 0.62 (12.97)
1+ 5 2
Hence the second natural mode of vibration is: • Orthogonality with respect to the mass matrix
The modal mass matrix M* is:
5 – 1- 5–1
1 --------------- 1 – ----------------
T 2 1 0 2
M* = Φ MΦ = ⋅ ⋅
5–1 0 1 5 – 1-
– ---------------- 1 --------------- 1
φ 12 1– 5
- = – 0.62
2 2
= ------------
2 (12.98)
φ 22 1 5 – 1- 5–1
1 1 --------------- 1 – ----------------
= 2 ⋅ 2
5–1 5–1
– ---------------- 1 ---------------- 1
2 2
2
1 + § ----------------·
5–1
0
© 2 ¹
3) Orthogonality of modes = = 1.38 0
2 0 1.38
1 + § ----------------·
5–1
In the following the orthogonality of the modes of vibration 0
© 2 ¹
should be checked.
(12.100)
Hence, following matrix of the eigenvectors is needed:
i.e. the matrix M* is diagonal.
5–1
φ 11 φ 12 1 – ----------------
Φ = = 2 (12.99)
φ 21 φ 22 5 – 1-
--------------- 1
2
The influence vector ι represents the dis- • Additional important modal quantities
placement of the masses resulting from the
The modal participation factor Γ n is defined as:
static application of a unit ground displace-
ment u g = 1 : Ln
Γ n = ---------- (12.114)
mn *
ι = 1 (12.109)
1 and substituting L n and m n * into this definition gives following
values for Γ 1 and Γ 2 :
The substitution of L into the equation of
motion in modal coordinates L1 1.62
Γ 1 = ---------- = ---------- = 1.17 (12.115)
m1 * 1.38
·· + K* ⋅ q = – L ⋅ u·· ( t )
M* ⋅ q (12.110)
g
L2 0.382
leads to: Γ 2 = ---------- = ------------- = 0.28 (12.116)
m2 * 1.38
··
1.38 0 ⋅ q 1 + 0.528 0 ⋅ q 1 = – 1.62 ⋅ u·· ( t ) The effective modal mass is defined as:
g
0 1.38 q··2 0 3.618 q2 0.382 2
m n*, eff = Γ n ⋅ m n* (12.117)
(12.111)
and substituting Γ n and m n * into this definition gives following
Checking the circular natural frequency computed using modal
values for m 1*, eff and m 2*, eff :
coordinates against the results of Section 12.4.2:
2
k 11 * m 1*, eff = Γ 1 ⋅ m 1* = 1.17 2 ⋅ 1.38 = 1.894 (12.118)
ω1 = 0.528
------------ = ------------- = 0.62 OK! (12.112)
m 11 * 1.38 2
m 2*, eff = Γ 2 ⋅ m 2* = 0.28 2 ⋅ 1.38 = 0.106 (12.119)
k 22 * 3.618
ω2 = ------------ = ------------- = 1.62 OK! (12.113) m 1, eff + m 2*, eff = 1.894 + 0.106 = 2.000 OK! (12.120)
m 22 * 1.38
12.4.4 Response spectrum method 2) Natural modes of vibration and natural frequencies
The 2-DoF system analysed in the previous Sections shall be The results of the previous Sections computed using unit mass-
used to illustrate the response spectrum method. For this reason es and unit stiffnesses shall be multiplied by the factor:
real masses and stiffnesses shall be assumed. The seismic ac-
2
tion on the 2-DoF system is represented by the elastic response k 100kg ⁄ s –1
(12.123)
---- = ------------------------ = 100s
spectrum of the “El Centro” earthquake. m 1kg
1) Model • Fundamental mode
Similar to Section 12.4.1, –1
Natural frequency: ω 1 = 0.62 ⋅ 100s = 6.2 Hz
however with a new definition
of masses and stiffnesses: 2π 2π
Natural period: T 1 = ------ = ---------------- = 1.02 s
m 1 = m 2 = 1kg ω1 6.2 Hz
The stiffness chosen for each story is k 1 = k 2 = k = 100 N/m The eigenvectors are dimensionless quantities and remain un-
and an appropriate units transformation leads to: changed:
The spectral values given by the elastic acceleration response 5) Inverse transformation
spectrum of the “El Centro” earthquake for the periods T 1 and
The peak deformations and internal forces belonging to each mode of vibra-
T 2 are: tion in the original reference system are obtained by multiplying the relevant
eigenvector with the corresponding peak value of the modal coordinate.
S pa1 = 4.25 m/s 2 and (12.132)
• Fundamental mode
S pa2 = 7.34 m/s 2 (12.133)
(1) 1 = 130 mm
u max = q 1, max ⋅ φ 1 = 0.130m ⋅ (12.136)
10 0.62 81
ζ = 5%
Pseudo acceleration [m/s2]
(1) (1)
Spa2 = 7.34 m/s2 f max = K ⋅ u max = 100 – 100 ⋅ 0.130 = 13.0 – 8.1 = 4.9 N
100 200 0.081 – 13.0 + 16.2 3.2
(12.137)
Alternatively (allow an approximate consideration of nonlinearities):
5 S 2
pa1 = 4.25 m/s
s 1 = Γ 1 Mφ 1 = 1.17 ⋅ 1 0 ⋅ 1 = 1.17 (12.138)
0 1 0.62 0.725
(1)
f max = s 1 ⋅ S pa1 = 4.25 1.17 = 4.9 N (12.139)
0.725 3.2
T2 = 0.39 s T1 = 1.02 s
0
0.01 0.10 1.00 10.0
Period [s]
1.62kg 1
q 1, max = ----------------- ⋅ ---------------------2- ⋅ 4.25 m/s 2 = 0.130m (12.134)
1.38kg ( 6.2Hz )
0.38kg 1
q 2, max = ----------------- ⋅ ------------------------2- ⋅ 7.34 m/s 2 = 0.008m (12.135)
1.38kg ( 16.2Hz )
(2) The total peak response is obtained from the peak response of
u max = q 2, max ⋅ φ 2 = 0.008m ⋅ – 0.62 = – 5.0 mm (12.140) the single vibration modes using e.g. the SRSS combination rule
1 8.0
(SRSS = Square Root of the Sum of Squares).
(2) (2) • Peak displacements
f max = K ⋅ u max = 100 – 100 ⋅ – 0.005 = – 0.5 – 0.8 = – 1.3 N
100 200 0.008 0.5 + 1.6 2.1
2
(12.141) (n) 2 2 2
u 1, max = ¦ ( u1 ) = ( 130mm ) + ( – 5 mm ) = 130mm
Alternatively (allow an approximate consideration of nonlinearities): k=1
(12.144)
s 2 = Γ 2 Mφ 2 = 0.28 ⋅ 1 0 ⋅ – 0.62 = – 0.173 (12.142)
01 1 0.28 2
(n) 2 2 2
u 2, max = ¦ ( u2 ) = ( 81mm ) + ( 8mm ) = 81mm
(1)
f max = s 2 ⋅ S pa2 = 7.34 – 0.173 = – 1.3 N (12.143) k=1
0.28 2.1 (12.145)
In this case the total peak displacements are almost identical to
the peak displacements of the fundamental mode. The relatively
small contributions due to the second vibration mode basically
disappear because of the SRSS combination rule.
• Peak sectional forces (Shear force V) 12.4.5 Response spectrum method vs. time-history analysis
Upper shear force: 1) Model
2 2
V 1, max = ( 4.9N ) + ( – 1.3N ) = 5.1N
(12.146)
Lower shear force:
2 2
V 2, max = ( 8.1N ) + ( 0.8N ) = 8.1N
(12.147)
2) Results • Demand
• Note that in this case the eigenvectors are normalized to yield Time-history: V = 5.69N Time-history: V = 1.51N
unit displacement at the top of the second story. Therefore, the
eigenvectors and the participation factors of case study 1 differ Lower shear force: Lower shear force:
from the values obtained in previous sections.
SRSS: V = 8.05N SRSS: V = 4.40N
Displacement [cm]
10 10
0 0
−10 −10
−20 −20
0 5 10 15 20 0 5 10 15 20
20 20
Second mode Second mode
Displacement [cm]
Displacement [cm]
10 10
0 0
−10 −10
−20 −20
0 5 10 15 20 0 5 10 15 20
20 20
Sum Sum
Displacement [cm]
Displacement [cm]
10 10
0 0
−10 −10
• Time-histories: Summary
Blank page
20
Case study 1
10
Displacement [cm]
20
Case study 2
10
Displacement [cm]
a) People-induced vibrations
13.1 Introduction
- Pedestrian bridges
There are more and more vibration problems in structures - Floors with walking people
because: - Floors for sport or dance activities
• Higher quality materials with higher exploitation - Floor with fixed seating and spectator galleries
- slender constructions - High-diving platforms
- smaller stiffnesses and masses
b) Machinery-induced vibrations
• More intensive dynamic excitations
- Machine foundations and supports
• Increased sensitivity of people - Bell towers
- Structure-borne sound
- Ground-transmitted vibrations
Nevertheless vibration sensitive structures are
often designed for static loads only c) Wind-induced vibrations
- Buildings
Goal of this chapter - Towers, chimneys and masts
- Bridges
• Give an overview of possible causes of vibration problems in
buildings and of potential countermeasures - Cantilevered roofs
• Description of practical cases with vibration rehabilitation d) Vibrations induced by traffic and construction activity
- Roads and bridges
- Railways
- Construction works
[BB88] Baumann K., Bachmann H.: “Durch Menschen verursachte • Frequency tuning
dynamische Lasten und deren Auswirkungen auf Balkentrag-
werke”. IBK Bericht Nr. 7501-3, 1988. 10
0.00
[BW95] Bachmann H., Weber B.: “Tuned Vibration Absorbers for 9
0.01
Damping of Lively Structures”. Structural Engineering Inter- 8
national, No. 1, 1995. ζ = 0.05
Trrasmissibility TR
7
[SIA06] SIA: “Wind – Kommentar zum Kapitel 6 der Normen SIA 261 6
und 261/1 (2003) Einwirkungen auf Tragwerke”. SIA Doku-
mentation D0188. Zürich 2006. 5
0.10
4
[SIA03] SIA: “Aktuelle Probleme der Brückendynamik”. SIA Doku-
mentation D0198. Zürich 2003. 3
0.20
[SIA96] SIA: “Dynamische Probleme bei Brücken- und Hochbauten”. 2
0.50
SIA Dokumentation D0138. Zürich 1996. 1
0.70 2
[VB87] Vogt R., Bachmann H.: “Dynamische Kräfte beim Klatschen, 0
Fussstampfen und Wippen”. IBK Bericht Nr. 7501-4, 1987. 0 0.5 1 1.5 2 2.5 3 3.5 4
ω/ωn
The amplitude of the nth harmonic component of a force gener- 13.3 People induced vibrations
ated by people excitation is proportional to the mass of the per-
son ( F o = G ⋅ α n = g ⋅ M ⋅ α n , see Equation (13.8)). 13.3.1 Excitation forces
2 g ⋅ M ⋅ αn 1 k g ⋅ M ⋅ αn 1 M g ⋅ αn In Chapter 6 “Forced Vibrations” it has been already mentioned
u··max = ω ⋅ ----------------------- ⋅ ------ = ---- ⋅ ----------------------- ⋅ ------ = ----- ⋅ ------------- (13.5)
k 2ζ m k 2ζ m 2ζ that excitation due to people, like e.g. walking, running, jumping,
and so on, can be represented as Fourier-series:
M g ⋅ αn
u··max = ----- ⋅ ------------- (13.6)
m 2ζ ∞
F ( t ) = a0 + ¦ [ a n cos ( nω 0 t ) + b n sin ( nω 0 t ) ] (13.7)
• Remarks
n=1
- A soft structure is more prone to vibration than a rigid one.
Equation (13.7) can also be represented in a form according to
See Equations (13.2) to (13.4).
Equation (13.8):
- The acceleration amplitude is directly proportional on the
N
ratio of the mass of the people to the building mass.
F(t) = G + ¦ G ⋅ α n ⋅ sin ( n ⋅ 2πf 0 ⋅ t – φ n ) (13.8)
n=1
Where:
- G = Weight of the person
- α n = Fourier coefficient for the nth harmonic
- G ⋅ α n = Amplitude of the nth harmonic of the excitation force
- f 0 = Step frequency of the excitation force
- φ n = Phase shift of the nth harmonic ( φ 1 = 0 )
- n = Number of the nth harmonic
- N = Number of considered harmonics
The steady-state response of a SDoF system under periodic ex- • Jumping (left) and walking (right), see [BB88]
citation can be computed in analogy to Chapter 6 as:
N
u ( t ) = u0 ( t ) + ¦ un ( t ) (13.9)
n=1
Where
G
u 0 ( t ) = ---- (Static displacement) (13.10)
k
2
G ⋅ α n ( 1 – β n ) sin ( nω 0 t – φ n ) – 2ζβ n cos ( nω 0 t – φ n )
u n ( t ) = --------------- ⋅ ---------------------------------------------------------------------------------------------------------------
- (13.11)
k 2 2 2
( 1 – β n ) + ( 2ζβ n ) • Clapping, foot stomping and rocking, see [VB87]
nω 0
ω 0 = 2πf 0 , β n = --------- (13.12)
ωn
• Walking (see [Bac+97] Figure G.1) • Jumping (see [Bac+97] Figure G.2)
• Clapping (see [Bac+97] Figure G.3) • Jumping: Fourier amplitude spectrum (see [Bac+97] Figure G.2)
13.3.2 Example: Jumping on an RC beam Contact time: t p = 0.16s (phase angle computation)
(13.15)
Here the same example as in Section 6.1.3 is considered again:
Weight of the person: G = 0.70kN (13.16)
• Results
Excel Table: u max = 0.043m (13.17)
F 1 1.8 ⋅ 0.70 1
Equation (13.2): u max = --- ⋅ ------ = ---------------------- ⋅ --------------------- = 0.042m
k 2ζ 886 2 ⋅ 0.017
(13.18)
• Remarks
- Shape of the excitation “similar” as half-sine
- Maximum deflection very close to the solution obtained by
means of the half-sine function
1.5
- Horizontal transverse: Avoid natural frequencies be-
1.0
tween 0.7 and 1.3Hz (absolutely safe: fht,1 > 3.4Hz).
0.5
0.0
00
- Horizontal longitudinal: Avoid natural frequencies be-
-0.5
tween 1.6 and 2.4Hz.
-1.0 • Amplitude limitation
-1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
- Calculation of the acceleration maximum amplitude.
Time (s)
2
a max < ca. 0.5m ⁄ s = 5% g (13.19)
• Response
• Special features of the amplitude limitation
0.0500
Static component (n=0)
0.0400 First harmonics (n=1) - When walking or running, the effectiveness of people is lim-
Second harmonics (n=2)
0.0300 Third harmonics (n=3)
ited, because the forces are not always applied at midspan;
Total (3 harmonics)
[m]
0.0200
System
Displacement
0.0100
0.0000
-0.0100
-0.0200
0 0200
D
-0.0300 Equivalent
-0.0400 SDoF system
-0.0500
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time (s)
ζ = 0.01
abs(uj) / umax
0.6
0.4
0.2
0
0 5 10 15 20 25 30 35 40 45 50
Cycle
- Not all people walk in the step (Exception: Lateral vibra- The acceleration a max given in Equation (13.20) is the accelera-
tions → Synchronisation effect) tion generated by one person crossing the footbridge. If n people
To take into account the specificities of the amplitude limitation, are on the bridge at the same time, the maximum acceleration is
sophisticated methods are available. From [Bac+97] the follow- typically less than n ⋅ a max because not all people walk in step
ing one is adopted: across the bridge.
2 2 2 .
a max = 4π ⋅ f ⋅ y ⋅ α ⋅ Φ [ m ⁄ s ] (13.20)
Where: The square root of the number of people is often
- y : Static deflection at half the span chosen as the multiplication factor, i.e. n ⋅ a max
- α : Fourier coefficient
- Φ : dynamic amplification factor
G ⋅ α1 1 1 ⋅ 0.4 1
u max,1 = --------------- ⋅ ------ = --------------- ⋅ --------------------- = 0.0133m = 1.33cm
K n 2ζ 886 2 ⋅ 0.017
(13.22)
u max = 0.11 + 1.33 = 1.44cm (13.23)
Discretisation for FE Analysis
2 2 2
a max = ω u max,1 = ( 2π ⋅ 2 ) ⋅ 0.0133 = 2.10m ⁄ s (13.24)
• Estimate of the maximum displacement and acceleration using
the improved method
The maximum acceleration is computed by means of Equation
(13.20) as follows:
-0.5
-1
-1.5
0 5 10 15 20 25 30 35 40
Time [s]
acceleration am [m/s2]
1.5
ce / Static weight [-]
F2
1.2
1
F3
1
F15 0.5
0.8 F28 0
0.6
-0.5
Forc
Vertical a
0.4
-1
0.2
-1.5
0
-2
0 2 4 6 8 10 12 14 16
0 5 10 15 20 25 30 35 40
Time [s]
Time [s]
- Position (standing, sitting, lying) • G2 limit for vibrations parallel to the spinal column
- Direction of the action compared to the spinal column
- Activity (resting, walking, running, ...)
- Type of vibration
- ....
Frequency 1 to 10 Hz Frequency 10 to 100 Hz
Description
amax [m/s2] vmax [m/s]
Barely noticeable 0.034 0.0005
Clearly noticeable 0.1 0.0013
Disturbing 0.55 0.0068
Not tolerable 1.8 0.0138
Vertical harmonic vibration action on a standing person. Accepted averages; scat-
ters up to a factor of 2 is possible (from [Bac+97])
T
--1- ³ a ( t ) dt
2
a eff = (13.31)
T 0
- Walking: α 1 = 0.4 , α 2 = 0.1 , α 3 = 0.1 - Limits depend on the activity, if e.g. people are sitting in the
dance hall, as well, this limit shall be reduced.
- Running: α 1 = 1.6 , α 2 = 0.7 , α 3 = 0.2
- Because of the large forces that can be generated through
- Jumping: α 1 = 1.9 , α 2 = 1.6 , α 3 = 1.1
these activities, the dynamic characteristics of the structure
- Dancing: α 1 = 0.5 , α 2 = 0.15 , α 3 = 0.1 shall be estimated as precisely as possible.
(however: a - many people moving rhythmically. b - certain
dances are very similar to jumping) 13.3.6 Concert halls, stands and diving platforms
• Frequency tuning
See [Bac+97].
- If the excitation is generated through jumping ( f max ≅ 3.4Hz )
or dancing ( f max ≅ 3.0Hz ), then the following natural frequen- 13.4 Machinery induced vibrations
cies shall be exceeded:
Gyms Dance halls It is not possible to carry out here a detailed treatment of machin-
Construction Natural frequency Natural frequency ery induced vibrations. Therefore, reference to [Bac+97] is
[Hz] [Hz] made.
Reinforced concrete > 7.5 > 6.5
Prestressed concrete > 8.0 > 7.0
Composite structures > 8.5 > 7.5
Steel > 9.0 > 8.0
13.5 Wind induced vibrations • Vortex shedding: Periodic action transversely to the wind di-
rection
Wind-induced vibrations cover a challenging and wide area. It is
not possible to carry out here their detailed treatment. Therefore, - Vortex are not shedded left and right at the same time. If the
reference is made to the relevant literature: time-interval of the vortex shedding is equal to the oscilla-
tion period of the structure, resonance excitation occurs.
• [Bac+97]
• Simiu E., Scanlan R.H.: “Wind Effects on Structures”. Third
Edition. John Wiley & Sons, 1996.
• References
[BW95] Bachmann H., Weber B.: “Tuned Vibration Absorbers for
Damping of Lively Structures”. Structural Engineering Inter-
national, No. 1, 1995.
[Den85] Den Hartog J.P.: “Mechanical Vibrations”. ISBN 0-486-
64785-4. Dover Publications,1985. (Reprint of the original
fourth edition of 1956)
The system of equation can be easily solved using “Maple”, and Thus, the norm of the dynamic amplification function U H ⁄ U H0
we obtain the following expression for the amplification function can be easily calculated:
U H ⁄ U H0 : 2 2 2 2
UH = ( β – Ω ) + ( 2Ωβζ T )
2 2 --------
- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
-2
( β – Ω ) + 2iΩβζ T 2 2 2 2 2 2 2 2 2 2 2
UH U H0 [ ( β – Ω ) – Ω β ( 1 – γ ) + Ω ( Ω – 4βζ H ζ T ) ] + 4 [ ( β – Ω )ζ H + ( 1 – Ω – Ω γ )βζ T ]
--------
U H0
- = ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2 2 2 2 2 2 2 2 2 2
- (13.38)
[ ( β – Ω ) – Ω β ( 1 – γ ) + Ω ( Ω – 4βζ H ζ T ) ] + 2i [ ( β – Ω )ζ H + ( 1 – Ω – Ω γ )βζ T ]
(13.44)
The complex expression given in Equation (13.38) shall now be A similar procedure can be followed to compute the dynamic am-
converted into the form: plification function U T ⁄ U H0 .
z = x + iy or U H = U H0 ( x + iy ) (13.39) Next figure show a representation of Equation (13.44) in function
of Ω for an undamped structure ζ H = 0 . Curves for different val-
The displacement U H has therefore two components: 1) One
ues of the parameters β , γ and ζ T are provided.
that is in phase with the displacement U H0 and 2) one with a
phase shift equal to π ⁄ 4 . From the vectorial sum of these two 16
components the norm of U H can be computed as: ]T=0 ]T=inf.
U H = U H0 x + y
2 2
(13.40) J = 1/20
12
E=1
Equation (13.38) has however the form
UH/UH0 [-]
P
( A + iB ) 8
UH = U H0 --------------------- (13.41)
( C + iD ) ]T=0.32
and must be first rearranged as follows:
4
( A + iB ) ⋅ ( C – iD ) ( AC + BD ) + i ( BC – AD ) ]T=0.10 Q
UH = U H0 ----------------------------------------------- = U H0 ----------------------------------------------------------------
-
( C + iD ) ⋅ ( C – iD ) C +D
2 2
(13.42) 0
2 2 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
A +B : [-]
U H = U H0 -------------------
2 2
(13.43)
C +D
13.6.3 Optimum TMD parameters • Amplification function with TMD: Variation of TMD frequency
Based on observations and consideration at the previous image 50
No TMD (ζT=inf.)
Den Hartog found optimum TMD parameters for an undamped
γ = 0.01
structure: 40
ζΗ = 0.01
fH fH 1
f T, opt = ----------------------------- = ----------- or β opt = ----------- (13.45)
UH/UH0 [-]
30
1 + mT ⁄ mH 1+γ 1+γ
20
3m T ⁄ m H 3γ -
ζ T, opt = --------------------------------------- = --------------------- (13.46) fT=1.02f
1.02fT,opt
T opt
fT=0.98fT,opt
3 3
8 ( 1 + mT ⁄ mH ) 8(1 + γ) 10
fT=fT,opt
• TMDs are most effective when the damping of the structure is low
UH/UH0 [-]
30
10 ζT=ζT,opt
• Meaningful mass ratios γ are 3-5%
• The exact tuning of the TMD occurs experimentally, therefore 0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
great care should be paid to construction details.
Ω [-]
14 Pedestrian Footbridge with TMD Page 14-1 14 Pedestrian Footbridge with TMD Page 14-2
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
MT
Mass ratio: μ = -------- = 0.0585 = 5.85%
MH
fH
Natural frequency: f opt = ------------ = 1.92Hz
1+μ
(Measured: f T = 1.91Hz )
2
Stiffness: K T = M T ⋅ ( 2πf opt ) = 50.9kN/m
3μ
Damping rate ζ opt =
----------------------- = 0.14 = 14%
8( 1 + μ )3
(Measured: ζ T = 13% )
14 Pedestrian Footbridge with TMD Page 14-3 14 Pedestrian Footbridge with TMD Page 14-4
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
14.3 Test programme 14.4 Free decay test with locked TMD
Following tests are carried out: Time history of the displacement at midspan
No. Test Action location TMD
40
14 Pedestrian Footbridge with TMD Page 14-5 14 Pedestrian Footbridge with TMD Page 14-6
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
1.5
In order to excite the different modes of vibration of the foot-
bridge, the test is repeated several times changing the position
1.0 of the impact of the sandbag on the bridge. The considered lo-
cations are:
0.5 - at midspan (Section 14.5.1)
- at quarter-point of the span (Section 14.5.2).
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 These tests are carried out with locked TMD. In order to investi-
Frequency [Hz] gate the effect of the TMD on the vibrations of the system, the
Figure 14.4: Free decay test with locked TMD: Fourier-spectrum of the test of Section 14.5.1 is repeated with free TMD (see Section
displacement at midspan. 14.5.3).
14 Pedestrian Footbridge with TMD Page 14-7 14 Pedestrian Footbridge with TMD Page 14-8
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
1.0 1.0
0.0 0.0
-0.5 -0.5
-1.0 -1.0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time [s] Time [s]
Figure 14.5: Sandbag test with locked TMD: Acceleration at midspan. Figure 14.7: Sandbag test with locked TMD: Acceleration at quarter-
point of the span.
0.040 0.040
f1=2.00Hz
"Spectral acceleration"
"Spectral acceleration"
0.030 0.030
f1=2.00Hz
0.020 0.020
0.000 0.000
0 5 10 15 20 0 5 10 15 20
Frequency [Hz] Frequency [Hz]
Figure 14.6: Sandbag test with locked TMD: Fourier-spectrum of the Figure 14.8: Sandbag test with locked TMD: Fourier-spectrum of the
acceleration at midspan. acceleration at quarter-point of the span.
14 Pedestrian Footbridge with TMD Page 14-9 14 Pedestrian Footbridge with TMD Page 14-10
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
• With the sandbag test in principle all frequencies can be excit- 1.0
"Spectral acceleration"
0.030
ured first natural frequency f 1 = 2.0Hz in good agreement with f1=2.00Hz
the computation provided in Section 14.2.1.
0.020
0.010 f3=18.31Hz
0.000
0 5 10 15 20
Frequency [Hz]
Figure 14.10: Sandbag test with locked TMD: Fourier-spectrum of the
acceleration at midspan.
14 Pedestrian Footbridge with TMD Page 14-11 14 Pedestrian Footbridge with TMD Page 14-12
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
Remarks
0.040
"Spectral acceleration"
0.030 f2=8.84Hz
0.020 f1=2.00Hz
0.010 f3=18.31Hz
0.000
0 5 10 15 20
Frequency [Hz]
Figure 14.12: Sandbag test with locked TMD: Fourier-spectrum of the
acceleration at quarter-point of the span.
14 Pedestrian Footbridge with TMD Page 14-13 14 Pedestrian Footbridge with TMD Page 14-14
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
1.0 • With active (free) TMD the “first” and the “third “natural fre-
Acceleration at midspan [m/s2]
0.020
“f1=2.05Hz”
“f3=19.04Hz”
0.010
0.000
0 5 10 15 20
Frequency [Hz]
Figure 14.14: Sandbag test with free TMD: Fourier-spectrum of the
acceleration at midspan.
14 Pedestrian Footbridge with TMD Page 14-15 14 Pedestrian Footbridge with TMD Page 14-16
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
Remarks 0.0
14 Pedestrian Footbridge with TMD Page 14-17 14 Pedestrian Footbridge with TMD Page 14-18
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
14 Pedestrian Footbridge with TMD Page 14-19 14 Pedestrian Footbridge with TMD Page 14-20
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
5 5
0 0
-5 -5
-10 -10
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Time [s] Time [s]
Figure 14.18: One person walking with 2 Hz: Displacement at midspan Figure 14.20: One person walking with 2 Hz: Displacement at midspan
with locked TMD. with locked TMD. (ABAQUS-Simulation).
2.0 2.0
Acceleration at midspan [m/s2]
14 Pedestrian Footbridge with TMD Page 14-21 14 Pedestrian Footbridge with TMD Page 14-22
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
The curves in Figures 14.20 and 14.21 were computed using the 14.7.3 Free TMD
FE program ABAQUS. A similar calculation is described in detail
14 Pedestrian Footbridge with TMD Page 14-23 14 Pedestrian Footbridge with TMD Page 14-24
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
14.7.4 Remarks about “One person walking with 2 Hz” The results of the experiments with several people walking on
the bridge are commented by using the results of tests with one
• The effect of the TMD can be easily seen in Figures 14.22 and person walking (see Section 14.7) as comparison. For this rea-
14.23. The maximum acceleration at midspan reduces from son the maximum vibration amplitudes shown in Figures 14.18,
1.63m ⁄ s 2 to 0.34m ⁄ s 2 , which corresponds to a permissible 14.19, 14.22, 14.23 and 14.24 to 14.27 are summarised in Ta-
value. bles 14.1 and 14.2.
14 Pedestrian Footbridge with TMD Page 14-25 14 Pedestrian Footbridge with TMD Page 14-26
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
Case Group 1 person ratio of the displacement is larger that the amplification factor of the
accelerations, because the static deflection caused by the
Maximum displacement at
20.52 mm 12.04 mm 1.70 group is significantly larger than that caused by the single per-
midspan. Locked TMD
son.
Maximum displacement at
12.28 mm 3.27 mm 3.76 • The activation of the TMD results in a reduction of the maxi-
midspan. Free TMD
mum acceleration caused by the single person by a factor of
Ratio 1.67 3.68
4.79. In the case of the group the reduction factor is only 2.14.
Table 14.2: Comparison of the displacements at midspan. It should be noted here that when the TMD is active (free), the
vibrations are significantly smaller, and therefore it is much
It is further assumed that only about 16 of the 24 persons are on
easier for the group to walk in step. It is therefore to be as-
the footbridge at the same time.
sumed that in the case of the free TMD, the action was strong-
The following remarks can thereby be made: er than in the case of the locked TMD. This could explain the
seemingly minor effectiveness of the TMD in the case of the
• The maximum acceleration measured at midspan of the bridge
group.
with locked TMD is only about 1.26-times greater than the ac-
celeration which has been generated by the single person. Ac-
cording to section 13.3.3 we could have expected a larger
acceleration from the group ( 16 = 4 ). One reason why the
maximum acceleration is still relatively small, is the difficulty to
walk in the step when the “ground is unsteady.” With a little
more practice, the group could probably have achieved much
larger accelerations. It is further to note that the person who
walked of the bridge for the test presented in Section 14.7 was
with his 95 kg probably much heavier than the average of the
group.
• The maximum displacement measured at midspan of the
bridge with locked TMD is 1.70 times larger than the displace-
ment generated by the single person. The amplification factor
14 Pedestrian Footbridge with TMD Page 14-27 14 Pedestrian Footbridge with TMD Page 14-28
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
0 0
-10 -10
-20 -20
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Time [s] Time [s]
Figure 14.24: Group walking with 2 Hz: Displacement at midspan with Figure 14.26: Group walking with 2 Hz: Displacement at midspan with
locked TMD. free TMD.
Acceleration at midspan [m/s2]
1.0 1.0
0.0 0.0
-1.0 -1.0
-2.0 -2.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
time [s] Time [s]
Figure 14.25: Group walking with 2 Hz: Acceleration at midspan with Figure 14.27: Group walking with 2 Hz: Acceleration at midspan with
locked TMD. free TMD.
14 Pedestrian Footbridge with TMD Page 14-29 14 Pedestrian Footbridge with TMD Page 14-30
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
14 Pedestrian Footbridge with TMD Page 14-31 14 Pedestrian Footbridge with TMD Page 14-32
Course “Fundamentals of Structural Dynamics” An-Najah 2013 Course “Fundamentals of Structural Dynamics” An-Najah 2013
60
40 Maximum dyn. displ.: d dyn,max = 0.82 ⋅ 1.8 ⋅ 5.5 = 8.12mm
20
Maximum displacement: d max = 8.12 + 0.82 = 8.94mm
0 (Measured: d max = 8.12mm )
-20
2 2
Maximum acceleration: a max = 4π ⋅ 1.95 ⋅ 0.0008 ⋅ 1.8 ⋅ 5.5
-40
= 1.22m ⁄ s 2
-60
0 5 10 15 20 25 30 35 40 (Measured: a max = 1.04m ⁄ s 2 )
Time [s]
Figure 14.30: One person jumping with 2 Hz: Displacement at midspan 14.9.3 Remarks about “One person jumping with 2 Hz”
with free TMD.
• When jumping, the footbridge can be much strongly excited
8.0 than when walking.
Acceleration at midspan [m/s2]
6.0
4.0 • The achieved acceleration a max = 7.18m ⁄ s 2 = 73% g is very
2.0 large and two jumping people could easily produce the lift-off
0.0 of the footbridge.
-2.0 • The effect of the TMD can be easily seen in Figures 14.30 and
-4.0 14.31. The maximum acceleration at midspan reduces from
-6.0 7.18m ⁄ s 2 to 1.04m ⁄ s 2 , what, however, is still perceived as un-
-8.0 pleasant.
0 5 10 15 20 25 30 35 40
Time [s]
Figure 14.31: One person jumping with 2 Hz: Acceleration at midspan
with free TMD.
14 Pedestrian Footbridge with TMD Page 14-33 14 Pedestrian Footbridge with TMD Page 14-34