scieee Science in your language
[en] (orig)

Parameter identification of large-scale magnetorheological dampers in a benchmark building

Abstract

Magnetorheological (MR) dampers are devices that can be used for vibration reduction in structures. However, to use these devices in an effective way,a precise modeling is required. In this sense, in this paper we consider a modified parameter identification method of large-scale magnetorheological dampers which are represented using the normalized Bouc-Wen model. The main benefit of the proposed identification algorithm is the accuracy of the parameter estimation. The validation of the parameter identification method has been carried out using a black box model of an MR damper in a smart base-isolated benchmark building. Magnetorheological dampers are used in this numerical platform both as isolation bearings as well as semiactive control devices.

Read accessible full text

Parameter identification of large-scale magnetorheological dampers in a benchmark building

Author: Bahar, Arash,Pozo Montero, Francesc,Acho Zuppa, Leonardo,Rodellar Benedé, José,Barbat Barbat, Horia Alejandro
Year: 2010
DOI: 10.1016/j.compstruc.2009.10.002
Source: https://upcommons.upc.edu/bitstream/2117/6413/3/bahar_et_al.pdf
Pa ame e iden i ica ion o la ge-scale
magne o heological dampe s in a benchma k building✩
A ash Baha a, F ancesc Pozo∗,b, Leona do Achob, Jos´e Rodella a, Alex
Ba ba c
aCoDAlab, Depa amen de Ma em`a ica Aplicada III, Escola T`ecnica Supe io
d’Enginye s de Camins, Canals i Po s de Ba celona (ETSECCPB), Uni e si a
Poli `ecnica de Ca alunya (UPC), Jo di Gi ona, 1-3, 08034 Ba celona. Spain.
bCoDAlab, Depa amen de Ma em`a ica Aplicada III, Escola Uni e si `a ia d’Enginye ia
T`ecnica Indus ial de Ba celona (EUETIB), Uni e si a Poli `ecnica de Ca alunya,
Com e d’U gell, 187, 08036 Ba celona, Spain.
cCIMNE, Depa amen de Resis `encia de Ma e ials i Es uc u es a l’Enginye ia, Escola
T`ecnica Supe io d’Enginye s de Camins, Canals i Po s de Ba celona, Uni e si a
Poli `ecnica de Ca alunya, Jo di Gi ona, 1-3, 08034 Ba celona, Spain.
Abs ac
Magne o heological (MR) dampe s a e de ices ha can be used o ib a ion
educ ion in s uc u es. Howe e , o use hese de ices in an e ec i e way,
a p ecise modeling is equi ed. In his sense, in his pape we conside a
modi ied pa ame e iden i ica ion me hod o la ge-scale magne o heological
dampe s which a e ep esen ed using he no malized Bouc-Wen model. The
main bene i o he p oposed iden i ica ion algo i hm is he accu acy o he
pa ame e es ima ion. The alida ion o he pa ame e iden i ica ion me hod
has been ca ied ou using a black box model o an MR dampe in a sma
base-isola ed benchma k building. Magne o heological dampe s a e used in
his nume ical pla o m bo h as isola ion bea ings as well as semiac i e con-
ol de ices.
✩Suppo ed by CICYT (Spanish Minis y o Science and Inno a ion) h ough g an
DPI2008-06463-C02-01.
∗Co esponding au ho
Email add esses: [email p o ec ed] (A ash Baha ), [email p o ec ed]
(F ancesc Pozo), leona [email protected] (Leona do Acho), [email p o ec ed]
(Jos´e Rodella ), [email p o ec ed] (Alex Ba ba )
URL: h p://www-ma3.upc.es/codalab (F ancesc Pozo),
h p://www.cimne.upc.es (Alex Ba ba )
P ep in submi ed o Else ie July 30, 2009
Key wo ds: MR dampe , pa ame e iden i ica ion, benchma k building
1. In oduc ion
Magne o heological (MR) dampe s a e de ices ha change hei mechan-
ical p ope ies when hey a e exposed o a magne ic ield. The magne o he-
ological luid o hese ac ua o s is cha ac e ized by a g ea abili y o a y,
in a e e sible way, om a ee- lowing linea iscous liquid o a semi-solid
one wi hin milliseconds [1]. Mo eo e , MR dampe s ha e a low cos , low
powe equi emen s, la ge o ce capaci y, obus ness and can be con olled
wi h a low ol age a he coils [1]. All hese ea u es make MR dampe s
e y a ac i e and p omising as ac ua o s con olled by he ol age ha can
be used in di e en enginee ing ields, such as dampe s and shock abso be s
(p essu e d i en low mode de ices), as well as clu ches, b akes, chucking,
and locking de ices (di ec -shea mode de ices) [9]. F om a s uc u al con-
ol poin o iew, MR dampe s a e usually employed as ac ua o s ope a ed
by low ol ages. In his espec , semi-ac i e con ol sys ems seem o com-
bine he bes comp omise be ween passi e and ac i e con ol: hey o e he
eliabili y o passi e de ices oge he wi h he e sa ili y and adap abili y
o ac i e sys ems [3, 5, 20]. Howe e , he i s s ep in he design o a semi-
ac i e con ol s a egy is he de elopmen o an accu a e model o he MR
de ice. I is wo h no ing ha he sys em-iden i ica ion issue plays a key
ole in his con ol p oblem [19]. High-accu acy models can be designed us-
ing wo di e en model amilies: semi-physical models [16, 20], and black box
models [10, 23]. Some o he mos known semi-physical models o desc ibe
he hys e e ic beha iou o MR dampe s a e he Bingham model and i s ex-
ended e sions, he Bouc-Wen model, he Dahl model, he modi ied LuG e
model and some o he non-pa ame ic models [12, 17]. I is impo an o e-
ma k ha hese models a e no linea -in-pa ame e s and, he e o e, classical
pa ame e iden i ica ion me hods, such as he g adien o he mean squa e
algo i hms, canno be applied.
Using a no malized e sion o he Bouc-Wen model, Ikhouane e al. [7]
p esen an iden i ica ion algo i hm which is di ec ly used o MR dampe s
in shea -mode [17]. Howe e , his me hodology can p oduce la ge pa ame-
e iden i ica ion e o s i he iscous ic ion is much smalle han he d y
ic ion [17]. To cope wi h his d awback, a modi ied s ep was p oposed by
Rod ´ıguez e al. [17] and es ed in a small-scale MR dampe . When he
2
iden i ica ion is applied o a la ge-scale MR dampe , he pa ame e iden i i-
ca ion e o s inc ease [18]. The aim o his pape is o imp o e he accu acy
o he iden i ica ion algo i hm. This is based on augmen ing he no malized
Bouc-Wen model wi h an addi ional e m. The alida ion o his modi ied
pa ame e iden i ica ion me hod has been ca ied ou using a black box model
o an MR dampe in a sma base-isola ed benchma k building [14]. This
model is unknown o he use o his benchma k and o he designe o
con ol sys ems. The benchma k pla o m is hen conside ed as a i ual
labo a o y expe imen . The nume ical esul s show ha he p oposed mod-
i ied me hod is able o imp o e signi ican ly he accu acy o he pa ame e
iden i ica ion.
The pape is o ganized as ollows. In Sec ion 2, he magne o heological
dampe model is p esen ed. In Sec ion 3, he key poin s o he modi ied
iden i ica ion me hod a e discussed. In Sec ion 4, he applica ion o he
p oposed iden i ica ion me hod o a la ge-scale MR dampe in a benchma k
building is conside ed. Finally, some concluding ema ks a e s a ed in Sec ion
5.
2. The magne o heological dampe model
The no malized e sion o he Bouc-Wen model [7] is an equi alen ep-
esen a ion o he o iginal Bouc-Wen model [23]. Fo MR dampe s in shea
mode i akes he o m:
Φn( ˙x, w)( ) = κ˙x( ) ˙x( ) + κw( )w( ),(1)
˙w( ) = ρ( ˙x( )−σ|˙x( )||w( )|n−1w( ) + (σ−1) ˙x( )|w( )|n),(2)
whe e Φn( ˙x, w) is he ou pu o ce o he MR dampe , ˙x( ) and a e he
eloci y and ol age inpu s, espec i ely. The ol age inpu is he applied
ol age a he coil o he MR dampe . The sys em pa ame e s, which a e
ol age-dependen , a e κ˙x( )>0, κw( )>0, ρ > 0, σ > 1/2, and n≥1.
These pa ame e s con ol he shape o he hys e esis loop and hei meaning
can be ound in [6]. The s a e a iable w( ) has no a physical meaning so
ha i is no accessible o measu emen s.
Since he no malized Bouc-Wen ep esen a ion desc ibed in equa ions
(1)-(2) is no a linea -in-pa ame e model, classical pa ame e iden i ica ion
me hods canno be applied. In his ega d, a new pa ame e iden i ica ion
algo i hm has been p oposed in [7, p. 38], which is based on a physical un-
de s anding o he de ice along wi h a black box desc ip ion. Rod ´ıguez e
3
al. [18] used his me hodology o a la ge-scale MR dampe . The me hod is
based on applying a pe iodic inpu eloci y ˙x( ) a a cons an ol age coil
and obse ing he pe iodic s eady-s a e o ce esponse o he MR dampe .
None heless, la ge ela i e e o s in he iden i ica ion p ocess can be obse ed
when he MR dampe has a iscous ic ion (κ˙x( ) ˙x( )) small enough wi h
espec o he d y ic ion (κw( )w( )). To cope wi h his d awback, when
he displacemen is la ge enough, an al e na i e me hod based on he plas ic
egion o he o ce- eloci y diag am o he MR dampe has been p oposed in
[17]. Howe e , he model in equa ions (1)-(2) may no gi e an accu a e ep-
esen a ion o la ge-scale MR dampe s which do no belong o he shea - ype
ca ego y. Fo ins ance, conside he black box model o an MR dampe in
he sma base-isola ed benchma k building [14]. Figu e 1 con ains he o ce-
displacemen and o ce- eloci y diag ams when his MR dampe is exci ed
by a sinusoidal displacemen and eloci y. In o de o es he goodness o
he Bouc–Wen model, his igu e also con ains he esponse o he dynamic
model in equa ions (1)-(2) wi h some app op ia e pa ame e s. I can be ob-
se ed ha he esul ing plas ic b anch in he o ce- eloci y diag am is wide
ha he co esponding b anch o he Bouc–Wen model. In he li e a u e, he
same ype o cycles ha e been expe imen ally epo ed, o ins ance, in [4,
Figu e 4(b)], [11, Figu e 9] and [13, Figu e 7]. The e o e, i can be de i ed
ha his kind o hys e e ic loops a e unable o be ep oduced wi h he o igi-
nal Bouc–Wen model. To imp o e he accu acy o he model ep esen a ion
and, consequen ly, he accu acy o he pa ame e iden i ica ion, we use he
ollowing ex ended Bouc-Wen model:
Φe(x, ˙x, w)( ) = κx( )x( ) + κ˙x( ) ˙x( ) + κww( ),(3)
˙w( ) = ρ( ˙x( )−σ|˙x( )||w( )|n−1w( ) + (σ−1) ˙x( )|w( )|n),(4)
whe e he e m κx( )x( ), which ep esen s a linea elas ic o ce, has been
added. We conside ha he coe icien κxis ol age-dependen , as he o he
pa ame e s. The e ec o his e m can be obse ed no only in he o ce-
eloci y diag am, bu also in he o ce-displacemen : he esul ing plo is
inclined. This ea u e has also been expe imen ally epo ed in, o ins ance,
[2, Figu e 8], [11, Figu e 9] and [13, Figu e 7].
3. Model pa ame e iden i ica ion
This sec ion is conce ned wi h he compu a ion o he pa ame e s o he
ex ended model in equa ions (3)-(4). The p oposed algo i hm will be di ided
4
−200
−150
−100
−50
0
50
100
150
200
250
Fo ce (kN)
−2 −1 0 1 2
−400
−300
−200
−100
0
100
200
300
400
Fo ce (kN)
−250
Displacemen (m) Veloci y (m/s)
0−0.2−0.4 0.40.2
plas ic b anch
Figu e 1: Fo ce-displacemen (le ) and o ce- eloci y ( igh ) diag ams o he black box
model o he MR dampe when i is exci ed by a sinusoidal displacemen and eloci y
(dashed). The esponse o he dynamic Bouc–Wen model in equa ions (1)-(2) unde he
same exci a ion is ep esen ed by a solid line.
in wo s eps: (a) he es ima ion o he alue o κxand (b) he es ima ion o
he es o he pa ame e s based on he iden i ica ion algo i hm in [18].
A cons an ol age, he compu a ion o he pa ame e κx( ) can be pe -
o med g aphically by conside ing he o ce-displacemen diag am o he MR
dampe . When his de ice is exci ed by a sinusoidal displacemen wi h a
la ge enough ampli ude, he a e age inclina ion o he esul ing plo gi es an
es ima ion o his pa ame e . As an example, conside he black box model
o an MR dampe in he sma base-isola ed benchma k building. When
his MR dampe is d i en wi h ze o coil command ol age, we ob ain he
o ce-displacemen diag am in Figu e 2. The es ima ed alue o κxis hen
compu ed as κx=82.8
0.4= 207 kN.
To es ima e he es o he pa ame e s, we use he knowledge o he
pa ame e κxand he ac ha
Φn( ˙x, w)( ) = Φe(x, ˙x, w)( )−κx( )x( ).
Figu e 3 depic s (in solid line) he o ce- eloci y diag am o he esul ing
ou pu o ce Φeo he MR dampe minus he linea elas ic o ce κx( )x( ),
when his de ice is exci ed by a sinusoidal displacemen and eloci y. I can
be ecognized om his igu e ha his new cycle has he same shape as
5

he o ce- eloci y cycle o he model in equa ions (1)-(2), which is plo ed in
Figu e 1 igh . As a esul , o he iden i ica ion o he es o he pa ame e s,
ha is, he pa ame e s o he model
Φn( ˙x, w)( ) = κ˙x( ) ˙x( ) + κw( )w( ),
˙w( ) = ρ( ˙x( )−σ|˙x( )||w( )|n−1w( ) + (σ−1) ˙x( )|w( )|n),
we can ollow basically he same idea as in [18]. The de ails o his me hod
a e omi ed he e bu can be ound in he Appendix.
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5
−400
−300
−200
−100
0
100
200
300
400
Displacemen (m)
Fo ce (kN)
82.8 kN
Figu e 2: The a e age inclina ion o he o ce-displacemen diag am, when he MR dampe
is exci ed by a sinusoidal displacemen , gi es an es ima ion o he pa ame e κx.
4. Applica ion o a benchma k building
This sec ion is conce ned wi h he applica ion o he p oposed me hod on
a i ual MR dampe . Mo e p ecisely, he p oposed iden i ica ion algo i hm
is es ed using a black box model o an MR dampe , which is a pa o a
sma base-isola ed benchma k building p oblem [14]. Consequen ly, we use
his nume ical pla o m as a i ual labo a o y es (Figu e 4). To alida e
6
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2
−250
−200
−150
−100
−50
0
50
100
150
200
250
Veloci y (m/s)
Fo ce (kN)
Figu e 3: Fo ce- eloci y diag am o he esul ing ou pu o ce Φeo he MR dampe
(dashed) and he o ce Φn= Φe−κx( )x( ) (solid).
he esul s, he ou pu o ces o he i ual de ice and he iden i ied one will
be compa ed using se en p ede ined ea hquake eco ds o he benchma k
p oblem wi h hei co esponding luc ua ing ol age du ing ull simula ions.
The MR dampe is used in his con ex as a semi-ac i e de ice o educe he
s uc u al esponse o he building.
4.1. Sma base-isola ed benchma k building
The sma base-isola ed benchma k building [14] is employed as an in e -
es ing and mo e ealis ic example o u he in es iga e he e ec i eness o
he p oposed design app oach. This benchma k p oblem is ecognized by he
Ame ican Socie y o Ci il Enginee s (ASCE) S uc u al Con ol Commi ee
as a s a e-o - he-a model de eloped o p o ide a compu a ional pla o m
o nume ical expe imen s o seismic con ol a enua ion [15, 21].
The benchma k s uc u e is an eigh -s o ey ame building wi h s eel-
b aces, 82.4 m long and 54.3 m wide, simila o exis ing buildings in Los
7
MR dampe
x
˙x
Φ
Figu e 4: Inpu -ou pu a iables o he i ual MR dampe .
Angeles, Cali o nia. S o ies one o six ha e an L-shaped plan while he
highe loo s ha e a ec angula plan. The supe s uc u e es s on a igid
conc e e base, which is isola ed om he g ound by an isola o laye , and
consis s o linea beam, column and b acing elemen s and igid slabs. Below
he base, he isola ion laye consis s o a a ie y o 92 isola ion bea ings. The
isola o s a e connec ed be ween he d op panels and he oo ings below, as
shown in Figu e 5.
Figu e 5: Ele a ion iew wi h de ices
4.2. Iden i ica ion esul s
In o de o implemen he iden i ica ion p ocedu e in Sec ion 2 i is nec-
essa y o apply a pe iodic exci a ion displacemen and obse e he co e-
8
sponding MR dampe o ce. Figu e 6 illus a es hese wo signals o a ze o
ol age. A se o expe imen s ha e been pe o med o di e en ol ages in
he ange [0,1] ol s.
0 2 4 6 8 10
−80
−60
−40
−20
0
20
40
60
80
Time (s)
Fo ce (kN)
0 2 4 6 8 10
−0.03
−0.02
−0.01
0
0.01
0.02
0.03
Time (s)
Displacemen (m)
Figu e 6: Response o he MR dampe model in he benchma k building pla o m.
The esul ing alues o he pa ame e s o he model in equa ions (3)-(4)
a e lis ed in Table 1. Figu e 8 plo s hese pa ame e s as a unc ion o he
ol age. To ind an accu a e ol age-dependen ela ion o hese pa ame e s,
and acco ding wi h he unc ional dependence in Figu e 8, we conside ha
κx( ) is cons an , κ˙x( ) is linea and n( ), ρ( ) and σ( ) a e exponen ial:
κx( ) = κx(5)
κ˙x( ) = κ˙x,a +κ˙x,b (6)
n( ) = na+nbexp(−13 ) (7)
ρ( ) = ρa+ρbexp(−14 ) (8)
σ( ) = σa+σbexp(−14 ) (9)
Because o he impo ance o he pa ame e κwdue o i s g ea in luence in
he esul ed o ce ( he ange o i s magni ude is, app oxima ely, om 50 kN
o 1000 kN, as can be seen in Table 1), i s ol age dependence unc ion has
been es ima ed in h ee di e en egions
κw( ) = 


κw1+κw2 1.15, ≤0.3
κw3+κw4sin(π( −0.3)
0.8) + κw5sin(3π( −0.3)
0.8),0.3≤ ≤0.7
κw6+κw7 +κw8 3+κw9 5,0.7≤
,
(10)
based on he a ia ion o he esul ed alues (Figu e 7).
9
0 5 10 15 20 25 30
−1000
−500
0
500
1000
Time (s)
Dampe Fo ce E o (kN)
0 5 10 15 20 25 30
−1000
−500
0
500
1000
Time (kN)
Dampe Fo ce E o (kN)
Figu e 11: Gene a ed dampe o ce e o s o p oposed model (abo e), and o iginal me hod
[17] (below) , unde Kobe g ound mo ion (FP-y).
iden i ica ion me hod has been es ed using he MR dampe as a semi-ac i e
de ice unde ime- a ying ol age and ea hquake exci a ion.
A. Appendix
The pa ame e iden i ica ion in [18] depa s om he nex shea -mode
model:
Φn( ˙x)( ) = κ˙x( ) ˙x( ) + κw( )w( ) (13)
˙w( ) = ρ( ˙x( )−σ|˙x( )||w( )|n−1w( ) + (σ−1) ˙x( )|w( )|n) (14)
whe e κ˙x>0, κw>0, ρ > 0, σ > 1/2, and n≥1. All o hese pa ame e s
can be ol age o cu en dependen (he e he case o ol age dependency
is unde conside a ion, as emphasized o κ˙xand κw). I has been shown in
[7] ha his model is meaning ul in he sense ha he limi cycle depends
16

di ec ly on he pa ame e s ha appea in he no malized o m, and hus
depends only indi ec ly on he pa ame e s o he s anda d o m as
ρ=A
Dz0
>0,
σ=β
β+γ≥0,
κ˙x=αk > 0,
κw= (1 −α)Dkz0>0,
whe e A, D, α, β, γ and kcomes om he s anda d Bouc–Wen model
ΦBW(x)( ) = αkx( ) + (1 −α)Dkz( ),
˙z=D−1A˙x−β|˙x|z|z|n−1−γ˙x|z|n.
Fo pa ame e iden i ica ion, a T-pe iodic inpu ˙x( ) (see Figu e 12) is ap-
plied o he Bouc-Wen sys em unde cons an ol age . I has been p o ed
[7] ha he ou pu o ce o he Bouc-Wen model goes asymp o ically o a
pe iodic s eady-s a e so ha a limi cycle is ob ained. The iden i ica ion
me hod assumes he knowledge o he ela ion ¯w(x) ha desc ibes his cycle
as illus a ed in Figu e 13. The whole iden i ica ion p ocess can be summa-
ized as ollows.
The pa ame e κ˙xis i s de e mined using he plas ic egion ( ¯w≈1) o
he hys e esis loop by a linea eg ession o each cons an ol age:
¯
F(τ) = κ˙x( ) ˙x(τ) + κw( ).
To con inue wi h pa ame ic es ima ion, a unc ion θis compu ed as:
θ(x(τ)) = ¯
F(x(τ)) −κ˙x
dx(τ)
dτ , τ ∈[0, T+],(15)
which has a unique ze o, i.e, he e exis s a ime ins an τ∗∈[0, T+], and a
co esponding alue x∗=x(τ∗)∈[Xmin, Xmax], such ha he unc ion θis
ze o. Because θis known, hen ˙x∗is also known. De ine he quan i y
a=dθ(x)
dx x=x∗
.(16)
17
0T+TmT mT +T+(m+ 1)T
Xmin
Xmax
Inpu signal x
Time
Figu e 12: A sample T-wa e pe iodic signal
Then, he pa ame e nis de e mined as:
n=
log (dθ(x)
dx )x=x∗2
−a
(dθ(x)
dx )x=x∗1
−a
log θx=x∗2
θx=x∗1(17)
whe e x∗2> x∗1> x∗a e design pa ame e s. De ine
b=
a−dθ(x)
dx x=x∗2
θ(x∗2)n.(18)
Then, he pa ame e s κwand ρa e compu ed as ollows:
κw=n
a
b,(19)
ρ=a
κw
.(20)
18
Figu e 13: Symme y p ope y o he hys e esis loop o he no malized Bouc–Wen model.
The unc ion ¯w(x) can be compu ed as:
¯w(x) = θ(x)
κw
.(21)
Finally, he emaining pa ame e σis de e mined as:
σ=1
2


(d¯w(x)
dx )x=x∗3
ρ−1
(−¯w(x∗3)n)+ 1

(22)
whe e x∗3is a design pa ame e such ha x∗3< x∗.
Re e ences
[1] G. Bossis , P. Khuzi , S. Lacis, and O. Volko a, Yield beha io o mag-
ne o heological suspensions, Jou nal o Magne ism and Magne ic Ma e-
ials,258-259:456-458, 2003.
19
[2] W.W. Chooi, S.O. Oyadiji, Design, modelling and es ing o magne-
o heological (MR) dampe s using analy ical low solu ions, Compu e s
& S uc u es,86(3-5):473–482, 2008.
[3] A. Dominguez, R. Sedagha i, and I. S iha u, Modeling and applica ion
o MR dampe s in semi-adap i e s uc u es, Compu e s & S uc u es,
86(3-5):407–415, 2008.
[4] S.J. Dyke, B.F. Spence J ., M.K. Sain, and J. D. Ca lson, An expe -
imen al s udy o MR dampe s o seismic p o ec ion, Sma Ma e ials
and S uc u es,7(5):693–703, 1998.
[5] Z.Q. Gu, and S.O. Oyadiji, Applica ion o MR dampe in s uc u al
con ol using ANFIS me hod, Compu e s & S uc u es,86(3-5):427–
436, 2008.
[6] F. Ikhouane, J.E. Hu ado, and J. Rodella , Va ia ion o he hys e e-
sis loop wi h he Bouc-Wen model pa ame e s, Nonlinea Dynamics,
48(4):361–380, 2007.
[7] F. Ikhouane, and J. Rodella , Sys ems wi h Hys e esis: Analysis, Iden-
i ica ion and Con ol Using he Bouc-Wen Model, John Wiley & Sons,
Inc., 2007.
[8] L.M. Jansen, and S.J. Dyke, Semiac i e con ol s a egies o MR
dampe s: compa a i e s udy, Jou nal o Enginee ing Mechanics,
126(8):795–803,2000.
[9] M. R. Jolly, J. W. Bende , and J. D. Ca lson, P ope ies and applica ions
o comme cial magne o heological luids, Jou nal o In elligen Ma e ial
Sys ems and S uc u es,10(1):5–13,1999.
[10] J. Gang, M. K. Sain, K. D. Pham, B. F. Spence J ., and J. C. Ramallo,
Modeling MR dampe s: A nonlinea black box app oach, P oceedings o
he Ame ican Con ol Con e ence: 429–434, 2001.
[11] H. Ga in, J. Hoagg, and M. Dobossy, Op imal design o MR dampe s,
P oceedings o he US-Japan Wo kshop on Sma S uc u es o Im-
p o ed Seismic Pe o mance in U ban Regions, Sea le WA, 225–236,
2001.
20
[12] R. Jim´enez-Fabi´an, and L. Al a ez-Icaza, Simul aneous s a e es ima ion
and pa ame e uning in a shea building wi h a magne o- heological
dampe , S uc u al Con ol and Heal h Moni o ing,16(4):483–502,
2008.
[13] G. Jin, M. K. Sain, and B. F. Spence J ., Nonlinea blackbox modeling
o MR-dampe o ci il s uc u al con ol, IEEE T ansac ions on Con ol
Sys ems Technology,13(3):345–355, 2005.
[14] S. Na asimhan, S. Naga ajaiah, E. A. Johnson, and H. P. Ga in, Sma
base-isola ed benchma k building. Pa I: P oblem de ini ion, S uc u al
Con ol and Heal h Moni o ing,13(2-3):573–588, 2006.
[15] Y. Oh o i, R.E. Ch is enson, B.F. Spence and S.J. Dyke, Benchma k
p oblems in seismically exci ed nonlinea buildings. Jou nal o Engi-
nee ing Mechanics,130(4):366–385, 2004.
[16] F. Pozo, L. Acho, A. Rod ´ıguez, and G. Pujol, Nonlinea modeling o
hys e e ic sys ems wi h double hys e e ic loops using posi ion and ac-
cele a ion in o ma ion, Nonlinea Dynamics,57(1-2):1–12, 2009.
[17] A. Rod ´ıguez, F. Ikhouane, J. Rodella , and N. Luo, Modeling and
iden i ica ion o a small-scale magne o heological dampe , Jou nal o
In elligen Ma e ial Sys ems and S uc u es,20(7):825–835, 2009.
[18] A. Rod ´ıguez, N. Iwa a, F. Ikhouane, and J. Rodella , Model iden i ica-
ion o a la ge-scale magne o heological luid dampe , Sma Ma e ials
and S uc u es,18(1), doi: 10.1088/0964-1726/18/1/015010, 2009.
[19] S. M. Sa a esi, S. Bi an i, and M. Mon iglio, Iden i ica ion o semi-
physical and black box nonlinea models: The case o MR dampe s o
ehicles con ol, Au oma ica,41(1):113–127,2005.
[20] B. F. Spence J ., S. J. Dyke, M. K. Sain, and J. D. Ca lson, Phenomeno-
logical model o a magne o heological dampe , Jou nal o Enginee ing
Mechanics,123(3): 230–238, 1997.
[21] B.F. Spence , and S. Naga ajaiah, S a e o he a o s uc u al con ol.
Jou nal o S uc u al Enginee ing,129(7):845–856, 2003.
21

[22] D. H. Wang, and W. H. Liao, Neu al ne wo k modeling and con olle s
o magne o heological luid dampe s, P oceedings o he 10 h IEEE In-
e na ional Con e ence on Fuzzy Sys ems,3: 1323–1326, 2001.
[23] Y. K. Wen, Me hod o andom ib a ion o hys e e ic sys ems, Jou nal
o Enginee ing Mechanics,102(2):249-263, 1976.
[24] P.Q. Xia. An in e se model o MR dampe using op imal neu al ne wo k
and sys em iden i ica ion, Jou nal o Sound and Vib a ion,266(5):1009-
1023, 2003.
22