1- Independent Harmonic Control for Structural Engineering

download 1- Independent Harmonic Control for Structural Engineering

of 19

Transcript of 1- Independent Harmonic Control for Structural Engineering

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    1/19

    Journal of Earthquake Engineering, Vol. 6, No. 3 (2002) 297314c Imperial College Press

    INDEPENDENT HARMONIC CONTROL FOR

    STRUCTURAL ENGINEERING

    ALESSANDRO BARATTA and OTTAVIA CORBI

    Department of Scienza delle Costruzioni,

    University of Naples, via Claudio 21 Naples, 80125, Italy

    Received 17 May 2001Revised 31 October 2001

    Accepted 14 November 2001

    In a previous paper [Baratta and Corbi, 1999] one has defined a procedure allowing toidentify a closed-loop control algorithm with feedback based on the whole record of theresponse time-history rather than on instantaneous response parameters. The controlforce results from control of each harmonic component of the forcing function, simplyintegrated over the frequency domain. Every harmonic is controlled, independently ofeach other, by a classical linear control whose coefficients are calibrated in way to makethe relevant response component a minimum compatibly with the control effort one

    wants to apply at the corresponding frequency. The distribution of this control inten-sity over the frequency range remains a arbitrary choice; such a choice however lendsitself to be effectively assisted by intuition, much more than similar choices in otherprocedures (e.g.: the coefficients of the quadratic norms in the J-index optimization).The result is that every harmonic remains controlled by a different couple of optimalcoefficients (corresponding to the proportional and to the derivative terms in the linearcontrol law), and the overall control force for an arbitrary disturbance, after Fourierinverse transformation, is produced by feedback integration over the whole responsetime-history.

    The procedure, tested with reference to simple and composed harmonic excitationsincoming a s.d.o.f. structural system, has proved a good agreement of the numerical re-sults with the theoretical treatment; furthermore it has shown that the main limit of such

    an approach consists of referring the dynamic equilibrium solution to a particular solu-tion, that, neglecting the initial conditions, may introduce some unstable componentsin the oscillation. In the paper the effects induced in the controlled structural systemresponse by the adoption of the proposed procedure are deepened and an improved strat-egy is presented, able to overcome the detrimental transient effects determined by theoriginal algorithm. The final adopted control law is shown to achieve an improved timeresponse, both in the transient and in the steady-state field, in comparison to a controlstrategy based on classical linear control minimizing the response norm conditioned bya bounded control.

    Keywords: Harmonic control; frequential decomposition; norm algorithm.

    1. Introduction

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    2/19

    298 A. Baratta & O. Corbi

    interest on this argument is proven by the large literature that is appearing on

    many journals and by the many conferences that are holding all over the world. A

    very important state of the art paper has been redacted by a number of promi-nent scientists in the field, and it documents the encouraging results that have been

    obtained and that are currently being produced [Housner et al., 1996].

    At least four possible approaches to Structural Control can be distinguished:

    passive, active, semi-active and hybrid, each possessing its own characteristics and

    requiring different hardware and software equipment [Housner and Masri, 1996;

    Kobori, 1998]. In Active Control (AC) technology [Soong, 1990; Soong et al., 1994],

    the control action is added by external forces generated by contrast with oscillating

    masses (Active Mass Driver) or by contrast with an oscillating system that can be

    damaged and possibly replaced after the quake (Inter-Active Control). Such forcesare driven during the motion by a CPU central on the basis of the instantaneous

    sensing of the response parameters and according to a pre-established rule, thus

    conferring a kind of artificial intelligence to the system. The rule governing the

    control force can be set on the basis of a mathematical algorithm or on the basis of

    learning devices such as neural networks or genetic algorithms, all of them aiming at

    achieving an objective response that is considered optimal in some sense, essentially

    to yield the maximum possible mitigation of the structures strength, at the expense

    of the minimum engagement of the control force.

    Particular AC systems make recourse to linear control laws (Linear Control)whose objectives are identified in some well defined functionals of the response and

    of the control force (Quadratic Performance Index, Norm Response Functionals,

    Instantaneous Optimal Control IOC, etc.) [Soong, 1990]. A different control

    approach expressly studied to save the energy employed in the control process is the

    Bounded State Control, that finds application in a class of algorithms designed to

    reduce the controlled response variables (expressed in the state space) to predefined

    bounds: a suitably chosen control force activates only when an overcoming of the

    predefined threshold of any of the state variables is detected and lasts just for the

    length of time necessary to reduce the variable to the allowed magnitude range.In previous papers the authors [Baratta and Corbi, 1999] have proposed a pro-

    cedure finalised to define a linear control feedback algorithm for structural systems,

    able to effectively counteract the incoming forcing function that, as well known,

    can be viewed as the sum of a number of harmonic components. The procedure is

    based on the superposition of control laws independently controlling every single

    harmonic component of the external disturbance. The method allows a wide free-

    dom in the distribution of the control effort over the frequency range, that can be

    decided on the basis of some expectation concerning the power spectrum of the

    excitation. Actually the algorithm, designed in the frequency field by developing aprocedure referred just to the steady-state domain, is able to achieve really satis-

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    3/19

    Independent Harmonic Control for Structural Engineering 299

    effectiveness of the proposed procedure is widely investigated by simulation and the

    chance to attenuate the influence of the transient motion by coupling the harmonic

    control with a classical linear algorithm is tested.

    2. The Harmonic Control Algorithm

    2.1. Constrained optimal linear control for a harmonic excitation

    Let consider a controlled s.d.o.f. structural system

    mu(, t) + cu(, t) + ku(, t) + w(, t) = f(t) , (1)

    subject to an harmonic forcing function f(t) = fo()e

    jt

    and a linear control laww(, t) = wu(, t) + wu(, t) = p()u(, t) + q()u(, t) (2)

    being p() and q() the control parameters; a particular solution u(, t) of Eq. (1),

    should satisfy

    mu(, t) + cu(, t) + ku(, t) +p()u(, t) + q()u(, t) = f(t) (3)

    and it can be set in the form

    u(, t) = uo()ejt , u(, t) = juo()e

    jt , u(, t) = 2uo()ejt ; (4)whence one deduces that uo() satisfies

    uo() = H()fo() (5)

    with

    H() =1

    [k +p() m2] +j[c + q()]

    =k +p() m2 j[c + q()]

    [k +p() m2]2 + 2[c + q()]2

    |H()| = 1[k +p() m2]2 + 2[c + q()]2 .

    (6)

    Note that, assuming c > 0, |H()| is bounded for = 0, for any value of p() andfor any q() > c; it follows that the control Eq. (2) cannot introduce instability inthe solution u(, t). This does not mean that divergent motion cannot be triggered,

    as it will be discussed in Sec. 3.1.

    Note that, in the following, for simplicity of notation, the symbol x2 will be used

    signifying |x|2 for any complex variable x. The control parameters p() and q()must be selected in way to be compatible with some design and functional issues.

    A possible approach in Structural Engineering problems is to aim at the maximum

    attenuation of the structure response by keeping the energy supply below a given

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    4/19

    300 A. Baratta & O. Corbi

    w(|p,q) = [p() +jq()]H()fo() ;

    w2(|p,q) = |p2() + 2q2()|H2()f2o ()(8)

    the search for the control parameters, optimal for an harmonic excitation, can be

    set in the form of the following problem of constrained minimum

    minp,q

    u2o(|p,q)f2o ()

    = minp,q

    H2(|p,q)

    subw2(|p,q)

    f2o ()= [p2() + 2q2()]H2(|p,q) C2o ()

    (9)

    where Co() is, for any given , a number less than unity, in that, as a rule, the

    control force is generally required to be instantaneously smaller in magnitude than

    the active force. With the position

    G2(|p,q) = [k +p() m2] + 2[c + q()]2

    V2(|p,q) = [p2() + 2q2()](10)

    problem (9) can be written in the equivalent inverse formulation

    maxp,q G

    2(|p,q)sub G2(|p,q) C2Io()V2(|p,q) 0

    (11)

    with CIo = 1/Co. In order to solve the constrained extremum problem, let consider

    the KuhnTuckers conditions, with the Lagrangian multiplier

    (1 + )G2(|p,q)

    p C2Io()

    V2(|p,q)p

    = 0

    (1 + )G2(

    |p,q)

    q C2Io()V2(

    |p,q)

    q = 0

    [G2(|p,q) C2Io()V2(|p,q)] = 0

    0

    (12)

    yielding

    (1 + )[k +p() m2] p()C2Io() = 0

    (1 + )2[c + q()] C2Io()2q() = 0(13)

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    5/19

    Independent Harmonic Control for Structural Engineering 301

    The case = 0 shall be discarded, in that it leads to the absolute minimum of the

    objective function, i.e. G2(

    |p,q) = 0. Assuming > 0 Eqs. (13) yield

    p() = (k m2)(1 + )

    (1 + C2Io())

    q() = c(1 + )(1 + C2Io())

    G2(|p,q) C2Io()V2(|p,q) = 0

    > 0 .

    (14)

    By solving the third of (14), one gets the roots

    1 =1 + CIo()

    C2Io() 1, 2 =

    1 CIo()C2Io() 1

    . (15)

    Since C2Io() > 1 and the Lagrangian multiplier must be positive, the second root

    is definitely discarded and the first two of (14), with = 1, yield the control

    parameters that comply with the optimal criterion expressed by (9)

    p() =k m2

    CIo()

    1

    ; q() =c

    CIo()

    1

    . (16)

    Assuming that c > 0, the coefficient q() is always positive, while p() is negative

    for values of larger than 2o =

    k/m.

    Note that for = o, p() = 0 and at this pulsation the control is exclusively

    demanded to the derivative term. This result agrees with the fact that the deriva-

    tive linear term turns out to be prominent in any control algorithm, especially at

    resonance [Baratta and Di Paola, 1996]. Note anyway that this is not a choice, but

    it is the optimal solution of the conditioned optimization problem in Eq. (9).

    Note that if Co() is introduced as a even function of , also 1 is a even

    function, and both the control coefficientsp() and q() are real and even functionsof .

    2.2. The optimal harmonic control for a generic forcing function

    Consider the structure subject to any generic forcing function f(t), whose Fourier

    transform fo() exists

    fo() =12

    +

    f(t)ejtdt ;

    f(t) =12

    +fo()e

    jtd .(17)

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    6/19

    302 A. Baratta & O. Corbi

    one gets

    m 12

    +

    2uo()ejtd + c 12

    +

    juo()ejtd

    + k12

    +

    uo()ejtd +

    12

    +

    p()uo()ejtd

    +12

    +

    jq()uo()ejtd =

    12

    +

    fo()ejtd (18)

    whence one concludes that

    u(t) = 12

    +

    uo()ejtd = 1

    2

    +

    u(, t)d (19)

    with

    u(t) =12

    +

    u(, t)d = j12

    +

    uo()ejtd

    u(t) =12

    +

    u(, t)d = 12

    +

    2uo()ejtd

    (20)

    is a particular solution of the equation

    mu(t) + cu(t) + ku(t) + w(t) = f(t) (21)

    yielding the response to the excitation f(t) of the structure, controlled by the active

    force

    w(t) =12

    +

    p()uo()ejtd +

    12

    +

    jq()uo()ejtd

    =1

    2+

    p()u(, t)d +1

    2+

    q()u(, t)d = wu(t) + wu(t) .

    (22)

    For a generic bounded-support forcing function (like in case of seismic action) the

    existence of all the above transforms is ensured, and uo() and u(t) constitute a

    Fourier transform pair

    uo() =12

    +

    u()ejd; juo() =12

    +

    u()ejd . (23)

    It is possible to manipulate the algorithm yielding the control force in Eq. (22).

    In fact, considering the inverse Fourier transforms of p() and q()

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    7/19

    Independent Harmonic Control for Structural Engineering 303

    introducing Eqs. (23) into (22) and distinguishing the two terms wu(t) and wu one

    can write

    wu(t) =12

    +

    p()uo()ejtd =

    1

    2

    +

    p()

    +

    u()ejd

    ejtd

    =12

    +

    12

    +

    p()ej(t)d

    u()d

    =12

    +

    P(t )u()d (25)

    wu(t) = 12

    +

    jq()uo()ejtd

    =1

    2

    +

    q()

    +

    u()ejd

    ejtd

    =12

    +

    12

    +

    q()ej(t)d

    u()d

    =1

    2 +

    Q(t

    )u()d . (26)

    Note that P(x) and Q(x) are real and even functions of x, in that both p() and

    q() are real and even functions of .

    One should remark that the response u(t) [Eq. (19)] results from the superposi-

    tion of harmonic components, each optimised in the sense of the problem (9), with

    the relevant component of the control obeying the constraint inequality. Hence the

    final control force remains bounded by

    |w(t)| =1

    2 +

    w(|p,q)e

    jt

    d 12

    +

    Co()|fo()|d . (27)

    Whence one understands how the function Co() can be properly designed in such

    a way to have the desired threshold level of the control force.

    With reference to the practical implementation of the proposed control

    algorithm, note that any control rule requires some prediction; this can, then, be

    reviewed as a common problem, met even in the optimal calibration of control coef-

    ficients in classical linear control that is nevertheless widely applied (see i.e [Soong,

    1990; Chap. 3. Baratta and Corbi, 2000]). So, practical feedback implementation

    will be forced to accept some manipulation and/or approximation. The problemis under current development and it is not yet fully afforded in this paper, which

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    8/19

    304 A. Baratta & O. Corbi

    2.3. Purely real forcing function

    In most cases (i.e. when a structure is invested by a strong earthquake) the dis-turbance f(t) is a purely real function, and it starts finding the structure in some

    known status at time t = 0. The structure response on the semi-axis t > 0 can

    be made independent on the previous time-history, by setting proper conditions at

    t = 0. Hence one can consider the function f(t) to be a real even function of t

    defined on the whole real axis. In this case fo() turns out to be as well a real even

    function of: the real uoR() and the imaginary uoI() parts ofuo() are an even

    and an odd function of respectively

    uoR() =

    k +p()

    m2

    [k +p() m2]2 + 2[c + q()]2 fo()uoI() = [c + q()]

    [k + p() m2]2 + 2[c + q()]2 fo()(28)

    so that also u(t), as expressed by Eq. (19), is real (but it is not even, in general).

    On such an understanding, one can directly refer to the one-sided transforms, so

    that all above formulas can be written in the form

    fo() = 2

    +

    0

    f(t)cos tdt ;

    f(t) =

    2

    +0

    fo()cos td

    (29)

    u(t) =

    2

    +0

    [uoR()cos t uoI()sin t]d (30)

    u(t) = j1

    2 +

    uo()ejtd

    =12

    +

    [juoR() uoI()] [cos t +j sin t]d

    =

    2

    +0

    [uoR()sin t + uoI()cos t]d (31)

    u(t) = 12

    +

    2uo()ejtd

    = 12

    +

    2[uoR() +juoI()] [cos t +j sin t]d

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    9/19

    Independent Harmonic Control for Structural Engineering 305

    wu(t) =12

    +

    p()uo()ejtd

    =

    2

    +

    p()[uoR()cos t uoI()sin t]d

    wu(t) =12

    +

    jq()uo()ejtd

    =

    2

    +0

    q()[uoR()sin t + uoI()cos t]d (33)

    P(x) =

    2

    +

    0

    p()cos xd ; Q(x) =

    2

    +

    0

    q()cos xd (34)

    wu(t) =

    2

    +0

    P(t )u()d ; wu(t) =

    2

    +0

    Q(t )u()d (35)

    Since the initial conditions at t = 0 for u(, t) are

    u(, 0) = uo() =

    [k +p() m2]

    G2(|p,q) j[c + q()]

    G2(|p,q)

    fo()

    u(, 0) = juo() =

    2[c + q()]G2(|p,q) +j

    [k +p() m2]G2(|p,q)

    fo()

    (36)

    the particular solution u(t) obeys the following initial conditions

    u(0) =

    2

    +0

    uoR()d =

    2

    +0

    [k +p() m2]G2(|p,q) fo()d

    u(0) =

    2

    +0

    uoI()d =

    2

    +0

    2[c + q()]

    G2(|p,q) fo()d(37)

    3. Considerations on the Initial Conditions Effect

    3.1. The role of initial conditions

    By looking at the results of the optimisation given by (16) and depicted in Fig. 1,

    where the objective function and the constraint condition are displayed through

    their level curves, it is possible to observe that in many cases the control coefficient

    p() is negative; it follows that the controlled response for a single harmonic force

    component [Eq. (3)] is less robust than the original one, or even that it becomes

    unstable (this is the case if p() < k).This happens although |H()| in Eq. (6) is smaller than the modulus of the

    f f ti f th t ll d t t |H ( )| I th d

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    10/19

    306 A. Baratta & O. Corbi

    100

    p10008006004002000

    q200

    150

    50

    0

    Co

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    11/19

    Independent Harmonic Control for Structural Engineering 307

    can be unstable or, at least, trigger somewhat different responses with varying the

    initial conditions.

    In order to illustrate this effect, consider the structure acted on by the harmonicforce

    f(t) = a1 sin t + a2 cos t . (38)

    The optimally controlled (in the sense of Sec. 2.1) structure response obeys the

    equation

    mu(, t) + c()u(, t) + k()u(, t) = f(t) (39)

    where

    k() = k +p() ; c() = c + q() . (40)

    The particular solution in the form equivalent to (4) is

    u(, t) = u1()sin t + u2()cos t

    u(, t) = [u1()cos t u2()sin t](41)

    with

    u1() =[k() m2]a1 + c()a2[k

    () m2

    ]2

    + c

    ()2

    2

    ;

    u2() =[k() m2]a2 c()a1[k() m2]2 + c()22 .

    (42)

    Equation (41) will be referred to as the stationary solution; it obeys the following

    initial conditions at t = 0

    u(, 0) = u2()

    u(, 0) = u1() .(43)

    Equation (42) proves that the stationary solution remains stable provided thatboth k and c in the original equation are larger than zero, as can be easily verified

    considering that p(0) > 0, q() > 0 [see (16)].By contrast, let consider the general integral of Eq. (39). The characteristic

    equation is

    m2 + c() + k() = 0 (44)

    whose roots are

    1 =c() + c()2 4k()m

    2m; 2 =

    c() c()2 4k()m2m

    (45)

    and the general integral of Eq. (39) is

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    12/19

    308 A. Baratta & O. Corbi

    If k() 0 it results 1 0 and the general integral is unbounded unless A = 0.This means that in this case only a one-parameter class of bounded solutions exists,

    corresponding to initial conditions

    U(0|B) = B + u2 ; U(0|B) = B + u1 any B (47)If k() > 0 but k()m c2 it results 1 0 and 2 < 0. If k()m > c2, 1and 2 are complex numbers and the free component of the motion, as well known,

    has a decaying oscillatory feature. In both cases the general integral is bounded.

    Nevertheless, the motion can be very heavily affected by initial conditions other

    than the stationary ones in (43).

    If, for instance, one searches for the solution verifying homogeneous initial con-

    ditions, as is the case if it is necessary to study the effect of a transient disturbanceon a structure, one has

    A =u1 2u2

    2 1 ; B =1u2 u1

    2 1 . (48)

    Some plots of the structure response vs. time are quoted in Figs. 24, corresponding

    to the following values of the structure and force parameters

    k = 900 kg/cm; m = 1 kgs2/cm; c = 1.5 kgs/cmo = 30 s1 ; = 5% ; To

    0.21 s ;

    = 2o = 60 s1; a1 = a2 = 1 kg .

    (49)

    In (49), o and To are the own pulsation and the own vibration period of the

    structure, and is the damping coefficient. From Fig. 2, wherep() = 900.3 kg/cmand q() = 0.5 kgs/cm are assumed, one can notice that, even if the systembecomes unstable (k() < 0), the stationary response remains bounded while,

    by contrast, the homogeneous solution diverges. In the limit case when k() = 0

    0.00 0.4

    0 0.80 1.20 1.60

    -4E-3

    0E+0

    4E-3

    8E-3

    1E-2

    t

    u

    (

    t

    )homogeneou

    s init

    ial condit

    ionsst

    at

    ionary init

    ial condit

    ions

    -8E+0

    -6

    E+0

    -4E+0

    -2

    E

    +!

    0

    0

    E

    +!

    0

    2

    E

    +!

    0

    t

    c" t#

    rl on hom. init#

    ial c" ond$

    it#

    ionsc

    "

    t#

    rl on st#

    at#

    . init#

    ial c"

    ond$

    it#

    ionsb

    %

    ase& ac" c" e& le& rat#

    ion

    0

    .0

    0

    0

    .40

    0

    .80

    1.20

    1.6

    0

    ..w(t),' u( (

    )

    t)g0

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    13/19

    Independent Harmonic Control for Structural Engineering 309

    0.00 0.40 0.80 1.20 1.60

    -4E-3

    0E7

    +0

    4E-3

    8E-3

    1E7

    -2

    t

    u (

    t) homogeneou s init

    ial condit

    ionsst

    at

    ionary init

    ial condit

    ions

    (a)

    -8E+0

    -6E+0

    -4E+0

    -2E+0

    0E+0

    2E7

    +0

    t

    0.00 0.40 0.80 1.20 1.60

    ct

    r8 l on9 h@

    omA . iB

    n9 iB

    t

    iB

    al con9 diB

    t

    iB

    on9 sct

    rl on st

    at

    . init

    ial condit

    ions

    bC

    ase acceler8

    at

    iB

    on9

    w(t),u

    (t)gD..

    (b)

    Fig. 3. k() = 0: structural time response (a), control action and forcing function (b).

    0.00 0.40 0.80 1.20 1.60

    -2E-3

    -1E-3

    0E7

    +0

    1E-3

    2E7

    -3

    t

    u

    (t) h@

    omA ogen9 eou s iB

    n9 iB

    t

    iB

    al con9 diB

    t

    iB

    on9 sst

    at

    ionary init

    ial condit

    ions

    (a)

    -2F

    EG

    +H

    0I

    -1EG

    +H

    0I

    0I

    E+0I

    1E+0I

    2E+0I

    t

    0I

    .0I

    0I

    0I

    .40I

    0I

    .80I

    1.20I

    1.60I

    cP

    tQ

    rl oR

    n hoR

    m. initQ

    ial cP

    oR

    ndS

    itQ

    ioR

    nsc

    P

    tQ

    rT

    lU

    oR

    nV

    stQ

    atQ

    . iW

    nV

    iW

    tQ

    iW

    alU

    cP

    oR

    nV

    dS

    iW

    tQ

    iW

    oR

    nV

    s

    bX

    aseY

    acP

    cP

    eY

    leY

    ratQ

    ioR

    n

    .` .`w(t),u (t)gD

    (b)

    Fig. 4. k() > 0: structural time response (a), control action and forcing function (b).

    (p() = k and q() = 0.5 kgs/cm are now considered) the homogeneous responseis bounded but it takes values much higher than the stationary one [Fig. 3], while

    Fig. 4 (where one assumes p() = 675 kg/cm and q() = 0.375 kgs/cm) provesthat, even if the system remains stable (k() > 0), the homogeneous response can

    be very different from the stationary one, for a rather long transient time ( t 8To).If one looks at the control force, one realizes that in the transient phase a surplus

    control may be required, disturbing the optimal character of the algorithm.

    3.2. Attenuation of the initial conditions effect

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    14/19

    310 A. Baratta & O. Corbi

    proposed control algorithm, consisting of the addition of a free- oscillation compo-

    nent to the stationary solution. One can believe that it is possible to attenuate such

    disturbance by adding a further control specifically designed for mitigation of freeoscillations. It is well known [Soong, 1990; Baratta and Corbi, 2000] that a classical

    linear control law is optimal to this purpose; the only problem consists of suitably

    calibrating the relevant coefficients.

    This can be done following the procedure set forth in [Baratta et al., 1998] under

    the name norm-control, based on the constrained minimization of the energy of the

    impulse response function of the controlled structure

    hc(x) =1

    cdecx sin cdx ; c =

    c + c

    2m

    ; cd = k + km

    2c (50)

    with k and c the coefficients of a classical linear control law

    w(t) = ku(t) + cu(t) (51)

    that are calibrated in way that+0

    h2c(x)dx = mink,c

    (52)

    under the condition

    wmax Wo (53)with Wo any given upper bound on the control force.

    In Fig. 5 the controlled response with a linear control, whose coefficients

    (k = 572.2 kg/cm, c = 32.3 kgs/cm, obtained for Wo = 1/4 of the forcing functionenergy) are calibrated according to the above norm procedure, added on the free

    component of the oscillations (with p() = 675 kg/cm and q() = 0.375 kgs/cm),

    -8E-4

    -4E-4

    0E+0

    4

    E

    -4

    8E

    -4

    t

    homogeneou s init

    ial condit

    ionsst

    at

    ionary! init

    ial condit

    ions

    u("

    t)

    0.00 0.40 0.80 1.20 1.60

    (a)

    -2E+0

    -1E

    +0

    0E+0

    1E+0

    2E+0

    t

    0.00 0.40 0.80 1.20 1.60

    ct

    rl on hom. init

    ial condit

    ionsct

    r# l on$ st

    at

    . i%

    n$ i%

    t

    i%

    al con$ di%

    t

    i%

    on$ s

    b&

    ase acceler#

    at

    i%

    on$

    .' .'w(

    "

    t),( u ("

    t)g)

    (b)

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    15/19

    Independent Harmonic Control for Structural Engineering 311

    -2E

    -3

    -1E

    -3

    0E+0

    1E-3

    2E

    -3

    t

    u("

    t

    )7

    0.00 0.4

    0 0.80 1.20 1.60

    n$

    or#

    m8

    con$

    t

    r#

    olledu n

    $

    con$

    t

    r#

    olled

    (a)

    -2E+0

    -1E

    +

    0

    0

    E

    +

    0

    1E+0

    2E+0

    t

    n

    o

    r

    m

    c

    o

    n

    t

    r

    o

    l

    u

    n

    c

    o

    n

    t

    r

    o

    l

    l

    e

    d

    b

    ase

    ac

    c

    e

    le

    rat

    io

    n

    0

    .0

    0

    0

    .4

    0

    0

    .80

    1.2

    0

    1.60

    w(t),u (t)g. .

    (b)

    Fig. 6. Response (a) and control force (b) only by norm control, with optimal coefficientsk = 572.2 kg/cm, c = 32.3 kgs/cm, obtained for Wo = 1/4 of the forcing function energy.

    is plotted; it shows that, when necessary, it is possible to resort to such provision

    in order to compensate for the initial conditions effect, as one can observe by com-

    parison with Fig. 4, where the additional control is absent.

    Finally, from Fig. 6, where the results derived by the adoption of the only norm

    control algorithm are reported, the convenience of adopting such a combined ap-

    proach, both in terms of effectiveness and economy of the employed control energy,

    can be appreciated.

    4. Application to an Earthquake-Type Forcing Function

    It is expected anyway, and it is confirmed by the numerical investigation (not

    extensively reported here for sake of brevity), that when the active force has a

    long duration (compared to the structure own period To) and has a initial build-

    up phase in which its intensity increases gradually, the effect produced by initialconditions is largely attenuated.

    The numerical investigation has been carried on considering the time-history

    recorded at Sturno in the Campania region in Southern Italy, during the 23 Novem-

    ber large earthquake in 1980. It is a quake characterised by a long duration (more

    than 70 s) and two subsequent strong motion phases, spaced out by a decay phase,

    in which the shake attains a peak acceleration of 0.225 g.

    Its time-history and power spectrum are plotted in Fig. 7. In Fig. 7(b), it is

    also reported the smoothed power spectrum S(), responding to the expression

    resulting from the superposition of the two KanaiTajimi curves

    S() = [G(|a a) + G(|b b)] (54)

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    16/19

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    17/19

    Independent Harmonic Control for Structural Engineering 313

    where is a normalisation factor, yielding the equality of the total energy between

    the actual and the smooth spectrum [Baratta and Zuccaro, 1999], and the param-

    eters a, a, b, b are given in Fig. 7(b).The control strategy can be designed in way that the control action is more

    intense at the frequency that can transmit more energy to the structure. This

    scope can be achieved giving the function Co() the following expression

    Co() =

    ComS()|Ho()| if ComS()|Ho()| ComaxComax otherwise .

    (56)

    In Figs. 8 and 9, the results of the simulation are quoted assuming Com =

    Comax = 0.9.

    5. Conclusions

    In the paper a procedure, able to optimise the control action with respect to the

    foreseen forcing function and constrained by an upper bound on the control action,

    has been developed. The procedure is based on the harmonic decomposition of

    the forcing function and on the ideal distribution of the control action over the

    frequency range in an optimal fashion with respect to the expected power density

    distribution of the active force. The set up of the rationale allows to distribute thecontrol power over the frequency range, by simply assigning the function Co(),

    in way to save the control power where the control action is expected to be less

    effective. Furthermore the significant influence of the initial conditions on what one

    could deduce if referring to the frequency domain has been deeply analysed. A wide

    numerical investigation (here only partially reported) has been finally carried out,

    demonstrating the effectiveness of the procedure, that allows a consistent response

    mitigation with a minimum control energy employ.

    References

    Baratta, A. and Di Paola, M. [1996] Optimal Non-Linear Control Law for LinearSdof Structures, Proc. 1st European Conference on Struct. Control, Barcelona,pp. 5663.

    Baratta, A., Cennamo, C. and Voiello, G. [1998] Design of linear active control algorithmsfor non-stationary unknown forcing function, J. Struct. Control2, pp. 731.

    Baratta, A. and Corbi, O. [1999] Improved control of structures by time-delayed algo-rithms, Proc. of the 5th International Conference on the Application of ArtificialIntelligence to Civil and Structural Engineering, Oxford, pp. 159168.

    Baratta, A. and Corbi, O. [1999] Algoritmo di controllo per forzanti armoniche neldominio delle frequenze, Proc. of the 12nd Italian Congress of ComputationalMechanics, GIMC, Napoli, pp. 1116.

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    18/19

    314 A. Baratta & O. Corbi

    Baratta, A. and Corbi, O. [2000] On the optimality criterion in structural control,Earthq. Engg. Struct. Dyn. 29, pp. 141157.

    Baratta, A. and Zuccaro, G. [1999] How to utilize the anti-optimization analysis to treatuncertainty in seismic excitation?, in Whys and Hows in Uncertainty Modeling.Probability, Fuzziness and Anti-optimization I. Elishakoff Ed., CISM Courses andLectures, 388, Springer-Verlag, pp. 159.

    Housner, G. W. and Masri, S. F. [1996] Structural control research issues arising outof the Northridge and Kobe earthquakes, Proc. of the 11th World Conference onStructural Control, Acapulco, paper n.2009.

    Housner, G. W., Bergman, L. A., Caughey, T. K., Chassiakos, A. G., Claus, R. O., Masri,S. F., Skelton, R. E., Soong, T. T., Spencer Jr., B. F. and Yao, T. P. [1997] Structuralcontrol: Past, present, and future, Eng. Mech. ASCE123(9), pp. 897971.

    Kobori, T. [1998] Mission and perspective towards future structural control research,

    Proc. of the 2nd World Conference on Structural Control, Kyoto, 1, pp. 2534.Soong, T. T. [1990] Active Structural Control: Theory and Practice, Longman, England

    and Wiley, New York, pp. 1194Soong, T. T., Reinhorn, A. M., Aizawa, S. and Higashino, M. [1994] Recent structural

    applications of active control technology, J. Struct. Control1(2), pp. 521.

  • 8/8/2019 1- Independent Harmonic Control for Structural Engineering

    19/19