Scribd Logo
Upload

Welcome to Scribd!

  • 82 views

    State Space Model of A Sructure

    Uploaded by

    vderti
    state space

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    State Space Model of A Sructure

    Uploaded by

    vderti
    82 views5 pages
    state space

    Original Title

    State Space Model of a Sructure

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    state space

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    82 views5 pages

    State Space Model of A Sructure

    Uploaded by

    vderti
    state space

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 5

    State Space Model of a Structure

    R. Kashani, Ph.D. www.deicon.com


    Traditionally, structural dynamics has been formulated using a set of secondorder ODEs. As familiar to the practitioners as this formulation is, it has some drawbacks. For example, decoupling the differential equations of a multiDOF system with viscous damping using this formulation is not possible. Moreover, this formulation does not lend itself to active multiinputmultioutput control either. These limitations are addressed using state space formulation, in which a set of rst order ODEs describe the system dynamics, depicted in the generic form of x y


    Ax Cx

    Bu Du

    where x, u, and y are the vectors of states, inputs, and outputs, respectively. A, B, C, and D are the constant dynamics, input, output, and direct transmission matrices, respectively. In structural dynamics, displacements and velocities (in physical and modal domains) are the most commonly used states. Readers unfamiliar with the state space formulation of dynamics systems are referred to State Space Modeling tutorial. The formulation presented by Equations 1 and 2 is the basis for statespace modeling of exible structures, having point force(s) as the input(s) and point displacement(s) as the measured output(s).

    y where state vector: number of modes: number of inputs: number of outputs: modal displacement: modal velocity: input: spatial coordinate i: output: natural frequency:

    zt

    
    =

    W O z t t

    Nm Nu Ny t t ut ri yt

    = 1 t 2 t 1 t 2 t = = u1 t u2 t

    = x r1 t x r2 t = diag 1 2 1

                           


    O 2 I 2 z O Q u Du Nm t Nm t uNu t
    T T T

    (1) (2)

    x rNy t

    Nm

    modal damping: eigenfunction i at location j : modal input matrix:

    i j Q

    = diag 1 2 1 1 . . .

    modal output matrix:

     !!! " & ' ' ' & $%# & ' ' ' & & ' ' ' & $%# & ' ' ' &
    Nm .. . Nm 1 1 1 . . . .. . 1 Ny

    Nm Nu Nm 1 . . .

    1 Nu . . .

    () 0 () 0

    Nm Ny

    The A, B, C, and D matrices describing the exible structures statespace model, shown in Equation 1 and 2, are functions of the parameters of the system (natural frequencies, damping ratios, and mode shapes, i.e., i i and i i 1 n ,) resulting in the following model format z y

    132

    45 66

    11 77 88 99 7 7 8 8
    Az C z

    Bu Du

    (3) (4)

    Input/Output Formulation of the Equation of Motion


    Without the loss of generality, the structural state space formulation is carried out for a one dimensional structure, i.e., a beam. Input Matrix When the beam is excited by a point force p0 and a point moment m0 , the modal equation of motion is

    where Qm i r and Qi r are the generalized modal forcing functions due to point moment and point force inputs, respectively. The general expansion of this forcing function is Qi r

    78

    78

    i t

    7 8@9

    i t 2i i

    7 8A9 7 8B1 7 8 7 8A9 7 8 7 8


    2 i i t Qm i r m0 t Qi r p 0 t
    p

    (5)

    where L is the length of the beam and F r could be viewed as the spatial representation of the forcing function; see Equations (7) and (8). When input to the beam is a point force p0 t acting at location r a, the forcing function f r t is f rt p0 t r a (7)

    and when the input is a point moment m0 t acting at location r f rt m0 t r

    7 8B137 7 8 7 8 8B1DC 7 8 7 8 78 78 1 7 8B1 7 8 F 7 G EPH Q I 8 78 1 7 8B1 7 8 F R 7 G PHE Q I8


    i r F r
    L 0 F r

    i r F r dr

    (6)

    78

    b, the forcing function is (8)

    F r

    The generalized modal forcing functions, using Equations (7) and (8) are Qi r
    p L

    Thus

    i t

    Equation (11) conned to a nite number of modes is the basis for state space modeling of a beam/structure. When a pair of point moments of equal magnitude and opposite sign, resembling piezoelectricpatch actuation, act at r1 and r2 , Equation (11) becomes i t Output Matrix Displacement/velocity measurement When the measured output is displacement or velocity, Equation 13 or its derivative is used to formulate the output equation. w Pt

    W STSY S TVUX U ST S TVU`Y W S T a S Y T U`Y S T S Y T bb c W U c aS T S T c S T c S TBUdY a S T


    i a
    0

    i r r

    a dr (9)

    Qm i r

    i r r b

    b dr

    i r r i b

    L 0

    i r r

    i t 2ii

    2 i i t

    i b m0 t

    aS T S Y T STc S T ST
    b dr i a p0 t

    (10) (11)

    S T c S TBUdYfe a S gT Y a S T h S T
    2 i i t i r2 i r1 m0 t

    (12)

    Note that D in the output equation is a zero matrix.

    S i TpU q S T ST

    Wn P n t

    (13)

    n 1

    Example 0.2: Model of a 2-DOF system Formulate the state space model of the 2DOF springmassdashpot system shown in Figure 1.

    The

    F c2

    c1

    m1 k2 x1

    m2

    k1

    x2

    Figure 1: A 2DOF discrete system

    force F exciting the mass m1 is the input and the displacement of mass m2 , i.e., x2 , is the ouput of the system. The following parameters are considered in the model: k1 5, k2 10, m1 8, and m2 15. 3

    U U

    The massnormalized eigenvalues are 1 The natural frequencies are 1 0 4197 and 2 cies are used to construct the A matrix as O 2 I 2

    The input and output matrices are formulated using the elements of the eigenvectors corresponding to mass m1 where the excitation is inputing the system and mass m2 where the displacement is measured. O Q 0 0 0 1673 0 3114

    and

    w w x y v t u t w t w t t w w t t ww t t w w
    0 0 0 4197 0
    2

    0 1673 0 2274 0 3114 0 1222 . and 2 1 5380. In the absence of damping the natural frequen0 0 0 1 5380 1 0 0 0 0 1 0 0

    tvu w w xy

    (14)

    (15)

    W O

    0 2274

    0 1222 0 0

    (16)

    Strain measurement: Strain, measured by a patch of piezoelectric ceramic, at any small region of a structure is proportional to the curvature at that region. In a beam the curvature is the second derivative of the displacement, i.e., r

    t t

    where , the proportionality factor, is the distance from the neutral axis of the beam. For an individual mode i 2i r i r i t (19) r2 where is a constant. The signal generated by a patch of piezoelectric strain gauge installed along the length of the beam between r r1 and r2 is proportional to the integral of strain over the domain of the gauge, i.e. between r1 and r2 . So the contribution of the mode i to measured output is

    t
    r2

    2 y r2 2 2 i r i t r i 1

    (17) (18)

    Thus the element of the W matrix mapping the modal coordinate system to the measured signal by the i r1 . Note that this is the same as the element of strain measuring device is proportional to i r2 the Q matrix transforming the physical piezoelectricpatch actuation, represented by a pair of moments of 4

    t t dd tt e f g f g e f g f g
    r2 r1

    i r

    2 i r i t dr r2 r1 i r r2 i t r r1 i r2 i r1 i t Wi i t

    (20)

    (21)

    equal magnitude and opposite signs, to the mode i. For collocated arrangements W QT , which similar to the relationship between Q and W when point force is the input and displacement is the output, with collocated arrangement. The model structure described above is based on the classical (proportional) damping assumption, i.e., damping coefcient matrix is proportional to the mass M and stiffness K matrices. Although this is more of a mathematical convenience than physical reality, the assumption is widely used in modeling exible structures. In the absence of classical damping assumption, the statespace model of the exible structure can still be represented in terms of eigenvalues and eigenvectors of the system.

    Model Structure of Systems with NonProportional Viscous Damping


    The equation of motion of an ndegreeoffreedom system in physical coordinates is:

    Multiplying Equation (22) by M

    where X t

    xt is the state vector. This system in the absence of rigid body motion and repeated x t eigenvalues has n pairs of complex conjugate eigenvalues and their corresponding complex conjugate eigenvectors: 1 1 1 1 2 2 2 2

    r sBhut rr sswv

    l h mo l o l p q n km lp
    M x C x K x F
    1

    ijkijkijh
    O M 1K I M 1C

    (22)

    and rewriting it in state space yields O M 1

    (23)

    Transforming the coordinates of the system from X to using the transformation matrix P, i.e., y P, dened as R I R P I (24) 1 1 2 2 results in P 1 AP P 1 Bu

    where the superscripts I and R signify imaginary and real parts of the complex i . P 1 AP is the block diagonal real matrix, shown below, with block entries consisting of real and imaginary parts of the eigenvalues R I R I 1 1 2 2 P 1 AP blockdiag (26) I R I 1 1 2 R 2 The output equation is y P (27)

    x x yx yx x hq| h l k l h tm o h

    x yx y{ x zzz z z z} l o m x p p xzzzv

    (25)

    You might also like