ContSys1 L5 SDOF RespAll

Embed Size (px)

Citation preview

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    1/59

    19/03/20141

    Mechanical Engineering Science 8

    Dr. Daniil Yurchenko

    Response of a second

    order system

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    2/59

    19/03/20142

    Remote position controller

    A device which is used to position aninertia load according to a desired angular

    position from a remote point.

    Input, i

    Angular

    position,

    Output, o

    Position

    transducer

    Error unit

    Electric

    Motor Inertia, J

    Viscous friction, F

    Torque, mo

    motor

    Inertial load

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    3/59

    19/03/20143

    Remote position controller

    Desired angular position input to error unit

    Error unit subtracts actual angular position togive error signal

    Error signal passed to electric motor whichgenerates a torque directly proportional to errorvoltage applied to it

    Such a system is known as a proportional action

    controllersince the controlling action (motortorque in this case) is directly proportional tosystem error

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    4/59

    19/03/20144

    Remote position controller Open loop descriptions of system elements

    Motor

    Where K is a constant of proportionality & m isthe motor torque

    Thus motor torque is proportional to error.

    Leads to

    Km

    Km

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    5/59

    19/03/20145

    Remote position controller

    Applying Newtons second law to inertialload

    Where F is viscous friction constant and Jis mass moment of inertia of load

    Re-arranging

    2

    2

    dt

    dJ

    dt

    dFm OO

    mdt

    dF

    dt

    dJ OO

    2

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    6/59

    19/03/20146

    Remote position controller

    Applying L-transform and assuming zero ICs

    This may be represented by the system blockdiagram in the next slide

    FsJsM

    2

    0 1

    ][2

    2

    mLdt

    dF

    dt

    dJL OO

    )()()( 002 sMsFssJs

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    7/5919/03/20147

    Remote position controller

    The system closed loop equation

    GH1

    GFunctionTransfer

    is

    K

    +

    -

    o

    Proportional

    Controller

    FsJs 21Mi

    1;2

    HFsJs

    KG

    Oi

    KFsJs

    K

    FsJs

    KFsJs

    K

    i

    2

    2

    20

    1

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    8/5919/03/2014

    8

    Remote position controller

    The relationship between system outputand the error is

    and

    Re-arrange to give system characteristicequation:

    )()()( sEss iO

    )()( 2 sFsJs KsO

    )()( 22 sFsJssEKFsJs i

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    9/5919/03/2014

    9

    Standard forms

    Similar to the standard form concept westudied when we worked with vibrating

    systems. In this case system naturalfrequencies and damping ratios will bemade up of a mixture of mechanical andelectronic components rather than springs,

    masses & dampers

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    10/5919/03/2014

    10

    Characteristic equation

    Or

    Leads to standard form

    iFsJssEKFsJs 22 )(

    )()(

    22

    ssJ

    F

    ssEJ

    K

    sJ

    F

    s i

    )(2)(2 222

    ssssEss innn

    where

    J

    Kn

    and

    KJ

    F

    J

    Fn

    22

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    11/5919/03/2014

    11

    Similarly the governing equation becomes

    Or

    Leading to the standard form

    )()(2

    sKsKFsJs iO

    )()(2 222 ssss inOnn

    )()(2

    sJ

    K

    sJ

    K

    sJ

    F

    s iO

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    12/5919/03/2014

    12

    Control system response

    This can be defined using 2 parameters,the error and the output O.

    The error response can be obtained fromsolving the characteristic equation for aspecified input and the output responsecomes from solving the governingequation for a specific input.

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    13/5919/03/2014

    13

    Control system response

    Both the error and the output can each beconsidered to be the sum of two components:

    a transient response and a steady stateresponse.

    The transient response is the complementaryfunction solution to the describing equationand the steady state response is the particularintegral solution (c.f. first order systems)

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    14/59

    19/03/201414

    Control system response

    Thus for the error:

    For a satisfactory control system weshould have:

    And

    statesteadytransienttotal _

    0statesteady _

    transientShould be of minimum duration and

    may need to have limited overshoot

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    15/59

    19/03/201415

    Control system response 1)unit step

    As before the input is defined as:

    We will use the remote position controlleras an example.

    First we will investigate the error responseof this system.

    00 tti , 01 tti ,

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    16/59

    19/03/201416

    Control system response 1)unit step

    Repeating the characteristic equation:

    L-transform of the unit step response is:

    Rearranging:

    s

    tULsi1

    )()(

    )()( 22 sFsJssEKFsJs i

    sKFsJs

    FsJssE

    1)(

    2

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    17/59

    19/03/201417

    Control system response 1)unit step

    Thus:

    The steady-state error can be easily obtainedby using Theorem 3 (Lecture 3):

    Thus,

    sKFsJsFsJssE 1)(

    2

    2

    )(lim)()(lim0

    ssQqtqst

    0lim)(lim)(2

    2

    00

    KFsJs

    FsJsssEe

    ssstst

    0statesteady

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    18/59

    19/03/201418

    Control system response 1)unit step

    The system response will depend on thevalue of

    To show that lets get the characteristicequation, which is obtained by assumingzero input:

    Thus:02 22 nnss

    0)(2 22

    sss Onn

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    19/59

    19/03/201419

    Control system response 1)unit step

    This is a quadratic with solution:

    Consider 3 cases:1

    2 nns

    10

    1

    Underdamped

    Critical

    1 Overdamped

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    20/59

    19/03/201420

    Control system response 1)unit step

    Case

    Both roots are real:

    Then

    1,1 222

    1 nnnn ss

    1

    11

    2)(

    )(

    22

    2

    22

    2

    nnnn

    n

    nn

    n

    i

    O

    ss

    sss

    s

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    21/59

    19/03/201421

    Control system response 1)unit step

    For unit step response

    Which can be written as

    Using the table of inverse L-transform

    1

    ssss

    nnnn

    nO

    1

    11

    )(22

    2

    bsasss nO

    2

    )(

    btatnOO aebe

    baab

    sLt 1

    1)()(2

    1

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    22/59

    19/03/201422

    Control system response 1)unit step

    Which after simplification gives:

    Then the error is:

    And the

    1

    212

    21

    121)(

    se

    set

    tsts

    nO

    212

    0

    21

    12)()()(

    s

    e

    s

    ettttsts

    ni

    0)(lim

    tt

    statesteady

    No Vibrations

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    23/59

    19/03/201423

    Control system response 1)unit step

    Then

    Which leads to

    And for unit step input

    1

    nnns 12

    2,1

    22

    22

    2

    2)(

    )(

    n

    n

    nn

    n

    i

    O

    ssss

    s

    ss

    s

    n

    nO

    1)(

    2

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    24/59

    19/03/201424

    Control system response 1)unit step

    Then

    Then the error is:

    And

    1

    tettt nt

    in 1)()()( 0

    0)(lim

    tt

    statesteady

    tesLt nt

    OOn 11)()( 1

    No Vibrations

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    25/59

    19/03/201425

    Control system response 1)unit step

    Then we have two complex conjugate roots

    Which leads to

    And for unit step input

    10

    22

    2

    2)(

    )(

    nn

    n

    i

    O

    sss

    s

    ssss

    nn

    nO

    1

    2

    )(22

    2

    222,1 111 nnnn is

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    26/59

    19/03/201426

    Control system response 1)unit step

    Then

    te

    sLt n

    t

    OO

    n

    2

    2

    1 1sin1

    1)()(

    10

    21 1tan

    21

    nd

    The system response is oscillatory with

    damped natural frequency and decays

    exponentially

    d

    ttet dd

    t

    On

    sin

    1cos1)(

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    27/59

    19/03/201427

    Control system response 1)unit step

    The error is

    And

    t

    e

    ttt n

    t

    i

    n

    2

    20 1sin1)()()(

    0)(lim

    tt

    statesteady

    10

    0

    2sin1)(

    tt nO

    2

    sin)( tt n

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    28/59

    19/03/201428

    Control system response 1)unit step

    21 1tan

    n

    ncos

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    29/59

    19/03/201429

    Control system response 1)unit step

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    30/59

    19/03/201430

    samethestaysst

    samethestayspM

    samethestaysn

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    31/59

    19/03/201431

    Control system response 1)unit step

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    32/59

    19/03/201432

    Control system response 1)unit step

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    33/59

    19/03/201433

    Control system response 1)unit step

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    34/59

    19/03/201434

    Control system response 1)unit step

    Mp - maximum percent overshoot

    where tp - peak time

    td - delay time. The time required for theresponse to reach half of the final value.

    tr - raise time. The time required for theresponse to rise from 10% to 90% or from0 to 100% of its final value

    %100)(

    )()(

    0

    00

    pp

    t

    M

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    35/59

    19/03/201435

    Control system response 1)unit step

    tssettling time. The time required for theresponse curve to reach and stay within arange about the final value

    (usually 2% or 5%)

    In control system design we are largely

    interested in obtaining a rapid responsewhilst minimising overshoot.

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    36/59

    19/03/201436

    Control system response 1)unit step

    The peak time may be obtained bydifferentiation

    tetetete

    ttette

    ttedt

    d

    dt

    td

    dnt

    dd

    t

    d

    tnd

    t

    n

    d

    d

    dd

    t

    dd

    t

    n

    dd

    tO

    nnnn

    nn

    n

    cos1

    1sinsin

    1cos

    cos1sinsin1cos

    sin1

    cos1)(

    2

    2

    2

    2

    22

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    37/59

    19/03/201437

    Control system response 1)unit step

    Equating to zero yields

    The first peak corresponds to

    p

    nn

    ttdd

    t

    d

    tnO tete

    dt

    td

    sinsin

    1

    0)(

    2

    2

    0sin1 2

    2

    pdd

    n t

    0sin pdt ,...2,1,0, nntpd

    d

    pt

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    38/59

    19/03/201438

    Control system response 1)unit step

    The first peakd

    pt

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    39/59

    19/03/201439

    Control system response 1)unit step

    The maximum overshoot occurs at tp=/d

    sin1

    sin1

    1sin1

    1)()(

    22

    2

    d

    n

    d

    n

    dn

    ee

    et

    d

    dOO

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    40/59

    19/03/201440

    Control system response 1)unit step

    2

    22

    221 1tan

    1tan

    1tan

    22

    2

    2

    2

    2

    2

    2

    2 1sin11

    1

    1

    tan1tansin

    dn

    dn

    eet OO

    sin1

    )()(2

    %100

    21

    eMp

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    41/59

    19/03/201441

    Control system response 1)unit step

    By increasing K we increase the overshoot

    %10021

    eMp KJF

    2 K

    K

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    42/59

    19/03/201442

    Control system response 1)unit step

    The rise time tr ,

    rd

    t

    rO t

    e

    t

    rn

    sin111)( 2

    1)( rO t

    0sin1 2

    rd

    t

    te rn

    0sin rdt

    ,...2,1,0, nntrd 21

    nd

    rt

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    43/59

    19/03/201443

    Control system response 1)unit step

    By increasing K we decrease the rise time

    K

    21

    nd

    rt

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    44/59

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    45/59

    19/03/201445

    Control system response 1)unit step

    In MATLAB

    dcsbsa

    s

    s

    i

    O

    2

    )(

    )(

    ][

    ]00[

    dcbden

    anum

    ),( dennumstep

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    46/59

    19/03/201446

    Unit ramp response of RPC

    In this case we present the system with aunit ramp input:

    The L-transformtti )(

    2

    1)()(

    s

    tLs ii

    22

    2 1)(

    sKFsJs

    FsJssE

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    47/59

    19/03/201447

    Unit ramp response of RPC

    222

    2 1

    2

    2)(

    sss

    sssE

    nn

    n

    J

    Kn

    KJ

    F

    2

    02

    2

    2lim

    )(lim)(

    222

    2

    0

    0

    nnn

    n

    s

    sstst

    s

    s

    ss

    ss

    ssEe

    This is aconstant

    steady

    state error

    )(lim)()(lim0

    ssQqtqst

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    48/59

    19/03/201448

    Unit ramp response of RPC

    Transient error response.

    To find this we need to examine thecomplementary solution to the characteristic

    equation. i.e. the solution to:

    This is identical to the equation we used to find

    the complementary function solution for the stepinput. Therefore the complementary solution isidentical to that case

    02 22 nnss

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    49/59

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    50/59

    19/03/201450

    Unit ramp response of RPC

    For two real roots

    No oscillatory response.

    1

    2222 1

    11)(

    ssss

    nnnn

    nO

    22

    2

    2)(

    )(

    nn

    n

    i

    O

    sss

    s

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    51/59

    19/03/201451

    Unit ramp response of RPC

    For two real roots

    It results in

    No oscillatory response

    1

    nnn

    s 122,1

    22

    2 1

    )( sss n

    n

    O

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    52/59

    19/03/201452

    Unit ramp response of RPC

    Then we have two complex conjugate roots

    Which leads to

    Partial fraction expansion

    10

    222

    2 1

    2

    )(

    sss

    s

    nn

    nO

    222,1 111 nnnn is

    222222

    2

    2

    1

    2 nnnn

    n

    ss

    DCs

    s

    BAs

    sss

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    53/59

    19/03/201453

    Unit ramp response of RPC

    10

    222

    2322

    222

    2

    )2(

    )2)((

    2 sss

    DsCsssBAs

    sss nn

    nn

    nn

    n

    220

    21

    2

    3

    :

    02:

    02:

    0:

    nn

    nn

    n

    Bs

    BAs

    DBAs

    CAs

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    54/59

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    55/59

    19/03/2014

    55

    Unit ramp response of RPC

    10

    222

    2 2

    242)(

    nnn

    nn

    n

    nO

    ss

    s

    s

    ss

    212

    2

    2

    2

    2

    2222

    2

    2

    1

    1tan,1sin

    1

    2

    1sin1

    142

    2

    2

    2

    1421)(

    te

    tet

    ss

    s

    ssssLt

    n

    t

    n

    n

    t

    nn

    nnnnnn

    O

    n

    n

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    56/59

    19/03/2014

    56

    Unit ramp response of RPC

    When t goes to infinity the second term

    vanishes

    10

    12cos

    ,1sin1

    2)(

    21

    2

    2

    te

    tt n

    n

    t

    n

    O

    n

    te

    ttt nn

    t

    n

    i

    n

    2

    20 1sin

    1

    2)()()(

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    57/59

    19/03/2014

    57

    Unit ramp response of RPC

    The error is

    And

    n

    tstatesteady t

    2)(lim

    10

    te

    ttt nn

    t

    n

    i

    n

    2

    2

    0 1sin

    1

    2)()()(

    J

    Kn

    KJ

    F

    2

    K

    F

    K

    J

    KJ

    F

    n

    stst

    2

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    58/59

    19/03/2014

    58

    Unit ramp response of RPC

    At t=0

    If then

    0sin1

    120 2

    nn

    0 0t

    te

    ttt nn

    t

    n

    i

    n

    2

    20 1sin

    1

    2)()()(

    )1(4sin,12cos 222222

  • 8/12/2019 ContSys1 L5 SDOF RespAll

    59/59

    Unit ramp response of RPC

    Either the output response or the errorresponse may be plotted against time