nonlineardynamicmodel-basedadaptivecontrolof asolenoid ... · 2 journal of control science and...
TRANSCRIPT
-
Hindawi Publishing CorporationJournal of Control Science and EngineeringVolume 2012, Article ID 846458, 13 pagesdoi:10.1155/2012/846458
Research Article
Nonlinear Dynamic Model-Based Adaptive Control ofa Solenoid-Valve System
DongBin Lee, Peiman Naseradinmousavi, and C. Nataraj
Department of Mechanical Engineering and Center for Nonlinear Dynamics & Control (CENDAC), Villanova University,Villanova, PA 19085, USA
Correspondence should be addressed to DongBin Lee, [email protected]
Received 6 December 2011; Revised 8 March 2012; Accepted 30 March 2012
Academic Editor: Lili Ma
Copyright © 2012 DongBin Lee et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In this paper, a nonlinear model-based adaptive control approach is proposed for a solenoid-valve system. The challenge is thatsolenoids and butterfly valves have uncertainties in multiple parameters in the nonlinear model; various kinds of physical appear-ance such as size and stroke, dynamic parameters including inertia, damping, and torque coefficients, and operational parametersespecially, pipe diameters and flow velocities. These uncertainties are making the system not only difficult to adjust to the environ-ment, but also further complicated to develop the appropriate control approach for meeting the system objectives. The main con-tribution of this research is the application of adaptive control theory and Lyapunov-type stability approach to design a controllerfor a dynamic model of the solenoid-valve system in the presence of those uncertainties. The control objectives such as set-pointregulation, parameter compensation, and stability are supposed to be simultaneously accomplished. The error signals are first for-mulated based on the nonlinear dynamic models and then the control input is developed using the Lyapunov stability-type analysisto obtain the error bounded while overcoming the uncertainties. The parameter groups are updated by adaptation laws using aprojection algorithm. Numerical simulation results are shown to demonstrate good performance of the proposed nonlinear model-based adaptive approach and to compare the performance of the same solenoid-valve system with a non-adaptive method as well.
1. Introduction
In order to achieve advanced automation [1] in systemssuch as marine vessels or ship-based machinery system [2],solenoid actuators and valves are often used [3] to increasesurvivability and capability. One typical type of actuatordriven by solenoids is shown in Figure 1, which is operated bythe electromagnetic force. The electric-driven solenoid valvesystem [4] and its sophisticated control can provide high lev-els of automation in large systems. The useful function of thesolenoid-valve, once an electrical signal (current or voltage)is applied, is to activate a mechanical motion such as dis-placement or rotation via the solenoid magnetic forces andtorques. The proportional solenoids normally require inte-grated electronics for controlling the plunger to give such asignal. Hydrodynamic torque of a butterfly valve comprisesthe core knowledge of fluid valve system design [5], andit is known that most of the valves in real systems havestrongly nonlinear characteristics between the force and
displacement [6, 7]. The use of an intelligent approach [8, 9]such as adaptive, robust, optimal, or nonlinear control of theactuator-valve machinery systems will benefit a wide spec-trum of nonlinear systems, compensating for nonlinearities[10] and dynamic characteristics. This approach will not onlydecrease the amount of cost and casualties but also improvethe performance of the mechatronic system. To investigatethe particular application, it is important to emphasizethe nonlinear dynamic modeling analysis of such actuator-valve systems because the accuracy and reliability of thesesystems depend highly on the mathematical system modeling[11] and its validation. In [12], the authors developed andanalyzed the nonlinear dynamic model of a solenoid-valvesystem; the reader is also referred to [13, 14] for recent mod-eling and analysis of solenoid actuators.
This paper will focus on model-based nonlinear adaptivecontrol of an actuator-butterfly valve. The solenoid-valvesystem is described based on the exact model knowledgeof the system. Figure 1 shows the integrated system, which
-
2 Journal of Control Science and Engineering
Pinion
Stem
Disk
Plu
nge
r
Coi
l
Coi
l
Yoke
k B1
V0 DpVJ
x
α ∼ 5
Figure 1: System Configuration.
consists of an electric-driven solenoid and a butterfly valve.The valve operates by solenoids that use a magnetic coilto move a movable plunger connected with the valve stemby means of a gear train and linkage. The control inputis designed by substituting the current signal from themodel of the electromagnetic force, pulling the plunger, andthen controlling the angular position of the butterfly valve.The system has uncertainties in multiple parameters in thedynamic model, which requires the system to continuouslyadjust to the environment and consequently requires adap-tation for sustainability and capability. The integrated systemis highly nonlinear in addition to its parameter uncertainties.Hence, an adaptation law is proposed [8, 9, 17] and anadaptive control method is developed for the solenoid-valvesystem in multiple parametric uncertainties. A closed-loopstable controller is designed for the set-point trajectorytracking by introducing a Lyapunov-based stability analysis[9] based on the error signals of the nonlinear solenoid-valvesystem. The numerical results in the simulation are used forinitial verification and performance evaluation.
2. Model-Based Nonlinear System
2.1. System Model. The dynamic equations of motion of theplunger and butterfly valve are given by [15]
mẍ + B1ẋ + kx = Fmag − Fc,J ä + B2α̇ = rgFc − Ttot,
(1)
where x(t) is the displacement of the solenoid plunger, α(t) isthe angle of butterfly disk, and Ttot(t) is the sum of the hydro-dynamic and bearing torques expressed as Ttot = Tb + Th.The bearing torque is given as Tb = (π/8)μDsD2pΔPvCR(α),where the bearing toque coefficient, CR(α) =
√C2L + C
2D =√
(1.1 4√
sin((α/90)3180))2 + ( 3√
cosα)2, is obtained from the
valve modeling (see [5, 15]) and the two subterms, lifting
force CL = 1.1 4√
sin((α/90)3180) and drag force CD = 3√cosα,are nonlinear functions of the valve angle rotation α(t). Thehydrodynamic torque Th is obtained by reviewing three-dimensional hydrodynamic torque coefficient based on [7,16] as Th = (8/3π)ρD3pV 2OTc(α)[(VJ /VO)(α)]2, where bothTc(α) and (VJ /VO)(α) depend on the closing angle (α) of thebutterfly valve and D3p term is a nonlinear term accordingto the pipe size. Solving the two equations in (1) with thecontact force, Fc(t), and substituting the magnetic force,Fmag(t), into the equation yields
(m +
J
r2g
)rg ẍ +
(B1 +
B2r2g
)rg ẋ + krgx
= rg C2N2
2(C1 + C2x)2 i
2 − Ttot,(2)
where the magnetic force Fmag = (C2N2/2(C1 + C2x)2)i2used to lift the plunger of the solenoid actuator. The actuatoris a current-controlled solenoid [15], proportional to thesquare of the current i(t), and C1, C2 are reluctances ofthe magnetic paths, obtained from the geometry of electricactuator [4, 15]. It is assumed that the pinion and the valveare moving at the same speed, that is, the gear ratio is 1 : 1 andx(t) = rgα is the simple geometric relationship between thedisplacement of the pinion and the valve angle. Hence, thecurrent source i2(t) is substituted for designing the closed-loop control input, u(t), and then the following equation isobtained:
(m +
J
r2g
)rg ẍ +
(B1 +
B2r2g
)rg ẋ + krgx
= rg C2N2
2(C1 + C2x)2 u− Ttot.
(3)
For the subsequent controller design, multiplying (3)with the inverse term of the control input, 2(C1 + C2x)
2/(rgC2N2), yields a compact form of the dynamic equation as
M(x, θ)ẍ + C(x, θ)ẋ + D(x, θ)x = u− B(Tb + Th), (4)
where θ describes a lumped expression of parametersobtained from (3) and (4), where each parameter is shownin Table 2 and the substituted terms are defined as follows:
M(x, θ) =(m +
J
r2g
)rgB, B = 2(C1 + C2x)
2
rgC2N2,
C(x, θ) =(B1 +
B2r2g
)rgB, D(x, θ) = krgB.
(5)
2.2. Error Signals Formulation. The following set-point con-trol approach is used. Let xd(t) define the set-point trajectoryand then the error can be defined as
e ≡ xd − x, ė = ẋd − ẋ, ë = ẍd − ẍ, (6)
-
Journal of Control Science and Engineering 3
0 0.5 1 1.5 2 2.5 3 3.5 40
10
20
30
40
50
60
70
80
90
α◦ :
val
ve r
otat
ion
an
gle
(deg
)
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 2: Valve rotation angle α(t) (adaptive approach).
where ẋd(t) and ẍd(t) are the first and second time derivativesof xd(t), which are assumed to be bounded. Premultiplyingë(t) in the last error signal of (6) with M(x, θ) yieldsM(x, θ)ë = M(x, θ)ẍd −M(x, θ)ẍ. Substituting M(x, θ)ẍ in(4) into the above equation produces
M(x, θ)ë = M(x, θ)ẍd + C(x, θ)ẋ + D(x, θ)x − u+ B(Tb + Th).
(7)
A filtered error signal and its derivative are defined as
r ≡ ė + λ1e, ṙ = ë + λ1ė, (8)where λ1 ∈ �+ is a positive adjustable control gain. Multi-plying (8) with M(x, θ) and then substituting for M(x, θ)ë(t)in (7) yields
M(x, θ)ṙ = M(x, θ)ẍd + C(x, θ)ẋ + D(x, θ)x − u
+ B(Tb + Th) + M(x, θ)λ1ė +1r2ge − 1
r2ge,
(9)
where the last term e(t)/r2g is added and subtracted for fur-ther development of the control design based on Lyapunov’smethod.
3. Lyapunov-Based Adaptive Feedback Control
Let V(t) be a Lyapunov candidate function
V = 12
(rTMr + eTe + eTα eα + Θ̃
TΓ−1Θ̃)
, (10)
where the last term of the Lyapunov candidate function, Γ =γIp×p, is a constant diagonal matrix with the gain value γ,Ip×p is a p× p identity matrix, and the parameter estimationerror, Θ̃, is defined as Θ̃ = Θ− Θ̂, where Θ ∈ �p is a known
Table 1: Regression and parameter estimation terms.
No. Terms No. Terms
W101 ẍd W113 CR(α)x
W102 ẍdx W114 Tc(α)[VJ(α)/VO]2x2
W103 ẍdx2 W115 CR(α)x2
W104 ẋ W116 ẋė
W105 ẋx W117 ẋe
W106 ẋx2 W118 xẋė
W107 x W119 xẋe
W108 x2 W120 e
W109 x3 W121 λ1ė
W110 Tc(α)[VJ (α)/VO]2 W122 λ1xė
W111 CR(α) W123 λ1x2ė
W112 Tc(α)[VJ(α)/VO]2x — —
Θ̂101 M̂s2Ĉ21/(Ĉ2N̂2) Θ̂113 T̂b14Ĉ1/(r̂g N̂2)
Θ̂102 M̂s4Ĉ1/N̂2 Θ̂114 T̂h12Ĉ2/(r̂g N̂2)
Θ̂103 M̂s2Ĉ2/N̂2 Θ̂115 T̂b12Ĉ2/(r̂g N̂2)
Θ̂104 Ĉs2Ĉ21/(Ĉ2N̂2) Θ̂116 M̂s2Ĉ1/N̂2
Θ̂105 Ĉs4Ĉ1/N̂2 Θ̂117 M̂s2Ĉ1/N̂2
Θ̂106 Ĉs2Ĉ2/N̂2 Θ̂118 M̂s2Ĉ2/N̂2
Θ̂107 k̂2Ĉ21/(Ĉ2N̂2) Θ̂119 M̂s2Ĉ2/N̂2
Θ̂108 k̂4Ĉ1/N̂2 Θ̂120 1/r2gΘ̂109 k̂2Ĉ2/N̂2 Θ̂121 M̂s2Ĉ21/(Ĉ2N̂2)
Θ̂110 T̂h12Ĉ21/(r̂g Ĉ2N̂2) Θ̂122 M̂s4Ĉ1/N̂2
Θ̂111 T̂b12Ĉ21/(r̂g Ĉ2N̂2) Θ̂123 M̂s2Ĉ2/N̂2
Θ̂112 T̂h14Ĉ1/(r̂g N̂2) — —
Table 2: List of simulation parameters.
m J rg B1 B2 k N ΔPv
0.1 1.04e−6 1e−2 10 20 4e2 8.8e3 0.5
C1 C2 ρ Dp VO μ Ds p
1.57e6 6.32e8 1e3 5′′ ∼ 8′′ 3.7 0.1 0.1 23
constant parameter vector and Θ̂ ∈ �p is the estimatedconstant parameter vector (see Table 1). Differentiating (10)yields
V̇ = rTMṙ + 12rTṀr + eT ė +
1r2geT ė − Θ̃TΓ−1 ˙̂Θ, (11)
where the time derivative of the inertia matrix is obtainedas Ṁ = (m + J/r2g )(4(C1 + C2x)/N2)ẋ, where Ḃ = (4(C1 +C2x)/rgN2)ẋ, eTα ėα = (1/r2g )eT ė as eα ≡ αd − α = xd/rg −x/rg = e/rg , ėα ≡ α̇d − α̇ = ẋd/rg − ẋ/rg = ė/rg , in which theerror signals of the valve angle, eα(t), can be defined using
the geometric relationship and ˙̃Θ(t) = − ˙̂Θ comes from thedefinition of Θ̃.
-
4 Journal of Control Science and Engineeringx:
plu
nge
r di
spla
cem
ent
(m)
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
00 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 3: Plunger displacement x(t) (adaptive approach).
0 0.5 1 1.5 2 2.5 3 3.5 40
10
20
30
40
50
60
70
80
90
α◦ :
val
ve r
otat
ion
an
gle
(deg
)
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 4: Valve rotation angle α(t) (no adaptation scheme).
3.1. Design of Control Input. Substituting M(x, θ)ṙ(t) into(11) yields
V̇ = rT(Mẍd + Cẋ + Dx − u + B(Tb + Th) + Mλ1ė + 1
r2ge
)
− (ė + e)Te
r2g+
12rTṀr + eT ė +
1r2geT ė − Θ̃TΓ−1 ˙̂Θ,
(12)
where the last term in (9) premultiplied by r(t) came out ofthe parenthesis in (12) and is used for the definition of r(t)
0.015
0.01
0.005
00 0.5 1 1.5 2 2.5 3 3.5 4
x: p
lun
ger
disp
lace
men
t (m
)
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 5: Plunger displacement x(t) (no adaptation scheme).
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
−0.01
r(t)
: filt
ered
err
or
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 6: Filtered Error r(t) (adaptive approach).
in (8). Then, combining the parameterized terms in (12) andsubstituting them into WΘ yields
WΘ =Mẍd + Cẋ + Dx + B(Tb + Th) + Mλ1ė + 1r2ge +
12Ṁr,
(13)
where W(ẍd, ẋ, x, r,α, ė, e) ∈ �1×p is a known regressionvector, which is shown in the left side of Table 1 via theprocess given later (see (17)) and Θ as the nominal value ofthe lumped parameter vector. Rearranging (12) produces
V̇ = rT(WΘ− u)− eTe
r2g+ eT ė − Θ̃TΓ−1 ˙̂Θ, (14)
-
Journal of Control Science and Engineering 5
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
−0.01
e(t)
: dis
plac
emen
t er
ror
(m)
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 7: Displacement tracking error e(t) (adaptive approach).
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
−0.01
x′: t
he
rate
of
disp
lace
men
t (m
/s)
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 8: The rate of displacement ẋ(t).
where ėTe/r2g is canceled in the last second term in (12),having the opposite sign because they are scalar, ėTe = eT ė.
The control input can be designed based on Lyapunovstability analysis, making V̇ negative definite to be shown inthe end, as
u =WΘ̂ + k1r + e, (15)
where r(t) is a feedback error term, k1 is a positive constantas the control gain, e(t) is another feedback error term addedto cancel the term having the opposite sign, eT ė, outside theparenthesis by utilizing the definition of r(t) given in (8), andWΘ̂ captures the uncertainties associated with the elements
0 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
0
50
100
150
u(t
): c
ontr
ol in
put
(A2)
Figure 9: Control input u(t) equivalent to square current i2(t)[A2].
0
1
2
3
4
5
6
7
8×104
Fm
ag: e
lect
rom
agn
etic
forc
e (N
)
0 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 10: Electromagnetic force Fmag(t).
of M, C, D, B(Tb1 +Th1), Mλ1, 1/r2g , and Ṁ, which is definedas
WΘ̂ = M̂ẍd+Ĉẋ+D̂x+B̂(T̂b1CR(α)+T̂h1Tc(α)
[VJVO
(α)]2)
+ M̂λ1ė+1r̂2ge+
12
˙̂Mr,
(16)
where the estimated parameter sets are given as M̂(x, θ) =(m̂ + Ĵ/r̂2g )r̂g B̂, where B̂ = 2(Ĉ1 + Ĉ2x)
2/r̂g Ĉ2N̂2, Ĉ(x, θ) =
(B̂1 + B̂2/r̂2g )r̂g B̂, D̂(x, θ) = k̂r̂g B̂, T̂b1 = (π/8)μ̂D̂sD̂2pΔP̂v,
-
6 Journal of Control Science and Engineering
0
50
100
150
200
250
Tt=Th
+Tb: t
otal
torq
ue
(Nm
)
0 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 11: Total torque Tt(t).
and T̂h1 = (8/3π)ρ̂D̂3pV̂ 2O. In order to develop the estimateparameter vector Θ̂ in (16), we need to first define theregression and the estimate terms. Thus, the first term, M̂ẍd,in (16) can be defined as
M̂ẍd =(m̂ +
Ĵ
r̂2g
)r̂g
2(Ĉ1 + Ĉ2x
)2
r̂g Ĉ2N̂2ẍd
= M̂s 2Ĉ21
Ĉ2N̂2ẍd + M̂s
4Ĉ1N̂2
xẍd + M̂s2Ĉ2N̂2
x2ẍd
=[ẍd xẍd x2ẍd
]︸ ︷︷ ︸
W101∼W103
⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣
M̂s2Ĉ21Ĉ2N̂2
M̂s4Ĉ1N̂2
M̂s2Ĉ2N̂2
⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦
︸ ︷︷ ︸Θ̂101∼Θ̂103
,
(17)
where M̂s = (m̂ + Ĵ/r̂2g ), Ĉs = (B̂1 + B̂2/r̂2g ) (this is shown inTable 1), and r̂g , Ĉ1, or Ĉ2 are canceled. W101 ∼W103 are themeasurable regression terms and Θ̂101 ∼ Θ̂103 is the estimat-ed parameters, defined in Table 1, respectively. Similarly to(17), the rest of the terms in (16) are also given in Table 1.
3.2. Online Adaptation Laws for Parameter Updates. Thefollowing is constructed to define the known upper andlower bounds but with an unknown parameter of Θ̂(t) in thesense that
Θ̂j≤ Θ̂ j(t) ≤ Θ̂ j , (18)
0
0.005
0.01
0.015
0.02
0.025
0.03
Tc(α
): to
rqu
e co
effici
ents
0 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 12: Hydrodynamic torque coefficient Tc(α).
0
20
40
60
80
100
120
140
160
VJ/VO
(α):
th
e ra
tio
of m
ean
flow
an
d je
t ve
loci
ties
0 0.5 1 1.5 2 2.5 3 3.5 4
= 5= 6
78
Time (s)
DpDp
==
DpDp
Figure 13: The ratio of VJ and VO: (VJ /VO)(α).
where Θ̂ j(t) are the estimated parameters as shown in Table 1
and Θ̂j
and Θ̂ j are the lower and upper bounds of the
estimated parameters, respectively, which will be set to the
amount of percentage of their true values. The vector ˙̂Θ j(t)is designed to update using a projection-based algorithm as
˙̂Θ j = Proj
{ΓWTr, Θ̂ j
}, (19)
-
Journal of Control Science and Engineering 7
0 1 2 3 4
0 1 2 3 4
0 1 2 3 4
0 1 2 3 4
1.1
1.2
1.3
1.4
1.5×10−5 θ101
8.944
8.946
8.948
8.95
8.952
θ102×10−3
1.8027
1.8027
1.8027
1.8027
1.8027
θ103
20.1068
20.1069
20.107
20.1071θ104
Time (s) Time (s)
Time (s)Time (s)
= 5= 6
78
DpDp
==
DpDp
= 5= 6
78
DpDp
==
DpDp
Figure 14: Parameter estimates: Θ̂101 ∼ Θ̂104.
where Proj{·} is the projection operator [8] and each lumpedparameter is adaptively updated using the adaptation laws[17] for online estimation of unknown parameter as follows:
Proj{
˙̂Θ j
}= Proj
{ΓWTr, Θ̂ j
}
=
⎧⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩
ΓWTr if Θ̂ j > Θ̂j, Θ̂ j < Θ̂ j ,
ΓWTr if Θ̂ j = Θ̂j, if ΓWTr > 0,
ΓWTr if Θ̂ j = Θ̂ j , if ΓWTr ≤ 0,
0 elsewhere.(20)
Thus, substituting u(t) in (15) into V̇ in (13) yields
V̇ = rTWΘ̃− rTk1r − (ė + e)Te − eTe
r2g+ eT ė − Θ̃TΓ−1 ˙̂Θ
= − rTk1r −(
1 +1r2g
)eTe + Θ̃T
{WTr − Γ−1 ˙̂Θ
}.
(21)
Here, WΘ̃ is defined as
WΘ̃ = M̃(ẍd + λ1ė) + C̃ẋ + D̃x + B̃T̃h1Tc(α) VJVO
(α)
+ B̃T̃b1CR(α) + r̃g e +12˜̇Mr,
(22)
where M̃ = M − M̂, C̃ = C − Ĉ, D̃ = D − D̂, B̃T̃h1 =BTh1 − B̂T̂h1, B̃T̃b1 = BTb1 − B̂T̂b1, r̃g = 1/r2g − 1/r̂2g ,
-
8 Journal of Control Science and Engineering
Time (s)
Time (s)
Time (s)
Time (s)
0 1 2 3 4
0 1 2 3 4
0 1 2 3 4
0 2 4
1.6207
1.6207
1.6207
1.6207
1.6207
×104 θ105
3.2659
3.2659
3.2659
×106 θ106
0.04
0.0401
0.0402
0.0403
0.0404θ107
32.4127
32.4128
32.4128
32.4128
32.4128
θ108
= 5= 6
78
DpDp
==
DpDp
= 5= 6
78
DpDp
==
DpDp
Figure 15: Parameter estimates: Θ̂105 ∼ Θ̂108.
and ˜̇M = Ṁ − ̂̇M, in which T̂h1 = (8/3π)ρ̂D̂3pV̂ 2O and T̂b1 =(π/8)μ̂D̂sD̂2pΔP̂v. Actually, Θ̃(t) can be expressed, for exam-
ple, using the definitions of Θ̂ in (16), as follows: owing to thesubsequent adaptation law, the time derivative of V(t) yieldsa negative definite function except the origin and upperbound by
V̇ ≤ −k1‖r‖2 −(
1 +1r2g
)‖e‖2, (23)
which can be written as
V̇ ≤ −k2‖z‖2, (24)where k2 = min{k1, (1 + 1/r2g )} is a positive constant and
z =[rT , eT
]T. (25)
Using Barbalat’s lemma [18], the set-point tracking error‖z(t)‖ → 0, thus ‖r(t)‖ → 0 and ‖e(t)‖ → 0 as t → ∞.
Remark 1. According to the analysis from (11) to (25) ofV̇ , the property of V , and the control law of (15) with theparameter updates of (19) and the projection-based methodof update laws of (20), it is straightforward to derive aconclusion that the tracking error vector z(t) in (25) is drivento zero. Thus, the set-point errors r(t), e(t), and eα(t) alsovanish and the parameter estimation error vector Θ̃ in (10)is bounded where Θ̂, defined after (10), is bounded due tothe projection-based update method and the constant knownparameter, Θ. Owing to the bounds of r(t), e(t) in (8), ė(t)is bounded and then x(t), resulting in α(t), and ẋ(t) arebounded, respectively, where all desired trajectories such asxd(t) and ẋd(t) are assumed to be bounded. M(·), B(·), C(·),and D(·) matrices in (4) are bounded because θ (due toΘ) and x(t) are bounded, and Ḃ and Ṁ after (11) are alsobounded owing to the bounds of ẋ(t). Ttot is thus boundedbecause α(t) is bounded. Hence, W(·) and the control inputu(t) are bounded and, thus, the current is bounded. Thisleads to the boundedness of ẍ(t) in the dynamic model given
-
Journal of Control Science and Engineering 9
0 1 2 3 4
0 1 2 3 4 0 2 4
0 1 2 3 4
6531.5313
6531.5313
6531.5313
6531.5313
×10−8
θ109 θ110
θ111 θ112
Time (s) Time (s)
Time (s) Time (s)
= 5= 6
78
DpDp
==
DpDp
= 5= 6
78
DpDp
==
DpDp
0.2
0.4
0.6
0.8
1
0.5
1
1.5
2
2.5
0
200
400
600
800
Figure 16: Parameter estimates: Θ̂109 ∼ Θ̂112.
in (4), which enables the boundedness of ë(t) in (6), and thenthe set-point tracking error dynamics ṙ(t) in (8) is bounded.Therefore, we can conclude that all signals are bounded.
Remark 2. WΘ̂ in (16) where the known regression W(ẍd, ẋ,x, r,α, ė, e) ∈ �1×23 terms and the parameter estimates Θ̂ ∈�23 are given in this system as in Table 1.
4. Simulation Results
Based on the dynamic model in (4), the numerical sim-ulation is performed to verify the proposed controllerswith consistently changing parameter values. The parametervalues can be divided into two categories: operational valuessuch as Dp and VO and uncertain values such as B1, B2,and μ. After the flow velocity VO is kept on 3.7 [m/s], thepipe diameter Dp shown in Table 2 is used to vary from 5inches to 8 inches as Dp = {5.0, 6.0, 7.0, 8.0} [in], whereDp = {0.1270, 0.1524, 0.1778, 0.2032}[m], by assuming thatthe pipe or transmission lines are different according totheir applications. With each Dp size, the variations of all
the parameters are set to 30% for the simulation resultsprovided here. The determined parameter vectors Tc(α) and[VJ /VO](α) of the butterfly valve model are borrowed from[16] for the simulation given as look-up tables from theexperimental data. The amount of upper and lower variationof the unknown parameter sets is 30% of their real values.The control gains were chosen selectively as γ = 10, k1 =250, and λ1 = 1.0 for all cases. MATLAB and Simulink areused for the simulation. The desired set-point distance of thesolenoid, xd(t), is given as A · (1 − e−Bt), where A = 0.0148and B = 5.
A typical parameter set for this simulation is given byTable 2. Figures 2 and 3 show the rotation of the valve angleand the actual displacement of the plunger, respectively. Thisadaptive control approach shows better results compared tothe results obtained from the previous research [15] usingnonadaptive method, which are shown in Figures 4 and 5 forthe displacement of the plunger and the angle of the butterflyvalve. The developed mathematical model is the same asthat of adaptive method based on the nonlinear models andthe new specific approach in this paper is that the error
-
10 Journal of Control Science and Engineering
0
0.5
1
1.5
2
0
0.5
1
1.5
2
1
2
3
4
4.4725
4.473
4.4735
4.474
0 1 2 3 4
0 1 2 3 4 0 2 4
0 1 2 3 4
×10−5 ×105
×10−3 ×10−3
θ113 θ114
θ115 θ116
Time (s) Time (s)
Time (s) Time (s)
= 5= 6
78
DpDp
==
DpDp
= 5= 6
78
DpDp
==
DpDp
Figure 17: Parameter estimates: Θ̂113 ∼ Θ̂116.
signals are formulated by introducing the desired trajectoryto reducing the displacement error in the adaptive methodwhile overcoming the complicated parametric uncertainties.And also the response of the adaptive method shows that themotion process is more consistent in both variables, showingoverdamped phenomena and short rising times, and thenit quickly reaches the steady state. The actual movement issmoother and faster while all the parameters are varying 30%in each flow velocity.
Figures 6 and 7 are the filtered error r(t), defined in (8),and set-point error e(t), given in (6), respectively. Figure 8shows the rate of the displacement to Figure 3. From Figure 8,all figures are presented without the notation of the adaptiveapproach. Figure 9 shows the control input u(t) of the sole-noid actuator for each pipe diameter, Dp, designed in (15)with the nonlinear adaptive controller by substituting thesquare of the current, i2(t). The electromagnetic force Fmag(t)given in (2) and the total torque Tt(t) given in (1) by sum-ming up the hydrodynamic and bearing torques are plotted in Figures 10 and 11, respectively. The torque coefficient Tc(t)
shown in Figure 12 and the ratio of input and output jetvelocities shown in Figure 13 are changed for every 5◦ of thebutterfly angle. As the strokes are increased by the controlinput, the angles get larger and then accordingly the valueof the inlet jet velocity increases, which affects the slowermotion of the strokes and angles but the variables (strokesand angles) reach the steady state and the ratio as well.
The challenge is that most parameter terms in Table 1 arecombined and lumped together due to the dynamic modelin the presence of parameter uncertainty and the model iscomplicated owing to the control objectives, set-point regu-lation, and parametric adaptation. As given in the right sideof Table 1, the unknown bounded parameter estimate vectorΘ̂ ∈ �23 is shown in Figures 14, 15, 16, 17, 18, and 19. It canbe seen that the parameters such as 101, 110, 111, 113, 116,118, and 121 are quickly updated in the form of premultipli-cation by its regression term, W and go to steady state. Thus,they are more parametric-centric terms and related to thefiltered error r(t). The parameters such as 102, 103, 104, 105,117, 119, and 122 are updated according to the motion of
-
Journal of Control Science and Engineering 11
4.473
4.473
4.473
4.4731
1.8027
1.8027
1.8027
1.8027
1.8027
1.8027
1.8027
1.8027
1.8027
1
1
1
1
1
0 2 4 0 1 2 3 4
0 1 2 3 40 1 2 3 4
×104
×10−3 θ117 θ118
θ119 θ120
Time (s) Time (s)
Time (s) Time (s)
= 5= 6
78
DpDp
==
DpDp
= 5= 6
78
DpDp
==
DpDp
Figure 18: Parameter estimates: Θ̂117 ∼ Θ̂120.
the angle and stroke of the solenoid-butterfly valve system.Some parameters such as 106, 112, 114, and 115 are notupdated but other parameters such as 107, 108, and 120 keepgetting updated. Any parameters changing their values mustbe related to the control objective because these terms areincorporated into the control input in the form of the esti-mates.
5. Conclusion and Future Work
For developing advanced automation systems such as ship-based hydraulic systems, a typical solenoid-butterfly valve,which is driven by electromagnetic, fluid mechanics, andhydrodynamic forces and torques, is chosen as a continuationof previous research and an adaptive stable control approachwith adaptation laws is developed accounting for uncertain-ties in multiple parameters on the nonlinear dynamic model.A stable adaptive controller of the solenoid-valve system isdesigned positioning the angle of the butterfly valve via aLyapunov-based approach. The approach yields bounded
error while adapting to the environment in the presence ofcomplex uncertainties such as different physical appearances,uncertain parameters, operational characteristics, and para-metric nonlinear dynamic models.
The parameter estimation for the unknown boundedparameters is performed using a projection algorithm whoseoutput yields the upper and lower bounds. Numericalsimulation is used to verify the performance of the proposedapproach to show its effectiveness by comparing to thesame dynamic model without adaptation from the previousresearch; when compared to the nonadaptive method, theresponses of the plunger displacement and the rotatingangle are steadier, smoother, and faster. Future work willbe focused on demonstrating the results of hardware-in-the-loop or experiments for the nonlinear solenoid valve systemas well as applying the suggested adaptive method based onLyapunov-based control approach to the real-world system.Further research on developing control techniques usingrobust or optimal method would be continued to overcomenonlinearities such as hysteresis or nonlinear dynamics.
-
12 Journal of Control Science and Engineering
1.1
1.2
1.3
1.4
1.5
8.9459
8.9459
8.9459
8.9459
8.9459
1.8027
1.8027
1.8027
1.8027
1.8027
0 2 4 0 1 2 3 4
0 0.5 1 1.5 2 2.5 3 3.5 4
×10−3×10−5 θ121 θ122
θ123
Time (s) Time (s)
Time (s)
= 5= 6
78
DpDp
==
DpDp
Figure 19: Parameter estimates: Θ̂121 ∼ Θ̂123.
Nomenclature
B1, B2: Damping coefficients of thesolenoid and butterfly valve,respectively (Ns/m)
C1, C2: Reluctances of the magnetic pathsobtained from plunger geometry ofsolenoid actuator
CR(α): Bearing torque coefficient as afunction of the valve angle (α)
Ds: Stem diameter (m)Dp: Pipe diameter (inch or m)Fmag: Magnetic force (N)Fc: Contact force or resultant force
Fr(N)i: Current of solenoid actuator (A)J : Inertia moment (kgm2)k: Spring stiffness (N/m)m: Mass of solenoid plunger (kg)μ: Friction coefficient of bearing areaN : Number of turns of the coilΔPv: Valve differential pressure (psi)p: Number of estimatesrg : Radius of pinion gear (m)ρ: Fluid mass density (kg/m3)u: Control input (A2)
Tc(α): Hydrodynamic torque coefficient as afunction of the valve angle (α)
Tb, Th: Bearing and hydrodynamic torques,respectively (Nm)
VJ , VO, (VJ /VO)(α): Jet velocity, mean flow velocity, andtheir ratio as a function of the valveangle (α), respectively.
Acknowledgments
This research is supported by the Office of Naval Research(N00014-08-1-0435), which the authors gratefully acknowl-edge. Thanks are in particular due to Mr. Anthony SemanIII. The authors would also like to thank Dr. Stephen Mastroand Mr. Frank Ferrese of Naval Surface Warfare Center(NSWC, Philadelphia) for help with many aspects of thepaper. They are grateful to the anonymous reviewer forcritical comments, which led to substantial improvement ofthe paper.
References
[1] A. Seman, “Adaptive automation for machinery control,” inProceedings of the Office of Naval Research (ONR) Presentation,2007.
-
Journal of Control Science and Engineering 13
[2] R. Hughes, S. Balestrini, K. Kelly, N. Weston, and D. Mavris,“Modeling of an integrated reconfigurable intelligent system(IRIS) for ship design,” in Proceedings of the ASNE Ships & ShipSystems Symposium, 2006.
[3] P. Tich, V. Mark, P. Vrba, and M. Pechoucek, “Chilled watersystem automation,” Rockwell Automation Case Study Report,2005.
[4] J. R. Brauer, Magnetic Actuators and Sensors, Wiley IEEE Press,Hoboken, NY, USA, 2006.
[5] Z. Leutwyier and H. Dalton, “A CFD study of the flow field,resultant force, and aerodynamic torque on a symmetric diskbutterfly valve in a compressible fluid,” Journal of PressureVessel Technology, vol. 130, Article ID 021302, pp. 1–10, 2008.
[6] T. Sarpkaya, “Oblique impact of a bounded stream on a planelamina,” Journal of the Franklin Institute, vol. 267, no. 3, pp.229–242, 1959.
[7] T. Sarpkaya, “Torque and cavitation characteristics of butterflyvalves,” Journal of Applied Mechanics, vol. 29, pp. 511–518,1961.
[8] M. Krstic, I. Kanellakopoulos, and P. Kokotovic, Nonlinearand Adaptive Control Design, Series on Adaptive and LearningSystems for Signal Processing, John Wiley & Sons, New York,NY, USA, 1st edition, 1995.
[9] D. B. Lee, C. Nataraj, and P. Naseradinmousavi, “Nonlinearmodel-based adaptive control of a solenoid-valve system,” inProceedings of the ASME Dynamic Systems and Control Con-ference (DSCC ’10), pp. 1–8, Boston, Mass, USA, September2010.
[10] J. Pohl, M. Sethson, P. Krus, and J.-O. Palmberg, “Modellingand validation of a fast switching valve intended for combus-tion engine valve trains,” Proceedings of the Institution of Mech-anical Engineers. Part I: Journal of Systems and Control Engi-neering, vol. 216, no. 2, pp. 105–116, 2002.
[11] C. Nataraj, Vibration of Mechanical Systems, Cengage, 1st edi-tion, 2011.
[12] P. Naseradinmousavi and C. Nataraj, “Nonlinear mathemati-cal modeling of butterfly valves driven by solenoid actuators,”Applied Mathematical Modelling, vol. 35, no. 5, pp. 2324–2335,2011.
[13] M. K. Zavarehi, P. D. Lawrence, and F. Sassani, “Nonlinearmodeling and validation of solenoid-controlled pilot-operat-ed servovalves,” IEEE/ASME Transactions on Mechatronics, vol.4, no. 3, pp. 324–334, 1999.
[14] R. R. Chladny, C. R. Koch, and A. F. Lynch, “Modeling auto-motive gas-exchange solenoid valve actuators,” IEEE Transac-tions on Magnetics, vol. 41, no. 3, pp. 1155–1162, 2005.
[15] C. Nataraj and P. Mousavi, “Nonlinear analysis of solenoidactuators and butterfly valve systems,” in Proceedings of the14th International Ship Control Systems Symposium, pp. 1–8,Ottawa, Canada, September 2009.
[16] J. Y. Park and M. K. Chung, “Study on hydrodynamic torqueof a butterfly valve,” Journal of Fluids Engineering, Transactionsof the ASME, vol. 128, no. 1, pp. 190–195, 2006.
[17] J. B. Pomet and L. Praly, “Adaptive nonlinear regulation: esti-mation from the Lyapunov equation,” IEEE Transactions onAutomatic Control, vol. 37, no. 6, pp. 729–740, 1992.
[18] H. K. Khalil, Nonlinear Systems, Prentice Hall, Upper SaddleRiver, NJ, USA, 3rd edition, 2002.
-
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2010
RoboticsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Journal ofEngineeringVolume 2014
Submit your manuscripts athttp://www.hindawi.com
VLSI Design
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
DistributedSensor Networks
International Journal of