Академический Документы
Профессиональный Документы
Культура Документы
I. I NTRODUCTION
With the rapid advancement of science and technology, the
design and development of complex critical systems facilitating and aiding the humankind has gained high momentum.
These complex systems are often subjected to rugged environmental conditions and in course of time may experience
faults and failures during operation which can prove to be
catastrophic. Therefore, ensuring an acceptable performance
with guaranteed stability of the overall system in the event
of such failures by compensating the adverse effects of such
eventualities, is the motivation behind fault tolerant control
(FTC) [1]. In accordance with the past literature, fault tolerant
control design has two major directions, namely, passive and
active approaches. Passive FTC methodologies rely on framing
the FTC design as a robust control problem offline and
ultimately an optimal solution is attained assuring robustness
towards anticipated faults and failures. Typical passive FTC
approaches can be found in [2][5]. On the contrary, active
FTC depends on a fault detection and isolation (FDI) scheme
for the online detection, isolation and subsequent estimation of
the fault. According to the fault estimated by the FDI module,
reconfiguration or adaptation of the controller takes place to
maintain an acceptable post fault performance and a considc
978-1-4799-5364-6/14/$31.00
2014
IEEE
x i
x n
y
x i = xi+1 + fi (x1 , x2 , ..., xi1 )
m
X
kj gj (x1 , ..., xn )uj
x n = fn (x1 , . . . , xn ) +
(1)
j=1
x 1
= x2 + f1 (x1 )
x 2 = x3 + f2 (x1 , x2 )
x 1
x 2
= x1
= x1
j=1,PtotF
(3)
m
X
kj gj (x)uj
j=1,PtotH
1 2
z
2 1
The time derivative of V1 is given by,
V1
V 1
z1
x1 yr
zi
xi i1 yr(i1)
i = 2, ....., n
(4)
(5)
(13)
z1 z1 = z1 (x2 + f1 (x1 ) y r )
z1 (z2 + 1 + y r + f1 (x1 ) y r )
= z1 z2 + z1 (c1 z1 )
= c1 z12 + z1 z2
V 1
= 1 V1 + z1 z2
(15)
vj
wnom
k1
1
(Tsign() + wnom sign())d(6)
gj 0
1 sign(s1 ) | s1 |1 2 sign(s 1 ) | s 1 |2 (7)
k k+1
, 1 , 2 > 0,
(8)
2k+1 k
2, ....., q,
q+1 = 1, and here q = 2.
Z
s2 (t) s2 (0) wnom dt
(9)
Z
= (T + ||)
(10)
(14)
V i
(16)
zi zi = zi (x i i1 yr(i) )
i1
X
i1
zi (zi+1 + i + yr(i) + fi
xk+1 yr(i)
xk
=
=
k=1
i1
X
i1
(k)
k=0
yr
yr(k+1) )
(17)
V i
= zi (zi+1 + i + fi
i1
X
i1
k=1
i1
X
i1
(k)
k=0 yr
xk
xk+1
yr(k+1) )
(18)
m
X
kj
(11)
i Vi + zi zi+1 zi1 zi ,
i = 2ci
j=1,PtotH
s 1
s2
mp
X
s 2
=
=
() + (() 1)w + w
() + w
j=1,PtotF
kj gj (x)
uj
(12)
(19)
(20)
1 2
1 2
+
T
2
2
n1
X
i Vi + V
1 2
T
4
(24)
= ||
4
So a decrease in V eventually drives the closed loop system
||
kk
(25)
4
4
n1
1
1X 2
1 2
z + 2 +
T
2 i=1 i
2
2
(27)
||
(26)
(21)
1
V = ( + w wnom ) + TT
(22)
Vi + V < 0
i=1
i=1
Vk1 =
n1
X
for k = 1, 2, ...., (m p)
(28)
Vk1 (t+
k1 ) implies the boundedness of Vk1 (tk ). Now, from
= x2 + cos(x21 )
x 2
y
(32)
0.6
0.4
0.2
Output y = x1
0
0.2
0.4
Output y = x
0.6
Reference y
10
Fig. 1.
15
20
25
Time (s)
30
35
40
45
50
0.03
0.02
Tracking error ( y yr )
0.8
0
0.01
0
0.01
0.02
0.03
0
10
15
20
25
Time (s)
30
35
40
45
50
4
2
3
5
0
2
4
6
8
10
12
0
Fig. 3.
u1
5
10
15
20
25
Time (s)
30
35
u2
40
u3
45
50
10
wnom
3
4
10
u1
0
10
15
20
25
Time (s)
30
35
u2
40
u3
45
Fig. 4. Control input with Tang et al.s method [10] in the event of actuator
failures
50
Methodologies
Proposed
Tang et al. [10]
0.0015
0.0105
1.62 s
0.23
0.2385
0
0.02336
VI. C ONCLUSION
In this paper, an actuator failure compensation strategy for
nonlinear uncertain systems is proposed and designed using
an adaptive second order sliding mode control developed in a
backstepping framework. The design has resulted in a simpler
control structure in comparison to the control law given by
[10]. The proposed methodology has been utilized in the
design of failure compensators in the event of actuator stuck
failures, which are unknown in time, pattern and magnitude.
The closed loop stability of the proposed fault tolerant control
is proved using Lyapunovs criterion and is guaranteed at
every time instances of the occurrence of actuator failures.
Moreover, the method presented provides an increased robustness towards nonparametric uncertainties and actuator failures.
On the other hand, the design freedom is increased as the
transient performance of the output can be improved due to its
direct dependence on the virtual control gain (set by the user).
Finally, since the controller structure is simplified, there is
significant reduction in the number of computations needed in
the application of the proposed control law with respect to the
procedure in [10]. Simulation results illustrate the efficiency
of the control method in the case of actuator failures.
R EFERENCES
[1] Y. Zhang and J. Jiang, Bibliographical review on reconfigurable faulttolerant control systems, Annual Reviews in Control, vol. 32, no. 2, pp.
229 252, 2008.
[2] M. Benosman and K.-Y. Lum, Application of absolute stability theory
to robust control against loss of actuator effectiveness, IET Control
Theory Applications, vol. 3, no. 6, pp. 772788, 2009.
[3] Q. Hu, B. Xiao, and M. Friswell, Fault tolerant control with H
performance for attitude tracking of flexible spacecraft, IET Control
Theory Applications, vol. 6, no. 10, pp. 13881399, 2012.
[4] , Robust fault tolerant control for spacecraft attitude stabilization
under actuator faults and bounded disturbance, Journal of Dynamic
Systems, Measurement and Control, vol. 133, pp. 051 0061051 006
8, 2011.