A New Hybrid Robust Fault Detection of Switching Systems by Combination of Observer and Bond Graph Method

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 8, No. 4, August 2018, pp. 2157 – 2171
ISSN: 2088-8708 2157
Institute of Advanced Engineering and Science
w w w . i a e s j o u r n a l . c o m
A New Hybrid Robust Fault Detection of Switching Systems
by Combination of Observer and Bond Graph Method
Mohammad Ghasem Kazemi and Mohsen Montazeri
Department of Electrical and Computer Engineering Shahid Beheshti University, A.C., Tehran, Iran
Article Info
Article history:
Received August 4, 2017
Revised March 15, 2018
Accepted: April 5, 2018
Keyword:
Switching system
Robust fault detection
Fault sensitivity
Disturbance attenuation
Bond Graph
ABSTRACT
In this paper, the problem of robust Fault Detection (FD) for continuous time switched
system is tackled using a hybrid approach by combination of a switching observer and
Bond Graph (BG) method. The main criteria of an FD system including the fault sen-
sitivity and disturbance attenuation level in the presence of parametric uncertainties are
considered in the proposed FD system. In the first stage, an optimal switching observer
based on state space representation of the BG model is designed in which simultaneous
fault sensitivity and disturbance attenuation level are satisfied using H−/H∞ index. In
the second stage, the Global Analytical Redundancy Relations (GARRs) of the switch-
ing system are derived based on the output estimation error of the observer, which is
called Error-based Global Analytical Redundancy Relations (EGARRs). The paramet-
ric uncertainties are included in the EGARRs, which define the adaptive thresholds on
the residuals. A constant term due to the effect of disturbance is also considered in the
thresholds. In fact, a two-stage FD system is proposed wherein some criteria may be
considered in each stage. The efficiency of the proposed method is shown for a two-tank
system.
Copyright c 2018 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
M. G. Kazemi
Shahid Beheshti University
East Vafadar Blvd., Tehranpars, Tehran, Iran
+989177428946
Email mg kazemi@sbu.ac.ir
1. INTRODUCTION
Fault diagnosis has become an important topic in industrial applications, which further leads to more
attention in research community. Model based fault detection, isolation and identification has received much
attention in the literature [1],[2],[3],[4],[5],[6].
Generally, obtaining a proper model of the system is crucial for fault diagnosis design. Presenting a
proper model by considering multi physics nature and the interaction between continuous and discrete dynamics
of the system is a challenging task in the model based FD system design for hybrid systems. Bond Graph (BG)
method may be considered as a convenient tool for modeling and fault diagnosis of dynamic systems. Different
applications of the BG method can be found in the literature for modeling [7],[8],[9],[10],[11],[12] and fault diag-
nosis [12],[13],[14],[15],[16] applications. Because of its efficiency in modeling and fault diagnosis of continuous
dynamic systems, the BG method is developed to represent switching phenomenon in hybrid dynamic systems,
which leads to numerous applications of the method for modeling and fault diagnosis of hybrid systems in the
literature.
Due to the advancement of computer technology in control systems, switching systems have drawn much
attention in recent years. Switching systems may be assumed as an important class of hybrid systems wherein a
finite number of dynamical subsystems and a logic rule define the overall dynamic of the system. The dynamic of
the subsystems and thus, the overall system may be continuous or discrete, linear or nonlinear and so on. In this
current paper, the linear continuous case of the switching system with average dwell time is considered.
Fault detection of the switching systems draws much attention in recent years. In [17], fault detection
Journal Homepage: http://iaescore.com/journals/index.php/IJECE
Institute of Advanced Engineering and Science
w w w . i a e s j o u r n a l . c o m
, DOI: 10.11591/ijece.v8i4.pp2157-2171

2158 ISSN: 2088-8708
problem for a class of linear switched time varying delay system is formulated as H∞ filtering problem with
ADT in LMI form. A filter based fault detection approach for continuous time switching systems with ADT is
considered in [18]. Simultaneous Fault Detection and Control (SFDC) for continuous and discrete time linear
switched systems with ADT based on dynamic filter is tackled in [19] wherein a fault sensitivity is achieved for a
given disturbance attenuation level based on H−/H∞ criteria.
The robustness of an FD system, which is defined as the considering the effects of uncertainties, distur-
bances and noises in the system, may be regarded as the most important criteria in an FD system. However, it
is impossible to completely removal the effects of uncertainties, disturbances and noises in an FD system. Thus,
many studies are devoted to attenuate the effects of disturbances, uncertainties and noises in an FD system. In other
hand, the attenuation of disturbance effects may lead to reduce the fault sensitivity in an FD system. Therefore, the
fault sensitivity must also be considered in an FD system design.
The effects of uncertainties in the model may lead to missed or false alarms in the FD system. Toward this
end, two approaches are proposed in the literature including active and passive approaches [20]. In active approach,
the robustness of the FD system is defined in the residual generation stage, while the robustness is considered in
the residual evaluation stage in the passive approach [20].
In [21], the robust fault detection filter design of continuous time switched delay systems is considered.
Design of observer-based robust power system stabilizers by considering parametric uncertainties is noticed in
[22]. The parametric uncertainty representation in the BG method in Linear Fractional Transformation (LFT) form
is presented in [23] and noticed in thereafter studies for robust fault diagnosis [24],[25]. Robust fault diagnosis
of energetic system with parameter uncertainties is tackled in [24]. The nonlinear modeling, structural analysis,
residual generation with adaptive thresholds and sensitivity analysis are done using the BG method. The proposed
method is then used for a boiler system. The modeling and fault diagnosis of a DC motor relying on the BG method
is given in [16]. The parameters of the BG method are obtained using the real data of the system. The average
values of the parameters and their standard deviations are assumed as the nominal values and uncertain parts of the
parameters, respectively. In [25], the residuals and thresholds generation in presence of parameter uncertainties
in LFT form is considered and is applied for mechatronic systems. Robust fault diagnosis and prognostics of a
hoisting mechanism based on the BG method considering the LFT form of the parametric uncertainties is addressed
in [26].
Several studies are concentrated on the fault diagnosis of switched system and robust fault diagnosis of
continuous time systems as well, while there are few works on the robust fault diagnosis of switched systems.
Incremental BG approach is given for hybrid systems in [27], which is an extension of the proposed method
in [28] for continuous time dynamic systems. The adaptive thresholds are obtained considering the parametric
uncertainties in the system, which are dependent on the modes of the hybrid system as well. In [29], a robust
fault detection and isolation on the basis of pseudo BG model for hybrid systems is presented. The parametric
uncertainties are given in the LFT form and are assumed 2% of their nominal values at maximum.
Robust FD systems based on the BG method may be considered as the passive approach, since some
thresholds are assumed in the FD system to detect or isolate the occurred faults. In this paper, based on the
BG method, a new active robust FD system is proposed in which the disturbance is attenuated, the effects of
uncertainties are considered and the fault sensitivity is enhanced as well. The proposed method has the benefits of
both the BG and the observer method.
In summary, the main contributions of the paper may be stated as follows:
• A new robust active fault detection based on the combination of the observer and BG method for linear
switched system is presented, which simultaneously attenuates disturbance level and enhances fault sensi-
tivity.
• Error form of the GARRs is used as the residuals which is based on the output estimation error of the
observer.
• The effects of disturbance and parametric uncertainties in the thresholds are considered as a fixed threshold
and an adaptive threshold, respectively.
The remainder of the paper is organized as follows. The problem formulation and some lemmas and re-
marks based on these lemmas are given in section 2. The Research Method of the paper is given in five subsections
of section 3 as disturbance attention level, enhanced fault sensitivity, simultaneous optimal fault sensitivity and
disturbance attenuation level, the GARRs in error form based on the BG method and robustness of the FD system
against parametric uncertainties. The simulation results and discussion for a two-tank system are given in section
4, follows by conclusion in section 5.
IJECE Vol. 8, No. 4, August 2018: 2157 – 2171

IJECE ISSN: 2088-8708 2159
2. PROBLEM FORMULATION AND PRELIMINARIES
Consider the continuous time linear switching system state space representation in the form of
˙x(t) = Aix(t) + Biu(t) + Bdid(t) + Bfif(t) (1)
y(t) = Cix(t) + Ddid(t) + Dfif(t) (2)
which can be given as Eq.(3) - Eq.(4) based on the BG model of the switching system.
˙x(t) = A(a, θ)x(t) + B(a, θ)u(t) + Bd(a)d(t) + Bf (a)f(t) (3)
y(t) = C(a, θ)x(t) + Dd(a)d(t) + Df (a)f(t) (4)
where x ∈ Rn
is the states, y ∈ Rp
is the outputs, u ∈ Rk
is the inputs, d ∈ Rq
is disturbances and f ∈ Rr
is faults
in the switched system. A ,B and C are respectively system, input and output matrices with appropriate dimensions
denote the system, input and output matrices, respectively. . Bd, Bf , Dd and Df are the matrices of disturbance
and fault distribution on the states and on the outputs of the system, respectively. The i, i ∈ l = 1, 2, . . . , N index
is considered to define modes of the switched system. The a = [a1 a2 . . . az] is the state of the controlled junctions
in the BG model, which will define the active mode of the switching system. The number of controlled junctions
in the BG model is assumed as z. The system matrices are dependent on the parameters of the BG model (R, C,
L, ...) which represented as θ in the model.
The switching Luenberger observer is assumed in the form of Eq.(5).
˙ˆx(t) = Ai ˆx(t) + Li(y(t) − ˆy(t)) (5)
ˆy(t) = Ci ˆx(t)
where Li is the observer gain and is dependent on the active mode of the BG model. The observer gain is
designed for each mode in such a way that some criteria are satisfied.
The output estimation error, which is defined as Eq.(6), is used in the new form of the GARRs.
ey(t) = y(t) − ˆy(t) (6)
The asymptotical stability of the observer is investigated on the error dynamic of the observer as Eq.(7).
˙e(t) = (Ai − LiCi)e(t) + (Bfi − LiDfi)f(t) + (Bdi − LiDdi)d(t) =
Aclie(t) + Bclfif(t) + Bcldid(t)
(7)
The output estimation error is achieved as Eq.(8).
ey(t) = Cie(t) + Dfif(t) + Ddid(t) (8)
In this paper, two criteria are considered for the output estimation error of the observer for disturbance
attenuation level and fault sensitivity, which is presented in Eq.(9) and Eq.(10), respectively.
sup
||ey(t)||2
||d(t)||2
< γ, γ > 0 (9)
inf
||ey(t)||2
||f(t)||2
> β, β > 0 (10)
The following definitions and lemmas are used in the current paper.
Definition. Let Nδ(t1, t2) stand for the switching number of δ(t) on the interval [t1, t2) for a given
switching signal δ(t) and any t2 > t1 > t0. If Eq.(11) is satisfied for N0 ≥ 0 and τa > 0, then τa is called the
ADT and N0 is chatter bound [30].
Nδ(t1, t2) ≤ N0 + (t2 − t1)/τa (11)
Lemma 1: For the continuous time switching system Eq.(1) and Eq.(2), let α > 0, µ ≥ 1 and γi >
0, ∀i ∈ l be constant scalars. Suppose there exists positive definite C1
function Vδ(t) : Rn
→ R, δ ∈ l with
Vδ(t0)(x(t0)) = 0 such that [31]:
˙Vi(xt) ≤ −αVi(xt) − yT
t yt + γ2
i uT
t u,∀i ∈ l (12)
A New Hybrid Robust Fault Detection of Switching Systems by ... (Mohammad Ghasem Kazemi)

2160 ISSN: 2088-8708
Vi(xtk) ≤ µVj(xtk), ∀(i, j) ∈ l × l, i = j
then the switching system is Globally Uniformly Asymptotically Stable (GUAS) and will satisfied Eq.(9), which
is called H∞ performance, with index γ no greater than max(γi) for any switching signal with ADT as Eq.(13).
τa > τ∗
a = ln(µ)/α (13)
V is called the Lyapunov function.
Remark 1: In order to use the Lemma 1 in this paper, the u and y are substituted by d(t) and ey(t),
respectively, which indicate the disturbance attenuation level on the output estimation error of the observer.
Lemma2: For the continuous time switching system Eq.(1) and Eq.(2), let α > 0, µ ≥ 1 and βi >
0, ∀i ∈ l be constant scalars. Suppose there exists positive definite C1
function Vδ(t) : Rn
→ R, δ ∈ l with
Vδ(t0)(x(t0)) = 0, which is called Lyapunov function, such that [32]:
˙Vi(xt) ≤ −αVi(xt) + yT
t yt − β2
i uT
t u,∀i ∈ l (14)
Vi(xtk) ≤ µVj(xtk), ∀(i, j) ∈ l × l, i = j
then the switching system is GUAS and will satisfied H− performance as Eq.(10) with index β no smaller than
min(βi) for any switching signal with ADT as Eq.(13).
Remark 2: In order to use the Lemma 2 in this paper, the u and y are substituted by f(t) and ey(t),
respectively, which indicate the fault sensitivity on the output estimation error of the observer.
Lemma 3 [33]: For a given m × m symmetric matrix Z ∈ Sm and two matrices U and V of column
dimension m, there exists matrix X that is unstructured and will satisfy Eq.(15).
UT
XV + V T
XT
U + Z < 0 (15)
if and only if the following inequalities as Eq.(16) and Eq.(17) with respect to X are satisfied.
NT
U ZNU < 0 (16)
NT
V ZNV < 0 (17)
NU and NV are arbitrary matrices that their columns are a basis for null spaces of U and V , respectively.
3. RESEARCH METHOD
In this section, the LMI formulation for fault detection problem of linear continuous time switched system
Eq.(1)-Eq.(2) is given. The overall results are given in five subsections.
3.1. Disturbance attenuation performance in the output estimation error of the switched observer
The switching Lyapunov function for disturbance attenuation criterion of the observer is considered as
Eq.(18), which its derivative must be negative definite as Eq.(19).
V (t) = e(t)T
Pie(t) > 0 (18)
˙V (t) = ˙e(t)T
Pie(t) + e(t)T
Pi ˙e(t) < 0 (19)
The switching Lyapunov function derivative may be obtained as Eq.(20) in fault free case.
˙V (t) = (e(t)T
AT
cli + d(t)T
BT
cldi)Pie(t) + e(t)T
Pi(Aclie(t) + Bcldid(t)) (20)
which can be written as Eq.(21) for simplicity.
˙V (t) = e(t)T
(AT
cliPi + PiAcli)e(t) + 2e(t)T
PiBcldid(t) (21)
According to Remark 1, we need ey(t)T
ey(t) for using lemma 1, which can be given as:
ey(t)T
ey(t) = e(t)T
CT
i Cie(t) + 2e(t)T
CT
i Ddid(t) + d(t)T
DT
diDdid(t) (22)

IJECE ISSN: 2088-8708 2161
Thus, Eq.(12) may be obtained as follows using Lemma 1.
e(t)T
(AT
cliPi + PiAcli)e(t) + 2e(t)T
PiBcldid(t) + αVi(t) + ey(t)T
ey(t) − γ2
d(t)T
d(t) < 0 (23)
Equation (23) can be given as Eq.(24) by regarding Eq.(18) and Eq.(22), which may be represented in matrix
inequality form as Eq.(25).
e(t)T
(AT
cliPi + PiAcli + αPi)e(t) + 2e(t)T
PiBcldid(t) + e(t)T
CT
i Cie(t)+
2e(t)T
CiDdid(t) + d(t)T
DT
diDdid(t) − γ2
d(t)T
d(t) < 0
(24)
AT
cliPi + PiAcli + αPi PiBcldi
∗ 0
+
CT
i Ci CT
i Ddi
∗ DT
diDdi − γ2
I
< 0 (25)
One may obtain Eq.(26) by some simpliﬁcations in Eq.(25).
I AT
cli
0 BT
cldi
αPi Pi
Pi 0
I 0
Acli Bcldi
+
0 CT
i
I DT
di
−γ2
I 0
0 I
0 I
Ci Ddi
< 0 (26)
By assuming Eq.(26) as NT
U ZNU , we have[32]:
I AT
cli 0
0 BT
cldi I


αPi + CT
i Ci Pi CT
i Ddi
∗ 0 0
∗ ∗ −γ2
I + DT
diDdi

 .


I 0
Acli Bcldi
0 I

 < 0 (27)
Thus, it can be concluded that:


αPi + CT
i Ci Pi CT
i Ddi
∗ 0 0
∗ ∗ −γ2
I + DT
diDdi

 < 0 (28)
Therefore, U may be achieved as [Acli − I Bcldi] taking into account that NU columns are basis for the null space
of U. Furthermore, according to [32], NV and V are assumed as:
NV =


I 0
0 0
0 I

 V = [0 I 0]
By using abovementioned results, Eq.(15) in Lemma 3 may be obtained as:


AT
cli
−I
BT
cldi

 Xi[0 λI 0] +


0
λI
0

 Xi[Acli − I Bcldi] +


αPi + CT
i Ci Pi CT
i Ddi
∗ 0 0
∗ ∗ −γ2
I + DT
diDdi

 < 0 (29)
which can be cast into one LMI as Eq.(30).
LMI(1) :


Θ11 Θ12 Θ13
∗ Θ22 Θ23
∗ ∗ Θ33

 < 0 (30)
where
Θ11 = αPi + CT
i Ci + AT
i Xi + XT
i Ai − NiCi − CT
i NT
i
Θ12 = Pi − XT
i + λAT
i Xi − λCT
i NT
i
Θ13 = CT
i Ddi + XT
i Bdi − NiDdi
Θ22 = −λXi − λXT
i
Θ23 = λXT
i Bdi − λNiDdi
Θ33 = DT
diDdi − σI
σ = γ2
, Ni = XT
i Li

2162 ISSN: 2088-8708
3.2. Fault sensitivity performance in the output estimation error of the switched observer
By considering the Lyapunov function as Eq.(31) for faulty case, its derivative may be obtained as Eq.(32).
V (t) = e(t)T
Pfie(t) (31)
˙V (t) = ˙e(t)T
Pfie(t) + e(t)T
Pfi ˙e(t) = e(t)T
(AT
cliPfi + PfiAcli)e(t) + 2e(t)T
PfiBclfif(t) (32)
According to Remark 2, we need ey(t)T
ey(t) for using lemma 1, which can be given as Eq.(33):
ey(t)T
ey(t) = e(t)T
CT
i Cie(t) + 2e(t)T
CT
i Dfif(t) + f(t)T
DT
fiDfif(t) (33)
According to Remark 2 and Lemma2, one may obtain that:
e(t)T
(AT
cliPfi + PfiAcli)e(t) + 2e(t)T
PfiBclfif(t) + αe(t)T
Pfie(t)−
e(t)T
CT
i Cie(t) − 2e(t)T
CT
i Dfif(t) − f(t)T
DT
fiDfif(t) + β2
f(t)T
f(t) < 0
(34)
which can be arranged into Eq.(35).
e(t)T
(AT
cliPfi + PfiAcli + αPfi)e(t) + 2e(t)T
PfiBclfif(t) − e(t)T
CT
i Cie(t)−
2e(t)T
CT
i Dfif(t) + f(t)T
(DT
fiDfi + β2
I)f(t) < 0
(35)
The matrix inequality representation of Eq.(35) is given as Eq.(36).
AT
cliPfi + PfiAcli + αPfi PfiBclfi
∗ 0
+
−CT
i Ci −CT
i Dfi
∗ β2
I − DT
fiDfi
< 0 (36)
Equation Eq.(36) is given as Eq.(37) in form of NT
U ZNU .
I AT
cli 0
0 BT
clfi −I


αPfi − CT
i Ci Pfi −CT
i Dfi
0 0 0
∗ 0 β2
I − DT
fiDfi




I 0
Acli Bclfi
0 −I

 < 0 (37)
Thus, Z matrix may be obtained as follows.
Z =


αPfi − CT
i Ci Pfi −CT
i Dfi
∗ 0 0
∗ ∗ β2
I − DT
fiDfi

 (38)
U = [Acli − I Bclfi], V = [0 λI 0]
Using Eq.(15) in Lemma 3, we have:


AT
cli
−I
BT
clfi

 Xi[0 λI 0] +


0
λI
0

 Xi[Acli − I Bclfi] +


αPfi − CT
i Ci Pfi −CT
i Dfi
∗ 0 0
∗ ∗ β2
I − DT
fiDfi

 < 0 (39)
Altogether, a linear matrix inequality as Eq.(40) is achieved by assuming τ = β2
.
LMI(2) :


Ψ11 Ψ12 Ψ13
∗ Ψ22 Ψ23
∗ ∗ Ψ33

 < 0 (40)
where
Ψ11 = αPfi − CT
i Ci + AT
i Xi + XT
i Ai − NiCi − CT
i NT
i
Ψ12 = Pfi − XT
i + λAT
i Xi − λCT
i NT
i
Ψ13 = −CT
i Dfi + XT
i Bfi − NiDfi
Ψ22 = −λXi − λXT
i
Ψ23 = λXT
i Bfi − λNiDfi
Ψ33 = τI − DT
fiDfi
in which, τ, Ni, Pfi and Xi are unknown variables of the problem.

IJECE ISSN: 2088-8708 2163
3.3. Simultaneous Optimal Fault sensitivity and Disturbance Attenuation Level
According to Eq.(30) and Eq.(40), the disturbance attenuation level and fault sensitivity are defined in two
separate LMI problems, which must be considered simultaneously in the observer design at a given disturbance
attenuation level or fault sensitivity. In this paper, by defining a weighted LMI optimization problem, the optimal
values of the disturbance attenuation and fault sensitivity are obtained. The theorem is given as follows.
Theorem: Consider the continuous time switched system as Eq.(1)-Eq.(2) with ADT along with the
switched observer as Eq.(5). The state estimation error dynamic Eq.(7) is asymptotically stable and satisfies the
performance criteria as Eq.(9) and Eq.(10) for any nonzero d(t) ∈ l2[0, ∞), if there be the scalars values τ > 0,
µ > 0 and the matrices Pi > 0 Qi > 0 Xi Ni such that the subsequent weighted LMI optimization problem has
solution.
min(Pi,Qi,Xi,Ni,µ,τ)
wσ − (1 − w)τ
s.t.LMI(1) < 0, LMI(2) < 0
Pi < µPj, Pfi < µPfj
where 0 ≤ w ≤ 1 is the weighting factor for simultaneous fault sensitivity and disturbance attenuation perfor-
mances and its value is defined by minimization of the following cost function:
min
w
γ
β
Lastly, the gains of the observer for different modes of the switched system, optimal disturbance attenuation level
and optimal fault sensitivity of the output estimation error of the observer are achieved as:
Li = X−T
i Ni, γ =
√
σ, β =
√
τ
3.4. Global Analytical Redundancy Relations in Error Form Using BG method
The generic form of a GARR may be given in Eq.(41) [34].
GARR : ψ(y(n)
, . . . , y, u(m)
, . . . , u, θ, a) (41)
where y, u and θ are denoted for outputs, inputs and parameters of the system, respectively. As in Eq.(3)-Eq.(4),
a = [a1 a2 . . . az] is a binary vector that determine the active mode of the switched system..
The GARRs are differential equations, which their orders are dependent on the sensor locations in the
system. The lower order GARRs are more efficient for monitoring of a system. The generic form of the GARRs
may be presented in matrix form by maximum second order assumption of derivatives.
GARR : ψ(ÿ, ˙y, y, ü, ˙u, u, θ, a) = M1(θ, a)ÿ(t) + M2(θ, a) ˙y(t) + M3(θ, a)y(t)+
Z1(θ, a)ü(t) + Z2(θ, a) ˙u(t) + Z3(θ, a)u(t)
(42)
The effects of disturbances, noises and parametric uncertainties on the GARRs may lead to false or missed
alarms in the FD system. In this paper, a new method by combination of the BG method and observer method is
proposed to overcome these problems in the FD system based on the BG method, which further leads to some
benefits in the FD system.
In normal case i.e. in the case that there are no disturbances, faults, noises and uncertainties in the system,
the estimated states and outputs of the observer converge to the states and outputs of the system. Therefore, the
GARRs by using estimated outputs may be given as:
GARRh : ψ(¨ˆy, ˙ˆy, ˆy, ü, ˙u, u, θ, a) = M1(θ, a)¨ˆy(t) + M2(θ, a) ˙ˆy(t) + M3(θ, a)ˆy(t)+
Z1(θ, a)ü(t) + Z2(θ, a) ˙u(t) + Z3(θ, a)u(t)
(43)
By subtracting Eq.(43) from Eq.(42), the GARRs in the error form, which is called EGARRs, are obtained.
EGARR : M1(θ, a)ëy(t) + M2(θ, a)˙ey(t) + M3(θ, a)ey(t) (44)
The output estimation errors in Eq.(44) come from the observer, which is designed in such a way that be
more sensitive to faults and simultaneously the effect of disturbance is attenuated. Therefore, the effects of the
disturbances are reduced in the GARRs, while the fault sensitivity is enhanced.

2164 ISSN: 2088-8708
Figure 1. The two-tank system
3.5. Robustness of the FD system against parametric uncertainties
The fault detection and isolation are based on the EGARRs, which are dependent on the parameters of
the BG model. In order to obtain a robust FD system against parametric uncertainties, the proposed method in
[23] is considered in this paper. In robust fault detection based on the BG method, a GARR can be decoupled into
nominal (GARRn) and uncertain part (UGARR) as follows.
GARR = GARRn(a, θn, u, y) + UGARR(a, δθ, u, y) (45)
where θn is the nominal part and δθ is relative deviation compared to the nominal value of the parameter θ.
The nominal part is the obtained GARRs, which contain the nominal values of the parameters and made
the residuals. The uncertain part contains the uncertain parameters of the nominal GARRs, which further deter-
mines the upper and lower bounds of the residuals.
An EGARR may be given as Eq.(46) regarding this fact that the EGARRs are not directly dependent on
the inputs and their distribution matrices.
EGARR = EGARRn(a, θn, ey) + UEGARR(a, δθ, ey) (46)
The adaptive thresholds on the residuals are defined as UEGARRs, which are dependent on the parametric
uncertainty (δθ), mode of the system (a) and the output estimation error of the observer (ey).
As mentioned above, the proposed method is not directly dependent on the inputs and distribution matrices
of the inputs (Z1 to Z3). Thus, the robustness of the proposed method may be improved.
Finally, the residual evaluation function and thresholds for fault detection are assumed as Eq.(47) and
Eq.(48), respectively.
JL(t) = ||rk||2 = (
l0+L
l0
rT
k rk)0.5
(47)
Jth = sup
d(t),u(t)∈l2,f(t)=0
JL(t) + || k||2 (48)
in which JL(t) is the residual evaluation function, Jth is the threshold on the residual evaluation function, l0 is the
initial time, L is the window length for residual evaluation function calculations and k is the effect of parametric
uncertainties on the residual, which is obtained from UEGARR. Finally, the faulty or fault free conditions are
declared by comparison between the residual evaluation function and considered threshold.
4. RESULTS AND DISCUSSION
Consider the two-tank system in Fig.1, which involve a flow source, two output valves named as R1 and
R2 and one interconnection valve (R3). The valves could be in open or closed state, which determined the active
dynamic of the switched system. Three configurations as Table I are assumed [35].
The BG models of the understudy two-tank system in integral and derivative causality in 20-sim software
are shown as Fig.2 and Fig.3, respectively.
As mentioned earlier, the valves could be in open or closed state. The coefficients a1, a2 and a3 are used
to show the valve states of R1, R2 and R3, respectively, which are set to ”0” in closed state and ”1” in open state.

IJECE ISSN: 2088-8708 2165
Table 1. Considered modes of the two-tank system
Mode R1 R2 R3 A = [a1 a2 a3]
Mode 1 off on off [0 1 0]
Mode 2 on on off [1 1 0]
Mode 3 off on on [0 1 1]
Figure 2. BG model of two-tank system in integral causality.
In hydraulic domain, the effort and flow variables are regarded as pressure and flow, respectively. The subscript
c is considered to determine the controlled junctions of the model. Three controlled junctions are assumed for
three valves. In BG theory of hybrid systems, the flow variable of the connected bonds to the ith
1-type controlled
junction is multiplied by ai. Two pressure sensors are used to measure pressures of the two tanks which is given
as P1 and P2.
The state space representation of a hybrid system may be obtained in compact form using the BG model.
The states of the system are considered as generalized momentum of inertia elements (pI) and generalized dis-
placement of capacitive elements (qC). Hence, the state vector of the two-tank system is achieved as [q2 q9]T
.
The flow through the valves is assumed laminar. This assumption leads to linear equations of the system
[35]. Using the state space derivation method in [34], the governing equations in compact form are:
˙q2 = (−
a1
R1CT 1
−
a2
R2CT 1
)q2 +
a2
R2CT 2
q9 + qin (49)
˙q9 =
a2
R2CT 1
q2 + (−
a2
R2CT 2
−
a3
R3CT 3
)q9
y1 =
1
CT 1
q2, y2 =
1
CT 2
q9
The parameters definition and their values are given in Table II.
It is also noted that the tank capacity is defined as CT = A/g wherein g = 9.81m/s2
is the acceleration
of gravity.
Thus, the system matrices for three considered modes of the two-tank system are:
A1 =
−0.0007 0.0011
0.0007 −0.0011
, C1 =
0.1962 0
0 0.3270
Figure 3. BG model of two-tank system in derivative causality.

2166 ISSN: 2088-8708
Table 2. Parameters of two-tank system
Variable Deﬁnition Value
A1 Cross sectional area of tank 1 50m2
A2 Cross sectional area of tank 2 30m2
R1 Output pipe of tank 1 resistance 300s/m2
R2 Interconnection pipe resistance 300s/m2
R3 Output pipe of tank 1 resistance 100s/m2
0 0.2 0.4 0.6 0.8 1
0.18
0.2
0.22
0.24
0.26
0.28
0.3
0.32
weight
Disturbanceattenuationlevel
0 0.2 0.4 0.6 0.8 1
0.34
0.36
0.38
0.4
0.42
0.44
weight
Faultsensitivity
Figure 4. Disturbance attenuation level and fault sensitivity versus weighting factor
A2 =
−0.0013 0.0011
0.0007 −0.0011
, C2 =
0.1962 0
0 0.3270
A3 =
−0.0007 0.0011
0.0007 −0.0044
, C3 =
0.1962 0
0 0.3270
B1 = B2 = B3 =
1
0
The other matrices of the system as in Eq.(1) and Eq.(2) are assumed as follows.
Bd1 = Bd2 = Bd3 = B1, Bf1 = Bf2 = Bf3 =
1 0
0 1
Dd1 = Dd2 = Dd3 =
0.1
0.1
, Df1 = Df2 = Df3 =
0.3 0
0.1 −0.5
The GARRs of the system may be obtained using the Causality Inversion Method (CIM) in [34].
GARR1 = CT 1
dP1
dt
+
a2
R2
(P1 − P2) +
a1
R1
P1 − qin = 0 (50)
GARR2 = CT 2
dP2
dt
−
a2
R2
(P1 − P2) +
a3
R3
P2 = 0
The matrices Mi and Zi can be calculated as:
M1 = 0, M2 =
CT 1 0
0 CT 2
, Z1 = Z2 = 0
M3 =
a2
R2
+ a1
R1
− a2
R2
− a2
R2
a2
R2
+ a3
R3
, Z3 =
1
0
The upper bound of parameter uncertainty is assumed as 2% of their nominal values [29].
Solving the weighted LMI optimization problem for α = 0.08, µ = 1.2 and λ = 0.9, the disturbance
attenuation level and fault sensitivity versus weighting factor are given in Fig.4.
By minimization the cost function of the Theorem, we have:
w = 0.8

IJECE ISSN: 2088-8708 2167
0 10 20 30 40 50 60
1
1.5
2
2.5
3
Time(’s’)
SwitchingSignal
Figure 5. Switching signal
The disturbance attenuation level, fault sensitivity and ADT are also given as:
β = 0.4116, γ = 0.2941, τ∗
a = ln(µ)/α = 2.279
One may obtain the following observer gains for three defined modes of the switched system.
L1 =
2.0887 1.3471
−0.0522 0.3873
, L2 =
2.0880 1.3512
−0.521 0.3864
, L3 =
2.0789 1.3987
−0.0453 0.3514
The disturbance signal in the switched system is considered as follows [32].
d(t) = (0.07 + 0.5e−0.04t
cos 0.7πt)u(t) (51)
The switching signal is plotted in Fig.5. The EGARRs may be calculated as:
EGARR1 = CT 1
dey1
dt
+
a2
R2
(ey1 − ey2) +
a1
R1
ey1 = 0 (52)
EGARR2 = CT 2
dey2
dt
−
a2
R2
(ey1 − ey2) +
a3
R3
ey2 = 0
Thus, the uncertain parts of the EGARRs, which are considered in the thresholds on the residuals, are
obtained as:
1 = |δCT1
CT 1
dey1
dt
| + |δ 1
R2
a2
R2
(ey1 − ey2)| + |δ 1
R1
a1
R1
ey1| (53)
1 = |δCT2
CT 2
dey2
dt
| + |δ 1
R2
a2
R2
(ey1 − ey2)| + |δ 1
R3
a3
R3
ey2|
The upper bound of thresholds due to the parametric uncertainties is defined by evaluating || k||2 for each
residual.
The simulation results are obtained for the case that the uncertainties in the parameters (maximum 2 %)
are included in the system, but the observer and EGARRs are based on the nominal values of the parameters. At
first, the first part of the thresholds according to disturbances is obtained in fault free case. The values for the
thresholds are obtained as 0.0434 and 0.0063 for r1 and r2 respectively. Second, the thresholds on the residual
evaluation function are achieved by calculating the parametric uncertainty parts i.e. || k||2 .
The obtained residuals and the residual evaluation function for fault free case in the presence of distur-
bance and parametric uncertainties are shown in Fig.6 and Fig.7, respectively.
As can be seen from the figures, the residual evaluation functions are not exceeded from their two-part
thresholds (fixed and adaptive) and thus, no fault is declared.
A fault signal (f1) is considered as a pulse signal with unit amplitude that occurred in 40 seconds and
remains until 50 seconds for faulty situation. The residuals and residual evaluation functions are plotted in Fig.8
and Fig.9, respectively. It can be observed from the figures that the fault sensitivity is enhanced in an efficient
manner and can be efficiently discriminated from disturbance. The residual evaluation function of the first residual
is exceeded from the defined threshold, which declares a faulty situation in the system. It is also clear that the first
residual is sensitive to f1 fault.

2168 ISSN: 2088-8708
0 10 20 30 40 50 60
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
Time(’s’)
Residualsignalr
1
0 10 20 30 40 50 60
−0.1
−0.05
0
0.05
0.1
0.15
0.2
Time(’s’)
Residualsignalr
2
Figure 6. Residuals in fault free case
0 10 20 30 40 50 60
−0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
Time(’s’)
Residualevaluationfunction
0 10 20 30 40 50 60
0
2
4
6
8
10
12
14
16
18
20
x 10
−3
Time(’s’)
Residual evaluation function of r1
Threshold
Residual evaluation function of r
2
Threshold
Figure 7. Residual evaluation function in fault free case
0 10 20 30 40 50 60
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
Time(’s’)
Residualsignalr
1
0 10 20 30 40 50 60
−0.1
−0.05
0
0.05
0.1
0.15
0.2
Time(’s’)
Residualsignalr
2
Figure 8. Residuals in faulty case
0 10 20 30 40 50 60
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Time(’s’)
0 10 20 30 40 50 60
−0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Time(’s’)
Residual ealuation function of r1
Threshold
Residual ealuation function of r2
Threshold
Figure 9. Residual evaluation function in faulty case

IJECE ISSN: 2088-8708 2169
0 10 20 30 40 50 60
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
Time (’s’)
56 57 58 59 60
0.0434
0.0434
0.0435
Residual evaluation function
Threshold
Figure 10. Residual evaluation function in conventional method for fault free case in the presence of parametric
uncertainties
In comparison to the case that the parametric uncertainties are not considered in the FD system, the
proposed method leads to robustness of the FD system. The method without two-term threshold may lead to
false alarm, which is shown in what follows. The obtained results for the non-faulty case in the presence of
disturbance and parametric uncertainties are shown in Fig.10. The considered thresholds are based on the effects
of disturbances in the system and the adaptive terms due to parametric uncertainties are not included.
As can be seen from the figure, the residual are exceeded from the considered thresholds and occurrence
of a fault is declared, while there is no fault in the system. The residuals for the proposed method may be given in
Fig.9 in which there is no exceeding from the thresholds in this case.
According to the simulation results, the proposed FD system is robust against parametric uncertainties
and simultaneously attenuates disturbance level and enhances fault sensitivity. These criteria are achieved using a
two-stage FD system by combination of a robust switching observer and BG method.
The effects of the parametric uncertainties may lead to false alarm in the fault diagnosis system, as their
effects are not included in the thresholds. The result in this case is shown in Fig.10, which declares a fault
occurrence in comparison to the proposed method as in Fig.9. In fact, the new defined term as the effect of
disturbance on the residuals will improve the robustness aspect of the FD system. The effects of disturbances are
attenuated in the residuals by the switching observer in the first stage, which leads to smaller thresholds in the
second stage. Smaller thresholds will also improve the efficiency of the FD system for fault detection in the case
that the fault has small amplitude and small effect on the residuals consequently.
5. CONCLUSION
In this work, a new robust fault diagnosis system for continuous time switching system with average dwell
time has been proposed based on combination of a switching observer and hybrid bond graph method, in which
different criteria including fault sensitivity, disturbance attenuation and robustness against parametric uncertainties
has been considered. Different criteria have been included in each stage of the two-stage proposed fault diagnosis
system. In the first stage, a switching observer has been proposed that simultaneously enhances fault sensitivity
and attenuates disturbance in the output estimation error of the observer in an optimal manner by defining and
solving a weighted LMI optimization problem. Afterwards, the output estimation error of the observer has been
used in a new form of the GARRs, which called EGARRs and are according to the output estimation error of the
observer. The thresholds on the residuals contain two different parts including a fixed threshold due to the effect
of disturbance and an adaptive threshold considering the parametric uncertainties of the systems in the EGARRs.
A two-tank system has been considered and the simulation results have been shown the efficiency of the proposed
method.
REFERENCES
[1] R. Isermann, “Model-based fault-detection and diagnosis–status and applications,” Annual Reviews in con-
trol, vol. 29, no. 1, pp. 71–85, 2005.
[2] V. Venkatasubramanian, R. Rengaswamy, K. Yin, and S. N. Kavuri, “A review of process fault detection and
diagnosis: Part i: Quantitative model-based methods,” Computers & chemical engineering, vol. 27, no. 3,

2170 ISSN: 2088-8708
pp. 293–311, 2003.
[3] R. Patton and J. Chen, “Observer-based fault detection and isolation: Robustness and applications,” Control
Engineering Practice, vol. 5, no. 5, pp. 671–682, 1997.
[4] A. Chahmi, M. Bendjebbar, B. Raison, and M. Benbouzid, “An extender kalman ﬁlter-based induction ma-
chines faults detection,” International Journal of Electrical and Computer Engineering, vol. 6, no. 2, p. 535,
2016.
[5] S. Asgari, A. Yazdizadeh, and M. Kazemi, “Robust model-based fault detection and isolation for v47/660kw
wind turbine,” AUT Journal of Modeling and Simulation, vol. 45, no. 1, pp. 55–66, 2015.
[6] R. Abbasi, M. Ghazal, and M. Kazemi, “Stator winding fault detection in induction motors using wiener
ﬁlter,” International Review of Electrical Engineering, vol. 7, no. 4, 2012.
[7] D. C. Karnopp, D. L. Margolis, and R. C. Rosenberg, System dynamics: modeling, simulation, and control of
mechatronic systems. John Wiley & Sons, 2012.
[8] R. Sanchez and A. Medina, “Wind turbine model simulation: A bond graph approach,” Simulation Modelling
Practice and Theory, vol. 41, pp. 28–45, 2014.
[9] N. Krikelis and F. Papadakis, “Gas turbine modelling using pseudo-bond graphs,” International journal of
systems science, vol. 19, no. 4, pp. 537–550, 1988.
[10] B. O. Bouamama, K. Medjaher, A. Samantaray, and M. Staroswiecki, “Supervision of an industrial steam
generator. part i: Bond graph modelling,” Control Engineering Practice, vol. 14, no. 1, pp. 71–83, 2006.
[11] J. Ben Salem, M. N. Lakhoua, and L. El Amraoui, “Modeling of dynamical system piloted by discrete
subsystem based on bond graph approach,” International Journal of Electrical and Computer Engineering,
vol. 7, no. 5, 2017.
[12] M. A. Djeziri, B. O. Bouamama, and R. Merzouki, “Modelling and robust fdi of steam generator using
uncertain bond graph model,” Journal of process control, vol. 19, no. 1, pp. 149–162, 2009.
[13] S. Benmoussa, B. O. Bouamama, and R. Merzouki, “Bond graph approach for plant fault detection and
isolation: Application to intelligent autonomous vehicle,” IEEE Transactions on Automation Science and
Engineering, vol. 11, no. 2, pp. 585–593, 2014.
[14] Y. Touati, R. Merzouki, and B. O. Bouamama, “Robust diagnosis to measurement uncertainties using bond
graph approach: Application to intelligent autonomous vehicle,” Mechatronics, vol. 22, no. 8, pp. 1148–1160,
2012.
[15] M. G. Kazemi and M. Montazeri, “A new robust fault diagnosis approach based on bond graph method,”
Journal of the Brazilian Society of Mechanical Sciences and Engineering, vol. 39, no. 11, pp. 4353–4365,
2017.
[16] M. A. Djeziri, R. Merzouki, B. O. Bouamama, and G. Dauphin-Tanguy, “Robust fault diagnosis by using
bond graph approach,” IEEE/ASME Transactions on Mechatronics, vol. 12, no. 6, pp. 599–611, 2007.
[17] D. Zhang, L. Yu, and W.-a. Zhang, “Delay-dependent fault detection for switched linear systems with time-
varying delaysthe average dwell time approach,” Signal processing, vol. 91, no. 4, pp. 832–840, 2011.
[18] D. Wang, P. Shi, and W. Wang, “Robust fault detection for continuous-time switched delay systems: an linear
matrix inequality approach,” IET control theory & applications, vol. 4, no. 1, pp. 100–108, 2010.
[19] M. Davoodi, A. Golabi, H. Talebi, and H. Momeni, “Simultaneous fault detection and control design for
switched linear systems based on dynamic observer,” Optimal Control Applications and Methods, vol. 34,
no. 1, pp. 35–52, 2013.
[20] J. Chen and R. J. Patton, Robust model-based fault diagnosis for dynamic systems, vol. 3. Springer Science
& Business Media, 2012.
[21] D. Wang, W. Wang, and P. Shi, “Robust fault detection for switched linear systems with state delays,” IEEE
Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), vol. 39, no. 3, pp. 800–805, 2009.
[22] H. M. Soliman and M. Soliman, “Design of observer-based robust power system stabilizers,” International
Journal of Electrical and Computer Engineering, vol. 6, no. 5, p. 1956, 2016.
[23] C. S. Kam and G. Dauphin-Tanguy, “Bond graph models of structured parameter uncertainties,” Journal of
the Franklin Institute, vol. 342, no. 4, pp. 379–399, 2005.
[24] B. O. Bouamama, M. Djeziri, and R. Merzouki, “Robust fault diagnosis of energetic system with parameter
uncertainties,” IFAC Proceedings Volumes, vol. 41, no. 2, pp. 6114–6119, 2008.
[25] M. A. Djeziri, R. Merzouki, B. O. Bouamama, and G. Dauphin-Tanguy, “Bond graph model based for robust
fault diagnosis,” in American Control Conference, 2007. ACC’07, pp. 3017–3022, IEEE, 2007.
[26] S. Ghoshal and S. Samanta, “Robust fault diagnosis and prognostics of a hoisting mechanism: a simulation
study,” International Journal of Engineering Science and Technology IJEST, vol. 3, no. 2, pp. 962–980, 2011.
[27] W. Borutzky, “Fault indicators and adaptive thresholds from hybrid system models,” IFAC Proceedings Vol-

IJECE ISSN: 2088-8708 2171
umes, vol. 45, no. 2, pp. 392–397, 2012.
[28] W. Borutzky and G. Dauphin-Tanguy, “Incremental bond graph approach to the derivation of state equations
for robustness study,” Simulation modelling practice and theory, vol. 12, no. 1, pp. 41–60, 2004.
[29] S. K. Ghoshal, S. Samanta, and A. K. Samantaray, “Robust fault detection and isolation of hybrid systems
with uncertain parameters,” Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Sys-
tems and Control Engineering, vol. 226, no. 8, pp. 1013–1028, 2012.
[30] D. Liberzon, Switching in systems and control. Springer Science & Business Media, 2012.
[31] B. Lu, F. Wu, and S. Kim, “Switching lpv control of an f-16 aircraft via controller state reset,” IEEE transac-
tions on control systems technology, vol. 14, no. 2, pp. 267–277, 2006.
[32] M. Davoodi, A. Golabi, H. Talebi, and H. Momeni, “Simultaneous fault detection and control design for
switched linear systems: A linear matrix inequality approach,” Journal of Dynamic Systems, Measurement,
and Control, vol. 134, no. 6, p. 061010, 2012.
[33] P. Gahinet and P. Apkarian, “A linear matrix inequality approach to h control,” International journal of robust
and nonlinear control, vol. 4, no. 4, pp. 421–448, 1994.
[34] A. K. Samantaray and B. O. Bouamama, Model-based process supervision: a bond graph approach. Springer
Science & Business Media, 2008.
[35] A. Mahmoudi, A. Momeni, A. G. Aghdam, and P. Gohari, “Switching between ﬁnite-time observers,” Euro-
pean Journal of Control, vol. 14, no. 4, pp. 297–307, 2008.
BIOGRAPHIES OF AUTHORS
Mohammad Ghasem Kazemi received his B.Sc. degree with ﬁrst class honors
in Electrical Engineering from the University of Semnan, Semnan, Iran in 2005. He also received
the M.Sc. degree in Electrical Engineering from the Tarbiat Modares University (TMU) in 2007.
He is currently pursuing a Ph.D. degree in Automation in Shahid Beheshti University (SBU), AC,
Tehran, Iran. His research activities focus on control systems, fault detection and isolation, hybrid
systems and intelligent systems. ORCID iD: 0000-0002-0236-9974
Mohsen Montazeri received his B.Sc. and M.Sc. degrees in Control Engineering
from the University of Tehran in 1987 and 1989, and the D.E.A and Ph.D. degrees in Automatic
and Signal Processing, both from Orsay University, France, in 1991 and 1994, respectively. From
1994, he has been as an assistant professor in Shahid Beheshti University (SBU), AC, Tehran, Iran.
He has 20 years of experience in power plants.

A New Hybrid Robust Fault Detection of Switching Systems by Combination of Observer and Bond Graph Method

More Related Content

Similar to A New Hybrid Robust Fault Detection of Switching Systems by Combination of Observer and Bond Graph Method (20)

More from IJECEIAES (20)

Recently uploaded (20)

A New Hybrid Robust Fault Detection of Switching Systems by Combination of Observer and Bond Graph Method