International Journal of Quality & Reliability Management

Availability and cost analysis of an engineering system involving subsystems in series configuration
Anuj Kumar, Sangeeta Pant, S. B. Singh,
Article information:
To cite this document:
Anuj Kumar, Sangeeta Pant, S. B. Singh, (2017) "Availability and cost analysis of an engineering system involving
subsystems in series configuration", International Journal of Quality & Reliability Management, Vol. 34 Issue: 6, doi: 10.1108/
IJQRM-06-2016-0085
Permanent link to this document:
http://dx.doi.org/10.1108/IJQRM-06-2016-0085
Downloaded on: 25 April 2017, At: 23:19 (PT)
References: this document contains references to 0 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 5 times since 2017*
Access to this document was granted through an Emerald subscription provided by emerald-srm:543096 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

 AVAILABILITY AND COST ANALYSIS OF AN ENGINEERING SYSTEM
INVOLVING SUBSYSTEMS IN SERIES CONFIGURATION

Abstract.

Purpose: The Purpose of this paper is the evaluation of various reliability measures like

availability, reliability, M.T.T.F (mean time to failure) and profit function.

Design/methodology/approach: We present a novel method for availability analysis of an

engineering system incorporating waiting time to repair. The considered system consist two

subsystems, namely, A and B connected in series. The subsystem B has two identical units in
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

standby arrangement. Each unit of the subsystem has two modes i.e. normal efficiency or failed.

The two standby units of the subsystem B are connected by an imperfect switching. The system

is analyzed by supplementary variable technique, Laplace transformation and Gumbel–Hougaard

family of copula.

Findings: Numerical examples with a way to highlight the important results have been appended

at last. Numerical calculation shows that availabilty and reliability of the system is decreasing

with respect to time when failure rates are fixed at different values. Finally, cost analysis of

system reveals that the expected profit decreases with increase in service cost.

Originality/value: This paper presents a mathematical model in which an important aspect of

switching has been taken into conideration, which is consistent with actual failures of switching

by assuming two different types of failure between adjacent transition. It is evaluated with the

help of the Gumbel–Hougaard family of copula.

Key words: Availability, Reliability, Mean Time to failure (M.T.T.F.), profit function.

Paper type Research paper

 1. Introduction

As far as planning, design and operations of any complex system are concerned the reliability

measures are the foremost concern. Reliability measures are very effective and efficient tool for

probabilistic risk assessment in system design, operation, and maintenance. The determination of

the reliability measures of a complex system composed of a number of interconnected

subsystems has received much attention over the years. Gupta and Sharma (1993), Singh and

Goel (1996), Jain et al. (2002), Jain et al. (2004), Rehmert and Nachlas (2009), Ghasemi et al.
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

(2010), Singh et al. (2011) develop different mathematical models and computed various

reliability measures such as availability, mean time to failure, reliability of complex engineering

system with different assumptions. In many cases it is not always possible to have an immediate

repair facility. Gopalan and Naidu (1982) discussed reliability system in which the repair facility

exists at random interval of time as it is not immediately available after detection of failure. Pirie

and Bendell (1984) have extended the study on the model taken by Bendell and they proposed

failures are assumed to be dependent on the actual operating time as well as elapsed calendar

time. The model was applied in the situation where all the system components are regularly

switched between the ‘on’ and ‘off’ position. Pan (1997) discussed two types of imperfect

switching system: (1) Non repairable system with only one standby component, one switch and

one sensor. (2) A non-repairable system with two standby components, one switch and one

sensor. They considered three modes of switch failure: (1) Failure under energized condition. (2)

Failing open. (3) Failing closed.

Initially reliability models were based on the assumption that only one type of repair/failure is

possible between two adjacent states. However, there are many situations where more than one

repair/failure is possible between two adjacent states, and, when this happen, the system is
 studied by copula. The use of copula in reliability analysis has been extensively studied by

Gennheimer (2002), Nelsen (2006), kumar and Singh (2008), Brunel et al. (2010) and Ram and

Singh (2010). Copula provides coupling of different distributions. Kumar and Singh (2008) has

considered a complex engineering system in which two different types of repairs possible

between two adjacent states and he successfully employing the concept of copula for the

determination of various reliability measures. Later Ram and Singh (2008) and Ram et al.

(2013), solved a complex system and evaluated various reliability measures with the help of

copula. Singh and Rawal (2014) have done the cost analysis of a complex system consisting two
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

subsystems, in series configuration and handling by a human operator. Kumar et al. (2015) and

Pant et al.(2015) have done the reliability optimization of some complex systems by using nature

inspired optimization algorithms.

The present contribution is structured as follows. Section 2 gives the description of the

mathematical model. It provides the details of assumptions, describes the formulation and

solution of the mathematical model and asymptotic behavior of the system. Section 3 presents

the particular cases of reliability measures, and Section 4 gives the conclusions of the proposed

analysis. The block diagram and transition state diagram of the system is shown in Figure 1 and

Figure 2 respectively.

INSERT Table 1. Nomenclature

2. Mathematical Model Details

2.1 Model Description

Here we have considered a complex system consisting of two subsystems A and B which

connected in series. The subsystem B consists of two identical units in standby arrangement
 (Figure 1). Each unit of the subsystem has two modes namely normal efficiency or failed. The

whole system can fail, if one of the subsystems A and B fails or due to environmental

failure/critical human error. Dhillon (1989), Dhillon and Liu (2006) define human error as the

failure to perform a specified task that results in disruption of scheduled operations or result in

damaging equipment. Repair is not immediately available after detection of failure and repair

facility exist at random interval of time.

The two standby units of the subsystem B are connected by an imperfect switching. All the

repairs follow general time distributions whereas all the failures follow exponential time
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

distributions.

In contrast to earlier models here authors have taken an important aspect of switching which is

consistent with actual failures of switching by assuming two different types of failure between

adjacent transition states S3 and S4 and it is evaluated with the help of the Gumbel–Hougaard

family of copula. At the same time by employing Supplementary variable technique[Cox (1995)

and Oliveira et al. (2005)] and Laplace transformation technique; various transition state

probabilities, Availability, reliability and mean time to failure (M.T.T.F) of the system have been

evaluated. Some particular cases also have been discussed for different value of failure rates. The

results are demonstrated by graphs and conclusions have been drawn.

The following assumptions are associated with the model:

(1) Initially, the system is in good state of normal efficiency.

(2) Each unit has two modes namely- normal efficiency or failed.

(3) In one step only one change can take place.

(4) The system can fail on the failure of either subsystem, or environmental failure or human

error.
 (5) Switching over device is imperfect.

(6) Joint probability distribution of imperfect switching over device from state S3 to S4 is

obtained by Gumbel-Hougaard family of copula.

(7) All repairs follow general time distribution whereas all failures follow exponential time

distribution.

(8) After repair, the system works like a new one and never damage anything.

(9) The system waits for repair if repair facility is not available.
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

INSERT Figure 1. Block Diagram of the System

INSERT Figure 2. Transition State Diagram

2.2 Formulation of Mathematical Model

if u1 = log b x and u2 = a then Joint probability (state S3 to state S4 ) according to Gumbel-

Hougaard family is given by

[ ]
1
= exp {ln(log b x )} θ
( )
+ ln a θ θ (1)

By probability considerations and continuing arguments, we can obtain the following set of

difference-differential equations governing the present mathematical model:

∂  ∞ ∞
∂t + λ + λ1 + λE1 + λh1 PO,O,S (t) = ∫ PF,O,S (z, t)β (z)dz + ∫ PE (K,t)βE (K)dK
  0 0

∞ ∞
+ ∫ PO, F ,F ( y, t )β1 ( y)dy + ∫ PH ( N , t )β H ( N )dN (2)
0 0

∂

{
 ∂t + a + exp (ln(logb x) ) + (ln a )
θ θ 1/ θ 
}
 PO, F ,S (t ) = λPO,O, S (t )

(3)

∂ ∂ 
 ∂x + ∂t + β ( x) PO, F ( x, t ) = 0 (4)
 
 ∂  ∞
 ∂t + λ ′ + λE2 + λh2  PO, F ,O (t ) = aPO, F , S (t ) + ∫ PO, F ( x, t ) β ( x)dx (5)
  0

∂ ∂ 
 + + β1 ( y ) PO , F , F ( y, t ) = 0 (6)
 ∂y ∂t 

∂ ∂ 
 ∂z + ∂t + β ( z ) PF ,O , S ( z , t ) = 0 (7)
 

 ∂ ∂ 
 ∂K + ∂t + β E ( K ) PE ( K , t ) = 0 (8)
 

 ∂ ∂ 
 ∂N + ∂t + β H ( N ) PH ( N , t ) = 0 (9)
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

 

Boundary Conditions

[
PO , F (0, t ) = exp {ln (log b x )}θ + (ln a )θ ]
1/θ
PO , F , S (t )
(10)

PO , F , F (0, t ) = λ ′PO , F ,O (t ) (11)

PE (0, t ) = λ E1 PO,O, S (t ) + λE 2 PO, F ,O (t ) (12)

PH (0, t ) = λh1 PO,O, S (t ) + λh2 PO, F ,O (t ) (13)

PF ,O , S (0, t ) = λ1PO,O, S (t ) (14)

Initial conditions are

P O ,O , S ( 0 ) = 1 otherwise, zero. (15)

2.3 Solution of the Model

On taking Laplace transforms of equations (2) through (14) and using (15), we obtained

[s + λ + λ1 + λE1 + λh1 ]PO,O,S (s) = 1 + ∞∫ PF ,O,S ( z, s)β ( z)dz

0

∞ ∞
+ ∫ PE (K, s)βE (K)dK + ∫ PO, F, F ( y, s)β1( y)dy
0 0

∞
+ ∫ PH ( N , s ) β H ( N )dN (16)
0
 

{
 s + a + exp (ln(log b x) ) + (ln a )
θ θ 1/ θ 
}
 PO, F , S ( s ) = λPO ,O ,S ( s )

(17)

∂ 
 ∂x + s + β ( x) PO, F ( x, s ) = 0 (18)
 

[s + λ ′ + λE2 + λh2 ]PO, F ,O (s) = aPO, F ,S (s) + ∞∫ PO, F ( x, s)β ( x)dx (19)
0

∂ 
 + s + β1 ( y ) PO , F , F ( y, s ) = 0 (20)
 ∂y 

∂ 
 ∂z + s + β ( z ) PF ,O, S ( z , s ) = 0 (21)
 
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

 ∂ 
 ∂K + s + β E ( K ) PE ( K , s ) = 0 (22)
 

 ∂ 
 ∂N + s + β H ( N ) PH ( N , s ) = 0 (23)
 

[
PO, F (0, s ) = exp {ln (log b x )}θ + (ln a )θ ]
1/θ
PO , F , S ( s ) (24)

PO, F , F (0, s ) = λ ′PO, F ,O ( s ) (25)

PE (0, s ) = λE1 PO ,O , S ( s ) + λE2 PO , F ,O ( s ) (26)

PH (0, s ) = λh1 PO,O, S ( s ) + λh2 PO, F ,O ( s ) (27)

PF ,O, S (0, s ) = λ1PO,O, S ( s ) (28)

Now on integrating equation (16) through (23) and using (24) through (28) one may obtain

1
PO,O, S ( s ) = (29)
γ (s)

−1

PO, F , S ( s ) =
1 
{
⋅ λ  s + a + exp (ln(log b x) )θ + (ln a )θ
γ (s) 
}
1/θ 


(30)
 −1

PO, F ( s ) =
1 
{
⋅ s + a + exp (ln(logb x) )θ + (ln a )θ
γ ( s ) 
}1/θ 
 ×


{
× exp (ln(logb x) )θ + (ln a )θ } θ .D ( s )
1/
x (31)

−1

1 

{
s + a + exp (ln(logb x) )θ + (ln a )θ }
1/θ 
 ⋅λ

PO, F ,O ( s ) = ⋅ ×
γ (s) [
s + λ ′ + λE2 + λh2 ]

{
× a + exp (ln(logb x) )θ + (ln a )θ

} θ ⋅ S (s)
1/
x (32)
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

−1

1

{
λλ ′ ⋅  s + a + exp (ln(logb x) )θ + (ln a )

}
θ 1/θ 


PO, F , F ( s ) = ⋅ ×
γ (s) [s + λ ′ + λE2 + λh2 ]

{
× a + exp (ln(logb x) )θ + (ln a )θ

} θ ⋅ S (s).D (s)
1/
x 1 (33)

λ1
PF ,O, S ( s ) = ⋅ DZ ( s ) (34)
γ (s)

 −1

1 



{
λλ E 2 ⋅  s + a + exp (ln(logb x) )θ + (ln a )θ

1 / θ 

 ×
}
PE ( s) = ⋅ λ E1 +
γ (s) 

[
s + λ ′ + λ E 2 + λ h2 ]




{
× a + exp (ln(logb x) )θ + (ln a )θ

} θ ⋅ S (s) ⋅ D
1/
x E (s) (35)
  −1

1 



{
λλh2 ⋅ s + a + exp (ln(logb x))θ + (ln a )θ

1 / θ 

 ×
}
PH (s) = ⋅ λh1 +
γ ( s) 

[
s + λ ′ + λE2 + λh2 ]




{
× a + exp (ln(logb x) )θ + (ln a )θ

1/θ
}

⋅ S x ( s ) ⋅ DH ( s )

(36)

Where γ ( s) = s + λ + λ1 + λE1 + λh1 − λ1S z ( s)

−1

{
λλ ′ ⋅  s + a + exp (ln(logb x) )θ + (ln a )θ

1/θ 


}
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

− ×
[
s + λ ′ + λ E2 + λh2 ]

{
× a + exp (ln(logb x) )θ + (ln a )θ

1/θ 
}
⋅ S x ( s ).S1 ( s )


 −1



{
λλ E2 ⋅  s + a + exp (ln(logb x) )θ + (ln a )θ

1 / θ 

 ×
}
− λE1 +


[
s + λ ′ + λE2 + λh2 ]




{
× a + exp (ln(logb x) )θ + (ln a )θ

1/θ 
}
⋅ S x ( s )  ⋅ S E ( s )


 −1

 λλh2

{
⋅  s + a + exp (ln(logb x) )θ + (ln a )

}
θ 1/θ 



− λh1 + ×


[
s + λ ′ + λE2 + λh2 ]




{
× a + exp (ln(logb x) )θ + (ln a )θ

1/θ 
}
⋅ S x ( s )  ⋅ S H ( s )


1 − Si ( s )
And Di ( s) = ∀ i. (37)
s

Also it s interesting to note that

 1
PO,O , S ( s ) + PO, F , S ( s ) + PO, F ( s ) + PO, F ,O ( s ) + PO, F , F ( s ) + PF ,O, S ( s ) + PE (s ) + PH (s ) = (38)
s

2.4. Asymptotic Behavior of the System

Using Abel’s lemma viz. lim s →0 sF ( s) = limt →∞ A(t ) = F ( say ) , provided the limit on R.H.S exists, in

equation (29) through (36), the time independent probabilities are obtained as follows:

1
PO,O, S = (39)
γ ′(0)

−1

PO, F , S =
1 
{
⋅ λ a + exp (ln(logb x) )θ + (ln a )θ
γ ′(0) 
1/θ 


} (40)
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

PO, F =
{
exp (ln(logb x) )θ + (ln a )θ } θ ⋅ λ ⋅η
1/

x (41)
γ ′(0)

1 λ
PO, F ,O = ⋅ (42)
γ ′(0) (λ ′ + λ E2 + λh2 )

1 λλ ′
PO , F , F = ⋅ η1 (43)
γ ′(0) (λ ′ + λ E2 + λ h2 )

λ1
PF ,O, S = ⋅ηZ (44)
γ ′(0)

1  λλ E2 
PE = ⋅ λE1 +  ⋅ηE
′
γ (0)  ′ [
s + λ + λE2 + λh2 ]
(45)

1  λλh2 
PH = ⋅ λh1 +  ⋅ηH
γ ′(0)  [
λ ′ + λE2 + λh2  ] (46)

Where γ ′(0) =  
d
γ ( s) (47)
 ds  s =0

And ηi = − Si (0), for all i. (48)

3. Some Particular Cases

3.1. Availability, Reliability and Unreliability

 β β1
When repairs also follow exponential time distributions, Setting S x ( s) = , S1 ( s ) = ,
(s + β ) ( s + β1 )

βE
S E (s) = etc. in equation (29) through (36), we get the following results
(s + β E )

1
PO ,O, S ( s ) = (49)
E (s)

−1

PO, F , S ( s ) =
1 
{
⋅ λ  s + a + exp (ln(logb x) )θ + (ln a )θ
E ( s) 
1/θ 


} (50)
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

PO, F ( s ) =
1
E ( s)
{
⋅ exp (ln(logb x) )θ + (ln a )θ
1/θ
}
×

−1

{
× λ  s + a + exp (ln(logb x) )θ + (ln a )θ

1/θ 


} 1
(s + β )
(51)

−1

{
λ  s + a + exp (ln(logb x) )θ + (ln a )θ
1/θ 
 }
⋅  
1
PO, F ,O ( s ) = ×
E (s) [
s + λ ′ + λE2 + λh2 ]

{
× a + exp (ln(logb x) )θ + (ln a )θ
1/θ
⋅ }β 

(s + β ) 
(52)


−1

1

{
λλ ′ s + a + exp (ln(logb x) )θ + (ln a )θ

1/θ 


}
PO, F , F ( s ) = ⋅ ×
E (s) [
s + λ ′ + λE2 + λh2 ]

{
× a + exp (ln(logb x) )θ + (ln a )θ
1/θ
⋅} β 
.
1
( s + β )  ( s + β1 )
(53)


λ1 1
PF ,O, S ( s ) = ⋅ (54)
E (s) (s + β z )
  −1

1 
 
{
λλ E 2  s + a + exp (ln(log b x) )θ + (ln a )θ

}
1/θ 


PE ( s ) = ⋅ λE + ×
E (s)  1

[
s + λ ′ + λE 2 + λh2 ]




{
× a + exp (ln(logb x) )θ + (ln a )θ
1/θ
⋅ } β 
 ⋅
1
( s + β )  ( s + β E )


(55)

 −1

1 



{
λλh2 s + a + exp (ln(logb x))θ + (ln a )θ

1 / θ 

 ×
}
PH (s) = ⋅ λh1 +
[ ]
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

E (s)  s + λ ′ + λE2 + λh2






{
× a + exp (ln(logb x) )θ + (ln a )θ }
1/θ
.
β  1
(s + β ) s + β
]. (56)


−1

βz

{
λλ ′ s + a + exp (ln(logb x ) )θ + (ln a )θ

}1/θ 


Where E ( s) = s + λ + λ1 + λE1 + λh1 − λ1 − ×
(s + β z ) [
s + λ ′ + λ E 2 + λ h2]

{
× a + exp (ln(logb x) )θ + (ln a )θ }
1/θ
⋅
β  β1
.
( s + β )  ( s + β1 )


 −1



{
λλ E 2  s + a + exp (ln(log b x ) )θ + (ln a )θ

}1/θ 


−  λ E1 + ×
 [
s + λ ′ + λ E 2 + λ h2 ]





{
× a + exp (ln(logb x) )θ + (ln a )θ
1/θ
⋅ } β 
 ⋅
βE
( s + β )  ( s + β E )

  −1



{
λλh2  s + a + exp (ln(logb x) )θ + (ln a )θ

1 / θ 

 ×
}
− λh1 +


[
s + λ ′ + λE2 + λh2 ]




{
× a + exp (ln(logb x ) )θ + (ln a )θ
1/θ
⋅
+
}
β 
β
 ⋅
βH
+ βH )
(57)
 ( s )  ( s

Availability of the system is given by

1  λa 
Pup ( s ) = 1 +  (58)
( s + λ + λ1 + λE1 + λh1 )  ( s + 1)( s + λ ′ + λE2 + λh2 ) 
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Investing equation (58) we obtain

− (λ + λ1 + λ E1 + λh1 )t − (λ + λ E 2 + λh2 )t
Pup (t ) = (1 + B )e + Ae − t + Ce (59)

Where

aλ
A= (60)
(λ + λ1 + λE1 + λh1 − 1)(λ ′ + λE2 + λh2 − 1)

aλ
B= (61)
(1 − λ − λ1 − λ E1 − λh1 )(λ ′ + λE2 + λh2 − λ − λ1 − λE1 − λh1 )

aλ
C= (62)
(1 − λ ′ − λE2 − λh2 )(λ + λ1 + λE1 + λh1 − λ ′ − λE2 − λh2 )

1
Also R (s) = (63)
( s + λ + λ1 + λE1 + λh1 )

The reliability of the system is given by

− (λ + λ1 + λ E1 + λh1 )t
R (t ) = e (64)

The unreliability of the system can be obtain by the expression 1 − Pup (t ) , i.e.,

Unreliability =1 − Pup (t ) (65)

 On Setting the numerical values of different parameters say a = 0.6, λ = 0.01, λ ′ = 0.03,

λ1 = 0.05, λE1 = λE 2 = 0.01 and λh1 = λh2 = 0.04 in equation (59) & (64), one may obtain

Pup (t ) = 0 .7753 e (− 0 .11t ) + 0 .0073 e (− t ) + 0 .2174 e (−0 .06 t ) (66)

R(t ) = e (−0.11 t ) (67)

One may obtain different values of Pup (t ) and R(t ) from equation (66) & (67), for t = 0, 1, 2, 3, 4,

5, 6, 7, 8 and 9 unit of time, as exhibits in Table 2 and shown by Figure 3 Figure 4.

INSERT Table 2. Availability & Reliability of system

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

INSERT Figure 3. Availability vs Time

INSERT Figure 4. Relibility vs Time

3.2. Mean Time to Failure (M.T.T.F.)

Taking all repairs zero and the limit as ‘s’ tending towards zero in equation (58) for the

exponential distribution; one can obtain the M.T.T.F. as

1
M .T .T .F = lim s →0 Pup ( s ) = (68)
(λ + λ1 + λE1 + λh1 )

One may obtain variation of M.T.T.F. with respect to failure rates as shown in Figure 5 just by

setting, λ = 0.01, λ1 = 0.05, λE1 = 0.01 and λh1 = 0.04 , and varying λ , λ1 , λE1 , λh1 one by one

respectively 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09 in equation (68).

INSERT Table 3. Failure Rates Vs MTTF

INSERT Figure 5. Mean Time to Failure with respect to Failure Rates

3.3. Cost Analysis

Suppose that the service facility is always available, then the expected profit during the interval

[0, t[ is
 t
E p (t ) = K1 ∫ Pup (t ) dt − K 2t (69)
0

For the same set of parameter of equation (66) & (67) one can obtain equation (70).

Therefore

[ ]
E p (t ) = K1 − 7.048e (− 0.11t ) − 0.0073e (− t ) − 3.623e (− 0.06 )t + 10.678 − K 2t (70)

Setting revenue cost ( K1 ) at 1 and varies service cost ( K 2 ) as 0.01, 0.05, 0.10, 0.15, 0.20, 0.25

and 0.5 respectively and varying t = 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10 unit of time in equation (70),

one may get expected profit as a function of time as shown in Figure 6. Table 3 that again in this
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

particular case expected profit decreases with increase in service cost

INSERT Table 4. Expected profit

INSERT Figure 6. Expectected profit vs Time

4. Conclusion

Figure 3 & Figure 4 provide information about the changes in availability and reliability of the

complex engineering system with respect to time when failure rates are fixed at different values.

It is clear from the data presented in Table 2 that availability as well as reliability of the system

decreases as ‘t’ increases and it became steady after a sufficient long interval of time so the

future behavior of a complex system is easily predictable for any given set of parametric values.

Table 3 and Figure 5 yields the M.T.T.F. of the system with respect to variation in λ / λ E1 , λ1 and

λh1 respectively, when the other parameters have been taken as constant. From these graphs, one

can conclude that the Mean time to failure of the system decreases smoothly with each failure

rate.

Table 4 and Graphs in Figure 6 demonstrate the relation between time ‘t’ and expected profit

‘EP(t)’ when the revenue cost (K1) is fixed at 1 and service cost (K2) varies as 0.01, 0.05, 0.10,
 0.15, 0.20, 0.25, and 0.5. One can clearly observe that expected profit decreases very far with

respect to time when the service cost is fixed at a very high value 0.5. So one may analyze the

behavior of profit function with respect to time for different set of values of parameters and it

can be concluded that profit decreases as service cost increases.

5. References

Brunel, N. J.-B., Lapuyade-Lahorgue J., and Pieczynski W. (2010), “Modelingand

unsupervised classification of multivariate hidden Markov chains with copulas”, IEEE

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Trans. Autom. Control, Vol. 55, No. 2, pp. 338–349.

Cox, D. R. (1995), “The analysis of non-Markov stochastic processes by the inclusion of

supplementary variables”, Proceedings of the Cambridge Philosophical Society.

Mathematical and Physical Sciences, Vol. 51 , pp. 433–441.

Dhillon, B. S. (1989), “Human errors: A review”, Microelectronic Reliability, Vol. 29, No.

3, pp. 299–304.

Dhillon, B. S. and Liu, Y. (2006), “Human errors in maintenance: A review”, Journal of

Quality in Maintenance Engineering, Vol. 12, No. 1, pp. 21–36.

Gennheimer, H. (2002), “Model Risk in copula based default pricing models, working paper

series”, Swiss Banking Inst., Univ. of Zurich and NCCR, FINRISK, Zurich, Switzlerland,

Working Paper No. 19.

Ghasemi, A., Yacout, S., and Ouali, M. S. (2010), “Evaluating the reliability function and the

mean residual life for equipment with unobservable states”, IEEE Transaction on Reliability,

Vol. 59, No. 1, pp. 45–54.

 Gopalan, M.N., Naidu, R. S. (1982), “Cost—benefit analysis of a one-server two-unit cold

standby system subject to inspection”, Microelectronics Reliability, Vol. 22, No.4 , , pp.

699-705.

Gupta, P.P., Sharma, M.K. (1993), “Reliability and MTTF evaluation of a two duplex-unit

standby system with two types of repair,” Microelectronic Reliability, Vol. 33, No. 3, pp.

291-295.

Jain, M., Kumar, A. and Sharma, G. C. (2002), “Maintenance Cost analysis for Replacement

Model with Perfect/Minimal Repair”, International Journal of Engineering, Vol. 15, No. 2,
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

pp. 161-168.

Jain, M., Maheshwari, R. and Maheshwari S. (2004), “Reliability Analysis of Redundant

Repairable System with Degraded Failure”, International Journal of Engineering, Vol. 17,

No. 2, pp. 171-182.

Kumar, A. and Singh, S.B. (2008), “Reliability analysis of an n-unit parallel standby system

under imperfect switching using copula”, Computer Modeling and New Technologies, Vol.

12, No. 1, pp. 47-55.

Kumar, A., Pant, S., Singh, S.B., (2015), “ Reliability Optimization of Complex System by

Using Cuckoos Search algorithm” accepted for publication in Mathematical Concepts and

Applications in Mechanical Engineering and Mechatronics.

Melchiori, M.R. (2003), “Which Archimedean copula is right one?”, yieldcurve.com e-

journal, third version, pp. 1-18.

Nelsen, R. B. (2006), “An Introduction to Copulas”, 2nd ed. New York: Springer-Verlag,

(2006).
 Oliveira, E. A., Alvim, A. C. M., and Melo, P. F. F. (2005), “Unavailability analysis of

safety systems under aging by supplementary variables with imperfect repair”, Annals of

Nuclear Energy, Vol. 32, pp. 241–252.

Pan, J.N. (1997), “Reliability prediction of imperfect switching systems subjected to multiple

stresses”, Microelectronic Reliability, Vol. 37, No. 3, pp. 439-445.

Pant, S., Kumar, A., Kishor, A., Anand, D., & Singh, S.B. (2015), “Application of a Multi-

Objective Particle Swarm optimization Technique to Solve Reliability Optimization

Problem”, In the proceeding of IEEE Int. Conf. on Next generation Computing Technologies,
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

pp. 1004-1007. DOI: 10.1109/NGCT.2015.7375272.

Pirie, K. and Bendell, A. (1984), “Optimum component switching for systems that operate

and idle”, Microelectronic Reliability, Vol. 24, No. 4, pp. 769-780.

Ram, M. and Singh, S.B., Singh .V.V. (2013), “Stochastic Analysis of a Standby System

with Waiting Repair Strategy”, IEEE transactions on systems, man, and cybernetics, systems,

Vol 43, No. 3, pp. 698-707.

Ram, M. and Singh, S.B. (2010), “Availability, MTTF and cost analysis of complex system

under preemptive-repeat repair discipline using Gumbel-Hougaard family of copula”,

International journal of Quality & Reliability management, Vol. 27, No. 5, pp. 576-595.

Ram, M. and Singh, S.B. (2008), “Availability and Cost Analysis of a parallel redundant

complex system with two types of failure under preemptive-resume repair discipline using

Gumbel-Hougaard family of copula in repair”, International Journal of Reliability, Quality

& Safety Engineering, Vol. 15, No. 4, pp. 341-365.

 Rehmert, I. J., Nachlas, J. A. (2009), “Availability analysis for the quasirenewal process”,

IEEE Transaction on Systems, Man, Cybernetics. A, Syst., Humans, Vol. 39, No.1, pp. 272–

180.

Singh, J. and Goel, P. (1996), “Availability analysis of a standby complex system having

imperfect switching-over device”. Microelectronic Reliability, Vol. 35, No. 2, pp. 285-288.

Singh, S. B., Ram.M. and Chaube, S. (2011) , “Analysis of the reliability of a three

component system with two repairmen”, International Journal of Engineering Transactions

A: Basics, Vol. 24, No. 4, , pp. 395-401.

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Singh, V.V., Rawal, D.K. (2014), “Availability, MTTF and Cost analysis of complex system

under Preemptive resume repair policy using copula distribution”, Pakistan journal of

Statistics and Operation Research, Vol. 10, No. 3, pp. 299-312.

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 1. Block Diagram of the System

 Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 2. Transition State Diagram

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 3. Availability vs Time

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 4. Reliability vs Time

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 5. Mean Time to Failure with respect to Failure Rates

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

Figure 6. Expectected profit vs Time


λE1 / λE2
λ1
PO , O , S (t ) / PO , F , O (t )
Downloaded by Fudan University At 23:19 25 April 2017 (PT)
PH ( N , t )∆ / PE ( K , t )∆
β H (N ) / β E (K )
β ( z )∆ / β1 ( y )∆
a / log b x
Table 1. Nomenclature
Constant
λh1 / λh2 Constant human error
environmental
rates.
failure rates.
Constant failure rate of
Constant failure rate
λ′ / λ main unit/standby unit
of subsystem A.
of subsystem B.
The probability that
at time‘t’, the system The probability that at
is in the “good time‘t’ system is in
state/good state PO , F , S (t ) failed state due to
while main unit of failure of one unit of
subsystem B is subsystem B.
already failed.”
The probability that The probability that at
at time‘t’, the system time‘t’, system is in
is in failed state failed state due to
due to human error/ PO , F ( x, t ) ∆
failure of switch and
environmental cause, elapsed repair time for
elapsed repair time this lie in the interval
lies in the interval (x, x+∆).
(N, N+∆)/ (K, K+∆).
The probability that at
General repair rate time‘t’, the system is
for the state of in the failed state due
human error/ PO , F , F ( y , t ) ∆ to failure of subsystem
environmental B, elapsed repair time
failure. for the same lies in the
interval (y, y+∆).
The first order
probability that the
The probability that at
system will be
time‘t’, the system is
repaired completely
in the failed state
in the time interval
PF ,O , S ( z , t ) ∆ due to failure of
(z, z+∆)/(y,
subsystem A, elapsed
y+∆),conditioned repair time for this lies
that it was not
in the interval (z, z+∆).
repaired completely
up to time z/y.
Constant/
logarithmic
imperfect switching θ Parameter in [0,1].
rates over device
(b<1).
 β (s) General repair rate Laplace transformation
A (s )
for switch. of function A (t).
Constant switching Laplace transform
a s
rate over device. variable.
Downloaded by Fudan University At 23:19 25 April 2017 (PT)
 Table 2. Availability & Reliability of System vs Time

Time Availability Reliability

0 1 1

1 0.901965 0.895834

2 0.815997 0.802519

3 0.739333 0.718924

4 0.670468 0.644036

5 0.608412 0.57695
Downloaded by Fudan University At 23:19 25 April 2017 (PT)

6 0.552408 0.516851

7 0.501823 0.463013

8 0.456107 0.414783

9 0.414774 0.371577

10 0.377387 0.332871
Downloaded by Fudan University At 23:19 25 April 2017 (PT)
 Table 4. Expected profit

Time (t) Expected profit EP(t) (with K1 = 1)

K 2 =0.01 K 2 =0.05 K 2 =0.10 K 2 =0.15 K 2 =0.20 K 2 =0.25 K 2 -0.50

0 0 0 0 0 0 0 0

1 0.939463 0.899463 0.849463 0.799463 0.749463 0.699463 0.449463

2 1.787547 1.707547 1.607547 1.507547 1.407547 1.307547 0.807547

3 2.554478 2.434478 2.284478 2.134478 1.984478 1.834478 1.084478

Downloaded by Fudan University At 23:19 25 April 2017 (PT)

4 3.248745 3.088745 2.888745 2.688745 2.488745 2.288745 1.288745

5 3.877624 3.677624 3.427624 3.177624 2.927624 2.677624 1.427624

6 4.447532 4.207532 3.907532 3.607532 3.307532 3.007532 1.507532

7 4.964197 4.684197 4.334197 3.984197 3.634197 3.284197 1.534197

8 5.432755 5.112755 4.712755 4.312755 3.912755 3.512755 1.512755

9 5.85783 5.49783 5.04783 4.59783 4.14783 3.69783 1.44783

10 6.24358 5.84358 5.34358 4.84358 4.34358 3.84358 1.34358