International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

Estimation of Reliability Parameters of a Complex Repairable System

V.K. Pathak1 , Ashish Namdeo2 & Anil Tiwari3
1.
2.

3.

B.C.S. Govt. P.G. College, Dhamtari , (Chhattisgarh), India

Comp-Tech Degree College, Sorid Nagar, Dhamtari(Chhattisgarh), India
Disha College, Ramnagar Kota, Raipur, (Chhattisgarh), India

Abstract:-In this paper estimation of reliability parameters of a complex repairable system is considered and semi-markov process is used in
analyzing various reliability parameters such as Transition Probabilities, Mean sojourn times, MTSF, Availability and Busy period of repairman
in repairing the failed units. In the past, Arora et-al[2] has done reliability analysis of two unit standby redundant system with constrained repair
time. Gupta et-al [6] has worked on a compound redundant system involving human failure. Rander et-al [2] has evaluated the cost analysis of
two dissimilar cold standby systems with preventive maintenance and replacement of standby units. A pioneer work in this field was done by
Gopalan [1] and Osaki [3] by performing analysis of warm standby system and parallel system with bivariate exponential life respectively.
Earlier, Pathak et al [7&8] studied reliability parameters of a main unit with its supporting units and also compared the results with two different
distributions. In this paper, Chapman-Kolmogorov equations are used to develop recursive relations. Also the involvement of preventive
maintenance in the model increases the reliability of the functioning units. In the end a particular case is also taken for discussion.
Keywords: Regenerative Point, MTSF, Availability, Busy Period Subject Classification: 53A04, 53A05

__________________________________________________*****_________________________________________________
System Description about the model:
The system consists of three units namely one main unit A and two associate units B & C. Here the associate unit B and
C dependents upon main units A and the system is operable when the main unit and at least one associate unit is in operable.
Main unit are employed to rotate B and C. As soon as a job arrives, all the units work with load. It is assumed that only one job is
taken for processing at a time. There is a single repairman who repairs the failed units on first come first served basis. Using
regenerative point technique several system characteristics such as transition probabilities, mean sojourn times, availability and
busy period of the repairman are evaluated. In the end the expected profit is also calculated.
Assumptions used in the model:
a. The system consists of one main unit and two associate units.
b. The associate unit A and B works with the help of main units.
c. There is a single repairman which repairs the failed units on priority basis.
d. After random period of time the whole system goes to preventive maintenance.
e. All units work as new after repair.
f. The failure rates of all the units are taken to be exponential whereas the repair time distributions are arbitrary.
g. Switching devices are perfect and instantaneous.
Symbols and Notations:

pi j Transition probabilities from S i to S j

i Mean sojourn time at time t

E 0 =State of the system at epoch t=0
E=set of regenerative states

S0 S5
of transition time from S i to S j
qi j (t ) Probability density function
Qi j (t ) Cumulative distribution function of transition time from S i to S j

i (t ) Cdf of time to system failure when starting from state E0 S i E

i (t ) Mean Sojourn time in the state E0 S i E
Bi (t ) Repairman is busy in the repair at time t / E0 S i E
r1 / r2 / r3 / r4 =Constant repair rate of Main unit A /Unit B/Unit C/Unit A&B or A&C
/ / =Failure rate of Main unit A /Unit B/Unit C
6493
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

g1 / g 2 / g 3 / g 4 =Probability density function of repair time of Main unit A/Unit B/Unit C/Unit A&B or A&C
G1 / G2 / G3 / G4 =Cumulative distribution function of repair time of Main unit A/Unit B/Unit C/Unit A&B or A&C
a(t) = Probability density function of preventive maintenance .
b(t) = Probability density function of preventive maintenance completion time.

A (t ) = Cumulative distribution functions of preventive maintenance.

B (t ) = Cumulative distribution functions of preventive maintenance completion time.

s = Symbol for Laplace -stieltjes transforms.
= Symbol for Laplace-convolution.

Symbols used for states of the system:

A0 / Ag / Ar -- Main unit A under operation/good and non-operative mode/ repair mode

B0 / Br / Bg -- Associate Unit B under operation/repair/ good and non-operative mode
C0 / Cr / C g -- Associate Unit C under operation/repair/good and non-operative mode
P.M. -- System under preventive maintenance.
Up states: S 0 ( A0 , B0 , C0 ); S 2 ( A0 , Br , C0 ); S 3

( A0 , B0 , Cr )
Down States: S1 ( Ar , Bo , Co ); S4 (S.D.); S5 ( P.M .)
Transition Probabilities:

Simple probabilistic considerations yield the following non-zero transition probabilities:

t

1.

Q01 (t ) e

( ) t

A (t )dt

2.

Q02 (t ) e ( )t A (t )dt
0

0
t

Q03 (t ) e ( )t A(t )dt

3.

Q10 (t ) g1 (t )dt

4.

5.

Q20 (t ) e

( ) t

g 2 (t )dt

Q24 (t ) ( )e ( )t G2 (t )dt

6.

7.

Q30 (t ) e ( )t g3 (t )dt

Q34 (t ) ( )e ( )t G2 (t )dt

8.

0
t

9.

Q40 (t ) g 4 (t )dt

10.

Q50 (t ) b(t )dt

11.

Q05 (t ) a(t )e ( )t dt
0

Where x1 , Now letting t , we get Lim Qij (t ) pij

t

12.

p01 e

( ) t

A (t )dt

14.

p03 e ( )t A (t )dt
0

x1

x1

[1 a ( x1 )] , 13. p02 e ( )t A (t )dt

*

x1

[1 a * ( x1 )]

[1 a * ( x1 )] ,

15.

p10 g1 (t )dt 1
0

p05 a(t ) e ( )t dt a* ( x1 )
0

16.

17.

p20 e ( )t g 2 (t )dt g 2 ( )
*

6494
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

18.

p24 ( )e

( ) t

G2 (t )dt 1 g 2 ( ) 19. p30 (t ) e ( )t g3 (t )dt g3 ( )

*

20.

p34 (t ) ( )e ( )t G3 (t )dt 1 g3 ( )
*

21 .

p40 (t ) g 4 (t )dt 1

22.

p50 (t ) b(t )dt 1

23.

p10 p40 p50 1

[6.1-6.23]

It is easy to see that

p01 p02 p03 p05 1 , p20 p24 1 ,

p30 p34 1

[6.24-6.26]

And mean sojourn time are given by

1
27. 0
[1 a * ( x1 )] ,
x1
29.

28.

1 G1 (t )dt

1
*
[1 g 2 ( )] ,

30.

1
*
[1 g3 ( )]

31.

4 G4 (t )dt
0

32.

5 B(t )dt
0

[6.27-6.32]
We note that the Laplace-stieltjes transform of Qij (t ) is equal to Laplace transform of qij (t ) i.e.

~
*
Qij ( s) e st Qij (t )dt L{Qij (t )} qij ( s)

[6.33]

~
( s x1 ) t
A (t )dt
34. Q01 ( s) e
0

s x1

35.

~
Q02 ( s) e ( s x1 )t A (t )dt
0

s x1

36.

~
Q03 ( s) e ( s x1 )t A (t )dt
0

s x1

[1 a * ( s x1 )]
[1 a * ( s x1 )]
[1 a * ( s x1 )]

37.

~
Q05 ( s) e ( s x1 )t a(t )dt a * ( s x1 )
0

38.

~
*
Q10 ( s) e st g1 (t )dt g1 ( s)
0

39.

~
*
Q20 ( s) e ( s )t g 2 (t )dt g 2 ( s )
0

40.

~
( )
*
Q24 ( s) ( ) e ( s )t G2 (t )dt
[1 g 2 ( s )]
s

~
*
( s ) t
g3 (t )dt g3 ( s )
41. Q30 ( s) e
0

42.

~
( )
*
Q34 ( s) ( ) e ( s )t G3 (t )dt
[1 g 3 ( s )]
s
0
6495

IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

~
*
st
43. Q40 ( s ) e g 4 (t )dt g 4 ( s )
0

44.

~
*
Q50 ( s) e stb(t )dt b ( s)
0

[6.34-6.44]

We define mi j as follows:-

d ~
/
Qi j ( s)]s 0 Qi j (0)
ds
It can to show that m01 m02 m03 m05 0 ; m20 m24 2 ; m30 m34 3
mi j [

Where

[6.45]

x1

[6.466.49]
7. Mean time to System failure-

Time to system failure can be regarded as the first passage time to the failed state. To obtain it we regard the down state
as absorbing. Using the argument as for the regenerative process, we obtain the following recursive relations.

0 (t ) Q01 (t ) Q02 (t ) s 2 (t ) Q03 (t )

2 (t ) Q20 (t )
0 (t ) Q24 (t )
s
3 (t ) Q30 (t ) s 0 (t ) Q34 (t )

3 (t ) Q05 (t )

[7.1-7.3]

Taking Laplace -stieltjes transform of above equations and writing in matrix form.
We get

~
~
~
~
1
Q02 Q03 0 Q01 Q05
~
~

0
Q20 1
2 Q24

~
~

1
Q30 0
3 Q34

~
~
1
Q02 Q03
~
~ ~
~ ~
D1 ( s) Q20 1
0
1 Q20Q02 Q03Q30
~
Q30 0
1
~
~
~
~
Q01 Q05
Q02 Q03
~
~
~
~ ~
~ ~
And N1 ( s ) Q24
1
0
(Q01 Q05 Q02Q24 Q03Q34 )
~
Q34
0
1
Now letting

[7.4]

s 0 we get

[7.5]

D1 (0) 1 p02 p20 p03 p30

The mean time to system failure when the system starts from the state
MTSF= E (T )

[7.6]

S 0 is given by

D / (0) N1/ (0)

d ~
0 ( s)]s 0 1
ds
D1 (0)

To obtain the numerator of the above equation, we collect the coefficients of relevant of mi j in

m02 m03 m05 ) = 1

Coeff. of ( m20 ) = p 02
Coeff. of ( m30 m34 ) = p 03

[7.7]

D1/ (0) N1/ (0) .

Coeff. of ( m01

[7.8-7.10]

From equation [7.7]

6496
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

MTSF= E (T )

D / (0) N1/ (0) 0 2 p02 3 p03

d ~
=
0 ( s)]s 0 1
1 p02 p 20 p03 p30
ds
D1 (0)

[7.11]
8. Availability Analysis:
Let

M i (t )(i 0,1,2) denote the probability that system is initially in regenerative state S i E is up at time t without

passing through any other regenerative state or returning to itself through one or more non regenerative states .i.e. either it
continues to remain in regenerative S i or a non regenerative state including itself . By probabilistic arguments, we have the
following recursive relations

M 0 (t ) e ( ) t A (t ),

M 2 (t ) e ( ) t G2 (t ),

M 3 (t ) e ( ) t G3 (t ),

[8.1-8.2]

Ai (t ) given as follows:

Recursive relations giving point wise availability

A0 (t ) M 0 (t )
A2 (t ) M 2 (t )

A4 (t ) q40 (t )

i 1, 2, 3, 5

i 0, 4

0i

2i

(t )

Ai (t )

(t )

Ai (t )

A0 (t )

A1 (t ) q10 (t )

A3 (t ) M 3 (t )

A5 (t ) q50 (t )

A0 (t ) ;

3i

i 0, 4

(t )

Ai (t ) ;

A0 (t ) ;
[8.3-8.8]

Taking Laplace stieltjes transformation of above equations; and writing in matrix form, we get

q6 X 6 [ A0 , A1 , A2 , A3 , A4 , A5 ]/ [M 0 ,0, M 2 , M 3 ,0,0]/
*

*
q10

q
20
*

q30
q40

*
q50

[8.9]

6 X 6

q01

q02

q03

q05

q24

q34

*
q10

q20
Therefore D2 ( s )
*

q30
q40

*
q50

q01

q02

q03

q05

q24

q34

Where

q6 X 6

6 X 6

1 q01 q10 q02 (q20 q24 q40 ) q03 (q30 q34 q40 ) q05 q50
*

6497
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________
If

s 0 we get D2 (0) 0 which is true

M 0*

0
M *
Now N 2 ( s ) 2
M *
3
0
0

[8.10]

q01

q02

q03

q05

q24

q34

6 X 6

Solving this Determinant, we get

N 2 (s) M 0 M 2 q02 M 3 q03

*

If

[8.11]

s 0 we get

N 2 (0) 0 2 p02 3 p03

[8.12]
/

To find the value of D2 (0) we collect the coefficient mi j in

Coeff. of

(m01 m02 m03 m05 ) 1 L0

Coeff. of

(m10 ) p01 L1

Coeff.of

(m30 m34 ) p03 L3

Coeff. of

m50 p05 L5

Coeff. of

D2 ( s) we get

(m20 m24 ) p02 L2

Coeff.f

m40 p03 p34 p03 p34 L 4

[8.13-8.18]

Thus the solution for the steady-state availability is given by

A0 () Lim A0 (t ) Lim sA0 ( s)

*

S 0

N 2 (0)
/

D2 (0)

0 L0 2 L2 3 L3
i Li
i 0 ,1, 2 , 3, 4 , 5

[8.19]
(a) Let

Wi (t ) (i

9. BUSY PERIOD ANALYSIS:

1,2,3,4,5,6) denote the probability that the repairman is busy initially with repair in regenerative state S i and

remain busy at epoch t without transiting to any other state or returning to itself through one or more regenerative states.
By probabilistic arguments we have

W1 (t ) G1 (t ) , W2 (t ) G2 (t ) , W3 (t ) G3 (t )

[9.1-

9.3]
Developing similar recursive relations as in availability, we have

B0 (t )

i 1, 2, 3, 5

0i

B2 (t ) W2 (t )

B4 (t ) q40 (t )

(t ) c

i 0, 4

2i

Bi (t )
(t ) c Bi (t )

B0 (t )

B1 (t ) W1 (t ) q10 (t )
B3 (t ) W3 (t )

;
;

B5 (t ) q50 (t )

i 0, 4

3i

B0 (t ) ;

(t )
c

Bi (t ) ;

B0 (t ) ;
[9.4-9.9]

Taking Laplace stieltjes transformation of above equations; and writing in matrix form, we get
6498
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

q6 X 6 [ B0 , B1 , B2 , B3 , B4 , B5 ]/ [0,W1 ,W2 ,W3 ,0,0]/

*

[9.10]
Where

q6 x 6 is denoted by [8.9] and therefore D2 / ( s) is obtained as in the expression of availability.

0
*
W1
*
W
Now N 3 ( s ) 2
W *
3
0

q01

q02

q03

6 X 6

q05

q34

*
24

Solving this Determinant, In the long run, we get the value of this determinant after putting

N 3 (0) (1 p01 2 p02 3 p03 )

1 L1 2 L2 3 L3 =

s 0 is

L
i

i 1, 2 , 3

[9.11]

Thus the fraction of time for which the repairman is busy with repair of the failed unit is given by:
1*

1*

1*

B0 () Lim B0 (t ) Lim sB0 ( s)

t

s 0

N 3 (0)
/

L
L

i 1, 2 , 3

D2 (0)

i i
i 0 ,1, 2 , 3, 4 , 5

[9.12]
(b)Busy period of the Repairman in preventive maintenance in time (0, t], By probabilistic arguments we have

W5 (t ) B (t )

[9.13]

Similarly developing similar recursive relations as in 9(a), we have

B0 (t )
B2 (t )

i 1, 2, 3, 5

i 0, 4

0i

2i

(t )

(t )

B4 (t ) q40 (t ) c

Bi (t )

B1 (t ) q10 (t )

c Bi (t )

B0 (t )

B3 (t )

3i

i 0, 4

(t )

B5 (t ) W5 (t ) q50 (t )

B0 (t )

c Bi (t )

Bc 0 (t )

;
;
[9.14-9.19]

Taking Laplace stieltjes transformation of above equations; and writing in matrix form, we get

q6 X 6 [ B0 , B1 , B2 , B3 , B4 , B5 ]/ [0,0,0,0,0,W5 ]/
*

Where

q6 x 6 is denoted by [8.9] and therefore D2 / ( s) is obtained as in the expression of availability.

0
0
Now N 4 ( s )
0*

0
*
W5

q01

q02

q03

q05

q24

q34

6 X 6

Solving this Determinant, In the long run, we get the value of this determinant after putting

N 4 (0) 5 p05 5 L5

s 0 is

[9.20]
Thus the fraction of time for which the system is under preventive maintenance is given by:

6499
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________
2*

2*

2*

B0 () Lim B0 (t ) Lim sB0 ( s)

t

s 0

N 4 (0)
=
/
D2 (0)

5 L5
i Li

[9.21]

i 0 ,1, 2 , 3, 4 , 5

(c)Busy period of the Repairman in Shut Down repair in time (0, t], By probabilistic arguments we have

W4 (t ) G4 (t )

[9.22]

Similarly developing similar recursive relations as in 9(b), we have

B0 (t )
B2 (t )

i 1, 2, 3, 5

i 0, 4

0i

2i

(t )

(t )

Bi (t )

c Bi (t )

B4 (t ) W4 (t ) q40 (t )

B1 (t ) q10 (t )

B3 (t )

3i

B5 (t ) q50 (t )

B0 (t ) ;

i 0, 4

(t )

B0 (t )
c Bi (t )

B0 (t )

;
;
;
[9.23-9.28]

Taking Laplace stieltjes transformation of above equations; and writing in matrix form, we get

q6 X 6 [ B0 , B1 , B2 , B3 , B4 , B5 ]/ [0,0,0,0,W4 ,0]/
*

Where

q6 x 6 is denoted by [8.9] and therefore D2 / ( s) is obtained as in the expression of availability.

0
0
Now N 5 ( s )
0*

W5*

6 X 6
s 0 is

q01

q02

q03

q05

q24

q34

In the long run, we get the value of this determinant after putting

N 5 (0) 4 ( p02 p24 p03 p30 ) 4 L4

[9.29]
Thus the fraction of time for which the system is under shut down is given by:
3*

3*

3*

B0 () Lim B0 (t ) Lim sB0 ( s)

t

s0

N5 (0)
=
/
D2 (0)

4 L4
i Li

[9.30]

i 0 ,1, 2 , 3, 4 , 5

10. Particular cases: When all repair time distributions are n-phase Erlangian distributions i.e.

nri (nri t ) n1 e nrit

Density function g i (t )
n 1!
(nri t ) j e nrit
j!
j 0

n 1

And Survival function G j (t )

[10.1-10.2]

And other distributions are negative exponential

a(t ) e t , b(t ) e t , A (t ) e t , B (t ) e t
r t
r t
For n=1 g i (t ) ri e i , Gi (t ) e i
If i=1, 2, 3, 4

[10.3-10.6]

g1 (t ) r1e r1 t , g 2 (t ) r2 e r2 t , g 3 (t ) r3 e r3 t , g 4 (t ) r4 e r4t
G1 (t ) e r1 t , G2 (t ) e r2 t , G3 (t ) e r3t , G4 (t ) e r4t
Also

p01

x1

, p02

x1

, p03

x1

, p05

[10.7-10.14]

x1

6500
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

,
r2
r2
r3

1
1
p30
, p34
, 0
, 1 ,
r3
r3
x1
r1
r2

p10 1, p20

, p24

, 3

, 4

1
1
, 5
r4

[10.15-10.30]

r2 4
r3
2 p02 3 p03
L 2 L2 3 L3
MTSF= 0
, A0 () 0 0
,
1 p02 p 20 p03 p30
i Li
1*

B0 ()

i 0 ,1, 2 , 3, 4 , 5

L
L

i 1, 2 , 3

, B0

i i
i 0 ,1, 2 , 3, 4 , 5

Where L0

2*

5 L5
3*
, B0 ()
i Li

( )

i 0 ,1, 2 , 3, 4 , 5

4 L4
i Li

[10.31-10.35]

i 0 ,1, 2 , 3, 4 , 5

1 ; L1 p01; L2 p02 ; L3 p03 ; L4 p02 p24 p03 p34 ; L5 p05 ;

[10.36-10.41]

11. Profit Analysis:The profit analysis of the system can be carried out by considering the expected busy period of the repairman in repair of
the unit in (0,t].
Therefore, G(t) = Expected total revenue earned by the system in (0,t] -Expected repair cost of the failed units
-Expected repair cost of the repairman in preventive maintenance -Expected repair cost of the
Repairman in shut down

C1 up (t ) C2 b1 (t ) C3 b 2 (t ) C4 b3 (t )

= C1 A0 C2 B
Where up (t )

C3 B 2 0 C 4 B 3 0
t

[11.1]
t

A0 (t ) dt ; b1 (t ) B01 (t ) dt ; b 2 (t ) B02 (t ) dt ; b3 (t ) B0 (t )dt

3

C1 is the revenue per unit time and C2 , C3 , C4

[11.2-11.5]
are the cost per unit time for which the system is under simple repair,

preventive maintenance and shut down repair respectively.

12. References:[1] Gopalan M.N.[1975]. Probabilistic analysis of a single server n-unit system with (n-1) warm standbys. Operation
Research, Vol. 23, pp 591-595.
[2] Arora J.R. [1976]. Reliability of a two unit standby redundant system with constrained repair time. IEEE Transactions on
Reliability, Vol. 25(3), pp 203-205.
[3] Osaki [1980].A two unit parallel redundant system with bivariate exponential life. Microelectron & Reliability Vol. 20,
pp521-523.
[4] Rander M.C., Kumar Ashok and Suresh K [1994]. Cost analysis of two dissimilar cold standby system with preventive
maintenance and replacement of standby unit. Microelectron & Reliability, Vol. 34, pp 171-174.
[5] Chandrasekar P. and Nataraja R. [1994]. Confidence limits for a steady state availability of a two units standby system.
Microelectron & Reliability, Vol. 34(7), pp 1249-1251.
[6] Gupta R.,Tyagi P.K. and Goel L.R.[1995]. A cold standby system with arrival time of server and correlated failures and
repairs. Microelectron & Reliability, Vol. 35(4), pp 739-742.
[7] Pathak V.K. [2013]. Profit analysis of a system having one main unit and two supporting units. International Journal of
Engineering and Computer Science. Vol. 2, pp 1-13.
[8] Pathak V.K. [2014]. Comparative study of reliability parameter of a system under different types of distribution
functions. African Journal of Mathematics and Computer Science Research, Vol. 7(4) pp 47-54.

6501
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 3 Issue: 12

ISSN: 2321-8169
6493 - 6502

_______________________________________________________________________________________________

Figure: state transition diagram

S2
AO
BO

S5

Cr

G3(t)

SO

S1
Ar

AO

a(t)
PM

BO

CO

b(t)

g1 (t)

g2(t)

Bg

Cg
Bg

g4(t)

g
Ao 2
Br

Co
S.D.

+
S2

up state

S4

Down state

Regenerative State

6502
IJRITCC | December 2015, Available @ http://www.ijritcc.org

_______________________________________________________________________________________________