Performance analysis of production line with

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 02 Issue: 12 | Dec-2013, Available @ http://www.ijret.org 178
PERFORMANCE ANALYSIS OF PRODUCTION LINE WITH
BERNOULLI’S MACHINES
Soumen Paul1
, Somnath Ganguly2
1
Assistant Professor, Mechanical Engineering, 2
Assistant Professor, Electrical Engineering, Bankura Unnayani Inst. of
Engg, West Bengal, India
sou_indicom@rediffmail.com, somnath_aec82@yahoo.co.in
Abstract
In flexible manufacturing environments, the performance of a production system is often affected by the sequence of operation. While
performance evaluation, improvement and lean design of production system have been studied extensively, the joint effect of
productivity and quality parameters on operation sequencing remains practically unexplored. Indeed, determining the optimal
operation sequence has significant implication from both theoretical and practical perspectives. In this work the frame work of
Bernoulli reliability and quality models, we develop effective indicator that area simple and easy to implement in practice to
determine the optimal operation sequence that maximize the system production rate.
---------------------------------------------------------------------***---------------------------------------------------------------------
1. INTRODUCTION
Performance evaluation, continuous improvement, lean
design, and bottleneck identification of production systems
have been studied extensively during the last 50 years . In this
studies , it is often assumed that all parts are posed according
to predetermine sequence of operations. The effect of the
operation sequence (OS) on the performance of production
systems, however, has not been systematically investigated.
Indeed in flexible manufacturing environments, one type of
final product can be produced by several sequence of
operations. Thus, determining the optimal sequence of
operations to be performed by all parts is of importance for the
entire production processes. Practical example of operation
sequencing can be widely found in product assembly [(3)], as
well as in machining processes [(4)].
The development based on a recently developed improvement
methodology for production system with quality quantity
couple operation where by increasing the probability to
complite a job during a cycle time leads to decreasing job
quality. Production lines with unreliable machines usually
contain finite capacity buffers intended to attenuate mutual
perturbations of the machines due to breakdowns. It is well
known that the capacity of the buffers should be as small as
possible, that is, lean.
In the present work the production machine considered are
unreliable machine and have non perfect quality. The buffers
used between machines have finite capacity. In the several
production lines the machines follow Bernoulli reliability and
quality model.
The production system with unreliable machines, non perfect
quality and finite buffer are considered. The serial production
lines With M machines having Bernoulli reliability and quality
models. According to these models, machine mi, i ϵ {1,…..,M
}, when neither blocked nor starved, produces a part during a
cycle time with probability pi and fails to do so with
probability ( 1 - pi); in addition, for each part produced by this
machine, it is of good quality with probability gi and is
defective with probability (1- gi). Parameters pi and gi are
referred to as the efficiency and quality of machine mi,
respectively. The Bernoulli reliability model is applicable to
manufacturing operations where the unscheduled downtime is
comparable to the cycle time (e.g., assembly and painting
operations, conveyor pallet jams, etc.). The Bernoulli quality
model is applicable when the defects are due to random and
uncorrelated events (e.g., dust and scratches, etc.).
2. MODELING OF PRODUCTION LINE
p1,g1 N p2g2
CR1-2 -----→⃝----→[]----→⃝----→PR1-2
m1 ↓ b m2 ↓
SR1
1-2 SR2
1-2
Fig: 1.1 sequence m1 – m2
→⃝→[]→⃝→[]→…….→⃝→[]→⃝→
m1 b1 m2 b2 mM−1 bM−1 mM
Fig: 1.2 Serial production line.
The following assumptions are consider for a production
system with M machine shown in Fig. (i)The system consists

__________________________________________________________________________________________
of M machines arranged serially, and M−1 buffers separating
each consecutive pair of machines.(ii)The machines have
identical cycle time Tc. The time axis is slotted with the slot
duration Tc. Machines begin operating at the beginning of
each time slot.(iii)Each buffer is characterized by its capacity,
Ni <∞,1 ≤ i ≤M−1.(iv)Machine i is starved during a time slot
if buffer i−1 is empty at the beginning of the time slot.
Machine 1 is never starved for parts.(v) Machine i is blocked
during a time slot if buffer i has Ni parts at the beginning of
the time slot, and machine i+1 fails to take a part during the
time slot. Machine M is never blocked by ready goods
buffer.(vi)Machines obey the Bernoulli reliability model, that
is, machine i, i = 1, . . . ,M, being neither blocked nor
starved during a time slot, produces a part with probability pi
and fails to do so with probability 1− pi. Parameter pi is
referred to as the efficiency of machine i.
3. MATHEMATICAL EXPRESSION FOR
PERFORMANCE MEASURES
Production Rate (PR): The production rate of a flexible
manufacturing system may be defined as the average number
of parts produced by the downstream machine per cycle time.
Consumption Rate (CR): It is defined as the average
number of raw material consumed by the upstream machine
per cycle time.
Scrap Rate (SR): Scrap rate means the average number of
parts scrap by the machines per cycle time.
Work-in-process (WIP): work-in-process defined as the
average number of parts in buffer b. i.e., WIP, is given by
PR 1-2 =p2g2 [ 1 – Q (p1g1 , p2 , N)]
PR1-2 p2 [ 1 – Q (p1g1 , p2 , N)]
CR1-2 = ---------- = --------------------------------
g1g2 g1
P2 ( 1 – g1 )[ 1 – Q(p1g1,p2,N)
SR1
1-2 = CR1-2 ( 1 – g1 ) =-----------------------------------
g1
SR2
1-2 =CR1-2 g1 (1 – g2 )= p2(1 – g2)[1 – Q (p1g1,p2,N)]
SR1-2 = SR1-2 + SR2-1
If p1g1 = p2 = p then the equation of WIP given below
. N (N + 1)
WIP = ---------------------
2 ( N + 1 – p )
If p1g1 ≠ p2 , then the equation of WIP given below:
1 – αN
(p1g1,p2)
p1g1 [--------------------- - NαN
(p1g1,p2)]
1 – α (p1g1,p2)
WIP1-2 = -----------------------------------
P2 – p1g1 αN
(p1g1,p2)
Where i – j ϵ { 1 – 2 , 2 - 1} denotes Operation Sequence mi -
mj and if x ≠ y then
(1 - x)[ 1 – α(x , y)]
Q (x, y, N) = ---------------------------
1 – x/y αN
(x , y)
If x = y then the equation is :
1 - x
Q (x, y, N) = ----------------
N + 1 – x
x ( 1 – y )
α (x , y) = -----------------
y ( 1 – x )
Note that Q (x, y, N) is equal to the probability that the buffer
is empty in the steady state of a two machine with Bernoulli
line with upstream machine efficiency x, downstream machine
efficiency y, buffer capacity N , and no quality issues . For a
Bernoulli line define by (i) to (ix) under Operation Sequence
mi - mj , since pig I is the “effective” probability that the
upstream machine sends a part to the buffer during a time slot,
it follows immediately that Q (pigi,pj,N) is equal to the
probability of empty buffer in the steady state for the system
consider in this paper.
Idea of the aggregation: Consider the M-machine line and
aggregate the last two machines, mM-1, and mM, into a single
Bernoulli machine denoted as mb
M-1, where b stands for
backward aggregation (see Figure 2.1). The Bernoulli
parameter, pb
M-1, of this machine is assigned as the production
rate of the aggregated two-machine line, calculated using the
first expression. Next, aggregate this machine, i.e., mb
M-1, with
mM-2 and obtain another aggregated machine, mb
M-2.The
forward aggregation define that the first machine, m1, with the
aggregated version of the rest of the line, i.e., with mb
2. This
results in the aggregated machine, denoted as mf
2 , where f
stands for forward aggregation (see Figure 2.2). This process
is continue to the last one which will forward aggregrated. We
show below that the steady states of this recursive procedure
lead to relatively accurate estimates of the performance
measures for the M-machine line.
p1 N1 p2 N2 pM-2 NM-2 pM-1 NM-1 pM
→⃝→[]→⃝→[]→……→⃝→[]→⃝→[]→⃝→
m1 b1 m2 b2 mM−2 bM−2 mM−1 bM−1 mM

__________________________________________________________________________________________
p1 N1 p2 N2 pb
M-2
→⃝→[]→⃝→[]→…………[]→⃝→
m1 b1 m2 b2 mb
M−2
pb
1
→⃝→
mb
1
Fig 2.1 Backward aggregation
p1 N1 pb
2
→⃝→[]→⃝→
m1 b2 mb
2
pf
M-1 NM-1 pM pf
M
→⃝→[]→⃝→ →⃝→
mf
M-1 bM-1 mM
Fig 2.2 Forward aggregation
Blockages and starvations: Since these probabilities must
evaluate blockages and starvations of the real, rather than
aggregated, machines, taking into account expressions, the
estimates of these performance measures, BLi and STi, are
introduced as follows:
BLi = piQ(pb
i+1; pf
i ;Ni); i = 1,……..M - 1;
STi = piQ(pf
i-1; pb
i ;Ni-1); i = 2,…….M:
4. RESULT AND ANALYSIS OF PERFORMANCE
MEASUREMENT:
Table 1 Performance measure of model I. (m1-m2)
CASE p1 p2 N PR CR SR WIP
I
0.3 0.5
2
0.275 0.275 0.55 0.71
0.4 0.6 0.372 0.372 0.744 0.8
0.6 0.8 0.576 0.576 1.152 0.91
Table 2 Performance measure of model II (m1-m2)
CAS
E
p1 p2 g1 g2 N PR CR SR WI
P
II
0.
3
0.
3
0.
3
0.
5
4
0.04 0.29 0.25 0.38
0.
4
0.
4
0.
4
0.
6
0.09 0.39 0.29 0.55
0.
6
0.
6
0.
6
0.
8
0.28 0.59 0.30 0.93
Table 3 Performance measure of model II (m2-m1)
CASE p1 p2 g1 g2 N PR CR S
R
WI
P
II 0. 0. 0. 0. 4 0.04 0.2 0. 0.7
3 3 5 3 2
0.
4
0.
4
0.
6
0.
4
0.09 0.3 0.
2
0.9
0.
6
0.
6
0.
8
0.
6
0.2 0.5 0.
3
1.4
Performance Measure for g1 vs PR, WIP, CR, SR
0
0.2
0.4
0.6
0 0.5 1
Y-Values
PR vs g1
0
0.5
1
0 0.5 1
Y-Values
WIP vs g1
0
0.2
0.4
0 0.5 1
Y-Values
CR vs g1
0
0.2
0.4
0 0.5 1
Y-Values
SR vs g1

__________________________________________________________________________________________
CONCLUSIONS AND FUTURE WORK:
When the machine have different efficiency but quality does
not exits the higher efficiency machine should be placed
upstream to achieved higher production rate. In the same sense
the results obtain from previous chapter when the machines
have efficiency and quality parameter both, the lower quality
machine placed upstream to achieved higher production rate.
If the machine efficiency same it is shown that selecting the
optimal operation sequence increased the production rate by
6% with typically reduction of work-in-processes by 15%
The future of the system theoretic properties of production
system includes: (1)Extension of the results obtained to
production system with machines having other reliability and
quality model e.g., exponential, Webull, Gamma, Lognormal
etc.(2)Generalization of the results for production system with
different topologies, e.g., assembly system, closed line, re-
entrant line etc.(3)Investigation of the effect of operation
sequencing on the trade off among deferent performance
measures.(4)Generalization of the results for small volume
job-shop production environment with high product varity.
ACKNOWLEDGMENTS
The author is very grateful for the valuable comments and
suggestions, especially on the proof of Lemma A.1 by the
anonymous reviewers.
REFERENCES
[1]. N. Viswanadham and Y. Narahari, Performance Modeling
of Automated Manufacturing Systems. Englewood Cliff, NJ:
Prentice-Hall, 1992.
[2]. R. G. Askin and C. R. Standridge, Modeling and Analysis
of Manufacturing Systems. New York: Wiley, 1993.
[3]. H. L. Lee and C. S. Tang, “ Variability reduction through
operations reversal,” Manage . Sci., vol. 44, no. 2, pp. 162-
172, 1998
[4]. J. F. Bard and T. A. Feo,” operation sequencing in discrete
part manufacturing,” Manage. Sci., vol. 35, no. 2, pp. 249-
255,1989
[5]. S. Biller, S. Marin, S. M. Meerkov, and L. Zhang,”
Bottleneck in Bernoulli serial lines with rework,” IEEE Trans.
Autom. Sci. Eng., vol.7, no.2 pp 208-217, Apr. 2010.
[6]. A.B. Hu and S.M. Meerkov,” Lean Buffering in serial
production line with Bernoulli machines , April 07.
[7]. J. Arinez, S. Biller, S. Marin, S. M. Meerkov and L.
Zhang, “ Quantity/quality improvement in an Automotive
paint shop: A case study,” IEEE Trans. Autom. Control. Vol7.
No. 4, pp 755-761, Oct. 2010.
[8]. J. G. Shanthikumar, S. Ding and M. T. Zhang ,“ Queueing
theory for semiconductor manufacturing system: A servey and
open problems,” IEEE Trans. Automat. Sci. Eng., vol 4, pp.
513-522, oct. 2007
[9]. J. Li,” Throughput analysis on Automotive pain shop: A
case study,” IEEE Trans. Automat. Sci. Eng., vol. 1 pp. 90-98,
Jul. 2004.
[10]. Chang Wang and Li JIngshang,” Approximate Analysis
of Reentrant Lines With Bernoulli Reliability model.” IEEE
Trans. Automat. Sci. Eng., vol. 7, no. 3,July 2010.
[11]. Liang Zhang and Xiaohang Yue,” Operation sequencing
in Flexible Production Lines with Bernoulli’s Machines,”
IEEE Trans. Automat Sci. Eng., vol. 8, no. 3, Jul. 2011

Performance analysis of production line with

More Related Content

What's hot (6)

Viewers also liked (20)

Similar to Performance analysis of production line with (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Performance analysis of production line with