Songklanakarin J. Sci. Technol.

36 (3), 345-357, May - Jun. 2014

http://www.sjst.psu.ac.th

Original Article

Modifying production line for productivity improvement:

A Case Study of Vision Lens Factory
Roongrat Pisuchpen1* and Wongsakorn Chansangar2
1
Department of Industrial Engineering,

2
Engineering Program in Engineering Management,
Kasetsart University, Chatuchak, Bangkok, 10900 Thailand.

Received 8 December 2012; Accepted 3 February 2014

Abstract

This study aimed to modify the plastic vision lens production line CR39 in order to improve its productivity. Work
study and line balancing techniques were employed in order to improve the bottleneck point in the manufacturing process.
Standard time was measured. The result shows that productivity increased by 1,257 pieces per day and labor productivity
increased from 82 to 88 pieces per man-hour. However, this improved production capacity was still below the target capacity.
Therefore three additional alternatives were proposed. A simulation model was employed in order to analyze these alterna-
tives. The results found that, by using the lowest unit production cost scenario, additional labor and machines were added
into the bottleneck point of the production line. The productivity was then increased by 7,738 pieces per day or a 44%
improvement, which reached the target productivity, and the average resource utilization was 89%, which satisfied the objec-
tives.

Keywords: productivity, simulation, work study, utilization, line balance, WIP

1. Introduction overtime due to changes in eyesight. Compared to contact

lenses or laser eye surgery, vision lenses are also more eco-
Thailand is now considered one of the largest vision nomical. Nowadays, the demand for glass lenses continuously
lens production bases in the world. Seven main manufactur- falls since customers tend to use plastic lenses more. As a
ers operate their production facility in Thailand and are result, there has been a major increase in demand for plastic
responsible for 22% of global demand. Last year the propor- lenses. For our case study, the company of interest produces
tion of basic product business to value-added product various types of both glass and plastic lens, e.g. single vision
businesses was 65:35. In this period of economic crisis and lens, bifocal lens, and progressive additional lens. Most of
intense price competition, customers tend to use basic the manufacturing capacity aims to serve the export market.
products more than value-added products; however, vision Since plastic lens demand is on the increment, the company
lenses are essential to customers with common eyes has attempted to increase the manufacturing capacity of
problems. Moreover, vision lenses are needed to be replaced plastic lenses by using the concept of waste reduction manu-
facturing and effective allocation of resources. In this study,
the focus is on the CR39 model (Plastic Lens Casting). This
CR39 model is the main product of the plastic lens manu-
* Corresponding author. facturing line and also has the highest sales volume. As a
Email address: [email protected] consequence of government policy, the minimum wage rate
346 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

has increased, so the company is forced to carefully manage as shown in Figure 1.

labor cost and economic investment. Therefore, this study
aims to improve the efficiency of the production line by using 2.3 Just-in-time
work study and also applying the line balancing concept to
the bottleneck of the production line. Three different alterna- Just-in-time (JIT) is a production system that strives
tives for improving the manufacturing process are proposed. to eliminate production waste and non-value added activities
ARENA software is employed to simulate and analyze these in order to improve material and parts management. JIT is
alternatives. Performance indicators are compared to obtain simply the production of materials, parts or finished goods
the optimal result. to meet their demand exactly on time, with the required levels
of quality and quantity. By implementing JIT, the in-process
2. Literature Review inventory and its associated costs can be reduced. One
important element of JIT is to produce in small lot size (or
2.1 Work study ideally one piece per lot). This system also increases pro-
duction flexibility to improve customer satisfaction. Different
Work study can be divided into method study and lot size production directly affects inventory and lead time.
work measurement. It is used to systematically study and Small lot production results in reduced inventory and reduced
improve human working methods by considering all factors lead time.
that affect the working efficiency and conditions. After the Raymond (2010) improved work flow in a semicon-
job of interest has been selected, time study can be examined ductor production line using small lot size. The result shows
by 1) recording all information about the job, 2) breaking the a significant reduction in production lead time. This improve-
job down into elements, 3) examining those elements and ment concept can be applied in other production line with
determining the sample size, 4) recording the time to perform short lead time and high production volume. That is the
each element using a stop-watch, 5) assessing the speed of batch size is adjusted to smaller lot size in order to reduce
working, 6) converting the observed time to basic time, 7) the lead time and increase the throughput. Hadi et al. (2005)
determining the allowances, and 8) determining the standard proposed a U-type plant layout. Many production lines use
time. straight-line production along with a just-in-time system.
Shuang et al. (1997) improved the bottleneck process Duran et al. (2008) studied the effects of the line balancing
in a lamp assembly line using the basic concepts of work study, problem in assembly lines, the production system, and the
ECLS, and line balancing. The result shows the improvement continuous development of man and machines.
for both product and operator. The reduction in production
time resulted in better productivity. Khalid and Saleh (2011) 2.4 Simulation
studied the manufacturing process in the automobile industry
both before and after the improvement had been made, by Savsar and Al-Jawini (1995) studied the just-in-time
using an operation process chart. The bottleneck of the production system using a simulation technique. The simu-
process was selected to study and improve in order to lated model analyzed the system with different levels of
increase manufacturing capacity. ARENA software was
employed to select the suitable option.
Resnick and Zanotti (1997) incorporated ergonomic
concepts in the design of a manufacturing process in order
to improve productivity. Equipment weight, distance of
movement, and working from heights were factors to be
considered. The results show that all three factors had signi-
ficant impact on the standard time. This means that the design
of the manufacturing process does have an impact on
productivity.

2.2 Line balancing

Keytack (1997) proposed the concept of expert line

balancing system (ELBS). The study shows that by combin-
ing two or more jobs and workstations in order to eliminate
unnecessary jobs and reduce time between machines, an
improved productivity of the production line can be
achieved. By using the line balancing concept, decision
makers can determine a suitable system for their workload.
This also results in maximum total efficiency and productivity Figure 1. Flow chart of an expert line balancing system (ELBS).
 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 347

various factors, such as processing time, Kanban size,

customer demand, the length of the production line, and
operation policy. The objective was to study the effects of
these factors on operation performances, such as throughput
rate, WIP, and station utilization. The result showed that
processing time and customer demand significantly affects
throughput rate and station utilization. Moreover, by using
different Kanban policies, the system also yields different
Figure 2. Type of plastic lens CR39.
throughput rates, WIP, and station utilization. The simulated
model was also used to compare a push and pull system
under the same conditions. However, a pull system always
outperforms push a system regarding the WIP level. Andrew
(2008) used simulation to study the bottleneck point in a
production line. ARENA12 software was employed to simu-
late an intermittent production system in order to find the
appropriate WIP level and optimal production design and to
reduce the problem at the bottleneck process.

3. Case Study

Plastic lens CR39 is divided into finished lenses and

semi-finished lenses. Finished lenses are lenses with finished
optical surfaces on the front and back of the lens and are
completed to meet the requirements of a prescription (Figure
2). Semi-finished lenses are lenses with an unfinished back
Figure 3. Plastic lens CR39 operation process chart.
surface and require an RX lab for surfacing or finishing
according to each individual customer’s need. The manu-
facturing process of plastic lens CR39 is illustrated in Figure
3.
Based on the data collection in a plastic lens CR39
production line the average production capacity per day
(overtime is not included) at some stations does not reach
the target of 25,000 pieces per day (1 shift/8 hrs/day) as
shown in Figure 4. It can be seen that the bottleneck
processes are the assembly and filling stations. Both depart-
ments use process layout by placing different machines in
different locations as shown in Figure 5.
Both stations are job shop productions with large
batch sizes. This means that a lot of time is wasted in the
transportation of WIP between five stations, which results in
high WIP as shown in Figure 6.There are five workstations Figure 4. Production capacity of each Station in CR39 production
which are divided into two main processes as follows. line.

Figure 5. Process layout of assembly and filling station of current line.

348 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

Table 1. Resources allocation. where ti is time for task i, j is the number of tasks, Cd is cycle
time, and
Workstation No. of Staff No. of Machines
(4.35+22+10.9+3.13+11.6) 51.98
N= = =2.36  3 workstation .
Mould Inspection 4 - 22 22
Mould Assembly 10 10 In this case, the number of workstation should be
Gasket Assembly 2 - reduced from 5 to 3 is to increase the production line effi-
Spring Assembly 2 - ciency according to the concept of expert line balancing
Filling Monomer 10 10 system (ELBS).
Summary 28 20 In this stage, standard time is calculated using work
study. Line balancing is then conducted by reducing the
number of workstation to three stations by combining the
mould assembly and filling monomer into one station. Gasket
(1) Mould Assembly with Machine: This process assembly, spring assembly, and filling monomer are also
starts at the mould inspection station. Moulds are then sent grouped together as a single station. By doing so, the
to the mould assembly station and the filling station respec- distance of movement is decreased. One piece flow produc-
tively. tion can be implemented and the batch size of mould inspec-
(2) Mould Assembly with Gasket: This process starts tion station is reduced by half as shown in Figure 7.
at the mould inspection station and then the mould is After the workstations are combined, some unneces-
assembled with a gasket. A spring is added to the mould at the sary activities can be removed and, hence, the standard times
spring assembly station. Moulds are then sent to the same of some assembly steps are decreased. The standard time of
filling station as in mould assembly with a machine line. The improved line is presented in Figure 8.
resources allocation is shown in Table 1. From a method and According to the improvement, the workstations of
time study, the standard time can be calculated as shown in mold assembly operation and filling monomer operation
Figure 6. are combined into new workstation. The total number of
machines required by the new workstation is eighteen. Ten
4. Methodology of them are for the mold assembly operation and eight of
them are for the filling monomer operation. Also, the work-
4.1 Work study method stations of gasket assembly operation, spring operation, and
filling monomer operation are combined into the new
The analysis of the production improvement use workstation. Here, the total number of machines required is
several performances measured to compare the results which two. They are for the monomer filling operation. From the
are productivity and efficiency. From the data in Figure 6, the improved standard time, resources allocation can now be
job time for mould assembly is equal to 22 seconds. The modified as shown in Table 2. From the line balancing and
number of workstation can be calculated from equation (1) improved working method, production layout can be
j improved as shown in Figure 9.
 ti The efficiency for the production line calculated from
N i 1 (1) Equation 2:
Cd

Figure 6. Value stream mapping of current line and standard time of workstations of the current line.
R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 349

Figure 7. Value stream mapping of improved line.

Figure 8. Standard time of workstations of improved line.

Figure 9. Process layout of assembly and filling station of improved line.

Table 2. Number of operators and machines of the improved line.

Workstation No. of Staff No. of Machines

Mould Inspection 5 -
Mould Assembly + Filling Monomer 18 18
Gasket Assembly + Spring +Filling Monomer 3 2
Summary 28 20
350 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

Table 3. Results of the improvement from the work study.

[(2)-(1)]
KPI (1) (2) *100 =
(1)
Before Improvement After Improvement % of Difference
No. of Staff 28 28 -
Productivity (pieces/day) 18,344 19,601 7%
Productivity (pieces/hr/man) 82 88 7%
Utilization Average (%) 80% 85% 6%
Total Efficiency (%) 47.25% 78% 55%

j
fore, an additional improvement is needed in order to reach
 ti the target. It is worth noting that the above analysis is
E= i=1 (2)
NCa studied with average values only. The result might be differ-
ent if data with variances are considered here. Thus it is
where ti is time for task i, j is number of tasks, Ca is actual necessary to conduct a computer simulation in order to verify
cycle time, N is the number of workstation, and the correctness of the experiment and to analyze the amount
(4.35  22  10.9  3.13  11.6) of WIP. Moreover, different line balancing scenarios can be
E  0.4725 . simulated in order to obtain the optimal solution for the
(5*22)
production line by considering system throughput, work in
The result shows that the efficiency for the current process and production cost per unit.
production line is 47.25%. Table 3 shows the operation
performance results both before and after the improvement. 4.2 Simulation model by ARENA
From the above results, the production capacity
increased by 1,257 pieces per day. However this capacity is Figure 10 shows the simulation flow diagram of the
still below the target value of 25,000 pieces per day. There- studied production line.

Figure 10. Simulation flow diagram.

 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 351

Table 4. Distribution processing time of each workstation (time in seconds).

Workstation Activities Lot size Processing time P-value

Cleaning Product Arrival 240 Constant (2)
Assembly Pick and Place mould on the tape 180 Constant (61.23)
inspection table
Inspect Mould 1 1.1 + LOGN(2.26, 1.25) >0.15
Place mould in the holding cabinet 180 Constant (66.05)
Assembly staff picks mould from 180 Constant (89.34)
the holding cabinet
Assembly mould Tape 2 16 + GAMM(1.61, 2.34) >0.15
Bring mould Assembly to filling department 100 Constant (108.83)
Pick and Place mould on the gasket 180 Constant (77.40)
inspection table
Inspect and assemble mould gasket A 2 8.04+3.96*BETA(1.18, 1.39) >0.15
Inspect and assemble mould gasket B 2 7 + WEIB(1.46, 1.12) >0.15
Bring mould gasket to spring installation 100 Constant (126.56)
department
Install spring 1 1.59+LOGN(0.756, 0.481) >0.15
Filling Bring mould gasket to waiting area for filling 100 Constant (28.86)
Pick and Place mould assembly 100 Constant ( 73.26 )
on the filling machine
Filling Gasket A 1 8+3.96*BETA(1.09, 1.67) >0.15
Filling Gasket B 1 6+WEIB(1.48, 1.31) >0.15
Filling Tape SF 1 NORM(11.8, 0.892) >0.15
Filling Tape F 1 6+GAMM(0.458, 2.3) >0.15

An input analyzer was used in order to measure the From the Kolmogorov-Smirnov Test, the results in Table 4
statistical dispersion of data. The Kolmogorov-Smirnov Test show that the p-value >0.15. Therefore, we fail to reject the
with a sample size N=30 was conducted under a 95% confi- null hypothesis (H0). The data is the representation of the
dence interval (significant level = 0.05). The null hypotheses real processing time. The simulation model based on the
will be accepted if P-Value >0.05. The input distribution input data can be constructed as illustrated in Figure 11. The
processing time of each workstation is represented in Table 4. simulation model was used to find the productivity, work in

Figure 11. Simulation model of the assembly and filling stations of the current line.
352 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

process, number of operators and machines, and utilization (FGB). In this department, the two of same production type
percentage both before and after the improvement had been will be batched together permanently using batch module
made. before assembling in process module. Batches of 180 pieces
In the simulation model, the raw materials are created will be delivered to the spring department. After that batches
into the system using separate module. The raw materials are of 100 pieces will be transferred to the fill monomer. In fill
created into the system every two minutes 240 pieces with monomer operation, the batches will be separated in pieces
a total of 50,000 pieces per day to produce 25,000 finished and processed with filling machine using process module.
products per day. The finished products consist of two raw The finished products leave the system by the dispose
materials. The assign module is used to distinguish the type module.
of raw materials. The decide module is used to separate 70% Figure 12 represents the model concept of expert line
of raw materials for inspection mould and 30% of raw balancing system (ELBS). The number of workstation is
materials for gasket assembly. The raw materials will be reduced from 5 to 3. The gasket assembly operation, spring
batched into 180 pieces using batch module before sending operation, and filling monomer operation are combined with
them to different departments. At the destination depart- no batch size during workstations with initiating the one-
ments, the batches will be separated using separate module piece flow continuous process as seem by using three
and followed by process module. process modules in the model.
For example, after inspection mould workstation, the The simulation model is run as a terminating system
batch module will be used to batch the products with batch starting from 8:00-17:00 (8 hrs) without a warm up period.
size shown in Figure 9. At the following station, mould An initial run is used to obtain the half-width for the confi-
assembly station, the products will be separated by separate dence interval to consider the number of replication. The
module. The decide module separates lenses into finished- appropriate number of replication can be calculated from
lenses and semi-lenses. The subclasses for finished-lenses Equation 3 when R0 is an initial number of replication, h 0 is
are finished lens assembly TS machine(FTS), finished lens the half-width from the initial run, and h is the required half-
assembly TP machine(FTP), and finished lens assembly TJ width.
machine(FTJ). The subclasses for semi-lenses are semi lens
assembly TS machine(STS) and semi lens assembly TJ
h o2
R  Ro (3)
machine(STJ). The percentages for each type of lenses are h2
shown in Figure 10. The batch module is used to permanently The initial run with ten replications results in an h 0 of 17,365
batch the products before assembling in process module. ±75 pieces/day. The half-width (h) of 50 required R = 23
After mould assembly station, the products are temporary replication from Equation 3. The research sets the number of
batched again for 100 pieces before sending them to filling replication of 30 to reduce the error variance. The validation
monomer. of the simulation model is completed by comparing the
As the job in gasket assembly, decide module will be throughput of the actual system with the simulation system.
used to separate lenses to be finished lens assembly gasket The hypothesis testing for comparing mean is conducted at
type A (FGA) and finished lens assembly gasket type B  = 0.05 using MINITAB14.

Figure 12. Simulation model of assembly and filling station of improved line.
 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 353

Figure 13. Results for the hypothesis testing.

In Figure 13, the p-value of 0.105 implies that we fail scenarios as shown in Table 7.
to reject the null hypothesis at = 0.05. The data from the
current production and from the simulation model are Alternative 3: Resources allocation
normally distributed. The throughput from both systems is Alternative 3 employs the concept of expert line
not significantly different. Therefore, the simulation model balancing system (ELBS) which is the simulation model as
can be used to represent the current production system. shown in Figure 12. The resource allocation of each process
was established so that the production capacity reaches
Production Line Improvement Alternatives 25,000 pieces per day. A process analyzer was also used.
Resources are control variables. System output and utiliza-
Alternative 1: Work study and line balancing tion are responses. Alternative 3-1 is performed what-if
In this alternative, a work study and line balancing analysis for batch size. Alternative 3-2, is performed what-
concept were employed as described above. The simulation if analysis for resource allocation. Alternative 3-3 is the
model is presented in Figure 12. analysis of the combination of both parameters from what-if
analysis.The simulation model of the assembly and filling
Alternative 2: Production batch size stations of the improved line was run and analyzed for 22
In this alternative, the batch sizes between stations scenarios.
are adjusted using current layout. The optimal batch size was
calculated using the process analyzer in ARENA13. Batch 5. Results
size is considered as a control variable. Responses are seven
types of work in process which are finished lens assembly 5.1 Production improvement through line balancing:
gasket type A (FGA), finished lens assembly gasket type B Alternative 1
(FGB), finished lens assembly TJ machine(FTJ), finished lens
assembly TP machine(FTP), finished lens assembly TS The results of improvement alternative 1 are shown
machine(FTS), semi lens assembly TJ machine(STJ), and in Tables 4 and 5. The average resource utilization and pro-
semi lens assembly TS machine(STS) as well as the output ductivity increased by 7% and 9% respectively. Moreover,
system. The simulation model of the assembly and filling WIP also decreased by 12%.
stations of the current line was run and analyzed for five

Table 5. Amount of WIP.

[(2)-(1)]
Entity Name (1) (2) *100 =
(1)
Current Model Alternative 1 % of Difference
Mould 3,405 3,363 1.2%
FGA 84 33 60.7%
FGB 84 20 76.2%
FTJ 53 1 98.1%
FTP 57 1 98.2%
FTS 63 1 98.4%
STJ 84 2 97.6%
STS 61 1 98.4%
354 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

Table 6. Improvement results of alternative 1.

[(2)-(1)]
KPI (1) (2) *100 =
(1)
Current Model Alternative 1 % of Difference
No. of Staff 28 28 -
Productivity (pieces/day) 17,400 19,003 9%
Labor Productivity (pieces/man/hr) 78 85 9%
Utilization Average (%) 72% 78% 8%
WIP (pieces) 3,899 3,422 -12%

Table 7. Optimal batch size according to alternative 2.

Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5

Response Batch 1 = 100 pieces Batch 1 = 50 pieces Batch 1 = 30 pieces Batch 1 = 10 pieces Batch 1 = 10 pieces
Batch 2 = 180 pieces Batch 2 = 90 pieces Batch 2 = 60 pieces Batch 2 = 60 pieces Batch 2 = 30 pieces
Mould 3,406 3,246 3,160 3167 3137
FGA 85 51 38 25 25
FGB 85 51 38 25 25
FTJ 54 30 19 9 9
FTP 58 32 22 11 11
FTS 64 37 25 14 14
STS 62 36 25 14 14
STJ 85 54 42 30 30
Total WIP 3,899 3,537 3,369 3,295 3,265
Output System 17,400 17,679 17,795 17,890 17,894
WIP Decrease(%) - 9% 14% 15% 16%
Output Increase (%) - 1.60% 2.27% 2.82% 2.84%

5.2 Production improvement through optimal batch size: alternative 1. This scheme includes scenario 1, 2, 3, and 4.
Alternative 2 Alternative 3-2: The batch size remains unchanged and
resource allocation is modified from alternative 1. This
The results are shown in Table 7. Batch 1 is the batch scheme includes scenario 5, 6, 7, 8, 9, 10, 11, 12, and 20. Alter-
transshipped from the mould assembly station to the filling native 3-3: Both batch size and resource allocations are
monomer station and from the gasket assembly station to the modified from alternative 1. This scheme includes scenario
filling monomer station. Batch 2 is the batch transshipped 13, 14, 15, 16, 17, 18, 19, 21, and 22. Table 9 summarizes the
from the inspection mould station to the mould assembly additional number of machines and operators required in
station and from the gasket assembly station to the spring each process.
assembly station. The simulation model was then run with overtime
As can be seen from Table 7, the optimal batch size is being considered. The new output was used in order to
in scenario 5 (Batch 1=10, Batch 2=30) which resulted in a calculate the cost per unit. Filling machines cost is 1,000,000
16% reduction in WIP. Moreover, the system output Baht per machine and the cost for new layout design is
increased by 2.84%. 100,000 Baht. Therefore, as in scenario 17, the total set up
cost for four machines is 4,100,000 Baht. In this study, annual
5.3 Production improvement through resource allocation: worth criterion is employed. Operational cost per day (OC) is
Alternative 3 computed from the cost of machine per day plus labor wage
per day (LC) and can be presented as follows.
Alternative 3 can be divided into three schemes as
follows. Alternative 3-1: The batch size is modified and OC = LC+[Machine Cost (A/P, 10%, 5)/
production resource allocation remains unchanged from number of operating days per year] (4)
 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 355

Table 8. Resource Allocation according to Alternative 3

Scenario Resource Batch size Utilization Average Output System

Capacity (%) (pieces/day)
Scenario 1 28 180 78% 18,976
Scenario 2 28 90 78% 19,003
Scenario 3 28 60 78% 19,022
Scenario 4 28 30 78% 19,032
Scenario 5 29 90 81% 20,462
Scenario 6 30 90 83% 21,885
Scenario 7 31 90 85% 23,335
Scenario 8 33 90 85% 24,205
Scenario 9 32 90 85% 24,190
Scenario 10 33 90 83% 25,137
Scenario 11 31 90 89% 25,088
Scenario 12 32 90 86% 25,135
Scenario 13 31 30 86% 24,306
Scenario 14 31 30 86% 24,300
Scenario 15 32 30 83% 24,294
Scenario 16 31 60 89% 25,106
Scenario 17 31 30 89% 25,138
Scenario 18 32 30 86% 25,139
Scenario 19 31 180 88% 25,031
Scenario 20 34 90 81% 25,172
Scenario 21 34 30 81% 25,191
Scenario 22 35 30 79% 25,189

Table 9. Additional resources required in the system.

Resource
Scenario
No. of Staff Tapping TJ Machine Tapping TP Machine Tapping TS Machine Fill Machine
Scenario 1 0 0 0 0 0
Scenario 2 0 0 0 0 0
Scenario 3 0 0 0 0 0
Scenario 4 0 0 0 0 0
Scenario 5 1 0 0 1 0
Scenario 6 2 0 0 2 0
Scenario 7 3 0 0 3 0
Scenario 8 5 0 0 4 0
Scenario 9 4 0 0 4 0
Scenario 10 5 0 0 4 1
Scenario 11 3 0 0 4 0
Scenario 12 4 0 0 4 0
Scenario 13 3 0 0 3 0
Scenario 14 3 0 0 3 0
Scenario 15 4 0 0 3 0
Scenario 16 3 0 0 4 0
Scenario 17 3 0 0 4 0
Scenario 18 4 0 0 4 0
Scenario 19 3 0 0 4 0
Scenario 20 6 0 0 5 0
Scenario 21 6 0 0 5 0
Scenario 22 7 1 0 5 0
356 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014

Table 10. Unit manufacturing cost of each alternative.

Model No. of OTCost NT Cost Labor Cost/ Annual Worth/ Total Cost/ Output system
Staff (Baht) (Baht) Day(Baht) Day(Baht) Unit(Baht) (pieces)
Current 28 - 8,400 8,400 - 0.4828 17,400
Current + OT 28 6,300 8,400 14,700 - 0.6851 21,457
Policy 1 28 - 8,400 8,400 - 0.4420 19,003
Policy 1+OT 28 6,300 8,400 14,700 - 0.6628 22,179
Policy 2 28 - 8,400 8,400 - 0.4694 17,894
Policy 2+OT 28 6,300 8,400 14,700 - 0.6718 21,880
Policy 3-1 28 6,300 8,400 14,700 - 0.6623 22,197
Policy 3-2 31 - 9,300 9,300 2,170 0.4572 25,088
Policy 3-3 31 - 9,300 9,300 2,170 0.4563 25,138

Table 11. Comparison of the operational performances of the improved models.

KPI Current Alternative 1 Alternative 2 Alternative 3-3

No. of Staff 28 28 28 31
Productivity (pieces/day) 17,400 19,003 17,894 25,138
Productivity (pieces/hr/man) 78 85 80 101
Utilization Average (%) 72% 78% 72% 89%
WIP (pieces) 3,899 3,422 3,265 1,036

It can be seen that alternative 3-3, adding workers capacity is still below the target capacity. An additional
into the production line is necessary as in Table 9. The improvement alternative was proposed in order to increase
output reached the goal of 25,000 pieces per day. The best the utilization percentage to 100% or full capacity. In this
scenario is scenario 17 which adds 31 workers and gains 89% third alternative, additional operators and machines were
utilization. This means that the unit manufacturing cost is added to the bottleneck point of the process. The best
lowest and the production capacity reaches the target. scenario 17 with the lowest unit production cost provided a
44% increase of production capacity. This means that the
5.4 Comparison of operational performances of the capacity increased to 25,138 pieces per day or increased by
improvement alternatives (without overtime) 7,738 pieces per day from the current production line, which
reached the target capacity. The average utilization percent-
Table 11 compares the operational performances of age is 89% which satisfied the objectives. The economic
the current model and the improved models. analysis for alternative 3 shows the payback period of 2.33
years. This is an interesting option to apply to the production
6. Conclusions and Recommendations line with high possibility to be implemented and reached
customer needs and company policy to increase the produc-
All three improvement alternatives provided better tivity of the mean production line. The further research can
production performances. Alternative 1 used line balancing consider other types of costs as factors in production line
and work study concepts. The batch size was decreased by and can apply to similar case study.
half. The production capacity increased from 17,400 to
19,003 pieces per day which means the capacity increased by References
1,603 pieces per day. Moreover, labor productivity increased
from 78 to 85 pieces per man-hour. Resources utilization Duran T., Selcuk, K., Ertan, G. and Baykoc, O.F. 2008. Simple
improved by 8% and WIP decreased by 12%. In alternative and U-type assembly line balancing problems with a
2, small lot size production was implemented. As a result in learning effect. Applied Mathematical Modeling. 32,
Scenario 5, WIP decreased by 16% and the production 2954–2961.
capacity increased from 17,400 to 17,894 pieces per day Hadi, G., Kursat,A.,Cevriye, G and Emel, K. 2005. A shortest
which means the capacity increased by 494 pieces per day route formulation of simple U-type assembly. Applied
or 2.84%. The production resources of both alternatives Mathematical Modeling. 29, 373-380.
remained unchanged. However, this improved production
 R. Pisuchpen & W. Chansangar / Songklanakarin J. Sci. Technol. 36 (3), 345-357, 2014 357

Keytack, H.O. 1997. Expert Line Balancing System (ELBS). Resnick, M.L. and Zanotti, A. 1997. Using Ergonomics to
Computers and Industrial Engineering. 33, 303-306. target Productivity Improvement. Computers and
Khalid, S. and Saleh, A. 2011.Productivity improvement of Industrial Engineering. 33, 185-188.
a motor vehicle inspection station using motion and Savsar, M. and Jawini, A.Al. 1995. Simulation analysis of
time study techniques. King Saud University – Engi- just-in-time production. International Journal of
neering Sciences. 23, 33-41. Production Economics. 42(1), 67-78.
Mikell, P.G. 2008. Automation,Production System, and Com- Shuang, L., Wang, X. and Lixin, M. 1997. Optimization of
puter – Integrated Manufacturing. Pearson Education Assembly Line Based on Work Study. Harbin Univer-
Inc., New Jersey, U.S.A. sity of Commerce, Harbin, P.R. China.
Raymond, G. 2010. Leveraging new SEMI standard to reduce
waste and improve flow for Semiconductor manu-
facturing. Simulation Modeling Practice and Theory.
26, 658–664.