Productivity Improvement of A Assembly Line
Productivity Improvement of A Assembly Line
Productivity Improvement of A Assembly Line
LINE
A Thesis
by
PRANAVI YERASI
August 2011
A Thesis
by
PRANAVI YERASI
Approved by:
Chair of Committee,
Committee Members,
Head of Department,
Jorge V. Leon
Guy L. Curry
David E. Claridge
Brett A. Peters
August 2011
iii
ABSTRACT
Productivity Improvement of a Manual Assembly Line.
(August 2011)
Pranavi Yerasi, B.En., Sri Krishnadevaraya University
Chair of Advisory Committee: Dr. V. Jorge Leon
The current project addresses the productivity improvement of a manual
assembly line by making use of operations analysis in the framework of Lean
production. A methodology is proposed that helps to improve the productivity of any
production process. The methodology consists of selecting a product or product family
to be studied followed by current process study. Once the existing process is
documented, all the assembly tasks involved must be timed using time study techniques.
Operations analysis enables the reduction of non-productive tasks and results in a set of
standardized work elements along with the set of standard procedures for performing the
operations.
Assembly line balancing along with the associated operations analysis assists in
constructing or re-configuring an assembly system, which is the key step in improving
the overall performance of an assembly line. Following this approach, two manual
assembly line configurations (single stage parallel line and five-stage serial line) are
constructed for a case study. The results show that by changing over to the single stage
assembly line configuration the operator productivity is doubled when compared to the
existing assembly method.
iv
DEDICATION
To My Family
ACKNOWLEDGEMENTS
I would like to thank Dr. Jorge Leon for his time and advice during the course of
this project. I would also like to thank Dr. Guy Curry and Dr. David Claridge for
agreeing to be on the thesis committee and providing their input. I would also like to
thank Judy Meeks and Laura Reinisch for their patience and support throughout the
process. I would like to thank my parents and family for providing me with the best of
opportunities and having faith in my abilities. Finally, I wish to thank all my friends who
have made my stay away from home a joyous experience.
vi
TABLE OF CONTENTS
Page
ABSTRACT ..................................................................................................................... iii
DEDICATION .................................................................................................................. iv
ACKNOWLEDGEMENTS ............................................................................................... v
TABLE OF CONTENTS .................................................................................................. vi
LIST OF TABLES ......................................................................................................... viii
LIST OF FIGURES ........................................................................................................... ix
1 INTRODUCTION ........................................................................................................... 1
1.1 Objectives ......................................................................................................... 2
2 LITERATURE REVIEW ................................................................................................ 3
2.1 Lean Manufacturing ......................................................................................... 3
2.2 Assembly Line Balancing ................................................................................ 5
3 ASSEMBLY LINE PRODUCTIVITY IMPROVEMENT METHODOLOGY .......... 12
3.1 Product Selection for Study............................................................................ 12
3.1.1 A-B-C Classification Method.......................................................... 13
3.1.2 Part-Process Matrix Method............................................................ 15
3.2 Time Study ..................................................................................................... 17
3.2.1 What and How to Time? ................................................................. 18
3.2.2 Time the Work Elements................................................................. 20
3.3 Operations Analysis ....................................................................................... 21
3.3.1 Process Chart ................................................................................... 22
3.3.2 Purpose of the Operation ................................................................. 24
3.3.3 Material ........................................................................................... 25
3.3.4 Setup and Tools ............................................................................... 26
3.3.5 Working Conditions ........................................................................ 27
3.3.6 Material Handling ........................................................................... 27
3.3.7 Line Layout ..................................................................................... 29
3.3.8 Principles of Motion Economy ....................................................... 31
3.4 Assembly Line Balancing Problem ................................................................ 32
4 CASE STUDY .............................................................................................................. 39
4.1 Current Assembly Method ............................................................................. 40
vii
Page
4.2 As-Is Study ..................................................................................................... 41
4.3 Analysis and To-Be System ........................................................................... 42
4.3.1 Single Stage Parallel Line Configuration ........................................ 45
4.3.2 Five-Stage Serial Line Configuration ............................................. 47
4.4 System Evaluation .......................................................................................... 50
4.4.1 Objectives of the Simulation Analysis ............................................ 50
4.4.2 Material Handling System - Proposed Operation ........................... 51
4.5 Simulation Results.......................................................................................... 55
4.5.1 Material Handling Cart Capacity .................................................... 55
4.5.2 Material Handlers Required Supply Side ..................................... 56
4.5.3 Input Buffer Size ............................................................................. 56
4.5.4 Output Buffer Size .......................................................................... 56
4.5.5 Material Handlers Required Bar Coding Side.............................. 57
4.5.6 Analysis of Results .......................................................................... 58
5 SUMMARY AND RECOMMENDATIONS ............................................................... 61
5.1 Summary ........................................................................................................ 61
REFERENCES ................................................................................................................. 63
APPENDIX A .................................................................................................................. 66
APPENDIX B .................................................................................................................. 72
VITA ................................................................................................................................ 78
viii
LIST OF TABLES
Page
ix
LIST OF FIGURES
Page
Page
Figure 21 Results Comparing Two Configurations ......................................................... 58
Figure 22 Operator Utilization ......................................................................................... 59
Figure 23 Production Rate for Improved Assembly Line ................................................ 61
Figure A. 1 Station Time Distribution ............................................................................. 70
Figure A. 2 Loading and Unloading Time Distribution ................................................... 71
Figure B. 1 Assembly Layout With Dimensions ............................................................. 72
Figure B. 2 Material Handling Cart Routing Logic ......................................................... 73
Figure B. 3 Improved Assembly Layout ......................................................................... 74
Figure B. 4 Material Handling Supply Side .................................................................. 75
Figure B. 5 From Assembly to Bar-Coding ..................................................................... 76
1 INTRODUCTION
Assembly lines are one of the most widely used production systems. Productivity
of a manufacturing system can be defined as the amount of work that can be
accomplished per unit time using the available resources. Pritchard (1995) defines
assembly line productivity as how well a production system uses its resources to achieve
production goals at optimal costs. The conventional productivity metrics, namely
throughput and utilization rate gives a substantial measure of the performance of an
assembly line.
These two metrics alone are not adequate to completely represent the behavior of
a production system Huang et al (2003). A set of other measures such as assembly line
capacity, production lead time, number of value added (VA) and non-value added
(NVA) activities, work-in-process, material handling, operator motion distances, line
configuration and others, along with the throughput and utilization rate, completely
characterize the performance of a production system. An assembly line yields optimal
performance by an optimal setting of all these factors.
Flexibility and agility are the key factors in developing efficient and competitive
production systems. For products involving light manufacturing and assembly, this level
of flexibility can be easily achieved through the use of manual assembly systems.
Manual assembly lines are most common and conventional and still provide an attractive
and sufficient means production for products that require fewer production steps and
___________
This thesis follows the style of International Journal of Lean Six Sigma (IJLSS).
To establish the material handling system for the manual assembly line.
2 LITERATURE REVIEW
2.1 Lean Manufacturing
Lean Manufacturing or simply Lean is a production philosophy that targets the
identification and elimination of any waste in the production processes; especially
reduce waste in human effort, inventory, time to produce and production space etc. The
concept of Lean was originally developed by Toyota (TPS) for their automobile
manufacturing replacing mass production Womack and Jones (1990). According to
Womack, the primary focus of Lean is to maintain the value of the product with less
work. Lean drives a self-directed work-force and is driven by output-based goals aligned
with customer satisfaction criteria Elizabeth and Cassandra (2010).
Waste is generally caused due to unnecessary delays, processes, costs and errors.
The seven types of wastes associated with Lean are overproduction, transportation,
processing, inventory (work-in-process and finished goods), waiting, motion and defects.
These wastes are also associated with support functions involved in a production system.
The main focus of Lean is to address the value-added and non-value added activities. A
non-value added activity (NVA) is most commonly defined as any activity for which the
customer is not willing to pay. Lean necessitates the reduction of these NVAs by
making the system perform better while consuming lesser resources Czarnecki and Loyd
(2001). Some of the widely recognized benefits of Lean manufacturing include:
Productivity Improvement.
Quality Improvement.
Better utilization.
Safer operations.
Todays market environment demands for high quality products with low costs
with a greater variety in products and at faster response times. The manufacturer faces
the challenge to meet these demands while maintaining a profit. Implementing Lean is
an ongoing and long term goal. Proper defining of the goals suitable to a production
process and setting baselines is the key to productivity improvement.
2.2 Assembly Line Balancing
Moberly and Wyman (1973) propose the approach of using simulation two
compare two assembly line configurations. According to Moberly, the study of
production line configurations along the length of the line is called assembly-line
balancing. The set of work stations along the line that results from this balancing is the
generated line configuration. They demonstrate splitting the assembly line width wise
rather than length wise i.e., one workstation is replaced by two identical parallel stations
and they named it as dual production line. A comparison of two assembly line can be
seen in Figure 1.
Figure 1 Two Assembly Configurations and Comparisons (Moberly and Wyman, 1973)
The author presents the concept of expedited dual production line, the feature
which doubles the service rate of a non-failed work station. If one of the two parallel
workstations is failed, the operator from the failed station moves to assist the operator at
non-failed station and hence doubling the service rate. This is the main difference
between single independent line and dual assembly line configurations. The objectives
of this paper are to decide the best configuration to choose at the same given cost, two
single independent lines or one dual line. Also, the configurations are compared based
on the output rate. The model constraints are on the workstation failure rate and service
rate for finding the output rate. Simscript II was used to simulate the model.
This paper presents close resemblance to the topic under consideration, the only
difference being the assembly line configurations and the model constraints. The thesis
model proposed considers the configurations along the length of the assembly line. Apart
from the output rate, the operator or workstation utilization and the material handling
requirements are also considered.
Bartholdi (1993) designs a computer program to balance a two-sided assembly
line. The paper mainly focuses on the case study of a small utility vehicle manufacturing
line. The important point to be noted about two-sided assembly lines is that the operators
at each pair of work stations (mated-station) work on different tasks but on the same
individual component. Hence, Bartholdi puts forward that two-sided assembly lines are
more practical for large products like vehicles and heavy machinery than small products.
In contrast to Kim et al s (2009) proposition Bartholdis model tries to minimize
the number of work stations for a given cycle time, by restricting the positions where
tasks can be placed. The standard ALB problem considers assigning tasks only based on
the processing times. This paper poses constraints on certain tasks that they should be
always kept together. By doing so, the operator can learn more quickly and perform a
particular set of tasks efficiently. This is good as long as the model yields sufficient
results. The author mainly focuses on balancing two sided assembly lines where the
operators at each station share the work elements assigned to that station and work on
the same component simultaneously. This does not discuss about station imbalances and
their impact on assembly utilization.
Becker and Scholl (2006) survey the simple assembly line balancing problem and
several mathematical techniques that can be applied to solve this problem. They give an
Integer Programming and Dynamic Programming approach to solve an assembly line
balancing problem. Scholl and Becker (2006) define an assembly line balancing problem
as optimally distributing the assembly work among the m workstations with respect to
some objective. Given the number of work stations, m and assembly cycle time, c,
several assembly line balancing problems arise by varying the objective function. A few
examples are to minimize m given a c; minimize c given m, maximize the line
efficiency, E, by considering the interrelationship between c and m.
Kim, Song and Kim (2009) propose a genetic algorithm approach to solve a two
sided assembly line known as two-ALB. They present a mathematical formulation of the
two sided assembly line balancing problem with the objective of minimizing the cycle
time for a given number of paired workstations. They call these parallel paired stations
which perform similar tasks as mated-station. The advantages of the two-sided assembly
line over one-sided assembly are shorter line length, lesser material handling, reduced
tools and fixtures cost and better throughput. The performance of the GA is compared
with the heuristic approach and found to be better.
Rachamadugu and Talbot (1991) propose a heuristic procedure based on mixed
integer programming formulation of an assembly line balancing problem. They present a
procedure that rebalances an already balanced line such that the workload on all
number of workstations, and the mean work load W, is defined as W = (Total Work
Content/ m).
The authors suggest an iterative procedure to reduce this MAD by transferring
elemental work tasks from a station with higher than mean work load (W) content to a
station with lower than mean workload content. It must be made sure that this
transferring is between precedence tasks only. If there is no precedence restriction on
any task it can be assigned to any station so that MAD is reduced. The flowchart
showing the workload variation minimization procedure is given in Figure 2.
Although this method proves to smoothen the workload on the assembly
workstations, the model does not take into account the initial constraints or grouping of
tasks to be performed at a single workstation. This model cannot be applied successfully
under such circumstances.
10
Merengo et al (1999) analyses some of the most common issues associated with
manual mixed-model assembly lines. This paper focuses on reducing the number of
incomplete jobs at each assembly workstation. The problem formulation similar to
Rachamadugu and Talbot (1991), but the objective is to minimize the mean number of
incomplete jobs and even out the variation at each station to improve productivity of a
mixed-model assembly line.
Boysen, Fliedner and Scholl (2007), put together the variety of assembly line
balancing problems. The authors discuss about parallelization of assembly work i.e.
increasing operator productivity by partitioning the total work content among different
11
production units. This also includes division of work across several stations within the
same serial line by making sure that the average station time does not exceed the cycle
time. They survey a series of assembly line balancing problems faced and issues
associated with assembly line design problem.
12
As-Is Study
Operations Analysis
Assembly line Balancing
Material Handling Analysis
System Evaluation
Performance Evaluation
13
Different product family classification methods are available, the most dominant
in usage being the following methods
3.1.1 A-B-C Classification Method
The A-B-C classification process is a method that helps to identify products
families based on three importance ratings namely A-Outstandingly important, BModerately important and C-Least important. This classification makes use of Paretos
Principle which can be generally told as 20% of the products account for 80% of the
total dollar usage.
This method mainly focuses on:
With the demand and sales data in hand the classification procedure is as follows:
1. List the products along with the respective demand and sales values. This is
normally represented in terms of annual demand or for any relevant time
period.
2. Calculate the product of demand, D, and value, v, i.e., Dv for each product.
14
3. Arrange the products in descending order, starting with the product with
largest Dv value.
4. Then calculate the corresponding cumulative dollar usage and cumulative
percentage of total usage for all the products.
5. Then the products are classified into A, B or C classes according to the dollar
usage (Dv) values obtained. Initially products can be classified using the
following guidelines:
a. Class A: The first 5 to 10% of the products, as ranked by total dollar
usage, fall under this most important category. They generally
represent the 20% of the total inventory. Generally the account for
50% or more of the total dollar usage.
b. Class B: The products accounting for more than 50% of the remaining
dollar usage fall under this not so important class. Generally they
represent the 30% of the total products.
c. Class C: A majority of the remaining products fall in this least
important category and they represent only a minor part of the total
dollar usage.
Once an initial classification is obtained using the rules above, the decision
maker can modify it to take into consideration other important business criteria; i.e., a
product that has particularly long lead times, new critical products, products associated
with new products, but has a low Dv can be moved into Class A.
15
Process Sequence
Step 2
Step 3
Step 4
Step 5
Step 6
Step 7
Step 8
16
should be done for all the products. The Table 2 shows part-process matrix with process
steps marked against each product.
Process Sequence
Step 2
Step 3
x
x
x
x
x
x
x
x
Step 4
Step 5
Step 6
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Step 7
Step 8
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
The next step is to sort the parts based the processing steps. This sorting brings
together all the products with almost similar processing steps. The example in table 3
shows this sorting and grouping of products.
Table 3 Part-Process Matrix Showing Product Families
Process Sequence
Demand Actual Sales
Product Name
(Units/time) ($/Unit)
Product 2
Product 4
Product 7
Product 1
Product 5
Product 6
Product 9
Product 8
Product 3
Product 10
Step 1
Step 2
Step 3
Step 4
Step 5
Step 6
Step 7
Step 8
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
17
After identifying the desired product groups, they can be labeled for easy
identification. For the above example part-process matrix four product families can be
defined and they are labeled for easy identification as:
P: Products 2, 4 and 7
Q: Products 1, 5, 6 and 9
R: Product 8
S: Products 3 and 10
Out of these product families, the one that is expected to yield maximum output
(based on sales data) is selected. Also, under some special circumstances, a product from
a different family can be dropped in the product family considered for study if the total
work content for those families lies above 30%. For example, the product 8 can be
grouped with family P, if it is a critical component. The wastes or improvement
opportunities identified by following a single product family are likely to be translated in
equal proportions in every other product or family of products.
3.2 Time Study
Time study is a technique used to establish a time standard to perform a given
assembly operation. It is based on the measuring the work content of the selected
assembly, including any personal allowances and unavoidable delays.
It is the primary step required to determine the opportunities that improve
assembly operations and set production standards. The key objectives of a time study
are:
18
To increase productivity.
The following points are observed before starting the actual Time Study,
19
The obvious problems are taken care of first. There is no reason to spend
time describing and timing work elements that are obviously unnecessary or
redundant.
The time study analyst should familiarize with the assembly operations
Start the study capturing all work elements first (VA and NVA) once all
work elements have been captured, and then proceed to time them one by one.
Trying to do both simultaneously can be overwhelming and confusing.
The first step is documenting all the assembly tasks in their work elements before timing
them by observing the following points Ortiz (2006):
Each work element should be listed in the sequence it is performed. One part
of assembly might have many work elements associated with it. Any
repeating steps should be listed down as many times they occur throughout
the process.
The level of detail of the work element is such that it allow you to capture
enough detail to provide useful insight about the process, but not too much to
overwhelm you with unnecessary information.
Work elements should be defined in a way that they can be reliably timed. If
the level of detail is such that the work element is very short in duration, then
20
The work elements are basically categorized as set up tasks, actual processing
tasks and system or administrative tasks. Also, any movements associated
with performing these work elements (ex: moving the subassemblies, looking
for parts and tools etc.,) should be clearly noted. It is convenient to separate
the study of setups, processing and system processes.
The work elements are further marked as value added and non-value added
activities. Value added work is the actual work that is valuable and is
reflected in the final product. Examples of non-value added activities include
searching for tools, moving the sub-assembly to a different location for next
process, unnecessary moving of parts etc.
21
situation. In snapback timing method, the watch is set back to zero at the end of each
work element and the time for next work element is recorded.
It is recommended to take several time measurements for each work element
(e.g. At least 12 time samples for each work element). Recording multiple samples
allows us to better estimate the average time and also capture the variability associated
with a given work element. Work elements can be timed two different ways: as they
occur sequentially during the assembly process, or element-by-element. The Time Study
Data Sheet can be used to record the times during the time study. If abnormal events
occur while taking a sample, the sample must be discarded or annotated.
Calculations:
Remove outliers (those outside 3s) and recomputed until there are no
outliers. It is important to document and explain the reason why an outlier
occurred. Explaining outliers provides useful insight and knowledge about
the assembly processes Ortiz (2006).
22
Any operation improvement is a continuing process and with sufficient study of all the
operations, they can be practically improved. Some of the elements of Lean operations
analysis are as follows.
3.3.1 Process Chart
A process chart is a graphical representation of any manufacturing process or an
assembly operation. It contains the sequence of all operations in the order in which they
are performed and includes inspections, time allowances and materials used in any
business process from the arrival of raw material to the final product.
There can be a variety of process charts depending upon the specific application
such as the operation process chart, the man and machine process chart, the flow process
chart, the operator process chart etc. It is essential to document all the work elements
performed involved in an assembly process. The procedure that follows process charts
analyzes all the work elements and the non-value added activities are given special
attention with the goal of process improvement. During the operation analysis special
consideration must be given to
Material Handling
Plant Layout
Delay Times
Storage
23
and everything should be questioned. Answers should be given based on facts and actual
available data. The most important question that one should ask while analyzing an
operation is why?, this immediately leads to other questions related to the process that
take the form as shown below in Figure 4.
Answering these questions may generate ideas that will lead to the process
improvement. Questions should not be asked at random, but it should proceed
systematically, in the order in which they should be acted upon. For example, it is
unwise to question upon the tools and setup times is not recommended before defining
the purpose of the operation. When this systematic questioning approach is followed,
possibilities for improvements will be uncovered. The above discussed questioning
approach should be carried out along with the following operation analysis approaches
to analyze each operation and recognize improvements Maynard (2007).
24
25
Is the operation added to improve the appearance of final product? Is this cost
justified?
Can any feasible changes be requested from the supplier to avoid any
unnecessary operations?
Efforts should be made to obtain true answers to the questions. The result of the
analysis would eliminate any obvious unnecessary operations and provides a base to
subsequent operation analysis approaches.
3.3.3 Material
The following points should be considered related to direct and indirect materials
utilized in the process:
26
27
Improve the temperature and comfort conditions of the work area. This
reduces heat fatigue and cramps to the operator. Uncomfortable working
conditions sometimes cause operator stress and reduce the productivity.
28
Reduced handling costs: The labor costs, material costs and overhead costs
can be reduced due to effective material handling.
Following are the few points that help analyze and improve the material handling
system
material on the floor must be avoided. Material can be stored on pallets or trays
that can be picked up directly and moved to desired location. Advanced material
handling equipment like conveyers, portable elevators may be installed if
necessary.
29
30
31
The both hands should begin and end work at the same time. Work station
design should be improved so that operator can work with both hands at same
time. The work reach area is shown in Figure 6.
Hand movements should be limited to smaller areas and long reaches should
be avoided. This can be enabled by placing the frequently used objects close
to the operator and by installing efficient material and tool storage units.
32
33
Line balance
Material handling
Work-in-process management
Man power
Line size
Line configuration
All these factors are closely related with one another and have a considerable
impact on the assembly line performance as well as production cost. Various line
configurations would demand different material handling strategies and multiple levels
of line re-balancing so that the desired performance level can be achieved. Assembly
line design involves step-by-step approach by varying and analyzing each of these
factors and arriving at a best feasible design.
The operations analysis of a manual assembly system results in a set of
standardized production and assembly operations. The next step is to organize these
assembly tasks in an optimal manner to achieve the required targets. The important
decision problem that arises when constructing or re-configuring an assembly line is
assembly line balancing. The assembly line balancing consists of distributing the total
manufacturing work load uniformly among all the workstations present along the
assembly line. The overall performance of the production system is greatly affected by
this distribution of work.
34
Van Zante-de
Fokkert and de Kok (1997). For a given product, the entire assembly operation is broken
down into a number of work elements or tasks. Each work station performs some of
these assigned tasks. The product assembly is completed by sequential completion of all
the tasks.
Every product goes through the same sequence of assembly tasks in the same
order. The precedence relationship between the assembly work elements can be well
represented using a precedence network diagram. This on the other hand forms the basic
step in solving an assembly line balancing problem. Then, the strategic assignment of the
work elements to consecutive work stations in the assembly line follows with respect to
some objective. While doing so it must be ensured that the precedence constraints are
met.
Some of the terms associated with general assembly lines and their definitions
are given below Niebel (1982).
Precedence Diagram : The precedence diagram is a network showing the order of
assembly tasks in a sequence in which they are carried out and including the restrictions
on the performance of these tasks (such as position, precedence relationship etc.)
Minimum Rational Work Element: A minimum rational work element is an assembly
task that cannot be sub-divided into any further feasible tasks. The time taken by kth
work element can be denoted by Tek. Before starting the line balancing process, all the
35
work elements involved in an assembly must be clearly defined and the time taken for
each work element must be estimated.
Let,
S
Number of workstations, 1, 2, , m
Total Work Content: The total work content (Twc) of an assembly is equal to the sum of
all work element (k) processing times associated with that assembly.
Twc = nk Tek , for all values of k
Station Time : Station time is the total time available at each work station. It is the sum
of all the times of work elements that are being processed on a single work station (S).
Station time, Tsi
Cycle Time : Cycle time (Tc )can be defined as the rate of production. This is the time
between two successive assembled units coming out of the line. The cycle time can be
greater than or equal to the maximum of all times, taken at any particular station.
i.e. Tc max {Tsi}
If Tc = max {Tsi}, then there exist ideal times at all the stations, i.e. having station time
less than cycle time. It can be understood that the cycle time can be never less than the
station time.
Cycle time Tc = (available time or total work content time)/(Target production rate)
Line balance efficiency: This denotes the performance of the assembly line. It is given
by the ratio of total work content time to the total cycle time multiplied by number of
work stations.
36
Van Zante-de Fokkert and de Kok (1997) mention that the assembly flow lines can
be classified into the following classes.
37
Mixed Model Lines Handles the assembly of more than one product or
model. Each assembly station is designed such that various tasks needed to
produce any model that moves along the line can be performed. Most
consumer products have mixed model lines.
method,
network
and
assignment
problem
methods,
dynamic
programming etc. Rekiek et al (2002). Ranked Positional Weight (RPW) heuristic was
one of the first proposed heuristics to solve assembly line balancing problem. The
ranked positional weight is defined as the sum of the operation time of the work element
and the operation times of all work elements that come after it in the precedence network
sequence. Two things need to be considered while an assignment is being made (1) The
precedence relationship is maintained at all times and (2) The overall station time does
not exceed cycle time. The manual assembly line problem considered in the case study is
balanced using RPW heuristic approach. The procedure followed using RPW heuristic
approach is given below:
38
1. For all the work elements precedence events are identified and the network
diagram is drawn showing next to it the corresponding time taken to perform
the task.
2. The total work content time, workstation cycle time and the theoretical
number of work stations required are calculated.
3. The ranked positional weight for each node of the precedence network as
given below.
Let n i represent all the nodes in the path proceeding from node i in the
precedence diagram. Therefore, the RPW for node i is,
RPW (i) = Tei + Tej, where j n (i)
= (work element time) + (sum of work element times of all tasks
following node i)
4. Arrange the work elements in the decreasing order of the RPW value.
5. Following the ranked order, start assigning the work elements to work
stations, one station at a time. While the tasks are assigned to the first station
make sure that (1) the precedence relationship is maintained and (2) the
overall station time does not exceed cycle time.
element is making the station time value go higher than the cycle time,
proceed to check with the next work element. A new station opens when
there is no possibility of continuing assigning operations to the currently open
station. Proceed with assigning tasks to the next station in a similar way.
39
4 CASE STUDY
The three step productivity improvement methodology is applied to a real
problem consisting of a manual assembly line. The assembly line contains mobile phone
package assembly operations. The process involves initial disassembly, light assembly
and inspection operations. Each package comes in a master box which contains ten such
packages as shown in Figure 7. Once all the packages are ready they are placed in an
empty master box and the master box is moved to bar-coding area and then to the
shipping area.
Handset
Battery
Charger
Installation disc
40
Package box
Outer wrapper
Bar-coding stickers
41
42
43
with time study data and cycle time is shown in Appendix A. The precedence network
diagram for the standardized assembly is given in Figure 9.
It can be seen from the precedence diagram that certain tasks are grouped (like
tasks 20 to 26; 15 to 19) which enables improve operator efficiency. It is preferable to
have a single operator complete a sequence of operations and pass a finished subassembly to next stage. Setting such sub goals enables operators learn the tasks more
quickly and help perform them more dependably and faster. Also grouping of tasks
facilitates better material storage at each stage, supports visual management of work
space and mistake proofing (Poka-Yoke).
4.16
START
3.85
30
0.89
1.76
15
0.83
0.9
1.59
18
19
1.7
2.32
10
0.96
0.74
11
0.62
0.77
12
0.97
20
22
21
1.2
23
24
25
1.19
2.53
26
0.74
27
28
1.06
1.99
29
1.68
2.65
13
1.23
16
3.87
14
1.02
17
4.40
31
32
2.24
4.51
33
FINISH
1.48
44
Operations analysis step also results in selecting the most suitable assembly line
layout, which further helps in planning a good material handling system. Taking into
account the total assembly time required to produce one package (which is considerably
small), the simplicity of the assembly operations, the feasibility to modify the existing
layout without causing much effect on current production, the traditional straight line
configuration is chosen. A straight line configuration is well suitable for assemblies
involving operators perform a set of tasks continuously in a given sequence for all the
products (Aase et al, 2004).
The two proposed assembly line configurations for the current assembly method
are shown in Figure 10. The next step to improve the assembly line productivity is to
design and balance the assembly line accordingly to satisfy the cycle time and demand
requirements.
Both the configurations take into consideration Lean manufacturing principles
such as Standard Work, 5-S, Visual Controls, Kaizen (Continuous Improvement) and
knowledge sharing, to improve productivity, reduce work-in-process inventory, floor
space reduction, minimize operator unnecessary motion and reworks. A brief description
of each configuration with the workstation specifications follows.
45
46
The light signals at each table serve as Lean visual management tool and allow efficient
material handling.
Green light ON denotes that there is a box in the buffer at the table.
47
Having the input buffer shelf helps reduce the excess inventory of master boxes
at the assembly station as compared with original method. Also, with the single stage
layout the floor space usage is reduced from 42.07 m2 to 24.22 m2.
4.3.2 Five-Stage Serial Line Configuration
The assembly table consists of five work stations and each stage is assigned with
a defined set of work elements. The work elements are assigned to each station using
Ranked Positional Weight (RPW) heuristic method (Section 3.4). The balanced line with
five assembly stages is shown in Figure 13.
48
Start
3.85
Station 1
0.89
15
18
3
1.7
0.96
0.83
0.9
30
1.76
4.18
19
1.59
10
11
2.32
0.74
0.62
0.95
0.77
12
20
22
1.48
21
24
25
26
27
2.53
0.74
1.06
28
29
31
32
1.68
2.22
4.51
33
1.22
23
1.19
2.65
Station 3
1.99
Station 4
Station 5
FIN
1.23
13
1.02
3.87
14
4.4
16
17
Station 2
After the completion of tasks at each stage, the components or sub-assemblies are
pushed on to a conveyor located along the center of work table by using a material tray.
The operator at the next stage pulls the tray from the conveyor and completes the
assembly. Once the package reaches the end of assembly table it is placed in the master
box and then the master box is moved to bar-coding area by a material handler. The light
signals and input buffers help make the assembly line Lean. Figure 14 and Figure 15
give the design of assembly workstation and assembly line for the five-stage
configuration.
49
The conveyor at each assembly stage can hold only two material trays. This
prevents excess work-in-process inventory in terms of packages. The stopper acts as
mistake proofing tool by avoiding accidental tray movement to the next stage.
50
The number of material handling carts required to deliver the master boxes
from storage area to assembly tables.
51
52
fill material. The cart always moves along the specified path and in one direction only.
And also the material is always filled in the order in which the cart moves. The finished
master boxes are transported to bar-coding area in a similar fashion. See Appendix B for
detailed material handling logic followed by carts.
The parameters used by the cart and related assumptions are given below
The cart speed is determined experimentally so that it is not too fast or too
slow. Cart speed = 0.7 m/s.
The carts loading and unloading times at each assembly table are Triangulardistributed with mean 4 seconds and lower limit and upper limit as 3 seconds
and 5 seconds respectively (Appendix A).
Once this logic is set, the buffer size required at each assembly table, the cart
capacities and the quantity needs to be determined. The simulation model is tested for
two assembly line configurations with different material handler inputs. The results in
terms of average station utilization and the part output are plotted against the quantity
carried by material handling cart. The detailed description of material handling and
routing logic is given in Appendix B.
The schematics of the simulation model are shown in Figure 16 and Figure 17
respectively.
53
Figure 16 Simulation Model for Single Stage Parallel Line Configuration Showing Pick-up
Point
The model contains 4 assembly modules with 16 tables and operators in each
module.
The material handlers move along the path shown in grey across the tables.
The conveyor is located between the tables.
Cycle Time has Normal Distribution with mean = 538.7 seconds and std. dev.
= 10 (for one complete master box assembly) (Appendix A).
54
Figure 17 Simulation Model for Five Stage Serial Line Configuration Showing Pick-up
Point
The model contains 11 assembly modules with two assembly tables in each
module. Each table contains five stations and one operator working at each
stage.
Each module contains a worktable at the front of the line which pulls out
packages from master box and places sim card on each package. The
packages are then pulled by assembly workstation one at a time.
55
The material supply carts move along the path shown in green and the
finished master boxes are moved to bar-coding area along the path shown in
grey. A conveyor is located between the tables.
4.5 Simulation Results
4.5.1 Material Handling Cart Capacity
For single stage line it can be seen from Figure 18 that at cart capacity as 6 boxes
maximum utilization is achieved. The idle time for material carts increase when the
capacity exceeds 6 units although utilization is 100%, which is not recommended.
Similarly for five-stage line, maximum table utilization is observed at a capacity of 6
boxes. So, for both the configurations the material handling cart loads 6 boxes per trip.
100%
100%
88%
95%
87%
90%
86%
86.8%
86.8%
86%
85%
86.8%
85%
80%
75%
100%
73%
84%
70%
83.5%
83%
65%
82.8%
82%
60%
81%
55%
80%
50%
4
56
100%
88%
95%
87%
90%
86%
99.70%
85%
85%
86.90%
84%
85.50%
80%
83%
82.30%
75%
Buffer Size - 1
Buffer Size - 2
82%
Buffer Size-1
Buffer Size-2
57
buffer capacity is obtained for single stage line as 5 units and for five-stage line as 2
units per table
4.5.5 Material Handlers Required Bar Coding Side
The single stage line requires two operators to carry finished master boxes to bar
coding area. The five-stage line requires three material handlers with carts to transfer
master boxes to bar coding area. This is determined based on how the finished box
removal from output buffer affects the assembly utilization. The material handling
requirements based on the table utilization is shown in Figure 20.
Avg. Table Utilization
100%
100%
96%
96%
95%
88%
87%
90%
86%
86.8%
86.8%
86.8%
85%
85%
80%
80%
75%
84%
70%
83%
65%
60%
83.5%
82.8%
82%
60%
81%
55%
54%
80%
50%
1
2
No. of Operators
58
Parameter
2 Operators
Cart capacity
6 Boxes
6 Boxes
2 Boxes
2 Boxes
5 Boxes
2 Boxes
Box Output/Operator/Hour
60
4.5
59
58
3.5
4.4
57
59.77
58.1
56
55
59
The consolidated results comparing the two assembly line configurations are as
follows.
Tables Served Per Material Handler: Number of tables served by each material handling
unit is higher for five stage serial line configuration. Figure 21 shows that the five stage
serial line requires less material handlers than the single stage line. The number of tables
to be served is lesser in five stage configuration compared to the single stage
configuration. But it can be observed that the difference is not highly dominating.
Productivity: The single stage configuration gives output as 59.7 boxes/operator/hour
where as five stage line gives 58 boxes/operator/hour. There is a considerable
improvement in productivity in both the assembly lines from the original method.
Operator Utilization: Figure 22 shows that the average operator utilization for single
stage line is about 99% and for five stage line is 86.9%. It can be seen that for a fivestage line all the operators at different stages of assembly line are not uniformly utilized.
100%
120%
98%
96%
100%
94%
92%
90%
98.05%
89.50%
88.17%
86.73%
80%
63.51%
99.10%
88%
60%
86%
84%
86.90%
40%
82%
20%
80%
0%
Stn. 1
Stn. 2
Stn. 3
Stn. 4
Stn. 5
60
Operation
Take individual box
Avg. Time
0.96
3.85
7
8
9
10
11
12
13
15
16
17
20
21
22
23
24
25
26
27
14
18
19
28
29
30
31
32
33
0.83
0.90
1.70
2.32
0.74
0.77
1.23
0.89
3.87
4.40
1.22
0.95
0.62
2.65
1.19
2.53
0.74
1.06
1.02
1.76
1.59
1.99
1.68
4.18
2.22
4.51
1.48
Work Stn.
Station Times
Cycle Time
Imbalance
Stage 1
11.31
-0.54
stage 2
11.16
-0.39
10.77
stage 3
10.97
-0.20
stage 4
8.04
2.73
stage 5
12.39
-1.62
Hence, it is recommended to implement the single stage parallel line in order to achieve
higher productivity and better overall assembly performance.
61
Box/Operator/Hour
70
59.7
60
50
40
35
29.77
30
20
10
0
Initial Target
Measured
Improved Line
62
63
REFERENCES
Aase, G. R., Olson, J. R. and Schniederjans, M. J. 2004 , U-shaped assembly line
layouts and their impact on labor productivity: An experimental study, European
Journal of Operational Research, Vol. 156 No. 3, pp. 698-711.
Banks, J., Carson, J. S., Nelson, B. L. and Nicol, D. M. (2000), Discrete-Event System
Simulation, Prentice Hall, Upper Saddle River, NJ.
Bartholdi, J.J. (1993), Balancing two-sided assembly lines: A case study, International
Journal of Production Research. Vol.31 No.10, pp. 2447-2461.
Becker C. and Scholl A. (2006), A survey on problems and methods in generalized
assembly line balancing, European Journal of Operational Research, Vol. 168
No. 3, pp. 694-715.
Boysen, N., Fliedner, M., and Scholl, A. 2007 , A classification of assembly line
balancing problems, European Journal of Operational Research, Vol. 183, pp.
674-693.
Chow, We-Min. (1990), Assembly Line Design: Methodology and Applications, Marcel
Dekker Inc., New York, NY.
Cudney. E. and Elrod. C. (2010 , Incorporating lean concepts into supply chain
management, International Journal of Six Sigma and Competitive Advantage,
Vol. 6 Nos. 1-2, pp. 12-30.
Czarnecki, H. and Loyd, N. 200 , Simulation of lean assembly line for high volume
manufacturing, Proceedings of the Huntsville Simulation Conference, Available
online at: http://www.scs.org/confernc/hsc/hsc02/hsc/papers/hsc037.pdf.
64
65
Ortiz, C.A. (2006), Kaizen Assembly: Designing, Constructing, and Managing a Lean
Assembly Line, Taylor and Francis Group, Boca Raton, FL.
Pritchard, R.D. (1995), Productivity Measurement and Improvement: Organizational
Case Studies, Praeger Publishers, New York, NY.
Rachamadugu, R. and Talbot, B.
66
APPENDIX A
Table A. 1 Work Element Sheet with Time Data
Date :
Company : ABC
Product code:
Work
Type of Activity
Time
Element
Work Element
Sequence
VA
NVA
Avg.
SD
No.
1
0.82
0.21
3.10
1.91
2.80
1.06
7.82
1.89
1.42
0.61
10
1.91
0.50
11
1.93
0.74
12
3.89
1.55
13
3.56
1.88
14
1.87
0.55
15
2.44
0.82
67
policy
16
17
18
Pass to person 2
19
20
21
22
23
24
25
26
27
28
29
1.05
0.50
0.80
0.32
0.57
0.16
0.69
0.24
2.36
1.05
1.78
0.65
0.76
0.34
5.49
2.53
1.43
0.75
1.00
0.82
1.29
0.69
2.01
0.51
1.69
0.45
14.53
44.56
30
1.70
0.69
31
2.21
1.52
32
Pass to person 3
0.71
0.39
33
1.40
0.66
34
0.76
0.30
2.12
0.98
35
pamphlets
36
2.96
1.54
37
1.87
0.86
68
38
0.98
0.59
39
2.41
0.73
40
3.10
1.32
41
7.09
1.53
42
Pass to person 4
0.77
0.40
43
0.78
0.21
44
4.02
1.88
45
6.72
3.44
46
1.36
0.43
47
48
Move to storage
Total
Time
Key:
S Set-up Tasks
A Administrative or System Tasks
NVA Non-Value Added Activities
107.99
Seconds
Note: Operation description and time data slightly modified from the original data.
10
11
12
1.33
4.24
1.13
1.22
3.78
2.49
0.95
1.00
1.12
1.08
1.70
3.54
3.91
2.27
0.84
1.91
1.04
0.61
2.42
1.08
3.51
0.88
2.38
2.63
1.90
4.92
1.56
5.25
1.56
1.18
3.30
0.65
0.67
1.66
3.31
0.68
0.91
1.16
0.68
1.30
3.62
7.51
1.37
1.48
1.83
0.57
0.65
2.11
1.27
2.21
0.59
1.81
1.78
1.99
4.42
2.00
4.32
1.18
0.68
3.10
0.52
1.15
1.83
1.64
0.41
0.59
1.14
1.07
0.73
3.88
4.53
1.43
1.19
1.16
0.83
0.30
1.94
0.94
2.06
0.72
0.71
2.50
1.47
5.12
2.56
4.50
1.39
0.87
2.68
0.63
0.52
1.68
1.59
0.70
0.57
2.09
1.09
1.17
6.89
5.14
1.45
2.12
1.23
0.94
0.73
1.98
1.08
2.48
0.50
0.81
1.52
1.28
4.42
1.06
4.14
1.62
1.00
5.90
0.86
0.74
1.42
2.40
0.33
1.13
0.93
1.30
0.30
2.92
4.06
2.26
3.23
0.98
1.11
0.80
2.89
0.70
2.36
0.66
0.80
1.55
1.64
2.46
1.67
3.90
1.62
0.87
2.72
0.72
0.72
1.26
2.45
1.77
0.68
0.79
1.07
0.86
4.19
5.39
1.53
1.14
1.02
1.06
0.59
3.51
0.88
2.89
0.62
0.87
1.66
1.56
3.64
1.75
6.86
1.36
0.87
4.62
0.62
0.72
1.36
2.26
1.38
1.63
1.43
0.76
0.76
3.21
4.11
1.80
1.60
1.24
1.12
0.44
2.96
1.56
1.39
0.79
0.68
2.28
1.81
5.95
2.10
4.62
1.41
0.85
4.58
0.61
0.94
1.64
2.40
0.57
0.31
1.28
0.90
0.82
3.43
3.20
1.63
1.56
1.14
1.18
0.64
3.11
1.91
3.01
0.81
0.81
2.69
1.56
4.37
2.75
4.08
2.35
0.89
4.09
0.92
1.12
1.74
2.36
0.56
0.43
1.13
1.37
0.80
3.47
5.10
1.72
1.92
0.97
0.89
0.76
3.01
1.09
2.01
0.67
1.13
1.89
1.86
3.20
4.70
4.19
1.18
1.01
3.97
1.30
1.50
1.25
1.12
0.56
0.70
1.24
1.02
0.74
3.87
3.27
1.54
1.07
0.85
0.90
0.56
2.62
1.06
3.11
0.87
0.65
0.88
1.37
4.07
2.13
3.19
1.23
0.83
4.31
1.22
0.69
1.71
3.73
0.39
0.66
1.16
0.88
0.73
3.69
3.02
1.97
1.17
0.91
0.85
0.65
3.00
1.80
2.00
1.00
1.18
2.34
2.24
3.82
2.21
4.31
1.31
1.17
2.71
0.75
0.81
1.08
2.12
0.63
0.64
1.30
0.97
0.74
3.72
3.57
2.16
1.76
1.43
0.96
0.74
2.30
0.89
3.32
0.73
0.91
2.16
1.50
3.74
2.16
4.78
1.56
0.96
3.85
0.83
0.90
1.70
2.32
0.74
0.77
1.23
1.02
0.89
3.87
4.40
1.76
1.59
1.22
0.95
0.62
2.65
1.19
2.53
0.74
1.06
1.99
1.68
4.18
2.22
4.51
1.48
0.18
0.94
0.25
0.28
0.67
0.68
0.41
0.34
0.3
0.19
0.34
0.96
1.21
0.32
0.61
0.33
0.16
0.14
0.48
0.36
0.61
0.13
0.5
0.52
0.27
0.88
0.86
0.85
0.3
0.05
0.27
0.07
0.08
0.19
0.20
0.12
0.10
0.09
0.05
0.10
0.28
0.35
0.09
0.18
0.09
0.05
0.04
0.14
0.10
0.18
0.04
0.14
0.15
0.08
0.25
0.25
0.25
0.09
53.87 seconds
Remarks
69
Allowances
Work Element
Standard
Error
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Time Samples
Std. Dev.
Work
Element
Sequence
No.
Date :
Avg. Time
70
The mean station times follow a Normal Distribution and the loading and unloading
times are estimated to have Triangular Distribution. The distribution is found from the
available data for station times.
Table A. 3 Time Study Data Sheet with Station Times
Avg. Station
Cycle Time in
Time in
Std. Dev.
Seconds
Seconds
Stn 1
11.31
1.52
Stn 2
11.16
1.72
10.97
1.12
Stn 4
8.04
0.96
Stn 5
12.39
1.6
Stn 3
10.77
12
10
Frequency
8
6
4
2
0
-1
71
The actual data is not available for loading and unloading times. Hence, it estimated that
these times lie between 3 seconds and 5 seconds. MINITAB Statistical Tool is used to
draw both the distributions.
Distribution Plot
Triangular, Lower=3, Mode=4, Upper=5
1.0
Density
0.8
0.6
0.4
0.2
0.0
3.0
3.5
4.0
4.5
5.0
72
APPENDIX B
Material Handling Logic
Single Stage Parallel Line Configuration
The dimensions and the layout used for simulation of Single Stage Parallel Line
configuration is given below.
73
Both the carts 1 & 2 pick-up boxes at the pick-up point and go to A.
At A,
Cart 1 checks for the total number of boxes across buffers in line 1 and
line 4. It then proceeds to the fill the line with less number of boxes .
At B,
Cart 2 checks for the total number of boxes across buffers in line 5 and
line 8. It then proceeds to the fill the line with less number of boxes.
74
At C,
Cart 1 proceeds to fill lines 2 and 3. If empty, goes to the pickup point.
If cart 2 is empty, it proceeds along line 3 to the pickup point, else goes
to fill lines 6 and 7.
At D,
Then, the finished master boxes are moved to bar-coding manually by operators.
PALLETS
0.75 m
1.50 m
1.50 m 1.40 m
0.5+0.8
2m
0.50 m
2.90m
12.75 m
14.25 m
0.7+0.8+0.5
3m
7 m
4 m
4 m
4.80 m
7 m
Bar-Coding area
4.80 m
Material Movement
from Tables to Bar
Coding
75
Carts 1&2 load master boxes at the pickup point and proceed to point A
At point A, cart 1 goes to fill tables numbered from 1 to 6. Cart 2 goes to fill
tables numbered from 7 to 11
76
After unloading at each table, if the empty carts go to the pickup point to load
master boxes through the path shown in red. Else proceeds to the next table
At point B, if the cart 2 has enough master boxes goes to fill tables 6 to 1, else
The carts wait until at least one master box is ready at each of the buffers
77
Cart
Cart
empty goes back along the same path to load boxes at tables ,2 and 3
Cart 2 unloads at bar-coding stations BC4, BC3 and in that order. Once
empty goes back along the same path to load boxes at tables 4,5 and 6
It is assumed that the empty master box is available at the end of table
78
VITA
Name:
Pranavi Yerasi
Address:
Email Address:
Education: