See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/234059244
A Simulation approach to improve assembly line performance
Article in International Journal of Industrial Engineering · January 2011
CITATIONS READS
13 2,822
3 authors, including:
Bernardo Villarreal
Universidad de Monterrey
37 PUBLICATIONS 511 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
EMS faciliy and ambulance location View project
EMS operations View project
All content following this page was uploaded by Bernardo Villarreal on 22 May 2014.
The user has requested enhancement of the downloaded file.
International Journal of Industrial Engineering, 18(6), 283-290, 2011.
A SIMULATION APPROACH TO IMPROVE ASSEMBLY LINE
PERFORMANCE
Bernardo Villarreal and María del Roble Alanís
Universidad de Monterrey, Departamento de Ingeniería, I. Morones Prieto 4500 Pte., San Pedro Garza García, N.L.,
México 62638
Corresponding author: Villarreal, Email: bvillarreal@udem.edu.mx
A fundamental characteristic of today´s competitive environment is the need for shorter product life cycles and
increased demands for customization. These aspects are difficult to satisfy operating with traditional production lines.
The development of JIT U-lines has been an emerging response to compete in this type of environment. The present
work describes the utilization of simulation to guide the improvement efforts during the redesign of a traditional
assembly line system in a Mexican manufacturing facility. The approach taken is a two-level one. A macro model that
simulates operations at plant level and assesses the synchronization of material flows between warehouses and
assembly lines and is used to determine materials handling resource requirements and overall layout options. The
second level is detailed and at the line level. Here, Balancing, operator assignments and buffer sizes are defined for
each line. The space required and detailed layout for each line are also determined.
Keywords: Simulation; assembly line performance; just in time
(Received 11 April 2009; Accepted in revised form 10 September 2010)
1. INTRODUCTION
A fundamental characteristic of the current competitive environment is that of short product cycles and a higher
demand of make to order items according to consumer desires. These conditions originate a significant increase in
product variety. These needs are difficult to satisfy operating with traditional production and/or assembly lines.
The development of Just in Time (JIT) production lines represents a valuable response to compete adequately under
the environment previously described (Miltenburg, 2001). Carrier de Mexico, S.A. de C.V., a Mexican company, faces
the challenge of surviving in a similar competitive environment. In addition, this firm presents an impressive growth
due to the re-location of production facilities from the United States of America to the facilities currently in place at
Santa Catarina, Nuevo Leon, Mexico.
In order to face this situation, Carrier decided to assess the current production line system with the purpose of
designing and implementing a new scheme that would improve flexibility, responsiveness and decrease waste levels.
The present work describes a methodological proposal that includes simulation as the main tool for designing the new
production scheme. The following section offers a review of relevant literature to the design of production and
assembly cells, and the use of simulation for this purpose. Next, section III illustrates in detail the design methodology.
Section IV provides an application of the methodology and a description of the results obtained. Finally, conclusions
about the project are presented in the final section.
2. REVIEW OF RELEVANT LITERATURE
The problem of designing a manufacturing cell has been treated exhaustively in the academic literature. Most of the
work can be classified into two areas; The first one concerns the development of ideas and concepts associated to the
design of lean manufacturing cells, (Ohno, 1988 and Monden, 1983), among others. On the other hand, a great bulk of
the literature focus on the solution of specific aspects related to the design of manufacturing cells. Most of it treats the
phase of cell formation (Beaulieu, 1997, Christy, 1986 and Greene, et al. 1984).
Simulation has been one of the original methodologies or tools of Operations Research. Initially, its use was very
limited because it required a profound knowledge of computational programming to simulate a model, in addition to
the required effort of modelling and corresponding knowledge of the problem in question. The computer languages
used were of the first level. As these languages were developed to facilitate the simulation (i.e., GPSS-H), the
application of the methodology increased significantly. Currently, the use of simulators such as ProModel, Witness and
others have facilitated enormously the application of simulation to the design, improvement and validation of systems
in a wide variety of areas of knowledge (Bowersox, 1972, Heinz, 1998, Park, et al. 1998, and Lee, et al. 2004).
The use of simulation to the problem of designing manufacturing cells is recommended to incorporate the uncertainty
and interdependency inherent in their operation. Ruiz-Torres, et al. 2008 apply simulation to investigate the
performance of a number of multi-cell configurations obtained by varying cell size, worker flexibility and shop type.
Cochran, et al. 1998 use it to verify the feasibility of operating a cellular system in a company of the automotive
industry. The purpose of the change was the reduction of the product manufacturing flow time. To achieve this
ISSN 1943-670X © INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING
Villarreal and Alanis
objective, the authors developed a simulation model for each alternative of the cellular system using Witness, a discrete
event simulation software. The focus of the analysis was micro, that is, at the cell level detail.
Rajamani, et al. 1990 describe their experiences in the design and balancing of a TV assembly line. In the initial
stage, the authors apply an algorithm based on costs under a stochastic environment to assign activities to work stations
(Kottas, et al. 1973). Then, the assignment is modified to consider the inter-dependencies among activities. Finally, the
line is simulated with a model developed using SLAM-II to estimate the work in process inventory levels needed to
guarantee an adequate flow in the line. In this case, the analysis is also at the micro level.
Suraj, et al. 1996 propose a simulation model developed at two levels to design a manufacturing cellular system. At
the macro level, the system is modelled and simulated at the plant level. At this level, the interest is in the equipment
layout and the achievement of a simple and synchronized material flow. The analysis of the operation of each
manufacturing cell represents the micro level of the model. The important issues to define at this level corresponds to
the line balancing, the level of WIP in the line, and the utilization of the resources such as operators and equipment.
The model is applied to analyze and redesign the manufacturing system of a plant that fabricates valves and pipe
accessories. Both simulation models, at the micro and macro levels, are developed with SIMAN. It is important to
emphasize that these models were disconnected, that is, they were run independently, and thus, part of the stochastic
behaviour generated at the cell level was not considered at the macro level.
It is important to clarify that when designing a system, the alternatives generated about its structure could be
numerous. These will depend upon the quantity of variables of interest that could be controlled, and that are key to
define the structure. The region of possible solutions could be very wide. For this reason, if we use simulation to look
for the best design, it will be necessary that we use a search strategy for finding the values of the relevant variables
associated with it. Carson, et al. 1997 give a general review of the methods used to achieve such purpose. The previous
process has been called simulation optimization. A model of this process is described in Figure No. 1. Here, the result
of the simulation model is used in the optimization strategy to provide feedback about the progress made towards the
optimum solution, which in turn is used as an input for the simulation model.
Input
Simulation Output Optimization
Model Strategy
Figure 1. Model of Simulation Optimization Process
The optimization strategy of the model includes normally a methodology such as genetic algorithms, stochastic
optimization, gradient search heuristics and the like. Due to the importance of this process, current simulation software
packages have developed modules that include special search procedures. ProModel includes a module called
SimRunner. AutoMod includes an additional package called AutoStat, and Micro Saint 2.0 incorporates OptQuest as a
module that carries out the process.
3. DESCRIPTION OF DESIGN METHODOLOGY
The methodology proposed for designing manufacturing systems consists of four stages (see Figure No. 2). The first
stage has the goal of tying corporate strategy to the design problem. Here, the competitive factors and their levels are
defined; Quality, cost, flexibility and response time levels are established. In addition, product volumes and mix are
also determined. This stage delineates the operating context to satisfy by the designed alternative production systems.
The second stage consists of a detailed analysis of the current production system, including the status of its
performance level. The third stage is structured to generate the production scheme that satisfies the fundamental
principles of the Just in Time Philosophy. It is in this stage that a scheme similar to that of Suraj¨s is suggested. Finally,
there is the implementation phase. This considers an evaluation step, and an iterative cycle to incorporate improvement
ideas. A brief description of each stage is given below.
Define Operations Analysis of System Re- New System
Strategy Current System design Implementatio
n
Figure 2. Description of General Design Methodology
284
Simulation of Assembly Line
• Operations strategy definition.
Previous to the design of the production system of a firm, it is required to establish the operating strategy. This phase
initiates with the identification of the needs of the market. One should know how the orders compete for a market
share. Is it through better quality? Or reduced prices? Delivering faster?.After identifying the relevant order winners
and qualifiers, and establishing their respective levels, the Administration must define the direction that should guide
the company´s efforts.
• Analysis of current situation.
At this phase a detailed analysis of the products assembled or fabricated in the production system under study is carried
out. For each product, the process sequence and flow is studied using layout analysis and time and motion studies for
each work station. These studies help us to identify improvement areas for each work station and the overall general
system. At this stage it is also recommended to develop a simulation model of the current system at both levels, micro
and macro. The model should help us to identify additional improvement areas at the plant level that result from the
stochastic behavior and the interaction and synchronization of the manufacturing cells.
• Design of manufacturing system.
With the previous model and additional information regarding future product design changes, the entry of new products
and sales forecasts, the design stage follows. Here, the methodology propossed incorporates an analysis scheme that
has the following elements:
- Simulation models at the macro (plant level) and micro (cell level), similar to the approach
suggested by Suraj, et al. 1996.
- Communication between models provided through the simulation language or the use of
information about processing time distributions that result from each micro model in the macro
model.
- The utilization of the simulation optimization process using heuristics and optimization tools ad
hoc to the cell design problem.
- The incorporation of JIT concepts in the design of the new system.
• Implementation of new system.
In this stage, the goal is to participate actively in the physical change and startup of the new production system. In this
section, the activities required to uninstall, transfer, install and carry out the pilot testing are described.
4. INDUSTRY APPLICATION
The methodology for designing manufacturing cells was applied in Carrier de Mexico, S.A. de C.V., A Mexican
company that is dedicated to fabricate and assemble air conditioned, ventilation and refrigeration units for the Mexican
and International markets. The firm is part of Carrier Corporation Inc. At the end of 1999, the top management of
Carrier de Mexico undertook a new strategic direction that emphasized the following points:
- Improve drastically customer response time.
- Achieve world class manufacturing levels.
- Decrease work in capital levels.
- Focus in cost reduction.
Carrier de Mexico assembles a high variety of products and spares for the residential, commercial and industrial
markets.
4.1 Operations Strategy Definition.
Carrier de Mexico determined its Operations Strategic Plan for the next five years starting in year 2004. This was
devised with the purpose of maintaining a world level of competitiveness. Two elements that have an important impact
in Carrier´s performance are fundamental. The first element has to do with the firm´s growth. Carrier Corporation took
the decision of transferring some of its assembly operations located at USA to Mexico, and more specifically to those
located at Santa Catarina and Garcia, N.L. The second element consists of strategies to achieve better competitiveness
levels. As part of these strategies, the company considered necessary the re-organization of the assembly lines
transferred with the objective of reducing operating costs and improve market response in terms of delivery time and
volume and variety flexibility.
The installations that would receive the assembly lines were the current plants at Santa Catarina and Garcia, and a
distribution center that would be transformed into a manufacturing site. In order to obtain a better level of
competitiveness, Carrier de Mexico decided that the new assembly resources were to be organized under the Just in
Time scheme. This implied that the organization should seek for a waste free operation with a high degree of personnel
participation and flexibility.
4.2 Analysis of current situation.
The purpose of this stage is to obtain a deep knowledge of the current production system and its level of performance.
The company counts with ten assembly lines; A Coil, N Coil, ICP, PTAC, GRAC, WRAC/GCDUS, Fan Coil, CDUS,
Coolers and 40RM. These lines are organized in the traditional manner, and most of the activities carried out are
285
Villarreal and Alanis
manual.. It is important to notice that the operators were not ready for a change in shop floor organization due to the
following reasons:
- They were specialized in the activities of the work station only.
- Just 27% of line operators are not temporary.
- Only 8% of the line operators are fully multi-functional.
The analysis of the operations included time and motion studies for each work station of each assembly line. The
coordination of the operators in each work station was also analyzed through Joint man-machine diagrams. With these
studies, the degree of balancing for each line was identified finding a lack of synchronization and high levels of
idleness in the stations.
Finally, the full analysis was complemented with the development of simulation models for each line, using
ProModel. The models were validated yielding the current performance conditions of the lines. As a result of the
analysis, several areas for improvement were identified, considering that the new production line should have Just in
Time characteristics.
- Operators with a lack of multi-functionality.
- Unbalanced assembly lines with a high degree of idleness and operators.
- A high WIP level.
- Low response assembly time.
4.3 Design of the new production system.
The simulation model of the traditional assembly system was used as the basis to initiate the design phase. Various
modifications were assessed using the model. Some of these are:
- Balanced lines.
- Degree of multi-functionality.
- Various changes in the work station layouts.
- Modifications in activities sequencing.
- Changes in product mix assembled in each line.
All of these modifications were analyzed considering their impact on cost, WIP level, response time, quantity of
operators required and the space necessary to accommodate the resources for each line. The final layout suggested for
the lines was in “U”. This form of layout is recommended throughout the literature because of its significant impact on
operating performance (Hyer, et al. 2002, Miltenburg, et al. 2001 and Sekine, 1990). The model representing this type
of layout for Carrier’s assembly lines is shown in Figure 3.
Figure 3. Simulation model for the proposed assembly cells.
During the process of cell balancing the team recommended the use of Sparling’s algorithm (Sparling, et al.
1998). The layout of the stations was defined utilizing the basic principles of plant distribution looking for the
reduction of trips and distance. The degree of multi-functionality was determined after evaluating several scenarios
with different amount of operators with different degree of multi-functionality. The SimRunner module was used to
make the previous step efficient.
286
Simulation of Assembly Line
The design at the macro level was developed with the simulation models for the selected cells located at each
installation; The current at Santa Catarina, the new one at the former distribution centre, and the one at Garcia. At this
level, the objective was to define the following:
- The general layout for each installation looking to minimize distance and trips.
- The required handling capacity.
- The raw materials inventories, finished product inventory and WIP levels, and the required space.
- The lot size and supply frequency from each supplier.
With the purpose of considering the impact of the variability and inter-dependency of all the cells included in each
facility, the design team considered the distribution of assembly times from each cell in the general facility simulation
model. There was another alternative to this approach evaluated that consisted in the development of a full complete
general model. This option was discarded because it required longer running times and a more complex analysis. Other
important input of the model was the arrivals distribution for each important supplier.
4.4 Implementation and results of the proposed systems.
The results of the design project were used to guide the detailed engineering required to insure the successful
implementation and startup of the new assembly system.
The new design would impact positively the operating performance in the following manner:
- A 35% reduction in the quantity of operators.
- A 40% increase in the utilization of productive resources.
- A 27% reduction of space requirements.
- Greater operator flexibility after an intense training.
- Response time improved in 16%.
- A 28% reduction in unit production cost.
The implementation plan was initiated at the end of year 2004 with the installation of two assembly cells in the
former distribution centre (see Figure No. 4). The cell pilot testing was also carried out at the end of the same year (see
Figure No. 5).
Figure 4. Implementation of assembly cell at former distribution center.
287
Villarreal and Alanis
Figure 5. Cell pilot testing.
5. CONCLUSIONS
The project described in this report presents a methodological scheme for designing manufacturing cells using
simulation as the principal tool. The approach is a two-level one; the macro level is developed towards the definition of
the group of cells and trying to achieve a stable material flow though the whole facility. The micro level is intended for
the detailed analysis of each cell. Both levels must “communicate” with each other to incorporate the variability and
inter-dependency among them.
The methodology was applied to the re-design of the assembly system of Carrier de Mexico achieving interesting
results in several operating performance indicators. The company initiated the implementation of the new system at the
end of year 2004 finishing it during the first quarter of year 2005. This process has been applied to other family product
lines successfully in the later years.
6. REFERENCES
1. Beaulieu, A., Gharbi, A. and Ait-Kadi (1997), An Algorithm for the Cell Formation and the Machine
Selection Problems in the Design of a Cellular Manufacturing System, International Journal of Production
Research, Vol. 35.
2. Bowersox, D.J. (1972), Planning Physical Distribution Operations with Dynamic Simulation, Journal of
Marketing, Vol. 36.
3. Carson, Y. and Maria, A. (1997), Simulation Optimization: Methods and Applications, Proceedings of the
1997 Winter Simulation Conference.
4. Christy, D.P. and Nandkeolyar, U. (1986), A Simulation Investigation of the Design of Group Technology
Cells, Proceedings of the Decision Science Institute.
5. Cochran, D.S., Duda, J.W., Linck, J. and Taj, S. (1998), Simulation and Production Planning for
Manufacturing Cells, Proceedings of the 1998 Winter Simulation Conference.
6. Harrel, Ch, Ghosh, B.K. and Bowden, R.O. (2004), Simulation Using ProModel, McGraw Hill, 2nd Edition.
7. Heinz, W.K. (1998), Simulation of Large Scale Brewery Distribution System, Proceedings of the 1998 Winter
Simulation Conference.
8. Hirano, H. (1990), JIT Implementation Manual: The Complete Guide to just-in-time manufacturing, Portland
OR: Productivity Press.
9. Hung, Ch, Miltenburg, J. and Motwani, J. (2000), The Effect of Straight- and U-Shaped Lines on Quality,
IEEE Transactions on Engineering Management, Vol. 47, No. 3.
10. Hyer, N. and Wemmerlöv, U. (2002), Reorganizing the factory, Competing though cellular manufacturing,
Productivity Press.
288
Simulation of Assembly Line
11. Kottas, J.R. and Lau, H. (1973), A Cost Oriented Approach to Stochastic Line Balancing, AIIE Transactions,
Vol. 5, No. 2.
12. Lee, Y.M., Cheng, F. and Leung, Y.T. (2004), Exploring the Impact of RFID on Supply Chain Dynamics,
Proceedings of the 2004 Winter Simulation Conference.
13. Miltenburg, J. (2001), U-Shaped Production Lines: A Review of Theory and Practice, International Journal of
Production Economics, Vol. 70.
14. Monden, Y. (1983), Toyota Production System, IIE Press.
15. Ohno, T. (1998), Toyota Production System: Beyond Large Scale Production, Productivity Press.
16. Park, Y.H., Matson, J.E. and Miller, D.(1998), Simulation and Analysis of the Mercedes Benz All Activity
Vehicle (AAV) Production Facility, Proceedings of the 1998 Winter Simulation Conference.
17. Rajamani, D. and Singh, N.(1990), A Simulation Approach to the Design of an Assembly Line: A Case Study,
International Journal of Operations and Production Management, Vol. 11, No. 25, 1990.
18. Ruiz-Torres, A.J. and Mahmoodi, F., (2008), Analysis of Multi-Cell Production Systems Considering Cell
Size and Worker Flexibility, International Journal of Industrial Engineering: Theory, Applications and
Practice, Vol. 15, No. 4.
19. Sekine, K., One-piece Flow, Productivity Press.
20. Sparling, D. and Miltenburg, J. (1998), The Mixed-Model U-line Balancing Problem, International Journal of
Production Research, Vol. 36, No. 2.
21. Suraj, M. A., Biles, W., Evans, G.W. and Zahran, I.M.(1996), A Macro/Micro Modelling Approach to the
Simulation of Cellular Manufacturing Systems, Working Paper No. 96-12, Department of Industrial
Engineering, University of Louisville.
289
Villarreal and Alanis
BIOGRAPHICAL SKETCH
Bernardo Villarreal is a full professor of the Department of Engineering of the
Universidad de Monterrey. He holds a PhD and an MSc of Industrial Engineering
from SUNY at Buffalo. He has 20 years of professional experience in strategic
planning in several Mexican companies. He has thaught for 17 years courses on
industrial engineering and logistics in the Universidad de Monterrey, ITESM and
Universidad Autónoma de Nuevo León. He has made several publications in journals
such as Mathematical Programming, JOTA, JMMA, and Transportation Journal. He
is currently a member of the IIE, INFORMS, POMS, and the Council of Logistics
Management.
Roble Alanis has an Industrial Engineering degree from University of Monterrey.
She has worked at Whirlpool in the Lean Manufacturing department. Currently, she
works for John Deere where she has been in several positions in the Supply
Management, Manufacturing, and Human Resources areas, contributing in several
projects in United States, Mexico, Brazil, Argentina, Spain and China. She is the
current head of the Quality and Production System Assessment Team worldwide.
She is also a part time Professor at University of Monterrey teaching courses related
to Quality and Lean Manufacturing.
290
View publication stats
