WSC 2005

WSC 2005 Final Abstracts

Manufacturing Applications Track

Monday 10:30:00 AM 12:00:00 PM
Assembly Systems

Chair: Edward Williams (Production Modeling Corporation)

Analyzing Printed Circuit Board Assembly Lines Using a Pcb Asembly Template
Pradip D. Jadhav and Jeffrey S. Smith (Auburn University)

PCB assembly lines are characterized by asynchronous transfer of parts resulting from variability in terms of random processing, failure, repair, exhaust, and replenishment times. The throughput for such manufacturing systems depends upon simultaneous availability of resources; namely, machines, material, and operators. Further, random break-downs along with the capacitated buffers cause blocking and starving which affects the throughput. Analytical models for such system require simplifying assumptions, hence simulation modeling is the popular choice. For simulation models to be successful, capturing the impact of operator interference is important. This paper describes a methodology developed for simulation modeling and analysis of PCB manufacturing lines, capturing the complex interactions between its components. A custom-developed PCB assembly template is used for modeling purposes. Though the analysis described in this paper pertains to PCB assembly lines, it is equally applicable to general class of serial production lines with capacitated buffers.

Operator Paced Assembly Line Simulation
Marvin S. Seppanen (Productive Systems)

This paper describes an Arena based operator paced assembly line simulation model. Besides the normal flow of assembly units between work cells, the model considers both the movement of operators between cells and intermittent duties, such as equipment repair and material stocking. Simulation results are presented for a 4-cell test case.

Intelligent Modeling and Simulation of Flexible Assembly Systems
SK Ahad Ali (University of Toledo), Hamid Seifoddini (University of WIsconsin Milwaukee) and Hong Sun (University of Toledo)

A combination of product mix and production volume is analyzed using a reconfigurable simulation model aiming to improve the performance and optimal designing requirements. The performance under different production scenarios is developed to find the optimal combination of product mix to meet future customer demands. This research provides a re-configurable assembly system modeling by adding flexibility and evaluates alternative designs. The best satisfaction of the production requirements under dynamic production is validated with real application.

Monday 1:30:00 PM 3:00:00 PM
High Volume Material Flow

Chair: Guodong Shao (NIST)

Advanced Decision Logic in Simulation of Material Flow Processing Networks
Douglas A. Bodner, Ke Wang, and Sheng Xu (Georgia Institute of Technology)

Material flow processing networks are ubiquitous in modern society. Such networks embody uncertainty, advanced decision logic, and increased performance expectations. As such, they have been subject to analysis by a variety of methods, with one of the most prominent being discrete-event simulation. Simulation enables detailed flow modeling, incorporates uncertainty and allows experiment-based performance assessment. Traditional approaches to simulation are not well-suited to modeling advanced decision logic, though. This paper explores the issue of representing advanced decision logic and presents a reference model for material flow processing networks to support such representations. Implementation issues are discussed, as well.

High Performance Simulation in Quasi-Continuous Manufacturing Plants
Lan Chen and Michael Pidd (Lancaster University)

Though the quality of discrete simulation software packages is high, most are aimed at systems in which discrete objects change state as they move from work-station to work-station. This generic model is a good fit for some manufacturing, for example, much automotive production. However is not well suited to very high-speed quasi-continuous manufacturing as found in the food and drinks industry. LanSkim is a prototype PC-based package designed for these applications. It is simple to use and runs very fast but is insufficiently detailed for all purposes. Adding full detail would make the simulations run very slowly, hence we examine the use of parallel computation to allow increased level of detail. We describe Lan-WARPED, based on WARPED, in which is a parallel simulation models of such plant.

Fluid Based Simulation Model for High Volume DC Conveyor Systems
Ying Wang and Chen Zhou (Georgia Institute of Technology)

In this paper, we present a fluid simulation methodology applying to high volume large conveyor networks operating in a slowly changing environment, often found in large distribution centers. Traditional discrete-event cell-based approach to simulate such networks becomes computationally challenging due to large number of events resulting from high WIP level, complex network and large conveyor footprint. The fluid simulation model is built on a Petri-Net based framework. We present the model and investigate the feasibility in modeling capability in terms of input and release control logics, performance evaluation and computational savings.

Monday 3:30:00 PM 5:00:00 PM
Facilities Planning & Work Cell Design

Chair: Kiran Mhajan (Hinz Nixdorf Institute)

For Effective Facilities Planning: Layout Optimization Then Simulation, or Vice Versa?
Esra E. Aleisa and Li Lin (University at Buffalo)

It is widely accepted that simulation is an integral part of any effective facilities planning or layout study. Traditional approaches claim that layout optimization produces strategic results and therefore should precede simulation analysis, which focuses on operational issues. On the other hand, more recent studies suggest that running simulation models prior to conducting layout optimization produces more realistic layouts. In this paper, we contrast these two paradigms, with respect to the general assumptions and the types of applications that advocates from each paradigm have used to support their claim. In addition, we propose guidelines on which approach to pursue according to the layout study objectives and the characteristics of the system under consideration.

Validation of The Re-Design of a Manufacturing Work Cell Using Simulation
Carrie Grimard (Robert Bosch Corporation), Jon H. Marvel (Gettysburg College) and Charles R. Standridge (Grand Valley State University)

Simulation can be used to validate the design or redesign of any complex system before it is implemented. Validation evidence is obtained if the simulation demonstrates that the system operation corresponds to its design. This evidence includes comparing both detailed system behavior and performance measure values to those stated in the design. The application of simulation to validating the redesign of an injector assembly and calibration production area is discussed. Simulation is necessary to validate the initial estimate of cell throughput since a single worker must perform multiple operations at multiple workstations. The feasibility of the pattern of movement by this worker between stations must be demonstrated and alternative patterns assessed. Controls on the amount of work in process inventory in the cell must be validated. Modeling challenges unique to part movement using one-piece flow, work in process inventory control, and the movement of both workers and parts are discussed.

Applying Simulation to Interative Manufacturing Cell Design
Sara L. Maas (Leon Plastics) and Charles R. Standridge (Grand Valley State University)

Launching a manufacturing cell to be efficient and lean, yet profitable, is a time-consuming process and is often based on many assumptions. The utilization of simulation models to help design the cell and the logistical structure to support it can expedite and streamline the development process. Assumptions and designs can be validated to help insure effective operations as soon as possible. We have developed a generic simulation model and associated capacity analysis, schedule planning, and target inventory setting software to support the computer based assessment of the operation of cells typical to the plastic manufacturing industry before capital investments are finalized. Model input describes a particular cell, the products it produces, and customer demand for these products. Results show the customer service level, product inventory levels, equipment utilization, and the daily production schedule. Spreadsheet software supports data entry and report examination. All software is integrated in a single simulation environment.

Tuesday 8:30:00 AM 10:00:00 AM
Production Analysis & Constraints

Chair: Durk-Jouke van der Zee (Universtiy of Groningen)

Applying Discrete Event Simulation and an Automated Bottleneck Analysis as an Aid to Detect Running Production Constraints
Patrick Faget and Ulf Eriksson (Volvo Car Corporation) and Frank Herrmann (Chalmers University of Technology)

Discrete event simulation is an important decision support tool to evaluate changes in manufacturing, distribution or process facilities. The challenge arises when it comes to the integration of simulation as an effective tool to detect manufacturing constraints and to suggest improvement alternatives. This paper describes the application of a method for detecting bottlenecks in discrete event models developed by Toyota Motor Company. The objective in this case is to automate the bottleneck analysis facilitating the understanding and adoption of simulation by decision makers without knowledge of simulation. The main results of this paper are the validation of the bottleneck detection method and its integration with MS Excel spreadsheets. Moreover system improvement alternatives are presented by the use of design of experiments.

Simulation Modeling and Analysis of a New Mixed Model Production Line
Servet Hasgül and Alptuğ Selçuk Büyüksünetçi (Osmangazi University)

Mixed model production lines are often used in manufacturing systems. In production lines, different product types are simultaneously manufactured by processing small batches. This paper describes a recently completed project involving the development of simulation models for a mixed model production line in a refrigerator company. Decision maker wants to determine the bottlenecks before changing the traditional line to a mixed model production line. Due to the enlarged number of models, the design of an assembly system becomes more complicated. Performance evaluation is an important phase in the design of assembly lines in a mixed model production environment. Simulation models helped us to identify production line bottlenecks and evaluate some number of suggested solutions.

A Kanban Module for Simulating Pull Production in ARENA
Mark A. Treadwell and Jeffrey W. Herrmann (University of Maryland)

In the short timeline of rapid improvement events (kaizen events), it is difficult to use the full power of simulation because of the time required to construct models, particularly if the system uses pull production control methods such as kanbans. This paper describes multiple ways to model pull production control and compares them on measures related to model construction and execution. A kanban workstation module significantly reduces the time required to develop a pull production model, which makes simulation more useful as a decision-making tool in rapid improvement events.

Tuesday 10:30:00 AM 12:00:00 PM
Production Policy & Cost Control

Chair: Onur Ülgen (Production Modeling Corporation)

Finding the Optimal Production Control Policy Using the Production Control Framework
Sean M. Gahagan (Northrop Grumman Corporation) and Jeffrey W. Herrmann (University of Maryland)

The production control policy affects the performance of a manufacturing system. Evaluating a production control policy usually requires simulation modeling due to the complex interactions that occur. This paper introduces a technique that optimizes production control of single product flow shops under hybrid production control by using the Production Control Framework. This simulation modeling template is designed to explore the production control domain. The paper demonstrates how this template can be used in conjunction with existing simulation optimization software to find an optimal production control policy. The decision variables are location of the push-pull interface and the number of kanbans at each workstation. The objectives include improving customer service and reducing work-in-process inventory.

Applying Energy Aspects on Simulation of Energy Intensive Production Systems
Petter Solding and Damir Petku (Swedish Foundry Association)

The electricity market within the European Union was liberalized in 2004, which means that Sweden, together with other countries with generally low electricity prices, will face higher electricity prices. This has triggered research with the aim to help energy-intensive companies reduce their electricity usage and total energy usage as well. This paper introduces a new concept for simulating energy-intensive production. Using specially built simulation models helps companies take more than the time aspect into consideration when planning their production. Often energy and power utilization is an important aspect to consider when planning, since the cost for energy and power usage can be high, especially if a “load subscription” is used and the usage is above the company’s subscribed level. The methodology described helps breaking down energy parameters into three groups and gives examples of how the simulation model can be built to take energy and power into consideration.

Tuesday 1:30:00 PM 3:00:00 PM
Simulation-Data Interface & Integration

Chair: Jan Oscarsson (University of Skövde)

Lot-sizing Within Capacity-Constrained Manufacturing Systems Using Time-phased Planning
Silvanus T. Enns and Pattita Suwanruji (University of Calgary)

Research on lot sizing has mostly assumed single echelon systems. Even when multiple echelon systems have been used, capacity constraints are seldom considered. However, in manufacturing capacity constraints can lead to significant queuing effects. Commonly used lot sizing policies like Lot-For-Lot (LFL) and Period Order Quantity (POQ) do not take these effects into account. This research compares these policies with a Fixed Order Quantity (FOQ) policy, within which lot sizes are based on minimizing estimated lot flowtimes at capacity-constrained machines. Simulation is used to study a small production and distribution network using time-phased planning. Results show that the FOQ policy performs better than both LFL and POQ when inventory levels and delivery performance are of concern.

Discrete-Event Simulation and Automatic Data Collection Improve Performance in a Manufacturing System
Arne Ingemansson and Jan Oscarsson (University of Skövde)

Data utilization from an automatic collection system helps simulation. DES (Discrete-Event Simulation) is a suitable tool in that respect. Improvements of a manufacturing system by reducing production disturbances in block engine line of a truck manufacturer were shown in a case study. The manufacturing line is equipped with an automatic data gathering system, collecting detailed information from each machine, such as DT (DownTime). The collected data and the simulation model were used in finding bottlenecks in the system. The improvement groups of the company suggested improvements to these machines in collaboration with DES work. An improvement of 12% was achieved in the block line during two years. The long-term improvement work was profitable. Further potential with the method is investigated in a new case study.

Data Exchange for Machine Shop Simulation
Y. Tina Lee and Yan Luo (National Institute of Standards and Technology)

Manufacturing simulation systems normally provide interfaces for data import and export using proprietary formats. Dependent data interfaces limit simulation applications in the manufacturing industry. This paper describes a mechanism for transferring data between a traditional database and XML files, which represent a machine shop data model developed at the National Institute of Standards and Technology (NIST). The data model is described by XML (eXtensible Markup Language) and UML (Unified Model-ing Language). The data transfer mechanism is based on Document Object Model (DOM), XML Path Language (XPath), and Open Database Connectivity (ODBC) database engines. A prototype implementation is also described in this paper.

Tuesday 3:30:00 PM 5:00:00 PM
Simuation and Control

Chair: Javier Faulin (Public University of Navarra)

Emulation with DSOL
Peter H.M. Jacobs and Alexander Verbraeck (Delft University of Technology) and William Rengelink (TBA Nederland)

Manufacturing control systems are extremely hard to design and test. Testing programmable Logic Controller (PLC) software in an on-line manufacturing setting can be costly, dangerous, and ineffcient. The availability of a seamless transition between the real manufacturing process and a simulated manufacturing process on the one hand, and a real PLC and a soft PLC on the other hand might help to solve these problems. Using the Java-based object oriented simulation library DSOL (Distributed Simulation Object Library), a case study was conducted for a concrete floor manufacturer to see whether these problems could be overcome. The full simulation and hardware-in-the-loop tests with DSOL, with the Modbus middleware protocol, and with real and soft PLCs went fine, and showed the added value of the distributed, service-oriented paradigm on which DSOL is based.

SAEDES++: Determining Complex System Availability Via Simulation
Javier Faulin (Public University of Navarra), Angel A. Juan and Carles Serrat (Technical University of Catalonia) and Vicente Bargueño (Universidad Nacional de Educación a Distancia)

Complex systems are everywhere among us: telecommunication networks, computers, transporting vehicles, and electrical appliances are well known examples. Designing reliable systems and determining their availability are both very important tasks for managers and engineers, since reliability and availability have a strong relationship to other concepts such as quality and safety. Furthermore, these tasks are extremely difficult, due to the fact that analytical methods can become too complicated, inefficient or even inappropriate when dealing with sophisticated systems. In this paper we present the basic ideas behind a simulation-based method, called SAEDES, which can be very useful when determining availability for a wide range of complex systems. The method is implemented in C/C++ using two different algorithms, SAEDES_A1 (component-oriented) and SAEDES_A2 (system-oriented). Two case-studies are introduced and analyzed using both algorithms, which allows us to compare the associated results.

An SDS Modeling Approach for Simulation-based Control
Sreeram Ramakrishnan and Mayur Thakur (University of Missouri-Rolla)

We initiate a study of mathematical models for specifying discrete simulation-based control systems. It is desirable to specify simulation-based control systems using a model that is intuitive, succinct, expressive, and whose state space properties are relatively easy computationally. We compare automata-based models for specifying control systems and find that all systems that are currently used (such as finite state machines, communicating hierarchical finite state machines (FSM), communicating finite state machines, and Turing machines) lack at least one of the abovementioned features. We propose using sequential dynamical systems (SDS)- a formalism for representing discrete simulations to specify simulation-based control systems. We show how to adapt the standard SDS model to specify cell-level controllers for a generic cell.