• Research Area 2 :

    System Design and Analysis of Automated Manufacturing Systems

    The goal of research in this area is to develop methods for the analysis, synthesis, and real-time operation of factories based on a fundamental, first-principles understanding of the dynamics of automated manufacturing systems, which operate with automated material handling systems (AMHS) and advanced production data acquisition and information support systems such as the manufacturing execution system (MES). The particular focus of our research is the unexpected random events that arise in modeling such a manufacturing system. The predominant such events are machine failures, quality deterioration, stochastic demand, supply uncertainties, and inaccurate inventory estimates, and the mathematical description of the system’s response to their occurrence forms an essential part of this research [1]. The long-term goal is to develop a coherent, comprehensive theory of manufacturing systems and a set of practical computational tools to aid factory designers, operators, and related professionals such as product designers.

    I have developed a queuing network model called the decomposition method for multiple part types. The method permits evaluation of the production rate and average buffer levels of a production line that produces multiple part types. In addition, using the queuing network concept, we have also implemented a real-time scheduling algorithm for a semiconductor fabrication facility. We have received support from Micron Technology, Deloitte Consulting, and LG Electronics for our research in this area.

    Previous results and accomplishments. Using the aforementioned method and concept, we have successfully implemented real-time scheduling and planning algorithms for the following production lines.

    • Micron Technology Inc., Hierarchical Production Planning System Analysis and Short-term Production Scheduling Algorithm Design (July-October 2010): real-time production system successfully implemented in Micron’s semiconductor wafer fabrication facility.
    • Woongjin Chemical Corp., Daily Production Scheduling System with Optimal Setup Change (April-September 2013): production daily scheduler designed and implemented, with verification from the company that the system reduced the work-in-process (WIP) inventory by 25% over six months of operation. This project received recognition from the company as the Best Practice of 2012.

    Current and future research

    we have applied the concept of the decomposition method to the layout and AMHS design of automated manufacturing systems. The main benefit of this method is that it can estimate the performance of a production line in a timely manner. It is particularly adept at analytically evaluating the average WIP inventory for a production line. As the amount of WIP inventory and inventory buffer space are major considerations in manufacturing system design, the decomposition method can be used effectively in designing the layout and AMHS of such a system.

    Process designers traditionally select the processing steps first. They then consider a layout design on the basis of these steps. Once the layout design is nearly finalized, AMHS engineers begin working on the AMHS hardware and software configurations. Unfortunately, because the layout and AMHS design steps are performed sequentially, the choices made during layout design leave few options for AMHS design. As a consequence, AMHS design tends to be driven by concerns over the transportation system itself. For instance, AMHS designers often focus on minimizing the travel time from one processing stage to the next, with the equipment locations already specified. However, travel time’s influence on overall system performance may not be as great as that of other factors such as the location of equipment. Hence, to design the most effective and efficient factory, AMHS and layout design should be combined. we plan to further develop the aforementioned algorithms by expanding the decomposition method to integrate these two design steps and eventually implement a tool for layout and AMHS designers.

     

    [1] Gershwin, Stanley B., Manufacturing Systems Engineering. Englewood Cliffs, NJ: Prentice Hall, 1997.

     

     

    Last updated : 2014/03/15