DETERMINING NETWORK CONTROL SOLUTION DESIGNS FOR MANUFACTURING SYSTEMS

摘要

In addressing the need for manufacturing systems that are more reconfigurable and flexible, there is an ever-increasing focus on utilizing networks at all levels to support control, diagnostics and safety functionality. Issues of performance and cost must be addressed. Further, making an appropriate network choice often involves examining the application environment rather than just universally applying a set of heuristics. The result is that network decisions, such as protocol, medium, and partitioning, are often made with incomplete or inappropriately prioritized information, leading to sub-optimal solutions in terms of both cost and performance. Three perspectives of information are needed for effective decision-making in network control system (NCS) deployment, namely theoretical, experimental and analytical. Theoretical involves understanding network operation, identifying metrics for NCS performance in the application environment, and mathematically evaluating the solution candidates with respect to these metrics. Experimental requires collecting and evaluating experimental data on particular aspects of network operation that are important to the application environment. Analytical involves a weighted cost analysis of the tradeoffs involved in network decision making and incorporates the results of theoretical and experimental analysis. In combining these three information perspectives, a methodology for making NCS design decisions is provided that allows the user to appropriately incorporate the application environment. An application of this methodology to a reconfigurable factory testbed demonstrates its use and effectiveness.