Low Cost Hierarchical Control of an Assembly Flow Line

Abstract

Much work has been carried out within the last few years to link together "islands of automation" on a factory floor using local area networks. This has led to the development of standards in communication systems, such as the Manufacturing Automation Protocol sponsored by the General Motors Corporation. At present a standard, low-cost approach to implementing manufacturing systems, ensuring compatability with the higher-cost MAP-based systems, is not available. If such a standard were available then small firms could install distributed automation control systems incrementally. This paper describes the application of multi-level, hierarchical control to a simple assembly flow line for the consumer electronics industry. There are four basic levels in the system, the lowest being an assembly operation carried out by human or robotic operators. A number of separate assembly operations, each controlled by work station controllers, are formed into a flow line with a number of flow lines forming the complete manufacturing facility controlled by a Computer Aided Manufacture (CAM) host. It is only at the level of the flow line controller and above that the MAP protocols can be applied cost-effectively in a small manufacturing environment. Adoption of the hierarchical approach throughout the system ensures that it is flexible, reliable, easily maintained and can be installed incrementally. The standardisation of low-cost local area networks (LANs) to interconnect the programmable logic controllers (PLCs) is crucial to this task because the lower level communication can then be implemented at much reduced cost as the performance required is minimised. An analysis of the requirements and implementations of LANs for this purpose is provided. To ensure correct operation of the workstation controllers and the associated communication to their flow line controller, a complete and novel Petri-net based logic controller design (PNLC) has been tested. This allows emulation and simulation of algorithms in the stations, to prove both the algorithms and their intercommunication are correct, prior to direct transfer to running on the factory floor.