MCSARCH: An architecture for the development of manufacturing control systems

Abstract

Various classes of computerised manufacturing control system (MCS) have emerged during recent decades. Potentially they have the capability to help co-ordinate and control product realising operations, by imposing the use of suitable business rules and organisational constraints. Therefore their use can improve the efficiency and effectiveness with which a company deploys its valuable resources. To date however that potential has been far from being realised. This paper explains that this situation can, to a significant degree, be attributed to unsatisfactory characteristic properties of contemporary classes of MCS, where these properties are inherited directly from the way in which such systems are designed and implemented. Indeed contemporary approaches to MCS design and construction are geared to producing essentially one-off systems. As a result they make limited use of modern system design methods and concepts and do not promote the reuse of standards and component building blocks. Invariably custom MCS design and build leads to very high cost systems and long installation lead-times [M.I. Barber, S. Jennis, R.H. Weston, J.D. Gascoigne, A Study of Business Process Re-engineering Practice in the UK, MSI Publication, Loughborough University, UK, 1996]. Furthermore resultant systems are often characterised by their lack of robustness and lack of flexibility. Therefore, seldom will contemporary forms of MCS readily enable change. This is a major impediment to the successful application of MCS in manufacturing environments which are dynamic in nature. It follows that contemporary forms of MCS have particularly limited application on the production shop floor in small and medium enterprises (SMEs). Even in static environments it is difficult to justify a major investment in any type of IT system. Hence, very seldom will it be possible to justify the use of inflexible systems in environments characterised by near constant change – this being common place in SMEs. An architecture for the development of manufacturing control systems, called MCSARCH, has been developed to reduce the effort needed to produce shop floor control systems. The approach is a natural extension of the use of object-oriented (OO) software development techniques. It unifies the use of several manufacturing and communication standards which individually were conceived to address “a particular aspect of the way in which systems should be designed and built”. These standards are STEP (ISO 10303), MMS (ISO 9506), and CORBA [OMG, The Common Object Request Broker: Architecture and Specification-revision 2.1, OMG Publication, August 1997]. This has required solutions to complex questions such as how MMS and CORBA “concepts” and “mechanisms” should be integrated in a flexible but effective way. Means of achieving such an integration are “embedded” into the MCSARCH architecture and approach. To implement the MCSARCH and approach a computer aided software engineering (CASE) tool has been specified. The particular implementation of the specification produced in this study is referred to as MCSTOOLS and has been developed in a “proof of concept” form in order to validate concepts embedded within the MCSARCH architecture, as well as to define encapsulation proposals for the different elements which comprise that architecture. The resultant systems engineering environment, i.e. MCSTOOLS, allows semi-automatic MCS development. The overall environment comprises means of supporting use of the unified modelling language (UML) and of C++ Developer Environments; using C++ language extensions for multithreading, persistence and distributed computing. This tool has also proved its use as a research tool to study various aspects of the MCS domain.