“Broadcaster”: An architectural description of a prototype supporting real-time remote data propagation in distributed manufacturing

Globalisation of manufacturing activities tend to geographically distribute manufacturing entities, resulting into each entity adopting its own mechanism, for aggregating and analysing real-time shop floor machines' information. The enterprise systems normally employ sophisticated and computationally expensive techniques to access this data, even if they operate remotely having limited network connectivity and system legacies. There is a need to propagate machine information in soft real-time basis to these entities regardless of their geographic locations and / or mechanisms. Authors are presenting an architectural description of a prototype system called the ldquoBroadcasterrdquo which efficiently distributes manufacturing machine information to a number of remotely located global engineering partners. This prototype addresses the emergent system issues like maintainability, reliability, integrity, robustness, flexibility and performance using a heterogeneous composition of ldquoBlackboardrdquo repository model with an event-driven invocation technique, implemented using interface-based strategy. The design and implementation assumes the control environment description to be engineered using the component-based system paradigm. Presently, the prototype is evaluated on a demonstration test rig provided by the Ford Motor Company, which is implemented using a fully Web services distributed control device called FTB, designed by the Schneider Electric Company. Based on the evaluation from the implementation stage, authors have justified and concluded the paper highlighting the key benefits of this approach, and described any future research that is to be carried out.

I. INTRODUCTION

In this industrial era, operation of manufacturing activities has evolved to follow the concept of a distributed paradigm [1]. One of the key factors initiating such a change is globalisation of manufacturing process. In order to provide the required flexibility and competitiveness, these manufacturing activities tend to be geographically distributed spanning various cities, countries and even continents. Every manufacturing entity possesses their individual mechanism for aggregating and analysing its machine information. Furthermore, the volume of interesting data emerging from the shop-floor machines is increasing and the management of it is becoming more complex. The manufacturing community often use sophisticated and computationally expensive techniques to access and analyse this data. This leads to the concept of satis- fying various requirements identified by this distributive business nature [2]. In addition, a number of challenges and limitations in terms of availability of real-time machine information exist [3]. The scenario becomes even complicated when the enterprise systems operate in a remote computing environment with limited features such as network connectivity and / or application legacy.