Distributed industrial automation systems: Automated communication patterns recognition

Abstract

Modern manufacturing systems have to continuously adapt to variable custom environment, satisfy to discerning expectations of customers and maintain in the increasing competition. Nowadays, manufacturing plants are very complex, in many cases distributed, composed of many subsystems with different functions that are exchanging a vast amount of data, i.e. events. In the case of observable malfunction, the amount of events grows substantially and during the system resurgence can lead to information explosion. We describe a few methods to help system's users to mitigate this information explosion and introduce an agent-based expert system proposal that will be able to provide a solution to a new arisen problem that is possible to be revealed from automated recognition of communication patterns. Such an expert system should also provide users with user-friendly manufacturing plant processes overview and management; allow searching for event patterns and simplifying the event processing.