Data intensive complexity modeling using Multi Agent System in industrial processes. From fault detection and diagnosis to energy efficiency

The emergence of information and artificial intelligence technologies together with data intensive techniques for big data analysis brought in the last decade to a radical change in different spheres of application. The research evolves towards the fourth paradigm of scientific research, making the data intensive technique the tools to address complexity in science. Simultaneously the system engineering develops from well ruled predefined design to a complex systems engineering where nothing is defined and the systems have to be able to adapt, change and renew. Such revolution produced a particular impact also in industrial processes, where, ever more companies are facing challenges in dealing with big data issues of rapid decision-making for improved productivity. This process is leading the industry to a transformation towards 4th generation Industrial Revolution (Industry 4.0); characterized by a variety of new technologies that are fusing the physical, digital and biological worlds, influencing all disciplines, economies and industries. In this context, the present work attempts to apply complex systems analysis tools together with data intensive techniques within industrial systems, with the aim to obtain process and energy performance indicators of the whole industrial system or of its components. Such indicators are able to take into account not only the monitored information but also the emergent features surfacing by the interaction of the parts of the systems. The proposed approach allows to analyse the system by a different point of view, no more as series of different isolated and well-designed components, but as a set of interweaved agents.