Abstract
In a volatile environment, the resilience of a manufacturing system is an element of high importance to avoid collapse and find a way to recover after disruptive events. In such an environment, having flexibility and adaptability are of higher long-term value than the attempt to maintain equilibrium or short-term stability at any sacrifice. Resilience is an intangible characteristic that cannot be measured directly, but its assessment requires the combination of measurable and probabilistic inputs. The aim of this work is to develop an architecture for resilience-aware monitoring and more knowledgeable management and control of manufacturing processes. The developed architecture combines the retrieval of real-time and event log data from the manufacturing shop floor and databases to allow the simulation of various manufacturing scenarios and data analytics for better decision-making. A modular real-time monitoring system creates the bases for higher predictability of internal disruptions. Therefore, four main groups of real-time and event log data were considered as input for process simulation and analytics: the status of machinery, workers and production orders; machinery and transportation system condition monitoring and in-process assessment of workpieces compliance; safety equipment usage and; items location and quantity tracing. The architecture was modeled to be applied on a cyber-physical demonstrator consisting of an assembly transfer line equipped with transport shuttles, workstations, a warehouse, a decision support system with analytics and simulation by creating an Industrial Internet of Things network. The developed resilience-aware architecture allows to optimize production and maintenance strategies execution for increased long-term resilience in cyber-physical production systems.