Reference Architecture for a Collaborative Predictive Platform for Smart Maintenance in Manufacturing

Abstract

Maintenance is a key factor to ensure the production efficiency, since the occurrence of unexpected failures leads to a degradation of the system performance, causing the loss of productivity and business opportunities, which are crucial roles to achieve competitiveness. The article aims to propose a reference architecture which will improve the way maintenance is considered in the current manufacturing world, by enabling an overall increase of production rates, while increasing the operational equipment effectiveness and decreasing the impact of maintenance needs. This objective would be accomplished by establishing an IoT infrastructure for the collection of the huge amount of available shop floor data, which can be analyzed, considering data analytics algorithms, predictive maintenance models and forecasting techniques, to perform the machine/system health assessment and prediction of maintenance needs, e.g. by detecting earlier the occurrence of possible failures and consequently the need to implement maintenance interventions. The scheduling of predictive maintenance needs will be integrated with the existing maintenance planning tools, and especially synchronized with the production planning tools to achieve a nondisruptive maintenance impact in the production system. A cloud-based collaborative maintenance services platform allows the secure collection, aggregation and analysis of a large amount of shared data from numerous manufacturers that use the same or similar machinery, and acts as an open market where companies can contract specialized maintenance services. This reference architecture aims to provide replicable architecture to be broadly applicable in a variety of industries, capable to improve the production efficiency through a real-time health monitoring and early detection of failures and outages, to speed up the maintenance delivery, and consequently mitigate their impact.