An Edge-Based Framework for Real-Time Prognosis and Opportunistic Maintenance in Leased Manufacturing System

Nowadays, with the wide adoption of Industry 4.0 and service-oriented manufacturing, leased manufacturing systems are facing new informatics challenges in the efficiency, scalability, and profitability of operation and maintenance. On the one hand, the centralized cloud needs to serve multiple high-value leased machines located remotely far from the lessor in real time. On the other hand, massive field data will be generated locally through proliferated Internet of Things (IoT) devices. The traditional cloud frameworks are limited by the computational capability and the scheduling complexity. To handle the above problems, we propose an edge-based framework that integrates real-time prognosis with opportunistic maintenance for the leased manufacturing system. Firstly, we extend cloud computing capabilities with edge computing resources for supporting condition prognosis and maintenance scheduling. Then, the edge-level prognosis will update time-to-failure (TTF) distributions and evaluate dynamic predictive maintenance intervals (DPMIs) for each leased machine in real time. By identifying maintenance opportunities from edge sides, the cloud-level PM scheduling will be dynamically executed for the whole leased system. A real case study of a leased engine crankshaft system deployed the proposed edge-based framework is conducted. Numerical results significantly demonstrate the computing efficiency, communication improvement, and economic advantage of our edge-based framework compared with the cloud-based solutions for the leased manufacturing system. Note to Practitioners—This work is motivated by improving the efficiency, scalability, and profitability of prognosis&health management (PHM) in the leased manufacturing system through edge computing. Currently, most existing PHM frameworks are based on the cloud to monitor the health condition and schedule the PM for a single machine which cannot be directly applied to multi-unit leased manufacturing systems. Meanwhile, due to the limited cloud capability and proliferated IoT devices, the timely prognosis updating and optimal maintenance scheduling are often constrained under the product-service paradigm. It is essential to coordinate the field-level resources, edge-level prognosis, and cloud-level PM in a hierarchical framework for the multi-unit leased manufacturing system to promote computational performance and maximize leasing profits.To fill this research gap, we develop an edge-based framework coupling with the real-time prognosis based leasing profit optimization (RP-LPO) strategy to achieve PHM for the whole system, as well as the optimal PM scheduling. Compared with the cloud-based frameworks and conventional maintenance policy, our edge-based framework with RP-LPO strategy greatly reduces the data transmission, the space storage, the synchronization time and maximizes leasing profits.