Integrating Production and Predictive Maintenance Planning Model with Industry 4.0 Technologies

In the present paper, we address the challenge of dynamically integrating production and predictive maintenance planning within the framework of Industry 4.0, employing a rolling horizon approach. Production planning involves determining the optimal levels of production, inventory holding, backorder, and set-up to meet the demand for all products within a defined planning horizon. Preventive maintenance action is employed to replace the system with a new one to ensure optimal performance and reliability. Corrective maintenance is executed to return the system to an "as-bad-as-old" condition when a failure occurs. Minimizing the total costs associated with maintenance and production planning is the objective of the integrated model. Based on having access to the data collected by sensors within the Industry 4.0 framework, a deep learning method is used to predict the system's health condition. Consequently, a Long Short-Term Memory (LSTM) method is utilized to determine the most suitable maintenance action based on the system's condition. By leveraging the rolling horizon approach, the model can simultaneously replan the production and maintenance decisions in real-time by incorporating new obtained sensor data. The numerical example highlights the efficacy of the data-driven integrated model in comparison to the model-based integrated model.