Condition-based Real-time Production Control for Smart Manufacturing Systems

In this paper, we present condition-based real-time production control for smart manufacturing which is aimed at improving system performance by automatically assessing a production system's condition and dynamically configuring the processing routes for smart products and parts. A machine's degradation condition is defined in discrete states and modeled as a Markov chain. By taking into account machines' degradation and buffers' occupancy, an optimization problem is formulated to maximize the production rate using Markov Decision Processes. The effectiveness of the method has been demonstrated on a three-machine flexible production system. Traditionally, condition monitoring and production control are designed, developed, installed and managed separately by different domain experts. Hence, in this paper, the implementation challenges of condition-based production control are also discussed, with the existing and missing enabling standards identified and analyzed.