A Simulation Environment to Support Prognostics and Health Management in Smart Manufacturing Systems

Maintenance activities are a significant part of an advanced manufacturer's budget and could potentially represent an organization's source of vulnerability in sustaining a stable supply chain. It is crucial for a manufacturer to understand the implications of maintenance on their organization's ability to produce quality products in an efficient time-frame at an acceptable cost. Furthermore, it can be expensive, time consuming, and create a risk to finished product supply to evaluate results from altering maintenance activities and policies during a production cycle. This is especially true when attempting to implement prognostics and health management (PHM) methods on existing smart manufacturing systems (SMS). This thesis presents research studying the trade-offs of various maintenance policies, including PHM methods, using a simulation with the purpose of providing decision support. The modeling approach involves building the structure for a simulation in which maintenance policies, probabilistic models of failure rates, processing speeds, and costs are inputs into the simulation environment. This generality allows for quickly altering a manufacturing model, and the ability to consider novel policies where control of machine parameters is needed. The outputs of the simulation evaluate policies in terms of cost per unit produced and overall equipment effectiveness (OEE), a widely used industry metric. Data from industrial partners was used in the validation of the simulation environment. This work is part of a larger attempt to equip plant managers and business leaders with the methods for understanding which PHM strategy is best suited for their manufacturing system.