Linxia Liao from System Science Laboratories discussed the future of self aware machines in manufacturing at the 2015 MTConnect meeting. The goal of pursuing self-aware machines is attempting to move from a preventative paradigm, such as checking performance and replacing parts on a set schedule, to a predictive paradigm where maintenance is auto scheduled for exactly when it is needed. Operational data accumulated through MTConnect can be used to see anomalies, detect outliers, and identify false positives. Using that human feedback, the machine will become smarter and more self aware— positively impacting production time, cost, and quality on the shop floor by reducing unplanned downtimes, adapting for work piece variability, and enabling specification of fault-tolerant process plans.
Furthermore, the predictive analytics from data gleaned using MTConnect can generate improved health condition estimation and predictions for the equipment and better shop floor planning recommendations. While there are infinite opportunities in combining data with manufacturing, there are also challenges including the need to design an effective hybrid approach for prognostics with humans in the loop and the need to continue minimizing misleading false positives, and adjusting the corresponding recommendations.
For more on the self aware machine, check out one of Linxia Liao’s articles for the PARC blog.