An evolvable model of machine tool behavior applied to energy usage prediction

Abstract

There is a growing requirement for models representing machine tool behavior to learn from operational data and adapt themselves in the use stage to various views defined by production managers, machining process designers, and operators. This study formulates these models as evolvable mapping from operational data to process definitions in views pertaining to multiple stakeholders at various temporal scales. A case study presents such a model for the behavior of a five-axis machining center defined as per multiple stakeholder views that predicts energy usage accordingly, while evaluating prediction accuracy of the model implemented using a supervised machine learning algorithm.