LEARNING TO MONITOR A MACHINE-TOOL

This paper deals with the issue of automatic learning and recognition of various conditions of a machine tool. The ultimate goal of the rese arch discussed in this paper is to develop a comparehensive monitor an d control (M&C) system that can substitute for the expert machinist an d perform certain critical in-process tasks to assure quality producti on. The M&C system must reliably recognize and respond to qualitativel y different behaviours of the machine tool, learn new behaviors, respo nd faster than its human counterpart to quality threatening circumstan ces, and interface with an existing controller. The research considers a series of face-milling anomalies that were subsequently simulated a nd used as a first step towards establishing the feasibility of employ ing machine learning as an integral component of the intelligent contr oller. We address the question of feasibility in two steps. First, it is important to know if the process models (dull tool, broken tool, et c.) can be learned (model learning). And second, if the models are lea rned, can an algorithm reliably select an appropriate model (distingui sh between dull and broken tools) based on input from the model learne r and from the sensors (model selection). The results of the simulatio n-based tests demonstrate that the milling-process anomalies can be te amed, and the appropriate model can be reliably selected. Such a model can be subsequently utilized to make compensating in-process machine- tool adjustments. In addition, we observed that the learning curve nee d not approach the 100% level to be functional.