Machine tool condition monitoring using sweeping filter techniques

Abstract

This paper presents an approach to milling process monitoring utilizing sweeping filter frequency analysis techniques. The paper demonstrates how critical information retrieved via the new technique may be used to ascertain the tool condition. This information can be used to suggest remedial actions, to increase the quality of work, and to reduce overall machining costs. The deployed system does not require any additional sensors to be attached to the machine. It consists of three front-end nodes based upon embedded controllers which are linked on a controller area network bus. These nodes acquire and analyse data dynamically from the spindle load, spindle speed, and cutting axis current signals as are typically available on milling machines. These nodes are controlled by a bridge node, which is interfaced to the internet for remote information display. The front-end nodes contain the implementation of the sweeping filter technique. Potential faults from the signature of the acquired signal are identified at this level and can be further verified by the bridge node. The overall system allows dynamic decision making based on a health index measure of the process. The system has the flexibility for any number of required front-end nodes to be deployed depending upon the monitoring requirements.