2012 Volume 20 Issue 6 Pages 55-64
The main spindle is an essential part of a machine tool, and main spindle malfunctions strongly impact productivity. In particular, unexpected random failures and malfunctions are a critical issue. Being able to predict the time when a main spindle needs to be replaced by detecting in advance that it is nearing the end of its life or is about to fail would greatly improve the efficiency of production planning by the machine tool user and maintenance service by the machine tool manufacturer. The ultimate goal of this study is the development of a system that can predict the life of a spindle. The feasibility of this goal was studied using the MT system. This first report concerns the development of a system for detecting failures efficiently. A unit space was selected using the S/N ratio of the pre-shipment electric power consumption waveforms of spindles, and decisions were made by the RT method from characteristic quantities extracted from the electric power waveforms used for calculation of the S/N ratios. As a result, a difference in distance was found between the unit space and unknown data, indicating the possibility of developing a system that can efficiently detect failures and predict spindle life.
