ニューラルネットワークによる切削工具摩耗状態の推定

概要

論文の詳細を見る
Neural network is applied to estimate the durability of a high-speed cutting tool for stainless steel. The lifetime of the tool is categorized into three conditions : 'good cutting condition', 'forced cutting condition' and 'end of tool lifetime'. The vibration data for classifying into the three cutting conditions are obtained from an acceleration sensor attached to the tool. These vibration data are transformed into power spectrum data by FFT. The neural network used in this study is constructed using three layers of processing units. Three typical patterns of the power spectrum data according to the three cutting conditions are used to train the neural network. The performance of the neural network in classifying into the three cutting conditions is tested by applying the power spectrum data for every sample. The experimental results show that the neural network is capable of estimating the lifetime condition of the cutting tool
一般社団法人日本機械学会の論文
1992-06-25