機械音スペクトルの出現頻度を用いた加工工程監視システムの開発(第1報,16進数化スペクトルパターン(HSP)表示による工具状態監視システムの構築)(機械力学,計測,自動制御)

抄録

Machining sounds has been told to be effective for monitoring of cutting tool conditions, but it has been difficult to avoid misunderstanding of the conditions in previous systems. There are two main causes. First, the peculiar sound to a condition dose not emanate continuously and an unexpected sound emanates occasionally. Secondly, in case of tool wear, the small difference between normal tool and tool wear makes the recognition of the wear hard. We propose the HSP (Hexadecimal Spectrum Pattern) as the definite expression of the spectrum patterns in this paper. And we also propose the new monitoring system by using the frequency of sound, which is applied by neural network, because we consider that the emanating time of the peculiar sound is as important as the spectrum pattern of it to recognize the condition exactly. In this system, cutting tool conditions, like normal, wear and fracturing, can be monitored without misunderstanding under a few cutting conditions.