Study on inprocess sensing of tool failure utilizing microphones

研究了车削过程的监控系统，并对不同切削条件下的车削过程进行了研究。切割声音由麦克风检测，并通过FFT分析仪转换为频谱。这关注的是切削声频谱上的峰值随着刀具磨损的增加而发生的变化。实验结果表明，当齿面磨损增大时，峰值增大，峰值频率略有减小。利用切削声检测刀具磨损是一种非接触式测量方法，是一种非常有效的方法，但同时也存在问题。在实际环境中有许多声源。周围噪声干扰了监测系统，降低了刀具磨损检测的准确性。因此，对于一个实用的利用切割声的监测系统来说，具有声源位置的指定能力是必要的，也是必不可少的。该研究还尝试赋予声传感器尖锐的指向性，以便检测来自指定声源的唯一声音。在产生指向性的方法上，采用了麦克风阵列和抛物面反射器两种方法。实验研究了每种方法是否能够确定多个声源的位置并与其他声源区分开来。在麦克风阵列实验中，用两台车床测试了两种切削过程。此时，通过改变麦克风阵列的方向来识别车床发出的声音。实验结果表明，该阵列系统能够保持阵列对非点源非预期声音的灵敏度，并能从多个声源中区分出预期声音。本文还研究了另一种利用抛物面反射镜的方法，得到了几乎相同的结果。

The study deals with the monitoring system for turning process, and investigates under various cutting conditions. Cutting sound is detected by microphones, and is transformed to frequency spectrum by a FFT analyzer. This pays attention to the changes of peak on frequency spectrum of cutting sound as tool wear increases. Experimental results show that when flank wear increases, the peak value increases and the peak frequency reduces slightly.The method utilizing cutting sound is very effective to detect tool wear as it is non-contact measurement, but at the same time it has a problem. There are many sound sources in practical environment. Surrounding sound interrupts the monitoring system and makes accuracy of detection of tool wear worse. Therefore, it is necessary and indispensable for a practical monitoring system utilizing cutting sound to have an ability to specify the sound source locations.The study also makes an attempt to give the acoustic sensor sharp directivity in order to detect the only sound from specified sound source. As the method to make directivity, two methods which use a microphone-array or a parabolic refledctor are adopted. Each method is experimentally investigated whether it can specify location of several sound sources and distinguish from other sound sources. In the experiment with the microphone-array, two cutting processes are examined with two lathes.At that time, sounds generated from the lathes are identified by varying direction of the microphone-array. The experiment shows that the array system can hold down sensitivity for unexpected sound from other source than a pointed sound source by the array, and can distinguish expected sound from several sound sources. Another method by the parabolic reflector is examined as well, and almost the same results are obtained.