S1310204 移動分散値とスライディング離散フーリエ変換を応用したセンサレスびびり振動検知システムの開発([S131-02]先進工作機械技術(2),生産加工・工作機械部門)

抄録

Chatter vibration deteriorates the machining accuracy and shortens the tool life. Although several prediction methods for chatter vibration have been proposed by creating a cutting model, their accuracy is not sufficiently high because of the parameter identification error. Thus, some studies presented in-process monitoring methods by using additional sensors such as acceleration sensors. However, using additional sensors leads to high cost and frequent maintenance. Previous research experimentally shows that chatter vibration in milling can be detected by using disturbance observer which estimates a fluctuation in cutting torque only from the servo information. For further demand, because chatter vibration is classified into self-excited vibration and forced vibration, an assorted detection method is strongly required. In this paper, we propose a novel frequency analysis method for real-time usage by combining moving variance and sliding discrete Fourier transform algorithms to detect self-excited and forced vibration separately. The proposed analysis method obtains the power spectrum density of each vibration from the estimated disturbance torque with small computational load. The experimental results clearly show the proposed method separately detects the chatter vibration by type without additional sensors.