基于人工神经网络的超声加工振幅控制的研究

针对超声振动加工中,系统振幅随外界条件的变化而衰减的问题,提出了利用人工神经网络分析换能器的输入电流,以保证超声加工的振幅始终工作在最大值附近。通过人工神经网络训练数据,建立换能器输入电流与超声波发生器频率调整间的关系,达到实时调节超声波发生器频率使整个系统处于谐振的作用。最后,将此方案应用在功率超声珩磨加工中,通过换能器输入电流的稳定性和加工精度证明此方案的效果。

Study on Amplitude Control of Ultrasonic Machining Based on Artificial Neural Network

Aimed at the problems of system amplitude decaying with the change of outside conditions during ultrasonic vibration machining, the method was proposed that analyzed input current of transducer by using artificial neural network, so as to ensure the amplitude of ultrasonic machining varied nearby a range of maximum. Through using artificial neural network training data, the relationship between input current of transducer and adjustment frequency of the ultrasonic generator was established, and the effect was arrived that made the whole system in the resonance state by adjusting frequency of the ultrasonic generator in real time. Finally, the scheme is proved effective through applied to powered ultrasonic honing machine by the stability of input current of transducer and machining accuracy.