基于神经网络的自适应相位纠偏器在电液伺服系统中的应用

ANN-based adaptive phase corrector applied in electro-hydraulic servo system

针对电液伺服系统正弦响应中存在相位滞后的现象,基于Adaline神经网络,使用LMS自适应滤波算法设计了自适应相位纠偏器(APC)。当响应信号对输入信号有相位滞后时,利用LMS算法对神经网络的权值进行调整,经加权后的输入信号作用于控制系统,从而消除相位滞后。该方法并不需要对控制对象进行辨识,能保证控制系统的实时性。仿真和试验结果表明,自适应相位纠偏器能有效消除相位滞后,快速跟踪输入信号。

Aiming at the phase delay phenomenon occurs in electro-hydraulic servo system corresponding to a sinusoidal input, based on Adaline neural network, adaptive phase corrector was designed for eliminating phase delay using LMS adaptive filtering algorithm. When there exists phase delay between the input and its corresponding output, LMS algorithm was used to adjust the weights of the network, the weighted input signal was added to the control system such that the output phase delay was cancelled. This method needs not to estimate the system model and has good real-time performance. The simulation and experiment results show that the adaptive phase corrector can efficiently eliminate the phase delay and quickly trace the input signal.