论文部分内容阅读
针对柔性机械臂结构振动控制中可能出现的压电器件故障问题,以提高系统可靠性和稳定性为主要研究目标,提出了一种集小波神经网络与取代控制技术相结合的容错控制方法.首先设计了3种粘贴不同故障压电片的机械臂结构;然后采用小波包对各种故障压电片进行特征提取,通过径向基函数网络进行特征识别;再根据故障类型,选用硬件取代控制或基于一种新型非线性滑模观测器的软件取代控制;最后通过NI CRIO平台进行的容错控制实验结果表明,传感器件故障诊断的置信度达到0.9,前两阶振动模态的抑制效果达到10 d B以上.
In order to improve the reliability and stability of the piezoelectric device, a fault tolerant control method based on wavelet neural network and substitution control technology is proposed to solve the possible piezoelectric device fault in the vibration control of flexible manipulator. Three kinds of robotic arm with different fault Piezoelectric pieces are designed. Then the wavelet packet is used to extract the features of all kinds of fault Piezoelectric pieces, and the characteristics are identified by Radial Basis Function Network. According to the type of fault, hardware is used instead of control or Based on a new type of nonlinear sliding mode observer software replacement control; Finally, the experimental results of fault-tolerant control using NI CRIO platform show that the confidence level of sensor fault diagnosis reaches 0.9 and the suppression effect of the first two modes reaches 10 d B above.