论文部分内容阅读
实验研究了基于ＡＮＮ离线学习在线检测的故障诊断方式的某实物液压舵机（伺服系统）的鲁棒故障检测。通过提出的自适应阈值来提高故障检测系统的鲁棒性和敏感性。实验表明，该方法能快速和鲁棒性地检测出舵机小幅值突变故障，早期、小幅值缓变故障。
The robust fault detection of some kind of hydraulic servo (servo system) based on fault diagnosis method of ANN off-line learning online test was experimentally studied. Through the proposed adaptive threshold to improve the fault detection system robustness and sensitivity. Experiments show that this method can detect the small amplitude sudden change fault of the steering gear quickly and robustly, and the early and small amplitude slowly change fault.