摘要
故障检测和诊断技术对提高系统可靠性具有重要意义,针对飞控系统中常见的传感器故障,提出了基于神经网络观测器的故障诊断方法;通过构造神经网络模型代替解析系统建模,利用神经网络的学习能力在线检测传感器故障,最后,应用BP神经网络算法对故障进行仿真;仿真结果表明,神经网络观测器方法对单一传感器故障及多个传感器故障均能够准确识别,并对故障的定位也有不错的效果。
The sensor fault detection and diagnosis technology are important to improve the reliability of system. Aimed at the common sensor fault for flight control system, the approach of fault diagnosis based on neural network observer is discussed. By constructing neural network model instead of modeling for analytic system, the learning ability of neural network is used to detect sensor fault on--line, Finally, BP Neural Network algorithm is applied to simulation. The simulation results indicate that the approach can accurately identify one or more sensor faults and exactly locate the fault.
出处
《计算机测量与控制》
CSCD
2008年第5期613-615,共3页
Computer Measurement &Control