基于神经网络的余度舵机系统故障诊断

Neural Network-based Fault Diagnosis for Redundancy Actuator Systems

对复杂的余度舵机系统进行故障诊断对提高飞行控制系统的可靠性和容错能力具有重要意义。文章提出1种变结构径向基神经网络学习算法,使用梯度下降法对网络参数进行优化,利用径向基神经网络对某型飞机余度舵机系统进行故障模式识别。仿真结果表明设计的神经网络具有较强的故障模式识别能力。

Failure diagnosis is the key technique of redundancy electric actuator systems to improve reliability. A scheme for neural network-based fault diagnosis is proposed. It is based on the Radio Basis Function network（RBF）which is designed by Gradient Descent Learning algorithm of parameter optimization. Application of dual-redundant electric actuator systems simulation is introduced. The simulation result show neural network failure diagnosis is achieved.