针对电液伺服作动器早期故障信号的微弱性和故障特征分布的未知性,提出了一种基于Parzen窗估计的故障检测和健康评估算法；采用多重判别分析提取作动器故障特征向量,利用窗函数法估计故障特征概率密度函数,通过监测作动器故障特征概率密度分布相对于正常状态的偏离,将不易检测的故障信号转化为容易观测到的距离信息评估作动器状态的健康状况；仿真结果证明了该方法的可行性和有效性.%Aim at that early fault signal of the electric hydraulic servo actuator is weak and the unknown of fault feature distribution, a a-rithmetic based on parzen window assessment is put forward for fault detection and health assessment. Multiple discriminant analysis is used to extract the actuator fault feature vector and parzen window estimation method is used to estimate the parameter of fault feature vector probability density function, the probability density function distributing departure between current state and normal state of electric hydraulic servo actuator is detected, the fault signal which is not easy to detect is transformed to deviation information which is easy to observe, the health state of the actuator is evaluated. The simulation results prove the validity and feasibility of the method.
展开▼
