基于滤波方法的航空发动机故障诊断研究

Research of Aero-Engine Fault Diagnosis Based on Filtering Methods

在某型航空涡轮发动机非线性模型的基础上,建立了简化非线性模型。对突变型和渐变型两类故障,分别采用EKF、UKF、PF和EKPF四种方法进行诊断。结果表明:四种方法对此类航空发动机故障均能进行准确诊断,整体效果从好到坏依次是UKF、EKPF、EKF和PF。对于高斯型、弱非线性系统,粒子滤波类方法相对于卡尔曼滤波类方法无明显优势。

A simplified nonlinear model was established based on nonlinear model of a certain type of aeroengine. Two kinds of the fault （The mutation type one and gradual type one） were diagnosed respectively by EKF, UKF, PF and EKPF. The result shows that the four methods can make accurate diagnosis for such aero- engine faults, and the whole effect ranks UKF, EKPF, EKF and PF from good to bad in sequence. For weak nonlinear Gaussian system, particle filter method has no obvious advantage over the kalman filter method.