航空发动机在高温、高压、高转速及较大振动等恶劣的条件下工作时,其控制系统中的传感器很容易受到干扰,所以发动机测量参数中常常包含较大的噪声。另一方面,发动机的测量参数多于其独立变量的数量,即在这些测量参数中存在冗余信息。AANN(自联想神经网络) 通过对信息的压缩及解压缩过程,能够利用冗余信息抑制其测量噪声。在发动机故障诊断过程中,应用自联想神经网络对测量参数进行预处理,可以大大提高故障诊断的准 确率。 关键词:航空发动机;故障诊断;神经网络;数字滤波 Abstract : Sensors in engine control system are easily disturbed , as aeroengine operates in an environment of high temperature , high pressure ,high speed and rough vibration. Therefore engine measurements contain nois2 es. On the other hand , the number of measurements is greater than that of independent variables in the sys2 tem , which implies that there is spare information in the parameters. Auto associative neural network (AANN) is introduced to reduce the noise level contained through mapping and decoding process. It is found that the noise can be greatly filtered to result in a higher success rate of fault diagnosis of aeroengine. Key words :aeroengine ; fault diagnosis ; neural network ; digital filter