基于小波神经网络的无刷同步发电机旋转整流器故障诊断研究

Study on Fault Diagnosis of Brushless Rotary Rectifier Based on Wavelet Neural Network

无刷励磁系统中旋转整流器因工作环境恶劣,发生故障时故障信息难以提取,因而故障诊断一直是一个难点,成为制约该励磁方式发展的障碍之一。针对大型无刷汽轮发电机故障诊断需要,提出应用小波神经网络对旋转整流器故障进行诊断,通过对旋转整流器的故障信号的频谱分析,提取故障信息频域特征量作为学习样本,通过训练使构建的小波神经网络能够准确反映频谱特征量和故障之间的映射关系,从而准确对故障进行诊断。提出了比较精确的数学模型,提高了旋转整流器故障的诊断能力和诊断的准确性。

The fault diagnosis of rotary rectifier is always difficult because of its hard working conditions and inconvenience of fauh information extraction, thus it is becoming one of the obstacle of the development of brushless excitation system. According to fault diagnosis of large brushless generator, a new method using wavelet neural network is proposed to be applied in fault diagnosis. By analyzing the spectrum of fault information, the fault information spectrum eigenvalue is extracted to be the study sample. After the wavelet neural network is trained with the sample, it can express the relationship of spectrum eigenvalue and fault precisely, therefore the diagnosis is done accurately. Precise mathematics model is proposed, and the ability and accuracy of fault diagnosis of rotary rectifier are improved.