基于免疫算法的无刷励磁发电机旋转整流器故障诊断

在大型发电机的无刷励磁系统中,旋转整流器承受着强大的离心力和电磁作用,故障发生率相对较高。采用旋转整流器故障时交流励磁机旋转电枢产生的谐波磁场在其定子励磁线圈中感应谐波电势的电磁特性,从中提取旋转整流器故障信号波形。运用傅里叶分解故障信号波形,对故障信号波形的特征值进行编码处理,并用免疫算法中的反向选择算法对傅里叶分解后的故障特征值进行计算,完成旋转整流器的故障诊断。利用人工免疫的反向选择算法,对一台15 kW无刷励磁同步发电机旋转整流器的短路与开路故障进行仿真诊断。仿真结果表明,基于免疫系统反向选择机制的人工免疫算法,与傅里叶分解算法相结合,进一步提高了无刷励磁同步发电机旋转整流器故障诊断的准确率,可达95%以上。

Immune algorithm in rotating rectifier fault diagnosis of brushless generator

In brushless excitation system of large generator,the rotating rectifier,which is under powerful centrifugal force and electromagnetic action,has higher incidence rate.Its fault signal waveform can be extracted from the harmonic voltage induced in the stator excitation windings of AC exciters by the harmonic magnetic flux produced by the fault current in its armature when the rotating rectifier is under any abnormal operating mode.The fault signal is expanded in Fourier series,and its characteristic values are encoded and then dealt with by the negative-selection algorithm of artificial immune system,the fault diagnosis for the rotating rectifier is thus completed.With this method,the short-circuit fault and open-circuit fault of the rotating rectifier have been diagnosed successfully for a 15 kW generator with brushless excitation system.The simulation result shows that,the fault diagnosis method for brushless excitation system of large synchronous generators based on the combination of the negative-selection immune algorithm and the Fourier series algorithm has higher correctness ratio of over 95%.