涡轮泵实时故障检测的频段幅值最大值比方法

根据液体火箭发动机涡轮泵出现故障后振动加速度信号的频谱变化,提出了一种基于频段幅值最大值比的故障检测方法用于涡轮泵实时故障检测。该方法采用快速傅里叶变换（fast fourier transformation,FFT）将历史信号从时域变换到频域并进行频段划分,选择两个有代表性的频段并计算其幅值最大值比。将该比值作为故障特征对待检信号进行实时故障检测。用某型号涡轮泵振动加速度信号对故障检测方法进行验证,结果表明对时长20.80 s的待检信号,设定故障特征阈值为3.5,该方法检出故障时刻为20.08 s,比故障真实出现时刻晚0.07 s（在0.5 s之内）,未出现虚警和漏警。该故障检测方法具备了良好的精确性与实时性。

Turbopump real-time fault detection method based on maximum amplitude ratio of frequency band

According to the spectrum change of vibration acceleration signal after appearing of the LRE（Liquid Rocket Engine） turbopump faults,a fault detection method based on maximum amplitude ratio of frequency band is proposed for turbopump real-time fault detection.The method transforms the historical signals from time domain to frequency domain through Fast Fourier Transformation（FFT）,divides frequency spectrum into several bands,chooses two typical bands and computes the maximum amplitude ratio.The ratio is used as fault feature in the real-time fault detection of the test signals.A part of the vibration acceleration signal of a certain type of turbopump is chosen as the test object to validate the method.The test results show that for the test signals within 20.80 s＇ duration,set the fault feature threshold as 3.5,the method detects the faults at 20.08 s without false alarm and missing alarm,and there is a delay of 0.07 s（much less than 0.5 s） after the real appearing of faults.The method has good accuracy and real-time performance.