An efficient approach for the diagnosis of faults in turbo pump of liquid rocket engine by employing FFT and time-domain features

Abstract

The turbo pump is the propulsion machinery used in liquid rocket engine. The fault caused due to the bearing and rotating components in the turbo pump would lead to a great damage to the pump or to the whole rocket engine and thus creates failure to the mission. For diagnosing the faults in the turbo pump, a technique based on the Fast Fourier Transform and time domain features is proposed in this study. The vibration produced in the turbo pump during dry run test gives the information about the condition of the turbo pump rotating components. The vibration of turbo pump is analysed by the extraction of statistical time domain features such as Mean, Standard Deviation, Root Mean Square, Kurtosis Factor, Crest Factor and Skew Factor. The faults present in the turbo pump are detected using support vector machine classifier, and the data are classified either as Normal or Abnormal data. The historical data-set obtained from Indian Space Research Organization is used for validation of this method. The obtained results indicate that this method diagnoses the fault effectively with 100% accuracy.