Signal extraction and fault identification of steam turbine vibration

Abstract

This paper, the vibration signal of steam turbine which are detected by fault diagnosis are influenced by environmental noise and detecting instrument itself, leading to vibration signal waveform distortion which contains a large number of non-stationary composition, and cannot effectively react turbine fault characteristics, and the coupling among different fault characteristics of unilateral fault features make it difficult to identify fault accurately. Aiming at solving this problem, this paper combine the axis of spectrum analysis with path analysis of vibration signal processing and recognition method, two kinds of detection method in the fault diagnosis process validation to ensure the accuracy of test results.