2016 年 3 巻 1 号 p. 15-00317
The detection of the presence of a structural abnormality in rotating machines with parts without gap, such as rigid couplings, by using the conventional method of vibration RMS, is difficult. This work investigates the application of Singular Spectral Analysis (SSA) to change detection in rotating machines. The performance of this technique is quantitatively evaluated in the scenario of misalignment in a turbopump assembly with rigid coupling. It is shown that a statistically significant change appears when using the Change Score calculated using the SSA method. Comparison with special indicators of a structural abnormality is also performed. Since the SSA detects changes in the shape of the signal, it is much more sensitive to changes related to abnormality than conventional methods or the use of special indicators.
