Stable subspace dimension reduced MUSIC for blade tip timing. (17th February 2023)

• Record Type: Journal Article
• Authors: Yang, Zhi-Bo; Wang, Zeng-Kun; Li, Hao-Qi; Tian, Shao-Hua; Chen, Xue-Feng
• Is Part Of: Journal of sound and vibration. Volume 545(2023)
• Publication Date: 2023-02-17
• Subjects: MUSIC -- Blade tip timing -- Condition monitoring -- Spectrum analysis -- Min-norm vector; Sound -- Periodicals; Vibration -- Periodicals; Son -- Périodiques; Vibration -- Périodiques; Sound; Vibration; Periodicals; Electronic journals; 620.205
• DOI: 10.1016/j.jsv.2022.117438
• Languages: English
• ISSNs: 0022-460X

Abstract

Abstract: Multiple signal classification (MUSIC) is a parameter extraction method for blade tip timing to suppress the under-sampled problem. Its faster version, subspace dimension reduced MUSIC (SDR-MUSIC), has been proposed, which uses a single noise vector instead of the whole noise subspace to reduce computational complexity. However, the principle of random noise vector selection leads to instability in frequency identification. Therefore, we replace the random vector with the min-norm vector to obtain a stable SDR-MUSIC (SSDR-MUSIC). The superiority of the min-norm vector can be mathematically proven by the location of the zeros of the polynomial. Simulated and experimental tests indicate that SSDR-MUSIC gains stability when the min-norm vector is used while maintaining computational efficiency. Highlights: SSDR-MUSIC boosts the stability of SDR-MUSIC under measurement noise. SSDR-MUSIC inherits the low computational complexity of SDR-MUSIC. The rationality of min-norm vector is proved based on the polynomial zeros. The rank of target frequency are proposed for stability comparison. The stability is verified by experiment and simulated MC tests under various SNRs.