Enhanced symplectic Ramanujan mode pursuit and its application in mechanical composite fault diagnosis

Abstract

In practical engineering applications, traditional signal decomposition methods are often affected by various factors such as strong noise, alternating periods, etc. When signal mode analysis is conducted, it is often found that the decomposition results do not meet engineering requirements. To address these issues, Enhanced Symplectic Ramanujan Mode Pursuit (ESRMP) method is proposed in this paper, which aims to improve the accuracy and reliability of signal decomposition and period estimation. First, the cyclic symplectic geometry similarity transform is used to separate the components of different modes in the signal, and the anti-noise autocorrelation function is used to estimate the period of different components. Then, the rectangular length of the intercepted signal is determined based on the estimated period, and the periodic compensation is achieved through related detection and cubic spline interpolation. Finally, the reconstructed signal is projected onto the Ramanujan subspace to extract and enhance periodical pulses. The experimental results of multimodal composite fault signals show that the ESRMP method can accurately separate components of different modes, especially periodic pulse signals.