Gong Wen-Bin, Li An, Wu Zhong-Hong, Qin Fang-Jun
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033, Hubei, China.
College of Weaponry Engineering, Naval University of Engineering, Wuhan 430033, Hubei, China.
ISA Trans. 2023 Aug;139:660-674. doi: 10.1016/j.isatra.2023.03.051. Epub 2023 Apr 6.
Analyzing the vibration features of a shipboard stabilized platform (SSP) is significant to the design or vibration compensation of a marine gravimetric survey vibration isolation system. Empirical Fourier decomposition (EFD) is a recently developed method for nonlinear and non-stationary signal decomposition. However, the spectral segmentation boundary needs to be set in advance according to sophisticated experience, and it is easy to be disturbed by noise, and the decomposition result is inaccurate. In order to accurately extract the nonlinear vibration characteristics of SSP, this paper proposes a new method called power spectrum envelope adaptive empirical Fourier decomposition (PSEEFD). Firstly, the number of selected modal decomposition is determined based on the mutual information to realize adaptive segmentation. Then, the improved power spectrum envelope segmentation method is adopted to effectively diminish the interference of noise since the segmentation boundary is formed by the minimum of the adjacent extreme points enveloped by the maximum value of the power spectrum. The spectrum segments obtained from segmentation contain less interference. Finally, the component signal in each frequency band is reconstructed by inverse fast Fourier transform, and the instantaneous frequency signal component with physical significance is obtained. Through the analysis of vibration simulation signals and measured data of SSP, the proposed method is compared with EMD, AFVMD, EWT and EFD. The results show that PSEEFD has a well suppression of noise interference and can effectively extract the characteristics of nonlinear vibration signals.
分析舰载稳定平台（SSP）的振动特性对于海洋重力测量隔振系统的设计或振动补偿具有重要意义。经验傅里叶分解（EFD）是一种最近开发的用于非线性和非平稳信号分解的方法。然而，频谱分割边界需要根据复杂的经验预先设定，并且容易受到噪声干扰，分解结果不准确。为了准确提取SSP的非线性振动特性，本文提出了一种新的方法，称为功率谱包络自适应经验傅里叶分解（PSEEFD）。首先，基于互信息确定所选模态分解的数量以实现自适应分割。然后，采用改进的功率谱包络分割方法，由于分割边界由功率谱最大值所包络的相邻极值点的最小值形成，从而有效减少噪声干扰。分割得到的频谱段干扰较小。最后，通过快速傅里叶逆变换对每个频段的分量信号进行重构，得到具有物理意义的瞬时频率信号分量。通过对SSP振动仿真信号和实测数据的分析，将所提方法与EMD、AFVMD、EWT和EFD进行了比较。结果表明，PSEEFD对噪声干扰具有良好的抑制作用，能够有效提取非线性振动信号的特征。
