A novel method for early looseness state identification of bolted joint beams is proposed in this paper based on generalized variational mode decomposition (GVMD) and a similarity index. In the proposed method vibration signals are decomposed by GVMD, which has the property of the multiscale and fixed-frequency decomposition. To effectively extract the desired modes, the frequency band allocation is designed through flexibly defining scale parameters and prior center frequencies according to the characteristics of the signal itself and real needs. A new similarity index is formed based on the centroid frequency ratio of sensitive vibration modes to reliably identify early looseness state. The effectiveness of the proposed method is verified by impulse experiments of bolted joint transverse beam and bolted joint vertical beam. The results indicate that compared with other methods, the proposed method can effectively identify the early looseness state of the bolted joint beams, and has a good repeatability.

提出了一种基于广义变分模态分解（GVMD）和相似度指标的螺栓连接梁早期松动状态识别的新方法。该方法利用GVMD对振动信号进行分解，具有多尺度和固定频率分解的特性。为了有效地提取期望的模式，通过根据信号本身的特性和实际需要灵活定义比例参数和先前的中心频率来设计频段分配。基于敏感振动模式的质心频率比形成新的相似性指标，以可靠地识别早期松动状态。螺栓连接横梁和螺栓连接竖向梁的脉冲试验验证了该方法的有效性。