模糊聚类算法稳定图应用于桥梁结构参数识别

摘要

将模糊聚类算法应用于稳定图理论,并将该稳定图与协方差驱动的随机子空间法相结合,进行桥梁结构模态参数识别.提出不考虑阻尼的影响,以频率为横坐标,以模态保证准则MAC中任一列数据为纵坐标的模糊聚类算法稳定图.通过比较稳定图中各聚类圆大小的方法进行稳定图真假模态的判别,从而使模态判别不再需要人的主观参与而变得更加智能和准确.通过简支梁和连续梁仿真分析验证了模糊聚类算法稳定图用于桥梁结构参数识别的准确性和较强的抗干扰性.%Parameter identification of a bridge structure was undertaken using a stabilization diagram with fuzzy clustering method and stochastic subspace method. The stabilization diagram was proposed taking frequency as the abscissa, and any column data of modal assurance criterion (MAC) as the ordinate without consideration of damping. The compared circle method was presented to distinguish spurious and physical modes so that the stabilization diagram became more intelligent, correct and the mode recognition no longer needed people's subjective participation. Numerical simulation results of a simply supported beam and a continuous beam demonstrated the correctness and the stronger anti-interference ability of the proposed approaches.