The paper puts forward a novel method of the damage detection using singular value decomposition (SVD) and artificial bee colony (ABC) algorithm for beam structures. The proposed method includes two steps: damage localization and depth estimation. In the first step, singular value decomposition is used to decompose the trajectory matrix of the attractor reconstructed from the shape data to localize the damage. In the second step, B-spline wavelet on the interval (BSWI) finite element method is usually applied to construct a damage extent estimation database at the detected damage locations, which can reveal the relationship between natural frequencies and its damage extent. Then ABC algorithm is adpoted to predict the damage depths by searching the frequency estimation database. Numerical simulations and experiment study are used to accurately verify the effectiveness of the proposed method for the case of cantilever beam.

提出了一种利用奇异值分解（SVD）和人工蜂群（ABC）算法对梁结构进行损伤检测的新方法。所提出的方法包括两个步骤：损伤定位和深度估计。第一步，使用奇异值分解分解从形状数据重建的吸引子的轨迹矩阵，以定位损伤。第二步，通常采用区间B样条小波（BSWI）有限元方法在检测到的损伤位置建立损伤程度估计数据库，可以揭示固有频率与其损伤程度之间的关系。然后采用ABC算法通过搜索频率估计数据库来预测损伤深度。