基于多种搜索策略的人工蜂群算法的结构损伤识别

人工蜂群算法是一种元启发式算法,具有构架简单,易于操作和鲁棒性较好的特点。本文对原始蜂群算法进行了改进,为蜜蜂们提供了更丰富的搜索策略。基于轮盘赌原则选择更优的迭代方式,进而使算法的收敛速度和精度有了显著的提高。基于频率残差和模态置信准则(MAC)建立结构损伤识别问题的目标函数,然后利用ABC算法,QABC算法和本文方法求解该目标函数,得到损伤识别的结果。利用一桁架结构做数值模拟,用简支梁结构进行实验验证。算例表明,改进的算法更能有效地检测出结构的局部损伤,具有对测量噪声不敏感、高效率以及高精度等优点。

Structural damage identification based on ABC algorithm with variable search strategy

A technique for structural damage detection based on the Artificial Bee Colony algorithm with variable search strategy based on modal data is presented.More search strategies are offered and the bee will choose one search mode based on the tournament selection strategy.Such change can enhance the algorithm''s convergence rate and accuracy.The frequencies residual and MAC are utilized to form objective function,the ABC,QABC and proposed method are used to solve the nonlinear optimal problem,and acquire identified results.The optimal solution can represent the real situation of the structure.A numerical example (a 61 bar truss structure) is utilized to investigate the efficiency of proposed method.An experimental work (a supported beam) is also referenced for further verification.Final results show the ABC with variable search strategy can acquire better identification outcome,compared with ABC and QABC,even when some modal data is polluted by the artificial noise.