步行荷载的两阶段遗传算法识别  

Walking load identification based on two-stage genetic algorithm

作　　者：王鹏程 陈隽[1,2] 王浩祺 WANG Pengcheng;CHEN Jun;WANG Haoqi(School of Civil Engineering,Tongji University,Shanghai 200092,China;State Key Lab of Civil Engineering Disaster Prevention,Shanghai 200092,China;Bridge and Structure Laboratory,University of Tokyo,Tokyo 113-8656,Japan)

机构地区：[1]同济大学土木工程学院,上海200092 [2]土木工程防灾国家重点实验室,上海200092 [3]东京大学桥梁与结构实验室,东京113-8656

出　　处：《振动与冲击》2019年第19期64-69,共6页Journal of Vibration and Shock

基　　金：国家自然科学基金(51778465);上海中医健康服务协同创新中心项目(ZYJKFW201811009)

摘　　要：以测力板直接测试结果为基础的步行荷载建模方式,存在有不能准确反映结构上行人真实步态的疑问。为此,利用遗传算法,研究由结构响应识别步行荷载的反问题建模方式。首先在傅里叶级数荷载模型基础上,考虑步行的双支撑特征,分别以结构位移响应功率谱及位移时程一致为目标函数,建立了两阶段遗传算法识别步行荷载参数的方法。进而通过荷载模型和实测荷载的识别分别检验了算法的准确性和适用性,识别结果表明前5阶动载因子、相位以及步行频率等十一个控制参数的识别效果较好,算法具有鲁棒性。The walking load modeling method based on direct test results of a force-measuring plate has a problem being not able to reflect pedestrian’s true gait on a structure.Here,a genetic algorithm was used to study the inverse problem modeling method of recognizing walking load with structural responses.Firstly,based on Fourier series load model,considering double-support feature of walking,taking structural displacement response power spectrum and displacement response time history as objective functions,the two-stage genetic algorithm was built to identify walking load parameters.Then,the correctness and applicability of the proposed algorithm were verified through identification of the load model and the actual measured load,respectively.The recognition results showed that the recognition effect of 11 control parameters including the first five orders dynamic load factors,phase and walking frequency,etc.is better;the proposed algorithm is robust.

关 键 词：荷载识别 遗传算法 傅立叶级数模型 步行荷载 

分 类 号：TH212[机械工程—机械制造及自动化] TH213.3