基于遗传算法的钢筋混凝土框架-剪力墙结构失效模式多目标优化  

MULTI-OBJECTIVE OPTIMIZATION OF GENETIC ALGORITHM-BASED FAILURE MODE FOR REINFORCED CONCRETE FRAME-SHEAR WALL STRUCTURES

作　　者：颜欣桐 徐龙河[1] YAN Xin-tong;XU Long-he(School of Civil Engineering,Beijing Jiaotong University,Beijing 100044,China)

机构地区：[1]北京交通大学土木建筑工程学院,北京100044

出　　处：《工程力学》2018年第4期69-77,共9页Engineering Mechanics

基　　金：国家重点研发计划项目(2016YFC0701101);国家自然科学基金项目(51578058;51322806);北京市自然科学基金项目(8172038)

摘　　要：该文提出了一种基于遗传算法(GA)的钢筋混凝土框架-剪力墙结构失效模式多目标优化方法。该方法以截面尺寸为优化变量,材料用量为约束条件,最大层间位移角及结构整体损伤指数为算法目标函数,利用基因序列的杂交及变异,实现有利基因的传递。将一5层钢筋混凝土框架结构简化为集中质量体系,验证了优化算法的正确性。以一10层钢筋混凝土框架-剪力墙结构为例,运用增量动态分析(IDA),确定其敏感地震动及相应峰值加速度(PGA),并作为优化过程中的地震动输入。对结构进行静力弹塑性分析,得到其屈服及极限位移,用于计算整体损伤指数。提出了多目标最小值优化问题的线性加权方法,并评价各性能指标的算法收敛性。历经4代共654个随机样本的弹塑性时程分析,结果表明:该方法在不增加材料用量的前提下,使得结构最大层间位移角减少了16.3%,整体损伤值减少了20.8%,各极限状态的年超越概率降低,结构抗倒塌储备系数提高,有效改善了结构的抗震性能。The multi-objective optimization method of genetic algorithm(GA)based failure mode is proposed for reinforced concrete(RC)frame-shear wall structure.The sectional dimensions and materials consumption are served as the optimization variable and constraint condition,respectively.The maximum drift ratio and global structural damage index are used to construct the objective functions of GA related,and the favorable gene is transmitted by the crossover and mutation of the gene sequence.With simplifying a 5-story RC frame structure to a lumped mass system,the validity of the algorithm is proved.A case study of a 10-story RC frame-shear wall structure is carried out.Applying Incremental Dynamic Analysis(IDA),the severest ground motion and corresponding peak ground acceleration are determined to serve as the seismic input during the process of optimization.Meanwhile,pushover analysis is implemented on the structure to obtain the values of the yield and ultimate displacements,which are used to calculate the global damage index.A linear weighted method for the multi-objective minimum optimization problem is proposed to evaluate the algorithm convergence speed for each evaluation criteria of the performance.After 654 random samples'elastic-plastic time-history analyses in 4 generations,on the condition of little increase of materials consumption,it is indicated that the maximum drift ratio of the structure and the global damage index are reduced by 16.3%and 20.8%,respectively.The mean annual exceeding probability of each limit state is decreased,the Collapse Margin Ratio(CMR)is increased at the same time,and the aseismic performance of structure is effectively improved.

关 键 词：遗传算法 失效模式优化 钢筋混凝土框架-剪力墙结构 增量动态分析 性能指标 

分 类 号：TU398.2[建筑科学—结构工程]