人工地震波遗传优化生成方法  

Artificial seismic wave generation using genetic optimization

作　　者：杜迎乾 王之军 周承浩 许汉兵 贾书海[2] Du Ying-qian;Wang Zhi-jun;Zhou Cheng-hao;Xu Han-bing;Jia Shu-hai(Pinggao Group Co.Ltd.,Pingdingshan 467001,China;Institute of Mechanical Engineering,Xian JiaoTong University,Xi'an 710049,China)

机构地区：[1]平高集团有限公司,河南平顶山467001 [2]西安交通大学机械工程学院,陕西西安710049

出　　处：《工程抗震与加固改造》2021年第6期149-156,95,共9页Earthquake Resistant Engineering and Retrofitting

摘　　要：传统三角级数法生成人工地震波过程易受误差顽固点干扰,向设计谱迭代逼近时存在收敛速度慢、拟合精度低等缺陷。对此,引入群优化思想,形成基于遗传优化的人工地震波生成方法,克服幅值谱修正过程中线性缩放方法弊端,引导反应谱拟合过程的定向优化。针对数据特征,设计矩阵化加速处理方案,提高计算效率。研究结果表明:通过定向清除顽固点,可以提高人工地震波的反应谱拟合精度。由此方法构造出的人工波数据符合抗震设计规范,真实还原自然地震动属性,增强了设备动力时程抗震分析工作的准确性和可信度。The traditional trigonometric series method of seismic wave generation is always interfered by the stubborn points in the process of approximating the design spectrum, therefore the iteration is difficult to converge and the fitting accuracy is low. Herein, the idea of group searching is introduced to form a new way to generate seismic wave based on genetic optimization. To reduce the calculation burden, the intermediate data is stored and calculated in form of matrix instead of redundant loops, in which way computation acceleration is achieved. By directional removal of stubborn points, the response spectrum fitting accuracy is significantly improved. The genetic optimization overcomes the drawbacks in linear scaling correction method of amplitude and guide the fitting process well. Matrix acceleration scheme provides effective support in computation efficiency, as the time consumption of optimization is obviously squeezed. The as-constructed artificial wave is in good coordination with the seismic design code, and truly restores the natural ground motion properties. They can be ideal input for dynamic time-history analysis of important equipment, with seismic examination credibility enhanced.

关 键 词：人工地震波 遗传优化 反应谱拟合 抗震校核 动力时程分析 

分 类 号：P315[天文地球—地震学]