基于小生境遗传算法的设计地震动反应谱标定方法

Calibrating method for seismic response spectra based on niche genetic algorithm

基于小生境遗传算法,提出了新的设计地震动反应谱标定方法,给出了反应谱标定的验证方法。设计地震动反应谱特征参数包括地震动峰值、第一拐点周期、第二拐点周期(特征周期)、平台高度值及衰减指数等,其标定属于多参数优化问题。利用小生境遗传算法具有多参数的全局寻优能力,提出了以平均标准差最小为目标的地震动反应谱标定模型。拟合给定特征参数的设计地震动反应谱,对比分析标定的特征参数与给定的特征参数,要求两者之间的误差小于给定的精度,以此作为新的验证方法。以某一场地计算反应谱作为算例,采用所提出的模型对其进行了标定。通过验证结果和工程算例,证明了所提出的地震动反应谱标定方法的有效性。

A new method, niche genetic algorithm （NGA）, is adopted to calibrate the seismic response spectra and a verification standard for calibrating the seismic response spectra is proposed. The characteristic parameters of seismic response spectra include the peak of ground motion, the first turning point of spectrum, the second turning point of spectrum, the flat height of spectrum and the attenuation index, and so on, and therefore the calibration of the seismic response specta is a problem of multi-parameter optimization. NGA, with the ablitity of global optimization of multi-parameter, is used to solve this problem, aiming at the least mean standard error of spectral values at appointed periods. A new verification method proposed is that the seismic response spectra with given characteristic parameters are used to verify the effectiveness of the method, that is, the seismic response spectra with given characteristic parameters are calibrated and the errors between the calibrated characteristic parameters and the given ones are smaller than the appointed precision. Three pieces of the calculated response spectra for one real site are calibrated as an application example. The obtained results show the effectiveness and efficiency of the proposed method.