基于粒子群优化投影寻踪概率模型的桥梁地震震害预测方法研究

发布时间：2018-02-23 03:42

本文关键词： 投影寻踪粒子群桥梁震害评价置信区间　出处：《世界地震工程》2017年03期 　论文类型：期刊论文

【摘要】：地震是我国最常见的自然灾害之一,强烈的地震往往会引起震区内大量建筑物倒塌或损害。桥梁是是社会生命线工程的重要构成部分,也是地震中最容易破坏的部件。如需对震区桥梁震害损失、交通可靠性、桥梁加固优化以及救援应急决策等方面研究,需要进行桥梁震前的震害评估。传统的震前震区桥梁震害预测的方法或多或少存在主观性的因素,且存在评估结果稳定性不高的情况。借助投影寻踪模型数理统计聚类方法,构建了基于投影寻踪方法的生命线桥梁震害预测模型,结合粒子群计算机算法优化投影方向参数,将桥梁震害预测的多维评价指标投影计算成一维数值,并根据投影值的大小以及置信区间对桥梁的震害进行预测,通过两组数据实验验证模型,结果基本达到了预期,泛化能力较好。
[Abstract]:Earthquake is one of the most common natural disasters in China. Strong earthquakes often cause a large number of buildings to collapse or damage in the earthquake area. Bridge is an important component of social lifeline engineering. It is also the most easily damaged component in the earthquake. For example, the damage loss of bridges in the earthquake area, the reliability of traffic, the optimization of bridge reinforcement and the emergency decision of rescue and so on are studied. It is necessary to evaluate the earthquake damage of bridges before earthquake. The traditional prediction method of bridge damage in seismic area before earthquake has subjective factors more or less, and the stability of evaluation results is not high. By means of projection pursuit model, mathematical statistical clustering method is used to predict the damage of bridges. The prediction model of seismic damage of lifeline bridges based on projection pursuit method is constructed. The projection direction parameters are optimized by using particle swarm optimization algorithm, and the multi-dimensional evaluation indexes of bridge damage prediction are projected into one-dimensional values. According to the magnitude of the projection value and the confidence interval, the earthquake damage of the bridge is predicted, and the model is verified by two groups of data. The results basically reach the expectation and the generalization ability is good.
【作者单位】：中国海洋大学海洋生态与环境教育部重点实验室;中国海洋大学工程学院;
【基金】：山东省优秀中青年科学家奖励基金(ZR2016DB25) 青岛市博士后应用研究项目中央高校基本业务研究经费
【分类号】：U442.55

【相似文献】