基于ACCRBF网络的多层砖房震害预测
|
摘要
针对传统震害预测方法逐栋抽样计算建筑物抗震性能的不足,本文提出了一种基于蚁群聚类径向基(ACCRBF)网络模型的建筑物震害预测方法。依据不同地震动峰值加速度下多层砖房的实际震害资料,对模型进行训练,在模型的输入和输出之间建立映射关系,并利用这种映射关系对未知样本进行分类,实现对多层砖房的震害分析和预测。模型的输入为反映结构的震害影响因子,输出为给定的地震动峰值加速度下结构震害等级。研究表明,基于ACCRBF网络模型的多层砖房震害预测结果与震害实例基本吻合,具有推广应用价值。
To overcome the deficiency in the conventional seismic damage prediction method, in which the anti-earthquake behavior is evaluated by sampling survey, here we present a construction seismic damage prediction method based on ant colony clustering radial basis function (ACCRBF) neural network model. Through training the network on the basis of real seismic damage data, this method sets up the mapping relationship of the parameters between the inputs and outputs. Then The mapping relationship then can be adopted for sample classification, which makes it possible to realize the seismic damage evaluation and hazard prediction. The inputs of the model are the factors that affect the seismic damage, and the output of the model is the seismic damage level under certain peak ground acceleration. Our results show that the seismic damage prediction results from the ACCRBF neural model are in good agreement with the real examples. Therefore, it is worthy of promotion and application in the future.
To overcome the deficiency in the conventional seismic damage prediction method, in which the anti-earthquake behavior is evaluated by sampling survey, here we present a construction seismic damage prediction method based on ant colony clustering radial basis function (ACCRBF) neural network model. Through training the network on the basis of real seismic damage data, this method sets up the mapping relationship of the parameters between the inputs and outputs. Then The mapping relationship then can be adopted for sample classification, which makes it possible to realize the seismic damage evaluation and hazard prediction. The inputs of the model are the factors that affect the seismic damage, and the output of the model is the seismic damage level under certain peak ground acceleration. Our results show that the seismic damage prediction results from the ACCRBF neural model are in good agreement with the real examples. Therefore, it is worthy of promotion and application in the future.
引文
成小平,胡聿贤,帅向华,2009.基于神经网络模型的房屋震害易损性估计方法.自然灾害学报,9(2):68—73.
姜伟,马令勇,刘功良,2011.基于MATLAB神经网络方法的多层砖房震害预测.西北地震学报,33(2):155—158.
康飞,李俊杰,许青,2009.堆石坝参数反演的蚁群聚类RBF网络模型.岩土力学与工程学报,28(增2):3640—3644.
刘本玉,叶燎原,江见鲸,2002.用模糊人工神经网络方法预测多层砖房震害.清华大学学报(自然科学版),42(6):843—846.
乔亚玲,闫维明,2005.建筑结构震害预测方法研究评述.工业建筑,35(6):1—5.
汤皓,陈国兴,2006a.灰关联与人工神经网络在建筑物震害预测中的应用.地震工程与工程振动,26(3):57—59.
汤皓,陈国兴,李方明,2006b.基于BP神经网络模型的多层砖房震害预测方法.地震工程与工程振动,26(4):141—146.
伊之潜,1996.地震灾害及损失预测方法.北京:地震出版社.
杨秀萍,康飞,李俊杰,2011.基于ACCRBF网络的面板堆石坝参数反演.水电自动化与大坝监测,35(2):56—58.
姜伟,马令勇,刘功良,2011.基于MATLAB神经网络方法的多层砖房震害预测.西北地震学报,33(2):155—158.
康飞,李俊杰,许青,2009.堆石坝参数反演的蚁群聚类RBF网络模型.岩土力学与工程学报,28(增2):3640—3644.
刘本玉,叶燎原,江见鲸,2002.用模糊人工神经网络方法预测多层砖房震害.清华大学学报(自然科学版),42(6):843—846.
乔亚玲,闫维明,2005.建筑结构震害预测方法研究评述.工业建筑,35(6):1—5.
汤皓,陈国兴,2006a.灰关联与人工神经网络在建筑物震害预测中的应用.地震工程与工程振动,26(3):57—59.
汤皓,陈国兴,李方明,2006b.基于BP神经网络模型的多层砖房震害预测方法.地震工程与工程振动,26(4):141—146.
伊之潜,1996.地震灾害及损失预测方法.北京:地震出版社.
杨秀萍,康飞,李俊杰,2011.基于ACCRBF网络的面板堆石坝参数反演.水电自动化与大坝监测,35(2):56—58.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心