基于LM-BP神经网络的钢筋混凝土框架结构震害快速预测模型
|
摘要
为了快速对建筑群进行震害预测,构建城乡地震易损性模型,为防灾减灾工程提供依据并助力韧性城乡建设,本文选取位于抗震设防烈度为7度区的某市,把该市已完成的震害预测项目的基础数据和结果作为数据库的样本来源,以框架结构作为研究对象,将层数、层高、楼高、柱面积率等12个易获取且关联度较高的关键数据作为震害影响因子,充分利用了Matlab可视性良好的建模特性及LM算法能够快速拟合的优势,训练了一个基于LM算法的BP神经网络震害预测模型。采用多组数据测试,对单次模拟结果数据进行对比,并通过多次随机选取数据样本建立模型,验证该模型的精度和稳定性。测试结果表明,本文所选取的震害影响因子能够对框架结构的震害预测结果进行准确映射,该方法准确快速高效,可以应用于建筑群的快速震害预测实际工作中。
In order to predict the earthquake damage of buildings quickly,to build seismic vulnerability models for urban and rural areas, to provide a basis for the disaster prevention and mitigation projects, and to build the resilient city,a city with the Ⅶ seismic intensity is taken for studies in this paper. As the source of the database sample, basic data and results of the completed earthquake damage prediction are used in the research of RC frame structures.12 high relational key parameters are easily obtained as the seismic damage impact factor, such as the number and the height of floors, the building height and the area ratio of column, and so on. A model of earthquake damage prediction based on LM-BP neural network is fully trained, by use of the advantage that on the one hand MATLAB has a good visibility and on the other hand the LM algorithm can fit very quickly. The results from a number of tests show that the model is very stable, and the seismic damage impact factor can accurately map results of the earthquake damage prediction for the RC frame structure. As a consequence, the new method proposed in this paper can be applied to predict earthquake damage of actual engineering rapidly and accurately.
In order to predict the earthquake damage of buildings quickly,to build seismic vulnerability models for urban and rural areas, to provide a basis for the disaster prevention and mitigation projects, and to build the resilient city,a city with the Ⅶ seismic intensity is taken for studies in this paper. As the source of the database sample, basic data and results of the completed earthquake damage prediction are used in the research of RC frame structures.12 high relational key parameters are easily obtained as the seismic damage impact factor, such as the number and the height of floors, the building height and the area ratio of column, and so on. A model of earthquake damage prediction based on LM-BP neural network is fully trained, by use of the advantage that on the one hand MATLAB has a good visibility and on the other hand the LM algorithm can fit very quickly. The results from a number of tests show that the model is very stable, and the seismic damage impact factor can accurately map results of the earthquake damage prediction for the RC frame structure. As a consequence, the new method proposed in this paper can be applied to predict earthquake damage of actual engineering rapidly and accurately.
引文
[1] 郭明珠,吕欣欣,王亚昆.国内外群体建筑物震害预测方法研究现状综述[J].防灾科技学院学报,2016,18(4):8-13.GUO Mingzhu,LV Xinxin,WANG Yakun.Research status review of earthquake damage prediction methods of building group at home and abroad [J].Journal of Institute of Disaster Prevention,2016,18(4):8-13.(in Chinese)
[2] 孙柏涛,孙得璋.建筑物单体震害预测新方法[J].北京工业大学学报,2008(7):701-707.SUN Baitao,SUN Dezhang.New method for predicting single earthquake damage in buildings[J].Journal of Beijing University of Technology,2008(7):701-707.(in Chinese)
[3] 张桂欣,孙柏涛.多因素影响的建筑物群体震害预测方法研究[J].世界地震工程,2010,26(1):26-30.ZHANG Guixin,SUN Baitao.A method for earthquake damage prediction of building groups based on multiple factors [J].World Earthquake Engineering,2010,26(1):26-30.(in Chinese)
[4] 张润,王永滨.机器学习及其算法和发展研究[J].中国传媒大学学报(自然科学版),2016,23(2):10-18+24.ZHANG Run,WANG Yongbin.Research on machine learning with algorithm and development [J].Journal of Communication University of China(Science and Technology),2016,23(2):10-18+24.(in Chinese)
[5] 成小平,胡聿贤,帅向华.基于神经网络模型的房屋震害易损性估计方法[J].自然灾害学报,2000(2):68-73.CHENG Xiaoping,HU Yuxian,SHUAI Xianghua.A method of building vulnerability estimation for earthquake damage based on neural network model[J].Journal of Natural Disasters,2000(2):68-73.(in Chinese)
[6] 刘本玉,叶燎原,江见鲸.用模糊人工神经网络方法预测多层砖房震害[J].清华大学学报(自然科学版),2002(6):843-846.LIU Benyu,YE Liaoyuan,JIANG Jianjing.Forecasting seismic damage in multistory masonry buildings with a neuro-fuzzy approach [J].Journal of Tsinghua University(Science and Technology),2002(6):843-846.(in Chinese)
[7] 尹之潜.地震灾害及损失预测方法[M].北京:地震出版社,1996.YIN Zhiqian.Earthquake Disaster and Loss Prediction Method [M].Beijing:Seismological Press,1996.(in Chinese)
[2] 孙柏涛,孙得璋.建筑物单体震害预测新方法[J].北京工业大学学报,2008(7):701-707.SUN Baitao,SUN Dezhang.New method for predicting single earthquake damage in buildings[J].Journal of Beijing University of Technology,2008(7):701-707.(in Chinese)
[3] 张桂欣,孙柏涛.多因素影响的建筑物群体震害预测方法研究[J].世界地震工程,2010,26(1):26-30.ZHANG Guixin,SUN Baitao.A method for earthquake damage prediction of building groups based on multiple factors [J].World Earthquake Engineering,2010,26(1):26-30.(in Chinese)
[4] 张润,王永滨.机器学习及其算法和发展研究[J].中国传媒大学学报(自然科学版),2016,23(2):10-18+24.ZHANG Run,WANG Yongbin.Research on machine learning with algorithm and development [J].Journal of Communication University of China(Science and Technology),2016,23(2):10-18+24.(in Chinese)
[5] 成小平,胡聿贤,帅向华.基于神经网络模型的房屋震害易损性估计方法[J].自然灾害学报,2000(2):68-73.CHENG Xiaoping,HU Yuxian,SHUAI Xianghua.A method of building vulnerability estimation for earthquake damage based on neural network model[J].Journal of Natural Disasters,2000(2):68-73.(in Chinese)
[6] 刘本玉,叶燎原,江见鲸.用模糊人工神经网络方法预测多层砖房震害[J].清华大学学报(自然科学版),2002(6):843-846.LIU Benyu,YE Liaoyuan,JIANG Jianjing.Forecasting seismic damage in multistory masonry buildings with a neuro-fuzzy approach [J].Journal of Tsinghua University(Science and Technology),2002(6):843-846.(in Chinese)
[7] 尹之潜.地震灾害及损失预测方法[M].北京:地震出版社,1996.YIN Zhiqian.Earthquake Disaster and Loss Prediction Method [M].Beijing:Seismological Press,1996.(in Chinese)