进化神经网络预测房屋爆破振动破坏程度研究  

Research on predicting the destructive degree of housing blasting vibration through advanced Neural Network

作　　者：郑松鹏 黄志波[1] 李辉 汪勇志 ZHENG Songpeng;HUANG Zhibo;LI Hui ;WANG Yongzhi(College of Jinshan,Fujian Agriculture and Forestry University,Fuzhou 350002)

机构地区：[1]福建农林大学金山学院,福建福州350002

出　　处：《福建建筑》2018年第7期120-123,共4页Fujian Architecture & Construction

基　　金：福建农林大学金山学院院级服务产业特色专业项目(Y160402);2017年院级大学生校外实践教育基地建设项目(S170401)

摘　　要：房屋在爆破振动作用下的损害状况是当前爆破工程界的研究热点和难点。考虑爆破振动特征参量和房屋结构特性的影响,引入进化神经网络,并通过对神经网络结构参数和权值的优化,建立爆破振动对房屋破坏程度的预测模型。研究表明,该方法预测结果与房屋破坏情况吻合较好,其量化的结果对准确评判爆破振动对房屋影响具有重要意义。Presently,housing damaging situation under blast vibration is a hot and difficult issue among the field of Blasting Engineering.Taking the characteristic parameters of blasting vibration and housing structural characters into consideration,Evolutionary Neural Network was put forward as well;plus,through optimizing the structural parameters and weights of Neural Network,a model predicting the housing damaging degree under blasting vibration was found.Researches indicate that the predicted results can keep consistent with the real destruction of housing very well.Hence,the quantitative results of this approach can provide basis for judging the degree of housing destruction,which shows important significance in accurately judging the influences on housing caused by blasting.

关 键 词：爆破振动 神经网络 预测 遗传算法 

分 类 号：TU751.9[建筑科学—建筑技术科学]