基于Web-BIM的装配式建筑结构选型智能决策支持方法  

Web-BIM-based Intelligent Decision Support Method for Prefabricated Building Structure Selection

作　　者：王彩雪[1] 张乐芳[1] WANG Caixue;ZHANG Lefang(School of Human Settlements and Civil Engineering,Xi’an Eurasia University,Xi'an 710065,China)

机构地区：[1]西安欧亚学院人居环境学院,陕西西安710065

出　　处：《工业加热》2022年第5期72-76,共5页Industrial Heating

基　　金：陕西省第二批新工科研究与实践项目(2020GKSJ001);西安欧亚学院校级重点课程(2019KC006);西安欧亚学院本科教学改革项目(2021YB019);西安欧亚学院校级科研项目(2021XJZK04)。

摘　　要：装配式建筑作为建筑行业的发展趋势之一,凭借节能环保等诸多优势,受到业界人士的普遍认可,因此,为提升此类建筑的应用价值,取得最优结构方案,将Web端与建筑信息模型技术融合,提出了一种智能决策支持方法。以Web万维网为基础,利用扩展应用等模块以及有关软件构建Web-BIM平台,基于此引入人工神经网络建立三层后向传播神经网络,明确变换函数、隐含层层数与各层节点数量,经归一化处理后,设计双向传播的监督学习算法,根据网络取值特点将决策事实或假设划分成非顺序取值与顺序取值,通过调整各节点权重得到期望的决策方案。实验环节选取三种主要建筑结构,从抗震性与价值指数角度验证了该方法的决策结果,实验结果表明所提方法具有较好的决策有效性,可以用于实际应用中。As one of the development trends of the construction industry,prefabricated buildings are generally recognized by the industry with many advantages such as energy saving and environmental protection.Therefore,in order to enhance the application value of such buildings and obtain the optimal structural scheme,an intelligent decision support method is proposed by integrating the web end and building information model technology.Based on the web world wide web,the web-bim platform is constructed by using extended application modules and related software.Based on this,the artificial neural network is introduced to establish a three-layer back-propagation neural network.The transformation function,the number of hidden layers and the number of nodes in each layer are defined.After normalization,the supervised learning algorithm of two-way propagation is designed,According to the network value characteristics,the decision-making fact or hypothesis is divided into non sequential value and sequential value,and the expected decision scheme is obtained by adjusting the weight of each node.Three main building structures are selected in the experiment to verify the decision-making results of the method from the perspective of seismic performance and value index.The demerit of the experiment shows that the proposed method has good decision-making effectiveness and can be used in practical application.

关 键 词：Web-BIM 装配式建筑 人工神经网络 建筑结构选型 智能决策支持 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]