基于神经网络的方钢管混凝土柱隔板贯通节点静力拉伸承载力分析  

ANALYSIS ON STATIC TENSILE BEARING CAPACITY OF DIAPHRAGM-THROUGH JOINT OF CONCRETE-FILLED SQUARE STEEL TUBULAR COLUMN BASED ON NEURAL NETWORK

作　　者：穆友政 荣彬[1] 张广泰[1] 韩建红[1] 

机构地区：[1]新疆大学建筑工程学院,乌鲁木齐830047

出　　处：《建筑技术》2015年第5期459-462,共4页Architecture Technology

基　　金：国家自然科学基金资助项目(51268054);自治区自然科学基金项目(201233146-4);新疆大学校院联合资助项目(XY110137)

摘　　要：方钢管混凝土柱隔板贯通节点静力拉伸承载力,存在现有计算方法与已有试验结果误差较大的问题。而采用BP神经网络方法,建立隔板贯通节点静力拉伸承载力的预测模型,预测值与试验结果吻合较好,验证了神经网络方法预测节点静力拉伸承载力的精度与可行性。采用该神经网络模型对节点静力拉伸承载力进行参数化分析,具体参数包括:钢管柱宽厚比、隔板强度、钢管强度、隔板厚度、隔板浇筑孔径和混凝土强度。参数化分析表明,宽厚比、隔板强度、钢管强度、隔板厚度和浇筑孔径对静力拉伸承载力影响较大;混凝土强度对承载力影响较小,可以忽略。The existing computational method of static tensile bearing capacity of diaphragm-through joint of concrete-filled square steel tubular column has large errors comparing for the existing test results. In this paper, a prediction model for static tensile bearing capacity of diaphragm-through joint of concrete-filled square steel tubular column is developed based on BP neural network method. Predicted values are in good agreement with test results, which verifies the prediction accuracy and feasibility of neural network for predicting static tensile bearing capacity. Then, the parametric analysis for tensile bearing capacity of the joint is conducted using this neural network model, the specific parameters including： width-to-thickness ratio, diaphragm intensity, steel tube intensity, diaphragm thickness, diameter of concrete cast hole and concrete intensity. The parametric analysis indicates that width-to-thickness ratio, diaphragm intensity, steel tube intensity, diaphragm thickness and diameter of concrete cast hole have greater impact on static tensile bearing capacity, while concrete intensity has less impact on the capacity and it can be ignored.

关 键 词：方钢管混凝土柱 隔板贯通节点 静力拉伸承载力 神经网络 

分 类 号：TU317[建筑科学—结构工程]