| 基于ANFIS的环境激励下桥梁结构应变响应预测分析 |
| |
| 引用本文： | 祝青鑫,王浩,茅建校,闵剑勇,刘建荣,王文君.基于ANFIS的环境激励下桥梁结构应变响应预测分析[J].中国公路学报,2019,32(11):62. |
| |
| 作者姓名： | 祝青鑫 王浩 茅建校 闵剑勇 刘建荣 王文君 |
| |
| 作者单位： | 1. 东南大学 混凝土及预应力混凝土结构教育部重点实验室, 江苏 南京 210096;2. 江苏省交通运输厅公路局, 江苏 南京 210004;3. 常州市公路管理处, 江苏 常州 213001 |
| |
| 基金项目： | 国家重点基础研究发展计划（"九七三"计划）项目（2015CB060000）；江苏省交通科学研究计划项目（8505001498）；江苏省重点研发计划项目（BE2018120）；江苏省研究生科研创新计划项目（KYCX19_0095） |
| |
| 摘 要： | 为准确预测复杂环境荷载作用下混凝土连续梁桥结构应变响应，基于结构健康监测系统长期实测数据，分析桥梁结构温度场变化规律，进而基于主成分分析及自适应神经网络模糊推理系统，建立桥梁结构温度场与桥梁结构应变响应的复杂非线性关系。首先，利用小波分解技术分离环境荷载及车辆荷载作用下的桥梁结构实测应变响应；然后利用平行坐标轴，分析混凝土连续梁桥结构温度场变化规律，并利用主成分分析提取结构温度场实测温度数据主成分；最后基于自适应神经网络模糊推理系统，以应变测点处温度数据、桥梁结构温度场实测温度数据主成分和采样时间点数据为输入数据，分别建立不同输入变量组合与应变响应的复杂非线性关系，并对比分析不同工况下结构应变响应的预测精度。结果表明：桥梁结构各测点处实测温度数据变化趋势基本一致，同侧测点实测温度数据高度相关，但桥梁结构上、下表面测点温度变化存在明显差异，仅考虑应变测点处温度变化，难以准确预测桥梁结构应变响应；当考虑桥梁结构温度场变化时，能更精确地建立温度与应变响应之间的关系模型，进而基于实测温度数据准确预测桥梁结构应变响应；当缺乏结构温度场实测温度数据时，将采样时间点作为反映桥梁结构温度场变化规律的参数，可取得较好效果。
|
| 关 键 词： | 桥梁工程 应变响应 自适应神经网络模糊推理系统 预测分析 主成分分析 温度 |
| 收稿时间： | 2019-03-13 |
Prediction of Strain Response of Bridges Under Ambient Excitation Based on Adaptive Network-based Fuzzy Inference System |
| |
| ZHU Qing-xin,WANG Hao,MAO Jian-xiao,MIN Jian-yong,LIU Jian-rong,WANG Wen-jun.Prediction of Strain Response of Bridges Under Ambient Excitation Based on Adaptive Network-based Fuzzy Inference System[J].China Journal of Highway and Transport,2019,32(11):62. |
| |
| Authors: | ZHU Qing-xin WANG Hao MAO Jian-xiao MIN Jian-yong LIU Jian-rong WANG Wen-jun |
| |
| Affiliation: | 1. Key Laboratory of Concrete and Prestressed Concrete Structures of Ministry of Education, Southeast University, Nanjing 210096, Jiangsu, China;2. Highway Administration of Jiangsu Communications Department, Nanjing 210004, Jiangsu, China;3. Changzhou Highway Administration Department, Changzhou 213001, Jiangsu, China |
| |
| Abstract: | This study aimed to accurately predict the strain of a concrete continuous girder bridge under complicated ambient loads. Temperature distributions were systematically investigated based on long-term field monitoring data. Accordingly, the relationships between temperature-induced strains and the temperature field were investigated using principal component analysis (PCA) and an adaptive network-based fuzzy inference system (ANFIS), which were applied to calculate the strain response under ambient excitation using field temperature measurements. First, the strain responses induced by temperature variations were extracted from the strain monitoring data based on wavelet decomposition and reconstruction. Then, the temperature variations were analyzed using parallel coordinates. In addition, the principal components of the temperature-monitoring data were extracted using PCA. The principal components, which provided a 99% cumulative contribution, were used to represent the structural temperature field. Subsequently, the temperature measurement corresponding to the strain data, the principal components of the temperature measurements, and the sampling time of the strain data were utilized as input data for ANFIS. Accordingly, models with different combinations of input variables were trained and then used to investigate the correlations of temperature-induced strain and the temperature field. Ultimately, the strain responses under ambient excitation were calculated using field temperature measurements. In addition, the prediction results from different models were compared with the field strain monitoring data. The results show that the variation trends of the temperature measurements from all the measurement points are consistent. Obvious relationships exist between the temperature data measured on the same side of box girder. However, obvious differences are observed between the temperature data measured on the opposite sides. Thus, calculating the strain data accurately using the temperature data from one measurement point is challenging. However, accurate results can be obtained using field temperature data from multi-measurement points. In addition, the sampling time of the strain data can be used to represent the characteristics of the structural temperature field during the calculations. |
| |
| Keywords: | bridge engineering strain response ANFIS prediction PCA temperature |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《中国公路学报》浏览原始摘要信息 |
|
点击此处可从《中国公路学报》下载全文 |
|
