基于多传感器信息融合的桥梁健康监测系统的研究与实现

发布时间：2018-05-19 05:43

本文选题：桥梁健康监测 + 传感器网络　；参考：《南京邮电大学》2017年硕士论文

【摘要】：随着我国桥梁建造技术的发展,各类大型桥梁相继建成,而桥梁的健康状况直接影响了交通、经济和社会生活,对桥梁进行长期的,实时的状态监测显得至关重要。多传感器信息融合桥梁监测系统是利用传感器技术、计算机软件技术以及模式识别技术,对桥梁结构状况进行监测与评估,为桥梁的维护决策提供依据。本文在分析桥梁监测系统研究现状及信息融合技术的基础之上,提出了基于多传感器信息融合的桥梁监测系统方案。通过对监测内容及桥梁结构特点的研究,确定了传感器网络的构成。分析推导了倾角挠度转换算法,光波长应变转换算法以及曲线拟合算法,设计了基于LabVIEW的传感器数据采集上位机程序,以及基于MVC模式的JavaEE程序。另外,因传感器数量较多,且测量时间周期长,设计了基于SQL SERVER的数据库系统,为融合处理提供数据依据,同时对数据库优化问题进行了分析。利用神经网络在数据识别中的优势,设计了三层BP神经网络,同时为了避免陷入局部最小值,使用了附加动量法训练函数,基于Matlab对倾角数据和加速度数据,进行多传感器与单传感器数据的对比融合分析。仿真结果表明,网络表现出较快的收敛速度及较高的数据识别准确率,说明设计的三层BP神经网络符合融合需求且具有较高的可靠性,同时验证了多传感器数据融合的分类识别准确度高于单传感器数据融合。本文的应用对象为天津海河大桥,通过对硬件模块、软件模块和融合算法的研究与设计,实现了基于多传感器信息融合的监测系统[1]数据采集、传输、计算和显示等功能,并在天津海河大桥现场得到实际应用。
[Abstract]:With the development of bridge construction technology in China, all kinds of large bridges have been built one after another, and the health of the bridge directly affects the traffic, economic and social life. It is very important to monitor the bridge status in real time and long-term. The multi-sensor information fusion bridge monitoring system uses sensor technology, computer software technology and pattern recognition technology to monitor and evaluate the bridge structure condition, which provides the basis for bridge maintenance decision. Based on the analysis of the status quo of bridge monitoring system and information fusion technology, a bridge monitoring system scheme based on multi-sensor information fusion is proposed in this paper. The structure of sensor network is determined by studying the monitoring content and bridge structure characteristics. The dip deflection conversion algorithm, optical wave length strain conversion algorithm and curve fitting algorithm are analyzed and deduced. The upper computer program of sensor data acquisition based on LabVIEW and the JavaEE program based on MVC mode are designed. In addition, because of the large number of sensors and the long measuring period, the database system based on SQL SERVER is designed, which provides the data basis for fusion processing. At the same time, the optimization problem of database is analyzed. Taking advantage of the advantage of neural network in data recognition, a three-layer BP neural network is designed. In order to avoid falling into local minimum, the additional momentum training function is used, and the inclination angle data and acceleration data based on Matlab are analyzed. The comparison and fusion analysis of multi-sensor and single-sensor data are carried out. The simulation results show that the network has a faster convergence speed and a higher accuracy of data recognition, which indicates that the designed three-layer BP neural network meets the needs of fusion and has higher reliability. At the same time, the classification and recognition accuracy of multi-sensor data fusion is higher than that of single sensor data fusion. The application object of this paper is Tianjin Haihe Bridge. Through the research and design of hardware module, software module and fusion algorithm, the functions of data acquisition, transmission, calculation and display are realized in the monitoring system based on multi-sensor information fusion. And in Tianjin Haihe Bridge site has been practical application.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U446

【参考文献】