民用建筑沉降监测与预报方法应用研究

发布时间：2018-07-11 11:14

本文选题：变形监测 + 沉降监测　；参考：《宁夏大学》2015年硕士论文

【摘要】：建筑变形监测及其数据分析处理与预报对于工程建设的安全非常重要。通过查阅、学习建筑变形监测技术及其数据分析处理与预报相关的文献资料,首先对深基坑和建筑物变形监测的意义和目的、内容及方法,以及监测方案技术设计的主要内容进行归纳。然后对几类常用的变形预报模型进行理论描述。最后结合4个参与完成的建筑工程沉降监测案例数据,分别运用MATLAB曲线拟合模型、神经网络模型、时间序列模型和灰色预报模型进行预报,并与实际监测数据对比分析,得出以下结论：(1)曲线拟合模型中,三次样条插值模型(cubic spline)具有一定的预报效果,当沉降量较大或沉降速度较快时,可以与其它模型联合预报。三次埃尔米特插值模型(shape-preserving)和多项式模型(Polynomial),不适于沉降监测预报。(2)神经网络模型中,GR神经网络模型预报效果好,BP神经网络模型和RBF神经网络模型预报效果较好。BP神经网络模型预报精度比RBF神经网络模型稍高,两者可与GR神经网络模型联合预报。(3)基于具有随机性的实测数据,运用时间序列模型进行预报,其精度比神经网络模型高。缺点是模型建立过程和程序编制较为复杂。(4)当有少量实测数据时,灰色预报模型也可实现较高精度的预报。但是如果数据存在噪音,则预报精度会受到影响,甚至达不到合格要求。
[Abstract]:Deformation monitoring, data analysis and prediction are very important for the safety of engineering construction. Through consulting and studying the building deformation monitoring technology and its data analysis and processing related to the literature data forecast, first of all, the significance and purpose, contents and methods of deep foundation pit and building deformation monitoring are studied. And the main contents of the technical design of the monitoring scheme are summarized. Then several kinds of commonly used deformation prediction models are described theoretically. Finally, combining with the data of four cases of building engineering settlement monitoring, MATLAB curve fitting model, neural network model, time series model and grey forecast model are used to forecast, and the results are compared with the actual monitoring data. The conclusions are as follows: (1) in the curve fitting model, the cubic spline interpolation model (cubic spline) has a certain prediction effect. When the settlement is large or the settlement velocity is fast, it can be combined with other models to forecast. The cubic Hermitian interpolation model (shape-preserving) and the polynomial model (Polynomial) are not suitable for subsidence monitoring and forecasting. (2) in the neural network model, the prediction effect of the GR neural network model is better than that of the BP neural network model and the RBF neural network model, and the prediction effect of the BP neural network model is better than that of the BP neural network model. The prediction accuracy of the network model is slightly higher than that of the RBF neural network model. The two models can be combined with the gr neural network model. (3) based on the measured data with randomness, the time series model is used to forecast, and the accuracy is higher than that of the neural network model. The disadvantage is that the process of model establishment and programming are more complicated. (4) when there are a small number of measured data, the grey prediction model can also achieve a higher precision prediction. However, if the data is noisy, the prediction accuracy will be affected, or even meet the requirements.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU433

【参考文献】