【作者】 张凯；

【导师】 刘铁新；

【作者基本信息】 大连海事大学 ， 工程硕士（专业学位）， 2022， 硕士

【摘要】 随着国民经济的快速发展,越来越多的工程建设涉及到深基坑施工。深基坑的变形是随施工进度逐步变化的过程,而监测数据可直接反映相应部位岩土体的受力和变形。基于监测数据分析,准确预测深基坑的变形趋势,进而及时了解工程整体的安全状态,对保证工程安全具有重要的工程价值和科学意义。本文以深圳市某深基坑工程为依托,综合考虑不同类型、不同部位的监测数据,对深基坑工程的变形趋势预测和安全性评估做了相关研究,具体工作如下:(1)依托深圳市某深基坑工程建设项目,以现场实测试验和采集工作为基础,系统地阐述了深基坑工程监测方案的设计,通过绘制时间数据变形曲线,深入挖掘各项监测项目的变形规律。同时,利用分段线性插值方法对原始监测数据进行了插值处理,并借助小波变换技术对变形数据进行了分解和重构。(2)针对现有变形预测方法稳定性差、预测精度不高的问题,在监测数据序列预处理的基础之上,利用遗传算法(GA)对支持向量机(SVM)的关键参数进行寻优,建立了基于小波变换的自动回归滑动平均算法(GASVM-ARMA)单步和多步滚动预测模型。通过与GASVM模型预测结果的对比分析发现,组合后的模型预测精度更高,对于复杂条件下非线性问题的处理适用性更强。(3)针对深基坑安全性评估缺乏系统、定量考虑的问题,基于空间插值技术提取的代表性变形监测数据,提出了一种改进的熵权-层次分析模糊综合评价模型,并根据新划分的安全状态5级评定指标,从权重角度对不同类型、不同部位的监测数据进行客观地整体性考量,研究深基坑工程的稳定性情况。通过对实际工程的安全性评估与结果对比,验证了该方法的可行性。(4)基于MATLAB-GUI平台开发了一套相应的深基坑工程安全性评估系统,实现了变形预测和安全性评估过程的可视化,使得预测和评估的结果能够以数字和图片的形式进行展示,有助于安全管理者的分析和决策。更多还原

【Abstract】 With the rapid development of national economy,more and more engineering construction involves deep foundation pit construction.The deformation of deep foundation pit is a process that changes gradually with the construction progress,and the monitoring data can directly reflect the stress and deformation of rock mass in the corresponding position.Based on the monitoring data analysis,it is of great engineering value and scientific significance to accurately predict the deformation trend of deep foundation pit,and then timely understand the safety state of the whole project.Based on a deep foundation pit project in Shenzhen,this thesis comprehensively considers the monitoring data of different types and different positions,and researches on the deformation trend prediction and safety assessment of deep foundation pit engineering are carried out.The specific work is as follows:(1)Relying on a deep foundation pit engineering construction project in Shenzhen,based on the field test and collection work,the system This paper expounds the design of monitoring scheme for deep foundation pit engineering,and digs out the deformation laws of various monitoring items by drawing the deformation curves of time data.At the same time,the piecewise linear interpolation method is used to interpolate the original monitoring data,and the wavelet transform technology is used to decompose and reconstruct the deformation data.(2)Deformation prediction method in view of the existing stability problem,the forecasting accuracy is not high,on the basis of monitoring data sequence pretreatment,using genetic algorithm(GA)to the key parameters of support vector machine(SVM)to carry on the optimization,set up automatic regression moving average algorithm based on wavelet transform(GASVM-ARMA)one step and multiple step rolling forecast model.Compared with the prediction results of GASVM model,it is found that the combined model has higher prediction accuracy and stronger applicability to nonlinear problems under complex conditions.(3)For deep foundation pit safety assessment quantitative problem,lack of system,based on spatial interpolation technique to extract the representation of the deformation monitoring data,puts forward an improved entropy weight-hierarchical analysis and fuzzy comprehensive evaluation model,and according to the new division of safety state evaluation index,level 5 from the Angle of the weights of different types,different parts of the monitoring data of holistic consideration objectively,The stability of deep foundation pit is studied.The feasibility of the proposed method is verified by the practical engineering safety evaluation and comparison.(4)Based on MATLAB-GUI platform,a set of safety evaluation system for deep foundation pit engineering was developed,which realized the visualization of deformation prediction and safety evaluation process.The results of prediction and evaluation could be displayed in the form of numbers and pictures,which was helpful for the analysis and decision of safety managers.更多还原