混凝土坝安全监控模型数值优化及变位预警指标研究

发布时间：2018-09-04 16:32

【摘要】：大坝原型观测数据处理是其安全监控的重要研究内容，本文针对大坝安全监控模型的拟合残差、分量提取及指标拟定等内容，综合运用统计学方法、遗传算法、人工神经网络、混沌理论、最小二乘支持向量机算法与有限元法等方法，以混凝土坝为研究对象，结合大坝位移原型观测资料，在建立混凝土坝安全监控模型的基础上，研究了监控模型的数值优化方法，探究了大坝主要物理力学参数的反演方法，并给出了大坝变位预警指标的拟定方法。主要研究内容如下： (1)研究了混凝土坝安全监控模型的构建方法，，利用遗传算法优化神经网络算法，融合混沌理论，运用相空间重构等技术，对大坝位移拟合残差进行预测，并将残差预测项作为位移监控模型的混沌因子，据此构建了考虑残差混沌因子的混凝土坝位移混沌混合监控模型，并验证了所建模型的有效性。 (2)探讨了混凝土坝综合弹性模量的反演方法，在此基础上，提出了一种能够反映大坝位移和坝体弹性模量间非线性映射关系的最小二乘支持向量机(LS-SVM)反演算法，并利用MATLAB平台，研制了基于LS-SVM算法的反分析程序。 (3)分析了服役期混凝土重力坝和拱坝的变形过程和转异特征，并在对大坝正反分析的基础上，进一步研究了混凝土坝变位预警指标的拟定方法。 (4)以某在役混凝土重力坝为例，在分析其水平位移变化规律的基础上，基于上述理论与方法，构建了该坝位移统计模型、混合模型及考虑残差混沌因子的混沌混合模型；并结合其典型坝段位移监测资料及正反分析成果，拟定了该坝变位预警指标，为评判大坝安全状态提供了理论依据。
[Abstract]:Dam prototype observation data processing is an important research content of dam safety monitoring. In this paper, the statistical method, genetic algorithm and artificial neural network are used synthetically to solve the problems of dam safety monitoring model, such as fitting residual, component extraction and index formulation, etc. Chaos theory, least square support vector machine (LS-SVM) and finite element method (FEM) are used to study concrete dam. Based on the observation data of dam displacement prototype, the safety monitoring model of concrete dam is established. The numerical optimization method of monitoring model is studied, the inversion method of the main physical and mechanical parameters of the dam is explored, and the method of drawing up the early warning index of dam displacement is given. The main research contents are as follows: (1) the construction method of concrete dam safety monitoring model is studied. The genetic algorithm is used to optimize the neural network algorithm, the chaos theory is fused, and the phase space reconstruction technology is used. The residual error of dam displacement fitting is predicted, and the residual prediction term is taken as the chaotic factor of displacement monitoring model. Based on this, the chaotic mixed monitoring model of displacement of concrete dam considering residual chaos factor is constructed. The validity of the model is verified. (2) the inversion method of composite elastic modulus of concrete dam is discussed. A least square support vector machine (LS-SVM) inversion algorithm which can reflect the nonlinear mapping relationship between dam displacement and elastic modulus of dam is proposed, and the MATLAB platform is used. The inverse analysis program based on LS-SVM algorithm is developed. (3) the deformation process and transition characteristics of concrete gravity dam and arch dam in service period are analyzed, and on the basis of positive and negative analysis of dam, In this paper, the method of determining early warning index for displacement of concrete dam is further studied. (4) taking a concrete gravity dam in service as an example, based on the above theory and method, the variation law of horizontal displacement is analyzed. The statistical model of displacement of the dam, the mixed model and the chaotic mixed model considering the residual chaos factor are constructed, and the displacement monitoring data of the typical dam segment and the positive and negative analysis results are combined to draw up the early warning index for the displacement of the dam. It provides a theoretical basis for judging dam safety state.
【学位授予单位】：南昌大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV698.1

【引证文献】