多测点多方向BP网络模型在大坝变形监测中的应用  

Application of Multi-point and Multidirectional BP Network Model in Dam Deformation Monitoring

作　　者：柴丽莎[1] 戚丹[1] 吴浩[1] 

机构地区：[1]河海大学水利水电学院,江苏南京210098

出　　处：《水电能源科学》2014年第3期94-97,共4页Water Resources and Power

基　　金：国家自然科学基金项目(51139001);河海大学水文水资源与水利工程科学国家重点实验室专项基金(20095860120)

摘　　要：鉴于大坝变形监测资料分析是大坝结构性态安全评价与预报的重要手段,针对单测点模型存在的缺点,建立了既考虑坝体不同方向的位移又考虑空间多个测点分布的多测点多方向位移模型,并利用BP神经网络较强的非线性映射能力,直接选取了对大坝变形有较大影响的自变量因子,解决了在建立大坝多测点多方向传统模型时自变量因子数众多、计算工作量大等问题。实例应用结果表明,多测点多方向BP网络模型可反映大坝变形的分布及变化规律,可见采用BP神经网络建立大坝多测点多方向变形监测模型具有可行性和有效性。Data analysis in dam deformation measuring is an important means of dam safety evaluation and prediction. Aiming at the shortcoming of single point model, this paper establishes multi-point and multidireetional displacement model, which considers both of different directional displacement of dam body and measure points distribution in space. By using strong nonlinear mapping ability of BP neural network, independent variable factors for influencing dam deformation were chosen to solve problems of establishing multi-point and multidirectional traditional model, such as many independent variable factors and large computing workload. An application example shows that the proposed model truly reflects the distribution and change law of dam deformation. Thus, it proves that the multi-point and multidirectional model of dam deformation measuring established by BP neural network is feasible and effective.

关 键 词：多测点多方向位移模型 BP神经网络 大坝 变形监测 

分 类 号：TV698.11[水利工程—水利水电工程]