海堤渗压神经网络分布模型的建立及规律分析

发布时间：2018-06-19 07:18

本文选题：海堤渗压 + 神经网络　；参考：《上海交通大学学报》2014年11期

【摘要】：为有效揭示海堤渗压的分布特征,掌握其在临海工作环境下的特殊规律,在利用神经网络建模优点的同时,采用多测点渗压监控信息,并将测点坐标因素加入到输入层,综合前期潮位因子、积分型降雨因子、时效因子,形成海堤渗压神经网络安全监控分布模型结构,以实测信息进行建模训练计算;在获得合理训练结果基础上,根据输入层因子补充插入坐标样本,获得不同位置的渗压模型值及渗压分布曲线.文中以广义回归神经网络为例,结合浦东海堤实测资料,以实例说明以上述方法在神经网络不提供显式的情况下,建立可获得分布曲线的监控模型,并以此对海堤渗压分布规律特色加以分析.
[Abstract]:In order to effectively reveal the distribution characteristics of seepage pressure of seawall and master its special law in the working environment of seawall, using the advantages of neural network modeling, the monitoring information of seepage pressure of multiple measuring points is adopted, and the coordinate factors of measuring points are added to the input layer. Combining the pre-tide level factor, integral rainfall factor and time-effect factor, the security monitoring and distribution model structure of seawall seepage pressure neural network is formed, and the modeling training calculation is carried out based on the measured information, and on the basis of obtaining reasonable training results, According to the input layer factor, the values of osmotic pressure model and the osmotic pressure distribution curve were obtained. Taking the generalized regression neural network as an example, combined with the measured data of Pudong seawall, it is illustrated that the monitoring model of the distribution curve can be obtained by the above method without the explicit expression provided by the neural network. The characteristics of seepage pressure distribution of seawall are analyzed.
【作者单位】：合肥工业大学土木与水利工程学院;上海交通大学船舶海洋与建筑工程学院;
【基金】：国家自然科学基金项目(50979056)资助
【分类号】：U656.314

【参考文献】