摘要
海堤渗压监测序列是反映堤身状况与规律的重要信息,为把握海堤渗压序列变化特点、进行有效分析预测,对浦东海堤实测渗压曲线进行了广泛分析,归纳出其多周期变化的特征,并从谐量分析角度给出针对周期特征变化的渗压监测模型因子,结合神经网络特点,提出相应的神经网络监测模型输入层单元形式,以广义回归神经网络为例建立监测模型,以监测资料和实例说明多周期特征对监测分析的作用,以及与神经网络模型相结合的具体应用.
Sea wall osmosis pressure monitoring sequences are important information reflecting sea wall state. In order to know osmosis pressure variety characteristic and do analysis and forecasting, Pudong sea wall monitoring osmosis pressure curves were studied to conclude their multi-periodic characteristic. Base on this, the factors of osmosis pressure monitoring model were presented referring to harmonic quantity analysis. Furthermore, the relevant input units of artificial neural network were advised, and the General Regression Neural Network was used as example to establish monitoring model with observed data. With the instance, the effect of multi-periodic characteristic on monitoring analysis has been shown; its application process together with artificial neural network to establish monitoring model is presented.
出处
《应用基础与工程科学学报》
EI
CSCD
2010年第2期330-336,共7页
Journal of Basic Science and Engineering
基金
国家自然科学基金项目(50979056
50609014)
上海市重点学科建设项目(编号B208)
关键词
海堤监测
渗压
周期特征
神经网络模型
sea wall monitoring
osmosis pressure
periodic characteristic
neural network model
