摘要
基于BP神经网络对国内某抽水蓄能电站下水库大坝主、次堆石区材料参数进行反分析,以面板堆石坝的实测沉降为基础,将堆石坝的沉降分解为瞬时变形沉降和流变沉降,从沉降监测值中减去蠕变分量,分解出沉降荷载变形分量的相对值,再从有限元计算得到的沉降变形绝对值中扣除检测仪器安装前的沉降值,得到沉降变形相对值,进行材料参数的反演计算。然后根据反演得到的堆石区材料参数进行有限元计算,通过计算所得结果与现场实测数据相一致,表明反演参数的合理性和面板堆石坝坝体材料参数反分析方法的有效性。
BP neural network is used to the back-analysis of the material parameter of the primary-secondary rock-fill dam area of a pumped-storage power station,on the basis of the measured settlement of the face rock-fill dam,the settlement of rock-fill dam is decomposed into instantaneous deformation and rheological settlement,subtract creep deformation from the monitored settlement value,then deduct the settlement value before the installation of detection equipment from the absolute value of the finite element calculation. Thus,conduct the inverse analys s of material paramet rs.Finally,starting the finite element calculate according to the material parameters by inverse analysis.Discussing the rationality of inversion data and the effective of back analysis method in face rock-fill dam by comparing calculation results with the field measured data,to provide the scientific basis and reference for rock-fill dam engineering design and site construction.
出处
《电子测试》
2013年第10X期35-37,共3页
Electronic Test
关键词
BP神经网络
反演
累积沉降量
监测沉降量
蠕变分量
BP neural network
back analysis
the accumulative settlement
the relative value of settlement
creep deformation