摘要
清江隔河岩水电站自1993年蓄水发电至今已运行近20年,各项性态趋于稳定。根据清江隔河岩大坝近年来的大坝安全监测资料,主要针对拱冠梁15号坝段的位移,在基于物理推断分析的基础上,采用统计模型和反向传播(BP)神经网络模型进行了分析与研究。通过对2种模型分析的结果进行比较,应用BP神经网络模型进行分析具有更高的拟合精度和预测精度。分析结果表明:隔河岩大坝拱冠梁径向位移与上游水位呈正相关,与气温呈显著的负相关;拱冠梁切向位移与上游水位和气温无明显的关系,且径向位移非常小,拱冠基本呈对称状态,符合拱坝变形规律。
Geheyan Hydropower Station on Qingjiang River has been in service for nearly 20 years since it begun to impound water and generate power in 1993,and every parameter tends to be stable.According to the monitoring data of Geheyan Dam in recent years,the statistical model and BP neural network model are employed to analyze the displacement of the crown cantilever of dam section No.15 based on the physical inference analysis.By comparing the results of the two models,it is found that the results of the BP neural network model exhibit higher fitting precision and prediction precision.The analytic results show that the radial displacements of the crown cantilever have a positive correlation with the upper water level and an obvious negative correlation with the air temperature.The tangential displacements of the crown cantilever have no obvious correlation with the upper water level and the air temperature.The values of the tangential displacements are very small.The crown cantilever is in a symmetric state and agrees with the deformation rules of arch dams.
出处
《水电自动化与大坝监测》
2010年第5期58-62,共5页
HYDROPOWER AUTOMATION AND DAM MONITORING
关键词
大坝
拱冠梁
环境量
统计模型
BP神经网络模型
dam
crown cantilever
environmental variable
statistical model
BP neural network model
