摘要
由于大坝变形影响参数较多且其变化多样,常见大坝变形预测方法有明显的局限性。为此,通过实例,在比较多元线性回归算法和BP神经网络算法在大坝变形预测中的优缺点的基础上,采用融合方法,进行大坝变形预测,取得了较好的预测效果,预测分析表明:一方面提高了预测精度,为大坝变形预测提供了一种可行的新方法;另一方面也为大坝变形关键影响因素分析,提供了理论依据。
Since dam deformation has impact on many dam parameters and different variations,traditional dam deformation prediction method has an apparent limitation.Therefore based on example and comparison with advantages and disadvantages of multiple regression analysis and BP neural net algorithm,dam deformation is predicted combing tow algorithms,and good prediction result is achieved.On one hand it enhances the accuracy and provides a practical method for dam deformation prediction,on the other hand it provides theoretical basis for the key factors analysis of dam deformation.
出处
《黑龙江水利科技》
2013年第2期1-3,共3页
Heilongjiang Hydraulic Science and Technology
关键词
大坝
神经网络
变形预测
回归分析
混合算法
精度提高
dam
neural net
deformation prediction
regression analysis
hybrid algorithm
accuracy increasment
