基于极限学习机的大坝多测点模型分析

Multiple Points Dam Analysis Model Based on Extreme Learning Machine

与单测点模型相比,大坝多测点模型能估计各测点间的空间联系,通过一个模型反映大坝整体变形态势。针对传统线性多测点回归模型中因子众多,回归方程中自变量多重共线性的问题,提出利用极限学习机(Extreme Learning Machine,ELM)建立非线性大坝多测点位移模型。实例表明,基于ELM的大坝多测点变形预报模型可在保证建模精度的基础上有效提高建模运算效率。

Compared with single model,multiple points dam model can reflect the connection of each deformation point on a dam and make prediction about the whole dam with one model.To solve the problem of multicollinearity in the traditional linear regression model,Extreme Learning Machine are proposed in the modeling of dam deformation.The results of the case study in this paper show that the multiple points dam model based on ELM isstable and much more efficient.