摘要
由非主断面观测站求取主断面上的边界角和移动角,目前采用等权最小二乘估计求参后反求,由于厚松散层失水等因素的影响及概率积分法具有边界收敛快的特点,所求边界角和移动角的精度得不到保证。从修正随机模型的角度,采用加权最小二乘估计求参。通过实例,从下沉残差序列、临界变形值和残差中误差等方面分析了等权和加权的拟合效果,得出等权估计的整体拟合效果优于加权估计,加权估计能有效提高求取边界角和移动角的精度的主要结论。
At present, the equal weight least square estimation is reversely used for the determination of the boundary angle and displacement angle of principle section observation station by observation station of non - principle sections. Owing to the water loss of thick unconsolidated layers and the fast convergence of probability integral method, the precision of boundary angle and move angle can not be guaranteed. In this paper, stochastic model was modified, using weighted least square estimation to resolve the parameters. In the present case, the fitting effect of equal weight and weighting were analyzed in the aspects of subsidence residual series, critical deformation value, residual mean square error, etc. The main conclusions were that, the fitting effect of equal weight least square estimation is better than the weighted least square estimation; and the weighted least square estimation can effectively enhance the calculating precision of boundary angle and move angle.
出处
《安徽理工大学学报(自然科学版)》
CAS
2011年第2期31-34,共4页
Journal of Anhui University of Science and Technology:Natural Science
基金
安徽高校省级自然科学研究项目(KJ2011Z085)
关键词
非主断面观测站
边界角
移动角
non -principle section observation station
boundary angle
displacement angle