Correlation Coefficient Control For Subgrade Settlement Curve Fitting

摘要

The measured settlement curves of high-speed railway are presented with the characteristic of "small scale, large fluctuations", which is difficult to reach the requirements of curve regression coefficient with 0.92 on specification. This was done by adopting the method of mathematical simulation and hyperbola fitting to study the change rules of correlation coefficient on subgrade simulated settlement curve, under different conditions. From the analysis result, smaller value of parameter a and b, will have great influence on curve correlation coefficient, otherwise, the parameter c becomes the major factor; When c value is greater than 1.5mm, the parameter b becomes the major factor; Considering the premise, it is recommended that in order to observe the total settlement amount for working point <10(5)mm that can determine the post-Settlement deformation trend condition, the correlation coefficient of the regression curve minimum standards can be relaxed from 0.92 to 0.8.