In applying Asaoka’s method to predict the post-construction settlement of a road foundation in Shenzhen

it was found that the predicted results can be different by three times

depending upon the chosen initial time and time spacing. In this paper

the mathematical formulation of Asaoka’s method is analyzed and the effect of the error of fitting parameters on the predicted results is discussed

from which the concept of the error amplification coefficient of fitting parameters is defined. The reliability of predicted results by Asaoka’s method is evaluated. The results show that one of the major reasons that affect the predicted results of Asaoka’s method is that the expression of settlement prediction is very sensitive to the fitting parameter ?1

and the sensitivity increases with the increase of ?1. Hence

without scarifying the accuracy of the linear fitting method

the time spacing is increased

and ?1 can be reduced

so that the influence of the fitting parameter error on the predicted results can be effectively reduced. In predicting the foundation settlement based on Asaoka’s method

the reliability of prediction is not only dependent upon the accuracy of linear fitting

but also on the error amplification coefficient of fitting parameters.