Abstract:

Hyperbola method, three-point method, exponent curve method, Asaoka method etc. are widely applied to settlement prediction for soft soil ground; but less study of applying to case of small settlement. With regard to the four methods, the small settlement data from the construction site of a passenger dedicated railway line are used to study the influence of starting point of calculation time and span of starting and ending time of the sample on relative error and prediction effect of the last observed values are checked. Comprehensively considering correlative coefficient and relative error, the applicability of every method is studied and the advantages and disadvantages for every method are analyzed. The research results show that the exponent curve method can not fit to the case of small settlement: other three methods all provide better fit to the real data when the settlement data fluctuation is small; however only three-point method and Asaoka method fit better if settlement data fluctuation is not small.

Key words: curve fitting methods, small settlement, ground settlement, ballastless track, relative error, correlative coefficient