The GA improved ANN-synergy-bifurcation model method on slope slide forecasting

Abstract

Based on friction law of single state variable, this paper establishes a nonlinear synergy-bifurcation model to forecast the sliding time of land slope. Then, by using of this model, combining with the displacements in-situ of slope, the sliding time of the slope can be calculated. Especially, the GA improved BP Neural Network is used firstly to forecast part of the future displacement series which can be got a more accurate result being improved by GA, then, using the more accurate future displacement series into the nonlinear synergy-bifurcation model, the sliding time of the slope can be deserved. This nonlinear synergy-bifurcation model belongs to a kind of scroll forecasting method which can avoid the continuing accumulation of forecasting displacement error that leads the forecasting data departing away from the true ones. At last, validated by a real case of slope sliding, the GA improved ANN-synergy-bifurcation model forecasting method can tell a more exact displacements of sliding, as well as work out the sliding time of slope. It can be used to forecast, prevent and rehabilitate the corresponding geo-disasters.