Abstract:

By taking the monitored horizontal displacements of the supporting structure in a deep foundation pit as samples, a prediction model of time series is established based on the back propagation （BP） neural network. Then, by using the Sim function, a simulation of the network is performed on the Matlab platform. Moreover, by employing the Plot function, an error analysis of the simulation is carried out. Thus, the horizontal displacement of the whole supporting structure is predicted. It is found that the predicted data accord well with the monitored and designed ones, thus showing that the proposed prediction method is feasible.

Key words: deep foundation pit, deformation, back propagation neural network, prediction