Prediction of the displacement in a foundation pit based on neural network model fusion error and variational modal decomposition methods

Prediction of foundation pit displacements is an important part of pit monitoring and early warning system, and accurate prediction of pit displacements can reduce pit disaster accidents. Herein, the Elman neural network model optimized by the mind evolutionary algorithm (MEA) is combined with the variational modal decomposition (VMD) and the fusion error method to accurately predict pit displacements at the top of the slope. Firstly, VMD is used to decompose the time series of pit displacements into trend, period, and random terms, and analyzed the rationality of the decomposition of the sub-displacement sequences; Secondly, the MEA-Elman model is used to predict the trend, period, and random terms and decomposition error in the decomposition of pit displacements, and finally the prediction results of the sub-displacements are superimposed to obtain the total displacement; The measured data of a pit in Qingdao, China, are used to verify the reliability of the model, and the calculated results show that the predictions are consistent with the measured data.