Research on Deformation Forecast of Deep Foundation Pit Based on Non-equidistant Monitoring Data

This paper introduces a solution of unequal interval deformation prediction by using non-equidistant Grey Model (1,1) (GM(1,1)) which is an effective tool to study uncertain system and can establish a mathematical model based on a spot of data. Firstly, the Grey System Theory is introduced briefly and the modeling process of non-equidistant GM(1,1) is shown. Secondly, the idea of real-time forward simulation is emphasized, which can improve prediction accuracy greatly. Thirdly, an algorithm for unequal interval deformation prediction is studied and realized by MATLAB, and a data fitting problem is verified as well for algorithm's correctness. Fourthly, a comparative analysis between the predicted and actual data is conducted based on a practical engineering, the result shows that the unequal interval grey model is effective whose prediction result is close to the reality and the feedback monitoring information is very important for the accuracy of the prediction. Then, a rebuilding of grey background level based on Lagrange interpolation is carried out. A better accuracy of simulation is got. Finally, some useful suggestions on prediction accuracy enhancement and some problems need to be noticed are mentioned.