Research On The Prediction Of Tunnel Vault Settlement Based On Gray DNGM(1,1)

In the process of tunnel excavation,the monitoring and measurement work has played an extremely important role in ensuring the safety of tunnel construction.The stability analysis of the stress and deformation of the tunnel surrounding rock and its supporting structure can verify the previous construction methods and The scientific rationality of supporting measures can further realize the purpose of optimizing design and guiding construction.In the process of deformation analysis for the settlement of the tunnel vault,the conventional regression analysis method needs to collect a large amount of deformation monitoring data before it can realize the fitting analysis of the overall deformation trend of the tunnel.The advantage of the model can realize the real-time dynamic deformation prediction during the monitoring process of the tunnel surrounding rock deformation,which is helpful to grasp the deformation law of the surrounding rock in time and detect abnormalities early,which is of great significance for ensuring the safety of tunnel engineering.This paper adopts the method of combining gray prediction theory model with practical engineering application,and applies the gray DNGM(1,1)model to a railway tunnel monitoring section vault settlement data for gray modeling prediction research.The main research results are as follows:(1)Based on the gray system related theory,the tunnel vault settlement value is predicted based on the classic GM(1,1)modeling,and the modeling method is optimized by the method of equal-dimensional and new-information dynamic prediction.The result is It shows that compared with the traditional modeling method,the equal-dimensional and new-information gray modeling prediction method has increased the prediction accuracy by 8.62%,14.82% and 18.28% in the three selected modeling dimensions.(2)Apply the equal-dimensional and new-information DNGM(1,1)model to the prediction of tunnel vault settlement deformation,and compare and analyze the prediction results with the classic GM(1,1)model under the same conditions.The average relative error is from The original 7.482% was reduced to 3.996%,and a better prediction effect was achieved in the early and mid-term when the deformation was severe.Then,in view of the characteristics that the selection of different modeling dimensions will have a greater impact on the deformation prediction results,through the comparison and analysis of the prediction accuracy of the 4-11 modeling dimensions of the two models,6-8 with higher prediction accuracy are given.Select the recommended interval for modeling dimensions.(3)Aiming at the non-stationary characteristics of the volatility of the tunnel settlement deformation value,the method of defining the upper and lower limits for the gray DNGM(1,1)model is used to deal with the prediction errors that exist in the gray modeling process.The model has a certain prediction width,so as to realize the expansion of the prediction form of the gray prediction model on the time sequence to the prediction area.Combining the relationship between the gray upper and lower limits and the tunnel deformation control value,from the perspective of preparing for the corresponding engineering countermeasures in advance,it has a certain reference for further providing the early preparation time for the tunnel deformation warning.