Novel High-Precision Grey Forecasting Model and its Application of Deformation of Tunnel Surrounding Rock

Abstract

Accurately estimating the deformation of tunnel surrounding rock is a very important work for surveyors, and we adopted grey model as a forecasting means because of its fast calculation with as few as four data inputs needed, however, the original GM (1, 1) model is not fit for dynamic and long data prediction. For this purpose, we propose a novel approach to improve prediction accuracy of GM(1,1) model through optimization of the initial condition and adoption the technique of rolling modeling, the new forecasting model termed RnGM(1,1). By using the optimized model to analyze and predict the deformation of tunnel surrounding rock and comparing this optimized model with other models, we finally draw a conclusion that this optimized model is able to improve the precision of prediction and therefore can be applied to deformation data analysis.

An Effective Hybrid Approach for Processing Deformation Monitoring Data

“…Now, the most simple and frequently employed grey forecasting model is the GM (1, 1) model that denotes a single variable and first-order differential equation model, is able to describe the interior character and the developing trend of discrete deformation data serials. It is unlike the regression analysis and other models, which needs accurate statistical distribution, and the determined prediction that needs many parameters [4][5][6][7][8]. Hence, the GM (1, 1) model can be used to rightly analyze the trend of deformation data serials which are influenced by various factors.…”

“…In order to improve the GM(1,1) model accuracy of forecast, several improved versions of GM(1,1) model have been discussed in the literature [4,5,7,8].In this project, artificial neural network have been used to model the stochastic phenomenon of the deformation. Currently, there are more than forty types of neural network models are availabl for deformation data processing, a multi-layer back propagation (BP) structure is selected.…”

Enhancing GNSS Deformation Monitoring Forecasting with a Combined VMD-CNN-LSTM Deep Learning Model

“…Multiple regression analysis is simple and easy to use with high accuracy [24][25][26][27][28], but it has issues with multicollinearity and lacks causal inference capability. Genetic algorithms are suitable for handling complex problems and situations lacking mathematical expressions [29][30][31][32][33], but they require special definitions, and parameter adjustments, and cannot guarantee the quality of solutions. Machine learning methods combined with historical data for prediction may amplify errors in high-dimensional data, leading to reduced training accuracy, such as LSTM [34][35][36], BP [37][38][39], CNN [40][41][42], and so on.…”

Research of the Surrounding Rock Deformation Control Technology in Roadway under Multiple Excavations and Mining

“…rough the comprehensive numerical simulation, three-dimensional laser scanning technology and field measurement method, Guo et al [12] analyzed the deformation characteristics of the roadway roof influenced by mining. Jia et al [13] concluded that there was a large plastic failure depth in the coal side of the roadway during mining, and the maximum failure depth of the roadway slope was inclined to the roadway side [14][15][16]. On this basis, the high-extension combined bolt supporting technology was proposed [17].…”