Multi-factorial Analysis on Vault Stability of an Unsymmetrically Loaded Tunnel Using Response Surface Method

Abstract

This study investigated the vault settlement characteristics of an unsymmetrically loaded tunnel which was excavated by annular excavation via core rock support method. Response surface methodology (RSM) was employed to design the experiments, evaluate the results with the purpose of optimizing the value of design parameters for reducing the vault settlement. The parameters such as horizontal distance, step length, tunneling depth, width of core rock, strength of surrounding rocks and support strength were firstly examined, and a second-order polynomial regression equation was then derived to predict the responses of vault settlement. The percentage contribution, validity of model and effects of different parameters as well as their interactions were assessed by analysis of variance (ANOVA). In the order from high to low effect, these parameters are strength of surrounding rocks, support strength, horizontal distance, width of core rock, tunneling depth, and step length. The results indicated that the influence of uncontrollable factors (i.e. strength of surrounding rocks, tunneling depth, and horizontal distance) on the vault settlement can be reduced through the adjustment of controllable factors (i.e. Width of core rock, step length, and support strength). Moreover, the proposed method in this paper was validated with results of field test measurement and simulation calculation which verified its feasibility.

Modeling and Optimization of Charge Materials Ranges in Converter Furnace with Enhanced Passivation Time in Copper Electrorefining Process: A Mixture Design Approach

“…Response surface methodology was proposed employed to design the experiments through the collecting data [15]. It is one of the most promising approaches in the case that the operational factors consider as ingredients of the mixture and the response is a dependent proportions function [16][17][18][19][20]. Low operating costs due to the reduction in the number of required tests and provide the possibility of the simultaneous changes in variables are the main advantages of this approach [21,22].…”