Study Of EPB Shield Tunnelling Parameters Of Changsha Metro Passing Beneath Xiangjiang River

With rapid development of the city subway, there are an increasing number of shield tunnels beneath river. It is worthy to study how to determine shield tunnelling parameters reasonably. The study was carried out on base of Changsha metro, which passes beneath Xiangjiang River. Project profiles and tunnelling parameters for each tunnel section were collected and summarized. Distribution characteristics, calculation methods, relationship, influencing factors and prediction of shield tunnelling parameters were researched correspondingly. The main research contents are as follows:(1) Based on the project profiles of Yingwanzhen-Juzizhou section and Juzizhou-Xiangjiangzhonglu section of shield tunnelling of Changsha metro passing beneath Xiangjiang River, typical stratigraphic features of tunnel interval were summarized, and challenges during the tunnelling were analyzed. Additionally, selection of shield, parameter configuration, design features of cutter head and cutter tool were summarized.(2) Distribution of shield tunnelling parameters for different strata in three intervals, including bank, west cross-river and east cross-river intervals, was statistically analyzed and compared. Reasonable range of tunnelling parameters in different sections was proposed, and correlation degree anmong main tunnelling parameters were analyzed.(3) The main tunnelling parameters of total thrust, cutter torque and earth pressure on bulkhead were calculated for Xiangjiang River tunnel. The characteristics of the relationship between total thrust and cutter torque were analyzed. Furthermore, the relation between specific thrust force and specific torque and its variation with stratigraphic condition for each section were investigated. The main influencing factors of tunneling parameters for the shield tunnelling were summarized.(4) Based on the site tunneling parameter data of Xiangjiang River tunnel, multiple linear regression method was adopted to derive the prediction models of penetration rate and cutter torque with evaluation of their superiority-inferiority. Besides, the artificial neural network prediction model of shield tunnelling parameters based on the geological parameters were established. Two models, including BP neural network and RBF neural network, were applied to predict main shield tunnelling parameters for different geological conditions by properly selecting the training samples. Finally, the neural network prediction model was assessed according to the prediction results.