A nonlinear multiple regression model for load identification of shield machine combined with mechanical analysis

The load identification of the shield machine is presented in this paper by introducing the mechanical analysis of shield excavating into the nonlinear multiple regression of on-site data. The analysis on mechanical characteristics of shield-soil system can decouple the nonlinear multi-parameter problem of load, so it is great helpful for the regression process to establish a load model. Then a load prediction model is obtained based on a group of on-site data of subway project in China. In addition the test combined with another group of independent data which is not involved in the regression is carried out. The result shows that the predicted load using the method proposed in this paper agrees well with on-site recorded data.