EXPERIMENT ON THE SHIELD TUNNEL DAMAGE IDENTIFICATION BASED ON ARMA MODEL RESIDUALS

Wen Fu,Hui Luo,Shengnan Wang,Hongping Zhu
2016-01-01
Abstract:As more and more metros are being constructed in the whole country to ease the traffic pressure, the construction and operation safety of the shield tunnel becomes a highlight concerning issue. A detection method based on autoregressive moving average (ARMA) model is proposed in this study. First, data pre-processing is conducted by partition and normalization to avoid noise. Secondly, the data processed is calculated to determine the order of time series model and then establish the ARMA model. Finally, based on the coefficients of ARMA model, a damage indicator (DI) is defined for damage identification and localization. In this paper, the validity of the proposed method is verified by the physical experiment. The physical experiment is a scale model of aluminium pipe simulating the segment of shield tunnel and gathering the vibration signal of acceleration. Finally, two pieces of data are handled to detect and localize. The results indicate that the proposed method is capable of detecting and localizing.
What problem does this paper attempt to address?