Long-term prediction of tunnel primary lining deformation based on wireless monitoring

Abstract. Due to the harsh construction environment such as blasting excavation and dust, conventional monitoring methods are challenging to be continuously deployed at the primary lining section of the drilling and blasting tunnel, which makes it difficult to capture the large deformation evolution trend and warn the deformation risk. Therefore, this paper takes a biased construction section of Xujiacun tunnel as a site case to implement a wireless sensor network, including gateway and laser distance senor, which is embedded into the primary lining to protect sensors in harsh construction environment for high frequency of primary lining deformation monitoring. Subsequently, a DLinear model is applied, which can be used for the long-term prediction of deformation time-series by splitting deformation sequence into trend sequence and residual sequence. Its prediction performance under different prediction steps is tested, so as to provide accurate and long-term risk early-warning in real-time for construction sites.