Computational Back-Analysis during Excavation of the Regional Connector Cavern

This paper presents a computational modeling effort carried out during the construction of an underground cavern to predict the ground and structural deformations. The Regional Connector Transit Corridor (RCTC) project consists of an 89-m long, 17.7-m wide, and 11-m high span crossover cavern constructed beneath critical infrastructure using the sequential excavation method (SEM) at relatively shallow depth (15 m below ground surface) in downtown Los Angeles. The cavern construction involved a complex three-drift, two-sidewall configuration and was excavated in the Fernando formation, a weak clayey siltstone. A 3D finite-difference model was developed to simulate the SEM construction explicitly considering 3D multi-face excavation effects and ground-structure interactions. Due to limited field/lab data available, the estimation of geotechnical parameters to be used in the model contains great uncertainty. During SEM construction, as field measurements were obtained continuously, the uncertain parameters were updated through a progressive back-analysis procedure. The predictions of the ground and structural deformations were updated accordingly. The performance of the analysis is demonstrated by comparing the computational results to the field observations. This provides a deliberate application of the observational method in urban SEM construction by quantifying and updating the knowledge of risk along the entire construction process.