Displacements on Existing Tunnels from Adjacent Deep Excavation in Los Angeles
A soldier pile deep excavation in Los Angeles was excavated next to twin tunnels. The original shoring was designed by Cefali & Associates Inc. The excavation consisted of two stepped walls and was reported by and was reported by Yao et. al “Prediction and Observation of Los Angeles Metro Red Line Subway Tunnel Response Due to Adjacent Deep Excavation”, 2016. Drilled soldier piles at 8ft spacing were installed and were supported by prestressed ground anchors. The lower wall was braced by one tieback level and one inclined steel raker. The shoring system was designed with a 25 H (in psf) apparent earth pressure diagram and displacements at the tunnel had to be limited to under 0.5 inches at any location. Figure 1 presents a typical cross section that was re-examined in this article.
Figure 1: Typical Excavation Section
Table 1: Soil and Rock Parameters from Geotechnical Report
B, MODEl REALNALYSIS WITH DEEPEX
The excavation was reanalyzed with DeepEX. Because the original sizes were not provided, some reasonable assumptions were made regarding beam and tendon numbers. Figures 2 & 3 presents the finite element results from the DeepEX model. Lateral displacements of approximately 0.4 inches at the tunnel were computed which were very close to the actual reported displacements.
Figure 2: DeepEX FEM Model Results with Lateral Wall Displacements
Figure 3: Closeup Horizontal Tunnel Displacements
Figures 4.1 through 4.3 compare wall bending envelopes for the FEM, non-linear Winkler spring method, and Limit-Equilibrium approach with the CALTRANS method. In the LEM approach, the 25% of the passive pressures from the upper wall was applied as a surcharge on the lower wall, with some minimal adjustments due to distance. Bending moments between the three methods for the lower wall are very similar, but greater differences exist for the upper wall. Maximum support reactions between the FEM and NL methods were at 26.9 klf and 26.3 klf respectively, while the LEM approach produced 17.6 klf. Due to the highly cohesive soils, the FHWA approach produces smaller apparent pressures vs. the 25 H apparent diagram. In such conditions, a reasonable approach to determining LEM apparent pressures can be to include portion of the at-rest earth pressures and active.
Figure 4.1: FEM Wall Bending Results
Figure 4.2: NL Wall Bending Results
Figure 4.3: LEM Wall Bending Moments
The finite element method can reasonably capture the impact of excavations on existing tunnels provided though that one takes care of modeling details. Engineers still need to be vigilant as a few parameters can lead to drastically different results. Winkler and LEM methods reasonably matched FEM load magnitudes and bending moments within reason. In LEM when deformation control is required in frictional soils with considerable cohesion, apparent pressures should consider portion of at-rest and not solely rely on FHWA apparent pressures.
With DeepEX we were able to quickly create this case history and examine three different analysis methods.