Deep Excavation in Stantiago, Chile

Measured vs. Theoretical Ground Anchor Loads for a 21.2m Deep Excavation in Santiago, Chile

Occasionally we get a chance to compare measured deep excavation performance against different methods. This 21.2m deep excavation case history in Santiago, Chile compares various conventional methods against measured ground anchor loads.

The excavation was constructed in competent cohesive gravels and was supported by two levels of ground anchors and 0.8m diameter reinforced concrete piles spaced at 3.2m. Reported displacements were under 2.0 cm. A 2m pile embedment below the final excavation subgrade was provided.

Various popular methods of conventional analysis were examined. In all cases wall to soil interface friction was ignored.

Most conventional methods produced reasonable agreement and came within 10% of the maximum measured loads. The PECK methods with the tributary height approach underpredicted reactions by as much as 25%. The CALTRANS approach with FHWA pressures overpredicted the upper ground anchor load by 17% and underpredicted the lower ground anchor load by 20%. The method producing the closest match was the German EAB approach which came in remarkably close.

The approach of using 50% of active and 50% of at-rest earth pressures severely overpredicted support reactions vs. the actual measured loads.

A critical aspect in case histories like this is the actual applied ground anchor prestress. In this project, the applied prestresses were in closer agreement to the German EAB approach. Should prestress levels had been applied in consistency with CALTRANS loads, measured ground anchor reactions could have been different.

With all due respect to German engineering, this case history clearly demonstrates the importance of evaluating different deep excavation analysis methods.