Information Feedback Analysis in Deep Excavations

Abstract

Deep excavation often causes displacement of adjacent structures. Hence, necessary construction measures must be taken in order to minimize such disturbances. Appropriate construction measures depend on effective and reliable estimation of the induced ground movement during an excavation. This paper presents a systematic procedure, referred to as “information feedback analysis,” which is used to predict excavation-induced deformation by collecting field information, such as displacements. With the use of optimization algorithms, the analyses result in a “best set” of soil parameters. These back-calculated soil parameters are then used to predict the deformation in the subsequent stages, one stage at a time, until the end of the excavation, which result in additional updated information continuously entered into the system, and hence, the prediction becomes progressively more and more accurate. This study has shown that the proposed approach exhibits at least two advantages over the conventional analysis. First, the use of field instrumentation to estimate geotechnical parameters allows the engineer to account for the global response of a soil-structure system. Second, since the information is collected throughout the length of construction, any departure from the original design should be reflected by the updated information, while the conventional analyses are conducted in the design stage only and always assume that construction proceeds as planned.