Failure Behaviour of Cantilever Retaining Walls - Soil-Retaining Wall Interaction

Abstract

The current state of preservation of cantilever retaining walls has attracted considerable attention in the last decade in Switzerland, as destructive tests have detected strongly lo-calised corrosion of the main reinforcement in many walls built in the 1970s. This has been identified as a potential threat that could lead to an unpredictable brittle collapse of the wall, which may cause severe damage to high-traffic roads and even victims. Typically, retaining walls are designed to withstand active earth pressure. This condition implicitly presupposes certain soil deformations, which require the wall to have a sufficient rotation capacity. How-ever, corrosion damage can significantly reduce structural rotation capacity. Therefore, quantifying the earth pressure acting on corrosion-damaged cantilever retaining walls is essential to assess their safety reliably. This work studies the evolution of the earth pres-sure as a function of corrosion-driven wall displacement. Analytical, numerical, and exper-imental analyses are performed to quantify the history of earth pressure, from the construc-tion of the wall to the moment of possible corrosion-induced collapse. The obtained results are generally valid for any problem involving the same failure mode as that resulting from a corrosion of the main reinforcement. The relevant failure mode is identified as a rigid-body rotation around its toe. The limit load is determined using a static and a kinematic solution based on the limit analysis theorems and compared to conventional design methods. This failure mode is further analysed in scaled experiments, where different initial conditions and soil parameters are investigated. Loose, contractive soil requires much larger rotations to reach the residual state than dense soil. In addition, the unloading process is influenced by the initial stress state in the backfill. In uncompacted soil, the initial earth pressure is bilinearly distributed, whereas higher stresses are measured close to the soil surface in statically compacted samples. Slightly larger wall rotations are required to reach the active state in compacted backfills. By imposing the rotation of a single wall section, it is shown how an inhomogeneous distri-bution of the corrosion degree over the wall length can lead to a decreased limit load on the failing wall section due to the stress redistribution occurring in the backfill. Conse-quently, neighbouring sections must withstand increased loads. A numerical framework for quantifying the earth pressure on cantilever retaining walls is developed based on experimental observations and widely known constitutive laws to guar-antee practical applicability. The framework is generally applicable and provides reliable results as it is validated using experimental data. The material behaviour is calibrated through virtual element tests performed using the Level Set Discrete Element Method. In plane strain tests, the mobilised soil strength is higher than in triaxial tests, which confirms the experimental observations. Furthermore, the Level Set Discrete Element Method is used to analyse the earth pressure coefficient at rest, showing a correlation between the coefficient and the peak friction angle, which does not imply causation. Then, the developed numerical models are applied to some case studies. Taking into ac-count a more accurate structural model, it is apparent that actions and reactions can be decoupled to assess the safety of walls, as the precise modelling of the elastoplastic wall behaviour does not significantly influence the earth pressure. Furthermore, the effects of cyclic atmospheric temperature changes are simulated and discussed, considering the im-plications for wall monitoring. Finally, a verification procedure for cantilever retaining walls is proposed, considering the wall and soil behaviour.