Delayed collapse of cut slopes in stiff clay

Abstract
Delayed failure of railway cutting slopes in stiff clays excavated in the nineteenth and early twentieth centuries has been studied for some forty years. Known failures have generally been deep-seated. More recently failures have been a problem in the slopes of motorway cuttings and embankments, although generally these have been shallow. The average operational strength at failure in these slides has been significantly less than the peak strength, and progressive failure has been postulated as the probable cause of this. Progressive failure can now be analysed using advanced numerical techniques. A series of coupled finite element analyses have been conducted assuming strain-softening soil with properties based on the Brown London Clay, and the results are reported. They show that progressive failure is considerable, and fully explain the observed field behaviour. The delays experienced in the field are also recovered by the analyses. Progressive failure is generated primarily by the high lateral stresses in the soil prior to excavation. The rupture surface generated spreads horizontally from the toe as the soil swells, and differs significantly from the critical surface predicted by limit equilibrium analysis. The average strength on it at collapse is significantly lower than that obtained by back analysis by limit equilibrium methods. It becomes lower as lateral stress increases, but the effect is compensated by the increasing depth of rupture. The roll of the hydraulic surface boundary condition, which is controlled by climate, is important in controlling collapse. The implications of the findings for monitoring and for remedial works in existing slopes is discussed. Delayed failure of railway cutting slopes in stiff clays excavated in the nineteenth and early twentieth centuries has been studied for some forty years. Known failures have generally been deep-seated. More recently failures have been a problem in the slopes of motorway cuttings and embankments, although generally these have been shallow. The average operational strength at failure in these slides has been significantly less than the peak strength, and progressive failure has been postulated as the probable cause of this. Progressive failure can now be analysed using advanced numerical techniques. A series of coupled finite element analyses have been conducted assuming strain-softening soil with properties based on the Brown London Clay, and the results are reported. They show that progressive failure is considerable, and fully explain the observed field behaviour. The delays experienced in the field are also recovered by the analyses. Progressive failure is generated primarily by the high lateral stresses in the soil prior to excavation. The rupture surface generated spreads horizontally from the toe as the soil swells, and differs significantly from the critical surface predicted by limit equilibrium analysis. The average strength on it at collapse is significantly lower than that obtained by back analysis by limit equilibrium methods. It becomes lower as lateral stress increases, but the effect is compensated by the increasing depth of rupture. The roll of the hydraulic surface boundary condition, which is controlled by climate, is important in controlling collapse. The implications of the findings for monitoring and for remedial works in existing slopes is discussed.