Small scale model studies were undertaken in association with the University of Strathclyde to investigate the relation between strain and mobilised strength in both unreinforced and reinforced soil structures. These studies clearly demonstrated that the controlled development of strain fields in reinforced soil backfill during construction permitted the soil and reinforcement strengths to be fully utilised and produced a significant reduction in post-construction wall movements. As the results had been obtained on small scale laboratory models, it was decided to test their validity by observing the behaviour of large scale walls constructed in the retaining wall facility at the Transport Research Laboratory. the Transport Research Laboratory. The data obtained from the large scale walls constructed using different methods confirmed the findings of the small scale model wall studies. In particular the data showed that, for both unreinforced and reinforced soil walls, the magnitudes and distributions of lateral earth pressures on the walls are greatly influenced by the amount and mode of lateral boundary yielding during and after construction. It was also shown that filling and compaction processes can "lock-in" significant stresses into the backfill. Thus, to ensure reliable estimates of the lateral earth pressures associated with unreinforced or reinforced backfills, consideration must be given to the anticipated movements of the structure and backfill and to the influence of the construction process. The separate effects of construction and boundary yielding on internal behaviour are not easily quantified, but the application of the yielding wall technique offers a relatively simple and economic means of ensuring greater effectiveness of the soil and reinforcements and for controlling post-construction wall movements. The report also describes the use of finite element analysis to simulate the effect of yielding in ameliorating the horizontal stresses on walls. The CRISP package (Britto and Gunn, 1987) was used in a two- dimensional form, with a Mohr-Coulomb constitutive model being adopted for the soil. The analysis was compared with results from the large scale study. These comparisons highlighted a number of problems with the CRISP analysis, in particular the inadequacy of the analysis in dealing with hysteresis loops, such as those that occur with compaction, and an inability to model very soft layers. However, the finite element analysis was useful in providing further support for the view that the yielding wall technique offers an economic and convenient solution for reducing the lateral forces on retaining structures. (A)
Abstract