BEHAVIOUR OF RETAINING WALLS UNDER DYNAMIC LOADS

Abstract

The current practice of considering tha dynamic behaviour of retaining walls is to take into account an increase in the lateral earth pressure as calculated from the modified Coulomb formula. This has serious limitations. In this study, model tests were conducted to study the effect of the flexibility of high retaining walls on the static and dynamic earth pressures. Also, the effect of the damage potential of different ground motions was studied by shock type loading. Three types of 'small prototype' walls were used in this study; l) a steel cantilever wall 1.0 m high to represent the effect of flexural bending of the wall only 2) a steel and brick wall 1.0 m high with provision for rotation, to separately investigate the effect of the movement of wall foundations and 3) a rigid wall 2.0 m high to study the problem at lower acceleration levels and steady-state vibra tions. The walls 1 and 2 above, were housed in a horizontal shake table excited by the impact of a loaded pendulum. The wall 3 was placed in a pit dug in the ground and vibrations were caused by the fall of a heavy weight at some distance from the wall. Sinusoidal vibration of the system was also examined. This wall had a provision for varying its weight to study the inertial effects. The back was filled using air-dried medium coarse sand, and pressures were measured using suitable pressure cells. The above tests were intended to remove some of the dis crepancies of the Monenobe-Okabe formula widely in use to-day (1973). This theory is based on plastic equilibrium in soils and hence does —. not give any idea of the displacements suffered by a retaining wall * • 11 during an earthquake. It is obvious that a rational design of retaining walls should be based on the displacements, for, many walls can be permitted to undergo some displacements during an earthquake without constituting failure. This is more so because severe earth quakes are not very frequent and an engineering structure need be designed only for a few severe earthquakes. A mathematical model has been proposed here to determine the displacement of retaining walls in translation. Since it is rather difficult to assess the mass of soil participating in vibrations, an experimental set up was designed and fabricated for the purpose. From these studies, it was concluded that i) The static earth pressure on cantilever retaining walls is given by Jaky's formula for at-rest pressures. ii) The dynamic earth pressures on all types of walls depend more on peak velocities of the ground motion than on accelerations. iii) The point of application of the dynamic increment is at about the mid-height of the flexible walls. On rigid walls it is lower and is at about 40 % height of the wall above its base. iv) The inertia force of the wall is also found to be a function of the energy input and hence the ground velocities because the resistance would be mobilised only at a finite displacement and some work has to be done. v) Amathematical model for predicting the earthquake induced displacement of a retaining wall in translation has been proposed based on the present day knowledge of the behaviour of the soil below and behind the wall under static loads.