Earthquake resistant design of tied-back retaining structures

Technical Abstract

This report considers design procedures for tied-back retaining walls under earthquake loading. Tied-back retaining walls are becoming widely used in NZ to support permanent excavations on sloping sites in order to provide level building platforms for residential and commercial developments. They are also widely used to support excavations for roadways and other key infrastructure.

Very little guidance is available for the design of tied-back retaining walls to resist earthquake shaking. Little observational data on the behavior of tied-back walls during earthquakes has been published, but, what there is suggests that they behave well.

A survey of New Zealand practice has showed that there is no consistency of approach and that most designers are relying on a range of different “black box” computer software with earthquake loading input simply as an additional horizontal force applied directly to the wall. The appropriateness of this approach is questionable because the full range of different failure modes is not necessarily addressed by the software nor is it always obvious what the software does.

In this study, a seismic design procedure for tied-back retaining walls was synthesized based on an existing, widely used, semi-empirical design procedure for gravity design of tied-back walls. The design procedure does not depend on specialist computer software.

The design procedure was tested by designing a range of case study walls and then subjecting them to simulated earthquakes by numerical time-history analysis using PLAXIS finite element software for soil and rock. The response of the walls to a variety of real earthquake records was measured including deformations, wall bending moments, and anchor forces.

From the results of these analyses, it was observed that all of the wall designs were robust and performed very well, including those designed only to resist gravity loads. In some cases large permanent deformations were observed (up to 400mm) but these were for very large earthquakes (scaled peak ground acceleration of 0.6g). In all cases the walls remained stable with anchor forces safely below ultimate tensile strength. Wall bending moments reached yield in some cases for the extreme earthquakes, but this is considered acceptable provided the wall elements are detailed for ductility.

Wall designed to resist low levels of horizontal acceleration (0.1g and 0.2g) showed significant improvements in performance over gravity only designs in terms of permanent displacement for relatively modest increases in cost. Walls designed to resist higher levels of horizontal acceleration (0.3g and 0.4g) showed additional improvements in performance but at much greater increased in cost.

Even when walls were designed to resist 100 percent of the peak ground acceleration of a particular earthquake record, significant permanent deformations were still observed. A tentative, detailed design procedure is provided based on the results of the study.