Displacement-Based Design Criteria for Gravity Retaining Walls in Light of Recent Earthquakes

In the last 20 years many large earthquakes have occurred giving the geotechnical community an abundance of data available for analysis. Richards and Elms (1979) developed a design method for gravity retaining walls based on finite displacements, in accordance with the Newmark (1965) sliding block analysis and the Franklin and Chang (1977) earthquake records analysis. Richards and Elms approximated an upper bound to Franklin and Chang’s curves with an expression that permits a designer to choose an allowable displacement to determine the required wall weight for a particular peak ground acceleration and peak ground velocity. A preliminary investigation of digitized records from the Loma Prieta, Northridge, and Kobe earthquakes shows that the upper bound suggested by the Richards and Elms procedure significantly under predicts the displacement that would occur during these recent earthquakes. Consequently, walls designed with the suggested upper bound may be subject to excessive displacement. Comparisons are made between the Whitman and Liao (1985) method and the Richards and Elms procedure. An upper bound developed from the Northridge data results in as much as a 25% increase in the required wall weight. This paper analyzes the records of recent earthquakes and discusses the implications of raising the upper bound of the Richards and Elms limited displacement design approach. The combined consideration of recent earthquakes suggests that the normalizing parameters of peak velocity and peak acceleration (as suggested by Richards and Elms) may not be sufficient to develop an upper bound without significant scatter