The current limitations of displacement based design

Student: Tim Sullivan 
Supervisors: Prof. G.M. Calvi, Dr. M.J. Kowalsky

 

Abstract

Displacement based design methods are emerging as the latest tool for performance based seismic design. Of the many different displacement based design procedures proposed in recent years there are few that have been developed to a standard suitable for implementation in modern design codes. Through application to various case studies this report identifies and discusses the difficulties a designer may encounter when trying to use displacement based design. It is hoped that by presenting these limitations efforts will be made to develop the methods further so that designers can begin using the methods with ease and confidence.

The report incorporates results for five different case studies designed in accordance with eight different displacement based design methods.

The report focuses on three aspects of the design methods:
1. The relative ease or difficulty with which a design method can be applied and any apparent limitations a method may have.
2. The required strength obtained for each method and how this compares with the other methods.
3. The performance of the methods assessed by comparing the predicted ductility or drift values for each case study with those obtained through time-history analyses.

The case studies indicate that the level of involvement required by the designer does vary considerably between methods. Some methods are only applicable to certain structural types and others encounter difficulties with irregular structural forms and flexible foundations. In most instances the designer is required to make assumptions in order to proceed further and in other instances the method does simply not facilitate design.

As a performance based design tool some methods are currently limited to specific earthquake levels.

The study proceeds by using each of the design methods to develop design strengths to satisfy a set of design parameters. Despite all methods using the same set of design parameters, a large variation in design strength is observed. It is shown that some methods require a design base shear more than five times that of other methods.

After the designs are complete time history analyses are conducted using computer models with the strength as obtained for each design. The time-history analyses indicate that all the methods successfully provide designs that ensure limit states are not exceeded. Despite the large variation in design strength the variation in drifts observed between methods is relatively low.

You may download a digital version of this MSc dissertation here.