Settlement effects on masonry structures

Historic masonry structures are often perceived to be vulnerable to differential settlement, which can lead to visible cracking and uncertainty about structural safety. However, others may perceive these same structures to be resilient to support movement, and able to withstand relatively large displacements before there is any need for concern. This paper considers how these perceptions can exist in tandem, and how both are accurate, by considering various problems within the context of settlement damage to masonry. Settlement effects on masonry structures are first considered from a collapse (or life safety) perspective. Discrete element modelling is used to evaluate the effects of soil movements on structural safety, both for common structural typologies and for specific case studies. These examples illustrate the capacity of masonry to withstand settlement, but also the difficulty in predicting when collapse will occur. Second, settlement effects are considered from a serviceability perspective, including prediction of minor damage (e.g. cracking) rather than stability alone.A computational and experimental research project that aims to improve damage prediction methods is discussed. As part of this project, centrifuge tests were conducted on brittle models of masonry buildings subjected to tunnel volume loss and resulting ground settlement. These experiments provide data to resolve questions about the nature of the soil-structure interaction, and highlight the difficulty of predicting damage in the form of cracking. They also provide data to improve both detailed finite element modelling procedures and simplified damage prediction methods. Finally, settlement effects are considered from a field measurement perspective. A case study involving the use of terrestrial laser scanning to understand the effects of settlement on an existing masonry rail bridge is presented. This case study is used to exemplify the value of distributed displacement data in understanding structural behavior, and the need for better integration of more detailed field data in assessment and prediction procedures