Mitigation of Damages to Residential Buildings Caused by Liquefaction Induced Settlement in Christchurch

Abstract

The 2010 Darfield earthquake and aftershocks that followed have caused widespread liquefaction in Christchurch and the resulting differential settlements have damaged many residential properties. Ground improvement can effectively reduce liquefactioninduced ground deformation to within acceptable limits. However, the depth of liquefiable soil in Christchurch extends to considerable depth and it is impractical to treat the entire layer of liquefiable soil for residential buildings. To investigate the depth of ground improvement required to meet the differential settlement requirements as set out by the Ministry of Business, Innovation and Employment, numerical analyses were carried out using the finite difference program FLAC. A verification exercise was first carried out in Riccarton and Shirley where considerable damage occurred, to provide some validation to the numerical analyses. Subsequently, a typical soil profile was modelled in the suburb of St Albans in Christchurch based on nearby borehole data. St Albans is a Foundation Technical Category 3 (TC3) site with moderate to severe liquefaction observed during the 2010 Darfield earthquake and following aftershocks and is assumed to represent a worst case scenario for residential buildings in Christchurch. Typical house size and loadings in New Zealand have been modelled and parameters including width and stiffness of ground improvement, seismic input motion and liquefiable layer thickness were varied to assess their effects on the required depth of improvement. The width of ground improvement did not appear to influence the required depth of improvement as long as it extends outside the foundation perimeter by more than 1 metre. Stiffness of the ground improvement also does not have any significant effect on the foundation differential settlement, provided sufficient stiffness is achieved to prevent liquefaction. Varied seismic motion records did produce different required depth of improvement, possibly due to different response spectrums and power contents. While as the liquefiable layer thickness increased, required ground improvement depth increased until a limiting depth was reached. When results were plotted against Ishihara’s chart for estimating the required thickness of non-liquefiable layer to prevent liquefaction-induced damage, the data fitted reasonably and seem to fall between the 0.3g and 0.4 – 0.5g curves. It appears that Ishihara’s chart may be used with caution as an estimation tool.