Liquefaction modelling of a strong motion station in Christchurch, New Zealand

Abstract

Advanced constitutive models can replicate several aspects of soil behaviour, but, due to their complexity and number of parameters, they need more sophisticated and realistic validation under general loading conditions. When modelling liquefaction phenome na, the lack of field monitoring data means that model testing, such as centrifuge experiments, is often used as benchmark for the numerical analyses. The 2010 - 2011 Canterbury earthquake sequence in New Zealand was recorded by a number of strong motion sta tions at various distances from the earthquake epicentre. Additionally, an extensive field and laboratory program me has since become available, adequately describing the geological, geotechnical and hydro geo logical conditions in the area. As such, the perf ormance of a two - surface bounding surface plasticity constitutive model for sands, calibrated based on site - specific laboratory data, is assessed using field evidence of a strong motion station in fully - coupled effective stress - based finite element analyse s. As the real stratigraphy is complex , with layers of silts and clays between the sandy strata, a simpler cyclic non - linear elastic model , which can adequately incorporate the basic aspects of dynamic soil behaviour, is also used to model the non - liquefia ble strata. To specify the input ground motion at the base of the deposit , the recorded ground surface motion at a site with no evidence of liquefaction is deconvolve d and compared with the outcrop predictions of a New Zealand - specific gr ound motion predic tion equation . The numerical results are compared with the recorded horizontal ground surface acceleration time - history of the 22 nd February 2011 seismic event , exhibiting very good agreement .