Development of LSN-based pipe repair rate models utilising data from the 2011 Christchurch earthquakes

Abstract

The Canterbury Earthquake Sequence (CES) adversely impacted built, economic and social environments. This included widespread physical damage to the water supply pipe network in Christchurch, resulting in long service disruptions. The transient and permanent ground deformations generated by the earthquakes in the CES caused a range of pipe damage, particularly in the MW 6.2 22 February 2011 and the relatively less damaging MW 6.0 13 June 2011 event. Damage to the pipes in both events was largely attributed to liquefaction and lateral spreading effects. Pipes made of ductile material (e.g. PVC, HDPE) sustained lesser damage (and therefore lower repair rates) compared to the pipes made of non-ductile material (e.g. AC, CI). In all cases, the repair rates (number of repairs per kilometre) typically increased with increasing liquefaction severity. Utilising the pipe repair dataset and Liquefaction Severity Number (LSN) maps generated from extensive geotechnical investigation following the CES events, new repair rate prediction models for water pipes subjected to liquefaction effects have been derived and are presented in this paper. Repair data from both earthquakes has been analysed independently and in combination, providing two sets of repair rate functions and different levels of uncertainty. Repair rate functions were first derived from pipes grouped by combination of diameter (i.e. ϕ < 75 mm or ϕ ≥ 75 mm) and material type (i.e. ductile or non-ductile). The models were then refined by adding correction factors for those material types and diameters with sufficient sample length. Correction factors were derived for AC, CI, PVC pipes of diameter ≥75 mm and for MDPE and HDPE80 pipes of diameter <75 mm. Galvanised Iron (GI) pipes performed poorly during the earthquakes, resulting in very high repair rates compared to the other non-ductile pipes of diameter <75 mm damaged in the network; this warranted a separate repair rate model to be developed for this pipe type. The proposed models can be used in risk assessment of water pipe networks; i.e. to estimate the number of pipe repairs from potential liquefaction damage from future earthquakes.