Statistical Model for Estimating Restoration Time of Sewerage Pipelines after Earthquakes: Fid-Bau Portal

Statistical Model for Estimating Restoration Time of Sewerage Pipelines after Earthquakes

Liu, M. / Scheepbouwer, E. / Gerhard, D.

A timely and effective postearthquake recovery could minimize the impacts and disruption that have arisen due to the induced physical damage to sewerage pipelines. The knowledge of the restoration time could facilitate the efficiency of the recovery. This paper presents a statistical model for predicting the time required to restore sewerage systems after earthquakes and to test the applicability and adaptability of the proposed approach to other areas. Using a database of 4,648 pipes that have been repaired or renewed in Christchurch, New Zealand, after the Canterbury earthquake sequence in 2010–2011, four candidate statistical models and approaches, namely, accelerated failure time, Cox proportional hazard, random survival forest (RSF), and multiple linear regression, are compared and validated. The RSF approach is found to have the best prediction power for estimating restoration time of the impaired sewerage pipelines postearthquake using the random sampling validation approach. Furthermore, key variables that have significant influence on predicting the restoration duration of sewer pipelines are identified. Finally, restoration curves are produced for informing decision makers and the public regarding pipe restoration rates after earthquakes.