Stochastic deterioration modelling of sidewalk performance using compound poisson processes

Abstract

ABSTRACT Pedestrian sidewalk plays an increasingly important role in urban transportation. Due to the traditional misconception of it being an infrastructure class of low risk, the current sidewalk asset management still largely relies on experiences and personal judgment. To promote rational asset management, the authors have recently developed two performance indicators namely, the maintenance repair index (MRI) and the sidewalk condition index (SCI). This paper advances the effort by developing two deterioration models using a compound Poisson process (CPP), which consists of a nonhomogeneous Poisson process for the jumping time and a Weibull distribution for the jumping size. Based on a real-life sidewalk defects database, the models also consider the heterogeneity of the sidewalks by including asset attributes such as sidewalk service age, foundation soil type, sidewalk size, pedestrian density, and intensity of construction activities in the vicinity of a sidewalk. The modelling approach presented in the paper – particularly the considerations in the selection of CPP model, the Bayesian statistical inference method, and the model validation process – serves as an excellent example for stochastic deterioration modelling for risk-informed infrastructure asset management.