Copula parameter estimation using Bayesian inference for pipe data analysis

Abstract

Water main systems are aging and becoming a growing concern for maintenance. The structural deterioration of water mains is affected by different factors, such as pipe age, pipe material, soil condition, and pipe size, among others. Various methods of modeling have been used to predict the failure of water mains. Since pipe networks are underground and obtaining data on pipe conditions is very costly, statistical modeling has been widely used for pipe condition assessment. An emerging statistical method known as copula modeling is used here for pipe data analysis. The copula method is very useful in cases where marginals belong to different families of distributions. It is also useful for generating a large number of data points when it is difficult to obtain a data set, as is the case for pipe condition assessment, and where data sets have random variables belonging to non-Gaussian family distributions. Different copula families are applied here to model the dependency between the pipe age and repair age of pipes. The paper uses a Bayesian framework to estimate the parameter values in the copula model. This approach offers an additional option for estimating copula parameters for pipe data.