Date of Award

Spring 5-2023

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computational Analysis and Modeling

First Advisor

John C. Matthews

Abstract

The dissertation research work described here has three primary objectives under risk-based decision making. (1) The development of a comprehensive sewer pipe condition rating model that incorporates many environmental, structural, and hydraulic parameters. (2) The development of a sewer pipe deterioration model used to predict future overall condition states of the pipe, as well as determining the probability of failure at any given age of the pipe. (3) The development of a comprehensive consequence of failure model that assesses the consequence of sewer pipe failure using economic, social, and environmental cost factors.

The Pipeline Assessment and Certification Program (PACP) was developed by the National Association of Sewer Service Companies, the industry-accepted protocol for condition rating sewer pipes in the US. The PACP method relies exclusively on visual inspections performed using Closed-Circuit Television (CCTV), where existing structural and operation and maintenance (O&M) defects are observed by certified operators. A limitation of the PACP method is that it does not use pipe characteristics, depth, soil type, surface conditions, pipe criticality, capacity, the distribution of structural defects, or history of preventative maintenance to determine the condition rating of the sewer pipe segment. Therefore, a comprehensive rating model with pipe characteristics, external characteristics, and hydraulic characteristics was developed. The calculating of a comprehensive rating is an entirely manual process.

Therefore, this research work addresses this limitation of Analytical Hierarchy Process (AHP) and suggests AHP is not a suitable method to calculate comprehensive rating. Develops a faster calculation of a comprehensive rating model using and K-NN that incorporates pipe characteristics, environmental characteristics, and information about PACP structural score and PACP O&M score in hydraulic factors. Factors such as pipe age, pipe material, diameter, shape, depth, soil type, loading, carried waste, seismic zone, PACP structural score, and PACP O&M score are used. Our proposed model is applied to the data received from the City of Shreveport, LA, which is currently under a Federal Consent Decree. The results of a comprehensive rating model showed a below-average validity percentage because linear regression assumes a linear relationship between the input and output variables. Still, the relationship between response and the predictor is not linear for AHP to prove AHP is not a suitable method and satisfactory results for K-NN.

As part of decision-making, for capital improvement planning and budgeting, the capacity to predict future sewer pipe conditions and potential breakdowns is essential. In contrast to the often-used Discrete Time Markov Chain approaches in the literature, the deterioration model created here uses a Continuous Time Markov Chain method to calculate the likelihood that a pipe will change from a better to a worse condition at given age.

The consequence of the pipe's failure is established to ascertain the risk of failure and to create a comprehensive framework for risk-based decision-making. To estimate the impact of the asset's failure, the established consequence of failure model considers a significant number of economic, social, and environmental cost elements. For budgeting future capital projects and improvements, the CTMC model and failure consequences for sewers are useful.

Download

Share

COinS