Defect-based ArcGIS tool for prioritizing inspection of sewer pipelines

Abstract

This paper presents a defect-based model for assessing risk of failure for sewer pipelines. The proposed model deploys a Sugeno fuzzy inference system to create a risk index from which inspection and replacement activities may be prioritized. To determine the likelihood of failure, dynamic Bayesian network (DBN) was used as an inference engine to predict the likelihood of sewer pipeline failure based on both probable defects that could occur and some pipeline characteristics. The consequences of failure were determined using an economic loss model that assumed both costs resulting from the failure of sewer pipelines and benefits from avoiding such failures. An ArcGIS tool was created using the Python programming language to perform the Sugeno fuzzy inference method and determine the risk of failure by combining both the likelihood and consequences of failure. Actual data for inspected sewer pipelines in Doha, Qatar, were used to validate the tool; in the validation, the pipelines from the model were compared with the inspected pipelines. It was found that, if deployed, the proposed tool could save more than 77% over the current inspection practices followed by municipalities. It is expected that the resulting risk map would help key personnel in municipalities to identify sewer pipelines that require immediate interventions and would assist in better planning for inspection programs, especially in cases of limited funds.