Prioritization of limited funding for pipelines maintenance is a major issue that concerns municipalities nationwide. Conducting a probabilistic assessment can provide a complete characterization of the performance of structural elements and systems along with optimizing the limited resources. The most widely used probabilistic performance indicator is reliability, a measure of the probability of failure corresponding to a particular limit state (e.g., ultimate strength or serviceability). Reliability methods can be used to identify which pipeline sections within a particular system require the most urgent inspection or repair. To this end, an automated data-driven framework for large diameter reinforced concrete pipes (RCPs) is developed that converts the raw unfiltered inspection readings to data that is used for estimating residual life and further reliability assessment purposes. In the current work, initially, the wall thickness erosion is determined based on the inspection data collected using Light Detection and Ranging (LiDAR). Furthermore, the best fit among several probability distribution functions for the wall thickness loss is obtained, which is integrated with a serviceability limit state that defines failure as the complete loss of 1-in concrete cover caused by environmental conditions such as sulfide-induced erosion. Considering this limit state and a prescribed probability of exceedance threshold, a reliability-based prediction of the remaining service life is proposed. The developed framework requires minimal user interference and is, therefore, less time consuming and more consistent compared to previous research. From an asset management point of view, the most vulnerable pipeline sections are identified that will require further inspection and attention. This will provide decision makers crucial information regarding the current state of the pipeline network, to better allocate the already scarce maintenance funding of these pipelines.

Tran-SET Project: 20STUTA25