Abstract

Cross bore is an intersection of an existing underground pipeline such as a sewer pipeline by a second utility such as a gas pipeline installed using the trenchless technique, which could lead to severe incidents like explosions. There are hundreds of thousands of existing cross bores in the U.S., at an estimated average rate of 0.4 per mile of pipeline reported by Cross Bore Safety Association, which is at high risk and a significant threat to public safety if not well detected, characterized, and mitigated. The research on high-accuracy nondestructive evaluation (NDE) methods that can identify cross bore areas more effectively and efficiently under challenging field environment is of vital importance. This paper proposes a multi-channel capacitive sensing system for cross bores detection and classification, which passes through the gas pipe to inspect the dielectric property changes of surrounding materials and indicates the existence of cross bores. A 3-D simulation modeling tool was developed to optimize the footprint of multi-channel sensor, and the effect of four cross bore types on the received capacitance values was investigated. The lab-scale experiments were performed using the developed multi-channel capacitive sensing system through a soil box setup, and the experimental results indicated that the developed system can identify the four types of cross bores. To achieve an automated decision making for the cross bores detection and classification, machine learning (ML) algorithms were developed through the experimental dataset, and it was found that the subspace k-nearest neighbors (SKNN) performed better with a high classification accuracy. Finally, field test validation was performed at the pipe farm (Gas Technology Institute, IL, USA) and the superior capability of the developed sensing system in identifying cross bores as well as other key parts in gas pipelines including butt fusion and saddle fitting with coil was demonstrated.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call