Improving methods for detecting leaks on gas pipelines
- Impact: Transformational
- Key benefits/themes: Customer Vulnerability, Financial, environmental
- Project partners: Northern Gas Networks, Synthotech & Synovate
The Challenge
The UK’s gas networks currently use the ‘above ground bar holing’ method to detect gas leaks, which often produces inaccurate results, leading to unnecessary excavations and reinstatement activities. The application of this technique is also physically demanding and can cause unwanted strains or injury to operatives performing the work.
Additionally, pinpointing the location of a gas escape and making repairs is often a time-consuming process. This can cause excessive disruption for local communities, especially if multiple excavations are required to locate the exact position of the gas escape.
The Solution
LeakVISION is a detection sensor that inspects internal walls of live gas pipelines using machine learning technology to pinpoint areas of interest.
Compared to traditional techniques, LeakVISION gives users much more information in an easy-to-digest format whilst being quick and easy to deploy on-site, speeding up how leakage is detected and repaired.
The novel sensor enables improved decision making and precise targeted remediation, as well as an improved understanding of the leakage process through advanced thermal modelling techniques.
Learnings from this project and others have enabled Cadent Gas to implement LeakVISION as part of its BAU operations.
Project Partners
LeakVISION was developed as part of a Network Innovation Allowance (NIA) project with Northern Gas Networks (NGN). The project was delivered collaboratively by Synthotech and Synovate, and included ROSEN, The Technology Partnership (TTP) and academics.
The EIC supported NGN and Synthotech with the project development and contract agreement in accordance with the NIA framework .
Following successful completion of the NIA project, the solution was further progressed through the Discovery and Alpha phases of the Strategic Innovation Fund (SIF). This project, Thermal Imagery Analysis, was supported by NGN and National Gas Transmission.
The Project
Live field trials demonstrated various benefits for network partners, including reduced operational expenditure (OPEX). One trial conducted on a sample section of the gas distribution network in the north of England demonstrated that LeakVISION reduced the overall time required to locate and repair multiple gas escapes, minimising disruption to customers.
As part of the SIF funded project, a trial was completed in a live hydrogen pipeline to assess the device’s functionality and its capability to safely detect leaks. The trial proved successful, providing confirmation that both the deployment and retrieval approach for the device are secure.
Outputs to Date
- The successful implementation of a live thermal imaging robotic system within an operational gas distribution network.
- The system potentially could identify specific ‘areas of interest’ that require remediation.
- The system has future capability to guide more agile asset investment decisions.
- A robotic device that uses a heating element to warm-up flowing gas, in a live distribution network, safely and without incident.
- Identification of design improvements for field operations.
The Impact
Customer Service: Minimised disruptions from excavations to assess pipes in difficult to reach places.
Societal: Reductions in the number of excavations and faster repair times.
Environmental: Reductions in the volume of gas vented to the atmosphere during complex-to-locate gas escapes.
Operational: A reduced number of excavations will lead to improved safety and will result in fewer injuries to operatives.
Next steps
Consideration is underway within NGN to create a comprehensive, quantified benefits case for each use case for the successful deployment of the solution.
Learnings from previous LeakVISION projects are being used by Cadent Gas in its SIF project, Digital Platform for Leakage Analytics (DPLA), which will improve the GDNs’ understanding, accuracy and granularity regarding the locations and volumes of leaks within assets. It is expected to enable a transformational change in the current shrinkage and leakage model used by GDNs, with completion anticipated by March 2026.