Digital Platform for Leakage Analytics (DPLA)

The Digital Platform for Leakage Analytics (DPLA) Project is an innovative project aimed at significantly reducing gas network leaks and emissions in the UK Gas Transmission and Distribution sectors through digital technology and data.

The project is led by Cadent in partnership with SGN (Southern Gas Networks), Northern Gas Networks (NGN), Wales and West Utilities (WWU), National Gas Transmission (NGT) and Guidehouse as technology delivery partner.

DPLA is funded through the Strategic Innovation Fund (SIF). Part of the RIIO-2 price controls, the SIF is delivered in partnership with Innovate UK, part of UK Research and Innovation (UKRI) with the objective of helping transform the UK into the ‘Silicon Valley’ of energy, making it the best place for high-potential businesses to grow and scale in the energy market. The SIF therefore aims to find and fund innovative projects with the potential to accelerate the transition to net zero.

Gas Distribution Networks (GDNs) and their customers face increasing economic and environmental costs due to natural gas losses from the transportation network (i.e. shrinkage). GDN shrinkage represents in fact 4.5% of total methane emissions and 1% of total GHG emissions in the UK. The current modelling methodologies are based on emission coefficients derived over 20 years ago and do not have the ability to locate leaks to the level of accuracy needed to drive emissions reduction through targeted asset management programmes.

DPLA aims to significantly reduce gas network leaks and emissions in a cost-effective way. The project will provide a new real time emissions platform that will improve the accuracy of leakage modelling and enable the proactive identification of leaks before the public.

DPLA presents several financial, environmental and consumer benefits for external and internal stakeholders. Financial benefits are linked to the capability of targeting larger leaks sooner, leading to lower gas losses per year and lower shrinkage costs. These costs are passed on from GDNs to consumers' bills, meaning that customers will benefit directly from lower gas leakage volumes.

By minimizing methane leaks, with methane having a far higher global warming potential than carbon dioxide (CO2), DPLA is projected to avoid 12,435 GWh of natural gas loss and 14,856 ktCO2e of greenhouse gas emissions from distribution shrinkage and leakage by 2050. This will positively impact government priorities around tackling methane emissions as a participant of the Global Methane Pledge at COP 26 in November 2021.

Other benefits include improved safety and incident prevention through fewer site visits and near real time monitoring and fewer disruptions by enabling quicker, more targeted leak repairs.

DPLA combines the innovation of novel in-field detection technologies, machine learning, and modelling techniques. We are currently testing and refining the optimal mix of in-field detection and modelling to identify the best approach to deliver our solution that will enable localised leak assessment and aggregated/ network-wide reporting.

A number of in-field technologies have been assessed for the trials of in-field methane detection technologies, ranging from fixed sensors on Above Ground Installations (AGIs) to mobile leak detection devices for vehicles. 

 Five providers were shortlisted and from July 2024 trials are being carried out to assess the effectiveness of a technology in detecting, quantifying, and reporting on total emissions from an asset.

These technologies include: 

• Sensirion: Modular low maintenance continuous monitoring IR Laser Spectroscopy sensor to detect methane leaks across our AGIs;
• Qube: Modular low maintenance continuous monitoring molecular property spectrometer sensor (MPS) for detecting emissions across our assets;
• Distran: Handheld ultrasonic imaging cameras to detect leak across our assets;
• Satelytics: Satellite captures to detect emissions on our network;
• Bohr: Tuneable Diode Laser Absorption Spectroscopy (TDLAS) methane/ethane modular vehicle mounted sensor.

Probabilistic models play an important role in Machine Learning, by taking a statistical approach to forecast the possibility of future result from the impacts of random occurrences or actions. It is a quantitative modelling method that projects several possible outcomes potentially going beyond recent events. For DPLA, this model will use the same principle to run several different scenarios and conditions within Cadent’s network, using flow and pressure data as the main assumptions.

The SIF Beta Phase will be completed in June 2025. From 2026, with the commencement of RIIO-GD3, the platform will begin to be rolled out across the other GDNs.