Forecaster for Embedded Generation (FEmGE)

Gas networks supply embedded power stations that support the electricity network. These embedded generators can fire up without any warning to GDNs and is causing significant challenges to gas networks.GDNs are required to submit hourly gas demand nominations to National Gas for each offtake point within specified time deadlines.Embedded generators are small. They are not included in the UNC’s requirements to notify their GDN of intended offtake activity due to their size being below the threshold for NExAs (network exit agreements). Despite this, GDNs must include the demand from these embedded generators in their nominations to ensure there is sufficient gas within their network. This causes numerous challenges for SGN and other GDNs.GDNs’ current forecasting process does not specifically forecast embedded gas generation, and current models do not take inputs from the electricity market. Embedded generators act in a variety of electricity markets, yet GDNs don’t have visibility of this demand.It is anticipated that additional embedded generators will connect in the coming months/years as the demand for electricity increases.The challenge of not having knowledge of embedded generator’s demand and its potential to contribute to a storage shortage has been acknowledged by both EGRIT (Electricity and Gas Resilience Task Group) and NESO (National Energy System Operator). The benefits of creating a notification platform supported by a ML engine are various. Namely to develop an ML-enabled forecasting tool to predict gas demand from embedded generators with increased accuracy as delivery time approaches. In addition to create a notification platform to improve real-time visibility of embedded generator activities within the electricity and gas networks.This NIA project aims to progress the FEmGE forecasting tool from TRL 1 to TRL 7, delivering a fully functional MVP. NGN will be funding this project to the value of £92,333 and SGN to £184,666 of the total of £276,999.