UK Power Networks has begun a new project to simulate the electricity network across London, the South and South East of England with the aid of machine learning.
By modelling the power flows across hundreds of thousands of miles of network, the company said it will be able to operate its assets more efficiently and unlock almost 70MW of spare capacity that can be used to connect low-carbon technologies such as electric vehicle chargers and heat pumps.
As part of its Envision project, the company will feed network data, including real-time data from monitoring devices connected to substations, into a machine learning algorithm that will use this information to create a simulation of its network and predict how it will behave. The software will then compare these forecasts with the actual loads and make adjustments where necessary, improving the accuracy of the model over time.
Ian Cameron, head of customer services and innovation at the distribution network operator, said: “Our customers rightly expect us to do everything we can to make the switch to electric cars and low carbon heating as affordable as possible. Through Envision, we’re thinking outside the box and re-imagining traditional ways of working, to make it happen.”
Simone Torino, head of product and business development at CKDelta, which is collaborating on the project, said: “The aim of the Envision model is to generate a ‘virtual sensing network’ that uses advanced data capabilities and machine learning to simulate the behaviour of the network at scale, accurately estimating changing network load profiles.
“In a world where the uptake of new distributed energy resources and the increasing electrification of transport are impacting electrical demand and distribution network constraints like never before, having this type of modelling and predictive analytics capabilities is a game changer for the utilities sector and has potential to reshape how we approach demand and supply in other sectors such as transport.”