AI powers DINO for smarter local energy networks

A pilot project that aims to show how artificial intelligence (AI) can be used to prevent local power networks being overloaded during the projected rapid rise in demand for electricity. The project has been launched with co-funding from Innovate UK, the UK’s innovation agency.

Programme

Project DINO (Domestic Infrastructure and Network Optimisation) — a partnership between Evergreen Smart Power, us, and Myenergi — will demonstrate in a new build residential setting how AI can safeguard network integrity through two-way communication with household appliances. It’s the first project of its type in the UK.

Currently, housing developments are designed to cope with an average peak load of 2kW per household. However, as electric vehicle (EV) charging and heat pumps become more commonplace, this could routinely exceed 10kW — and without intervention, this could lead to power outages.

“There is growing realisation that local power networks may not be able to cope with the huge increase in demand for electricity as we transition to electric vehicles, battery storage/charging and air source heat pumps alongside traditional household appliances,” says Jayson Whitaker, Managing Director of Energy Assets Networks, which operates local networks around the country.

“The DINO project aims to demonstrate how a network-to-device AI interface can manage loads dynamically by enabling appliances to automatically dial down consumption at peak times to relieve network stress and safeguard power to homes. Without such a solution, the country would need to invest in a hugely costly network reinforcement programme to increase capacity.”

Networks are predicted to come under pressure as electrification increases in every area of life as the country decarbonises on the road to Net Zero — from the adoption of electric vehicles, with the phase-out of petrol and diesel cars in 2035, to the transition from gas to electric heating.

Outcome

As part of Project DINO, Evergreen Smart Power aims to apply an AI solution that enables two-way communication between networks and devices, such as EV chargers, so that when the local system is under stress, energy consumption will reduce automatically to allow households to share available bandwidth.

“This will not only help keep costs down for consumers but also enable more sustainable use of available renewable energy within our electrical networks, reducing the need to fall back on fossil fuels to keep the lights on during peak times,” said Chris Williams, Project Manager for DINO, Evergreen Smart Power.

The pilot research project will last for two years, with the partners monitoring real-time power flow and consumption on a new build housing scheme at both feeder and substation levels, with the data augmented with models of EV charging and heat pump usage. This will lead to the development of new algorithms and proof-of-concept equipment that will automatically and precisely respond to system stress down to individual feeder cables. The project will also be further extended to model the effect of domestic battery storage to determine how these too could be similarly controlled.