Energy and powerNews

OpenSynth’s GB smart meter synthetic data now openly available

OpenSynth’s GB smart meter synthetic data now openly available

Image: Centre for Net Zero

Synthetic GB residential smart meter data created in the OpenSynth project is now openly accessible on the Zenodo open science platform.

The dataset, which was created by Octopus Centre for Net Zero’s Faraday generative AI model, contains 10 million synthetic load profiles trained on over 300 million half hourly smart meter readings from 20,000 Octopus Energy households sampled between 2021 and 2022.

It is conditioned on metadata including property type, energy performance certificate rating, low carbon technology ownership and seasonality.

OpenSynth is an open data community, originated by the Centre for Net Zero and sourced under the Linux Foundation through LF Energy.

Have you read?
Danes embrace energy data for greener and cheaper electricity
Octopus Energy boss Greg Jackson on ‘the magic of AI in the energy sector’

It is aimed to empower holders of raw smart meter data around the world to generate and share synthetic data, and for community members to generate, improve and share algorithms.

Access to smart meter data is considered essential to the rapid transition to electrified grids, underpinned by flexibility delivered by low carbon technologies, such as electric vehicles and heat pumps and powered by renewable energy.

But little of this data is available for research and modelling purposes due to consumer privacy protections.

The basis for the Centre for Net Zero’s approach is that while there are calls for raw datasets to be unlocked through regulatory changes, this is likely to be a lengthy process and that more rapidly created synthetic data can overcome the privacy issues.

The Faraday model produces household-level synthetic load profiles conditioned on the metadata.

The output data is reported to be highly faithful to the original training data it was trained on. On a population level, synthetic data and real data have similar mean and quantile values, i.e. peak consumption, and on the individual household level the synthetic profiles highly resemble that of real households.

The Centre for Net Zero has reported that Faraday to date has been used by more than 100 alpha testers globally, from universities to industrial partners to government and regulators.

Further work is underway to build out the backend infrastructure to host datasets.

Join the smart energy community in Milan to be a part of Europe’s premier energy transition event.

Leave a Reply

Your email address will not be published. Required fields are marked *