Planetary-scale machine learning is filling the gaps in solar data

Summary

Solar is the fastest growing power generation technology in history, however, inadequate data availability hides the true scale and hinders effective grid planning

To address the need for accurate, current, asset-level data, TransitionZero developed machine-learning and satellite imagery powered TZ-SAM

One quarter after launch, TZ-SAM registered 500 users and secured $1 million in funding for additional product development

Background

Solar power is the fastest-growing power generation technology in history. To effectively manage ongoing solar deployment, accurate and current facility-level data is essential. It enables the management of intermittency, facilitates grid planning, and allows for the assessment of trade-offs with biodiversity, conservation, and land protection priorities.

Unfortunately, available datasets are country-level, don’t adequately differentiate between large- and small-scale facilities, and struggle to keep up with the rapid pace of solar expansion. In order to meet the changing needs of grid planners, and their need for asset-level, up-to-date information, TransitionZero developed Solar Asset Mapper (TZ-SAM). Updated quarterly, TZ-SAM is an open-access, asset-level global dataset of commercial- and utility-scale solar facilities utilising satellite imagery and machine learning.

The TZ-SAM dataset currently contains the location and shape of 63,096 assets across 182 countries, covering close to 19,000 square kilometres with a total estimated capacity of 705 GW. TZ-SAM can be integrated into system models for electricity grid operations and planning, supporting more effective forecasting by filling gaps in traditional methods of solar asset reporting.

“Today there are over 140 countries installing solar and official data is terrible. I can’t tell you how uncertain we are of how much solar is actually out there. The idea of using satellite data to support is interesting, in order to get a more accurate picture on how many solar modules come onto the market."

Jenny Chase

Lead Solar Analyst, Bloomberg NEF

The product

TransitionZero developed algorithms using earth observation and machine learning to accurately identify the capacity, land area, and age of every large solar facility worldwide, in addition to a large number of small and medium-sized assets. TZ-SAM’s methodology builds on Kruitwagen, L., Story, K.T., Friedrich, J. et al. and uses the European Space Agency’s (ESA) Sentinel-2 dataset, combined with the community-driven and open-source OpenStreetMap (OSM) dataset for training labels.

Since launching our Solar Asset Mapper in May 2024, it has registered over 500 users from government departments, renewable developers, asset managers, energy consultancies, civil society organisations and global media including The Economist and New York Times.

In addition, it received a $1 million USD grant from the newly established Banyan Software Foundation to help accelerate the development of new features and improved methodology for TZ-SAM.

“Keeping tabs on the global solar build-out requires all hands on deck. At Global Energy Monitor, we compile a project-by-project, ground-up inventory of solar installations for scientists, policymakers and the public. But as progress quickens towards the global goal of tripling renewables, providing complete and accurate data on the growing number of smaller-scale projects across the globe is increasingly difficult. TransitionZero's Solar Asset Mapper addresses this challenge by rounding out the solar picture with geometries, locations, and capacity estimates.”

Diren Kocakuşak

Research Analyst, Global Energy Monitor

Looking ahead

With the help of the generous grant from the Banyan Software Foundation, TransitionZero is working on developing TZ-SAM further. This includes additional solar asset type metadata, increased accuracy of estimated solar capacities, and continued research into expanding detections to residential solar applications – a primary feature request for TZ-SAM.

“Banyan is helping us improve access to critical data for effective energy systems planning, empowering countries to pursue sustainable energy solutions independently.”

Matt Gray

CEO & Co-founder, TransitionZero