The Descartes Labs platform enables electric utilities and renewable power companies to make asset management, risk management and trading more efficient. They leverage our solution to:
Monitor vegetation and encroachment of new building construction
Forecast solar and wind loads during peak demand
Detect wildfires and alert stakeholders on a continental basis in real time
Determine “behind-the-meter” rooftop solar inventory
Copernicus Sentinel data 2020
The Descartes Labs Platform Provides Real-Time Wildfire Alerting for Electric Utilities
Defined by the US Forest Service as wildfires that burn over 100,000 acres, megafires are occurring far more frequently as a result of climate change and the present-day consequences of historical forest management policies. California residents used to experience a megafire once every few decades. Now they’ve witnessed at least one such blaze annually for the past seven years.
The Descartes Labs Platform helps electric power utilities in North America reduce wildfire risk. GOES-16 and 17 (satellites operated by NASA and NOAA) scan the western hemisphere every five minutes and collect both visible light and thermal emission (heat) data. The Descartes Labs Platform pre-processes this information, and our analytics-ready, real-time data feed enables rapid wildfire alerting.
The State of New Mexico receives alerts from the Descartes Labs Platform. By getting an advanced look at fires as soon as they ignite, forestry departments and electric utilities are able to mitigate asset and regulatory risk, and proactively respond to fire incidents before they become unmanageable.
Featured Video – Detecting Wind Turbines
Art of the Possible
With the Descartes Labs Platform, business and data science teams can rapidly hypothesis, test, and deploy predictive analytics. These examples showcase some of our customers’ success, but the possibilities are endless.
Datasets: National Agriculture Imagery Program (NAIP), United States Geological Survey (USGS) LiDAR data, vector pipeline data from Open Street Map
Summary: Automating tree detection can help electric utilities manage vegetation encroachment and significantly reduce wildfire risk. The Descartes Labs Platform provides post-processed USGS LiDAR and several collections of high-resolution imagery. By leveraging the platform’s vast processing power, users can train on hundreds of thousands of LiDAR images, combine high-resolution data sets, and quickly deploy computer vision-based encroachment alerts.
Global solar panel identification
Copernicus Sentinel data 2020
Datasets: SPOT 6 and 7, Sentinel-2
Summary:The Descartes Labs Platform is a powerful tool for computer vision on a global scale. Data scientists leveraged the platform’s sensor fusion and scalable computing capabilities to map solar panels across the globe. The semantic segmentation approach uses a combination of medium and high-resolution optical imagery in addition to training data from Open Street Map – all readily available on the platform.
Wind speed estimation and power forecasting
Datasets: NOAA Global Forecast System (GFS)
Summary: Wind power forecasts are valuable to electric utilities and power market participants. The Descartes Labs Platform provides NOAA’s GFS forecast data and wind turbine locations and users can calculate the amount of wind power that can be generated in the future. The platform’s stock of readily available data and its flexible modeling environment make endless variations possible, including multiple, daily gigawatt forecasts and estimates of the data that lie between current and future forecasts.
Datasets: SPOT 6 and 7, Sentinel-2, GOES 16, California Independent System Operator (CAISO) power data
Summary: Solar power generation varies wildly based on cloud cover, creating operational challenges for electric utilities and trading opportunities on spot markets. The GOES-16 and 17 satellites, pre-processed in real-time on the Descartes Labs Platform, provide observations over the northern hemisphere every five minutes. Data scientists can use proprietary or publicly available historical power output rates and solar panel location data to make near-term predictions of cloud cover location and power output.
Summary:Wind turbine location is critical to modeling power output based on wind speed. With the Descartes Labs Platform and its readily available training data and high-resolution imagery collections, the development of computer vision models that automatically map such objects is a snap. But it is also possible to further classify types and ages of turbines, making power estimates more accurate and more granular by capturing variations in turbine efficiency.
Customers include electric power utilities and renewable energy companies involved in all aspects of electricity production and distribution: