A solar irradiance prediction tool looking to enable more efficient management of electricity networks has been launched by Meniscus.
Near real time satellite imagery is used to create 1km2 gridded images of solar reflectance, including visible and infra-red images and metadata on estimated cloud height.
Images feed into the Meniscus Analytics Platform (MAP), the firm’s real time analytics platform that uses Internet of Things and artificial intelligence (AI) to calculate the direct, diffuse and total in-plane solar irradiance values.
The values are modified to account for cloud cover and actual and forecasted rainfall, as well as cloud height and the time of data. A set of solar irradiance predictions are then produced at 15 minute intervals for three hours.
Improved solar irradiance predictions allows PV sites to match generation to demand and reduce curtailment of generation through allowing co-located battery storage to charge from any solar above the export limit of the site, Meniscus said.
And for sites without storage, generation can be curtailed when the network is constrained in response to DSR signals, such as Demand Turn-Up, allowing DNOs and National Grid ESO to efficiently manage their network.
Earlier this year National Grid ESO announced it had developed its own solar forecasting method in collaboration with the Alan Turing Institute, focused on predicting solar generation rather than irradiance. Much like Meniscus’ irradiance forecasting, National Grid’s method utilised AI and resulted in an improvement in its forecasting by 33%.
The irradiance forecasting was first tested at a Cornwall Council-owned solar farm and is funded by Innovate UK. Other partners include Open Energi and BRE National Solar Centre.