Clouds Under Control of Algorithms. Mycroft Mind Changes Photovoltaic to Predictable Source
- Kristyna Ištvanfyova
- Feb 8
- 2 min read
The development of photovoltaic power plants increases the pressure on the stability of energy networks. The production of electricity from the sun is ecological, yet at the same time everchanging and dependant on local sky conditions. This complicates the operations of distribution systems as well as trading on short-term markets. Brno´s technology company Mycroft Mind, a member of the TTC Group, has reacted on this challenge with its solution SkyEye, the predictive solar AI camera, which provides a very precise short-term prediction of the production of photovoltaic power plants, and this with a success rate of up to 96 percent.
While common prediction of the production from photovoltaics come primarily from comprehensive meteorological models and satellite data, SkyEye works with detailed information directly from the place of its installation. It combines camera systems, local sensors and machine learning models, that enables it to monitor quick weather changes that have a significant impact on the instant productivity of the power plant. The development of the SkyEye solution also builds on the involvement of Mycroft Mind into IPCEI, the European initiative in the field of microelectronics and communication technologies, which supports strategic research and innovation in key technology areas.
“In photovoltaics today, we are not usually referring to the knowledge of what will happen in three days, but what will happen in the next five or fifteen minutes. Specifically on this horizon, classical models have already hit their limits,” says Petr Možný, Sales and Marketing Director of Mycroft Mind. “In practice, SkyEye can see approaching cloud cover sooner than indicated in the production decrease, thus enabling us to react in real time.”
Camera that learns a specific power plant
SkyEye is a solar AI camera with an integrated computing unit installed directly at a photovoltaic power plant. Thanks to its wide-angle lens, it continuously scans the sky in high resolution and combines the image data with information from local sensors, i.e. sunshine intensity, temperature and wind speed.
Artificial intelligence, which is gradually learning the specifications of a given location including its microclimate, plays a key role. Thanks to edge computing technology, the operations are running directly in the equipment enabling a very speedy reaction without the necessity of big data transfer to the cloud.
According to internal measurements, the prediction success rate over the course of the following 120 minutes reaches approximately 96 percent. This is essential primarily for intraday trading with electricity, deviation management or the optimalization of battery storage operations.
From photovoltaic to smart energy
SkyEye is designed mainly for mid and large solar parks, however its use also extends to the field of smart energy. It is possible to combine production with the management of battery storage, the charging of electricity cars or the optimisation of consumption in industrial areas. This solution is a part of the Energy Intelligence Center, a broader software platform enabling to work not only with individual power plants, but also with an entire portfolio of energy sources.
“SkyEye is not merely equipment but is a part of a broader system that connects data, algorithms and operation, as well as production management. It is precisely this combination that enables the energy sector to operate more intelligently, with fewer losses, lower costs, and greater stability,” concludes Petr Možný.
Comments