Chinese scientists say their AI model can mark the best spots for double-faced solar panels
- Research team devises tool that uses data from satellites and weather stations to optimise placement of photovoltaic modules
- Tibetan Plateau and deserts of Xinjiang are best spots in China to install double-sided PV panels to maximise energy output, team finds
01:36
China’s largest photothermal power facility drives development of new form of energy
Under the right sunlight conditions, two-sided solar panels can produce more power than conventional ones. However they are difficult to transport and maintain, so finding the optimal location for them is crucial to ensuring the best use of resources.
However, it lacks data that would help determine the best spots to place two-sided solar panels.
The country has just 17 radiation stations that collect data on the amount and type of “solar power” available in a given location. This includes data on direct radiation – beamed straight from the sun to the front surface of a solar panel – and diffuse radiation, which is scattered by the atmosphere and more likely to be picked up by the panel’s rear face.
The equipment at these stations requires annual adjustments and regular maintenance, resulting in considerable operating costs.
To overcome the lack of on-the-ground data, researchers from Tsinghua University in Beijing and the National Tibetan Plateau Data Centre created an AI model based on sunshine data from 2,500 weather stations across China.
The AI was trained on solar radiation data – collected either via ground observation or satellite remote sensing – and surface meteorological data, to predict the amount of direct and indirect radiation at any given spot.
“In principle, this model can be applied on the global scale without additional training with local data,” the team wrote.
China needs 10 times its solar and wind power to be carbon neutral, study finds
Senior author Yang Kun, a professor with the earth system science department at Tsinghua, said the lack of comprehensive radiation data meant there was little information to help the authorities and the solar industry to plan for panel installation sites.
“Now the output of the AI model supported by satellite data can inform decision-making on where and which type of panels to deploy to take full advantage of solar power,” Yang said.
Solar power accounted for close to 5 per cent of China’s electricity output in 2022.
Yang said the AI system also revealed the solar potential of remote areas of China that lack power line infrastructure, and this could inspire future research and policy planning.
China creates electric power in its western deserts equal to half that in US
First author Shao Changkun, a PhD candidate at Tsinghua, said the area surrounding the Taklamakan Desert in the southwestern part of Xinjiang autonomous region, and the eastern Tibetan Plateau were ideal spots for dual-sided panels.
“Direct solar radiation is high on the high-altitude plateau where the air is thinner, while diffuse solar radiation is also substantial thanks to its complex landscape and high cloud coverage,” Shao said.
“Both sides of solar panels will receive sizeable amounts of radiation in these regions.”
The team compared their estimates with radiation data from around the world and found that their AI model had high accuracy, Shao said, adding that combining its inputs with meteorological data from other countries could help the system be used for solar radiation projections globally.
Yang added that the data could also be applied in other fields such as agriculture, as plants have been found to carry out photosynthesis more efficiently under diffuse light conditions.
