State Grid Zhejiang Electric Power Co Ltd recently launched a prediction center leveraging artificial intelligence to precisely forecast electricity use.

The move came as an increasing amount of intermittent renewable power is being connected to the grid. As renewables such as coals are hard to manage, the former's connection to the grid would cause instability issues.

"Traditional power prediction still faces challenges of insufficient accuracy and limited flexibility when dealing with variables such as cyclical changes in power load, weather factors, and user behaviors during holidays," said Song Haomin, project leader of the AI-backed power use prediction center.

Song said the center deployed AI prediction on high-performance servers. By introducing deep learning and no-code machine learning, and analyzing historical power load and meteorological data, it effectively enhances the power prediction capability under complex and variable conditions.