Korea’s Hanwha Solutions speeds up AI technology development

Hanwha Qcells factory in Georgia, U.S. [Courtesy of Hanwha Solutions]

Hanwha Solutions Corp., a leading energy material and solution provider in South Korea, is accelerating efforts to develop artificial intelligence (AI) technology such as by establishing its own hyper-scale AI.

According to multiple industry sources on Wednesday, Hanwha Solutions is engaged in the development of hyper-scale AI for utilization in the development of various new materials.

Hyper-scale AI is the next-generation AI that can learn large quantities of data on its own and is capable of comprehensive and autonomous thinking, learning, and judgement like humans.

Hanwha Solutions was the first petrochemical company to jump into the development of hyper-scale AI in September 2021 amid fierce competition among information technology (IT) companies.

Hanwha Solutions is in the stage of demonstrating and updating its hyper-scale AI for application in actual research.

Hanwha Solutions has collected more than 100 million chemical formulas from academic databases to train its super-sized AI. This step was needed to make the AI understand chemical formulas through chemical data.

The AI, learning from experimental data, can quickly predict the chemical properties of new material candidates, significantly reducing the time required for commercialization.

It is also expected to facilitate the development of new catalysts by improving the accuracy of predicting the outcome of combining and organizing materials when developing new materials by up to seven times.

The Qcells division of Hanwha Solutions plans to increase solar cell productivity by introducing AI to its solar hub being built in Georgia, U.S.

It already has a smart factory in Jincheon, North Chungcheong Province that is powered by AI and the Internet of Things (IoT). The factory produces one solar cell every 1.8 seconds, and its quality is controlled by Tra.Q, a big data-based data analysis system, with a defect rate of 0.5 percent.

The Qcells division also applies AI to operate its energy storage systems in Europe and Australia. Pre-impact detection algorithms have been developed by collecting and analyzing ESS operation data to monitor more than 1,400 ESS products in real time.