Chongqing - A team from Yongchuan Vocational Education Center in Chongqing has won an Excellence Award at World Vocational College Skills Competition 2025 (WVCSC)  with a project that leverages the DeepSeek large language model to optimize photovoltaic (PV) power generation. The achievement marks the only team from Chongqing to advance to and win an award at the championship finals.

A team from Yongchuan Vocational Education Center in Chongqing has won an Excellence Award at the World Vocational College Skills Competition 2025 (WVCSC). (Photo provided by the interviewee) 

The competition's energy and power track covers cutting-edge fields such as wind and solar power, hydrogen energy, energy storage, and smart grids. It requires interdisciplinary skills for roles in energy production, equipment manufacturing, and integrated energy management.

The winning project, titled "Smart Grid Control: Construction and Maintenance of an Intelligent Decision-Making Photovoltaic System," uses DeepSeek to predict solar power output, enabling enterprises to achieve "source-grid-load-storage" integration through precise time-based scheduling. For example, during low-sunlight mornings, the system directs high-load workshops to draw power from the grid. At solar peak hours, it powers entire workshops with PV energy while charging storage systems. At night, stored battery energy supports peak grid demand, improving efficiency and reducing emissions.

The students take measurements. (Photo provided by the interviewee) 

The idea emerged during the team's visit to a local enterprise, where they learned about challenges with manual inspection inefficiency and unstable PV generation, leading to low utilization rates. This inspired them to combine drone inspections with DeepSeek's predictive capabilities to enable automatic power-mode switching.

Despite having less than two months to prepare, the four-member team - Pu Yang, Li Shimin, Wang Junhao, and Zou Xi - refined their collaboration through intensive training. With guidance from instructors Zeng Jingbo and Tan Qiansheng, they divided responsibilities based on expertise: Pu oversaw project coordination, Zou focused on installation, Wang specialized in PLC programming, and Li developed the digital twin and AI prediction system.

The student and the teacher work together. (Photo provided by the interviewee) 

The team pushed themselves to complete the required one-hour task in just 55 minutes, leaving time for contingencies. They recorded and analyzed every training session, fine-tuning details to stabilize their performance at around 54 minutes.

During the finals, the students simulated installation, fault diagnosis, and AI-driven energy scheduling for a rooftop PV station. When a sudden voltage anomaly occurred during an integration test, they swiftly executed an emergency plan: while two members checked the code, the others traced the physical connections. They quickly identified a loose contact and repaired it using one of their school's patented tools - a constant-temperature soldering station - earning high praise from the judges.

Group photo of the team members. (Photo provided by the interviewee) 

"This competition has not only sharpened the students' vocational skills but also provided invaluable hands-on teaching experience," said instructor Tan Qiansheng. He is now compiling technical documentation from the competition to create new training modules, bringing cutting-edge technology back into the classroom.

physical model of the Yongchuan Vocational Education Center. (Photo provided by the interviewee) 

Kong Lingyong, principal of Yongchuan Vocational Education Center, emphasized that the school will deepen industry-education integration and school-enterprise cooperation, transforming competition experience into teaching resources. "Our goal is to cultivate more high-quality technical talent that is practical, skilled, and adaptable," he said.