New Algorithm Aimed at Achieving Stable Electronic Power Supply and Energy Conversion Goals
Students Eunkyung Kang, Seonuk Yang, and Haryeom Jang at the Department of Big Data Analytics of the Graduate School won the grand prize in the field of Big Data Analysis in the Public Data Utilization Business Idea (BI) Contest hosted by the Ministry of Trade, Industry and Energy. Won the grand prize
The three students took on the task of developing an optimized predictive model for collective resource composition to improve the forecast accuracy of power generation using renewable-energy. They designed and proposed an algorithm that configures and optimizes multiple power plants into a collective power generation unit, the output of which can be sold on the electricity trading market based on the guideline of the government's renewable energy power generation prediction system. This new algorithm is expected to achieve a settlement amount close to 80% of the maximum theoretical value. Professor Sung-Byung Yang at the Department Big Data Analytics, who advised the students, explained, “The ideas presented by our students demonstrated strong business feasibility that could easily be converted into commercial application, which seems to have appealed to the judging committee.”
“Past experience and training in artificial intelligence algorithm development course gave me extra motivation”
If electricity is overproduced or underproduced, it can compromise the stability of the entire electrical grid. Currently, the only way to address this problem is by controlling the amount of electricity being generated. The main form of control is fossil fuel power generation, which has a more direct correlation between input and output. Renewable-energy power generation, on the other hand, is difficult to form consistently accurate forecast on how much electricity will be produced in a given period of time with the required degree of precision, as it can be significantly influenced by the environmental factors beyond human control. This remains a major obstacle against transitioning to renewable-energy power generation, which requires incorporating into the electrical grid an increasingly greater portion of power generation output produced by renewable resources.
To achieve dependable power supply and meet energy conversion goals, accurate forecasting of renewable-energy power generation is essential. As the amount of electricity obtained from a single unit of renewable-energy power plant is relatively small, multiple renewable-energy power plants are grouped and managed together as a collective power resource. Students Eunkyung Kang, Seonuk Yang, and Haryeom Jang focused on this aspect and conducted research in the direction of minimizing prediction errors in the state of collective power resources.
They explored several approaches and eventually decided to investigate various clustering algorithms and expand the central points of the data where those clustering algorithms intersected. Student Seonuk Yang said, “In an artificial intelligence class, I developed an algorithm, and that experience inspired me to create my own algorithms instead of drawing from well-established libraries. I was motivated to craft my own unique solution.”
Professor Sung-Byung Yang: “The role of an advisor is to create an atmosphere where students can grow together.”
The students, who are members of SMART Lab, participated in the contest at the recommendation of Professor Yang. He said, “Recently, a growing number of students enrolling in the Graduate School and the Department of Big Data Analytics are seeking employment opportunities over pursuing further academic goals. For those students, practical experience such as winning contests or internship engagement can be important. We actively advocate and promote more student participation in contests and projects.”
Professor Yang also gathered eight student teams preparing for diverse contests and hosted a workshop. “I think the role of the advisor is to create an atmosphere where students compete and cooperate and grow together,” he said, explaining the purpose of the workshop. “I think synergy occurred through students sharing ideas and problem-solving tips.”
He concluded, “Winning public contests is completely different from writing research papers. It is likely to become a battle of real-world ideas for developing practical solutions with business potential. I advised the students to demonstrate teamwork by opening up all possibilities and discussing them through brainstorming, and they responded splendidly.”
※ For further insights
SMART Lab Website
- University Communication & Press