Students from Pukyong National University have received top honors in a competition focused on electrical and chemical safety.
An undergraduate team from the Department of Safety Engineering identified gaps in the regulatory management of aging energy storage systems (ESS), while a graduate student developed an AI-based model for predicting acute toxicity of compounds.
The team, consisting of Ha Dong-geun, Park Ji-yoon, Oh Na-yeon, and Ji Ho-jun, was awarded first place in the policy proposal category of the 5th National Electric Safety Contest. They also received the Minister of Climate, Energy and Environment Award and a cash prize of 3 million won.
Hosted by the Ministry of Climate, Energy and Environment and organized by the Korea Electric Safety Corporation, the contest included a written preliminary evaluation, public voting, and an in-person final judging round.
The students highlighted that the same safety management standards apply to ESS regardless of their usage duration. They noted that the current Electric Safety Management Act lacks criteria that reflect the aging of equipment, making it difficult to proactively address performance degradation and safety risks associated with long-term operation of ESS.
Their proposal suggests introducing a 'recommended usage period' for ESS safety management and conducting 'detailed safety inspections' for aging equipment that meets certain criteria. This aims to enhance the management system to identify and address risks such as fires due to aging.
The team stated, "It seems we have been recognized for analyzing the issues in safety management systems using AI and data, and for the practicality of connecting this to improvements in current laws and regulations."
Graduate student Kang Ji-won won first place at the 1st AI Hazardous Materials Competition. The awards were presented recently at the 2026 Korean Society of Hazardous Materials Conference held at the Yeosu Expo Convention Center, with Professor Lee Chang-jun as his advisor.
The award-winning research focused on predicting acute toxicity values of aromatic compounds using a model based on the Group Contribution Method and Support Vector Regression.
The study utilized 55 functional groups and molecular weight-related input variables for aromatic compounds, including those with benzene rings. The method combines GCM, SVR, and PSO to predict acute toxicity values.
According to the Global Harmonization System (GHS), the accuracy of risk classification was found to be 90.6% for oral exposure, 100% for dermal exposure, and 80% for inhalation. This suggests the potential for rapidly predicting the acute toxicity classification of new substances that have not been tested for harmful effects without the need for animal or chemical experiments.
Kang Ji-won stated, "It seems that the integration of data-driven predictive models with chemical safety management was well-received."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.