Researchers in Korea succeeded in developing the world's first quantitative determination method for copper wire short circuit traces, which is the key to identifying the cause of the damage in the event of a fire, using microstructure data and machine learning. The research team is led by Dr. Joo-hee Kang in the Department of Materials Analysis at the Korea Institute of Materials Science (KIMS), a government-funded research institute under the Ministry of Science and ICT, and its joint research team is in the Electrical Disaster Research Center (Director Seon-bae Bang) of the Korea Electrotechnology Research Institute under the Korea Electrical Safety Corporation.

Copper wire molten traces found at electrical fire sites are different depending on the cause of the fire but there is a limit to identifying the cause of the fire in a complex and diverse fire environment with only the short-circuit mark identification method, which relies on the experience and intuition of fire investigators. The copper wire short circuit quantitative determination method developed by the joint research team uses digital microstructure and machine learning based on an electron microscope, not an intuitive determination method that has been judged only through the naked eye of fire investigators and an optical microscope.

Using the electron backscatter diffraction (EBSD) technique to extract digital microstructure factors, the research team presented a new determination method through data analysis and machine learning. The researchers actively utilized materials information science, such as proposing a quantitative discriminant using the microstructure data of the primary and secondary short circuit trace obtained through laboratory and fire experiments, and verifying it from various angles through machine learning.

By proposing the new method using digital microstructure for short circuit traces of unknown causes and verifying it through practical data of fire, the researchers reduced the unidentified short circuit rate and presented a new forensic investigation method that can be applied immediately.

A total of 9,474 electrical fires occurred in Korea as of 2021, and 2,415 cases due to unconfirmed short circuits account for 25% which can be highly contentious. As the method can identify the causes of short-circuit fires which account for the largest proportion of electrical fires, the research is expected to contribute greatly to the development of fire prevention and forensic investigation techniques.

Dr. Joo-Hee Kang, the principal investigator at KIMS, said, “By applying the newly developed short-circuit detection technology to the fire scene, the rate of unidentified short-circuit fires can be significantly lowered and the cause of the fire can be clearly identified. The technology is expected to minimize disputes related to the cause of fires in the future.”

The research was supported by the “Material Research Data Platform Project” of the Ministry of Science and ICT and the research support project of the Korea Electrotechnology Research Institute under the Korea Electrical Safety Corporation. Also, it was published in Materials on November 18th. The research team is currently conducting follow-up research with microstructure data of short circuit traces under various fire conditions.