Leverage machine learning to help predict s

Ships are a major mode of commercial transportation, contributing to 80% of global commodity and energy trade. However, they are emitted as exhaust gases – from the engines while sailing, and from the engines and the boiler when berthed in ports. These emissions negatively affect not only human health, but also the environment. Therefore, the International Maritime Organization has imposed regulations on the type of fuel used in ships. While efforts are being made to reduce the level of emissions from ships, a completely environmentally friendly fuel has yet to be developed. In the meantime, assessing and predicting the level of exhaust emissions from ships is crucial.

Given this background, a group of researchers from Korea National Maritime and Ocean University (NKMOU) led by Dr. Won-Joo Lee, associate professor in the institute’s Department of Marine Systems Engineering, measured the emissions of a continuously operating oil-fired boiler in a training ship under various air-to-air ratios. fuel. CO exhaust data₂No_xand then₂ The gases were collected for 18 states and used to predict emissions through data-driven modeling,” Dr. Lee explains.

Their work was made available online on September 18, 2022, and Published in Volume 375 of Cleaner Production Journal On November 15, 2022.

The researchers used unsupervised learning to compress the original data to generate three new datasets. They combined them to create the group data set. The performance of these five datasets was assessed – in terms of carbon dioxide₂No_xand then₂ Forecasts – using four basic models. Vector Machine based models with the original and combination datasets produced the best results.

Next, the researchers combined the core models to develop four core group models. These models, in turn, were used to build the dual group models. As expected, the dual ensemble models provided the most accurate predictions of emissions for all three gases.

Finally, the researchers applied the developed models to a new data set, validating the results and demonstrating the superiority and generalizability of the models.

How can this work help the shipping industry reduce carbon emissions? Dr. Lee discusses the future implications of their work. The results of this study can be used to predict exhaust gas emissions and will be applied to marine boilers soon. It will enable marine engineers to take the necessary measures to reduce emissions and reduce air pollution in port areas. As it is not economically feasible to install expensive equipment such as gas analyzers in boilers for shipping companies, the proposed technology would prove indispensable. Moreover, the ensemble data generation and dual ensemble model techniques can enhance the performance of various other machine learning applications. “

Here’s to achieving carbon neutrality, powered by technology and AI models!

***

reference

Authors: Min-Ho Park^1,2Jae Jung Hoor³Won Joo Lee^2,3

Title of original paper: Prediction of emissions from oil-fired boilers using combination methods considering variable combustion air conditions

magazine: Cleaner Production Journal

DOI: https://doi.org/10.1016/j.jclepro.2022.134094

Affiliations:

¹Department of Marine Engineering Korea Maritime and Ocean UniversityThe Republic of Korea

²Interdisciplinary major in Maritime Convergence and Artificial Intelligence, Korea Maritime and Ocean University, Republic of Korea

³Department of Marine Systems Engineering, Korea Maritime and Ocean University, Republic of Korea

About Korea National Maritime and Ocean University

South Korea’s most prestigious university for marine studies, transportation science, and engineering, Korea National Maritime and Ocean University is located on an island in Busan. The university was established in 1945 and has since merged with other universities to become currently the only post-secondary institution specializing in marine science and engineering. It has four faculties offering undergraduate and postgraduate courses.

website: http://www.kmou.ac.kr/english/main.do

About Dr. Won Joo Lee

Dr. Won-Joo Lee is Associate Professor in the Department of Marine Systems Engineering at Korea Maritime and Ocean University, South Korea. His research focuses on diesel engines, marine environment, combustion, recycling and marine engine CBM emissions. He obtained his Ph.D. He received his PhD in Mechanical Engineering from Korea Maritime and Ocean University in 2017. Prior to obtaining his Ph.D., he worked as a chief engineer for a control vessel and as a gas engineer for LNG carriers.