Application of machine learning in assessment of combustion of liquified natural gas.

This work focuses on the implementation on carbon capture on ships which run on liquified natural gas (LNG). LNG ships present a real-world example of LNG as well as a study case for carbon capture on LNG combustion. There is also special interest for that as well, since the International Maritime Organization (IMO), imposed a limit of 0.5% wt. of sulphur content in ship fuel has been imposed from 2020 to reduce pollution emissions from global shipping activities. This initiative will lead to major changes since the previous limit was set at 3.5% wt., which broadened fuel options for ships. In addition, the IMO is developing a long-term plan to completely nullify shipping’s impact on CO₂ emissions by 2030. Consequently, stricter regulations will be imposed to marine activities worldwide. LNG fuel seems to be a promising solution. The sulphur emissions are lower, in compliance with the latest IMO regulations. Additionally, it has a greater energy density in comparison to traditional fuels, like heavy fuel oil (HFO). This paper aims to study the feasibility of a project, which equips an LNG fuelled ship with a carbon capture system. The study includes an examination of an on-board carbon capture system, by simulating the LNG engine as well as the carbon capture system in simulation software. The engine model chosen is the Wärtsilä 6L34DF. The results of these simulations are analysed to examine the correlation between the system’s variables and to evaluate the possibility of heat integration within the system. The economic feasibility of the project is then assessed, using economic data. The results show that heat integration is possible. For example, the heat provided from the flue gas is calculated at 1.323MW when the reboiler duty is 0.3353 MW. However, the project is not sustainable under current market conditions.